Saccharomyces cerevisiae senses extracellular amino acids using two members of the family of amino acid transporters, Gap1 or Ssy1; aspects of the latter are reviewed here. Despite resemblance with bona fide transporters, Ssy1 appears unable to facilitate transport. Exposure of yeast to amino acids results in Ssy1-dependent transcriptional induction of several genes, in particular some encoding amino acid transporters. Amino acids differ strongly in their potency, leucine being the most potent one known. Using a selection system in which potassium uptake was made dependent on amino acid signalling, our laboratory has obtained and described gain-of-function mutations in SSY1. Some alleles conferred inducer-independent signalling; others increased apparent affinity for inducers. These results revealed that amino acid transport is not required for signalling and support the notion that sensing by Ssy1 occurs via its direct interaction with extracellular amino acids. Current work includes development of quantitative assays of sensing. We use the finding by Per Ljungdahl's laboratory that the signal transduction from Ssy1 involves proteolytic removal of an inhibitory part of the transcriptional activator Stp1. Protein-A Z-domain fused to the C-terminus of Stp1 and Western analysis using antibody against horseradish peroxidase allow quantification of sensing.
- amino acid
- nutrient sensing
- Saccharomyces cerevisiae
- transporter-like sensor
Free-living microorganisms cope with changes in the extracellular concentrations of ions, minerals and nutrients by controlling the rates of transport of these molecules across the plasma membrane. Such adaptability often involves responses to changes in the intracellular concentrations of the relevant solute. However, microbes have also evolved the ability to sense the extracellular concentrations of certain nutrients.
Amino acid sensing
In Saccharomyces cerevisiae, distinct mechanisms have evolved to sense the extracellular presence of amino acids. In one sensing system, cells in a fermentable growth medium, but starved for nitrogen, react to the addition of amino acids by a number of responses, including repression of stress-inducible genes and degradation of trehalose and glycogen. In this system the added amino acid interacts  with the general amino acid permease Gap1, which, in a process dependent on the kinase Sch9, leads to activation of protein kinase A , and subsequently to activation of trehalase and the many other responses. In this paper, however, we consider a different sensing system, which is active in non-starved cells: micromolar amounts of amino acids can be sufficient to increase arginase activities  and peptide uptake  and to induce transcription of the amino acid permease gene BAP2  and the peptide transporter gene PTR2 .
The SPS (Ssy1/Ptr3/Ssy5) system
Key components of this signal transduction pathway, Ssy1, Ptr3, Ssy5 and Stp1, were found by a screen for mutants that failed to take up branched-chain amino acids [7,8]. As first noted by Didion et al. , SSY1 is required for normal amino acid-dependent transcriptional induction of several target permease genes. As reviewed by Boles and André  and described in a later paper , other genes are also affected. Components of the amino acid signalling pathway were independently discovered through other investigations, including screens for mutations that confer resistance to toxic dipeptides , amino acid analogues  or high concentrations of histidine  and by mutations that confer synthetic lethality with nutrient requirements .
Ssy1 was singled out as the sensor based largely on the observation that it appeared to be a plasma membrane protein evolutionarily related to the amino acid permeases in S. cerevisiae [7,16,17]. Thus, Ssy1 is hypothesized to initiate signalling through direct interaction with extracellular amino acids. Ssy1 interacts at the plasma membrane with Ptr3 and Ssy5, presumably forming a complex designated SPS for Ssy1/Ptr3/Ssy5 . Amino acid binding is likely to confer a conformational alteration in Ssy1 producing a signal that is transmitted, via Ptr3 and Ssy5, to target promoters through the action of transcription factors. It was early suggested that Stp1 is an important transcription factor in this system , and Andréasson and Ljungdahl [19,20] indeed found that it is directly involved in signalling. Several other transcription factors have roles in the expression of one or more of the target genes (see, e.g. ). The discovery that overexpression of the amino terminus of Ssy1 actually interferes with amino acid signalling [13,18] revealed that the amino terminus of Ssy1 plays an important role in an intracellular aspect of this signalling.
Transporter or sensor?
Transport of amino acids by bona fide permeases fails to elicit the activation of amino acid permease promoters in ssy1 cells and thus supports the Ssy1 sensor hypothesis [8,9,14,16]. However, the inability of other permeases to elicit this type of response could be explained if sensing demanded recognition of amino acids by proteins, such as Ptr3 or Ssy5, found in close association with Ssy1. Such a scenario has been described in neurons in which calcium signalling is mediated by only one of two species of voltage-gated calcium channels because calmodulin is closely associated only with the sensing channel .
Although Ssy1 is structurally related to amino acid permeases, several experiments suggest that Ssy1 itself is unable to transport amino acids efficiently. Thus, although citrulline elicits the signalling response in SSY1 gap1 cells, such cells are unable to take up enough citrulline as a nitrogen source to confer growth . Similarly, although SSY1 is required for maximal uptake of branched-chain amino acids, SSY1 overexpression fails to restore their uptake in cells deleted for the four genes known to encode branched-chain amino acid transporters . Nor does the expression of SSY1 restore uptake of tryptophan in cells deleted for the authentic tryptophan transporters . However, such tests cannot rule out the possibility that Ssy1 transports small quantities of the signalling molecule, and if any transport occurs, the sensor that actually recognizes the signalling amino acid would not have to be Ssy1, but a closely associated protein. Thus, a key question is: does signalling require uptake and subsequent intracellular presence of inducer?
SSY1 mutant constantly signalling without inducer
To critically address this question, Gaber et al.  sought to determine if Ssy1 could be mutated into a constitutive sensor, capable of activating the pathway during growth on medium lacking amino acids. They first developed a system in which growth of yeast cells is dependent on transcriptional induction of the amino acid-responsive AGP1 promoter, but independent of amino acid uptake. This was done by construction of a strain in which cell growth depended on induction expression of KAT1, encoding a heterologous potassium channel. Thus, KAT1 was placed under control of the AGP1 promoter and expressed in cells deleted for the two endogenous potassium transporter genes, TRK1 and TRK2. In the trk1Δtrk2Δ background, growth on synthetic medium is nearly abolished unless supplemented with high concentrations of potassium . Expression of Kat1 restores potassium uptake, permitting growth on normal yeast medium. Gaber et al.  then used this reporter system in a screen to identify mutations in Ssy1 that conferred constitutive signalling. One such mutant, Thr382Lys, was found by PCR mutagenesis of the SSY1 open reading frame (ORF) and isolation of a mutant plasmid able to constitutively activate the AGP1-KAT1 reporter to allow growth in the absence of inducer. Subsequent measurement of AGP1-lacZ reporter activity showed full activation of the AGP1 promoter in the mutant in the absence of inducer. The mutant gene (SSY1-102) was found to be dominant over the SSY1 wild-type gene and to require Ptr3 and Ssy5 for function, as expected for gain of a function that otherwise follows the normal signal pathway.
A hyper-responsive mutant sensor
Prompted by this suggestion that the residue in position 382 of Ssy1 might have particular importance for signalling, Gaber et al.  next asked if other phenotypes might result from having other residues in this position. Whereas most residues gave phenotypes not easily distinguishable from that of the wild type, and while Thr382Arg was constitutive like Thr382Lys, the two substitutions Thr382Leu and Thr382His were very interesting. In both mutants the AGP1-KAT1 reporter was able to respond to lower concentrations of inducer than in the wild type. In principle, this could reflect an altered amino acid binding site with higher affinity in the mutant forms of Ssy1. However, the potency of any amino acid tested was higher towards the mutant than towards the wild type by about the same factor, rather suggesting that the mutants were affected in an equilibrium between a signalling conformation and a non-signalling conformation of Ssy1. The Thr382His mutant was subjected to quantitative dose–response analysis using the AGP1-lacZ reporter and exhibited about 10-fold higher affinity for the inducer citrulline than did the wild type. The basal level of β-galactosidase (i.e. when inducer is absent) was somewhat increased in the mutant, a feature that is indeed expected if the equilibrium between a signalling conformation and a non-signalling conformation of Ssy1 is affected. The existence of SSY1 mutants with increased apparent affinity for inducers provides a strong support to the notion that Ssy1 interacts directly with extracellular amino acids to initiate signalling.
A more direct assay for sensing
In further detailed studies of the in vivo significance of ligand–receptor interactions in this system, we believe that dose–response analyses will continue to be important. A new option for a read-out of signalling has been provided by Andréasson and Ljungdahl , who found that proteolytic removal of an inhibitory N-terminal part of the transcriptional activator Stp1 is an important step in the Ssy1-dependent signalling. Provided with our Thr382Lys mutant in SSY1, they  furthermore showed that a significant amount of Stp1 is activated in this way in the mutant, already in the absence of inducing amino acid. However, this amount was still clearly a minor part of the total Stp1, although full processing took place when inducer was added. Together with the fact that AGP1-lacZ expression is maximal in the mutant, this indicates that processing of just a minor part of Stp1 is sufficient for full AGP1 promoter activity. In other words, the mutant behaviour is fully constitutive with the promoter read-out but only partially constitutive when the processing of Stp1 is considered. We therefore expect that monitoring Stp1 processing will be a measure of signalling with a larger dynamic range than has the AGP1-lacZ reporter. In addition, this type of assay is a priori more direct, reflecting more closely what goes on with Ssy1. For this reason we constructed a translational fusion (Stp1-ZZ) in which a doublet of the IgG-binding Z domain of the Staphylococcus aureus Protein A is fused to the C-terminus of Stp1 . As expected from the modular behaviour of Stp1 processing , the fusion protein expressed in yeast is processed by amino acid signalling. This was monitored by Western blotting using a mixture of horseradish peroxidase and polyclonal rabbit anti horseradish peroxidase. Figures 1(A) and 1(B) show the dose–response relationship for signalling by leucine in cells grown in ammonium- and glucose-based minimal medium. Comparison with the AGP1-lacZ response (Figure 1C) reveals several interesting features that we have found typical for these assays. (i) Both curves can be reasonably fitted to a hyperbolic relationship (sigmoid in the semi-logarithmic plot), consistent with reversible, effective binding of single leucine molecules. (ii) However, there is a tendency for the lacZ response to be steeper than the theoretical curve, and (iii) to have a lower relative basal level than the Stp1-ZZ response; both of these features would be consistent with the possibility that activated Stp1 binds to promoters as dimers. (iv) The apparent Kd is lowest when determined by the AGP1-lacZ assay (1 μM, as compared with 12 μM with the Stp1-ZZ assay), which can be understood in view of full AGP1 promoter activity already by modest Stp1 processing. (v) Scatter is in our hands lowest in the Stp1-ZZ assay. (vi) There is a tendency for a drop in AGP1-lacZ activity at the highest ligand concentrations; this is also seen with amino acids other than leucine and makes the determination of the apparent Kd dependent on how far up in concentration one includes data. In conclusion, we see good reasons, both a priori and a posteriori, for choosing the Stp1-ZZ assay for investigations of events at or close to Ssy1. On the other hand, more global investigations of the regulation of a target gene such as AGP1 are of course expediently approached by monitoring promoter activity (see, e.g. ).
We thank Mrs. Lisbeth F. Petersen for skilful technical assistance, Dr Anders Brandt and Dr Frank van Voorst for valuable advice in the development of the Protein-A-based detection system and Dr Claes Andréasson and Dr Per Ljungdahl for sharing their results before publication. This work was supported by the Carlsberg Foundation and by a grant from the National Science Foundation (to R.F.G.). We acknowledge the permission by the American Society for Microbiology to cite extended parts of the article by Gaber et al. .
Nutrient Sensing through the Plasma Membrane of Eukaryotic Cells: Focused Meeting held at the Royal Agricultural College, Cirencester, 25–29 September 2004. Organized by S. Shirazi-Beechey (Liverpool, U.K.), J. Thevelein and F. Stolz (Leuven, Belgium). Edited by S. Shirazi-Beechey. Sponsored by Flanders Interuniversity Institute for Biotechnology and Nestlé U.K. Ltd.
Abbreviations: SPS, Ssy1/Ptr3/Ssy5
- © 2005 The Biochemical Society