Identification of Functional Domains in the Maize Transcriptional Activator CI: Comparison of Wild-Type and Dominant Inhibitor Proteins

Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Identification of functional domains in the maize transcriptional activator CI: comparison of wild-type and dominant inhibitor proteins

Stephen A. Goii,^'^* Karen C. Cone,^ and Michael E. Fromm^-^ ^Plant Gene Expression Center, U.S. Department of Agriculture/University of California at Berkeley, Albany, California 94710 USA; •'Division of Biological Sciences, University of Missouri, Columbia, Missouri 65211 USA

Genes encoding fusions between the maize regulatory protein CI and the yeast transcriptional activator GAL4 and mutant CI proteins were assayed for their ability to trans-activate anthocyanin biosynthetic genes in intact maize tissues. The putative DNA-binding region of CI fused to the transcriptional activation domain of GAL4 activated transcription of anthocyanin structural gene promoters in cl aleurones, cl Rscm2 embryos, and cl r embryogenic callus. Cells receiving these constructs accumulated purple anthocyanin pigments. The Cl acidic region fused to the GAL4 DNA-binding domain activated transcription of a GAL4-regulated promoter. An internal deletion of Cl also induced pigmentation; however, frameshifts in either the amino-terminal basic or carboxy-terminal acidic region blocked tra/is-activation, and the latter generated a dominant inhibitory protein. Fusion constructs between the wild-type Cl cDNA and the dominant inhibitor allele Cl-I cDNA were used to identify the amino acid changes in Cl responsible for the Cl-I inhibitory phenotype. Results from these studies establish that amino acids within the myb-homologous domain are critical for transcriptional activation. \Key Words: Anthocyanins; trans-activation; transcriptional activators; dominant inhibition] Received October 16, 1990; revised version accepted December 13, 1990.

Genetic analysis of maize has identified a large number encoded by the Al (NADPH-dependent reductase) and of genes, including at least four regulatory loci, involved Bronze 1 {Bzl, UDP-glucose flavonol 3-O-glucosyl in the synthesis of the purple anthocyanin pigments transferase) structural genes in aleurones (Dooner and (Coe et al. 1988). For example, the regulatory genes B Nelson 1979; Dooner 1983; K.C. Cone, unpubl.). (Styles et al. 1973) and Pi (Gerats et al. 1984) are required Sequence analysis of the coding regions of the R (Lud for pigmentation in most of the plant body, although wig et al. 1989; Perrot and Cone 1989) and Cl (Paz-Ares strong light can partially overcome the requirement for et al. 1987) genes provides molecular evidence in support PI (Coe et al. 1988). Cl is required for pigmentation of of the proposed regulatory roles of the encoded proteins the aleurone and the embryo but is not required in the (Cone et al. 1986; Paz-Ares et al. 1986; Ludwig et al. plant body (Chen and Coe 1977; Coe 1985). A functional 1989). The B gene is homologous to the R gene (Chandler R gene product is required for anthocyanin pigmentation et al. 1989), and both B and R encode proteins with ho in the aleurone, anthers, and coleoptile (Styles et al. mology to the helix-loop-helix subunit dimerization re 1973), although the B-Peru allele of B can substitute for gion of the myc family of proteins (Chandler et al. 1989; R in the aleurone (Styles et al. 1973). The Rscm2 allele of Ludwig et al. 1989; Perrot and Cone 1989). The Cl gene R conditions anthocyanin pigmentation in the maize encodes a protein with amino-terminal homology to the embryo (Styles et al. 1973). r or b mutations have been DNA-binding domain of members of the myh oncogene found to affect the levels of the mRNAs and/or enzymes family (Paz-Ares et al. 1987). The Cl protein also con encoded by some of the anthocyanin structural genes tains a carboxy-terminal region rich in acidic amino acid examined (Dooner and Nelson 1979; Dooner 1983; residues analogous to the acidic transcriptional activa Chandler et al. 1989; Ludwig et al. 1989). cl (colorless) tion domains found in many transcriptional activators mutants have low levels of the mRNAs and enzymes (for review, see Ptashne 1988). In addition to the various recessive mutant cl alleles, a dominant inhibitory allele ^Corresponding author. Ci-/has been identified (Emerson et al. 1935; Coe 1962). Present addresses: ^Department of Biology, Institute of Molecular Biol ogy, University of Oregon, Eugene, Oregon 97403 USA; ^Monsanto Com This dominant inhibitor allele was not deliberately iso pany, St. Louis, Missouri 63198 USA. lated by maize geneticists but is believed to originate

298 GENES & DEVELOPMENT 5:298-309 © 1991 by Cold Spring Harbor Laboratory ISSN 0890-9369/91 $1.00 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Trans-ictiyation of anthocyanin pigmentation

from ancient cultivated lines (East and Hayes 1911). The GAL4 domains have been used to identify the transcrip presence of this allele blocks pigmentation of the aleu- tional activation domain of the herpesvirus VP16 protein rone in Cl/Cl-I heterozygotes (Emerson et al. 1935; Coe (Sadowski et al. 1988; Chasman et al. 1989) and specific 1962), and pale pigmentation is occasionally observed in amino acid requirements within the Diosophila bicoid the aleurone of maize lines carrying an increased dosage DNA-recognition domain (Hanes and Brent 1989). of the wild-type CI allele (Coe 1962). Cl-I aleurone cells In addition to the analysis of the fimctional domains of contain higher levels of Cl-I mRNA relative to the level wild-type CI, fusions between CI and the dominant in of CI mRNA found in wild-type cells (Cone et al. 1988). hibitor encoded by the CI -I allele were used to analyze Recent sequence analysis of this inhibitory allele re the changes in CI responsible for the dominant inhibi vealed that amino acid substitutions, insertions, and tory phenotype of Cl-I. The results of these studies pro small deletions are present in the C:/-/-coding region vide evidence that a highly conserved amino acid present (Paz-Ares et al. 1990). An 8-bp insertion near the 3' end in the CI myb-like DNA-binding region, but altered in of the gene is predicted to result in premature termina Cl-I, is critical to the trans-activation function of CI tion of the CI-/-encoded polypeptide (Paz-Ares et al. and, therefore, may be essential to the function of other 1990). In vitro transcription/translation of a full-length myb-like proteins. Cl-I mRNA results in the production of a polypeptide that is smaller than the wild-type protein (K.C. Cone, unpubl.). Results High-velocity microprojectile delivery of DNA has Complementation of cl aleurones and embryos been used to develop a system to further analyze the regulation of the anthocyanin biosynthetic genes. The The activity of the Cl genomic subclone (pC/gen; for Al and Bzl genes have been shown to complement al plasmid constructs, see Fig. 1) and the cauliflower mo and bzl mutants when delivered into nonpigmented saic virus 35S (CaMV 35S) promoter-expressed Cl cDNA aleurone cells (Klein et al. 1989). This approach has been (p35SCi) were tested by gene transfer into cl aleurones used to demonstrate CI- and /^-regulated expression of or cl Rscm2 embryos using high-velocity microprojec- introduced Al and Bzl structural gene promoters (Klein tiles as described previously (Klein et al. 1989; Goff et al. et al. 1989) and trans-activation of the Al and Bzl pro 1990). Following incubation, pigmented cells were ob moters by the B regulatory genes (Goff et al. 1990). De served in cl aleurones receiving the Cl DNA constructs livery of a Bz/-luciferase reporter construct has also been (Fig. 2) but not in cl aleurones or cl Rscm2 embryos used to define the cis-acting sequences of the Bzl pro receiving the control vector pMF6 alone (data not moter required for CI- and /^-regulated expression (B.A. shown), cl aleurones receiving plasmids with frame- Roth and M.E. Fromm, unpubl.). Recently, microprojec shifts in the C/-coding region were also unpigmented tile delivery of DNA has been used to demonstrate that (data not shown). These observations demonstrate that the entire anthocyanin biosynthetic pathway can be in the Cl genomic and the CaMV 35S Cl cDNA constructs duced by the delivery of the regulatory genes R (Ludwig used in this study encode functional proteins able to et al. 1990) and B (Goff et al. 1990), and to demonstrate complement the cl regulatory mutation following deliv that the B-I {B-Intense) allele, which is not expressed in ery by microprojectile bombardment. the seed, and the B-Peru allele, which is expressed in the seed, encode functionally equivalent B regulatory pro teins (Goff et al. 1990). The present study was under Tians-activation of the maize Bzl promoter by C\ taken to analyze the functional domains of the CI pro constructs tein. To quantify the activation of anthocyanin structural A number of studies on yeast, invertebrate, and mam gene promoters by the transferred Cl constructs, a re malian transcriptional activators have demonstrated porter plasmid carrying the Bzl promoter driving the that these proteins are modular in structure and that synthesis of firefly luciferase (pfiz/Luc; Fig. 1) was specific domains can function independently and can be codelivered with either pCigen or p35SC/. An internal repositioned within the protein or exchanged between control plasmid expressing chloramphenicol acetyltrans- proteins without a loss of function (Ptashne 1988; Dynan ferase (CAT) from the alcohol dehydrogenase I promoter 1989). Codelivery of CI constructs with anthocyanin (pAdhCAT; Fig. 1) was included in each bombardment as structural gene reporters was used in this study to quan- a control for the efficiency of DNA transfer. Within the titate the functional activity of various CI constructs limits of detection, the activity of the Adhl promoter is encoding deletion, frameshift, and fusion proteins. To not influenced by anthocyanin regulatory loci (data not obtain evidence that the amino-terminal and carboxy- shown). The activation of the pBz/Luc reporter by the terminal domains of CI can function independently of Cl effector plasmids is therefore expressed as a ratio of one another, we have tested the trans-activation func luciferase to CAT activity. Homogenates of cl aleurones tion of an internal CI deletion. The CI basic and acidic or cl Rscm2 embryos receiving the p5z/Luc reporter domains were also analyzed for trans-activation function plasmid without a Cl expression plasmid contained low in fusion constructs with the yeast transcriptional acti ratios of luciferase to CAT activity, whereas tissues re vator GAL4 basic DNA-binding and acidic transcrip ceiving the pBz/Luc reporter plasmid as well as a Cl tional activation domains. Similar fusions with these effector plasmid contained, on average, 85- to >300-fold

GENES & DEVELOPMENT 299 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Goff et al.

Effector Constructs trans-activate the Bzl promoter (Table 1). These results pMF6 l-crtivaM->-T-'^Adhl Inl" clearly demonstrate that the presence of the CI gene product activates the Bzl promoter in cl aleurones and pClgen I CI Promoter> I—B'^^g^'^SMMH- in cl Rscm2 embryos, as expected from previous genetic studies (Dooner and Nelson 1979; Dooner 1983; Cone et p35SCl I —CiMV 368 -^ ^ al. 1986). j/Stop Codon p35SClfs26 I-ctw tiM->> CI cDK. ffiUji^TTH Adhl Int ^^/T- Frameshift Complementation and tians-activation with an p35SClfs27Q internal deletion of Cl Dclelion Frameshift To obtain evidence that the basic amino-terminal region of Cl (amino acids 1-114), with homology to the DNA- pCALCl [-c.llv»s-»^^^^^^^ ],^^~'"'^CAU DNA Bindin^Cl cWAS;;?igo;r7- binding domain of the myb oncogene products, and the carboxy-terminal acidic region (amino acids 234-273)

PGALVP16 l.---^--^^^ can function as independent domains, a Cl-coding se quence with an internal deletion (amino acids 117—144) pGALBindl was constructed and placed under control of the CaMV 35S promoter. This Cl internal deletion (pCIdel3; Fig. 1) pC 1 CAI,7,1 B^^^}''^7!^7~^^^E^M[Zi!^«i=*Lfi.}{IS 5 was delivered into cl aleurones and cl Rscm2 embryos and tested for its ability to complement the cl genotype pciCAi.e 1 Lzfi !?z){iis:^j:in as well as tra/is-activate the Bzl reporter p5zlLuc. Fol lowing delivery of pCldel3 to cl aleurones, purple cells CDC3«i;!r^s-li:ilIIK;jnil were observed (Fig. 2). Likewise, codelivery of pCidel3 with the pBzlLuc reporter into cl Rscm2 embryos re sulted in a 25-fold increase of luciferase to CAT activity relative to cells receiving control plasmids (Table 1). fv;;«:^n!^Hj:iJ These results demonstrate that amino acids 117-144 of the Cl protein are not essential for either trans-activa tion of a specific structural gene promoter nor activation of the entire anthocyanin biosynthetic pathway, al pC'lChilOII ••"NA (A...i..ji!).;ii though this deletion polypeptide appears to be less effi cient at Bzl promoter fri7/7s-activation than the intact p:if>sn I'ci-u [ Cl protein (Table 1). This observation provides evidence

Rcporler Const mots that the Cl amino-terminal basic region and the Cl car :i">:i.;v|llllllllll boxy-terminal acidic region can function as independent Bv.l Intl domains. :"Hff.~:?:iii

TAT*

pGAU,ur2 IIIIIIJIIIIU » Adhl Inl .jist:i;iii Activation of a GAL4-regulated promoter by a GAL4-CI fusion protein nornbardmrnt Control The carboxy-terminal acidic region of the Cl protein was pAdhCAT I 'ADHl'romolor>.} -j^),, ,„,' ' - |JlCATS§iH tested for its ability to functionally replace the transcrip tional activation domain of the yeast GAL4 protein. The coding sequence for the GAL4 DNA-binding domain, Figute 1. Plasmid structures. Plasmids used in this study are amino acids 1-147 (Ma and Ptashne 1987), was fused to not necessarily drawn to scale. Components of plasmids are the coding sequence for the carboxyl terminus of Cl described in Materials and methods. Effector plasmids encode (amino acids 173-273) and placed under CaMV 35S pro trfl/js-acting factors. Reporter plasmids were used to determine moter control [pGALCl; Fig. 1). As a positive control, a the activity of effector constructs. The bombardment control plasmid encoding a fusion between the DNA-binding do plasmid was used to normalize efficiency of gene delivery. Ab main of GAL4 and the transcriptional activation domain breviations are as follows: (Adhl Int) Maize alcohol dehydroge of herpes simplex virus VP16 was constructed nase 1 intron 1; (Nos 3') nopaline synthase 3' polyadenylation [pGALVPld; Fig. 1). As a negative control, a plasmid en region; (35SMin) minimal cauliflower mosaic virus 35S pro moter. The remaining abbreviations are described in the text. coding only the GAL4 DNA-binding domain was also constructed (pGAIBindl; Fig. 1). pGALCl and pGALVP16 were tested for their ability to trans-activate reporter plasmids containing a minimal 35S promoter higher ratios of luciferase to CAT activity (Table 1). Con with and without GAL4 recognition sites (pGALLuc2 structs with frameshift mutations in either the amino- and p35SMin, respectively; Fig. 1). terminal or carboxy-terminal coding region {p35SClfs26 Delivery of either pGALLuc2 or p35SMin without an or p35SClfs27, respectively; Fig. 1) were not observed to effector plasmid into embryogenic callus tissue resulted

300 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Trans-activation of anthocyanin pigmentation

^i •^,>. '',;f

Figure 2. Induction of anthocyanin synthesis in maize aleurones. Maize cl displaying purple anthocyanin-producing cells. Aleurones were bombarded with the following plasmids, described in Materials and methods, and photographed at 50x magnification: [A] Cl genomic plasmid, pClgen. [B] An expressed CI cDNA, p35SCI. (C) An expressed internal deletion of the Cl cDNA, pCldelS. (D) A fusion between the Cl amino-terminal basic coding sequence at Cl amino acid 117 and the GAL4 acidic transcriptional activation domain, pClGALJ.l. (£) A fusion between the Cl amino-terminal basic coding sequence at Cl amino acid 144 and the GAL4 acidic transcriptional activation domain, pClGAL6.l. [F] A fusion between the Cl amino-terminal basic coding sequence at Cl amino acid 258 and the GAL4 acidic transcriptional activation domain, pC2GAL10. in low levels of luciferase to CAT activity (Table 2). Em- yeast regulatory protein GAL4 [region II, amino acids bryogenic callus tissue has been shown to be a reliable 768-881 (Ma and Ptashne 1987)] was fused to Cl at substitute for aleurones or embryos in such trans-acti amino acids 117 (pCiGAI7.1), 144 (pCJGAI6.I), or 258 vation assays when the appropriate regulatory genes are (pCiGALlO). These constructs express fusion proteins transferred into these cells (Goff et al. 1990). Codelivery with the putative DNA-binding region of C1 and varying of p35SMin with any of the effector plasmids also re lengths of the remainder of Cl (the intact Cl protein is sulted in low ratios of luciferase to CAT activity (Table 273 amino acids) fused to a known transcriptional acti 2). The expressed GAL4 DNA-binding domain alone was vation domain (Fig. 1). Each of these constructs was de observed to induce reporter constructs with and without livered to cl aleurones by high-velocity microprojectile GAL4 DNA-binding sites approximately fivefold (Table bombardment, and each was observed to induce pigmen 2); therefore, this induction is considered nonspecific. tation of recipient cells (Fig. 2). Constructs expressing CodeUvery of pGALVPld or pGALCi with a reporter the GAL4 transcriptional activation domain fused in containing GAL4 DNA recognition sites resulted in a 53- close proximity to the putative DNA-binding region of and 76-fold increase in luciferase to CAT activity, re Cl (pClGAL7.1 and pClGALS.l] resulted in proteins spectively (Table 2). These observations demonstrate that stimulate only moderate pigment accumulation, that the Cl acidic region can function as a transcrip whereas constructs expressing either the intact Cl pro tional activation domain when fused to a well-character tein or a fusion of the GAL4 transcriptional activation ized heterologous DNA-binding domain. domain at amino acid 258 of Cl (pCIGALlO) stimulate intense pigmentation (Fig. 2). Previous work has demonstrated that both an ex Complementation and trans-activation with C1-GAL4 pressed Cl gene and an expressed B gene are required for fusion proteins pigment induction or trans-activation of the Bzl pro To obtain more definitive evidence that the Cl protein moter in cl r embryogenic callus (Goff et al. 1990). Code functions as a transcriptional activator, a coding se livery of plasmids containing an expressed B gene and quence for the transcriptional activation domain of the various C1-GAL4 fusion proteins with the Bzl reporter

GENES & DEVELOPMENT 301 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Gofi et al.

Table 1. Trans-activation of pBzlLuc by CI constructs

cl Rscm2 embryc) S cl aleurones Plasmids + pBziLuc + pAdhCA T Luc/CAT induction Luc/CAT induction pMF6 0.05 ± 0.01 1.0 0.04 ± 0.01 1.0 n = 12 n = 12 p35SCi 4.18 ± 0.94 85 ± 19 3.64 ± 0.68 91 ± 17 n = n n = 9 pCigen 16.6 ± 3.87 339 ± 79 13.8 ± 5.02 346 ± 125 n = 6 n = 6 pCJdel3 1.23 ± 0.06 25 ± 1.3 ND ND n = 3 pCifs26 0.07 ± 0.03 1.35 ± 0.6 0.02 ± 0.00 0.40 ± 0.03 n = 3 n = 3 pCifs27 0.17 ±0.02 3.44 ± 0.3 0.04 ± 0.00 0.82 ± 0.05 n = 3 n = 3 Plasmid constructs are described in Fig. 1 and were delivered to cl Rscm embryos and to cl aleurones by microprojectiles. Following incubation for 24—48 hr with illumination, tissues were homogenized and enzyme activities determined. Activities are expressed as a ratio of firefly luciferase (Luc) to CAT. Independent bombardments were analyzed and data are presented as mean values of n repetitions plus or minus S.E.M.. Induction values are determined as the ratio of luciferase to CAT activities to those activities observed with control plasmids. Induced activities are expressed as mean values of n repetitions plus or minus S.E.M. Induction of pBzJLuc by the Cl deletion pCidel3 was not determined (ND) in cl aleurones. plasmid pBzlLuc into cl r embryogenic callus cells re designated Cl-I (Emerson et al. 1935; Coe 1962), blocks sulted in a 63- to 145-fold increase in the ratio of lu anthocyanin biosynthesis in aleurones and encodes a ciferase to CAT activity relative to cells receiving re protein with several amino acid changes. The most ob porter and control plasmids (Table 2). pClGALlO was vious alteration is that the Cl-I protein is truncated at its also observed to trans-activate pBzlLuc > 100-fold in cl carboxyl terminus and thus lacks 21 amino acids that aleurones (data not shov^^n). However, both pClGALJ.l comprise much of the acidic domain (Paz-Ares et al. and pClGALS.l trans-activated pBziLuc on the average 1990). This raises the possibility that disruption of the of only 10- to 20-fold in cl aleurones (data not shown). transcriptional activation domain not only leads to a fail The cause of this variability in pBzlLuc trans-activation ure to trans-activate but also generates the inhibitory by the different C1-GAL4 fusions in cl aleurones versus phenotype of Cl-I. To test this possibility, the Cl con cl I embryogenic callus is not yet understood. Trans- struct containing a frameshift in the acidic domain activation of pBzlLuc by these C1-GAL4 fusions re (amino acid 258, p35SCifs27) was codelivered with an mains dependent on the presence of an expressed B gene expressed Cl construct (either p35SCl or pCJfgen) and in cl r embryogenic callus (data not shown), as was the the Bzl reporter plasmid pBziLuc. Codelivery of equal case for trans-activation of pBzILuc by wild-type Cl amounts (by weight) of an expressed Cl construct and (Goff et al. 1990). p35SClfs27 reduced pBzlLuc trans-activation by —50% A frameshift mutation at amino acid 258 of the Cl (Fig. 3); and at higher amounts of p35SCifs27, inhibition coding sequence (p35SCifs27) generates a Cl protein of trans-activation approached 90% (Fig. 3). These re with an additional 74 amino acids fused to amino acid sults demonstrate that it is possible to generate a domi 258 of Cl. The protein encoded by the gene with this nant inhibitor of anthocyanin structural gene transacti- frameshift mutation neither stimulated pigmentation vation by the introduction of a frameshift in the carboxy- (data not shown) nor trans-activated pBziLuc (Table 1). terminal acidic transcriptional activation domain of CI. Taken together with the data presented above, these re Analogous experiments using control plasmids lacking sults demonstrate that the Cl carboxy-terminal acidic Cl (either pUCI8 or pMF6) or a construct with a frameshift mutation in the coding sequence at amino acid 32 region is essential for CJ-mediated trans-activation but of Cl (p35SC126) did not result in a decrease in pBzlLuc can be replaced by the acidic transcriptional activation trans-activation (data not shown). domain of a known yeast regulatory protein. The Cl- GAL4 fusion construct encodes not only a functional activator of a specific anthocyanin structural gene pro moter but also the entire pigmentation pathway. In ad Analysis of the alterations in Cl-I that confer dition, these C1-GAL4 fusion constructs remain depen dominant inhibition dent on the presence of an expressed B construct (data To determine the location of the amino acid change, or not shown). changes, in the Cl protein that generates the Cl-I dom inant inhibition, in-frame exchanges were made between A frameshift mutant of Cl inhibits trans-activation of the coding sequence of the Cl and Cl -I cDNAs at vari the pBzlLuc ous positions, and the resulting constructs were tested A genetically identified dominant inhibitory allele of Cl, for both trans-activation and inhibition of trans-activa-

302 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Trans-activation of anthocyanin pigmentation

Table 2. Trans-activation of pBzl Luc by GAL4^C1 and trans-activate pBzlLuc at only 11% of the wild-type Cl C1-GAL4 fusion constructs level (Table 3) and was not observed to generate anthocyanin-producing cells (data not shown). Codelivery of GAL4-C1 cl I Embryogenic callus pClC/-l with intact Cl resulted in 60% inhibition of Plasmids pBzlLuc trans-activation (Table 4). + pAdhCAT Luc/CAT induction A complementary fusion construct between amino ac pMF6 0.42 ±0.12 1.0 ids 1 and 142 of Cl-1 and amino acids 143 and 273 of Cl p35SMin n = 6 (pC/C2-2; Fig. 1) has two alterations relative to Cl. pGALBindl 2.14 ±0.34 5.1 ±0.8 Amino acid Asp^°^ within the basic myh homology do p35SMin 73 = 6 main is changed to Glu^°\ and amino acids Ala^^^- pGALVP16 0.50 ± 0.02 1.2 ±0.0 Gly^^'^ are deleted (Paz-Ares et al. 1990). pCICl-1 tians- p35SMin n = 6 activated pBzlLuc at <1% of the wild-type Cl level (Ta pGALCl 0.73 ±0.13 1.7 ± 0.3 ble 3) and inhibited Ci-mediated trans-activation of p35SMin n = 6 pBzlLuc by —85% (Table 4). These results demonstrate pMF6 0.61 ± 0.05 1.0 that changes within both the basic DNA-binding domain pGALLuc2 n = 6 and the acidic transcriptional activation domain can pGALBindl 2.15 ±0.15 3.5 ± 0.2 block the trans-activation function of Cl. In addition, pGALLuc2 n = 6 pGALVP16 32.3 ± 2.8 52.7 ± 4.5 they demonstrate that the inhibitory property of Cl -I is pGALLuc2 n = 6 not due to a single change within one functional domain pGALCl 46.5 ± 11.7 76 ± 19 of the Cl protein but, rather, that the Cl-I dominant pGALLuc2 73 = 6 inhibitor phenotype is most likely due to an additive effect of multiple alterations. C1-GAL4 To test the possibility that the single conservative Plasmids change within the basic domain with myb homology is + pBzlLuc cl I Embryogenic callus partially responsible for the inhibitory phenotype of Cl- pAdhCAT I, site-directed mutagenesis was used to change Cl p35SBPeru Luc/CAT induction pMF6 0.03 ± 0.01 1.0 n = 6 p35SCi 2.8 ± 0.3 93 ± 9 n = 6 pClGALJ.l 1.9 ±0.4 63 ± 13 n = e pClGAL6.\ lA ± 0.2 80 ± 7 n = 6 CD pClGALlO 4.3 ± 0.9 145 ± 30 O ..-- n = 6 Plasmid constructs are described in Fig. 1 and were delivered to O 38 -- cli embryogenic callus tissue by microprojectiles. Following incubation for 24—48 hr without illumination, tissues were ho mogenized and enzyme activities were determined. Activities are expressed as described in the footnote to Table 1. In GAL4- o Cl constructs, plasmids delivered are those listed plus the in ternal control plasmid pAdhChT. In C1-GAL4 constructs, plasmids delivered are those listed plus pBziLuc, pAdhCAT, and p35SB-Peru.

123458789 10 li C-Terminal Frameshift/lnlacl Cl Figure 3. Inhibition of pBzJLuc trans-activation by a frame- tion of pBzlLuc. The expressed intact Cl-I protein shifted construct. Inhibition of CJ-mediated trans-activation of (p35SCi-/; Fig. 1) fails to trans-activate pBzlLuc (Table pBziLuc was determined by codelivery of a construct encoding 3), and inhibits Ci-mediated pBziLuc trans-activation a frameshift at amino acid 258 of the Cl cDNA (p35SC2fs27) by as much as 95% (Table 4). A construct encoding with a wild-type Cl construct (either p35SCJ or pCigen). The amino acids 1-220 of Cl fused to the corresponding car- results of individual bombardments and enzyme assays are ex boxy-terminal segment of Cl-1 (amino acids 221-252, pressed as a ratio of firefly luciferase activity to CAT activity above the background activity of tissues receiving expression pCIC7-l; Fig. 1) encodes a protein with amino acid sub vector plasmid without an inserted coding sequence. Determi stitutions at position 228 (P for L) and 249-252 (SWTT nation of the enzyme is as described in Materials and methods. for DVRA) and terminates following amino acid 252 The ratios of frameshifted construct to intact coding sequence rather than at the wild-type position following amino construct are weight ratios (i.e., |xg/(xg). Zero (0) represents no acid 273 (Paz-Ares et al. 1990). pCiCM was observed to frameshifted construct present in the bombardment.

GENES & DEVELOPMENT 303 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Goff et al.

Table 3. Tisins-activation of pBzlLuc by Cl-I and Cl/Cl-I fusion constructs cl I Embryogenic callus pBzlLuc p35SB-Peru Luc/CAT induction wild type (%) pMF6 0.04 ± 0.01 n = 6 p35SCJ 7.6 ± 0.22 180 ± 7.5 100.0 n = 3 p35SCi-I 0.04 ± 0.00 0.95 ± 0.12 0.0 n = 3 pC:?C/-l 0.85 ± 0.01 20.2 ±0.14 11.2 n = 3 pC7Ci-2 0.06 ± 0.01 1.44 ±0.11 0.8 23 = 3 pCiGlu'°' 0.17 ±0.02 3.95 ± 0.49 2.2 n = 6 Plasmid constructs are described in Fig. 1 and were delivered to cl i embryogenic callus tissue by microprojectiles. Following incubation for 24—48 hr without illumination, tissues were homogenized and enzyme activities determined. Plasmids delivered are those listed plus pBziLuc, Y>AdhCKT, and p35SB-Peru. Activities are expressed as described in the footnote to Table 1. The induction is expressed as the percentage of induction observed relative to the intact Cl cDNA construct. amino acid Asp^"' to Glu"" (pCiGIu^"S Fig. 1). This mutation, we have not observed an inhibitory phenotype single amino acid change in the Cl protein was observed associated with this single mutation (data not shown). to decrease trans-activation of pBz:/Luc to —2% of the wild-type level (Table 3). Reversion of this mutant cod Discussion ing sequence to Asp'"' by site-directed mutagenesis re stored the wild-type level of Ci-mediated trans-activa Microprojectile delivery of a Cl genomic clone or an tion (data not shown). Although C2-mediated pBziLuc expressed Cl cDNA into cl R aleurone or cl Rscm2 trans-activation is decreased dramatically by this single embryo tissues activated the anthocyanin biosynthetic

Table 4. Inhibition of pBzlLuc tians-activation by Cl-I and Cl/Cl-I fusion constructs cl r Embryogenic callus Pla«mid« +i*Adh(^ AT pBziLuc p35SB-Peru Ratio Luc/CAT induction Wild type (%) pMF6 0.03 0.01 n 6 pMF6 2.05 0.22 63.9 ± 6.8 100.0 pClgen n 6 p35SCi-/ 0.09 0.02 2.7 ± 0.51 4.2 pCigen n 6 p35SCJfs27 0.79 0.14 24.8 ± 4.2 38.8 pCigen n 6 pMF6 0.04 0.01 n 6 pMF6 11.5 1.00 279 ± 23.8 100.0 p35SCi n 6 pClCl-l 4.7 0.33 110 ± 7.9 39.4 p35SCl n 6 pMF6 0.02 0.01 n 3 pMF6 1.70 0.28 106 ± 17.4 100.0 pClgen n 6 pCICl-2 0.26 0.04 16.0 ± 2.4 15.1 pCigen n Plasmid constructs are described in Fig. 1 and were delivered to cl r embryogenic callus tissue by microprojectiles. Following incubation for 24—48 hr without illumination, tissues were homogenized and enzyme activities determined. Activities are expressed as described in the legend to Table 1. Plasmids delivered are those listed plus pBzlLuc, pAdhCAT, and p35SB-Peru. The induction is expressed as the percentage of induction observed relative to the intact Cl cDNA construct, divided into sections to represent values obtained in individually controlled experiments.

304 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Troas-activation of anthocyanin pigmentation pathway and resulted in purple pigmented cells. This an acidic transcriptional activation domain of a well- complementation of the cl mutation by the cloned CI studied transcriptional regulator. These results, together gene and expressed Cl cDNA confirms the identity of with the results discussed above, provide in vivo evi these constructs as functional Cl regulatory genes. dence that the product of the Cl gene functions as a These, and previously published results using the R transcriptional activator that utilizes a myb-like DNA- (Ludwig et al. 1990) or B (Goff et al. 1990) anthocyanin binding domain with an acidic transcriptional activation regulatory genes, demonstrate that a complex biosyn- domain. The basic domain of myb oncogene products thetic pathway can be activated by microprojectile de consists of three imperfect repeats of 51-52 amino acids livery of the appropriate regulatory genes. The intro (Gerondakis and Bishop 1986; Rosson and Reddy 1986), duced Cl gene functions in aleurones, embryos and, as and this domain has been shown to bind both specifi described previously, embryogenic callus (Goff et al. cally and nonspecifically to DNA (Oehler et al. 1990). 1990). Codelivery of an expressed Cl gene and a Bzl Deletion analysis of the myb DNA-binding region was reporter construct resulted in the activation of pBzlLuc. used to demonstrate that the first repeat (missing from This quantitative trans-activation assay allows for more the maize Cl protein) is not essential for DNA binding, detailed studies on the structure and function of plant whereas the second and third repeats are crucial for spe regulatory gene products. cific site recognition (Howe et al. 1990; Oehler et al. An intemal deletion of the Cl protein that does not 1990). Several myb oncogene products have been re remove any of the amino-terminal basic region or the ported to function as transcriptional activators that act carboxy-terminal acidic region trans-activated pBzlLuc through specific DNA-binding sites (Ness et al. 1989; and stimulated anthocyanin synthesis. Such deletion Nishina et al. 1989; Sakura et al. 1989; Ibanez and Lip- analysis has shown previously that up to 80% of either of sick 1990). It is noteworthy that a Bzl promoter muta the yeast regulatory proteins GAL4 or GCN4 can be tion in a 6-bp region with homology to the myb consen eliminated without a total loss of regulatory function sus DNA-binding site has recently been demonstrated to (Hope and Struhl 1986; Ma and Ptashne 1987). The in decrease expression of the Bzl promoter in Cl R maize temal deletion of Cl used in this study demonstrates embryos (B.A. Roth and M.E. Fromm, unpubl.). Multim- that amino acids 117-144 are not crucial for Cl trans- ers of the myb-consensus binding site (together with a activation function and provides evidence that the putative binding site for R or B] placed upstream of an amino-terminal basic region and the carboxy-terminal inactive CaMV 35S core promoter, confer Cl and B in- acidic region can function as independent domains. ducibility to this promoter (S.A. Goff and M.E. Fromm, A fusion containing the DNA-binding domain of unpubl.). Together, these data strongly suggest that Cl- GAL4 and the acidic carboxy-terminal domain of Cl regulated expression of the Bzl promoter is mediated {pGALCl ] was found to stimulate transcription of a min through a myb consensus DNA-binding site. imal CaMV 35S promoter containing GAL4 DNA recog Fusions between the wild-type Cl allele and the dom nition sites. An expressed GAL4 DNA-binding domain inant inhibitor Cl-I allele were used to analyze some of alone (pGALBindl) did not activate this promoter, and the numerous amino acid changes present in Cl-I. Re neither pGALCl nor pGALBindl activated a minimal sults from these experiments demonstrate that changes CaMV 35S promoter lacking GAL4 DNA recognition in both the amino-terminal basic and carboxy-terminal sites. These results demonstrate that the Cl acidic re acidic domains of Cl can decrease or eliminate the trans- gion can function as a transcriptional activation domain activation function of the resulting gene products. As when directed to a target promoter by the yeast GAL4 expected from genetic studies, the intact Cl-I protein DNA-binding domain. Many types of transcriptional ac fails to activate pBziLuc and is a potent inhibitor of Cl- tivation domains have been identified (Ptashne 1988). mediated pBzILuc tra/is-activation. A fusion between Domains rich in glutamine, proline, and acidic amino Cl 1-220 and CM 211-251 (pC7C/-l), and a frameshift acid residues are commonly found in regulatory proteins within the Cl acidic region at amino acid 258 (Ptashne 1988; Brendel and Karlin 1989). Acidic tran (p35SClfs27) both encode proteins with altered tran scriptional activation domains have no apparent rigid scriptional activation domains. Like Cl-I, both of these amino acid sequence requirements but are high in neg constructs failed to trans-activate pBzlLuc efficiently. ative charge and capable of forming amphipathic helices Lack of trans-activation by these constructs provides fur (Giniger and Ptashne 1987; Ma and Ptashne 1987). The ther evidence that the intact Cl acidic region is essential sequence of the acidic region of Cl is proposed to be for efficient transcriptional activation. Both the frame- capable of forming an amphipathic helix between resi shift (p35SClfs27) and the Cl-CI fusion [pClCI-l] en dues 246 and 260 (Paz-Ares et al. 1990), in agreement code proteins that weakly inhibit Cl-mediated pBzlLuc with the transcriptional activation properties of the Cl trans-activation relative to the potent inhibitor encoded protein. by Cl-/. Fusion proteins containing the myb-homologous basic The Cl-I protein and these altered Cl proteins could domain of Cl and the transcriptional activation domain inhibit pBzlLuc trans-activation via one of several dif of yeast GAL4 (region II; Ma and Ptashne 1987) were ferent mechanisms. They do not simply block wild-type found to stimulate transcription from pBzlLuc and acti Cl gene expression since they inhibit CaMV 35S- vate anthocyanin biosynthesis. These observations dem expressed Cl products, and Cl/Cl-I cells contain both onstrate that the acidic region of Cl can be replaced by Cl and Cl-I transcripts (Paz-Ares at al. 1990). Expression

GENES &. DEVELOPMENT 305 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

GoH et al. of a CI DNA-binding domain lacking a transcriptional the various mammalian or insect myb-homologous do activation domain could sterically block the function of mains could also decrease the transcriptional activation the intact protein, as demonstrated for both GCN4 and function of these proteins dramatically. Detailed studies GAL4 DNA-binding domains synthesized without acti on the structure of the pCIGlu^°^ DNA-binding domain vation domains (for review, see Struhl 1989), and for the could provide insight into the mechanism of myb-medi- Escherichia coh lac repressor when a repressor binding ated transcriptional activation. Single amino acid substi site is located within a heterologous promoter (Deuschle tutions within the helix-turn-helix DNA-binding do at al. 1990). Alternatively, if the CI protein normally main of bacteriophage X repressor have been shown to functions as a subunit in a heterodimer, the C1-I-, dramatically decrease transcriptional activation (Bush pClCI-l-, or p35SClfs27-encoded proteins could block man et al. 1989), possibly by alteration of an RNA poly the wild-type protein or protein complex by the forma merase direct contact point. Specific amino acid residues tion of mixed dimers (Herskowitz 1987). The predicted located within the MyoDl DNA-binding domain are polypeptide encoded by p35SCifs27 contains a basic car- also believed to be critical for transcriptional activation boxyl terminus (net charge of -1-3) rather than the wild- function (Davis et al. 1990). Unlike pCICl-2, however, type acidic domain (net charge of -9). It is also possible, pCJGlu^"^ was not observed to inhibit p5ziLuc trans- therefore, that this basic carboxyl terminus directly dis activation by intact CI. This result suggests that the rupts essential components of the transcriptional ma inhibitory property of pCICl-1 results from either the chinery. Further studies of these proteins are necessary Ala^^^ to Gly^" deletion adjacent to the myb-homolo- to distinguish between the possible modes of inhibition. gous region or from the combination of this deletion and A fusion between Cl-I 1-142 and CI 143-273 {pCICl- the Asp^°^ to Glu'°^ substitution. 2) contains a conservative substitution within the myb Although there is no definitive evidence that the prod homology domain (Asp^°^ to Glu^°^) and a deletion of ucts of CI and B/R act directly on anthocyanin struc amino acids Ala^^^ to Glu^" (Paz-Ares at al. 1990), and tural gene promoters, the results to date strongly suggest tr^ns-activates pBzlLuc at —1% the level of the wild- such a mode of action. It is noteworthy that the CI tran type protein. This construct was shown to inhibit trans- scriptional activation domain can activate transcription activation of pBziLuc by intact CI, although not to the of a GAL4-regulated promoter independent of the B ox R extent of the intact Cl-I protein. Together with the re gene products when fused to the GAL4 DNA-binding sults described above, these observations provide sup domain. Mounting evidence suggests that transcrip port for the interpretation that the inhibitory phenotype tional activation domains may function by contact with of Cl-I results from the additive effects of changes additional proteins, designated coactivators or adaptors within different domains of the protein. Mutations (Berger et al. 1990; Kellcher et al. 1990; Liu and Green within both the basic DNA-binding and acidic transcrip 1990; Pugh and Tjian 1990), or by direct contact with the tional activation domains decrease pBziLuc trans-acti TATA-binding factor, TFIID (Stringer et al. 1990). The vation and contribute to CI-/-mediated inhibition. Such CI DNA-binding domain fused to the transcriptional ac a model is consistent with the genetic evidence that re tivation domain of GAL4 remains dependent on an ex version of the Cl-I allele to the wild-type CI allele oc pressed B gene to achieve trans-activation of pBziLuc, as curs extremely infrequently if ever (Coe 1962) and also does CI. Neither CI, C1-GAL4, nor B/R alone can acti with the numerous sequence differences found between vate pBzlLuc, suggesting that a synergistic interaction of Cl-I and CI (Paz-Ares et al. 1990). Together, the avail these regulatory proteins may be necessary to achieve able evidence suggests that the dominant inhibitory CI -I anthocyanin structural gene promoter activity. Interac allele arose by multiple events that alone only partially tions between the rat glucocorticoid receptor and GAL4 eliminated CI trans-activation or generated Cl-I inhibi derivatives (Kakidani and Ptashne 1988), or the mamma tion. lian transcriptional activator ATF and GAL4 derivatives Site-directed mutagenesis of CI Asp'°' to Glu'°' have been described recently to synergistically activate (pC/Glu^°^) decreased pBz J Luc trans-activation to —2% their respective target promoters (Carey et al. 1990; Lin of the wild-type level. This result shows that a single et al. 1990). Likewise, the BASl gene of yeast, which conservative amino acid substitution in the DNA-bind contains a myb-like DNA-binding domain, requires an ing domain of CI can have very dramatic effects on tran additional protein encoded by BAS2 to activate HIS bio- scriptional activation function. It is possible that this synthetic genes (Tice-Baldwin et al. 1989). Both an ex conservative amino acid change decreases CI DNA bind pressed CI (or C1-GAL4 fusion) and B gene are required ing; however, the Cl-I protein (with Glu^°*) is believed for activity of a minimal CaMV 35S target promoter with to bind DNA (cited in Paz-Ares et al. 1990). This Asp a synthetic trimer of a myb and myc consensus DNA- residue is conserved evolutionarily between the various binding site (B.A. Roth, S.A. Goff, and M.E. Fromm, in members of the myb oncogene family (Bergmann et al. prep.). The results of this study demonstrate that if an 1981; Klempnauer et al. 1982, 1986; Gonda et al. 1985; interaction between CI and B/R is required for pBzlLuc Boyle et al. 1986; Shen-Ong et al. 1986; Peter et al. 1987) trans-activation, the interacting region in CI must be and may occupy a position in the DNA-recognition helix localized to the amino terminus. Alternatively, addi (Ohlendorf et al. 1982; Lehming et al. 1987; Aggarwal et tional adaptor or coactivator proteins may recognize R/B al. 1988; Jordan and Pabo 1988; Otting et al. 1988). This and the amino terminus of CI. Although B and R have conservation suggests that an analogous change in any of been shown to encode proteins with a potential helix-

306 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Trans-activation of anthocyanin pigmentation loop-helix dimerization motif, it is not likely that R/B Pstl hised to Cl-I 221-252, PstI-£coRI) and pCICl-2 {Cl-I 1- proteins use this motif to interact with CI, since the CI 142, £coRI-AatII fused to Cl 143-273, AatII-£coRI), and con protein has not been found to contain such a motif. In tain the 3'- and 5'-untranslated regions of the respective Cl and vitro analysis of the interactions between the CI and R/B Cl-I cDNAs. pClGlu'"^ was generated by site-directed mu tagenesis (Kunkel et al. 1987) of the wild-type Cl cDNA and proteins and DNA may clarify the nature of their mutual codes for glutamic acid at position 101 rather than the wild-type dependence. asparagine. The minimal 35S reporter and 35S reporter contain ing GAL4 recognition sites (Giniger et al. 1985) were con structed from pMA556 and pMA558 (Ma et al. 1988) by ligation Materials and methods of the vector and CaMV 35S promoter region (Pstl-Hiudlll) of Plant materials pMA556 (without GAL4 recognition sites) and pMA558 (with — 10 GAL4 recognition sequences) to a Pstl-Sacl Adh intron- Aleurones and embryos were isolated from cl or cl Rscm2 im luciferase fragment and a Sacl-Hindlll nos 3' end fragment. All mature kernels as described previously (Klein et al. 1989). An inserts are carried in plasmid vectors derived from pUC. Plas immature ear of A188 crossed with B73 was used as the source mids transferred into cells were isolated by equilibrium ultra- of 1-mm embryos for initiating the embryogenic callus as de centrifugation. scribed (Kamo and Hodges 1986). Embryogenic callus tissue was prepared for gene transfer by spreading a thin "lawn" of tissue in a circular area (3-cm diameter) on agarose plates containing N6 High-velocity microprojectile bombardment media (Lowe et al. 1985). Plasmid DNA was delivered to intact tissues by high-velocity microprojectiles using the Biolistics device as described (Klein Plasmids et al. 1988, 1989). A total of 8 ^lg of plasmid DNA (at 1.0 mg/ml) was precipitated onto microprojectiles by the addition of 25 yA Structures of all plasmid inserts used in this study are presented of 50 mg/ml 1.0 ji-m tungsten, followed by the addition of 25 in Figure 1. The pMF6 expression vector consists of the CaMV (xl of 1.0 M CaCli and 10 p-l 100 mM spermidine free base. Mi 35S promoter (0.5 kb), the Adhl first intron (0.5 kb), and the croprojectiles were allowed to settle for ~5 min at room tem nopaline synthase [nos] polyadenylation region (0.25 kb) as de perature, and 25 yA of overlying solution was removed. As an scribed previously (Callis et al. 1987), with additional restric alternate procedure, a fivefold increase in all components was tion sites between the Adhl intron and nos fragments. The Cl used, and 300 p-l of the supernatant was removed. The resulting cDNA (1.1 kb) was isolated from a XgtU library prepared from microprojectile solution was vortexed, and 2.5 \il of this slurry poly(A)^-selected mRNA from color-converted W22 (Brink der was loaded onto each macroprojectile for bombardment. Tis ivation) kernels 30 days postpollination (K.C. Cone, unpubl.). sues were analyzed after 24—48 hr incubation at 24°C. Aleurones The Cl-I cDNA was isolated from a Xgtll library (K.C. Cone, and embryos were incubated with illumination, and embryo unpub.) prepared from homozygous Cl-I kernels (Coe 1962). genic callus was incubated without illumination (see table foot The Cl genomic clone (pCigen) used was described previously notes). (Cone et al. 1988) and contains ~8 kb of genomic DNA harbor ing the Cl promoter, introns, and coding sequence. The Bzl promoter-luciferase (pBziLuc) plasmid was described previ Enzyme assays ously (Klein et al. 1989). The Adhl promoter/chloramphenicol In each experiment, the cl R aleurones or cl Rscm embryos acetyltransferase plasmid (pAdhCAT) used in this study was were bombarded with a mixture of the appropriate plasmid described previously as pAIlCN (Callis et al. 1987). The frame- DNAs and incubated under illumination for 24—48 hr before shifts in Cl to generate p35SCifs26 and p35SCifs27 were con assaying for luciferase and CAT activity. Anthocyanin biosyn structed by filling in restriction sites located at the coding re thesis was determined by visual detection of pigmented cells gion for amino acids 32 (Ncol) and 258 {Xhol], respectively. The using 20- to 50-fold magnification under a Zeiss dissecting mi fusions of the GAL4 transcriptional activation domain to the Cl croscope. Photographs were also taken at the same magnifica cDNA to generate pClGALJ.l, pClGALS.l, and pClGALlO tion using Kodak tungsten film (ASA 160) developed by stan were at the restriction endonuclease sites at positions 117 dard procedures. Luciferase was assayed in tissue extracts as {Narl], 144 {Narl}, and 258 {Xhol) of the C2-coding sequence, described previously (Callis et al. 1987). Luciferase activity is respectively. pCJdel3 was deleted between the C2-coding re expressed as light units detected by an Analytical Lumines gion for amino acids 117 (Narl) to 144 (Narl). The expressed cence model 2001 Luminometer for 10 sec at 25°C. CAT activ GAL4 DNA-binding region contains the coding sequence for ity was determined by the conversion of ''^C-labeled acetyl-CoA amino acids 1-147 of GAL4 (to Clal] inserted into the expres to ethylacetate soluble counts per minute for 1 hr at 37°C and sion vector pMF6. The GAL4 DNA-binding region fused to the counted in EcoScint scintillation fluor as described previously acidic region of CJ contains coding sequence for amino acids (Sleigh 1988) (CAT activity is expressed as cpm). 1-147 of GAL4 fused to coding sequence for amino acids 173- 273 of Cl [Clal to Narl ligation). The expressed B-Peru cDNA was described previously (Goff et al. 1990). The GAL4 DNA- binding domain fused to the herpes simplex virus VP16 tran Acknowledgments scriptional activation domain [pGALVPlS] contains the coding We thank Bradley A. Roth and Theodore M. Klein for assistance sequence for amino acids 1-147 of GAL4 fused to a coding se in this work and for plasmid constructs. We are grateful to Vicki quence for 80 amino acids of VP16. The VP16 transcriptional L. Chandler, Jun Ma, Mark Ptashne, David Ow, and Steven J. activation domain was kindly provided by S. Triezenberg Triezenberg for various plasmid constructs. We thank Jutta A. (Triezenberg et al. 1988). The expressed Cl-I cDNA (p35SC2-/) Tuerck, Sheila McCormick, David Twell, and Brenda Lowe for contains a 1.0-kb cDNA with the entire coding region inserted critical reading of the manuscript. We also thank E.I. DuPont into the pMF6 expression vector. Fusions between the Cl and Medical Products Division for the use of the microprojectile Cl-I cDNAs were used to generate pClCI-l {Cl 1-220, EcoRl- particle delivery apparatus. This work was supported by a U.S.

GENES & DEVELOPMENT 307 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Gofi et al.

Department of Agriculture C.R.G.O. grant 89-37280-4898 to endosperm development in maize. Genetics 91: 309-315. M.E.F. Dynan, W.S. 1989. Modularity in promoters and enhancers. Cell The publication costs of this article were defrayed in part by 58:1-4. payment of page charges. This article must therefore be hereby East, E.M. and H.K. Hayes. 1911. Inheritence in maize. Conn. marked "advertisement" in accordance with 18 USC section Agric. Exp. Sin. Bull. 167. 1734 solely to indicate this fact. Emerson, R.A., G.W. Beadle, and A.C. Eraser. 1935. A summary of linkage studies in maize. Cornell Univ. Agric. Exp. Stn. Memo 180. Gerats, A.G.M., J. Bussard, E.H. Coe, and R. Larson. 1984. In References fluence of B and PI on UDPG : flavonoid-3-O-glucosyltrans- Aggarwal, A.K, D.W. Rodgers, M. Drotter, M. Ptashne, and S.C. ferase in Zea mays. Biochem. Genet. 22: 1161-1169. Harrison. 1988. Recognition of DNA operator by the repres Gerondakis, S. and J.M. Bishop. 1986. Structure of the protein sor of phage 434: A view at high resolution. Science encoded by the chicken proto-oncogene c-myb. Mol. Cell. 242: 899-907. Biol. 6: 3677-3684. Berger, S.L., W.D. Cress, A. Cress, S.J. Triezenberg, and L. Giniger, E. and M. Ptashne. 1987. Transcription in yeast acti Guarente. 1990. Selective inhibition of activated but not vated by a putative amphipathic alpha helix linked to a DNA basal transcription by the acidic activation domain of VP16: binding unit. Nature 330: 670-672. Evidence for transcriptional adaptors. Cell 61: 1199-1208. Giniger, E., S.M. Varnum, and M. Ptashne. 1985. Specific DNA Bergmann, D.G., L.M. Souza, and M.A. Baluda. 1981. Vertebrate binding of GAL4, a positive regulatory protein of yeast. Cell DNAs contain nucleotide sequences related to the trans 40: 767-774. forming gene of avian myeloblastosis virus. /. Virol. 40: 450- Goff, S.A., T.M. Klein, B.A. Roth, M.E. Fromm, K.C. Cone, J.P. 455. Radicella, and V.L. Chandler. 1990. Trans-activation of an Boyle, W.f., J.S. Lipsick, and M.A. Baluda. 1986. Antibodies to thocyanin biosynthetic genes following transfer of B regula the evolutionarily conserved amino-terminal region of the tory genes into maize tissues. EMBO f. 9: 2517-2522. v-myb-encoded protein detect the c-myb protein in widely Gonda, T.J., N.M. Gough, A.R. Dunn, and J. de Blaguiere. 1985. divergent species. Pwc. Natl. Acad. Sci. 83: 4685-4689. Nucleotide sequence of the cDNA clones of the murine myb Brendel, V. and S. Karlin. 1989. Association of charge clusters proto-oncogene. EMBO J. 4: 2003-2008. with functional domains of cellular transcription factors. Hanes, S.D. and R. Brent. 1989. DNA specificity of the bicoid PTOC. Natl. Acad. Sci. 86: 5698-5702. activator protein is determined by the homeodomain recog Bushman, F.D., C. Shang, and M. Ptashne. 1989. A single glu nition hchx residue 9. Cell 57: 1275-1283. tamic acid residue plays a key role in the transcriptional Herskowitz, I. 1987. Functional inactivation of genes by domi activation function of lambda repressor. Cell 58: 1163-1171. nant negative mutations. Nature 329: 1\9-T11. Callis, J., M.E. Fromm, and V. Walbot. 1987. Introns increase Hope, I.A. and K. Struhl. 1986. Functional dissection of a eugene expression in cultured maize cells. Genes &) Dev. karyotic transcriptional activator, GCN4 of yeast. Cell 1: 1183-1200. 46:885-894. Carey, M., Y.S. Lin, M.R. Green, and M. Ptashne. 1990. A mech Howe, K.M., C.F.L. Reakes, and R.J. Watson. 1990. Character anism for synergistic activation of a mammalian gene by ization of the sequence-specific interaction of mouse c-myb GAL4 derivatives. Nature 345: 361-364. protein with DNA. EMBO /. 9: 161-169. Chandler, V.L., J.P. Radicella, T.P. Robbins, J. Chen, and D. Ibanez, C.E. and J.S. Lipsick. 1990. Trans-activation of gene ex Turks. 1989. Two regulatory genes of the maize anthocyanin pression by w-myb. Mol. Cell. Biol. 10: 2285-2293. pathway are homologous: Isolation of B utilizing R genomic Jordan, S.R. and CO. Pabo. 1988. Structure of the lambda com sequences. Plant Cell 1: 1175-1183. plex at 2.5 A resolution: Details of the repressor-operator Chen, S.M. and E.H. Coe. 1977. Control of anthocyanin synthe interactions. Science 242: 893-899. sis by the C locus in maize. Biochem. Genet. 15: 333-346. Kakidani, H. and M. Ptashne. 1988. GAL4 activates gene ex Coe, E.H. Jr. 1962. Spontaneous mutation of the aleurone color pression in mammalian cells. Cell 52: 161-167. inhibitor in maize. Genetics 47: 779-783. Kamo, K.K. and T.K. Hodges. 1988. Establishment and charac . 1985. Phenotypes in corn: Control of pathways by alle terization of long-term embryogenic maize callus and cell les, time and place. UCLA Symp. Mol. Cell. Biol. 35: 509- suspension cultures. Plant Sci. 45: 111-117. 521. Kelleher, R.J., P.M. Flanagan, and R.D. Komberg. 1990. A novel Coe, E.H., Jr., M.G. Neuffer, and D.A. Hoisington. 1988. The mediator between activator proteins and the RNA polymer genetics of corn. In Corn and corn improvement. Agronomy ase II transcription apparatus. Cell 61: 1209-1215. Monograph no. 18, 3rd ed., pp. 81-258. Klein, T.M., T. Gradziel, M.E. Fromm, and J.C. Sanford. 1988. Cone, K.C., F.A. Burr, and B. Burr. 1986. Molecular analysis of Factors influencing gene delivery into Zea mays cells by the maize anthocyanin regulatory locus CI. Proc. Natl. high-velocity microprojectiles. BioTechnology 6: 559-563. Acad. Sci. 83: 9631-9635. Klein, T.M., B.A. Roth, and M.E. Fromm. 1989. Regulation of Davis, R.L., P.F. Cheng, A.B. Lassar, and H. Weintraub. 1990. anthocyanin biosynthetic genes introduced into intact The MyoD DNA binding domain contains a recognition maize tissues by microprojectiles. Proc. Natl. Acad. Sci. code for muscle-specific gene activation. Cell 60: 733-746. 86:6681-6685. Deuschle, U., R.A. Hipskind, and H. Bujard. 1990. RNA poly Klempnauer, K.-H., T.J. Gonda, and J.M. Bishop. 1982. Nucle merase II transcription blocked by Escherichia coli Lac re otide sequence of the retroviral leukemia gene v-myb and its pressor. Science 248: 480-^83. cellular progenitor c-myb: The architecture of a transduced Dooner, H.K. 1983. Coordinate genetic regulation of flavonoid oncogene. Cell 31: 453-463. biosynthetic enzymes in maize. Mol. Gen. Genet. 189: 136- Klempnauer, K.-H., C. Bonifer, and A.E. Sippel. 1986. Identifi 141. cation and characterization of the protein encoded by the Dooner, H.K. and O.E.Nelson. 1979. Interaction among C, R, human c-myb proto-oncogene. EMBO f. 5: 1903-1911. and Vp In the control of the Bz glucosyltransferase during Kunkel, T.A., J.D. Roberts, and R.A. Zakour. 1987. Rapid and

308 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Trans-activation of anthocyanin pigmentation

efficient site-specific mutagenesis without phenotypic selec activation by Spl: Evidence for coactivators. Cell 61: 1187- tion. Methods Enzymol. 154: 367-382. 1197. Lehming, N., J. Satorius, G.G. Niemiller, B.V. Wilcken-Berg- Rosson, D. and P. Reddy. 1986. Nucleotide sequence of chicken mann, and B. Miiller-Hill. 1987. The interaction of the rec c-myb complementary DNA and implications for myb on ognition helix of lac repressor with lac operator. EMBO ]. cogene function. Nature 319: 604—606. 6:3145-3153. Sadowski, I., J. Ma, S. Triezenberg, and M. Ptashne. 1988. GAL4- Lin, Y.S., M. Carey, M. Ptashne, and M.R. Green. 1990. How VP16 is an unusually potent transcriptional activator. Na different eukaryotic transcriptional activators can cooperate ture 335: 563-564. promiscuously. Nature 345: 359-361. Sakura, H., C. Kanei-Ishii, T. Nagase, H. Nakagoshi, T.J. Gonda, Liu, F. and M.R. Green. 1990. A specific member of the ATF and S. Ishii. 1989. Delineation of three functional domains of transcription factor family can mediate transcription activa the transcriptional activator encoded by the c-myb protoon- tion by the adenovirus Ela protein. Cell 61: 1217-1224. cogene. Proc. Natl. Acad. Sci. 86: 5758-5762. Lowe, K.S., D.B. Taylor, P.L. Ryan, and K.P. Paterson. 1985. Shen-Ong, G.L.C., H.C. Morse III, M. Potter, and J.F. Mushinski. Plant regeneration via organogenesis and embryogenesis in 1986. Two modes of c-myb activation in virus-induced the maize inbred line B73. Plant Sci. 41: 125-132. mouse myeloid tumors. Mol. Cell. Biol. 6: 380-392. Ludwig, S.R., L.F. Habera, S.L. Dellaporta, and S.R. Wessler. Sleigh, M.J. 1986. A nonchromatographic assay for expression of 1989. Lc, a member of the maize R gene family responsible chloramphenicol acetyltransferase gene in eucaryotic cells. for tissue-specific anthocyanin production, encodes a pro Anal. Biochem. 156: 251-256. tein similar to transcriptional activators and contains the Stringer, K.F., C.J. Ingles, and J. Greenblatt. 1990. Direct and myc-homology region. Pioc. Natl. Acad. Sci. 86: 7092-7096. selective binding of an acidic transcriptional activation do Ludwig, S.R., B. Bowen, L. Beach, and S.R. Wessler. 1990. A main to the TATA-box factor TFIID. Nature 345: 783-786. regulatory gene as a novel visible marker for maize transfor Struhl, K. 1989. Molecular mechanisms of transcriptional reg mation. Science 247: 449-450. ulation in yeast. Annu. Rev. Biochem. 58: 1051-1077. Ma, J. and M. Ptashne. 1987. A new class of yeast transcriptional Styles, E.D., O. Ceska, and K.T. Seah. 1973. Developmental activators. Cell 51: 113-119. differences in action of R and B alleles in maize. Can. f. Ma, J., E. Przibilla, J. Hu, L. Bogorad, and M. Ptashne. 1988. Genet. Cytol. 15: 59-72. Yeast activators stimulate plant gene expression. Nature Tice-Baldwin, K., G.R. Fink, and K.T. Arndt. 1989. BASl has a 334:631-633. myb motif and activates HIS4 transcription only in combi Ness, S.A., A. Marknell, and T. Graf. 1989. The v-myb oncogene nation with BAS2. Science 246: 931-935. product binds to and activates the promyelocyte-spccilic Triezenberg, S.J., R.C. Kingsbury, and S.L. McKnight. 1988. mim-l gene. Cell 59: 1115-1125. Functional dissection of VP16, the trans-activator of herpes Nishina, Y., H. Nakagoshi, F. Imamoto, T.J. Gonda, and S. Ishii. simplex virus immediate early gene expression. Genes e) 1989. Trans-activation by the c-myb proto-oncogene. MJ- Dev. 2: 718-729. cleic Acids Res. 17: 107-117. Oehler, T., H. Arnold, H. Biedenkapp, and K.-H. Klempnauer. 1990. Characterization of the v-myb DNA binding domain. Nucleic Acids Res. 18: 1703-1710. Ohlendorf, D.H., W.F. Anderson, Y. Takeda, R.G. Fischer, Y. Takeda, and B.W. Matthews. 1982. The molecular basis of DNA-protein recognition inferred from the structure of cro repressor. Nature 298: 718-723. Otting, G., Y.-Q. Qian, M. Muller, M. Affolter, W. Gehring, and K. Wuthrich. 1988. Secondary structure determination for the Antennapedia homeodomain by nuclear magnetic reso nance and evidence for a helix-turn-helix motif. EMBO f. 7: 4305^309. Paz-Ares, J., U. Wienand, P.A. Peterson, and H. Saedler. 1986. Molecular cloning of the c locus of Zea mays: A locus reg ulating the anthocyanin pathway. EMBO ]. 5: 829-833. Paz-Ares,}., D. Ghosal, U. Wienand, P.A. Peterson, and H. Sae dler. 1987. The regulatory cl locus of Zea mays encodes a protein with homology to myb proto-oncogene products and with structural similarities to transcriptional activators. EMBO J. 6: 3553-3558. Paz-Ares, J., D. Ghosal, and H. Saedler. 1990. Molecular analysis of the Cl-I allele from Zea mays: A dominant mutant of the regulatory Cl locus. EMBO J. 9: 315-321. Perrot, G.H. and K.C. Cone. 1989. Nucleotide sequence of the maize R-S gene. Nucleic Acids Res. 17: 8003. Peter, C.W.B., A.E. Sippel, M. Vingron, and K.-H. Klempnauer. 1987. Drosophila and vertebrate myb proteins share two conserved regions, one of which functions as a DNA-binding domain. EMBO f. 6: 3085-3090. Ptashne, M. 1988. How eukaryotic transcriptional activators work. Nature 335: 683-689. Pugh, B.F. and R. Tjian. 1990. Mechanism of transcriptional

GENES & DEVELOPMENT 309 Downloaded from genesdev.cshlp.org on September 30, 2021 - Published by Cold Spring Harbor Laboratory Press

Identification of functional domains in the maize transcriptional activator C1: comparison of wild-type and dominant inhibitor proteins.

S A Goff, K C Cone and M E Fromm

Genes Dev. 1991, 5: Access the most recent version at doi:10.1101/gad.5.2.298

References This article cites 70 articles, 22 of which can be accessed free at: http://genesdev.cshlp.org/content/5/2/298.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.