Effects of glycosylation on the folding and stability of human, recombinant and cleaved alpha 1-antitrypsin

Glycosylation of human alpha 1-antitrypsin in Saccharomyces cerevisiae and methylotrophic yeasts

Human alpha 1-antitrypsin (alpha 1-AT) is a major serine protease inhibitor in plasma, secreted as a glycoprotein with a complex type of carbohydrate at three asparagine residues. To study glycosylation of heterologous proteins in yeast, we investigated the glycosylation pattern of the human alpha 1-AT secreted in the baker's yeast Saccharomyces cerevisiae and in the methylotrophic yeasts, Hansenula polymorpha and Pichia pastoris. The partial digestion of the recombinant alpha 1-AT with endoglycosidase H and the expression in the mnn9 deletion mutant of S. cerevisiae showed that the recombinant alpha 1-AT secreted in S. cerevisiae was heterogeneous, consisting of molecules containing core carbohydrates on either two or all three asparagine residues. Besides the core carbohydrates, variable numbers of mannose outer chains were also added to some of the secreted alpha 1-AT. The human alpha 1-AT secreted in both methylotrophic yeasts was also heterogeneous and hypermannosylated as observed in S. cerevisiae, although the overall length of mannose outer chains of alpha 1-AT in the methylotrophic yeasts appeared to be relatively shorter than those of alpha 1-AT in S. cerevisiae. The alpha 1-AT secreted from both methylotrophic yeasts retained its biological activity as an elastase inhibitor comparable to that of alpha 1-AT from S. cerevisiae, suggesting that the different glycosylation profile does not affect the in vitro activity of the protein.

Characterization of a human alpha1-antitrypsin variant that is as stable as ovalbumin

The metastability of inhibitory serpins (serine proteinase inhibitors) is thought to play a key role in the facile conformational switch and the insertion of the reactive center loop into the central beta-sheet, A-sheet, during the formation of a stable complex between a serpin and its target proteinase. We have examined the folding and inhibitory activity of a very stable variant of human alpha1-antitrypsin, a prototype inhibitory serpin. A combination of seven stabilizing single amino acid substitutions of alpha1-antitrypsin, designated Multi-7, increased the midpoint of the unfolding transition to almost that of ovalbumin, a non-inhibitory but more stable serpin. Compared with the wild-type alpha1-antitrypsin, Multi-7 retarded the opening of A-sheet significantly, as revealed by the retarded unfolding and latency conversion of the native state. Surprisingly, Multi-7 alpha1-antitrypsin could form a stable complex with a target elastase with the same kinetic parameters and the stoichiometry of inhibition as the wild type, indicating that enhanced A-sheet closure conferred by Multi-7 does not affect the complex formation. It may be that the stability increase of Multi-7 alpha1-antitrypsin is not sufficient to influence the rate of loop insertion during the complex formation.

Glycosylation and thermodynamic versus kinetic stability of horseradish peroxidase

The influence of N-linked glycans on the stability of glycoproteins has been studied using horseradish peroxidase isoenzyme C (HRP), which contains eight asparagine-linked glycans. HRP was deglycosylated (d-HRP) with trifluoromethanesulfonic acid and purified to an enzymatically active homogeneous protein containing (GlcNAc)2 glycans. The thermal stability of HRP and d-HRP at pH 6.0, measured by residual activity, was indistinguishable and showed transition midpoints at 57 degrees C, whereas the unfolding in guanidinium chloride at pH 7.0, 23 degrees C was 2-3-fold faster for d-HRP than for HRP. The results are compatible with a glycan-induced decrease in the dynamic fluctuation of the polypeptide chain.

Effect of glycosylation on the stability of alpha1-antitrypsin toward urea denaturation and thermal deactivation

The effects of glycosylation on the stability of human alpha1-antitrypsin were investigated. The transition midpoints in urea-induced equilibrium unfolding of a non-glycosylated recombinant, a yeast version of glycosylated, and human plasma alpha1-antitrypsin were 1.8 M, 2.2 M, and 2.5 M at 25 degrees C, respectively. Kinetic analyses of unfolding and refolding revealed that glycosylation retarded the unfolding without affecting the refolding rate significantly, suggesting that the stability increase is due to the stabilization of the native state as opposed to the destabilization of the unfolded state. In thermal deactivation, which is a heat-induced aggregation process, the unglycosylated recombinant alpha1-antitrypsin was deactivated most easily, which was followed in order by the yeast, and the plasma form. The results indicate that glycosylation confers the increase in stability of alpha1-antitrypsin, and that the oligomannose sugars present on the yeast form produce a less stable molecule than the complex type sugars on the plasma form. It appears that the effect of glycosylation on the enhancement of thermal resistance is exerted through the increase in conformational stability. However, a stable recombinant variant (Phe 51 --> Cys) that showed the same conformational stability as the plasma form was less resistant to thermal denaturation than the plasma alpha1-antitrypsin. The results suggest that the existence of carbohydrate moiety per se as well as the conformational stability contribute to the kinetic stability of alpha1-antitrypsin toward aggregation.

Reactivation of C1-inhibitor polymers by denaturation and gel-filtration chromatography

C1-inhibitor is a proteinase inhibitor in the serpin family. It is an important inhibitor of complement C1, plasma kallikrein, and factor XIIa, and as such is involved in regulating inflammatory pathways. Studies on the plasma-derived protein are hampered by the relative ease with which the protein converts to an inactive state on storage, under mild denaturing conditions, or by incubating in some unfavorable buffers. This inactivation is caused by formation of soluble polymers which can be visualized on native electrophoresis. In order to facilitate studies on both the plasma-derived protein and recombinant variants planned for the future, it was necessary to devise a method for the rapid reactivation of the polymers in high yield. It was found that nonionic detergents did not dissociate the polymers, but they were readily dissociated in 0.1% SDS. Treatment with 0.1% SDS followed by rapid removal of the SDS and refolding on an FPLC Superose 6 column allowed for recovery of about 15% of the protein in the active monomeric form. Eighty-five percent eluted as a range of higher order polymers. Using 8 M urea as the denaturant a 25% yield of active monomer was recovered. However, with 6 M guanidine hydrochloride as the denaturant, the yield of active monomer was almost 50%. The remaining material was not present as a range of polymeric species but was probably a dimer. Therefore this method is a useful technique to facilitate studies on C1-inhibitor. Moreover, the ability to produce monomer, dimer, and polymer forms of C1-inhibitor is useful for studies investigating the conformational changes which have occurred in the different forms.

Effect of individual carbohydrate chains of recombinant antithrombin on heparin affinity and on the generation of glycoforms differing in heparin affinity

Two major glycoforms of recombinant antithrombin which differ 10-fold in their affinity for the effector glycosaminoglycan, heparin, were previously shown to be expressed in BHK or CHO mammalian cell lines (I. Björk, et al., 1992, Biochem. J. 286, 793-800; B. Fan et al., 1993, J. Biol. Chem. 268, 17588-17596). To determine the source of the glycosylation heterogeneity responsible for these different heparin-affinity forms, each of the four Asn residue sites of glycosylation, residues 96, 135, 155, and 192, was mutated to Gln to block glycosylation at these sites. Heparin-agarose chromatography of the four antithrombin variants revealed that Gln 96, Gln 135, and Gln 192 variants still displayed the two functional heparin-affinity forms previously observed with the wild-type inhibitor, whereas the Gln 155 variant showed only a single functional high heparin affinity form. These results demonstrate that heterogeneous glycosylation of Asn 155 of recombinant antithrombin is responsible for generating the low heparin affinity glycoform. Analysis of heparin binding to the higher heparin affinity forms of the four variants showed that all exhibited increased heparin affinities of two- to sevenfold compared to wild-type higher heparin affinity form or to plasma antithrombin, with the Gln 135 variant showing the largest effect on this affinity. The extent of heparin-affinity enhancement was correlated with the distance of the mutated glycosylation site to the putative heparin-binding site in the X-ray structure of antithrombin. All variants displayed normal kinetics of thrombin inhibition in the absence and presence of saturating heparin, indicating that the carbohydrate chains solely affected heparin binding and not heparin-activation or proteinase-binding functions. These results indicate that all carbohydrate chains of recombinant antithrombin adversely affect heparin-binding affinity to an extent that correlates with their relative proximity to the putative heparin-binding site in antithrombin.

Structural transitions of human serum albumin: an investigation using electrophoretic techniques

The influence of pH and urea concentration on the electrophoretic mobility of native and reduced human serum albumin was evaluated by zonal electrophoresis across transverse urea gradients as well as by migration across transverse pH gradients in gels containing varying concentrations of urea. Exposure to urea results in a change of both pI and hydrodynamic volume of the albumin molecule. At acidic pH, the former effect is brought about by lower urea concentrations than the latter, as made evident by a biphasic denaturation curve; in alkaline buffers, all structural transitions occur at once, and a typical sigmoidal curve is observed. Below pH 6, the lower the pH, the lower the urea concentration causing albumin denaturation. For instance, in the presence of 3 M urea, below pH 5 > 95% of the protein is present in its denatured state, above pH 8 > 90% is in its native form, whereas in the 6.5-7.5 pH range the two components have similar abundance. Also, the reversibility of the transition between folding and unfolding depends upon pH, and is complete only above pH 6. After inclusion of beta-mercaptoethanol in the albumin sample the urea concentration required to bring about protein unfolding increases between pH 4 and 6 and decreases thereafter.

Kinetically controlled folding of the serpin plasminogen activator inhibitor 1

The serpin plasminogen activator inhibitor 1 (PAI-1) folds into an active structure and then converts slowly to a more stable, but low-activity, "latent" conformation [Hekman, C. M., & Loskutoff, D. J. (1985) J. Biol. Chem. 260, 11581-11587]. Thus, the folding of PAI-1 is apparently under kinetic control. We have determined the urea denaturation and refolding transitions of both latent and active PAI-1 proteins by using intrinsic tryptophan fluorescence. While folding of active PAI-1 is reversible, the denaturation and refolding of latent PAI-1 are not. Instead, denatured latent PAI-1 refolds in lower concentrations of urea to give the active protein. Thus, the high-stability latent conformation is kinetically inaccessible over a range of urea concentrations. Complete denaturation of latent PAI-1 occurs at 5.5 M urea [delta G(H2O) approximately 21 kcal] whereas active PAI-1 denatures in only 3.8 M urea [delta G(H2O) approximately 12 kcal]. The fluorescence emission profile, as a function of urea of both the active and latent forms of the protein, reveals intermediates with partial structure. Circular dichroism measurements and limited protease digestion with Lys-C suggest that the intermediate in the denaturation of latent PAI-1 retains most of the secondary structure of the fully folded protein, whereas the intermediate in the denaturation of active PAI-1 exhibits significant loss of secondary structure. The Lys-C digestion patterns show that the active protein is more susceptible to proteolysis near sheet A than is the latent form. The studies suggest a model for the kinetically controlled folding pathway of PAI-1.

Purification and characterization of alpha 1-antitrypsin secreted by recombinant yeast Saccharomyces diastaticus

The secreted human alpha 1-antitrypsin (alpha 1AT) produced by yeast was purified from the culture medium by ultrafiltration, ammonium sulfate fractionation (60-75% saturation), protamine sulfate treatment, and ion-exchange chromatography. Molecular mass of the purified alpha 1AT was 52 kDa, which is similar to that of human plasma alpha 1AT. Yeast-produced alpha 1AT was fully functional as an inhibitor compared with the plasma form. Unlike plasma alpha 1AT, however, treatment of the yeast-produced alpha 1AT with endoglycosidase H decreased the molecular mass to that of recombinant alpha 1AT produced in Escherichia coli, indicating the high-mannose type N-linked glycosylation of the secreted alpha 1AT. Glycosylation in yeast cells enhanced kinetic stability of alpha 1AT towards heat deactivation.

Glycosylation of alpha-1-proteinase inhibitor and haptoglobin in ovarian cancer: evidence for two different mechanisms

The change in glycosylation of the two acute-phase proteins, alpha-1-proteinase inhibitor (API) and haptoglobin (Hp), in progressive ovarian cancer is different. This has been shown by monosaccharide analysis and lectin-binding studies of proteins purified from serum. In the glycan chains of API, there is decreased branching (more biantennary chains), less branches ending in alpha 2-3 sialic acid, more branches ending in alpha 2-6 sialic acid and more fucose, probably linked alpha 1-6 to the core region. On the other hand, Hp shows increased branching (more triantennary chains), more branches ending in alpha 2-3 sialic acid, less branches ending in alpha 2-6 sialic acid, and more fucose, probably in the alpha 1-3 linkage at the end of the chains. This is surprising because API and Hp are thought to be glycosylated by a common pathway in the liver. We have also shown that the fucose-specific lectin, lotus tetragonolobus, extracts abnormal forms of both Hp and API in ovarian cancer, but the expression of this Hp is related to tumour burden and the expression of this API is related to lack of response to therapy. It is suggested that this difference in the behaviour of API and Hp in ovarian cancer may be associated with the different changes in their glycosylation. Of the many mechanisms that could explain these findings, a likely one is that a pathological process is removing API with triantennary chains from the circulation.(ABSTRACT TRUNCATED AT 250 WORDS)

Structure and function of corticosteroid-binding globulin: role of carbohydrates

To study the site-specificity of human corticosteroid-binding globulin (CBG) glycosylation and the functional significance of individual carbohydrate chains in its molecule, a panel of recombinant CBG mutants containing each of the six potential glycosylation sites alone and in various combinations has been expressed in Chinese hamster ovary (CHO) cells. Analyses of these mutant glycoproteins showed that three of the glycosylation sites are only partially utilized, and this may contribute to the production of glycoforms with distinct physiological functions. Processing of individual carbohydrate chains (branching and fucosylation) is site-specific and may, thus, account for the formation of structural determinants essential for the recognition of CBG by cell membranes. Glycosylation at the only phylogenetically conserved consensus site, Asn238-Gly239-Thr240, is essential for the biosynthesis of CBG with steroid-binding activity. Evidence has been obtained to support the hypothesis that transient carbohydrate-polypeptide interactions between Trp266 and the maturing carbohydrate chain at Asn238 occur during early stages of the CBG biosynthesis which affect protein folding and formation of the steroid-binding site. Another tryptophan residue, Trp371, has been found to be critical for CBG-steroid interactions and is likely located in the steroid-binding site.

A thermostable mutation located at the hydrophobic core of alpha 1-antitrypsin suppresses the folding defect of the Z-type variant

A thermostable mutation, F51L, at the hydrophobic core of human alpha 1-antitrypsin (alpha 1AT) increased the conformational stability of the molecule by decreasing the unfolding rate significantly without altering the refolding rate. The mutation specifically influenced the transition between the native state and a compact intermediate, which retained approximately 70% of the far-UV CD signal, but which had most of the fluorescence signal already dequenched. The mutant alpha 1AT protein was more resistant than the wild-type protein to the insertion of the tetradecapeptide mimicking the sequence of the reactive center loop, indicating that the mutation increases the closing of the central beta-sheet, the A-sheet, in the native state. The F51L mutation enhanced the folding efficiency of the Z-type (E342K) genetic variation, which causes aggregation of the molecule in the liver. It has been shown previously that the aggregation of the Z protein occurs via loop-sheet polymerization, in which the reactive center loop of one molecule is inserted into the opening of the A-sheet of another molecule. Our results strongly suggest that the hydrophobic core of alpha 1AT regulates the opening-closing of the A-sheet and that certain genetic variations that cause opening of the A-sheet can be corrected by inserting an additional stable mutation into the hydrophobic core.

Refolding of alpha 1-antitrypsin expressed as inclusion bodies in Escherichia coli: characterization of aggregation

Recombinant alpha 1-antitrypsin (alpha 1AT) produced as inclusion bodies in Escherichia coli was purified via several steps including solubilization of the inclusion bodies in 8 M urea and refolding by direct dilution of denaturant, followed by ion-exchange chromatography. The purified recombinant alpha 1AT has an activity comparable to human plasma alpha 1AT. During refolding, prolonged incubation of the alpha 1AT polypeptides at intermediate urea concentration favored production of inactive but soluble aggregates, which could regain activity after denaturation and renaturation. Nondenaturing polyacrylamide gel electrophoresis of the aggregates revealed the existence of dimers and higher oligomers. Immunological approach to characterize conformation showed that the oligomers were distinct from the native, the cleaved, or the denatured form, but was similar to the polymers induced from the native structure in mild denaturing condition. These results suggest that the oligomers are formed through specific interactions between aggregation-competent species which are stabilized at intermediate denaturant concentration.

Preparation and characterization of latent alpha 1-antitrypsin

Members of the serine proteinase inhibitor or serpin superfamily have a common molecular architecture based on a dominant five-membered A beta-pleated sheet and a mobile reactive center loop. The reactive center loop has been shown to adopt a range of conformations from the three turn alpha-helix of ovalbumin to the cleaved or latent inhibitor in which the reactive center loop is fully inserted into the A sheet of the molecule. While the cleaved state can be achieved in all inhibitory serpins only plasminogen activator inhibitor-1 and, more recently, antithrombin have been shown to adopt the latent conformation. We show here that the archetypal serpin, alpha 1-antitrypsin, can also be induced to adopt the latent conformation by heating at high temperatures in 0.7 M citrate for 12 h. The resulting species elutes at a lower sodium chloride concentration on an anion-exchange column and has a more cathodal electrophoretic mobility on non-denaturing polyacrylamide gel electrophoresis and isoelectric focusing than native M antitrypsin. Latent antitrypsin is inactive as an inhibitor of bovine alpha-chymotrypsin, is stable to unfolding with 8 M urea, and is more resistant to heat-induced loop-sheet polymerization than native but less resistant than cleaved antitrypsin. The reactive center loop of latent antitrypsin is inaccessible to proteolytic cleavage, and its occupancy of the A sheet prevents the molecule accepting an exogenous reactive center loop peptide. The activity of latent antitrypsin may be increased from < 1% to approximately 35% by refolding from 6 M guanidinium chloride.

Zone mapping of the binding domain of the rat low affinity nerve growth factor receptor by the introduction of novel N-glycosylation sites

The binding of NGF (nerve growth factor) to the rat low affinity nerve growth factor receptor (p75NGFR) has been studied by site-directed mutagenesis of the receptor. Introduction of non-native N-glycosylation sites within the binding domain indicates that the second of the characteristic cysteine-rich repeats may be particularly important to NGF binding. Two mutants of the second repeat, S42N and S66N, are glycosylated and bind NGF at a drastically reduced level, while still maintaining a conformation recognized by the monoclonal antibody against p75, MC192. Alanine substitution at these sites does not affect NGF binding. Two other mutations that result in local structural changes in the second repeat also greatly decrease binding. One of these altered residues, Ser50, appears to play an essential structural role, since it cannot be replaced by Asn, Ala, or Thr without loss of both NGF binding and MC192 recognition on a Western. Glycosylation of selected sites in the other repeats has little effect on NGF binding or antibody recognition. The introduction of non-native N-glycosylation sites may provide a generally useful scanning technique for the study of protein-protein interactions.

Molecular modelling of glycoproteins by homology with non-glycosylated protein domains, computer simulated glycosylation and molecular dynamics

OBJECTIVES

This study aims to visualise glycoproteins by computer graphics molecular modelling in order to research the dynamics of the oligosaccharide chains, determine their affects on protein conformation, antigenicity and function and to characterise oligosaccharide recognition determinants. With respect to the last, the modelling included the sialylpolylactosamine of thymocyte Thy-l and the sialyl Le(x)/Le(a) determinant present on brain Thy-l.

METHODS

The following techniques were used: 1) database searching for homologies with non-glycosylated protein domains; 2) protein modelling on the basis of homology and secondary structure prediction techniques; 3) oligosaccharide construction using a simulated annealing approach utilising the AMBER forcefield with appropriate parameters in the Biosyn software environment; 4) creation of glycoprotein conjugates for further investigation by energy minimisation and molecular dynamics.

RESULTS

This approach was successful in providing models of Thy-l and the carboxy terminus 27.5 kD of HIV-1 gp 120, by homology with immunoglobulin light chain folds and in one case (Thy-l) adding the oligosaccharide chains, phosphatidylinositol glycan anchor and lipid membrane and, in the other, adding additional highly glycosylated domains, and domains which folded by molecular dynamics. Significant affects on protein conformation were shown in the presence or absence of the lipid anchor and simulated membrane, but not by the N-linked oligosaccharide chains.

CONCLUSIONS

The highly glycosylated molecules Thy-l and gp120, which are not expected to crystallise in their native state, were modelled by computer graphics simulated annealing or molecular dynamics from which interactions could be predicted which agree with experimental data on antibody binding and in vitro activity.

Glycosylation of human corticosteroid-binding globulin. Differential processing and significance of carbohydrate chains at individual sites

Human corticosteroid-binding globulin (CBG) comprises 383 amino acids and six consensus sites for attachment of N-acetyllactosamine-type oligosaccharides. To study the extent of addition and processing of individual carbohydrate chains, we expressed CBG mutants, each containing only one of the six possible glycosylation sites, in Chinese hamster ovary cells and examined their electrophoretic, immunochemical, and lectin-binding properties. This indicated that Asn9, Asn308, and Asn347 are partially glycosylated and that oligosaccharides attached to Asn9, Asn238, Asn308, and Asn347 are predominantly biantennary, while more branched (most likely, triantennary) oligosaccharides are preferentially linked to Asn74 and Asn154. Only one of the biantennary chains (attached to Asn9) contains significant amounts of fucose. These data indicate that oligosaccharide processing is site-specific, and analyses of three other mutants, in which an additional glycosylation site was preserved, demonstrated that the processing of individual oligosaccharides occurs independently. Thus, the glycosylation of recombinant CBG appears to resemble that of natural human CBG. As we have previously found, glycosylation at Asn238 is essential for the production of CBG with steroid-binding activity, but when the mutant containing only one oligosaccharide at this position was enzymatically deglycosylated, its steroid-binding activity was unaltered. This suggests that interaction between this carbohydrate chain and the polypeptide is necessary for the folding and creation of the steroid-binding site only during CBG biosynthesis.

Expression of alpha 1-proteinase inhibitor in Escherichia coli: effects of single amino acid substitutions in the active site loop on aggregate formation

Overproduction of eukaryotic proteins in microorganisms often leads to the formation of insoluble protein aggregates which accumulate as intracellular inclusion bodies. alpha 1-Proteinase inhibitor (alpha 1-PI) when produced as a cytoplasmic protein in Escherichia coli (E. coli) forms inclusion bodies containing the majority of the inhibitor in an inactive form. Several variants of alpha 1-PI with single amino acid substitutions within their active site loop (amino acids 345-358) were produced in a bioreactor showing that substitution of Met351 with Glu resulted in significantly reduced aggregate formation compared to the other variants and to wild-type protein. In addition, this variant proved to be fully functional as a proteinase inhibitor. Based on these findings and on results of previous structural studies a mechanism for aggregate formation during expression of alpha 1-PI is suggested.

Single peptide bond hydrolysis/resynthesis in squash inhibitors of serine proteinases. 2. Limited proteolysis of Curcurbita maxima trypsin inhibitor I by pepsin

Porcine pepsin hydrolyzes the Leu7-Met8 (P2'-P3') peptide bond in Cucurbita maxima trypsin inhibitor I (CMTI I) in the pH range 2.0-4.8. The reaction proceeds to equilibrium between intact CMTI I and its cleaved form. The pH-independent value of the equilibrium constant (Khyd0 = 0.78) indicates that both forms of the inhibitor have similar Gibbs energies. The pH dependence of this constant shows that the peptide bond hydrolysis does not perturb ionization constants of any preexistent groups. The same equilibrium values can also be reached from the cleaved inhibitor side through pepsin-catalyzed resynthesis of the Leu7-Met8 peptide bond. Catalytic rate constants for the forward (hydrolysis) and reverse (resynthesis) reactions are similar. Both catalytic rate constants are strongly pH dependent, approaching the highest values at pH 2.0. Michaelis constant values for hydrolysis and resynthesis reactions depend much less on pH and are within values typical for oligopeptide substrates of pepsin. The influence of the binding loop rigidity on slow proteolysis by pepsin and other proteinases is discussed.

A study of the effects of altering the sites for N-glycosylation in alpha-1-proteinase inhibitor variants M and S

alpha-1-Proteinase inhibitor (A1Pi) is a monomeric secreted protein glycosylated at asparagines 46, 83, and 247. For this study cDNAs for M (normal) and S (Glu264-->Val) variants of A1Pi were altered by site-directed mutagenesis to produce the combinations of single, double, and triple mutants that can be generated by changing the codons normally specifying these Asn residues to encode Gln. The fates of the mutant proteins were followed in transiently transfected COS-1 cells. All variants with altered glycosylation sites are secreted at reduced rates, are partially degraded, accumulate intracellularly, and some form Nonidet P-40-insoluble aggregates. The carbohydrate attached at Asn83 seems to be of particular importance to the export of both A1PiM and A1PiS from the endoplasmic reticulum. All mutations affecting glycosylation of A1PiS notably reduce secretion, cause formation of insoluble aggregates, and influence degradation of the altered proteins. The variant of A1PiS missing all three glycosylation sites is poorly secreted, is incompletely degraded, and accumulates in unusual perinuclear vesicles. These studies show that N-linked oligosaccharides in A1Pi are vital to its efficient export from the endoplasmic reticulum and that the consequences of changing the normal pattern of glycosylation vary depending upon the sites altered and the variant of A1Pi bearing these alterations.

Glycosylation of annexin I and annexin II

Human placental annexin I and annexin II were shown to be glycosylated by one-dimensional affinity blotting with the lectin concanavalin A, which recognizes D-mannose and D-glucose residues. Further evidence that annexin I and annexin II are glycosylated was provided by the finding that these proteins incorporated D-[2,6-3H]mannose and D-[6-3H]glucose when they were biosynthesized by the human squamous carcinoma cell line SqCC/Y1. Annexin I and annexin II could be rapidly purified from a human placental membrane extract by concanavalin A-Sepharose, which indicated that these proteins contain two biantennary mannosyl residues.

Serpins: Implications of a mobile reactive centre

The serpins are unique among the families of serine proteinase inhibitors in having a reactive centre that is situated on a mobile loop. The structures of three alternative conformations are now known, and it can be deduced that the active form involves the partial insertion of the loop into the A sheet of the molecule. The ability of the loop to move in and out of this sheet has been adapted by evolution to allow the modulation of inhibitory activity. Manipulation of the structure of the loop and of other functional domains in the serpin superfamily enables the production of serpins with tailor-made activities. The ability of the loop to lock in latent conformations or to take part in intermolecular polymerization has implications for the production and stabilization of recombinant serpins. This review has been adapted from Current Opinion in Structural Biology 1992, 2:438-446.