The 90 000 dalton heat shock protein, a lot of smoke but no function as yet

What does the 90 000 dalton heat shock protein (hsp 90) really do? What cellular function does it carry out that is physiologically so important that its amino acid sequence and thus presumably its specific structure have been conserved in evolutionarily diverse organisms ranging from mammals, chickens, Drosophila, trypanosomes, and yeast, to bacteria? Is there a fundamental component of hsp 90 function that is similar or identical for other heat shock proteins? The elusive answer to these enigmatic questions has generated a not so small scientific industry that appears to be in a log phase of growth and publications. Proteins of the hsp 90 family are present in most, if not all, prokaryotic and eukaryotic cells in which they may constitute up to 1-2% of the total protein. The synthesis of some, but not all, hsp 90 species is strongly induced by heat shock and other forms of environmental stress, which leads to its classification as one of the major heat shock proteins. However, it is relatively abundant even in unstressed cells, presumably reflecting its normal function. In Saccharomyces cerevisiae two closely related genes, HSP-83 and HSC-83, are known that code for proteins of the hsp 90 family. HSC-83 is constitutively expressed at high levels and is moderately heat inducible. The HSP-83 product is constitutively synthesized at a lower level in unstressed cells, but is more strongly induced by heat (Lindquist and Craig, Annu. Rev. Genet., 22, 631-677, 1988). The latter is accumulated as yeast cells go into the stationary phase of growth or begin to sporulate. Double mutations in both genes are lethal; however, single mutations in either gene generate a heat-sensitive phenotype that cannot be grown at elevated temperature. This suggests that hsp 90 is required for growth at normal temperature, but that larger amounts of the protein are required under conditions of thermal stress. Two forms of hsp 90 are recognized in vertebrates, called a and /3 in humans and hsp 86 and hsp 84 in mice. These very similar proteins have about 86% homology. The hsp 90 forms appear to be primarily cytosolic and are present in a ratio of a:/3 of 4:6 (Lees-Miller and Anderson, J. Biol. Chem., 264,2431-2437, 1989). A third closely related gene codes for a very similar protein that is associated with the endoplasmic reticulum (ER). It has a 24 amino acid extension at its C-terminus, which has been shown to provide its retention on the membrane of the ER (Munro and Pelham, Cell, 48, 899-907, 1987). This ER protein is induced by glucose starvation and has thus been classified as GRP94 (glucose-regulated protein). Hsp 90 is an acidic phosphoprotein with an isoelectric point of about pH 5.5. Its most striking physical feature is its highly elongated form. In dilute solution it occurs as a dimer that chromatographs during gel filtration like a spherical protein of relative mass > lo6. From the Stokes radius of 83 A (1 A = 0.1 nm), the sedimentation coefficient (Sw,m) of 6.0, and the assumption that the protein is a prolate ellipsoid with partial specific volume of 0.74, we estimate the axial ratio to be 21 (Rose et al., Biochemistry, 26, 6583-6587, 1987). Others have determined similar frictional coefficient ratios which indicate that the protein is highly elongated. This is a primary characteristic of structural proteins, but is very unusual for enzymes. Hsp 90 has no recognized intrinsic enzymatic activity. However, it appears to form apparently specific complexes with a number of proteins. In vertebrates all steroid hormone receptors studied to date, including the estrogen, progesterone, and glucocorticoid receptors, can be isolated as an easily disrupted, inactive complex with hsp 90 from which they can be dissociated by the hormone, high salt, or metal chelators (reviewed by Lindquist and Craig, op. cit.). It has been hypothesized that the receptors bind to a highly charged DNA-like region near the amino terminus of hsp 90, thus forming an inactive complex until the release is triggered by binding of the hormone (Baulieu, J. Cell Biol., 35, 161-174, 1987). The transforming protein of Rous sarcoma virus, pp60SrC, was found to be associated with hsp 90 (Brugge et al., Cell, 26, 363-372, 1981; Oppermann et al., Proc. Natl. Acad. Sci. U.S.A., 78, 1067-1071, 1981). The inactive enzyme is synthesized and then, in unknown order, is associated with hsp 90, phosphorylated on tryosine, released, and inserted into the cell membrane as an active tyrosine kinase. A similar association has been reported for four other transforming proteins with protein kinase activity (Ziemiecki, Virology, 15, 265-273, 1986), indicating that a general mechanism may be involved. Our own interest in hsp 90 stems from its potent inhibitor effect on protein synthesis in the reticulocyte lysate system (Rose et al., J. Biol. Chem., 264, 6239-6244, 1989) and its stimulatory effect on the heme-controlled kinase that phosphorylates the a-subunit of peptide initiation factor 2 (HCR eIF-2a kinase). There is some indication that these effects are not entirely equivalent. The eIF-2a kinase is activated under conditions of heme deficiency in mammalian reticulocytes or their cell-free lysates. Phosphorylation of the a-subunit of eIF-2 leads to a block in protein synthesis, apparently by inhibiting the GDP-GTP exchange reaction that allows recycling of the factor (reviewed in Pain, Biochem. J., 235, 625-637, 1986). Increased phosphorylation of eIF-2a has also been observed in mammalian cells under conditions of heat shock or other stress that cause a reduction in protein synthesis and this eIF-2a kinase has been implicated. A 90 000 dalton protein was found to be physically associated with the highly purified reticulocyte eIF-2a kinase (Wallis et al., Biochemistry, 19, 798-804, 1980). It had no inherent enzymatic activity, but it enhanced eIF-2a phosphorylation when added to the kinase preparation. This protein was later identified as hsp 90. Previously