ownloaded from genome. cshlp org on November 3, 2010- Published by Cold Spring Harbor Laboratory Press Coevolution within a transcriptional network Number of yAP-1s protein interaction, which have demonstrated cases in which com- Species Lysine( K) Arginine(R) pensatory mutations are required to maintain protein interaction S cerevisiae 3 s paradoxus 3 f transcriptional networks, coevolution gives rise to"regulatory . S. mikatae homeostasis. "in which both mutations in a tf and its dna binding motif occur in compensatory fashion to maintain tran c. glabrata scriptional regulation. This series of compensatory mutations S castellii 4 which maintains both the transcriptional circuit and regulatory logic, parallels that of previous work demonstrating evolution of alternative transcriptional circuits producing identical logic( Tsong et al. 2006). Such systems of tightly coupled compensatory mu- tations might serve to counter the widespread divergence observed in transcriptional networks, and may constitute a general evolu- D hansenii tionary mechanism maintaining the regulation of transcriptional guilliermondii networks Methods L. elongosporus Yeast strains nidulans All immunoprecipitations were performed on strains where the appropriate gene has been endogenously fused to the TAP epitope Rigaut et al. 1999). Sc TAP-tagged strains were obtained from Oper S. octosporus Biosystems In Cg, the 2001HTU strain was used for TAP-tagging Spombe (see below) and deletions. Cg 2001HTU and NCCLS84 were ob- All yeasts in the species phylogeny(Wapinski et al. 2007) tained from atcc nome duplication event (orange star)are noted. Note that the Candida clade does not include Candida glabrate ScYapl. R79K and ScYap4 K252R mutants Endogenously epitope-tagged ScYapl: TAP and ScYap4: TAP strains(Open Biosystems) were used to introduce the appropriate Open questions still remain regarding how arginine and ly. mutation (ScYapl R79K, ScYap4 K252R) via the delitto perfette sine substitutions alter AP-1 DNA-binding motif preference. One method (Storici et al. 2001). In brief, ScYAPI and ScYAP4 were dis. hypothesis is that differences in their electrostatic charges alter the rupted with the URA3 selectable marker from pRS306(Brachman space required to accommodate other positively charged residues et al. 1998)using -100-bp homology Complementary 200-bp of the bZIp nding domain without electrostatic repulsion oligos with a mutation (ScYapl R79K or ScYap4 K252R)were then (Kim and Struhl 1995). This hypothesis suggests that the most transformed by electroporation(Thompson et al. 1998)to remove ine should be associated with YRE.A rather than YREo sites(Kim and Struhl 1995). How ing. This process creates strains possessing endogenous yAP-1 roteins having both the desired mutation and epitope. 11.15)was associated with YRE.A sites An alternative explanation involves a role for AP-1-induced CgApl TAP-tagged strain DNA flexibility. Complexes of yAP-1 protein with the YREO DNa The TAP tag was amplified with-100 otology(CgAP1) from sequence have been associated with an increase in incorporated pFA6a-TAP-HIS3MX6(Longtine et al flexibility (Kim and Struhl 1995)compared with yAP-1/YRE-A media(Amberg et al. 2005). C-terminal integration was verified complexes. A previous report has suggested that changes in dna by PCR and dNA sequencing with protein expression verified flexibility play a key role in determining half-site spacing prefer- by immunoblot (Amberg et al. 2005)with the peroxidase anti- ence and are responsible for differences between in vivo and in peroxidase antibody(Sigma P1291) vitro measurements(Suckow and Hollenberg 1998). Since residue 12 is in close proximity to DNA(Fig. 1B)within the protein-DNA Growth conditions, mRNA expression, and ChIP complex, residue changes may affect the ability of DNa to in- corporate water during binding, thus affecting both yAp-1 DNA Three(CgApl)or two (crapl R7 motif flexibility and binding Dragan et al. 2004b). Interestingly, complete media(Amberg et al. 2005)and treated with 0.03% the higher positive charge of arginine induces a stronger dipole methyl methanesulfonate(Sigma)for I h as performed previously than that of lysine, providing a possible mechanism for the in- (Tan et al. 2008). For mRNA expression analysis, total RNA w crease in the number of incorporated water molecules present at isolated by hot phenol/chloroform extraction and labeled with YRE-O sites and associated changes in DNA flexibility and binding Cy3 or Cys dyes (Invitrogen)(Kuo et al. 2010). Samples were hy- bridized to Agilent expression arrays and washed as recommended In summary, we have shown that conservation of the AP-1 by Agilent(Agilent Technologies regulatory program in yeast occurs through coordinated evolution For ChIP, all TAP-tagged strains were treated as previously of both the sequence of the TF(trans)and in its DNA-binding described (Tan et al. 2008). In brief, cells were fixed with 1% motifs(cis). This finding echoes that of previous studies of protein- formaldehyde for 20 min, inactivated with glycine and washedOpen questions still remain regarding how arginine and ly￾sine substitutions alter AP-1 DNA-binding motif preference. One hypothesis is that differences in their electrostatic charges alter the space required to accommodate other positively charged residues of the bZIP DNA-binding domain without electrostatic repulsion (Kim and Struhl 1995). This hypothesis suggests that the most positively charged residues such as arginine should be associated with YRE-A rather than YRE-O sites (Kim and Struhl 1995). How￾ever, in our findings lysine ( pI = 9.59) rather than arginine ( pI = 11.15) was associated with YRE-A sites. An alternative explanation involves a role for AP-1-induced DNA flexibility. Complexes of yAP-1 protein with the YRE-O DNA sequence have been associated with an increase in incorporated water molecules (Dragan et al. 2004a) leading to a decrease in DNA flexibility (Kim and Struhl 1995) compared with yAP-1/YRE-A complexes. A previous report has suggested that changes in DNA flexibility play a key role in determining half-site spacing prefer￾ence and are responsible for differences between in vivo and in vitro measurements (Suckow and Hollenberg 1998). Since residue 12 is in close proximity to DNA (Fig. 1B) within the protein–DNA complex, residue changes may affect the ability of DNA to in￾corporate water during binding, thus affecting both yAP-1 DNA motif flexibility and binding (Dragan et al. 2004b). Interestingly, the higher positive charge of arginine induces a stronger dipole than that of lysine, providing a possible mechanism for the in￾crease in the number of incorporated water molecules present at YRE-O sites and associated changes in DNA flexibility and binding preference. In summary, we have shown that conservation of the AP-1 regulatory program in yeast occurs through coordinated evolution of both the sequence of the TF (trans) and in its DNA-binding motifs (cis). This finding echoes that of previous studies of protein– protein interaction, which have demonstrated cases in which com￾pensatory mutations are required to maintain protein interaction over evolutionary time (Pazos and Valencia 2008). In the context of transcriptional networks, coevolution gives rise to ‘‘regulatory homeostasis,’’ in which both mutations in a TF and its DNA￾binding motif occur in compensatory fashion to maintain tran￾scriptional regulation. This series of compensatory mutations, which maintains both the transcriptional circuit and regulatory logic, parallels that of previous work demonstrating evolution of alternative transcriptional circuits producing identical logic (Tsong et al. 2006). Such systems of tightly coupled compensatory mu￾tations might serve to counter the widespread divergence observed in transcriptional networks, and may constitute a general evolu￾tionary mechanism maintaining the regulation of transcriptional networks. Methods Yeast strains All immunoprecipitations were performed on strains where the appropriate gene has been endogenously fused to the TAP epitope (Rigaut et al. 1999). Sc TAP-tagged strains were obtained from Open Biosystems. In Cg, the 2001HTU strain was used for TAP-tagging (see below) and deletions. Cg 2001HTU and NCCLS84 were ob￾tained from ATCC. ScYap1.R79K and ScYap4.K252R mutants Endogenously epitope-tagged ScYap1TTAP and ScYap4TTAP strains (Open Biosystems) were used to introduce the appropriate mutation (ScYap1.R79K, ScYap4.K252R) via the delitto perfetto method (Storici et al. 2001). In brief, ScYAP1 and ScYAP4 were dis￾rupted with the URA3 selectable marker from pRS306 (Brachmann et al. 1998) using ;100-bp homology. Complementary 200-bp oligos with a mutation (ScYap1.R79K or ScYap4.K252R) were then transformed by electroporation (Thompson et al. 1998) to remove URA3 by 5-FOA (US Biological) selection and verified by sequenc￾ing. This process creates strains possessing endogenous yAP-1 proteins having both the desired mutation and epitope. CgAp1 TAP-tagged strain The TAP tag was amplified with ;100-bp homology (CgAP1) from pFA6a-TAP-HIS3MX6 (Longtine et al. 1998), transformed by elec￾troporation (Thompson et al. 1998), and selected on complete –his media (Amberg et al. 2005). C-terminal integration was verified by PCR and DNA sequencing with protein expression verified by immunoblot (Amberg et al. 2005) with the peroxidase anti￾peroxidase antibody (Sigma P1291). Growth conditions, mRNA expression, and ChIP Three (CgAp1) or two (ScYap1.R79K, ScYap4.K252R, ScYap3) bi￾ological replicates were grown from OD600 0.2 to 0.8 at 30° C in complete media (Amberg et al. 2005) and treated with 0.03% methyl methanesulfonate (Sigma) for 1 h as performed previously (Tan et al. 2008). For mRNA expression analysis, total RNA was isolated by hot phenol/chloroform extraction and labeled with Cy3 or Cy5 dyes (Invitrogen) (Kuo et al. 2010). Samples were hy￾bridized to Agilent expression arrays and washed as recommended by Agilent (Agilent Technologies). For ChIP, all TAP-tagged strains were treated as previously described (Tan et al. 2008). In brief, cells were fixed with 1% formaldehyde for 20 min, inactivated with glycine and washed Figure 4. All yeasts in the species phylogeny (Wapinski et al. 2007) possess an AP-1 with an arginine. The Candida clade (shaded green) and the whole-genome duplication event (orange star) are noted. Note that the Candida clade does not include Candida glabrata. Coevolu tion wi thin a transcrip tional ne twork Genome Research 5 www.genome.org Downloaded from genome.cshlp.org on November 3, 2010 - Published by Cold Spring Harbor Laboratory Press