1046 Chemical Reviews, 1998, Vol. 98, No. 3
Table 1. Nobel Prizes in Chemistry for Work on Ribonuclease A
Christian B. Anfinsen (1916-1996) Stanford Moore (1913-1982) William H. Stein (1911-1980) Robert Bruce Merrifield (1921-)
1972 1972 1972 1984
11 “Studies on the principles that govern the folding of protein chains” “The chemical structures of pancreatic ribonuclease and deoxyribonuclease” “The chemical structures of pancreatic ribonuclease and deoxyribonuclease” “Solid-phase synthesis” 9
thoritative reviews on RNase A have disseminated thoughts and information.15-22 In this review, recent information on the structure and function of RNase A is added to the background of historic work. This review emphasizes applications of recombinant DNA technology and nucleic acid chemistry, which are shedding new light on the chemistry and biology of this venerable enzyme.
The cloned gene and cDNA that code for RNase A were expressed initially by relatively low-yielding systems in E. coli,32-34 Bacillus subtillus,35 and Saccharomyces cerevisiae.31,36,37 Similarly, rat pan- creatic ribonuclease was produced at low levels in cultured monkey kidney COS-1 cells.38 RNase 1 (human pancreatic ribonuclease) was produced at low
levels in S. cerevisiae39 hamster ovary cells.40
and in cultured Chinese
II. Heterologous Production
Changing the residues in a protein and analyzing the consequences of these changes is a powerful method for probing the role of particular functional groups in proteins.23,24 Although such changes can be made by either total synthesis or semisynthetic procedures, they can be much easier to effect by site- directed mutagenesis of a gene expressed in a het- erologous host.
The heterologous production of RNase A has been problematic. The difficulty has been due largely to three obstacles. First, the cDNA of RNase A is difficult to clone because the corresponding RNA must be isolated intact from the pancreas, an organ rich in ribonuclease.25 Second, RNase A is suscep- tible to proteolysis when unfolded. Third, high levels of native RNase A are cytotoxic. (See section XII.) These obstacles thwarted the creation of RNase A variants, and work on RNase A began to stall. This lag was made more frustrating by the notable success of early physical and chemical analyses of the en- zyme.
The first heterologous system for the expression of RNase A was based on the total synthesis of a gene that codes for RNase A (which followed the total synthesis of a gene that codes for the S-protein fragment26) and the expression of that gene in Escherichia coli to produce a fusion protein with -galactosidase.27 Purifying RNase A from this sys- tem was made more efficient by the elimination of the -galactosidase fusion tag.28 The RNase A pro- duced had a nonnatural N-formyl methionine residue at its N-terminus. The more recent addition of a murine signal peptide to this system directed active, mature enzyme to be secreted into the periplasm. This system allows approximately 5 mg of soluble RNase A (and 5 mg of insoluble RNase A) to be recovered from each liter of fermented culture. 29
After its synthesis, the gene that codes for RNase A as well as its cDNA were cloned by recombinant DNA methods.30,31 The DNA sequence that codes for the enzyme itself is preceded by a sequence that codes for a peptide of 26 residues.30 This peptide begins with a methionine residue, has a basic residue near the amino terminus, is hydrophobic, and terminates with a glycine residue. Each of these features is characteristic of peptides that signal the secretion of proteins. This signal sequence apparently directs the secretion of RNase A from pancreatic exocrine cells.
Perhaps the most important breakthrough in the heterologous production of RNase A was the develop- ment of pET systems.41 pET systems use the strong T7 RNA polymerase promoter to direct the expression of cloned genes. The resulting proteins are produced in such large quantities that they often aggregate into inclusion bodies. Because RNase A is easy to solubilize and refold, inclusion body formation is not problematic. Rather, the formation of inclusion bod- ies is beneficial because inclusion bodies are easy to isolate and contain almost pure target protein. More- over, unfolded RNase A in inclusion bodies lacks ribonucleolytic activity and thus cytotoxicity. By using a pET system, RNase A that is identical to that isolated from bovine pancreas has been produced with isolated yields of ∼50 mg per liter of culture. RNase 1 has been produced similarly in E. coli cells.39,42-44 Finally, a new system for the efficient production of active, mature RNase A in the peri- plasm of E. coli cells makes use of the alkaline phosphatase signal peptide and the λ PR promoter to produce 40 mg of enzyme per liter of culture. The pET and λ PR systems now make available virtually unlimited quantities of RNase A in which any amino acid residue is replaced with any other. 37 45,46
RNase A was first crystallized over 50 years ago,47,48 and these crystals were shown to diffract to a resolution of 2 Å.49 RNase A was the first enzyme and third protein (after insulin50 and hemoglobin51) for which a correct amino acid sequence was deter- mined,52,53 and the third enzyme and fourth protein (after myoglobin,54,55 lysozyme,56 and carboxypepti- dase A57) whose three-dimensional structure was determined by X-ray diffraction analysis.58 A general method for using fast atom bombardment mass spectrometry (FABMS) to assign completely the disulfide bonds of a protein was developed with RNase A.59 More recently, work on RNase A has yielded the first three-dimensional structure of a protein containing an isoaspartyl residue, which derives from the deamidation of an asparagine residue (here, Asn67).60,61 Finally, the use of NMR spectroscopy in elaborating both protein structure62 and protein folding pathways63 were developed with RNase A. The 1H NMR resonances of the enzyme have been assigned, and the structure of the enzyme in solution has been determined.64-68 NMR spectros-