Yeast galactose genetic switch

513

Epigenetics of the yeast galactose genetic switch

PAIKE JAYADEVA BHAT and REVATHI S IYER
Laboratory of Molecular Genetics, School of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Mumbai 400 076, Inidia (Fax, +91-22 25723480; Email, [email protected]) The transcriptional activation of enzymes involved in galactose utilization (GAL genes) in Saccharomyces cerevisiae is regulated by a complex interplay between three regulatory proteins encoded by GAL4 (transcriptional activator), GAL3 (signal transducer) and GAL80 (repressor). The relative concentrations of the signal transducer and the repressor are maintained by autoregulation. Cells disabled for autoregulation exhibit phenotypes distinctly different from that of the wild type cells, enabling us to explore the biological signicance of autoregulation. The redundancy in signal transduction due to the presence of GAL1 (alternate signal transducer) also makes it a suitable model to understand the phenomenon of epigenetics. In this article we review some of the recent attempts made to understand the importance of epigenetics in the establishment of cellular and transcriptional memory.
[Bhat P J and Iyer R S 2009 Epigenetics of the yeast galactose genetic switch; J. Biosci. 34 513522]

1.

Introduction

A long standing problem in biology has been the understanding of molecular interactions occurring with in the connes of time, space and concentrations that eventually lead to an animate system with unique properties. Historically, biologists have obtained valuable insights into this fundamental problem by studying the ability of single celled organisms such as bacteria and yeast to grow under varying environmental conditions. However, such experimental setups do not give us a perspective of how cellular populations adapt to dynamically changing environmental conditions. Unlike the cells of higher organisms, single celled organisms are exposed to the environment, the composition of which uctuates in an unpredictable manner. Thus, the evolutionary success of a unicellular organism such as yeast depends on its single minded approach of adapting to new conditions at the population level and not at the level of the single cell. Normally, all the members of a clonal population are expected to respond to changes in the environment in an identical fashion. Such a stereotypical response is clearly disadvantageous, in that the population as a whole can be wiped out if the genetic program is not equipped to adapt to

a new environment. How do genetically identical cells exist in phenotypically different states? In this review, we discuss how the genome and cytoplasmic factors collaborate to bring about myriad phenotypic outputs from a given genotype in the light of our recent understanding of the galactose genetic switch of Saccharomyces cerevisiae. 2. Working of the galactose genetic switch S. cerevisiae responds to galactose by activating the coordinated transcription of a family of genes, henceforth referred to as GAL genes (gure 1). Glucose (repressing carbon source) represses the ability of galactose to activate the transcription of GAL genes through multiple mechanisms. Carbon sources such as rafnose and glycerol (non-inducing, non-repressing carbon source) neither induce nor repress the ability of galactose to activate the GAL genes. This hierarchical utilization of carbon is tightly regulated at the genetic level (Bhat and Murthy 2001; Traven et al. 2006). The interplay between the proteins encoded by GAL3, GAL80 and GAL4 determines the ON/OFF status of the GAL switch (Traven et al. 2006). Gal4p binds the upstream activating sequences of GAL genes (UASg) but is unable to

Keywords. Autoregulation; epigenetics; GAL genes; stochasticity; transcriptional memory http://www.ias.ac.in/jbiosci J. Biosci. 34(4), October 2009, 513522, Indian Academy of Sciences 2009 J. Biosci. 34(4), October 513

514

Paike Jayadeva Bhat and Revathi S Iyer

Figure 1.

Galactose metabolic pathway.

activate the transcription in the absence of galactose. This is because, Gal80p, a nucleocytoplasmic shuttle protein, binds to the transcriptional activation domain of Gal4p. Upon interaction with galactose, Gal3p sequesters Gal80p in the cytoplasm, thereby relieving Gal4p to activate transcription. Two classes of GAL genes are distinguished based on the number of UASg elements present on the promoters. GAL genes with two UASg in their promoter have no detectable expression in the absence of galactose (but in presence of non-inducing and non-repressing carbon source) but are induced to very high levels in presence of galactose. The high induction of such genes is due to the co-operative binding of Gal4p to the adjacent sites, while the tight repression is due to the Gal80p dimer-dimer interaction occurring between the two adjacently located Gal4p dimmers (Melcher and Xu 2001). On the other hand, genes with one UASg have a detectable basal expression because of the lack of Gal80p dimer-dimer interaction. These genes are weakly or moderately induced because of the absence of co-operativity of Gal4p binding. 2.1 Autoregulation During cell division in a non-inducing, non-repressing carbon source such as ethanol, the relative concentrations of Gal3p, Gal80p and Gal4p in the daughter cells should be maintained such that the switch is OFF. If for some reason, daughter cells receive disproportionate amounts of these proteins, the
J. Biosci. 34(4), October 2009

switch can be fortuitously turned ON or OFF. For example, if the relative concentration of Gal80p, the repressor, is too low, then the transcription of GAL genes would get activated even in the absence of galactose. The switch should also have built in mechanisms to respond optimally to varying concentrations of galactose for its efcient utilization. To impart these features to the switch, expression of GAL2, GAL3 and GAL80 is autoregulated (gure 2). If a regulatory gene product regulates its own expression in addition to the expression of other genes normally under its control, then the gene under consideration is said to be autoregulated (Goldberg 1974). Autoregulation is brought about by positive or negative feedback circuits. Upon addition of galactose, expression of GAL2 encoded permease increases from undetectable to several hundred folds, which in turn increases the rate of uptake of galactose thus setting up a positive feedback loop (Hawkins and Smolke 2006). Among the regulatory genes, the expression of GAL3, which is required to initiate the induction cascade, is increased by approximately ten-fold in response to galactose, thus setting up a positive feed back loop. However, what is unexpected is that the expression of GAL80, the repressor, also increases ten-fold in response to galactose. To obtain insights into this observation, an in silico model of GAL genetic switch that predicts many of the experimentally observed results was developed (Verma et al. 2003). According to this, the concentration of Gal4p, Gal80p and Gal3p was estimated to be 0.005, 0.05 and

Yeast galactose genetic switch 0.25 M respectively in cells growing in non-inducing and non-repressing growth conditions. Under these intracellular concentrations, the GAL switch is not turned ON. Once the switch is turned on by the addition of galactose, expression of all the GAL structural genes increases as a function of time and reaches a steady state in two to three hours. Simultaneously, Gal3p concentration increases from the basal 0.25 M to 3.2 M and Gal80p concentration rises from 0.05 to 0.6 M. This increase in the concentration of Gal3p and Gal80p occurs through the same switch that they regulate. The obvious question then is what would happen if the autoregulatory loop is disabled. In silico studies indicated that disabling either GAL80 or GAL3 from auto-regulation resulted in a non-functional switch (Verma et al. 2003). However, disabling GAL3 as well as GAL80 from the autoregulatory loop resulted in an induction pattern similar to that of the wild type. If this is true, then why did yeast evolve the auto-regulatory loop of GAL3 and GAL80?

515

This was addressed by monitoring transcriptional induction in single cells where the autoregulation of GAL80 or GAL3 was disabled separately (Ramsey et al. 2006). In these strains, GAL80 and GAL3 were expressed from CYC1 promoter which results in expression comparable to that of the wild type strain under non-inducing conditions, but devoid of autoregulation. As compared to the wild type, the induction kinetics of GAL gene expression was initially delayed in cells disabled for GAL3 or GAL80 auto regulation, but reached a steady state value comparable to the wild type. This difference is due to an increased cell to cell heterogeneity in GAL gene expression. That is, the members of the wild type cell population did not differ very much from one another in terms of GAL gene induction, as compared to the cells disabled for the autogenous regulation. Does this initial delay in the disabled mutant in any way result in a growth disadvantage? Simulation studies showed that wild type would grow faster by 20% as compared to the mutant

Figure 2. Genetic circuitry of GAL genetic switch. Thick arrows indicate the interactions leading to induction while the thin arrows represent the reverse. In a cell that lacks Gal3p, Gal1p functions as the degenerate signal transducer. Gal1p mediated induction shows delayed induction kinetics (see text for details). J. Biosci. 34(4), October 2009

516

Paike Jayadeva Bhat and Revathi S Iyer period. If it did, then determine whether autoregulation played any role in this memory. It was observed that the expression of GAL1, as monitored in single cells at the end of 27 h, was independent of galactose or rafnose history, if galactose concentration during the 27 h growth period was below 0.012% or above 0.035%. However, within a concentration ranging from 0.012% to 0.35%, GAL1 expression was determined by the history of prior exposure (Acar et al. 2005). Like wild type cells, gal2 mutant cells continued to respond in binary fashion, except that the difference in expression between the two states is reduced. However, cells expressing Gal3p constitutively (that is, cells disabled for GAL3 feed back loop) did not exhibit the bimodal expression pattern, indicating that the cells were unable to remember the previous growth conditions. On the other hand, if GAL80 was expressed constitutively at the same level as that of the wild type (that is, the autoregulation of Gal80 is disabled), the memory persisted over a wide range of galactose concentrations. In this case, the memory was more pronounced than that of the wild type cells. However,

when the cells are challenged with alternating inducing and non-inducing carbon source. Thus, the autoregulatory loops seem to reduce the heterogeneity in expression of GAL genes within the members of clonal populations enabling the population as a whole to rapidly respond to the inducer. That is, autoregulatory loops force the individual cells of a clonal population to be as physiologically close as possible and remain fully committed to the job at hand. The following experimental paradigm was used to determine the role of the feed back loops of GAL3 and GAL80. Cells were separately disabled for GAL3 and GAL80 autoregulation, pregrown in rafnose (rafnose history) alone or rafnose plus 2% galactose (galactose history) for 12 h (gure3). Subsequently these cells were exposed to different concentrations of galactose ranging from 0.00 to 0.5% for a further period of 27 h. At the end of this growth period, GFP expression driven by the GAL1 promoter was monitored. The idea was to determine whether the history of prior exposure of the cells to galactose (Rafnose or galactose history) alter their response to subsequent exposure to galactose during the 27 h time

Figure 3. Schematic representation of the experimental paradigm for demonstrating cellular memory. (A) Cells were pre-grown in rafnose alone (rafnose history -) and rafnose plus galactose (galactose history ---) respectively for 12 h and subsequently grown for 27 h at different concentrations of galactose as the sole carbon source. (B) The GAL1 promoter activity in single cells at the end of 27 h of growth at the indicated galactose concentrations was determined using the expression of uorescent protein as a reporter. Although each strain was grown for 27 h at different galactose concentrations, results at one galactose concentration is indicated. J. Biosci. 34(4), October 2009

Yeast galactose genetic switch if the constitutive expression of GAL80 was reduced lower than that of the wild type, the population was predominantly represented by two types of cells (bimodal distribution), one with low and the other with high expression of GAL genes. If these two cell types were sorted and allowed to re-grow separately under the same conditions as before sorting, they relaxed back to the bimodal distribution. The time of relaxation depended on the concentration of galactose and GAL80 expression. An independent study demonstrated that Gal2p positive feed back loop is required for switch like response. If Gal2p is expressed from a constitutive promoter, the induction response as a function of galactose concentration is linear. Besides providing a switch like response, GAL2 feed back loop also confers binary or all or none type of induction while a gal2 knock out strain shows graded response (Hawkins and Smolke 2006). This graded response of a gal2 mutant strain is opposite to that observed by Acar et al. (2005). What could be the reason for this apparent discrepancy? The differences could arise because of the differences in the carbon sources used for pre-culturing the cells prior to monitoring the expression. (See conclusion for more details.) 3. Epigenetic regulation of signal transduction

517

M. In 0.4% of the cells it would be more than 0.075 M and these cells are not expected to express Gal1p at all in the absence of galactose. However, in the remaining 0.4%, Gal80p would be less than 0.025 M resulting in very low expression even in the absence of galactose. These cells, if exposed to galactose can initiate an induction cascade (gure 4). In fact quantitative analysis has supported this possibility. This is a classic example of epigenetics, wherein the change in the concentration of regulatory molecule initiates a signal transduction pathway by altering the concentration of a cytoplasmic protein. 3.1 Transcriptional and cellular memory

As early as 1948 it was observed that a gal3 mutant strain responds to galactose only after a lag of 48 to 76 h (Long term adaptation or LTA) unlike a wild type strain which responds to galactose with in few minutes (Winge and Roberts 1948). In the 1990s it became clear that this is because GAL1 encoded galactokinase, the rst enzyme in the galactose metabolism, has GAL3 like signal transduction activity. Further studies indicated that the signal transduction function of GAL1 encoded galactokinase is independent of its kinase activity indicating that galactokinase is a bifunctional protein (Bhat and Murthy 2001). However, signal transducing function of GAL1 in a strain defective for GAL3 expression remained a paradox. Initially, it was thought that gal3 cells, in the presence of galactose, accumulate sufcient Gal1p because of leaky expression. LTA is observed because the Gal1p dependent signal transduction is inefcient resulting in delayed expression. This idea did not receive experimental support and was eventually abandoned. Recently, it has been suggested that the stochastic variation in Gal80p, the repressor, could be responsible for the Gal1p dependent long term adaptation (Bhat and Venkatesh 2005). The distribution of Gal80p in gal3 cells growing in a neutral carbon source would vary from cell to cell as dictated by the statistical uctuation inherent to the system. According to this the Gal80p concentration could vary between 0.5 to 1.5 times the normal concentrations in 99.2% of the cells. That is, given the normal concentration of Gal80p is 0.05 M, the variation is around 0.025 to 0.075

We dene transcriptional memory as A gene that once induced for transcription, responds readily to subsequent induction, even in the absence of inducer for several generations after the rst induction. In the recent past, galactose genetic switch has been used to understand the molecular basis of transcriptional memory. The questions asked were what is the molecular mechanism and for how many generations can this transcriptional memory last? It has been recently observed that, GAL1 is recruited to the nuclear periphery upon activation of transcription. This recruitment appears to be mediated by the NUP2 gene product, a member of the nuclear pore complex (Brickner et al. 2007). The peripheral localization of GAL1 as a result of transcription activation by galactose persists for at least 6 generations of growth in glucose, a repressing medium. This peripheral localization promotes transcriptional activation on subsequent exposure to galactose. Histone variant H2A.Z is essential for the recruitment of GAL1 to the nuclear periphery and subsequent activation. Accordingly cells disrupted for H2A.Z do not exhibit transcriptional memory. In an independent study, this transcriptional memory is ascribed to the ATP dependent chromatin remodelling enzyme SWI2 (Kundu et al. 2007). Cells pre-cultured in rafnose, a neutral carbon source, induced GAL1 transcription with in 20 min of galactose exposure. After an hour of growth in galactose, cells were transferred to 2% glucose for one hour. Addition of galactose to these cells after removing glucose from the medium induced GAL1 with in ve minutes. That is, the cells exposed to galactose before glucose exposure retained their memory of galactose exposure. However, if glucose exposure is continued beyond 6 h this transcriptional memory is lost. It is demonstrated that this phenomenon is dependent on the kinase activity of Swi2p. The results of a more recent study (Zacharioudakis et al. 2007) are incompatible with the conclusions drawn from the above studies. In order to understand this, the re-induction kinetics of the GAL genes in response to galactose was monitored in cells initially grown in galactose and subsequently in glucose for different period of time
J. Biosci. 34(4), October 2009

518

Paike Jayadeva Bhat and Revathi S Iyer

Figure 4. protein.

Possible mechanism of LTA of gal3 cells. The shade represents the cell to cell difference in the concentration of Gal80

(gure 5A). As mentioned earlier, GAL1 and GAL3 are the two redundant signal transducers that transmit the signal from galactose to the Gal4p dependent transcriptional inducer. In cells grown in galactose, Gal1p is induced several hundred fold as compared to Gal3p. Using single cell analysis, it was demonstrated that, reinduction kinetics of GAL switch was faster in cells pre exposed to galactose even after six generation of growth in glucose, a repressing medium. This transcriptional memory persisted in cells disrupted for GAL3 but not GAL1 (Zacharioudakis et al. 2007). In an elegant experiment the authors rule out the role of peri-nuclear topology or chromatin modication in the transmittance of transcriptional memory. Here, the authors looked for a possible role for upstream cytoplasmic elements in the galactose signalling pathway responsible for transcriptional memory. Wild type haploid yeast cell forms heterokaryon (cells with two independent nuclei in the same cytoplasm) when mated with a kar1-1 haploid yeast cell of opposite mating type. The kar1-1 mutation is dominant and prevents the fusion of nuclei. Due to this, the nuclei exist as separate entities after diploidization. Thus, it is possible
J. Biosci. 34(4), October 2009

to generate a heterokaryon from haploid cells grown under different conditions. It was observed that induction was rapid in heterokaryons formed between the native GAL1GFP cells and galactose pre-cultured GAL1 kar1-1 cells (gure 5B). Under similar conditions, the GAL1-GFP induction was slower when heterokaryon was formed between native GAL1-GFP cells and gal1 kar1-1 cells. As expected, the induction of GAL1-GFP was much slower in heterokaryon formed between native GAL1-GFP cells and glucose grown GAL1 kar1-1 haploid cells. These results are difcult to explain based on the chromatin modication or peri-nuclear topology, since the GAL1-GFP cassette has not experienced a prior transcriptional induction. The mere presence of nuclei carrying this cassette in a cytoplasmic environment that had been previously exposed to galactose is sufcient to implement a faster induction response to subsequent galactose induction. Interestingly, exposure of cells to glucose beyond 6 generation erases this memory and this effect is most probably due to the loss of GAL1 gene product by simple dilution. At a time when the nature of gene itself was not clear, Spiegelman and his co-workers tried to address the question

Yeast galactose genetic switch

519

Figure 5. (A) Schematic representation of the experimental scheme for evaluating transcriptional memory. Cells were pre-grown in galactose and then in glucose (1) or glucose alone (2). Induction kinetics of GAL gene expression in these two cultures was determined in response to addition of galactose. (B) Schematic representation of heterokaryon assay for demonstrating the role of GAL1 protein in transcriptional memory.

of long term adaptation of gal3 cells. At that time, the only known regulatory gene of GAL genetic switch was GAL3. They tried to determine the number of generations a galactose adapted gal3 cell growing in a medium containing glucose, retains its ability to rapidly induce the switch in response to subsequent activation by galactose (reviewed in Spiegelman 1951). Recall that a gal3 cell normally does not respond to galactose immediately, but adapts to galactose, due to the accumulation of Gal1p, the alternate signal transducer. Obviously, Spiegelman and his coworkers were unaware of the redundant signal transduction

through GAL1 encoded galactokinase. However, they tried to calculate the number of generations necessary before the accumulated Gal1p is diluted out such that the re-induction kinetics returns to it original state, in a galactose adapted gal3 cell. They followed a simple experimental regimen, where in gal3 cells adapted to galactose (we now know that these cells have sufcient GAL1 encoded galactokinase) were transferred to glucose medium. A bud formed under these conditions was removed and transferred to galactose containing medium. It was noted whether the bud responded rapidly or slowly by observing the growth. It was observed
J. Biosci. 34(4), October 2009

520

Paike Jayadeva Bhat and Revathi S Iyer is thought to be resolved by gene duplication followed by sub-functionalization. After S. cerevisiae acquired GAL3 and GAL1 (the duplicated versions of the ancestral bi-functional GAL1 gene that had galactokinase as well as signal transduction activity), they were free to evolve. During this evolution, GAL3 lost the kinase activity completely but retained the signal transduction function. On the other hand, GAL1 retained partial signal transduction function along with the full galactokinase activity. Therefore, the sub-functionalization was not complete with respect to signal transduction function. On the other hand, subfunctionalization with respect to the expression pattern of GAL1 and GAL3 was complete. That is, GAL1 acquired high inducible expression but undetectable basal expression (this is unlike K. lactis GAL1, which is more like the ancestral GAL1). In contrast, GAL3 retained basal expression but acquired weak induction. It is demonstrated that the tness of K. lactis is enhanced on replacing the GAL1 promoter with that of S. cerevisiae. Conversely, the tness of S. cerevisiae was reduced when the GAL1 promoter was replaced with that of K. lactis. Swapping the promoter modules between GAL1 and GAL3 in S. cerevisiae resulted in lowered tness (Hittinger and Carroll 2007). These results indicate that the sub-functionlization which occurred between GAL3 and GAL1 in S.cerevisiae was not fortuitous, but had a denitive evolutionary purpose. The difference in the expression pattern of GAL1 and GAL3 is dependent on the number and disposition of UASg elements (see gure 6 for details).

that up to 6th or 7th bud isolated sequentially from the same gal3 mother cell responded to galactose very rapidly. After that, that is 8th bud onwards, lost this ability. Statistical analysis of this result indicated that a fully galactose adapted gal3 yeast cell has approximately 200 factors, which get diluted to reach a value below the threshold required for rapid induction on subsequent exposure to glucose for six to seven generations. We now know that the factors he was referring to is the GAL1 encoded galactokinase and in fact he was trying to understand the role of cytoplasm in the regulation of genetic switch. 4. Evolution of GAL genetic switch

The evolutionary signicance of redundant cytoplasmic signal transducers can be appreciated in the backdrop of GAL genetic switch of Kluveromyces lactis, a close relative of S. cerevisiae (Rubio-Texeria 2006). In K. lactis, GAL3 orthologue is absent. Instead, the GAL1 encoded gene performs the signal transduction as well as the kinase function. Accordingly, the GAL1 of K. lactis differs from that of S. cerevisiae in two respects. First, K. lactis GAL1 has a basal expression in the absence of galactose and second it is moderately induced by galactose. Recently, using a sensitive and direct competition assay, Hiittinger and Carrol have analysed the evolutionary divergence of GAL1 and GAL3 from a bifunctional ancestor (Hittinger and Carroll 2007). Genes with bifunctionality face adaptive conict in that optimization of one function may compromise or adversely affect the other function. Such adaptive conict

Figure 6. Schematic representation of duplication of ancestral GAL1 into GAL1 and GAL3. The black boxes represent the UASg elements. In ancestral GAL1, the UASg elements are on the opposite phases of the DNA. In GAL1 of S. cerevisiae, the UASg elements are on the same phase of the DNA. Due to this, Gal4 protein binds in a co-operative fashion resulting in high inducible expression. The expression of GAL1 under non-inducing conditions is undetectable because of GAL80 dimer-dimer interaction. Since GAL3 has only one UASg element, neither Gal4 protein binds co-operatively nor is the GAL80 dimer-dimer interaction stabilised. This results in detectable basal expression with weak induction. J. Biosci. 34(4), October 2009

Yeast galactose genetic switch 5. Conclusions

521

The GAL genetic switch is an excellent paradigm for unraveling how evolutionary forces shape and ne tune the regulation of biological process. Although the role of feed back loops in the GAL gene regulation has been studied in great detail, many questions remain unanswered. For example, Biggar and Crabtree (2001) demonstrated that the wild type cells show graded but not binary response to increasing concentration of galactose. This observation is in contrast to the binary response reported for wild type cells in other studies (Acar et al. 2005; Hawkins and Smolke 2006). Biggar and Crabtree (2001) pre-cultured the cells in glucose prior to induction with galactose. The difference in the response of wild type cells to induction, observed in different studies seems to indicate that growth history is an important determinant of whether the subsequent induction to galactose is graded or binary. The functional non-equivalence between members of a population of genetically identical single celled organisms is probably akin to the division of labour observed between cells of higher organisms. While many of the attributes of the GAL genetic switch provide valuable insights, their signicance can be fathomed only when the evolutionary context is deciphered. The potential to exhibit wide ranging phenotypic outputs appears to be much more than what the hard wired genetic program was originally thought to offer. It is pertinent to recall that elegant studies conducted in E. coli as early as 1950s, indicated that the autocatalytic induction of lac permease is the key for the heritable nature of the binary or bimodal response (Novick and Weiner 1957; Cohn and Horibata 1959a,b). Thus, organisms have evolved a plethora of molecular mechanisms to maintain and propagate different expression states between cells of a population within the constraints of a genetic program. The molecular mechanisms of epigenetic changes (Ptashne 2008) need to be explored in detail to understand the biological outcome of a given genetic program. Acknowledgements Thanks are due to the Department of Science and Technology, New Delhi for the nancial assistance provided to PJB (SR/ SO/BB-25/2004) and RSI (SR/WOS-A/LS-325/2004). References
Acar M, Becskei A and van Oudenaarden A 2005 Enhancement of cellular memory by reducing stochastic transition; Nature (London) 435 228231 Bhat P J and Murthy T V S 2001 Transcriptional control of the GAL/MEL regulon of yeast Saccharomyces cerevisiae: Mechanism of galactose mediated signal transduction; Mol. Micrbiol. 40 10591066

Bhat P J and Venkatesh K V 2005 Stochastic variation in the concentration of repressor activates GAL genetic switch: Implication in evolution of regulatory network; FEBS Lett. 579 597603 Biggar S R and Crabtree G R 2001 Cell signaling can direct either binary or graded transcriptional response; EMBO J. 20 31673176 Brickner D G, Cajigas I, Foundefe-Mittendorf Y, Ahmed S, Lee P, Widom J and Brickner J H 2007 H2A.Zmediated localization of genes at the nuclear periphery confers epigenetic memory of previous transcriptional state; PLoS 5 e81 Cohn M and Horibata K 1959a Inhibition by glucose of the induced synthesis of the -Galactosidase-Enzyme system of Escherichia coli. Analysis of maintenance; J. Bacteriol. 78 601612 Cohn M and Horibata K 1959b Analysis of the differentiation and of the heterogeneity within a population of Escherichia coli undergoing induced -Galactosidase synthesis; J. Bacteriol. 78 613623 Goldberg R F 1974 Autogenous regulation of gene expression; Science 183 810816 Hawkins M K and Smolke C D 2006 The regulatory roles of galactose permease and Kinase in the induction response of the GAL network in Saccharomyces cerevisiae; J. Biol. Chem. 281 1348513492 Hittinger C T and Carroll S B 2007 Gene duplication and the adaptive evolution of a classic genetic switch; Nature (London) 449 677682 Kundu S, Horn P J and Peterson C L 2007 SWI/SNF is required for the transcriptional memory at the yeast GAL gene cluster; Gene. Dev. 21 9971004 Melcher K and Xu H E 2001 Gal80-Gal80 interaction on adjacent Gal4p binding sites is required for complete GAL gene repression; EMBO J. 20 841851 Novick A and Weiner M 1957 Enzyme induction as an all or non phenomenon; Proc. Natl. Acad. Sci. USA 43 553566 Ptashne M 2007 On the use of the word Epigenetics; Curr. Biol. 17 R233R236 Ramsey S A, Smith J J, Orell D, Merille M, Peterson T W, de Atauri P, Bolouri H and Aitchison J D 2006 Dual feed back in the GAL regulon suppresses cellular heterogeneity in the yeast; Nat. Genet. 38 10821087 Rubio-Texeria M 2006 A comparative analysis of GAL genetic switch between not-s-distant cousins: Saccharomyces cerevsiae and Kluyveromyces lactis; FEMS Yeast Res. 5 11151128 Spiegelman S 1951 The particulate transmission of enzyme forming capacity in yeast; Cold Spring Harbor Symp. Quant. Biol. 16 8798 Traven A, Jelicic B and Sopta M 2006 Yeast GAL4: A transcriptional paradigm revisited; EMBO Reports 7 496499 Verma M, Bhat P J and Venkatesh K V 2003 Quantitative analysis of GAL genetic switch of Saccharomyces cerevisiae reveals that nucleocytoplasmic shuttling of Gal80p results in a highly sensitive response to galactose; J. Biol. Chem. 278 4876448769 J. Biosci. 34(4), October 2009

522

Paike Jayadeva Bhat and Revathi S Iyer

Zacharioudakis I, Gligoris T and Tzamarias D 2007 A yeast catabolic enzyme controls transcriptional memory; Curr. Biol. 17 20412046

Winge O and Roberts C 1948 Inheritance of enzymatic characters in yeast and the phenomenon of long term adaptation; C. R. Trav. Lab. Carlsberg. Ser. Physiol. 24 264315

ePublication: 18 June 2009

J. Biosci. 34(4), October 2009