Hsp90 and phosphorylation of the Slt2(Mpk1) MAP kinase activation loop are essential for catalytic, but not non-catalytic, Slt2-mediated transcription in yeast.

In yeast, the Slt2(Mpk1) stress-activated protein kinase directs the activation of two transcription factors, Rlm1 and Swi4/Swi6, in response to cell wall stress. Rlm1 is activated through a phosphorylation by Slt2, whereas the Swi4/Swi6 activation is noncatalytic and triggered by the binding of phosphorylated forms of both Slt2 and a catalytically inactive pseudokinase (Mlp1). Previous studies have delineated a role for the molecular chaperone Hsp90 in the activation of Slt2, but the involvement of Hsp90 in these events of catalytic versus non-catalytic cell integrity signaling has remained elusive. In cells lacking Mlp1, the Hsp90 inhibitor radicicol was found to inhibit the Slt2-mediated catalytic activation of Rlm1, but not the noncatalytic activation of Swi4/Swi6. Mutation of residues in the TEY motif of the Slt2 activation loop strongly impacted both Hsp90 binding and Rlm1-mediated transcription. In contrast, many of these same mutations had only modest effects on Swi4/6 (Slt2-mediated, non-catalytic) transcription, although one that blocked both the Slt2:Hsp90 interaction and Rlm1-mediated transcription (E191G) triggered a hyperactivation of Swi4/6. Taken together, our results cement the importance of the Slt2 activation loop for both the binding of Hsp90 by Slt2 and the catalytic activation of cell integrity signaling.

UCS protein function is partially restored in the Saccharomyces cerevisiae she4 mutant with expression of the human UNC45-GC, but not UNC45-SM.

A dedicated UNC45, Cro1, She4 (UCS) domain-containing protein assists in the Hsp90-mediated folding of the myosin head. Only weak sequence conservation exists between the single UCS protein of simple eukaryotes (She4 in budding yeast) and the two UCS proteins of higher organisms (the general cell and striated muscle UNC45s; UNC45-GC and UNC45-SM, respectively). In vertebrates, UNC45-GC facilitates cytoskeletal functions, whereas the 55% identical UNC45-SM assists assembly of the contractile apparatus of cardiac and skeletal muscles. A Saccharomyces cerevisiae she4Δ mutant, totally lacking any UCS protein, was engineered to express as its sole Hsp90 either the Hsp90α or the Hsp90ß isoforms of human cytosolic Hsp90. A transient induction of the human UNC45-GC, but not UNC45-SM, could rescue the defective endocytosis in these she4Δ cells at 39 °C, irrespective of whether they possessed Hsp90α or Hsp90ß. UNC45-GC-mediated rescue of the localisation of a Myo5-green fluorescent protein (GFP) fusion to cortical patches at 39 °C was more efficient in the yeast containing Hsp90α, though this may relate to more efficient functioning of Hsp90α as compared to Hsp90ß in these strains. Furthermore, inducible expression of UNC45-GC, but not UNC45-SM, could partially rescue survival at a more extreme temperature (45 °C) that normally causes she4Δ mutant yeast cells to lyse. The results indicate that UCS protein function has been most conserved-yeast to man-in the UNC45-GC, not UNC45-SM. This may reflect UNC45-GC being the vertebrate UCS protein that assists formation of the actomyosin complexes needed for cytokinesis, cell morphological change, and organelle trafficking-events also facilitated by the myosins in yeast.

Benzoate and Sorbate Salts: A Systematic Review of the Potential Hazards of These Invaluable Preservatives and the Expanding Spectrum of Clinical Uses for Sodium Benzoate.

Sodium benzoate and potassium sorbate are extremely useful agents for food and beverage preservation, yet concerns remain over their complete safety. Benzoate can react with the ascorbic acid in drinks to produce the carcinogen benzene. A few children develop allergy to this additive while, as a competitive inhibitor of D-amino acid oxidase, benzoate can also influence neurotransmission and cognitive functioning. Model organism and cell culture studies have raised some issues. Benzoate has been found to exert teratogenic and neurotoxic effects on zebrafish embryos. In addition, benzoate and sorbate are reported to cause chromosome aberrations in cultured human lymphocytes; also to be potently mutagenic toward the mitochondrial DNA in aerobic yeast cells. Whether the substantial human consumption of these compounds could significantly increase levels of such damages in man is still unclear. There is no firm evidence that it is a risk factor in type 2 diabetes. The clinical administration of sodium benzoate is of proven benefit for many patients with urea cycle disorders, while recent studies indicate it may also be advantageous in the treatment of multiple sclerosis, schizophrenia, early-stage Alzheimer's disease and Parkinson's disease. Nevertheless, exposure to high amounts of this agent should be approached with caution, especially since it has the potential to generate a shortage of glycine which, in turn, can negatively influence brain neurochemistry. We discuss here how a small fraction of the population might be rendered-either through their genes or a chronic medical condition-particularly susceptible to any adverse effects of sodium benzoate.

Mutation of the Ser18 phosphorylation site on the sole Saccharomyces cerevisiae UCS protein, She4, can compromise high-temperature survival.

Folding of the myosin head often requires the joint actions of Hsp90 and a dedicated UNC45, Cro1, She4 (UCS) domain-containing cochaperone protein. Relatively weak sequence conservation exists between the single UCS protein of simple eukaryotes (She4 in budding yeast) and the two UCS proteins of higher organisms (the general cell and smooth muscle UNC45s; UNC45-GC and UNC45-SM respectively). In vertebrates, UNC45-GC facilitates cytoskeletal function whereas the 55% identical UNC45-SM assists in the assembly of the contractile apparatus of cardiac and skeletal muscles. UNC45-SM, unlike UNC45-GC, shares with yeast She4 an IDSL sequence motif known to be a site of in vivo serine phosphorylation in yeast. Investigating this further, we found that both a non-phosphorylatable (S18A) and a phosphomimetic (S18E) mutant form of She4 could rescue the type 1 myosin localisation and endocytosis defects of the yeast she4Δ mutant at 39 °C. Nevertheless, at higher temperature (45 °C), only She4 (S18A), not She4(S18E), could substantially rescue the cell lysis defect of she4Δ mutant cells. In the yeast two-hybrid system, the non-phosphorylatable S18A and S251A mutant forms of She4 and UNC45-SM still displayed the stress-enhanced in vivo interaction with Hsp90 seen with the wild-type She4 and UNC45-SM. Such high-temperature enforcement to interaction was though lost with the phosphomimetic mutant forms (She4(S18E) and UNC45-SM (S251E)), an indication that phosphorylation might suppress these increases in She4/Hsp90 and UNC45-SM/Hsp90 interaction with stress.

Insights from yeast into whether the inhibition of heat shock transcription factor (Hsf1) by rapamycin can prevent the Hsf1 activation that results from treatment with an Hsp90 inhibitor.

UNLABELLED: In human cells TORC1 mTOR (target of rapamycin) protein kinase complex renders heat shock transcription factor 1 (Hsf1) competent for stress activation. In such cells, as well as in yeast, the selective TORC1 inhibitor rapamycin blocks this activation in contrast to Hsp90 inhibitors which potently activate Hsf1. Potentially therefore rapamycin could prevent the Hsf1 activation that frequently compromises the efficiency of Hsp90 inhibitor cancer drugs. Little synergy was found between the effects of rapamycin and the Hsp90 inhibitor radicicol on yeast growth. However certain rapamycin resistance mutations sensitised yeast to Hsp90 inhibitor treatment and an Hsp90 mutation that overactivates Hsf1 sensitised cells to rapamycin. Rapamycin inhibition of the yeast Hsf1 was abolished by this Hsp90 mutation, as well as with the loss of Ppt1, the Hsp90-interacting protein phosphatase that is the ortholog of mammalian PP5. Unexpectedly Hsf1 activation was found to have a requirement for the rapamycin binding immunophilin FKBP12 even in the absence of rapamycin, while TORC1 "bypass" strains revealed that the rapamycin inhibition of yeast Hsf1 is not exerted through two of the major downstream targets of TORC1, the protein phosphatase regulator Tap42 and the protein kinase Sch9--the latter the ortholog of human S6 protein kinase 1. SIGNIFICANCE: A problem with most of the Hsp90 inhibitor drugs now in cancer clinic trials is that they potently activate Hsf1. This leads to an induction of heat shock proteins, many of which have a "pro-survival" role in that they help to protect cells from apopotosis. As the activation of Hsf1 requires TORC1, inhibitors of mTOR kinase could potentially block this activation of Hsf1 and be of value when used in combination drug therapies with Hsp90 inhibitors. However many of the mechanistic details of the TORC1 regulation of Hsf1, as well as the interplay between cellular resistances to rapamycin and to Hsp90 inhibitors, still remain to be resolved.

Cdc37 engages in stable, S14A mutation-reinforced association with the most atypical member of the yeast kinome, Cdk-activating kinase (Cak1).

In most eukaryotes, Cdc37 is an essential chaperone, transiently associating with newly synthesised protein kinases in order to promote their stabilisation and activation. To determine whether the yeast Cdc37 participates in any stable protein interactions in vivo, genomic two-hybrid screens were conducted using baits that are functional as they preserve the integrity of the conserved N-terminal region of Cdc37, namely a Cdc37-Gal4 DNA binding domain (BD) fusion in both its wild type and its S14 nonphosphorylatable (Cdc37(S14A)) mutant forms. While this failed to identify the protein kinases previously identified as Cdc37 interactors in pull-down experiments, it did reveal Cdc37 engaging in a stable association with the most atypical member of the yeast kinome, cyclin-dependent kinase (Cdk1)-activating kinase (Cak1). Phosphorylation of the conserved S14 of Cdc37 is normally crucial for the interaction with, and stabilisation of, those protein kinase targets of Cdc37, Cak1 is unusual in that the lack of this Cdc37 S14 phosphorylation both reinforces Cak1:Cdc37 interaction and does not compromise Cak1 expression in vivo. Thus, this is the first Cdc37 client kinase found to be excluded from S14 phosphorylation-dependent interaction. The unusual stability of this Cak1:Cdc37 association may partly reflect unique structural features of the fungal Cak1.

Spotlight on the microbes that produce heat shock protein 90-targeting antibiotics.

Heat shock protein 90 (Hsp90) is a promising cancer drug target as a molecular chaperone critical for stabilization and activation of several of the oncoproteins that drive cancer progression. Its actions depend upon its essential ATPase, an activity fortuitously inhibited with a very high degree of selectivity by natural antibiotics: notably the actinomycete-derived benzoquinone ansamycins (e.g. geldanamycin) and certain fungal-derived resorcyclic acid lactones (e.g. radicicol). The molecular interactions made by these antibiotics when bound within the ADP/ATP-binding site of Hsp90 have served as templates for the development of several synthetic Hsp90 inhibitor drugs. Much attention now focuses on the clinical trials of these drugs. However, because microbes have evolved antibiotics to target Hsp90, it is probable that they often exploit Hsp90 inhibition when interacting with each other and with plants. Fungi known to produce Hsp90 inhibitors include mycoparasitic, as well as plant-pathogenic, endophytic and mycorrhizal species. The Hsp90 chaperone may, therefore, be a prominent target in establishing a number of mycoparasitic (interfungal), fungal pathogen-plant and symbiotic fungus-plant relationships. Furthermore the Hsp90 family proteins of the microbes that produce Hsp90 inhibitor antibiotics are able to reveal how drug resistance can arise by amino acid changes in the highly conserved ADP/ATP-binding site of Hsp90.

Maximising the yeast chronological lifespan.

When investigating aging it is important to focus on the factors that are needed to attain, and which can be manipulated to extend, the longest lifespans. This has long been appreciated by those workers who use Drosophila or Caenorhabditis elegans as model experimental systems to study aging. Often though it seems it is not a consideration in many studies of yeast chronological aging. In this chapter I summarise how recent work has revealed the preconditioning that is needed for yeast to survive for long periods in stationary phase, therefore for it to exhibit a long chronological life span (CLS). Of critical importance in this regard is the nature of the nutrient limitation that, during the earlier growth phase, had forced the cells to undergo growth arrest. I have attempted to highlight those studies that have focussed on the longest CLSs, as this helps to identify investigations that may be addressing - not just factors that can influence chronological longevity - but those factors that are correlated with the authentic processes of chronological aging. Attempting to maximize long-term stationary survival in yeast should also enhance the potential relevance of this organism as an aging model to those who wrestle with the problems of aging in more complex systems. Finally I also give a personal perspective of how studies on the yeast CLS may still yet provide some important new insights into events that are correlated with aging.

Activity of the yeast zinc-finger transcription factor War1 is lost with alanine mutation of two putative phosphorylation sites in the activation domain.

Saccharomyces cerevisiae acquires its resistance to carboxylate weak organic acids by inducing a plasma membrane ABC transporter, Pdr12. These acids activate a Zn(II)2Cys6 zinc-finger transcription factor, War1, which in turn induces the PDR12 gene. Mutation of the four potential sites of serine/threonine phosphorylation within the War1 activation domain revealed that Pdr12 induction was lost with mutations S923A or S930A, but not with the corresponding phosphomimetic mutations S923D or S930D. However, phosphorylation at these two sites has not been detected by mass spectrometry, so it still remains uncertain whether these are true sites of phosphorylation or merely serines whose side-chain hydroxyls are necessary for the proper structuring of the War1 activation domain. Mutation S923A prevented the sorbate-induced hyperphosphorylation of War1, while S930A caused War1 to be in a constitutively hyperphosphorylated state, irrespective of weak acid stress. Screening of non-essential protein kinase mutants of yeast failed to identify a kinase required for Pdr12 induction, or War1 hyperphosphorylation, in response to sorbate treatment. However, the mrk1∆ mutant was identified as having an elevated Pdr12 level in the absence of sorbate stress.

Resistance of yeasts to weak organic acid food preservatives.

Carboxylate weak acids are invaluable for large-scale food and beverage preservation. However, in response to safety concerns, there is now desire to reduce the use of these additives. The resistance to these compounds displayed by spoilage yeasts and fungi is a major reason why these preservatives often have to be used in millimolar levels. This chapter summarizes the mechanisms whereby yeasts are rendered resistant to acetate, propionate, sorbate, and benzoate. In baker's yeast (Saccharomyces cerevisiae), resistance to high acetic acid is acquired partly by loss of the plasma membrane aquaglyceroporin that facilitates the passive diffusional entry of undissociated acid into cells (Fps1), and partly through a transcriptional response mediated by the transcription factor Haa1. Other carboxylate preservatives are too large to enter cells through the Fps1 channel but instead penetrate at appreciable rates by passive diffusion across the plasma membrane. In Saccharomyces and Candida albicans though not, it seems, in the Zygosaccharomyces, resistance to the latter acids involves activation of the War1 transcription factor, which in turn generates strong induction of a specific plasma membrane ATP-binding cassette transporter (Pdr12). The latter actively pumps the preservative anion from the cell. Other contributors to weak acid resistance include enzymes that allow preservative degradation, members of the Tpo family of major facilitator superfamily transporters, and changes to the cell envelope that minimize the diffusional entry of the preservative into the cell.

Features of the Streptomyces hygroscopicus HtpG reveal how partial geldanamycin resistance can arise with mutation to the ATP binding pocket of a eukaryotic Hsp90.

Much attention is focused on the benzoquinone ansamycins as anticancer agents, with several derivatives of the natural product geldanamycin (GdA) now in clinical trials. These drugs are selective inhibitors of Hsp90, a molecular chaperone vital for many of the activities that drive cancer progression. Mutational changes to their interaction site, the extremely conserved ATP binding site of Hsp90, would mostly be predicted to inactivate the chaperone. As a result, drug resistance should not arise readily this way. Nevertheless, Streptomyces hygroscopicus, the actinomycete that produces GdA, has evolved an Hsp90 family protein (HtpG) that lacks GdA binding. It is altered in certain of the highly conserved amino acids making contacts to this antibiotic in crystal structures of GdA bound to eukaryotic forms of Hsp90. Two of these amino acid changes, located on one side of the nucleotide-binding cleft, weakened GdA/Hsp90 binding and conferred partial GdA resistance when inserted into the endogenous Hsp90 of yeast cells. Crystal structures revealed their main effect to be a weakening of interactions with the C-12 methoxy group of the GdA ansamycin ring. This is the first study to demonstrate that partial GdA resistance is possible by mutation within the ATP binding pocket of Hsp90.

Threonine 22 phosphorylation attenuates Hsp90 interaction with cochaperones and affects its chaperone activity.

Heat shock protein 90 (Hsp90) is an essential molecular chaperone whose activity is regulated not only by cochaperones but also by distinct posttranslational modifications. We report here that casein kinase 2 phosphorylates a conserved threonine residue (T22) in α helix-1 of the yeast Hsp90 N-domain both in vitro and in vivo. This α helix participates in a hydrophobic interaction with the catalytic loop in Hsp90's middle domain, helping to stabilize the chaperone's ATPase-competent state. Phosphomimetic mutation of this residue alters Hsp90 ATPase activity and chaperone function and impacts interaction with the cochaperones Aha1 and Cdc37. Overexpression of Aha1 stimulates the ATPase activity, restores cochaperone interactions, and compensates for the functional defects of these Hsp90 mutants.

Mechanisms of Resistance to Hsp90 Inhibitor Drugs: A Complex Mosaic Emerges.

The molecular chaperone Hsp90 holds great promise as a cancer drug target, despite some of the initial clinical trials of Hsp90 inhibitor drugs having not lived up to expectation. Effective use of these drugs will benefit greatly from a much more detailed understanding of the factors that contribute to resistance, whether intrinsic or acquired. We review how cell culture studies have revealed a number of different mechanisms whereby cells can be rendered less susceptible to the effects of Hsp90 inhibitor treatment. A major influence is Hsp90 inhibition causing strong induction of the heat shock response, a stress response that increases cellular levels of prosurvival chaperones such as Hsp27 and Hsp70. Another problem seems to be that these inhibitors do not always access the Hsp90 proteins of the mitochondrion, forms of Hsp90 that-in cancer cells-are operating to suppress apoptosis. It should be possible to overcome these drawbacks through the appropriate drug redesign or with the combinatorial use of an Hsp90 inhibitor with a drug that targets either heat shock factor or the chaperone Hsp70. Still though, cells will often differ in the key antiapoptotic versus proapoptotic activities that are dependent on Hsp90, in the key steps in their apoptotic pathways responsive to Hsp90 inhibition or Hsp70 level, as well as the extents to which their survival is dependent on oncogenic tyrosine kinases that are clients of Hsp90. A systems approach will therefore often be required in order to establish the most prominent effects of Hsp90 inhibition in each type of cancer cell.

The Fps1p aquaglyceroporin facilitates the use of small aliphatic amides as a nitrogen source by amidase-expressing yeasts.

Saccharomyces cerevisiae acquires a resistance to high, toxic levels of acetic acid by destabilizing Fps1p, the plasma membrane aquaglyceroporin through which this acid - in its undissociated state - enters the cell. In this study, Fps1p loss was shown to confer resistances to acetic acid, acrolein and allyl alcohol, not just in S. cerevisiae but also in the osmotolerant spoilage yeast Zygosaccharomyces rouxii. However, in Z. rouxii, the loss of Fps1p severely compromised the use of acetamide and several other small amides as sources of nitrogen, an indication that these amides enter the cells of this yeast by passive diffusion through the Fps1p pore. Saccharomyces cerevisiae cannot grow on acetamide, but was conferred with an ability to use this and other small amides as nitrogen sources by heterologous expression of a Z. rouxii ORF (ZrAMD1) with protein sequence identity to the amdS-encoded amidase of Aspergillus nidulans. This capacity of ZrAMD1-expressing S. cerevisiae to assimilate amide nitrogen was severely compromised by the loss of Fps1p. ZrAMD1 appears to encode the major amidase of Z. rouxii as a Zramd1Delta deletant mutant had, like the Zrfps1Delta deletant, lost the ability to assimilate small amides as sources of nitrogen.

A simple yeast-based system for analyzing inhibitor resistance in the human cancer drug targets Hsp90alpha/beta.

Heat shock protein 90 (Hsp90), a highly conserved molecular chaperone, is one of the most promising targets for cancer drug development. Whether any resistance to these Hsp90 inhibitor drugs could arise by Hsp90 mutation is still unknown. Yeast is readily engineered so that its essential Hsp90 function is provided by either isoform of the human cytosolic Hsp90, Hsp90alpha or Hsp90beta. However, its high intrinsic resistance to most drugs poses a major obstacle to the use of such Hsp90alpha- or Hsp90beta-expressing yeast cells as a model system to analyse whether drug resistance might arise by Hsp90 mutation. In order to overcome this problem, we have generated a strain that is both hypersensitive to Hsp90 inhibitors as it lacks multiple drug resistance genes, and in which different heterologous and mutant Hsp90s can be expressed by plasmid exchange. It is not rendered appreciably stress sensitive when made to express Hsp90alpha or Hsp90beta as its sole form of Hsp90. Should there be any development of resistance to the Hsp90 drugs now in cancer clinic trials, this system can provide a rapid initial test of whether any single nucleotide polymorphism appearing within the coding regions of Hsp90alpha or Hsp90beta could be a contributory factor in this resistance. We have used this strain to demonstrate that significant levels of resistance to the Hsp90 inhibitors radicicol and 17-allylamino-demethoxygeldanamycin (17-AAG) are generated as a result of the same single point mutation within the native Hsp90 of yeast (A107N), the human Hsp90alpha (A121N) and the human Hsp90beta (A116N).

Swe1Wee1-dependent tyrosine phosphorylation of Hsp90 regulates distinct facets of chaperone function.

Saccharomyces WEE1 (Swe1), the only "true" tyrosine kinase in budding yeast, is an Hsp90 client protein. Here we show that Swe1(Wee1) phosphorylates a conserved tyrosine residue (Y24 in yeast Hsp90 and Y38 in human Hsp90alpha) in the N domain of Hsp90. Phosphorylation is cell-cycle associated and modulates the ability of Hsp90 to chaperone a selected clientele, including v-Src and several other kinases. Nonphosphorylatable mutants have normal ATPase activity, support yeast viability, and productively chaperone the Hsp90 client glucocorticoid receptor. Deletion of SWE1 in yeast increases Hsp90 binding to its inhibitor geldanamycin, and pharmacologic inhibition/silencing of Wee1 sensitizes cancer cells to Hsp90 inhibitor-induced apoptosis. These findings demonstrate that Hsp90 chaperoning of distinct client proteins is differentially regulated by specific posttranslational modification of a unique subcellular pool of the chaperone, and they provide a strategy to increase the cellular potency of Hsp90 inhibitors.

Presence of the Fps1p aquaglyceroporin channel is essential for Hog1p activation, but suppresses Slt2(Mpk1)p activation, with acetic acid stress of yeast.

When grown at pH 4.5, Saccharomyces cerevisiae acquires a resistance to inhibitory acetic acid levels ( approximately 0.1 M) by destabilizing Fps1p, the plasma membrane aquaglyceroporin that provides the main route for passive diffusional entry of this acid into the cell. Acetic acid stress transiently activates Hog1p mitogen-activated protein (MAP) kinase, which, in turn, phosphorylates Fps1p in order to target this channel for endocytosis and degradation in the vacuole. This activation of Hog1p is abolished with the loss of Fps1p, but is more sustained when cells express an open Fps1p channel refractory to destabilization. At neutral pH, much higher levels of acetate ( approximately 0.5 M) are needed to inhibit growth. Under such conditions, the loss of Fps1p does not abolish, but merely slows, the activation of Hog1p. Acetate stress also activates the Slt2(Mpk1)p cell integrity MAP kinase, possibly by causing inhibition of glucan synthase activity. In pH 4.5 cultures, this acetate activation of Slt2p is strongly enhanced by the loss of Fps1p and is dependent upon the cell surface sensor Wsc1p. Lack of Fps1p therefore exerts opposing effects on the activation of Hog1p and Slt2p in yeast exposed to acetic acid stress.

The Hsp90/Cdc37p chaperone system is a determinant of molybdate resistance in Saccharomyces cerevisiae.

Saccharomyces cerevisiae lacks enzymes that contain the molybdopterin co-factor and therefore any requirement for molybdenum as a trace mineral supplement. Instead, high molybdate levels are inhibitory to its growth. Low cellular levels of heat shock protein 90 (Hsp90), an essential chaperone, were found to enhance this sensitivity to molybdate. Certain Hsp90 point mutations and co-chaperone protein defects that partially compromise the function of the Hsp90/Cdc37p chaperone system also rendered S. cerevisiae hypersensitive to high molybdate levels. Sensitivity was especially apparent with mutations close to the Hsp90 nucleotide binding site, with the loss of the non-essential co-chaperone Sti1p (the equivalent of mammalian Hop), and with the abolition of residue Ser14 phosphorylation on the essential co-chaperone Cdc37p. While it remains to be proved that these effects reflect direct inhibition of the Hsp90 of the cell by the MoO(4) (2+) oxyanion in vivo; this possibility is suggested by molybdate sensitivity arising with a mutation in the Hsp90 nucleotide binding site that does not generate stress sensitivity or an impaired stress response. Molybdate sensitivity may therefore be a useful phenotype to score when studying mutations in this chaperone system.

Structural basis of the radicicol resistance displayed by a fungal hsp90.

Heat shock protein 90 (Hsp90) is a promising cancer drug target, as multiple oncogenic proteins are destabilized simultaneously when it loses its activity in tumor cells. Highly selective Hsp90 inhibitors, including the natural antibiotics geldanamycin (GdA) and radicicol (RAD), inactivate this essential molecular chaperone by occupying its nucleotide binding site. Often cancer drug therapy is compromised by the development of resistance, but a resistance to these Hsp90 inhibitors should not arise readily by mutation of those amino acids within Hsp90 that facilitate inhibitor binding, as these are required for the essential ATP binding/ATPase steps of the chaperone cycle and are tightly conserved. Despite this, the Hsp90 of a RAD-producing fungus is shown to possess an unusually low binding affinity for RAD but not GdA. Within its nucleotide binding site a normally conserved leucine is replaced by isoleucine, though the chaperone ATPase activity is not severely affected. Inserted into the Hsp90 of yeast, this conservative leucine to isoleucine substitution recreated this lowered affinity for RAD in vitro. It also generated a substantially enhanced resistance to RAD in vivo. Co-crystal structures reveal that the change to isoleucine is associated with a localized increase in the hydration of an Hsp90-bound RAD but not GdA. To the best of our knowledge, this is the first demonstration that it is possible for Hsp90 inhibitor resistance to arise by subtle alteration to the structure of Hsp90 itself.