Chemical rescue of mutant proteins in living Saccharomyces cerevisiae cells by naturally occurring small molecules

Abstract

Intracellular proteins function in a complex milieu wherein small molecules influence protein folding and act as essential cofactors for enzymatic reactions. Thus protein function depends not only on amino acid sequence but also on the concentrations of such molecules, which are subject to wide variation between organisms, metabolic states, and environmental conditions. We previously found evidence that exogenous guanidine reverses the phenotypes of specific budding yeast septin mutants by binding to a WT septin at the former site of an Arg side chain that was lost during fungal evolution. Here, we used a combination of targeted and unbiased approaches to look for other cases of "chemical rescue" by naturally occurring small molecules. We report in vivo rescue of hundreds of Saccharomyces cerevisiae mutants representing a variety of genes, including likely examples of Arg or Lys side chain replacement by the guanidinium ion. Failed rescue of targeted mutants highlight features required for rescue, as well as key differences between the in vitro and in vivo environments. Some non-Arg mutants rescued by guanidine likely result from "off-target" effects on specific cellular processes in WT cells. Molecules isosteric to guanidine and known to influence protein folding had a range of effects, from essentially none for urea, to rescue of a few mutants by DMSO. Strikingly, the osmolyte trimethylamine-N-oxide rescued ∼20% of the mutants we tested, likely reflecting combinations of direct and indirect effects on mutant protein function. Our findings illustrate the potential of natural small molecules as therapeutic interventions and drivers of evolution.

Keywords: chaperone; chemical biology; drug action; mutant; yeast genetics.

Figure 1

Chemical rescue by GdnHCl of Arg-mutant ornithine transcarbamylase (OTC) in living yeast cells. (A) Superimposition of crystal structures of human and E. coli OTC. The zoomed-in image shows the locations of the side chains of Arg57 from the bacterial enzyme (blue) and Arg92 from the human enzyme. (B) Sequence alignment of portions of the OTC homologs from E. coli (EcOTC), human (HsOTC), and S. cerevisiae (ScOTC). Bold residues are identical. Blue highlights the leader sequence that directs import of HsOTC into mitochondria. Pink highlights the position of Arg57 in EcOTC, Arg92 in HsOTC, and Arg69 in ScOTC. (C) Dilution series of cells of the indicated genotypes on rich (YPD) or synthetic medium lacking arginine (“-arg”) and thus selective for OTC function. The carbon source in these plates was 2% galactose to induce OTC expression. Where indicated, ornithine (“+orn”) and/or GdnHCl were added to final concentration of 100 µg/mL or 3 mM, respectively, to attempt to improve HsOTC function. Selective plates were incubated at 30˚ for the indicated number of days (“3 d” or “9 d”) prior to imaging; YPD plate was incubated for 3 days. The strain was H06538 carrying plasmids pRS316 (“arg3Δ”), YCpU-Pgal-ASN1 (“ASN1”), G00648 [“arg3(R69G)”], pGALOTC (“hOTC”), G00639 [“hOTC(R92G)”], or G00796 (“hOTCΔ6-34”). (D) A lawn of H06538 cells carrying plasmid G00648 was plated to media selective for OTC function and a sterile paper disc soaked with GdnHCl or urea at the indicated concentrations was placed on the lawn. The plate was then imaged after 6 days incubation at 30˚.

Figure 2

In vivo chemical rescue by GdnHCl of Ura3(R235A). (A) Dilution series of cells of the indicated genotypes on rich (YPD) or synthetic medium lacking uracil and thus selective for Ura3 function. Plates were incubated at 30˚ for the indicated number of days prior to imaging. Strains were: “WT,” S288C; “WT recoded,” H06700; “ura3(R253A),” H06701; “ura3Δ,” BY4741. (B) As in (A) but on medium with 20 µg/mL uracil and 400 µg/mL FOA, which is converted by Ura3 to a toxic product that inhibits colony growth.

Figure 3

No evidence of chemical rescue of Asn1(R344A) by GdnHCl in living yeast cells. (A) Overlay of crystal structures of EcAsnB (PDB 1CT9) and HsASNS (PDB 6GQ3) showing the locations of EcAsnB Arg 325 and HsASNS Arg 339 and two bound substrates, glutamine (“Gln”) and AMP. (B) Sequence alignment of a region of asparagine synthase from S. cerevisiae (“ScAsn1”), E. coli (“EcAsnB”), and H. sapiens (“HsASNS”). The Arg residue corresponding to EcAsnB Arg 325 is highlighted in red. (C) Dilution series of cells of the indicated genotypes on complete (but lacking uracil) or synthetic medium lacking asparagine (and uracil) and thus selective for Asn1 function (and the URA3-marked plasmid). Plates were incubated at 30˚ for 6 days prior to imaging. Strain was H06544 carrying empty vector plasmid pRS426 (“asn1Δasn2Δ”), YCpU-Pgal-ASN1 (“ASN1”), or YCpU-Pgal-asn1(R344A) [“asn1(R344A)”].

Figure 4

GdnHCl fails to rescue p53(R249S). (A) Crystal structures of the DNA-binding domain of WT p53 (PDB 2AC0) and p53(R249S H168R) (PDB 3D0A) showing the locations of relevant side chains and, in particular, the guanidino group (“Gdm”) occupying the same position in both proteins. Glu171 is shown because in both cases it forms a salt bridge with the guanidino group. (B) Schematic of yeast reporter system. (C) Dilution series of cells of the reporter strain (RBy33) carrying HIS3-marked plasmids (“WT,” pMAM44; “R249S,” pMAM71; “R249S H168A,” pMAM70; “R249S H168A T123A,” and pMAM113) were plated on synthetic medium lacking uracil (“-ura”) or containing 50 µg/mL uracil and 0.1% FOA and incubated at 30˚ prior to imaging.

Figure 5

Chemical rescue by GdnHCl of Arg-mutant actin. (A) Cells from a collection of TS mutants were pinned using a screening robot to rich (YPD) plates with or without 3 mM GdnHCl and incubated at 37˚ prior to imaging. Circles are color-coded with text below to indicate specific mutants. Dashed circles indicate failure to form a colony. (B) As in (A) but imaging a portion of a different plate, and using squares to indicate specific mutants. (C) Dilution series of WT (BY4741) or act1-105 cells on rich (YPD) medium at the indicated temperatures (“RT,” room temperature) with or without GdnHCl for 3 days. (D) 100-µl cultures of act1-105 cells in rich (YPD) medium with the indicated additives (three cultures for each condition) were incubated with constant shaking at 37˚ in a 96-well plate and the absorbance at 600 nm (“A₆₀₀”) was measured every 5 minutes. (E) Crystal structure of yeast Act1 (PDB 1YAG) showing the location of Arg 312 and nearby residues, and other Arg residues substituted in other TS mutants present in the collection we screened. Inset shows the distances (all 2.9 Å) between atoms in the guanidino group and nearby amino acids, representing presumed intramolecular contacts.

Figure 6

GdnHCl rescue of a Lys-mutant allele of Pga1. (A) Dilution series on rich (YPD) medium with or without 3 mM GdnHCl at the indicated temperatures for strains carrying the indicated alleles of PGA1. Strains were H06474, KSY182, and AKY19. (B) Sequence alignment of Pga1 sequences from various fungal species, obtained from the Yeast Genome Database and NCBI server and aligned using COBALT. Residues identical in all WT sequences are in bold. Substitutions in TS mutant ScPga1 alleles are indicated above, those in Pga1-ts are highlighted by shading. Pink, K98N substitution in Pga1-ts at a position where Arg is found in other species. Sc, S. cerevisiae; Sp, S. paradoxus; Sk, S. kudriavzevii; Lk, Lachancea kluyveri; Nc, Naumovozyma castelli; Ag, Ashbya gossypii. (C) As in (A) but with, where indicated, 0.375 mM of GdnHCl (“guanidine”) or derivatives thereof. Where multiple drugs were added, the concentration of each was 0.375 mM.

Figure 7

Rescue of erg26 mutant by GdnHCl and sensitivity of WT cells and other erg mutants points to GdnHCl inhibition of ergosterol biosynthesis. (A) TS mutants rescued by GdnHCl identified by screening large collections. Dilution series of cells of the indicated genotypes on rich medium (YPD) with or without GdnHCl at the indicated temperatures (“RT,” room temperature). The pga1-ts cells were grown on a different part of the same plate as the others; the image was cropped and moved to preserve space. Strains were: “WT,” BY4741; “act1-105,” CBY07857; “orc2-1,” CBY08224; “stu1-5,” CBY08183; “stu1-8,” CBY09448; “hsp10-ts,” CBY10708; “erg26-1,” CBY11254; “mcm3-1,” RSY723; “pga1-ts,” H06474. RSY723 is the original mcm3-1 strain in which this mutant allele was first studied (Lei et al. 2002). (B) Schematic of the budding yeast ergosterol biosynthesis pathway with single reactions indicated as arrows, and labels for enzymes represented by mutants whose response to GdnHCl we tested, or, in italics, mutants known to suppress the TS defect of erg26-1 mutants. Red asterisks indicate steps at which defects generate toxic intermediates. (C) As in (A) but with the strains of the indicated genotypes and, where indicated, rich medium was supplemented with additional nutrients for which these strains are auxotrophic (uracil, histidine, leucine, and methionine). Strains were: “WT,” BY4741; “erg10-1,” CBY09883; “erg6Δ,” H06534; “erg8-1,” CBY08334; “erg11-td,” CBY04662.

Figure 8

GdnHCl induces proteostatic stress and synergizes with proteasomal inhibition to rescue orc2-1. (A) Dilution series of BY4741 (“WT”) and CBY08187 (“pre1-1”) plated on rich (YPD) medium with or without 3 mM GdnHCl and incubated at room temperature or 37˚. (B) Cells from liquid rich (YPD) cultures of the indicated strains (“WT,” BY4741; “pre1-1,” CBY08187; “orc2-1,” CBY08224) were plated to medium containing Pro as the nitrogen source and 0.003% SDS to allow uptake of MG132 (Liu et al. 2007). Immediately after plating, paper discs were placed on the plate surface and 10 µl of the indicated chemicals were spotted onto the discs: 100% DMSO, 40 mM MG132 in DMSO, or 5M GdnHCl in water. Plates were then incubated at 37˚ prior to imaging. The square image at right is a close-up of the MG132 disc from an independent experiment with no GdnHCl-soaked disc on the plate, ruling out any contribution from GdnHCl to the observed rescue. The sprinkling of orc2-1 colonies dispersed across the entire plate reflects non-TS derivatives that arise spontaneously at 37˚ independently of added chemicals and which we have not characterized further. (C) Induction of the Hsf1-dependent cytosolic stress response measured using a strain (YJW1741) expressing RFP (mKate) from the constitutive TEF2 promoter and GFP from an artificial Hsf1-inducible promoter containing four heat shock elements (4xHSE). Cells (23 per sample) were imaged following overnight growth at RT with or without 3 mM GdnHCl. Error bars, mean ± SEM. “A.U.,” arbitrary units.

Figure 9

Chemical rescue of TS mutants by DMSO. Dilution series of cells of the indicated genotypes on rich medium (YPD) with or without 5% DMSO incubated at the indicated temperatures (“RT,” room temperature). Strains were: “WT,” BY4741; “kar2-159,” CBY07833; “cdc23-4,” CBY06436; “nop4-3,” CBY10671; “lcb1-10,” CBY10175; “lcb1-2,” CBY10161; “lcb1-5,” CBY10156; “arp3-G15C,” CBY08235; “arp2-14,” CBY04958; “stu1-5,” CBY08183; “stu1-6,” CBY10325; “smc3-1,” CBY08012; “tub4ΔDSYLD,” CBY11146.

Figure 10

Chemical rescue of TS mutants by TMAO. (A) Dilution series of cells of the indicated genotypes on rich medium (YPD) with or without 0.5 M TMAO incubated at the indicated temperatures (“RT,” room temperature). Strains were: “WT,” BY4741; “dam1-1,” CBY07974; “dam1-9,” CBY07966; “duo1-2,” CBY07958; “pkc1-1,” CBY09957; “pkc1-2,” CBY09962; “pck1-3,” CBY09972; “pkc1-4,” CBY09967. The asterisk indicates where, due to a technical error, the undiluted pkc1-4 sample was not applied to the plate. (B) As in (A) but with different mutants and 1 M sorbitol instead of TMAO. Strains were: “WT,” BY4741; “cdc10-1,” CBY06417; “cdc10-2,” CBY06420; “cdc12-1,” CBY05110; “duo1-2,” CBY07958; “dam1-1,” CBY07974; “dam1-9,” CBY07966; “pkc1-1,” CBY09957. (C) As in (A) but with different mutants, 0.5 M TMAO, 3 mM GdnHCl and/or 5% DMSO, and all plates were incubated at 37˚. Strains were: “WT,” BY4741; “cdc10-2,” CBY06420; “arp3-G15C,” CBY08235; “kar2-159,” CBY07833; “smc3-1,” CBY08012; “lcb1-5,” CBY10156; “tub4ΔDSYLD,” CBY11146. (D) BY4741 was streaked on rich (YPD) medium with or without 600 mM urea or 300 mM TMAO and incubated at 37˚ for 4 days prior to imaging. (E) Chemical structures of small molecules were used in this study.