RESEARCH ARTICLE

Rlm1 mediates positive autoregulatory transcriptional feedbackthat is essential for Slt2-dependent gene expressionRaul Garcıa, Ana Belen Sanz, Jose Manuel Rodrıguez-Pen a, Cesar Nombela and Javier Arroyo*

ABSTRACTActivation of the yeast cell wall integrity (CWI) pathway induces anadaptive transcriptional programme that is largely dependent on thetranscription factor Rlm1 and the mitogen-activated protein kinase(MAPK) Slt2. Upon cell wall stress, the transcription factor Rlm1 isrecruited to the promoters of RLM1 and SLT2, and exerts positive-feedback mechanisms on the expression of both genes. Activation ofthe MAPK Slt2 by cell wall stress is not impaired in strains withindividual blockade of any of the two feedback pathways. Abrogationof the autoregulatory feedback mechanism on RLM1 severely affectsthe transcriptional response elicited by activation of the CWI pathway.In contrast, a positive trans-acting feedback mechanism exerted byRlm1 on SLT2 also regulates CWI output responses but to a lesserextent. Therefore, a complete CWI transcriptional response requiresnot only phosphorylation of Rlm1 by Slt2 but also concurrent SLT2-and RLM1-mediated positive-feedback mechanisms; sustainedpatterns of gene expression are mainly achieved by positiveautoregulatory circuits based on the transcriptional activation of Rlm1.

KEYWORDS: MAPK signaling, Positive feedback, Gene expression,Cell wall stress

INTRODUCTIONEukaryotic organisms, from yeast to mammals, regulate manycellular processes in response to environmental stimuli throughmitogen-activated protein kinase (MAPK) cascades. Activation ofMAPKmodules culminates in the phosphorylation and activation ofspecific transcription factors, leading to appropriate cellularresponses that allow cells to regulate key cellular events, such ascellular proliferation, differentiation, apoptosis or stress responses(Elion et al., 2005). The onset, amplitude and duration of theMAPKactivation combine to generate particular signaling profiles, leadingto specific output responses (Purvis and Lahav, 2013). Therefore,precise regulation of these pathways is required to modulate specificadaptive responses. Identifying the mechanisms that coordinatethese profiles is relevant to interpret and to intervene in the cellulartransmission process (English et al., 2015).Many regulatory mechanisms modulate signaling through

MAPK pathways in yeast (Chen and Thorner, 2005; Molinaet al., 2010; Zeke et al., 2009). These mechanisms include scaffoldproteins and docking interactions that contribute to providingsignaling specificity and the prevention of improper crosstalk(Flatauer et al., 2005; Saito and Posas, 2012; Witzel et al., 2012;Zalatan et al., 2012), protein phosphatases that directlydephosphorylate protein kinases to attenuate the response when

necessary (Martín et al., 2005), and positive- and negative-feedbackcircuits (FBLs), which are required to maintain a balanced output inthe stimuli response (Mobashir et al., 2014). Feedback mechanismsare mainly exerted through post-translational modifications, such asphosphorylation events performed by the MAPK on upstreamelements of the pathway (retrophosphorylation) (Feng and Davis,2000; Flotho et al., 2004; Hao et al., 2007; Kranz et al., 1994;Molina et al., 2010; Saito and Posas, 2012). Alternatively tophosphorylation regulatory mechanisms, regulatory circuits can bebased on the transcriptional regulation of their own elements, eitherinactivators (negative feedback) or activators (positive feedback)(Molina et al., 2010). The transcriptional induction of MAPKphosphatases by MAPK activation has been well documented as anegative-feedback mechanism of MAPK signaling in mammals(Caunt and Keyse, 2013; Keyse and Emslie, 1992). In yeast, theinduced expression of the phosphatases PTP2 and PTP3 afterosmotic shock also points to these phosphatases as part of negative-feedback loops (Jacoby et al., 1997), whereas activation of themating and filamentous growth pathways induces the expression ofthe transcription factor Ste12, indicating the existence of stimulus-specific positive-feedback mechanisms (Roberts et al., 2000;Zeitlinger et al., 2003).

The CWI pathway is the main system responsible for theorganisation and biosynthesis of the cell wall. This pathway isessential for the adaptation of the yeast stress conditions thatchallenge cell wall integrity (for a review see Levin, 2011).Membrane proteins Mid2 and Wsc1 act as the principal sensors ofthis pathway (Ketela et al., 1999; Rajavel et al., 1999). Under stressconditions, the protein kinase C, Pkc1, is activated as a consequenceof the activation of the small GTPase Rho1 by the guaninenucleotide exchange factor Rom2. Pkc1 triggers the sequentialphosphorylation of the MAPKKK Bck1, two redundant MAPKKs– Mkk1 and Mkk2 – and the MAPK Slt2 (also known as Mpk1)(Levin, 2011). The phosphorylated form of Slt2 in the conservedTxY motif (where x is any amino acid) leads to the activation of twotranscription factors: Rlm1 (Dodou and Treisman, 1997; Jung et al.,2002; Watanabe et al., 1997) and SCB-binding factor (Baetz et al.,2001).

CWI pathway activation leads to a transcriptional reprogrammingof the cell that has been characterised by global gene expressionprofiling (Boorsma et al., 2004; García et al., 2004, 2009; Jung andLevin, 1999; Lagorce et al., 2003; Reinoso-Martín et al., 2003;Roberts et al., 2000). This transcriptional response is largelydependent on the transcription factor Rlm1 and the MAPK Slt2(García et al., 2004, 2009; Jung and Levin, 1999; Sanz et al., 2012).Rlm1 is a MADS-box transcription factor, which shares homologyand in vitro DNA-binding specificity with mammalian MEF2orthologues. The transcriptional activity of Rlm1 is regulated bySlt2-mediated phosphorylation on residues Ser427 and Thr435 (Junget al., 2002; Watanabe et al., 1997). Therefore, phosphorylation ofthese residues of Rlm1 is necessary for the development of adequateReceived 16 September 2015; Accepted 22 February 2016

Departamento de Microbiologıa II, Facultad de Farmacia, UniversidadComplutense de Madrid, IRYCIS, Madrid 28040, Spain.

*Author for correspondence ( jarroyo@ucm.es)

1649

© 2016. Published by The Company of Biologists Ltd | Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

CWI transcriptional responses, although a non-catalytic mechanismby which Slt2 regulates the transcription of a small subset of CWI-responsive genes, through SBF, has also been recently established(Kim andLevin, 2011;Kim et al., 2008). In addition toRlm1 andSlt2,the SWI–SNF ATP chromatin remodelling complex is required foryeast transcriptional reprogramming under cell wall stress conditions.Through a direct interaction with Rlm1, this complex is recruited tothe promoters of genes that are responsive to cellwall stress,mediatingthe necessary nucleosome eviction (Sanz et al., 2012).Genome-wide transcriptional profiles have revealed that

expression of the genes encoding the MAPK Slt2 and thetranscription factor Rlm1 increases under conditions of CWIactivation, such as overexpression of PKC1 (Roberts et al., 2000),treatment with Congo Red (García et al., 2004; Sanz et al., 2012)and overexpression of a hyperactive allele ofMKK1 and temperatureupshift (Jung and Levin, 1999), indicating the existence of internalfeedback autoregulatory circuits of amplification of the cell wallstress response in order to control adaptation through this pathway.Simultaneous monitoring of gene expression programmes of othersignal transduction pathways in yeast by performing genome-widetranscript analysis has also uncovered that the genes encoding theMAPKs Fus3 and Kss1 are induced by stimulation of the respectivemating and invasive growth pathways (Roberts et al., 2000).However, the functional significance of these potential self-regulatory mechanisms is completely unknown.Here, we show that, under cell wall stress conditions, Rlm1

interacts with Rlm1-binding domains at the promoter regions ofRLM1 and SLT2, eliciting transcriptional positive-feedbackmechanisms. We eliminated the positive-feedback events onRLM1 or SLT2 by performing site-directed mutagenesis of theRlm1-binding boxes in the upstream regulatory sequences of therespective chromosomal loci. In the resulting strains, Rlm1 can beactivated by Slt2 but the regulatory transcriptional loops exerted bythe transcription factor on either RLM1 or SLT2 are blocked. Thefirst feedback event is crucial to set up a complete CWItranscriptional response, whereas the effect of the latter on thetranscriptional output is of a much smaller magnitude. Our resultsconfirm the existence of transcriptional amplification loops that aremediated by Rlm1 and that regulate the adaptive response mediatedby the CWI pathway, revealing a functional importance of thismechanism for the amplification of the cell wall stress response.

RESULTSRlm1 is recruited in vivo to the Rlm1-binding domain ofRLM1and SLT2 promoters as a consequence of cell wall stressIn response to Congo Red, a treatment which interferes with thecorrect assembly of yeast cell wall components, the CWI pathwayis transiently activated. Under these conditions, the MAPK Slt2 isphosphorylated and activates the transcription factor Rlm1, which isinterdependently recruited with the chromatin-remodelling complexSWI–SNF to the promoters of the CWI-responsive genes – such asKDX1 (also known as MLP1), SRL3 and YLR194C – to regulategene expression (Sanz et al., 2012). Expression of both RLM1 andSLT2 is also up-regulated under these stress conditions in an Rlm1-dependent manner (García et al., 2004). To further analyse thesepotential transcriptional positive-feedback mechanisms, we firstlooked for Rlm1-binding sites – CTA[T/A]4TAG (Dodou andTreisman, 1997) or TA[T/A]4TAG (Jung and Levin, 1999) – within800-bp sequences upstream from the start codon of both RLM1 andSLT2 using the Regulatory Sequence Analysis Tools (RSAT)software (van Helden, 2003). As shown in Fig. 1A, a singleconsensus motif was identified in the promoter region of both genes,

at positions −308 to −299 (CTAATAATAG) and −232 to −224(TAAAAATAG), respectively.

To confirm that Rlm1 occupies these Rlm1-binding sites in vivo,we performed chromatin immunoprecipitation (ChIP) assays.Chromatin from wild-type strains that expressed Rlm1 taggedwith hemagglutinin (HA) (yRLM1HA) or from a doubly taggedstrain Slt2–HA Rlm1–Myc (ySLT2HA-RLM1Myc), grown in thepresence or absence of cell wall stress, was immunoprecipitatedwith antibodies against HA and Myc, respectively, and analysedusing quantitative (q)PCR to check occupation at the regioncovering the Rlm1-binding sites of the RLM1 (RLM1BOX, fromposition −270 to −401) and SLT2 (SLT2BOX, from position −122to −267) promoters (Fig. 1A). As shown in Fig. 1B, Rlm1 waspresent at low levels in the absence of stress, as deduced fromcomparative Rlm1 binding in tagged and untagged wild-typestrains. However, cell wall stress induced by Congo Red notablyincreased Rlm1 recruitment (about threefold enrichment) to RLM1and SLT2 promoters (Fig. 1B).

To verify that Rlm1 was specifically bound to the Rlm1 DNAbinding motifs found in both genes, we constructed two strainsisogenic to those described above, but bearing mutated versionsof the Rlm1-binding sites at the RLM1 or SLT2 promoters. Inparticular, the internal AATA (RLM1) or AAAT (SLT2) consensusnucleotide sequences within the complete TA[T/A]4TAG Rlm1-binding motif were replaced by GCAG or CTCG, respectively(Fig. 1A). ChIP assays using both strains, yPmut-RLM1HA andyPmut-SLT2HA-RLM1Myc, revealed that occupation of theRLM1BOX and SLT2BOX by Rlm1 under Congo Red treatmentwas drastically reduced, showing only residual levels that were veryclose to those observed in the absence of cell wall stress (Fig. 1B).

Taken together, these results demonstrate that Rlm1 is recruited toits putative DNA binding domains in the promoters of RLM1 andSLT2 as a consequence of cell wall stress. Moreover, nucleotidesubstitutions at Rlm1-binding motifs in both genes essentially blockthe recruitment of Rlm1, enabling us to evaluate the functional roleof both feedback circuits (see below).

Overproduction of Rlm1 and Slt2 by cell wall stress dependson the recruitment of Rlm1 to RLM1 and SLT2 promotersTo characterise the functional relevance of the Rlm1 occupation atRLM1 and SLT2 promoters, we first quantified RLM1 and SLT2gene expression by means of real-time (RT)-qPCR assays. Asshown in Fig. 2A, expression of RLM1 (∼twofold) and SLT2(∼threefold) was upregulated in wild-type strains that had beengrown in the presence of Congo Red for 3 h. In accordance with ourprevious genome-wide transcriptional profiles (García et al., 2004),this induction was blocked in rlm1Δ and slt2Δ strains. Moreover, theinduction of RLM1 was also blocked in the strain that included themutated version of the RLM1 promoter (yPmut-RLM1HA), whereasSLT2 induction was severely compromised in this strain. In the caseof the strain bearing the altered SLT2 promoter (yPmut-SLT2HA),the transcription of SLT2 was only affected, maintaining a normaltranscriptional activation of RLM1 (Fig. 2A).

Then, the levels of Rlm1 and Slt2 were monitored in the strainsbearing native or mutated Rlm1-binding sites that had been grown inthe presence or absence of cell wall stress. Total protein extractswere analysed by western blotting using antibodies against HA andMyc to detect Rlm1–HA and Rlm–Myc, together with antibodies todetect Slt2 and phosphorylated Slt2 levels. As shown in Fig. 2B,C,Congo Red activated the CWI pathway in untagged, Rlm1-tagged(yRLM1HA) and Slt2-tagged (ySLT2HA) wild-type strains, asdenoted by the increased levels of phosphorylated Slt2.

1650

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

When the Rlm1-binding site of RLM1 was inoperative (yPmut-RLM1HA strain), the increase in the levels of Rlm1 that wereobserved in a wild-type strain (yRLM1HA) as a consequence of thestress was completely blocked, and Slt2 levels were significantlyreduced (Fig. 2B). However, the remaining pool of Rlm1 in theyPmut-RLM1HA strain exhibited the electrophoretic mobility shiftassociated with the phosphorylation by Slt2, as previouslydescribed for other CWI activation conditions (Marín et al.,2009), in agreement with the fact that activation of Slt2 by CongoRed was not impaired in this strain (Fig. 2B).Regarding the effects caused by blocking the binding of Rlm1 to

the promoter region of SLT2 (yPmut-SLT2HA strain), the increase inthe amount of Slt2 observed in wild-type cells (ySLT2HA strain)that had been exposed to Congo Red with respect to non-treatedcells, was totally dependent on a functional Rlm1-binding site,whereas phosphorylated Slt2 levels were not substantially affected(Fig. 2C). Contrary to the complete abrogation of the Rlm1increment detected in the strain bearing the mutated version of thepromoter region of RLM1 (yPmut-RLM1HA strain), accumulationof Rlm1 in a strain expressing SLT2 under the control of a non-Rlm1-responsive SLT2 promoter was similar to that obtained in thewild-type strain under stress conditions (Fig. 2D).These results indicate that the autoregulatory transcriptional

feedback mechanism exerted by Rlm1 on ‘cis-acting’ RLM1promoter elements is crucial for the transcriptional activation ofboth SLT2 and RLM1, and for the corresponding increase in theirproteins levels under stress conditions. In contrast, the feedbackmechanism mediated by Rlm1 on ‘trans-acting’ elements of the

SLT2 promoter modulates transcription of SLT2 and the resultantprotein levels but not those of RLM1.

In addition to SLT2 and RLM1, MKK1, one of the redundantMAPKKs of the CWI pathway, and BCK1, the MAPKKK of thispathway, also contain a putative Rlm1-binding domain in theirpromoters at positions −433 to −424 and −38 to −29, respectively.We checked if any of these elements could be a target for potentialRlm1-mediated feedback. However, as shown in Fig. 3, neitherMkk1 or Bck1, nor other upstream elements of the CWI pathway –such as Mkk2, Pkc1 or Rho1 –were overproduced as a consequenceof cell wall stress. Therefore, the positive-feedback transcriptionalregulation of this pathway is only mediated by Rlm1 on RLM1 itselfand on SLT2. Rlm1-binding consensus motifs at the promoters ofBCK1 and MKK1, the former very close to the ATG site and thelatter close to the 3′ non-coding region of the adjacent gene, do notseem to be operative.

The autoregulatory positive-feedback loop mediated byRlm1 is essential for the transcriptional response elicited byactivation of the CWI pathwayOne of the main functions of the CWI pathway is to regulate cellularadaptive responses against situations in which cell wall integrityis compromised. Through transcriptional reprogramming, yeastmodulate the expression of genes important for this process,including genes involved in cell wall remodelling, metabolism, andenergy and signaling (Arroyo et al., 2009). A key question waswhether the positive transcriptional feedback mechanismsdescribed above, mediated by the transcription factor Rlm1, are

Fig. 1. Rlm1 occupies the Rlm1-binding domain of RLM1 and SLT2 genes in vivo upon cell wall stress. (A) Schematic representation of the localisation ofthe Rlm1-binding domains within the HA-tagged RLM1 and SLT2 genes (strains yRLM1HA and ySLT2HA). To generate mutated Rlm1-binding domains,original nucleotides were replaced by nucleotides marked in bold by means of site-directed mutagenesis (strains yPmutRLM1HA and yPmutSLT2HA). Regionsamplified for ChIP experiments, RLM1BOX and SLT2BOX, are shown as broken lines. (B) Rlm1 recruitment to native and mutated promoters of RLM1 (left)and SLT2 (right) either in the presence [Congo Red (CR) 30 μg/ml for 3 h] or in the absence of cell wall stress was analysed by performing ChIP in untagged wild-type (WT), Rlm1–HA-expressing (yRLM1HA and yPmut-RLM1HA) and Rlm1–Myc-expressing (ySLT2HA-RLM1Myc and yPmut-SLT2HA-RLM1Myc) strains.Relative Rlm1 binding at specific DNA regions was calculated as detailed in Materials and Methods using the promoter region ofVMA8 as sequence control. Datarepresent the mean±s.d. of three independent biological replicates.

1651

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

relevant for the CWI transcriptional adaptive response. For thispurpose, we monitored expression levels of several Rlm1-dependent CWI-responsive genes, namely MLP1, CRG1 andYLR194C, together with those of RLM1 and SLT2, in the strainsbearing the mutated Rlm1-binding site at RLM1 (yPmutRLM1HA)or SLT2 (yPmutSLT2HA) promoter regions, as well as in theircorresponding isogenic wild-type strains, at different times aftertreatment with Congo Red.As shown in Fig. 4, the kinetics of transcriptional induction of the

three genes was similar in both tagged wild-type strains

(yRLM1HA and ySLT2HA). Congo Red caused a detectableincrease in the levels of expression of the reporters after 15 min oftreatment. After this time, the transcriptional activation increasedover the time, reaching a peak after 120–180 min. Activation ofgene expression was concomitant with an increase in the levels ofphosphorylated Slt2 over the whole time course (Fig. 5), inagreement with the previously described correlation between levelsof Slt2 phosphorylation and the magnitude of the transcriptionalCWI response (Arias et al., 2011). Additionally, in agreement withan increase of both RLM1 and SLT2 transcripts (Fig. 4), Rlm1 and

Fig. 2. Functional Rlm1-binding domains of RLM1 and SLT2 genes are required for the induction of both genes under cell wall stress. (A) Analysis byperforming RT-qPCR to determine RLM1 and SLT2 expression levels in wild-type (WT), rlm1Δ, slt2Δ, yRLM1HA, yPmutRLM1HA, ySLT2HA and yPmutSLT2HAstrains after 3 h of Congo Red treatment. The results are expressed as the ratio of Congo Red (CR)-treated versus untreated cells. Data represent the mean±s.d.of three independent biological replicates. (B–D) Rlm1 (Rlm1–HA and Rlm1–Myc), Slt2 and phosphorylated Slt2 (P-Slt2) levels were analysed by western blottingusing anti-HA, anti-Myc, anti-Mpk1 (Slt2) and anti-phosphorylated-p44/p42 MAPK antibodies, respectively, in the indicated strains growing exponentially in YPDat 24°C that had been exposed or not to Congo Red (30 μg/ml) for 3 h. Glucose-6-phosphate dehydrogenase (G6PDH) was used as loading control. Numbersindicate the relative amounts (quantified by densitometry analysis) of Rlm1, Slt2 and phosphorylated Slt2 in Congo-Red-treated versus non-treated cells,normalised with respect to the loading control. Graphics in the relevant lower panels show this quantification.

1652

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

Slt2 protein levels also increased in both tagged wild-type strainsafter 30–60 min of treatment as a consequence of the two feedbackloops mediated by Rlm1 (Fig. 5).When gene expression kinetics were monitored in the strain that

included a non-functional Rlm1-binding site at the promoter ofRLM1 (yPmut-RLM1HA), there were no differences between thisstrain and the corresponding wild-type strain in the expression of thethree reporters analysed after 15 min of Congo Red treatment(Fig. 4A), in agreement with the similar levels of Rlm1 observed inboth strains (Fig. 5A). At that time, Slt2 was already slightlyactivated, as measured by monitoring phosphorylated Slt2 levelsand, as a consequence, Rlm1 was also activated, as deduced fromthe Rlm1 shift (Fig. 5A). Although the high levels ofphosphorylated Slt2 reached at different times of treatment, from30 min to 3 h, were almost unaltered in the yPmut-RLM1HA strain,overexpression of RLM1 (Fig. 4A) and the correspondingoverproduction of Rlm1 (Fig. 5A) were blocked. Consequently,transcriptional upregulation by Congo Red of the genes under studywas severely impaired throughout the whole timecourse in theyPmut-RLM1HA strain in respect to that of the wild-type strain(Fig. 4A). Therefore, the high amount of phosphorylated Slt2 thathad accumulated after 1–3 h of Congo Red treatment wasinsufficient for a proper transcriptional CWI response. Instead, thelevels of Rlm1, through the Rlm1-autoregulatory positivetranscriptional feedback mechanism, are crucial for fullexpression of CWI target genes.In the strain bearing a mutated Rlm1-binding site at the SLT2

promoter (yPmut-SLT2HA), a partial reduction of about 30–40% inthe expression of MLP1, CRG1 and YLR194C was observed withrespect to the wild-type strain, but mainly after 120 and 180 min oftreatment, whereas no change or little effect was found after 30 and60 min (Fig. 4B). This effect is presumably related to the fact thatinduction of SLT2 expression was blocked throughout the wholetime course (Fig. 4B) and, consequently, Slt2 was not overproducedin this strain upon cell wall stress (Fig. 5B). The levels of

phosphorylated Slt2 slightly decreased, whereas RLM1 mRNA andRlm1 protein levels in this strain were comparable to those observedin the wild-type strain (Figs 4B and 5B).

On the basis of these results, we decided to construct a doublemutant strain yPmut-RLM1HA-Pmut-SLT2HA, bearing mutatedversions of the Rlm1-binding site at both RLM1 and SLT2promoters. Relative mRNA levels of MLP1, CRG1 and YLR194Cweremeasured in this strain in the absence and presence ofCongoRedfor 3 h, and compared to those found in a wild-type strain(yRLM1HA) and in a single-mutant yPmut-RLM1HA strain. Asshown in Fig. 6, the expression levels of the three genes in the doublemutant were similar to those found in the strain that was only affectedin the Rlm1 autoregulatory feedback. Therefore, the partialtranscriptional induction of CWI genes (∼30–40%) still present inthe absence of the Rlm1 autoregulatory feedback loop (strain yPmut-RLM1HA) is not dependent on the feedback mediated by Rlm1 onSLT2 but seems to depend on the constitutive levels of Slt2 and Rlm1.

Almost the whole yeast transcriptional response to cell wall stressmediated by Congo Red is dependent on the transcription factorRlm1 (Dodou and Treisman, 1997; García et al., 2004; Lagorceet al., 2003). However, an rlm1Δ yeast strain is not hypersensitive tocell wall stress, but it is resistant to Congo Red, Calcofluor White orzymolyase as compared to a wild-type strain (Dodou and Treisman,1997; García et al., 2004; Lagorce et al., 2003), probably becauseof compensatory responses in this mutant when grown under cellwall stress conditions. We tested cell-wall-related phenotypes ofyPmut-RLM1HA and yPmut-SLT2HA mutant strains, and of thecorresponding yRLM1HA and ySLT2HAwild-type strains, and wefound that, like the rlm1Δ yeast strain, both strains were moreresistant to Congo Red than the wild-type strains (Fig. 7), furthersupporting the relevance of the transcriptional feedback loops on thefunction of the CWI pathway.

DISCUSSIONModulation of specific adaptive responses by MAPK signalingpathways requires precise regulatory mechanisms to triggersustained or transient responses. These different dynamics arecrucial for determining their corresponding outputs (Marshall,1995; Santos et al., 2007). The yeast CWIMAPK pathway mediatesa compensatory response to reinforce the cell wall under conditionsaffecting cell wall integrity. The timing of this response, in terms ofactivation of the MAPK Slt2 and development of the correspondingtranscriptional programme, is very different to that of other yeastMAPK pathways like the HOG pathway or the mating signalingpathway. For the mating pheromone response pathway, Kss1 andFus3 are rapidly phosphorylated after a few minutes following theaddition of pheromone, with a maximal activation achieved by15 min, which declines by 2–3 h (Sabbagh et al., 2001), whereaschanges in gene expression have already begun by 15 min (Robertset al., 2000). The MAPK Hog1 is only transiently activatedfollowing osmotic stress, with levels reaching a maximum at 5 minthat then gradually decrease to near-basal levels within 30 min, witha rather early and largely transient transcriptional response (Haoet al., 2007; Maeda et al., 1995; Saito and Posas, 2012). In contrast,CWI signaling is activated persistently in response to growth atelevated temperatures (Kamada et al., 1995; Zarzov et al., 1996) andchemical agents that induce cell wall stress, such as Congo Red orzymolyase, activate CWI signaling by increasing levels ofphosphorylated Slt2 within 15 min upon stress, with a peak after1–3 h of treatment. Slt2 phosphorylation decreases later but it is stilldetected even after 6 h (Bermejo et al., 2008; García et al., 2004). Inagreement, the complete transcriptional response to cell wall

Fig. 3. Protein levels of elements of the CWI pathway upstream of Slt2 donot change upon cell wall stress. Protein levels of Mkk1, Mkk2, Bck1, Pkc1and Rho1 were evaluated in the corresponding Myc-tagged strains growingeither in the presence (30 μg/ml for 3 h) or in the absence of Congo Red (CR)by western blotting using an anti-Myc antibody. Numbers indicate the relativeamounts of the indicated proteins in Congo-Red-treated versus non-treatedcells, normalised with respect to the loading control (G6PDH). Activation of theCWI pathway was monitored by detection of phosphorylated Slt2 (P-Slt2) withan anti-phosphorylated-p44/p42 MAPK antibody.

1653

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

damage does not occur until hours after exposure to the cell-wall-perturbing agent (García et al., 2004; Rodríguez-Peña et al.,2005).Modulation of signaling through MAPK pathways requires many

regulatory mechanisms (Chen and Thorner, 2005; Molina et al.,2010; Zeke et al., 2009). One of these mechanisms involves

feedback loops that are exerted by components of one pathway onelements acting upstream of the same pathway. Negative-feedbackevents contribute to the attenuation of responses, whereas positive-feedback events should amplify signaling. Here, we demonstrate theexistence of two positive transcriptional feedback mechanismswithin the CWI MAPK pathway, one exerted by the transcription

Fig. 4. The transcriptional responsetriggered by cell wall stress is severelyaffected in strains lacking theautoregulatory positive transcriptionalfeedback loop mediated by Rlm1. mRNAlevels of several CWI-responsive genes(MLP1, CRG1, YLR194C, RLM1 and SLT2)were analysed by performing RT-qPCR afterdifferent times of Congo Red (CR) treatment(30 μg/ml) in yRLM1HA and yPmut-RLM1HAstrains (A), and in strains ySLT2HA andyPmut-SLT2HA (B). Values (mean±s.d. ofthree independent biological replicates)represent the expression ratio betweenCongo-Red-treated and non-treated cells ateach time point.

1654

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

factor Rlm1 on SLT2, which encodes the MAPK of the route, andthe other exerted by Rlm1 on RLM1 itself (Fig. 8).Individual blockade of either of these positive-feedback

mechanism through mutagenesis of the Rlm1-binding sites atSLT2 or RLM1 promoters does not impair activation of Slt2 andphosphorylation of Rlm1 induced by cell wall stress, allowing us toevaluate the real impact of these mechanisms on CWI outputresponses. In a wild-type strain, once Slt2 is activated, itphosphorylates Rlm1, which is also overproduced as aconsequence of the transcriptional feedback. Abrogation of theautoregulatory Rlm1 feedback loop results in the lack of RLM1induction and Rlm1 overproduction, as well as in an important

reduction of its transcriptional target SLT2 upon cell wall stress.Consequently, the transcriptional activation mediated by the CWIpathway under cell wall stress is severely impaired, as deduced froma decrease of 60–70% in expression levels of CWI-dependentgenes. Therefore, this positive feedback that is activated by the stressregulates the amplification of the cell wall stress response, and it iscrucially required for a proper transcriptional adaptive response.

Blockade of the transcriptional feedback exerted by Rlm1 onSLT2 affects the induction of SLT2 but not of RLM1. As aconsequence, levels of Slt2 do not increase, but Rlm1overproduction and phosphorylation are unaffected. Under theseconditions, expression levels of CWI-dependent genes are reduced

Fig. 5. Monitoring of Rlm1, Slt2 andphosphorylated Slt2 protein levelsin the absence of the transcriptionalpositive-feedback mechanismsmediated by Rlm1. Rlm1 (Rlm1–HAor Rlm1–Myc), Slt2 andphosphorylated Slt2 (P-Slt2) levelswere analysed by western blottingusing anti-HA, anti-Mpk1 (Slt2) andanti-phosphorylated-p44/p42 MAPKantibodies in strains yRLM1HA andyPmutRLM1HA (A), and strainsySLT2HA-RLM1Myc and yPmut-SLT2HA-RLM1Myc (B) that weregrowing exponentially in YPD at 24°Cand exposed or not to Congo Red (CR;30 μg/ml) for the indicated times.Glucose-6-phosphate dehydrogenase(G6PDH) was used as loading control.Numbers indicate the relative amounts(quantified by densitometry analysis)of Rlm1, Slt2 and phosphorylated Slt2in Congo-Red-treated versus non-treated cells, normalised with respectto the loading control. Graphics in therespective lower panels show thisquantification.

1655

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

by about 30%. Therefore, the increase in the amount of Slt2 that ismediated by the activity of Rlm1 through this feedback suggests apossible mechanism for potential amplification of the stressresponse through the maintenance of MAPK levels under stressconditions. This mechanism has less impact on the CWI outputresponse than the Rlm1 autoregulatory feedback but partiallycontributes to the CWI gene expression levels, particularly underlong-term stress, having almost no effect on transcription at earlytime points. Interestingly, no synergistic effect was found in thedecrease of the induction levels of CWI-dependent genes upon cellwall stress in a yPmut-RLM1HA-Pmut-SLT2HA double-mutantstrain, indicating that the partial transcriptional induction of CWIgenes (∼30–40%), still present in the absence of the Rlm1autoregulatory feedback loop, is not dependent on the feedbackmediated by Rlm1 on SLT2. Therefore, the two regulatorytranscriptional feedback mechanisms described here control themajority of the CWI transcriptional response, with a key role for theRlm1 autoregulatory feedback. In agreement, the phenotypes ofboth yPmut-RLM1HA and yPmut-SLT2HA strains are similar tothose of rlm1Δ. However, part of the response is not dependent onthe overproduction of Rlm1 and Slt2, but it does depend on theconstitutive levels of Slt2 and Rlm1 present in the cell. In fact, bothSlt2 and Rlm1 are activated in the absence of the feedback loops.Consequently, only the deletion of either SLT2 or RLM1 completelyblocks the transcriptional CWI response.The MAPK Slt2 regulates Rlm1 activation by phosphorylation of

Ser427 and Thr439, and this activation is required for transcriptionalactivation of CWI-responsive genes under cell wall stress that isinducedbyCalcofluorWhite (Jung et al., 2002).Our data demonstratethat phosphorylation of Rlm1 by the MAPK Slt2 is insufficient for a

full functional CWI transcriptional response. Self-sustaining patternsof gene expression are therefore mainly achieved by positiveautoregulatory circuits based on sustained activation of Rlm1. Asimilar autoregulatorymechanism involvingmyocyte enhancer factor2 (MEF2) has been described in Drosophila. Mammalian MEF2protein isoforms are MADS-box transcription factors, related toRlm1, and share a similar DNA-binding specificity (Dodouand Treisman, 1997; Jung et al., 2002). Interestingly, sustainedactivation of MEF2 through an autoregulatory enhancer seems to beimportant for muscle development in this organism (Cripps et al.,2004).

In contrast to negative autoregulation, which speeds up theresponse time of gene circuits (Becskei and Serrano, 2000), positiveautoregulation – which occurs when a transcription factor enhancesits own rate of production – generates bistability, amplifies geneexpression and slows down the kinetics of gene expression (Alon,2007; Maeda and Sano, 2006; Singh, 2014). Therefore, the positivetranscriptional feedback mechanism mediated by Rlm1 describedhere, which is crucial for the development of adequate cell walltranscriptional adaptive responses could be, at least in part,responsible for the unique dynamics of the CWI pathway output.Cell signaling systems that contain positive-feedback loops can, inprinciple, convert a transient trigger stimulus into an irreversibleresponse (Ferrell, 2002). So, in contrast to positive-feedback eventsthat amplify signaling through the CWI MAPK pathway, othernegative feedback events contribute to attenuate responses (Fig. 8).Particularly, Slt2 inhibits the pathway that downregulates theguanine nucleotide exchange factor (GEF) activity of Rom2 througha retrophosphorylation mechanism (Guo et al., 2009), in a mannersimilar to that of other negative feedback events exerted by Hog1and Fus3 on upstream elements of the HOG and mating pathways,respectively (Feng and Davis, 2000; Flotho et al., 2004; Hao et al.,2007; Kranz et al., 1994; Saito and Posas, 2012). Additionally,retrophosphorylation of the MAPKKs Mkk1 and Mkk2 by theirtarget Slt2 has also been reported, but its relevance in thetranscriptional response is lacking (Jiménez-Sánchez et al., 2007).By contrast, the CWI pathway can be negatively regulated by theactivity of diverse phosphatases such as Ptp2, Ptp3, Msg5, Sdp1(Martín et al., 2005) and Ptc1 (Du et al., 2006; González et al.,2006) on Slt2, whereas Pkc1 is deubiquitylated by Ubp3 (Wanget al., 2008). The transcriptional induction of many phosphatasesthat act upon MAPKs under the control of their target MAPK hasbeen well documented as a negative-feedback regulation loop ofMAPK signaling in humans (Caunt and Keyse, 2013; Ekerot et al.,2008; Keyse, 2000). Phosphatases Ptp2 and Msg5 aretranscriptionally induced in several conditions that activate theCWI pathway (García et al., 2004, 2009; Hahn and Thiele, 2002;Mattison et al., 1999) in an Slt2-dependent manner, indicating theexistence of transcriptional negative-feedback loops at this level(Fig. 8).

Because positive autoregulation is a common occurrence amongregulatory genes, our results on the CWI pathway provide insightsinto properties that might apply to other signal transduction

Fig. 6. Impairment of the induction of CWI-responsive genes in a strainlacking the Rlm1 autoregulatory positive transcriptional feedback loop isnot exacerbated by simultaneous blocking of the feedback exerted byRlm1 on SLT2. mRNA levels of MLP1, CRG1 and YLR194C genes wereanalysed by performing RT-qPCR at 3 h after Congo Red (CR) treatment(30 μg/ml) in yRLM1HA, yPmut-RLM1HA and yPmut-RLM1HA-Pmut-SLT2HAstrains. Values (mean±s.d. of three independent biological replicates)represent the expression ratio between Congo-Red-treated and non-treatedcells at each time point.

Fig. 7. yPmut-RLM1HA and yPmut-SLT2HA-RLM1Myc mutantstrains are more resistant to Congo Red than the correspondingwild-type strains. yRLM1HA, yPmutRLM1HA, ySLT2HA andyPmut-SLT2HAwere spotted onto YPD plates without or with 100 µg/ml of Congo Red (CR), and plates were incubated for 48 h at 30°C.

1656

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

networks. Similar to the CWI pathway, Fus3 (the MAPK of themating pathway) and the transcription activator Ste12 are targets forSte12 binding, and their expression is induced in the presence ofpheromones (Ren et al., 2000; Roberts et al., 2000), raising thepossibility of additional feedback mechanisms that control theMAPK pathways regulating mating and filamentation (Zeitlingeret al., 2003).

MATERIALS AND METHODSYeast strains and growth conditionsExperiments were performed with the Saccharomyces cerevisiae strainBY4741 (MATa; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0), and rlm1Δ and slt2Δmutant derivatives were provided by Euroscarf (Frankfurt, Germany).yRLM1HA and yPmut-RLM1HA strains were obtained using the methoddescribed previously (Sikorski and Hieter, 1989) by transforming the rlm1Δmutant with the integrative pRS303-RLM1HA and pRS303-Pmut-RLM1HA (HIS3) plasmids linearised at the unique EcoRV site (atposition −864 in the RLM1 promoter), respectively. ySLT2HA andyPmut-SLT2HA strains were achieved by transforming the slt2Δ mutantwith the integrative plasmids pRS306-SLT2HA and pRS306-Pmut-SLT2HA(URA3) linearised at the unique BlpI site (at position −323 in the SLT2promoter), respectively (Fig. 1A). Correct integration was confirmed byperforming PCR, amplifying a region between the positions −898 and +100(RLM1) and between −399 and +97 (SLT2), and sequencing thecorresponding DNA fragments. Wild-type (WT) WT-RLM1Myc, WT-BCK1Myc, WT-PKC1Myc, slt2Δ-RLM1Myc, ySLT2HA-RLM1Myc and

yPmut-SLT2HA-RLM1Myc strains were obtained by tagging the Mycepitope to the C-terminus of Bck1, Pkc1 or Rlm1 (BY4741, slt2Δ,ySLT2HA and yPmut-SLT2HA strains) using a PCR-based genemodification method (Longtine et al., 1998). The yPmut-RLM1HA-Pmut-SLT2HA double-mutant strain was obtained in two steps. First, the SLT2gene was replaced by the hphMX4 marker in the yPmut-RLM1HA strainusing the PCR-based method described by Wach et al. (1997). Second,the yPmut-RLM1HA slt2Δ strain was transformed with the integrativepRS306-Pmut-SLT2HA plasmid linearised at the unique BlpI site asdescribed above. MKK1–Myc and MKK2–Myc strains were providedby Victor J. Cid (Departamento de Microbiología II, UniversidadComplutense, Madrid, Spain). To obtain the RHO1–Myc strain, BY4741was transformed with the integrative Myc-RHO1-pRS406 plasmidlinearised at the MluI site, provided by Yu Jiang (Department ofPharmacology, University of Pittsburgh School of Medicine, Pittsburgh,PA) (Guo et al., 2009).

Cells were grown onYPD (1% yeast extract, 2% peptone and 2% glucose)in liquid medium at 220 rpm and 24°C to an optical density of 0.8–1 (A600).The culture was refreshed to 0.2 (A600) in YPD, grown for an additional 2 h30 min and then divided into two parts. One part was allowed to continuegrowing under the same conditions (non-treated culture) and the other onewas supplemented with 30 μg/ml of Congo Red (Merck, Darmstadt,Germany). Cells were then collected after 3 h of incubation.

Plasmids and site-directed mutagenesisThe plasmid pRS303-RLM1HA (integrativeHIS3) was obtained by cloningthe insert from plasmid YEp352-RLM1-3×HA (Sanz et al., 2012),

Fig. 8. Regulation of the CWI pathway. Upon cell wall stress, the MAPK Slt2 mediates, through Rlm1 and the SWI–SNF chromatin remodelling complex, theinduction of the CWI transcriptional programme (Sanz et al., 2012). Modulation of CWI signaling outputs requires both positive- and negative-feedbackmechanisms. The transcription factor Rlm1 exerts transcriptional positive-feedback loops on the expression of RLM1 and SLT2 (blue arrows). The former isessential for the amplification and setting up of a complete CWI transcriptional output response, whereas the latter contributes to a lesser extent. In contrast,attenuation of the CWI pathway requires several post-translational negative-feedback mechanisms, including negative retrophosphorylation feedback loopsexerted by Slt2 on upstream components of CWI pathway, such as Rom2 (Guo et al., 2009), the redundant MAPKKs Mkk1 and Mkk2 (Jimenez-Sanchez et al.,2007) (red lines) and the Rlm1-dependent transcriptional induction of the Slt2 phosphatases Ptp2 and Msg5 (Garcıa et al., 2004, 2009; Hahn and Thiele, 2002;Mattison et al., 1999) (blue line). In addition, other Slt2 phosphatases – such as Ptp3, Sdp1 and Ptc1 (Martın et al., 2005) – and ubiquitylation of Pkc1 by Ubp3(Wang et al., 2008) also contribute to the attenuation of the induction of the CWI pathway.

1657

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

containing the RLM1 promoter and coding regions with a 3×HA epitopefused to the C-terminus of Rlm1, into a pRS303 shuttle vector betweenPvuII and SacI restriction sites. To obtain the plasmid pRS306-SLT2HA(integrative URA3), the insert from plasmid p2188, which was kindlyprovided by David Levin (Department of Microbiology, Boston UniversitySchool of Medicine, Boston, MA), containing the SLT2 promoter andcoding regions with a 3×HA epitope fused to the C-terminus of Slt2 (Kimet al., 2008), was cloned into a pRS306 shuttle vector between EcoRI andSalI restriction sites. The plasmids pRS303-Pmut-RLM1HA and pRS306-Pmut-SLT2HA were obtained by site-directed mutagenesis of the Rlm1-binding domain in the upstream regions of RLM1 (CTAATAATAG) andSLT2 (TAAAAATAG), respectively, using the QuikChange II site-directedmutagenesis kit (Agilent Technologies). Primers design and PCRconditions were as described in the manufacturer’s instructions. Theprimers used to create pRS303-Pmut-RLM1HA plasmid were 5′-CGGAA-GAGATGAAATCTGCAGCCAGAAACAGCTCG-3′ and 5′-CGAGCT-GTTTCTGGCTGCAGATTTCATCTCTTCCG-3′. The presence of themutation was verified by restriction analysis with PstI because a restrictionsite for this enzyme had been introduced into the Rlm1-binding domain asa consequence of the mutagenesis. To create the pRS306-Pmut-SLT2HAplasmid, the following primers were used: 5′-GGAAAG TTTCAGTGTT-AACTCGAGAAACTGAAAAAGGAG-3′ and 5′-CTCCT TTTTCAGT-TTCTCGAGTTAACACTGAAACTTTCC-3′ including, in this case, aXhoI site in the Rlm1-binding domain to check for the correct mutation(Fig. 1A). In both cases, further verification of the mutations was performedby DNA sequencing.

Chromatin immunoprecipitation assaysChIP was performed as described elsewhere (Papamichos-Chronakis andPeterson, 2008; Sanz et al., 2012). The antibodies used in these experimentswere a polyclonal anti-HA antibody (ab9110, Abcam; 1:125 dilution ) and amonoclonal anti-Myc antibody (clone 9E10; MMS-150P, Covance; 1:125dilution). The immunoprecipitated DNA comprising the regionsRLM1BOX (position −401 to −270) and SLT2BOX (position −267 to−122), which includes the Rlm1-binding domain, was quantified byperforming qPCR using the following primers: RLM1-BOX-UP 5′-GCA-GCATCACCGGGTGA-3′ and RLM1-BOX-DOWN 5′-TTATTG TTCA-GAGGAAGATCGAGCT-3′ and SLT2-BOX-UP 5′-CTGCGA AATGTT-GGCAGAAT-3′ and SLT2-BOX-DOWN 5′-GGCAGACCAG GGTCTT-CAAT-3′. The fold enrichment at specific DNA regionswas calculated usingthe comparative Ct method (Aparicio et al., 2004) and the promoter region ofthe VMA8 gene, whose expression is not affected by cell wall stress, assequence control. Thus, the Ct of the input sample was subtracted from theCt of the immunoprecipitated sample to calculate the ΔCt, both in the controlsequence (ΔCtcont) and in the target DNA (ΔCtexp), for each condition.Finally, fold enrichment was calculated by using the formulaFE=2−(ΔCtexp −ΔCtcont).

Western blotting assaysThe procedures used for immunoblot analyses, including cell collection andlysis, collection of proteins, separation by SDS–PAGE and transfer tonitrocellulose membranes were performed as previously described (Bermejoet al., 2008) using the Odyssey Infrared Imaging System (LI-COR). Thedetection of phosphorylated Slt2 was accomplished using an anti-phosphorylated-p44/p42 MAPK monoclonal antibody (Thr202/Tyr204;4370; Cell Signaling Technology). The detection of HA and Myc epitopeswas performed using anti-HA (clone HA.11, MMS-101P, Covance) andanti-c-Myc (clone 9E10, MMS-150P, Covance) monoclonal antibodies,respectively. Slt2 was detected using the anti-Mpk1 monoclonal antibody(clone E9, sc-133189, Santa Cruz Biotechnology). To monitor proteinloading, glucose-6-phosphate dehydrogenase (G6PDH) levels weredetermined using an anti-G6PDH polyclonal antibody (A9521, Sigma-Aldrich). The secondary antibodies used were IRDye 800CW goat anti-rabbit (925-32211) and IRDye 680LT goat anti-mouse (925-68020), bothfrom LI-COR (Lincoln, NE). All antibodies were used at the dilutionsrecommended by the manufacturers. Quantification of protein bandswas determined by performing densitometry analysis using the softwareImageJ (Schneider et al., 2012). Relative amounts of the different proteins

in the presence of cell wall stress with respect to those in the absence ofstress were obtained after normalisation of the respective bands to theloading control.

RT-qPCR assaysRNA isolation and RT-qPCR assays were performed as previouslydescribed (García et al., 2004). For quantification, the abundance of eachtranscript was determined using the amount of the standard transcript ACT1for input cDNA normalisation, and final data on relative gene expressionbetween the two conditions tested were calculated following the 2−ΔΔCt

method, as described by Livak and Schmittgen (Livak and Schmittgen,2001). The following forward and reverse primers, respectively, were used:ACT1, 5′-ACGAAAGATTCAGAGCCCCA-3′ and 5′-GCAGATTCCAA-ACCCAAAACA-3′; RLM1, 5′-CCGCATATAATGGAAATACCG-3′ and5′-TCTCCTGAAATATCAGTCGAAAAA-3′, or 5′-TAATGGAAATAC-CGGGCTGA-3′ and 5′-CATAGGGATAGCCCGCATAG-3′ for HA-tag-ged RLM1 strains; SLT2, 5′-AAGGCATGATGCAGATTTCC-3′ and5′-AAGCGTGCCGTTATCATTCT-3′, or 5′-CCACTGGAAATACCGC-AGAT-3′ and 5′-TCATAGGGATAGCCCGCATA-3′ for HA-tagged SLT2strains; CRG1, 5′-TCGATTTGGAGATATTGAAGTCACA-3′ and 5′-G-CATTTGGGTCCGAAGGA-3′; YLR194C, 5′-GGTAGCGTCCGTAAT-GTCCAA-3′ and 5′-CCCGCACCATAAGCTATGTGA-3′; and MLP1,5′-TCCTTCCCTTCAAACATTGGTT-3′ and 5′-TGAATTATCAAGAA-TGCACAAAAGC-3′.

Congo Red sensitivity assayCells were grown overnight in YPD, and exponentially growing cultureswere adjusted to an optical density at 600 nm (A600) of 0.13 (∼2×103 cellsper µl). 1:5 serial dilutions of yeast cultures were spotted with a 48 pinMulti-Blot replicator (VP407AH model, Fisher Scientific, PA) onto YPDplates and YPD plates that had been supplemented with 100 µg/ml of CongoRed. Growth was monitored after 2 days of incubation at 30°C.

AcknowledgementsWe thank Sonia Dıez-Mun iz for technical assistance and Enrico Cabib for a criticalreading of themanuscript. Victor J. Cid, David Levin and Yu Jiang are acknowledgedfor the strains and plasmids provided. We are also in debt to Jesus Garcıa-Cantalejoand Rosa Perez at the Genomics Unit [Genomics and Proteomics Center,Universidad Complutense de Madrid (UCM)] for their help with the qPCRexperiments. All members of our research group (UCM-920640: Yeast FunctionalGenomics) at the Department of Microbiology II (UCM) are also acknowledged fortheir support.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsR.G. and A.B.S. performed experiments and analysed the data. J.M.R.-P. and C.N.designed experiments and analysed the data. J.A. directed the research, designedexperiments, analysed thedata andwrote themanuscript with theparticipation ofR.G.

FundingThis work was supported by the Ministerio de Economıa (MINECO) [grant numbersBIO2010-22146 and BIO2013-48136-P to J.A.]; and by the Comunidad de Madrid[grant number S2010/BMD-2414 to J.A.].

ReferencesAlon, U. (2007). Network motifs: theory and experimental approaches. Nat. Rev.

Genet. 8, 450-461.Aparicio, O., Geisberg, J. V. andStruhl, K. (2004). Chromatin immunoprecipitation

for determining the association of proteins with specific genomic sequences invivo. Curr. Protoc. Cell Biol. 23, 17.7.1-17.7.23.

Arias, P., Dıez-Mun iz, S., Garcıa, R., Nombela, C., Rodrıguez-Pen a, J. M. andArroyo, J. (2011). Genome-wide survey of yeast mutations leading to activation ofthe yeast cell integrity MAPK pathway: novel insights into diverse MAPKoutcomes. BMC Genomics 12, 390.

Arroyo, J., Bermejo, C., Garcıa, R. and Rodrıguez-Pen a, J. M. (2009). Genomicsin the detection of damage in microbial systems: cell wall stress in yeast. Clin.Microbiol. Infect. 15 Suppl. 1, 44-46.

1658

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

Baetz, K., Moffat, J., Haynes, J., Chang, M. and Andrews, B. (2001).Transcriptional coregulation by the cell integrity mitogen-activated proteinkinase Slt2 and the cell cycle regulator Swi4. Mol. Cell. Biol. 21, 6515-6528.

Becskei, A. and Serrano, L. (2000). Engineering stability in gene networks byautoregulation. Nature 405, 590-593.

Bermejo, C., Rodrıguez, E., Garcıa, R., Rodrıguez-Pen a, J. M., Rodrıguez de laConcepcion, M. L., Rivas, C., Arias, P., Nombela, C., Posas, F. and Arroyo, J.(2008). The sequential activation of the yeast HOG and SLT2 pathways is requiredfor cell survival to cell wall stress. Mol. Biol. Cell 19, 1113-1124.

Boorsma, A., de Nobel, H., ter Riet, B., Bargmann, B., Brul, S., Hellingwerf, K. J.and Klis, F. M. (2004). Characterization of the transcriptional response to cell wallstress in Saccharomyces cerevisiae. Yeast 21, 413-427.

Caunt, C. J. and Keyse, S. M. (2013). Dual-specificity MAP kinase phosphatases(MKPs): shaping the outcome of MAP kinase signalling. FEBS J. 280, 489-504.

Chen, R. E. and Thorner, J. (2005). Systems biology approaches in cell signalingresearch. Genome Biol. 6, 235.

Cripps, R. M., Lovato, T. L. and Olson, E. N. (2004). Positive autoregulation of theMyocyte enhancer factor-2 myogenic control gene during somatic muscledevelopment in Drosophila. Dev. Biol. 267, 536-547.

Dodou, E. and Treisman, R. (1997). The Saccharomyces cerevisiae MADS-boxtranscription factor Rlm1 is a target for the Mpk1 mitogen-activated protein kinasepathway. Mol. Cell. Biol. 17, 1848-1859.

Du, Y., Walker, L., Novick, P. and Ferro-Novick, S. (2006). Ptc1p regulates corticalER inheritance via Slt2p. EMBO J. 25, 4413-4422.

Ekerot, M., Stavridis, M. P., Delavaine, L., Mitchell, M. P., Staples, C., Owens,D. M., Keenan, I. D., Dickinson, R. J., Storey, K. G. and Keyse, S. M. (2008).Negative-feedback regulation of FGF signalling by DUSP6/MKP-3 is driven byERK1/2 and mediated by Ets factor binding to a conserved site within the DUSP6/MKP-3 gene promoter. Biochem. J. 412, 287-298.

Elion, E. A., Qi, M. andChen,W. (2005). Signal transduction: signaling specificity inyeast. Science 307, 687-688.

English, J. G., Shellhammer, J. P., Malahe, M., McCarter, P. C., Elston, T. C. andDohlman, H. G. (2015). MAPK feedback encodes a switch and timer for tunablestress adaptation in yeast. Sci. Signal. 8, ra5.

Feng, Y. and Davis, N. G. (2000). Feedback phosphorylation of the yeast a-factorreceptor requires activation of the downstream signaling pathway from G proteinthrough mitogen-activated protein kinase. Mol. Cell. Biol. 20, 563-574.

Ferrell, J. E., Jr (2002). Self-perpetuating states in signal transduction: positivefeedback, double-negative feedback and bistability. Curr. Opin. Cell Biol. 14,140-148.

Flatauer, L. J., Zadeh, S. F. and Bardwell, L. (2005). Mitogen-activated proteinkinases with distinct requirements for Ste5 scaffolding influence signalingspecificity in Saccharomyces cerevisiae. Mol. Cell. Biol. 25, 1793-1803.

Flotho, A., Simpson, D. M., Qi, M. and Elion, E. A. (2004). Localized feedbackphosphorylation of Ste5p scaffold by associated MAPK cascade. J. Biol. Chem.279, 47391-47401.

Garcıa, R., Bermejo, C., Grau, C., Perez, R., Rodrıguez-Pen a, J. M., Francois, J.,Nombela, C. and Arroyo, J. (2004). The global transcriptional response totransient cell wall damage in Saccharomyces cerevisiae and its regulation by thecell integrity signaling pathway. J. Biol. Chem. 279, 15183-15195.

Garcıa, R., Rodrıguez-Pen a, J. M., Bermejo, C., Nombela, C. and Arroyo, J.(2009). The high osmotic response and cell wall integrity pathways cooperate toregulate transcriptional responses to zymolyase-induced cell wall stress inSaccharomyces cerevisiae. J. Biol. Chem. 284, 10901-10911.

Gonzalez, A., Ruiz, A., Serrano, R., Arino, J. and Casamayor, A. (2006).Transcriptional profiling of the protein phosphatase 2C family in yeast providesinsights into the unique functional roles of Ptc1. J. Biol. Chem. 281, 35057-35069.

Guo, S., Shen, X., Yan, G., Ma, D., Bai, X., Li, S. and Jiang, Y. (2009). A MAPkinase dependent feedback mechanism controls Rho1 GTPase and actindistribution in yeast. PLoS ONE 4, e6089.

Hahn, J.-S. and Thiele, D. J. (2002). Regulation of the Saccharomyces cerevisiaeSlt2 kinase pathway by the stress-inducible Sdp1 dual specificity phosphatase.J. Biol. Chem. 277, 21278-21284.

Hao, N., Behar, M., Parnell, S. C., Torres, M. P., Borchers, C. H., Elston, T. C. andDohlman, H. G. (2007). A systems-biology analysis of feedback inhibition in theSho1 osmotic-stress-response pathway. Curr. Biol. 17, 659-667.

Jacoby, T., Flanagan, H., Faykin, A., Seto, A. G., Mattison, C. and Ota, I. (1997).Two protein-tyrosine phosphatases inactivate the osmotic stress responsepathway in yeast by targeting the mitogen-activated protein kinase, Hog1.J. Biol. Chem. 272, 17749-17755.

Jimenez-Sanchez, M., Cid, V. J. and Molina, M. (2007). Retrophosphorylation ofMkk1 and Mkk2 MAPKKs by the Slt2 MAPK in the yeast cell integrity pathway.J. Biol. Chem. 282, 31174-31185.

Jung, U. S. and Levin, D. E. (1999). Genome-wide analysis of gene expressionregulated by the yeast cell wall integrity signalling pathway. Mol. Microbiol. 34,1049-1057.

Jung, U. S., Sobering, A. K., Romeo, M. J. and Levin, D. E. (2002). Regulation ofthe yeast Rlm1 transcription factor by theMpk1 cell wall integrity MAP kinase.Mol.Microbiol. 46, 781-789.

Kamada, Y., Jung, U. S., Piotrowski, J. and Levin, D. E. (1995). The proteinkinase C-activated MAP kinase pathway of Saccharomyces cerevisiae mediatesa novel aspect of the heat shock response. Genes Dev. 9, 1559-1571.

Ketela, T., Green, R. and Bussey, H. (1999). Saccharomyces cerevisiaeMid2p is apotential cell wall stress sensor and upstream activator of the PKC1-MPK1 cellintegrity pathway. J. Bacteriol. 181, 3330-3340.

Keyse, S. M. (2000). Protein phosphatases and the regulation of mitogen-activatedprotein kinase signalling. Curr. Opin. Cell Biol. 12, 186-192.

Keyse, S. M. and Emslie, E. A. (1992). Oxidative stress and heat shock induce ahuman gene encoding a protein-tyrosine phosphatase. Nature 359, 644-647.

Kim, K.-Y. and Levin, D. E. (2011). Mpk1 MAPK association with the paf1 complexblocks sen1-mediated premature transcription termination. Cell 144, 745-756.

Kim, K.-Y., Truman, A. W. and Levin, D. E. (2008). Yeast Mpk1 mitogen-activatedprotein kinase activates transcription through Swi4/Swi6 by a noncatalyticmechanism that requires upstream signal. Mol. Cell. Biol. 28, 2579-2589.

Kranz, J. E., Satterberg, B. and Elion, E. A. (1994). The MAP kinase Fus3associates with and phosphorylates the upstream signaling component Ste5.Genes Dev. 8, 313-327.

Lagorce, A., Hauser, N. C., Labourdette, D., Rodrıguez, C., Martin-Yken, H.,Arroyo, J., Hoheisel, J. D. and Francois, J. (2003). Genome-wide analysis ofthe response to cell wall mutations in the yeastSaccharomyces cerevisiae. J. Biol.Chem. 278, 20345-20357.

Levin, D. E. (2011). Regulation of cell wall biogenesis inSaccharomyces cerevisiae:the cell wall integrity signaling pathway. Genetics 189, 1145-1175.

Livak, K. J. and Schmittgen, T. D. (2001). Analysis of relative gene expression datausing real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods25, 402-408.

Longtine, M. S., McKenzie, A., III, Demarini, D. J., Shah, N. G., Wach, A.,Brachat, A., Philippsen, P. and Pringle, J. R. (1998). Additional modules forversatile and economical PCR-based gene deletion and modification inSaccharomyces cerevisiae. Yeast 14, 953-961.

Maeda, Y. T. and Sano, M. (2006). Regulatory dynamics of synthetic gene networkswith positive feedback. J. Mol. Biol. 359, 1107-1124.

Maeda, T., Takekawa, M. and Saito, H. (1995). Activation of yeast PBS2MAPKK byMAPKKKs or by binding of an SH3-containing osmosensor. Science 269,554-558.

Marın, M. J., Flandez, M., Bermejo, C., Arroyo, J., Martın, H. and Molina, M.(2009). Different modulation of the outputs of yeast MAPK-mediated pathways bydistinct stimuli and isoforms of the dual-specificity phosphataseMsg5.Mol. Genet.Genomics 281, 345-359.

Marshall, C. J. (1995). Specificity of receptor tyrosine kinase signaling: transientversus sustained extracellular signal-regulated kinase activation. Cell 80,179-185.

Martın, H., Flandez, M., Nombela, C. andMolina, M. (2005). Protein phosphatasesin MAPK signalling: we keep learning from yeast. Mol. Microbiol. 58, 6-16.

Mattison, C. P., Spencer, S. S., Kresge, K. A., Lee, J. and Ota, I. M. (1999).Differential regulation of the cell wall integrity mitogen-activated protein kinasepathway in budding yeast by the protein tyrosine phosphatases Ptp2 and Ptp3.Mol. Cell. Biol. 19, 7651-7660.

Mobashir, M., Madhusudhan, T., Isermann, B., Beyer, T. and Schraven, B.(2014). Negative interactions and feedback regulations are required for transientcellular response. Sci. Rep. 4, 3718.

Molina, M., Cid, V. J. and Martın, H. (2010). Fine regulation of Saccharomycescerevisiae MAPK pathways by post-translational modifications. Yeast 27,503-511.

Papamichos-Chronakis, M. and Peterson, C. L. (2008). The Ino80 chromatin-remodeling enzyme regulates replisome function and stability. Nat. Struct. Mol.Biol. 15, 338-345.

Purvis, J. E. and Lahav, G. (2013). Encoding and decoding cellular informationthrough signaling dynamics. Cell 152, 945-956.

Rajavel, M., Philip, B., Buehrer, B. M., Errede, B. and Levin, D. E. (1999). Mid2 isa putative sensor for cell integrity signaling in Saccharomyces cerevisiae. Mol.Cell. Biol. 19, 3969-3976.

Reinoso-Martın, C., Schuller, C., Schuetzer-Muehlbauer, M. and Kuchler, K.(2003). The yeast protein kinase C cell integrity pathwaymediates tolerance to theantifungal drug caspofungin through activation of Slt2p mitogen-activated proteinkinase signaling. Eukaryot. Cell 2, 1200-1210.

Ren, B., Robert, F., Wyrick, J. J., Aparicio, O., Jennings, E. G., Simon, I.,Zeitlinger, J., Schreiber, J., Hannett, N., Kanin, E. et al. (2000). Genome-widelocation and function of DNA binding proteins. Science 290, 2306-2309.

Roberts, C. J., Nelson, B., Marton, M. J., Stoughton, R., Meyer, M. R., Bennett,H. A., He, Y. D., Dai, H., Walker, W. L., Hughes, T. R. et al. (2000). Signaling andcircuitry of multiple MAPK pathways revealed by a matrix of global geneexpression profiles. Science 287, 873-880.

Rodrıguez-Pen a, J. M., Perez-Dıaz, R. M., Alvarez, S., Bermejo, C., Garcıa, R.,Santiago, C., Nombela, C. and Arroyo, J. (2005). The ‘yeast cell wall chip’ - atool to analyse the regulation of cell wall biogenesis in Saccharomyces cerevisiae.Microbiology 151, 2241-2249.

1659

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience

Sabbagh, W., Jr, Flatauer, L. J., Bardwell, A. J. and Bardwell, L. (2001).Specificity of MAP kinase signaling in yeast differentiation involves transientversus sustained MAPK activation. Mol. Cell 8, 683-691.

Saito, H. and Posas, F. (2012). Response to hyperosmotic stress. Genetics 192,289-318.

Santos, S. D. M., Verveer, P. J. and Bastiaens, P. I. H. (2007). Growth factor-induced MAPK network topology shapes Erk response determining PC-12 cellfate. Nat. Cell. Biol. 9, 324-330.

Sanz, A. B., Garcıa, R., Rodrıguez-Pen a, J. M., Dıez-Mun iz, S., Nombela, C.,Peterson, C. L. and Arroyo, J. (2012). Chromatin remodeling by the SWI/SNFcomplex is essential for transcription mediated by the yeast cell wall integrityMAPK pathway. Mol. Biol. Cell 23, 2805-2817.

Schneider, C. A., Rasband,W. S. and Eliceiri, K. W. (2012). NIH Image to ImageJ:25 years of image analysis. Nat. Methods 9, 671-675.

Sikorski, R. S. and Hieter, P. (1989). A system of shuttle vectors and yeast hoststrains designed for efficient manipulation of DNA in Saccharomyces cerevisiae.Genetics 122, 19-27.

Singh, V. (2014). Recent advancements in synthetic biology: current status andchallenges. Gene 535, 1-11.

van Helden, J. (2003). Regulatory sequence analysis tools. Nucleic Acids Res. 31,3593-3596.

Wach, A., Brachat, A., Alberti-Segui, C., Rebischung, C. and Philippsen, P.(1997). Heterologous HIS3 marker and GFP reporter modules for PCR-targetingin Saccharomyces cerevisiae. Yeast 13, 1065-1075.

Wang, Y., Zhu, M., Ayalew, M. and Ruff, J. A. (2008). Down-regulation of Pkc1-mediated signaling by the deubiquitinating enzyme Ubp3. J. Biol. Chem. 283,1954-1961.

Watanabe, Y., Takaesu, G., Hagiwara, M., Irie, K. and Matsumoto, K. (1997).Characterization of a serum response factor-like protein in Saccharomycescerevisiae, Rlm1, which has transcriptional activity regulated by the Mpk1 (Slt2)mitogen-activated protein kinase pathway. Mol. Cell. Biol. 17, 2615-2623.

Witzel, F., Maddison, L. and Bluthgen, N. (2012). How scaffolds shape MAPKsignaling: what we know and opportunities for systems approaches. Front.Physiol. 3, 475.

Zalatan, J. G., Coyle, S. M., Rajan, S., Sidhu, S. S. and Lim, W. A. (2012).Conformational control of the Ste5 scaffold protein insulates against MAP kinasemisactivation. Science 337, 1218-1222.

Zarzov, P., Mazzoni, C. and Mann, C. (1996). The SLT2(MPK1) MAP kinase isactivated during periods of polarized cell growth in yeast. EMBO J. 15, 83-91.

Zeitlinger, J., Simon, I., Harbison, C. T., Hannett, N. M., Volkert, T. L., Fink, G. R.and Young, R. A. (2003). Program-specific distribution of a transcription factordependent on partner transcription factor and MAPK signaling.Cell 113, 395-404.

Zeke, A., Lukacs, M., Lim, W. A. and Remenyi, A. (2009). Scaffolds: interactionplatforms for cellular signalling circuits. Trends Cell Biol. 19, 364-374.

1660

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1649-1660 doi:10.1242/jcs.180190

Journal

ofCe

llScience