Most principles of protein folding emerged from refolding studies in vitro on small, soluble proteins, because large ones tend to misfold and aggregate. We developed a folding assay allowing the study of large proteins in detergent such that the extent of cellular assistance required for proper folding can be determined. We identified a critical step in the in vivo folding pathway of influenza virus hemagglutinin (HA). Only the formation of the first few disulfides in the top domain of HA required the intact endoplasmic reticulum. After that, HA proceeded to fold efficiently in a very dilute solution, despite its size and complexity. This study paves the way for detailed structural analyses during the folding of complex proteins.Protein folding is the process by which linear polypeptide chains acquire their biologically active three-dimensional conformation. Although all information needed to reach the native conformation is encoded by the amino acid sequence 1 , protein folding inside the cell generally is assisted by molecular chaperones and folding enzymes 2-4 . They mainly protect nascent chains and newly synthesized proteins from misfolding and aggregation. Chaperones and folding enzymes reside in almost every compartment of eukaryotic cells.The endoplasmic reticulum (ER) is highly specialized for the folding of membrane and secretory proteins. One of the bestcharacterized model proteins used to study protein folding in the ER is the influenza virus HA. This protein is not only crucial for influenza virus infection but also an important target for the immune system. HA is a multidomain glycoprotein of 84 kDa with six disulfide bonds and seven N-linked glycans. It folds in the ER, trimerizes and is transported to the plasma membrane, where it is incorporated into new virions. The folding of HA has been studied extensively both in complete cells [5][6][7][8][9][10][11] and in microsomes 12,13 , using radioactive labeling of newly synthesized HA or truncated ribosome-bound nascent chains 14 . Although these studies yielded fairly detailed information on the HA folding process, further molecular insight would require biophysical studies on purified protein, which would disregard the role of the cell.To start bridging the gap between folding studies in vitro and in intact cells, we developed an assay that allows the study of post-translational HA folding in a cell-free system without membranes. We replaced the chase period in the well-established pulse-chase folding assay 6 with a folding analysis in solution: the protein is translated and translocated in intact cells, after which post-translational folding is studied in a detergent cell lysate. Using this 'in vitro chase' assay, we found that HA can fold in a detergent cell lysate without ER assistance, provided that the first folding step occurs in an intact ER.
RESULTS
HA can fold in a detergent cell lysateWhen studied with the pulse-chase assay in intact cells, newly synthesized HA ran as a single band in a reducing SDS-PAGE gel (Fig. 1a, R) and as three ban...