CHEMICAL BIOLOGY / BIOLOGICAL CHEMISTRY 281 CHIMIA 2001, 55. No.4 Chimia 55 (2001) 281-285 © Schweizerische Chemische Gesellschaft ISSN 0009-4293 Protein Folding and Assembly Rudi Glockshuber* Abstract: One of the central dogmas in biochemistry is the view that the biologically active, three-dimensional structure of a protein is unique and exclusively determined by its amino acid sequence, and that the active conformation of a protein represents its state of lowest free energy in aqueous solution. Despite a large number of novel experiments supporting this view, including an exponentially increasing number of solved three- dimensional protein structures, it is still impossible to predict the tertiary structure of a protein from knowledge of its amino acid sequence alone. Towards the goal of identifying general principles underlying the mechanism of protein folding in vitro and in vivo, we are pursuing several projects that are briefly described in this article: (1) Circular permutation of proteins as a tool to study protein folding, (2) Catalysis of disulfide bond formation during protein folding, (3) Assembly of adhesive type 1 pili from Escherichia coli strains, and (4) Structure and folding of the mammalian prion protein. Keywords: Catalysis of disulfide bond formation' Prion proteins· Protein assembly' Protein folding' Type 1 pili During the last 20 years, the field of pro- molecular chaperones which assist the dom circular permutation analysis of a tein folding has gained enormous general protein folding process by preventing ir- protein using the periplasmic disulfide attention in the life sciences. First, this is reversible aggregation of unfolded and oxidoreductase DsbA from E. coli as a because large quantities of biologically partially folded polypeptide chains, and model system [2]. DsbA is a monomeric active, recombinant proteins are required iii) the prion hypothesis, suggesting the two-domain protein of 189 residues with for both modem diagnostic and therapeu- existence of self-propagating, exclusive- known three-dimensional structure. After tic medicine and structural biology. As ly proteinaceous infectious agents. Our linkage of the natural termini, each regu- many recombinant proteins can only be group is addressing basic questions on lar secondary structure and each loop re- produced as unfolded protein aggregates the mechanism of protein folding and as- gion of the protein was disrupted by in- (inclusion bodies) and thus have to be re- sembly within the following research troduction of new termini. Functional folded in vitro, there is a strong need for projects: analysis of about 70 different permuted the development of functional protein ex- variants showed that only about 30% of pression systems and protein refolding the polypeptide chain of DsbA may not protocols. Besides these practical as- 1. Circular Permutation of Proteins: be disrupted by new termini without loss pects, numerous human diseases, includ- A Tool to Study Protein Folding of folding competence (Fig. la). Surpris- ing Alzheimer's disease and prion diseas- ingly, novel termini were even tolerated es, are associated with protein misfolding One of the key questions about pro- within many regular secondary structure and formation of amyloid aggregates. tein folding is whether polypeptide elements [2]. All purified, biologically Therefore, investigations on the mecha- chains require unique nucleation sites to active permuted DsbA variants showed nisms underlying protein folding and fold to the native state. To identify such the same tertiary structure and spectro- protein misfolding are of key interest to nucleation sites, we are applying circular scopic properties as the wild type protein. many areas in biology and medicine. The permutation, i.e. connection of the natu- In contrast, all permuted variants with most important recent advances in the ral termini of a protein by a linker peptide new termini within the 'forbidden' re- field of protein folding [1] involve i) the and introduction of new termini by cleav- gions, comprising four a-helices of view that multiple pathways are available age of the polypeptide chain at a different DsbA (Fig. la), proved to be catalytically for a polypeptide chain to fold to its na- site. The rationale behind this approach is inactive and showed a significantly low- tive conformation, ii) the discovery of that folding nuclei formed by sequential- ered secondary structure content com- ly adjacent polypeptide segments may no pared to the wild type. This demonstrates longer be formed when the correspond- that random circular permutation can be ·Correspondence: Prof. Dr. R. Glockshuber ing segments are far apart in a circularly used as a general method to identify seg- Institut fOr Molekularbiologie und Biophysik Eidgenossische Technische Hochschule permuted protein variant. ments in a protein that are essential for Honggerberg Using genetic engineering in conjunc- folding and stability. As some of the cir- CH-8093 ZOrich tion with screening or selection proce- cularly permuted DsbA variants that we Tel.: +41 16336819 Fax: +411 6331036 dures at the bacterial colony level, 'we have generated are catalytically more ac- E-Mail: [email protected] have performed the first systematic ran- tive than the wild type protein [2], circu- CHEMICAL BIOLOGY / BIOLOGICAL CHEMISTRY 282 CHIMIA 2001,55, NO.4 lar permutation may also become a useful N tool for generating proteins with im- proved functions. N C In addition, we have applied a series of rational circular permutation experi- ments to study the folding of the green fluorescent protein (GFP) from Aeqora victoria. This monomeric 238 residue protein has the unique property that it spontaneously forms a p-hydroxyben- zylidene-imidazolidone chromophore dur- ing folding through main chain cycliza- tion and subsequent air oxidation [3]. In N the three-dimensional structure of native GFP, the chromophore is located within an II-stranded ~-barrel (Fig. Ib) and sol- a b c vent inaccessible. All these properties have made GFP one of the most impor- Fig. 1: Results of circular permutation studies on the disulfide oxidoreductase DsbA from E. coli tant reporter molecules available for stud- (a) and the green fluorescent protein (GFP)from A. victoria (b). (a)The disulfide oxidoreductase ying biological processes such as tran- DsbA from E. coli is required for disulfide bond formation in the bacterial periplasm. The monomeric protein consists of 189 residues. The ribbon diagram shows that DsbA is composed scription, protein targeting and protein/ of a thioredoxin-like domain (blue) and an a-helical domain (top, red) which is inserted into the ligand interactions inside the living cell thioredoxin motif. The sulphur atoms of the active-site disulfide bond are indicated in yellow. [3]. Analysis of 20 different circularly Random circular permutation of DsbA revealed that four a-helices (grey) may not be disrupted permuted GFP variants produced in E. by introduction of new termini in circularly permuted variants without loss offolding competence, coli showed that GFP is much less toler- The natural termini were connected by a 5-residue linker peptide in all permuted variants. (b) ant toward circular permutations than Ribbon representation of the three-dimensional structure of GFP (238 residues, monomer). The chromophore in the centre of the 11-stranded ~-barrel is shown in a ball-and-stick representa- OsbA [4]. Specifically, any disruption of tion. (c) Rational circular permutation of GFP. The sequential arrangement of regular secondary a strand from the ~-barrel by introduction structures is shown, as well as the hexapeptide linker used to connect the natural termini. The of new termini causes folding incompe- red and green symbols indicate the positions of newly introduced termini in permuted GFP tence. These data indicate that the 11- variants. Green circles: variants which retained folding competence during expression in E. coli. stranded ~-barrel of GFP represents an Red squares: variants which were either proteolytically degraded in the E.coli cytoplasm, or extremely cooperative unit which re- soluble and nonfluorescent, or insoluble in the cytoplasm. quires all interactions for stability (Fig. lc). 2. Catalysis of Oxidative Protein ent enzymes are responsible for oxidation actions with their substrate proteins in Folding by Bacterial Disulfide of folding proteins and isomerization of vivo determines their catalytic function, Oxidoreductases disulfide bonds, i.e. OsbA and OsbCI rather than the recycling of their reactive OsbG, respectively [5]. We are character- redox form [10][11]. We are presently Oisulfide bond formation generally izing the catalytic properties of bacterial extending our mechanistic studies on the constitutes the rate-limiting step during members of this enzyme family, in par- function of OsbC and are working on folding of secretory proteins. The reason ticular those of OsbA, an enzyme with bacterial expression systems that provide is that a disulfide bond cannot form auto- extraordinary biophysical properties. OsbA high yields of correctly folded proteins matically when two cysteine residues is a very strong oxidant, undergoes the with disulfide bonds in the periplasm come close during the folding process, fastest disulfide exchange reactions known through coexpression of OsbA and OsbC. because disulfide bond formation is an so far, and randomly transfers its reactive oxidation process which requires an ex- disulfide bond to folding polypeptides ternal