The Hopi gene is a member of the maize r1 gene family. By genetic and molecular analyses we report that Hopi consists of a single gene residing on chromosome 10 ~4.5 cM distal to r1. Hopi conditions anthocyanin deposition in aleurone, scutellum, pericarp, root, mesocotyl, leaves, and anthers, thus representing one of the broadest specifications of pigmentation pattern reported to date of all the r1 genes. A unique feature of the Hopi gene is that seeds are completely devoid of pigment at maturity but show a photoinducible germination-dependent anthocyanin accumulation in aleurone and scutellum. Our analysis has shown that the Hopi transcript is not present in scutellum of developing seeds but is induced only upon germination and that the simultaneous presence of both C1 and Hopi mRNAs is necessary to achieve A1 activation in scutella. We conclude that the expression pattern of the Hopi gene accounts for the germination-dependent anthocyanin synthesis in scutella, whereas the developmental competence of germinating seeds to induce anthocyanin production in scutella results from the combination of the light-inducible expression of C1 and the developmentally regulated expression of the Hopi gene.
ANTHOCYANINS represent the most widespread red and purple pigments in the plant kingdom. These pigments are produced in a variety of plant tissues, where they serve many diverse functions, such as attraction of pollinators and dispersal agents and protection against insects, phytopathogens, and UV irradiation. Anthocyanin accumulation in maize tissues requires the expression of many genes (at least 20), some that are involved in biosynthesis and others in tissue-specific regulation of biosynthetic loci (reviewed in Molet al. 1998). The structural genes c2, chi1, f3h, a1, a2, bz1, and bz2 encoding the biosynthetic enzymes are coordinately controlled at the transcriptional level by the products of at least two groups of regulatory genes, which are responsible for the developmental and tissue-specific pigmentation of plant and seed tissues. The first group is represented by the c1/pl1 family, whose members encode proteins with sequence homology to the tryptophan cluster, DNA-binding domain of MYB oncoproteins (Klempnaueret al. 1982; Cone et al. 1986, 1993a; Paz-Ares et al. 1986, 1987). Despite the c1 transcript being found at low levels in husks (Cooper and Cone 1997), the c1 gene acts to induce pigmentation only in seed tissues, such as the aleurone and the embryo (McCartyet al. 1989), whereas pl1 controls pigmentation in the seedlings, the plant body, and the pericarp, a maternally derived seed integument. The Pl-blotched allele, however, leads to anthocyanin expression in the aleurone also (Cocciolone and Cone 1993). Dominant light-independent and recessive light-dependent “sun red” alleles of pl1 are known (Coneet al. 1993b). The c1 and pl loci probably originate from a duplication event, since they share nearly identical coding regions and are likely to be functionally equivalent (Coneet al. 1993a).
The second group of regulatory genes, the r1/b1 gene family, consists of highly homologous genes encoding exchangeable proteins with an amino acid sequence containing the basic helix-loop-helix (bHLH) DNA-binding/dimerization domain found in MyoD protein (Daviset al. 1987; Chandleret al. 1989; Ludwiget al. 1989; Tonelliet al. 1991; Consonni et al. 1992, 1993). Although the r1 and b1 genes encode similar proteins, their structural organization differs. While b1 alleles consist of a single gene, many r1 alleles are genetically complex, consisting of many genes, each of which encodes a distinct pigmentation pattern. All members of the r1/b1 gene family are thought to have been derived from an as-yet-unidentified single ancestral gene that, following intrachromosomal duplication, gave rise to the r1 complex on chromosome 10 (Egglestonet al. 1995; Walkeret al. 1995), the Sn and Lc genes, lying about two units distal to r1 (Ludwiget al. 1989; Tonelliet al. 1991), and, after a postulated genome duplication event, to the b1 locus on chromosome 2 (Chandleret al. 1989). Molecular divergence during evolution may then have contributed to the establishment of distinct genes, each one performing a similar function in specific plant tissues. The duplicated origin of r1-like genes is inferred from the structural organization of the R-r complex. Genetic and molecular studies have revealed that R-r is composed of multiple units with sequence similarity, arranged in a tandem array: a plant (P) component that consists of a single gene specifying plant pigmentation; two seed genes (S1 and S2) arranged in opposite orientation, responsible for color in the seed; and a third cryptic component, referred to as q, consisting of a truncated gene sequence (Robbinset al. 1991; Walkeret al. 1995). The plant and seed pigmentation components can be lost independently by mutation or unequal crossing over, leading to different derivatives (Stadler 1948). While the expression pattern of complex r1 genes, such as R-r, is determined by the presence of multiple promoters, each one associated with a coding region, other r1 alleles and the b1, Sn, and Lc genes consist of a single genetic unit with a characteristic tissue-specific expression pattern probably depending on different controlling sequences located in the corresponding promoter regions, as demonstrated for two b1 alleles (Radicellaet al. 1992; Selingeret al. 1998).
Biochemical (Dooner 1983) and molecular analyses have revealed that the activity of a functional allele from both these two gene families is necessary for the transcriptional activation of the structural genes c2, chi1, f3h, a1, a2, bz1, and bz2 implicated in the anthocyanin biosynthetic pathway (Taylor and Briggs 1990; Bodeau and Walbot 1992; Grotewold and Peterson 1994; Grotewoldet al. 1998; Lesnick and Chandler 1998). Transient expression experiments have suggested that the B1 and C1 proteins interact to form a complex able to transactivate the bz1 promoter in maize embryogenic calli (Goff et al. 1991, 1992). Since the B1 protein does not contain either a transactivating domain or a DNA-binding activity but is required for transactivation of the bz1 promoter, Tuerck and Fromm (1994) proposed that the B1 protein might be required for an efficient binding of C1 to the MYB-like sites in the bz1 promoter, but not for the transactivating activity of the B1/C1 complex itself. However, it has been recently demonstrated that C1 does not require B or R to bind anthocyanin promoter sequences (Sainzet al. 1997; Lesnick and Chandler 1998). Furthermore, the observation that the a1 and a2 promoters contain sequences crucial for activation, which are not bound by C1, suggests that other factor(s) may be involved in activation through these promoter sequences (Sainzet al. 1997; Lesnick and Chandler 1998).
Hence, the pattern of pigmentation of a maize plant reflects its allelic constitution at r1/b1 and c1/pl1 regulatory loci. The activation of anthocyanin synthesis requires either c1 (in the seed) or pl1 (in the plant), while the r1/b1 genes, whose expression is tissue-specific, determine the tissue distribution of pigments (Consonniet al. 1993). Growth factors are also important in regulating pigment production. Anthocyanin accumulates in maize seed tissues starting at ~22 days after pollination (DAP) (Neillet al. 1987). These events are controlled by factors induced by abscisic acid (ABA) and by the product of the vp1 (viviparous1) gene that are required for the transcriptional activation of the c1 gene (Hattoriet al. 1992) and also of the maturation-specific wheat Em gene (McCarty 1995). vp1 mutants fail to repress precocious germination of developing seeds and are devoid of pigment in aleurone and embryo tissues. However, if rescued before desiccation, vp1 mutant seed is able to develop pigment in seed tissues during germination in presence of light (McCarty and Carson 1991), indicating that C1 expression also is regulated by light during kernel development. Supporting this idea, seeds homozygous for the c1-p allele are colorless but accumulate pigment if exposed to light during germination (Chen and Coe 1977), suggesting that in this allele of c1 the ABA/VP1 regulation has been lost but that the light regulation of the gene has been retained. Furthermore, C1 maize ears developed in total darkness produce colorless seeds that develop pigmentation after subsequent exposure to light (Dooner and Ralston 1994). It is likely that the transcription factors controlling tissue-specific accumulation of anthocyanin also are part of the light signal transduction pathway inducing its accumulation. Accumulation of A1 and C2 transcripts in seedling tissues, following light exposure, requires the activity of Sn (Tonelliet al. 1991). In addition “sun red” alleles of pl1 show a light-regulated expression pattern (Coneet al. 1993b; Procissiet al. 1997).
In this article, we report the genetic and molecular analysis of a new member of the r1/b1 gene family, called Hopi. This gene conditions the pigmentation of a wide variety of plant and seed tissues (Table 1). We show that the tissue-specific pigmentation conditioned by Hopi depends on a single gene residing on chromosome 10 ~4.5 cM distal to r1. A clear feature of the Hopi gene is its germination-dependent ability to determine anthocyanin accumulation in seed tissues, similar to the activity of the c1-p allele. In contrast to other r1 genes, such as R-sc, which promote pigment production in aleurone and scutellum during maturation on the ear (Table 1), homozygous Hopi seeds are completely devoid of pigment at maturity, but accumulate anthocyanins in scutellum and aleurone if germinated in the light. Light responsiveness is most effective in the first hours following the onset of germination. After a prolonged period of dark growth, light irradiation does not elicit pigmentation. It has been previously shown that in pericarp, both sn1 and pl1 expression is light-modulated, whereas in aleurone R-sc is constitutively expressed and C1 shows light inducibility; in both tissues, the MYB-like genes were found to be the limiting factors regulating the extent of pigment deposition (Procissiet al. 1997). Here, we demonstrate that Hopi gene expression in scutellum is not enhanced by light and is limited to the germination phase, whereas the accumulation of C1 transcript is under both developmental and light control. We also present evidence on the role played by Hopi and C1 in establishing the competence to induce anthocyanin synthesis in scutella during germination through control of the mRNA levels of structural and regulatory genes involved in pigment accumulation.
MATERIALS AND METHODS
Genetic stocks: All seed stocks used in this study were in the W22 background and were homozygous dominant for the a1, a2, c1, c2, bz1, and bz2 genes and homozygous recessive for the pl1 and b1 genes but differed in their r1 constitution.
r1 allele stocks: Detailed descriptions of the origin, phenotype, and structural characteristics of the r1 alleles used in this study can be found in Dooner and Kermicle (1974, 1976). Properties of the r1 alleles that are pertinent to this presentation are outlined below: R-st determines spotted aleurone and green plant. This is a germinally and somatically unstable allele of r1 composed of four r1 genes in direct orientation: one proximal r1 gene (Sc) producing a solid purple pigmentation of aleurone and three distal r1 genes (Nc) that determine near-colorless aleurone. The presence of a transposable element (I-R, Inhibitor of R) inserted in the Sc coding region produces the spotted seed phenotype of R-st (Egglestonet al. 1995).
r-Δ902: This symbol indicates an interstitial deletion involving a region of the long arm of chromosome 10 containing the r1 locus (kindly provided by J. Kermicle). Plant and seed tissues homozygous for the deficiency are totally devoid of pigment (Alleman and Kermicle 1993).
Hopi: A factor lying on chromosome 10, isolated from a maize stock given to G. Gavazzi by Dr. A. Brink and incorporated by backcrossing into the background of inbred W22; its origin is presumed to trace back to the Indian Hopi population. Hopi confers pigmentation to a wide variety of plant tissues (see Table 1), including pericarp, root, mesocotyl, and leaf blade in the seedling, midrib, ligule, leaf blade, and anthers in the mature plant. In addition, Hopi determines anthocyanin deposition in the scutellar and aleurone tissues of seeds, following germination in the presence of light. In the presence of the unlinked genetic factor Pl, it confers strong red (cherry) pigmentation in the pericarp of the seed and, for this reason, in the past it was termed r-ch:Hopi.
Germination and anthocyanin extraction: Seeds were incubated in sterile distilled water for 19 hr in rotating flasks kept in darkness at 25°, then plated in Plexiglas boxes on wet filter paper and germinated for increasing time periods (from 0 to 7 days) in the dark. Seedlings were then exposed to continuous white light for 24 hr at 21° and subsequently transferred to darkness for an additional 48 hr at the same temperature conditions. Anthocyanins from individual excised scutella were extracted with a fixed volume of 1% HCl in ethanol. The extracts were centrifuged twice and their absorption determined spectrophotometrically at 530 nm. Anthocyanin concentration is expressed as absorbance value at 530 nm per scutellum. Mean values represent seven independent replicates. Standard errors of the means were below 5%.
DNA and RNA analysis: Genomic DNA isolation and Southern analysis were performed as previously described (Tonelliet al. 1991). To perform a time course analysis of RNA accumulation, seeds were germinated in the dark for different time intervals (from 0 to 7 days) and then exposed to white light for 24 hr. Scutella were collected before and after light exposure. Cool white (F36T12/CW/HO) fluorescent lamps from GTE Sylvania (Lighting Products Group, Danvers, MA) were used. For the analysis of RNA during seed maturation, immature ears were harvested at 28 DAP and, after removal of husks, scutella were excised from immature kernels. Scutella were also excised from dry seeds previously water imbibed for 1–2 hr. Total RNA was extracted from scutella by grinding in liquid nitrogen and isolated as previously described (van Tunenet al. 1988). For Northern blot analysis, 30 μg total RNA was loaded on formaldehyde gels (Maniatiset al. 1982) and, after electrophoresis, blotted onto Biodyne nylon membrane (Pall). Hybridization and filter washing were performed as previously described (Tonelliet al. 1991). RNA molecular weight markers were from Bethesda Research Laboratory. Probes used for Southern and Northern hybridization were: (i) pSn1.4, 1.4-kb PstI-EcoRI fragment derived from the 3′ region of Sn cDNA (Tonelliet al. 1991); (ii) pSn0.4, a 0.4-kb EcoRI-PstI fragment derived from the 5′ portion of Sn cDNA (Tonelliet al. 1991); (iii) A1, a 700-bp BamHI fragment of A1 gene (Schwarz-Sommeret al. 1987); (iv) tub, a 1-kb EcoRI fragment from an α-tubulin cDNA from maize (Dolfiniet al. 1993); (v) restriction fragment length polymorphism (RFLP) probe, a 2.1-kb PstI fragment from bnl 7.49a, which maps 12.2 cM distal to r1 (kindly provided by Maize Genetics Cooperation-Stock Center).
Genomic cloning and sequence analysis: For library construction, genomic DNA from homozygous Hopi plants was digested with HindIII and cloned into λ NM1149 arms, as previously described (Tonelliet al. 1991). The library was screened with different Sn probes: pSn550 (a PCR fragment derived from positions −470 to +59 in the Sn promoter), pSn0.4, and pSn1.4 (Tonelliet al. 1991). Positive recombinant phages containing genomic fragments from Hopi (9 kb) were further purified. The 9-kb insert was subcloned into the plasmid vector pBS [Stratagene (La Jolla, CA) cloning system] and named pHopi9. Double-stranded DNA sequences were determined by the dideoxynucleotide chain termination method (Sangeret al. 1977) using Sequenase (U.S. Biochemicals, Cleveland) and employing oligonucleotides as primers. The Sn promoter sequence was obtained by sequencing the 3.8-kb HindIII genomic clone containing the promoter and the transcription start of Sn (Tonelliet al. 1991; Consonniet al. 1993). The alignment was performed using the ClustalV package (Higginset al. 1992). The sequences of Hopi and Sn promoters are available in GenBank (Hopi: no. AJ251720; Sn: no. X67619). Hopi cDNA sequence was determined by sequencing RT-PCR products (see RT-PCR analysis for experimental procedure) obtained with the following sets of primers: set 1, L1 (upstream primer 5′-CGCGGAGGAGAGCTCC TCCGGTT-3′, position +21) and CT19 (downstream primer 5′-TGGCGGCTGCAGCAAGCTGGCTC-3′, position +478); set 2, HO2 (upstream primer 5′-CAGCAGGCGCGTGATGGCGCTT-3′, position +369) and OB2 (downstream primer 5′-GGTGCTCGGCTGACCAAGT-3′, position +999); set 3, CT5 (upstream primer 5′-GTCAATCCTCTGCATCCCG-3′, position +906) and HO6 (downstream primer 5′-AGCTCCTTGAGGTAGGCT-3′, position +1781); set 4, HO3 (upstream primer 5′-ATCGAGGAGTTCTACAGC-3′, position +1255) and CT10 (downstream primer 5′-CATCTTTGCTTC GATCCC-3′, position +2296); set 5, oR31 and oR32 (position +2129 and +2531, respectively; see RT-PCR analysis for primer sequence). The positions of primers are given relative to the Sn cDNA (Consonniet al. 1992). The sequence of Hopi cDNA was confirmed by sequencing the exons of the Hopi genomic clone. The sequence of Hopi cDNA is available in GenBank (Hopi: no. AJ251719).
Transient transformation assay: The p35SC1 plasmid contains the 2.1-kb EcoRI C1 cDNA (Paz-Areset al. 1987) cloned into the EcoRI site of the plant expression vector pRT101 carrying the CaMV35S promoter (Topferet al. 1987), and the genomic clone pHopi9 was purified by using a Midi-Prep Kit (QIAGEN, Chatsworth, CA). Homozygous r-Δ902 seeds were surface-sterilized for 10 min in 5% sodium hypochlorite and kept in sterile water at 25° for 16 hr. After removal of the pericarp, seeds were allowed to germinate on MS medium containing 1% sucrose and 7% Bacto Agar (Difco Laboratories, Detroit) at 25° for 3 days prior to bombardment. Seeds were maintained in the dark during germination and after particle bombardment with the Biolistic PDS-1000/He particle gun (Bio-Rad, Hercules, CA). For each preparation of six shots, 3 μg of p35SC1 and 3 μg of pHopi9, or 6 μg of p35SC1 alone, or 6 μg of pHopi9 alone in a final volume of 5 μl were added to 50 μl of 60 mg/ml 1-μm gold microparticles (Bio-Rad) in 50% glycerol, 50 μl of 2.5 m CaCl2, and 20 μl of 0.1 m of spermidine-free base. After vortexing for about 3 min, the particles were centrifuged and the supernatant was removed. Microparticles were washed twice with 70% ethanol and then resuspended in 50 μl of 100% ethanol, followed by the spotting of 6 μl of coated particles onto Macrocarrier disks (Bio-Rad). DNA-coated gold microparticles were accelerated by the shock wave generated by the bursting of a rupture disk at ~900 psi of He gas. Bombarded seeds were kept in darkness for 72 hr at 22° and then subjected to visual inspection using a dissection microscope.
RT-PCR analysis: Reverse transcriptase polymerase chain reaction (RT-PCR) was used to detect Hopi, A1, and C1 gene transcripts. Assays for transcripts were also performed. First-strand cDNA was synthesized with an oligo(dT) primer from total RNA extracted from scutella of germinating seeds (see RNA isolation). The primer used was a 35-base oligonucleotide with 17dT residues and an adapter (5′-GGGAATTCGTCGA CAAGC-3′) sequence (Frohman 1990). All RNA samples were treated with DNase (1 unit/μg) before cDNA synthesis, and control reactions in which reverse transcriptase was omitted revealed no residual DNA contamination. To quantify cDNA yield in each reaction, an aliquot was labeled with a [32P]dCTP and incorporated radioactivity was determined with a counter (model LS 6000 Beckman Instruments, Palo Alto, CA). The different samples of cDNA were then diluted to obtain a uniform concentration. First-strand cDNA was used as template for PCR amplification. Amplification reactions containing an aliquot of cDNA synthesized from 5 μg of total RNA, 1× Promega polymerase buffer, 2.5 mm MgCl2, 200 μm each dATP, dCTP, dGTP, and dTTP, 0.1 μm each primer, and 1 unit of Taq DNA polymerase (Promega, Madison, WI) were performed in a final volume of 50 μl. Samples were overlaid with 100 μl of mineral oil. After the first denaturation step (5 min at 94°), the reaction mix underwent 20 cycles of denaturation at 94° for 1 min and 30 sec, annealing at 62° for 1 min and extension at 72° for 1 min. A final extension at 72° for 15 min was performed to complete the reaction. A set of primers specific for the orange pericarp-1 (orp-1) gene, which encodes the β subunit of tryptophan synthase (Wrightet al. 1992), was used to further standardize the concentration of the different samples. orp-1 was chosen as reference because the abundance of its mRNA is comparable to that of R-sc and A1 (Procissiet al. 1997). orp-1 specific sequences were amplified using the following primers: upstream primer, 5′-AAGGACGT GCACACCGC-3′, and downstream primer, 5′-CAGATACAGAACAACAACTC-3′. The length of the amplified product was 207 bp. To ensure that amplification reactions were within linear ranges, the reactions were carried out for 20 cycles. The PCR products were fractionated on 1.2% agarose gels and transferred onto Biodyne nylon membranes (Pall) and hybridized with random primed 32P-labeled fragments of orp-1. Hybridization signals were quantified by scanning the autoradiographic films with a Umax densitometer and cDNA samples were standardized accordingly. For mRNA detection of the genes under analysis, the following specific primer sets were used: for Hopi, oR31 (upstream primer 5′-ATGGCTTCATGG GGCTTAGATAC-3′) and oR32 (downstream primer 5′-GAAT GCAACCAAACACCTTATGCC-3′); for A1, A1 (upstream primer 5′-TTCTCGTCCAAGAAGCTCCAGGA-3′) and A2 (downstream primer 5′-CAATTCGTTGAACATGGAAGT AAG-3′); for C1, PL6 (upstream primer 5′-TCGGACGACTGC AGCTCGGC-3′) and Ac1 (downstream primer 5′-CACCGTGC CTAATTTCCTGTCCGA-3′). The size of the amplified products is 403 bp for oR31/oR32, 285 bp for A1/A2, and 313 bp for PL6/Ac1. For each set of primers, pilot experiments were performed to determine the melting temperature and the number of cycles needed to ensure that amplification reactions were within the linear range (Procissiet al. 1997). The PCR products were then separated on agarose gels (1.2%), transferred onto Biodyne nylon membranes (Pall), and hybridized with random primed 32P-labeled fragments. The Hopi PCR products were hybridized with a 1.4-kb PstI-EcoRI fragment of Sn cDNA (Tonelliet al. 1991); the C1 products with a 1.4-kb XhoI-EcoRI of the C1 cDNA (Paz-Areset al. 1987), and the A1 products with a 700-bp BamHI fragment of the A1 gene (Schwarz-Sommeret al. 1987).
Hopi consists of a single gene nonallelic to the r1 gene complex: The Hopi gene controls anthocyanin accumulation in an extensive number of plant and seed tissues (see Table 1), including pericarp, aleurone and scutellum in the kernel, root, mesocotyl, and leaf blade in the seedling, midrib, ligule, leaf blade, and anthers in the mature plant. The pigmentation of these tissues could be due to the activity of a complex locus whose different genetic elements could account for the tissue-specific pigmentation, as in the case of the R-r locus. However, our efforts to separate genetically the plant and seed phenotypic effects associated with Hopi have so far failed. Nonetheless, we could not discount the possibility that Hopi has a compound structure, since recombination in the region distal to r1 is reduced in Hopi-bearing chromosomes due to the presence of a heterochromatic knob (K10) (Racchi and Gavazzi 1988). To define the molecular structure of Hopi, we produced R-st Hopi/r-Δ902 plants, where R-st Hopi represents a crossover derivative devoid of K10, selected from the progeny of R-st/Hopi plants, and r-Δ902 is a viable deletion comprising the entire r1 locus (Alleman and Kermicle 1993). R-st Hopi/r-Δ902 plants were then used as females in testcrosses to male plants homozygous for r-Δ902. Since a diagnostic phenotype of Hopi is the induction of full pigmentation in the scutellum and irregular aleurone pigmentation following germination in light, seeds derived from the above cross (stippled and colorless in a 1 to 1 ratio) were germinated in darkness for 48 hr, exposed to continuous white light for 24 hr, and then scored for aleurone and scutellum pigmentation (Figure 1A). Four phenotypic classes were identified: two parental classes, one exhibiting red scutellum and patches of color in the aleurone superimposed on a stippled background (Figure 1A, class 1, R-st Hopi/r-Δ902) and the other one completely devoid of pigment (class 2, r-Δ902/r-Δ902), and two recombinant classes showing, respectively, stippled aleurone (class 3, R-st/r-Δ902) and one red pigmentation in scutellum and patches of color on a colorless background in aleurone (class 4, r-Δ902 Hopi/r-Δ902). Seedlings were then grown and the mature plants scored to establish whether the tissue-specific pigmentation determined by the Hopi gene depends on discrete separable genetic elements (Figure 1A). Presence of leaf, anther, and root pigmentation was observed in class 1 (R-st Hopi/r-Δ902) and class 4 (r-Δ902 Hopi/r-Δ902). The progeny from these plants were germinated and the pattern of pigmentation was analyzed. Progeny of class 4 (r-Δ902 Hopi/r-Δ902) individuals always showed colorless aleurone in dry kernel and, after germination in light, pigmentation in scutellum and aleurone and all other tissues usually pigmented in Hopi lines in a 3 to 1 ratio. Seeds derived from R-st/r-Δ902 plants (class 3) showed stippled and colorless aleurone in a 3 to 1 ratio and no pigmentation in seedling tissues (data not shown). The results indicate that both classes (class 1 and class 4, Figure 1A) and their progeny show the phenotype expected on the assumption of Hopi being a single gene responsible for plant and seed phenotypic effects. These data, together with the results obtained from several identical crosses (Table 2), also indicate that R-st and Hopi are separate genes lying 4.5 cM apart.
The Hopi phenotype is strictly correlated to a 9-kb genomic fragment: To establish whether Hopi consists of a single gene at the molecular level, 104 plants resulting from the above cross were individually analyzed by Southern analysis (Figure 1B). Since the Hopi gene shares DNA similarity with both Sn and r1 genes (Tonelliet al. 1991), we used a 0.4-kb EcoRI-PstI fragment derived from the 5′ end of Sn cDNA as a probe (Tonelliet al. 1991). The results showed that the 5- and 3.5-kb HindIII fragments are associated with R-st (Figure 1B, lanes 1 and 9), whereas a 9-kb HindIII fragment cosegregates with the Hopi phenotype (lanes 1 and 3–7). In fact, recombinant progeny carrying Hopi in cis with r-Δ902 (11 individuals; Figure 1A, class 4) retain the 9-kb HindIII fragment (Figure 1B, lanes 3–7) and, in agreement, loss of this fragment in the R-st/r-Δ902 recombinant plants (lane 9) is associated with absence of pigmentation in aleurone and scutellum after germination in light and in all other tissues normally pigmented by Hopi (13 individuals; Figure 1A, class 3). We included in the analysis DNA extracted from the original homozygous Hopi line, which contained a 9-kb HindIII fragment comigrating with that of r-Δ902 Hopi/r-Δ902 recombinant DNA (lane 8). As expected, no r1-homologous genomic fragments could be linked to r-Δ902, while the 4-kb HindIII fragment, weakly detected in all samples, was attributable to the recessive b1 gene. To confirm that the Hopi gene resides on a single HindIII genomic fragment, we used a DNA probe of 1.4 kb derived from the 3′ end of Sn cDNA to hybridize the same Southern blot filter. Again, a 9-kb HindIII fragment cosegregated with the Hopi phenotype and was detectable only in the R-st Hopi/r-Δ902 parental class, in the r-Δ902 Hopi/r-Δ902 recombinant class, and in the homozygous Hopi DNA (data not shown). These results suggest that Hopi consists of a single gene residing on a single 9-kb HindIII fragment.
To determine whether the Hopi gene is distal to r1 as are the other displaced r1 genes, Sn and Lc, an RFLP analysis was performed on individuals of the cross reported in Figure 1 using the probe bnl 7.49a, which maps 12.2 cM distal to r1 (Burret al. 1988). This probe distinguishes between the distal region of the R-st Hopi chromosome and the r-Δ902 one (Figure 2, lanes 1 and 3). The R-st Hopi/r-Δ902 parental plants used in testcrosses to r-Δ902/r-Δ902 were heterozygous for the RFLP marker, carrying both a 4.5- and a 7-kb SacI fragment (lane 2). As expected, all nonrecombinant progeny tested have the characteristic RFLP bands of the parental plants (lanes 4 and 7). The r-Δ902 Hopi/r-Δ902 recombinants retain both the 4.5-kb fragment of the distal region of the R-st Hopi chromosome and the 7-kb fragment of the r-Δ902 chromosome (Figure 2, lane 5), while R-st/r-Δ902 plants only retain the RFLP band characteristic of the r-Δ902 chromosome (Figure 2, lane 6). Southern analysis using an RFLP probe specific for the proximal region indicates no exchange of proximal markers (data not shown). These results demonstrate that Hopi is distal to r1 and suggest that Hopi, Sn, and Lc may be alleles of the same displaced r1 homologous locus.
Genomic cloning and sequence analysis of Hopi: Several members of the r1/b1 gene family, such as b1, Lc, and Sn, have been cloned by cross-hybridization using r1 sequences (Chandleret al. 1989; Ludwiget al. 1989; Tonelliet al. 1991). Using a similar approach, a HindIII genomic library from homozygous Hopi DNA was screened with different Sn probes. Eighteen recombinant phages were isolated and, on the basis of their size and restriction map, they were divided into two groups, one containing 10 9-kb HindIII fragments representing the Hopi gene (Figure 1C) and the other containing 8 3.5-kb fragments representing the recessive b1 gene (not shown). As shown in Figure 1C, the 9-kb genomic fragments showed three regions that cross-hybridized with three PstI fragments from the full-length Sn cDNA clone, thus confirming that Hopi resides on a single HindIII genomic fragment (Tonelliet al. 1991).
Sequence analysis of Hopi from positions −1489 to +442 (numbering as for the Sn gene; Consonniet al. 1993) revealed 95% nucleotide identity between Hopi and Sn. Sequence diversity mainly consisted of single base substitutions, although some small insertions and deletions were also revealed throughout the sequence (Figure 3).
In the 5′ untranslated leader sequence of the Sn cDNA five ATG triplets preceding the actual start codon have been previously described (Consonniet al. 1993). The first and the fourth of these are the start sites for two small upstream open reading frames (uORFs) of 38 and 15 residues, respectively. Compared to the Sn leader region, a triplet insertion (position +72) in the Hopi sequence that results in an additional codon and a single deletion (position +148) that causes a frameshift, together result in a shorter first uORF of 34 amino acids (Figure 3). The complete cDNA sequence confirmed that the major open reading frame (ORF) begins at the fifth ATG start codon at nucleotide position +379. It encodes a putative protein of 611 residues, 98% identical to SN (Consonniet al. 1992). Compared to the Sn ORF, four deletions are present in the Hopi cDNA, three of which result in single amino acid deletions, D 255, A 379, and N 500, and one in a two amino acid deletion, G 417 and T 418 (Figure 4). Twenty-two nucleotide substitutions between Hopi and Sn were also found throughout the coding region; however, only 9 of them led to an amino acid substitution. Seven of these amino acid changes are conservative and among them four are exclusively found in the HOPI protein, while B-PERU or R-S share the others. Two other amino acid substitutions are not conservative; the first one (N 418 instead of K in SN) is present in all R1 proteins except SN, while the second one (G 564 instead of S) is present exclusively in HOPI.
To determine whether the Hopi gene cloned was capable of activating anthocyanin pigmentation in maize tissues, the 9-kb HindIII genomic fragment was tested in transient transformation assays by microprojectile bombardment of colorless r-Δ902 germinating seeds. Cotransformation of this Hopi genomic fragment, together with a C1-expressing plasmid, p35SC1, restored pigmentation in cells of scutellar node, mesocotyl, coleoptile, and root (Figure 5). No pigmented cells were observed when transformations were performed with the C1-expressing plasmid alone (data not shown). This result demonstrates that the 9-kb fragment cloned includes the entire Hopi gene and that HOPI protein can substitute for the r gene products to activate the anthocyanin biosynthetic pathway in different seedling tissues.
Effect of light and germination phase on anthocyanin accumulation in the presence of Hopi: In contrast to other r1 genes, e.g., R-sc, which promotes pigment production in seed tissues during maturation on the ear, homozygous Hopi seeds are completely devoid of pigment at maturity but upon germination show a photoinducible anthocyanin accumulation in the scutellum and aleurone tissues of the seed. To establish more precisely how light and germination interact to trigger anthocyanin production in scutellar tissues, water-imbibed Hopi seeds were allowed to germinate in darkness from 1 to 7 days and were then exposed to continuous white light for 24 hr. The pigment content was measured following an additional 48 hr of darkness to allow anthocyanin synthesis to be completed. The results in Figure 6 (bottom) show that a peak in pigment content is reached when seeds are exposed to light after 48 hr of germination in darkness (Figure 6B, bar 3). Seeds germinated in darkness for longer periods (bars 4–7) show a decline in the response to light, leading to almost no induction after 5 days of dark germination (Figure 6B, bar 6). Thus, competence to respond to light is maximal during the first four days of dark growth but is lost if darkness is prolonged. Aleurone differs from scutellum in its requirement for light, since faint patches of aleurone pigmentation are observed even in the absence of light irradiation after 3 days of germination. However, an enhancing effect of light irradiation on aleurone pigmentation has also been observed (data not shown).
Expression pattern of Hopi and A1 during germination in presence of light: To ascertain whether the competent phase for anthocyanin accumulation in response to light was correlated with the expression pattern of genes responsible for anthocyanin accumulation, we performed Northern blot analysis, in which the transcript levels of the regulatory Hopi gene and of the A1 structural gene were tested. Imbibed Hopi seeds were germinated in the dark for increasing time intervals and then exposed to white light for 24 hr. Total RNA was extracted from dark-grown and light-exposed scutella. mRNA from unpigmented scutella excised from Hopi developing seeds close to maturity (28 DAP) and dry kernels were also included in the analysis. Figure 6A shows steady-state levels of mRNA in Hopi scutella during dark germination, while Figure 6, B and C, shows the transcript levels in response to light and during seed maturation, respectively. Increasing times of dark germination were associated with a transient increase in the amount of Hopi mRNA that reached a peak after 3 days of dark germination (Figure 6A, lane 3), while during seed maturation and in the dry seed the Hopi transcript was absent (Figure 6C). Following light irradiation the expression pattern of Hopi was similar to that observed during the dark growth, showing an increase of transcript level up to 3 days of germination (Figure 6B, lanes 1–3) and a decrease in the subsequent stages. These results clearly indicate that the transient increase in the steady-state levels of the Hopi mRNA is closely correlated to the developmental stage of the germinating seed but not to the induction of anthocyanin biosynthesis by light. In contrast to Hopi, the transcript levels of the A1 structural gene were strictly light dependent. In fact, A1 mRNA could be detected only after light exposure of the germinating seeds (Figure 6B). The maximum level of induction was reached when 2 days of dark germination preceded light irradiation (Figure 6B, lane 3), thus following the same time course of transcript accumulation as the Hopi gene. Anthocyanin content was therefore precisely correlated to A1 gene expression and, in turn, A1 gene expression was correlated to changing levels of Hopi transcript in the scutella during germination (Figure 6B, lanes 1–5). However, expression of Hopi was not sufficient to determine the expression of A1 during germination in darkness (Figure 6A). The absence of the Hopi transcript in dry seeds and in immature kernels could account for the dependence of anthocyanin synthesis in scutella upon germination (Figure 6C).
Analysis of the C1 expression pattern in scutellum during germination: To establish whether the activity of a MYB counterpart might be limiting the light-dependent pigmentation of scutella, we analyzed the expression pattern of the seed-specific C1 gene in these tissues. Since MYB-related genes, including C1, exhibit extremely low transcript abundance (Procissiet al. 1997), we performed a gene-specific RT-PCR analysis using cDNAs derived from the same total RNA samples analyzed in the previous Northern experiments. The steadystate levels of A1 and Hopi mRNA were also reassayed by RT-PCR to confirm the Northern data presented in Figure 6. As previously shown, the level of A1 steadystate mRNA, completely absent in scutella excised from dry seed and dark-germinated scutella, is strongly induced by light, particularly after 2 days of dark growth (Figure 7B, lane 3). The time course of Hopi mRNA accumulation is quite similar to that observed by Northern analysis, showing a peak after 3 days in dark-germinated scutella (Figure 7A, lane 3) and a similar expression pattern after light irradiation (Figure 7B). The light inducibility of the A1 structural gene appears to result from the activation of C1. In fact, the C1 transcript, which is completely absent in scutella excised from darkgrown seeds and barely detectable in dry seed, was strongly induced by light irradiation (Figure 7). The maximum level of C1 induction was reached at the start of germination in light (Figure 7B, lane 1), whereas the light responsiveness of C1 was less effective when light irradiation was preceded by dark growth (Figure 7B, lanes 2–5). Furthermore, the induction of the C1 regulatory gene after 24 hr of white light is clearly not influenced by the developmental stage within the first 4 days of dark germination, since the C1 transcript level did not change until the subsequent stages of growth were reached. Nonetheless, the lack of C1 gene expression when dark growth is prolonged (Figure 7B, lanes 6–7) is reflected in the absence of anthocyanin pigmentation, suggesting that the developmental competence of scutella to respond to light is also determined by the developmental window of expression of C1. The RT-PCR analyses also indicate that the simultaneous presence of both C1 and Hopi mRNAs is necessary to achieve A1 activation in scutella.
The r1/b1 gene family consists of a number of genes that independently control the tissue-specific distribution of anthocyanin in plant and seed tissue. In this study we report the cloning and characterization of a novel member of the r1/b1 family.
The Hopi gene maps to chromosome 10L, the same chromosome arm of the r1 locus. It conditions anthocyanin deposition in aleurone, scutellum, mesocotyl, leaves, and anthers and interacts with a dominant Pl allele to produce a “cherry” phenotype in the pericarp, thus representing a very broad specification of pigmentation pattern among all the r1 genes so far analyzed. The wide variety of tissues pigmented and the presence of a strong red pigmentation in the pericarp of the seed in the presence of the Pl gene are features common to other r1 genes, termed “cherry” for this reason (Table 1). Genetic studies indicated that the R-cherry genes (R-ch) are composed of a compound structure consisting of four genetic elements, each conferring a tissue-specific pigmentation (Sastry 1970). The extremely low frequency of R-ch mutants lacking one or more tissue-specific genetic determinants was due to the presence of a heterochromatic knob (K10) in the long arm of chromosome 10, which suppresses both genetic recombination and the chiasma frequency (Rhoades 1942). A similar compound structure has also been postulated for the Hopi gene (Racchi and Gavazzi 1988). Our genetic analysis has been performed using a Hopi line devoid of K10 and the recombinational data obtained, together with the RFLP analysis, show that Hopi consists of a single genetic element, lying 4.5 cM distal to r1 on the long arm of chromosome 10. In addition, through genetic and molecular analysis we have demonstrated that the broad pigmentation pattern conditioned by Hopi does not depend on the presence of multiple closely associated genetic units but on a single gene located in a 9-kb genomic fragment.
Nucleotide sequence analysis showed that Hopi share a high degree of similarity with Sn, both in the transcribed region (98% identity) and in the promoter sequence up to −1489 (95% identity). Microprojectile delivery of the Hopi genomic clone to colorless germinating embryos results in pigmented cells, thus establishing that the Hopi gene, cloned in this study, contains a functional coding region able to complement r1 regulatory mutation. Furthermore, the purple pigmented cells observed in scutellar nodes, mesocotyls, and coleoptiles following the delivery of the Hopi gene suggest that the 9-kb clone contains cis-acting sequences responsible for the expression of the gene in seedling tissues.
We have shown that sequence diversity between Hopi and Sn mainly consists of single base substitutions distributed throughout the promoter sequence (Figure 3) and the coding region (Figure 4). Compared to the SN protein, five amino acid deletions and nine amino acid substitutions were found in the HOPI protein. Most amino acid substitutions are conservative and, if not, are shared by the other R1 proteins. Similarly, most deletions are shared by other R1 proteins, except deletion of amino acid N500, which is located in the basic domain outside the bHLH motif.
Sn and Hopi share some territories of expression, such as pericarp, mesocotyl, and particularly leaf, where they display exactly the same temporal and cell-specific expression pattern and condition strong coloring in the midrib and ligule. However, Hopi conditions anthocyanin pigmentation in scutellum and aleurone, which are tissues typically pigmented by r1 alleles. Additionally, Hopi significantly differs from these r1 alleles in the timing of its control of anthocyanin synthesis in seed tissues, since pigmentation is induced only during germination in the presence of light. It would not be surprising if the promoter sequence diversity revealed by our analysis is responsible for the different expression patterns conditioned by Hopi and Sn. However, it is also possible that Hopi is located near a regulatory element strongly influencing its expression, providing new tissue-specific expression in the absence of relevant sequence changes in its promoter.
Hopi seeds are completely devoid of pigments at the end of seed maturation. However, if seeds are germinated in the light, anthocyanins accumulate in scutellum and aleurone. We have observed that anthocyanin accumulation in scutella occurs only if germinating seeds are transferred to light between the first and the fourth day of germination (Figure 6B). It has been previously shown that the different developmental competences of the pericarp and aleurone layers of immature seeds to respond to light result from the expression pattern of bHLH and MYB regulatory genes at different stages of seed development (Procissiet al. 1997). In the aleurone, the R-sc transcript was detectable both at 14 and 30 DAP, but the steady-state level of mRNA was slightly higher at 30 DAP, when the aleurone layer is more competent for light-dependent pigment accumulation. In both developmental stages, R-sc expression is not enhanced by light irradiation, whereas C1 shows light inducibility. However, the extremely low abundance of C1 transcript at 14 DAP was found to be the limiting factor for conferring the developmental competence of the aleurone to light responsiveness (Procissiet al. 1997). The R-sc and C1 expression pattern in aleurone cells is similar to that observed for Hopi and C1 in scutella during germination in the presence of light. As for R-sc, Hopi gene expression in scutellum is influenced by the developmental stage, but not by light. Its transcript is detectable only during germination with a similar accumulation pattern both in darkness and in light. Though the actual mechanism of regulation of Hopi in germinating seeds remains to be determined, the absence of the Hopi messenger appears to be the factor limiting anthocyanin pigmentation in developing seeds.
In contrast to Hopi, the expression of C1 in scutellum is clearly light inducible but is also under developmental control. It has been previously shown that C1 is abundant in embryo tissues of 22 DAP immature kernels, when anthocyanins begin to appear in the aleurone (McCartyet al. 1989). The activation of the C1 gene is dependent on the vp1 regulatory gene and ABA, which are known to act through cis-acting elements present in the C1 promoter (Hattoriet al. 1992). The presence of C1 mRNA has also been detected at the end of seed maturation in 35 DAP kernels (Paz-Areset al. 1986). Our analysis has shown that the C1 transcript is present in scutellum of dry seed and that the C1 transcript is not maintained after the onset of germination (Figure 7A). Light inducibility in the aleurone has been previously shown for germinating seeds of C1 mutants, such as c1-p (Scheffleret al. 1994). Our results clearly indicate that in the dark the absence of C1 transcripts inhibits anthocyanin pigmentation in germinating seeds. We conclude that the expression pattern of the Hopi gene accounts for the germination-dependent anthocyanin synthesis in scutellum, whereas the developmental competence of germinating seeds to induce anthocyanin production in scutella results from the combination of the light-inducible expression of C1 and the developmental regulated expression of the Hopi regulatory gene.
The r1/b1 gene family consists of a number of genes whose products perform similar roles in the control of anthocyanin synthesis in different maize tissues. Comparison of the Lc, Sn, R-S, and B-Peru cDNAs has shown that the deduced protein sequences share >80% amino acid identity (Consonniet al. 1993). This functional and sequence similarity has led to the hypothesis of a common evolutionary origin of the r1/b1 gene family components from a single ancestral gene. A detailed molecular analysis of the r1 gene family orthologs in different plant species has confirmed that the r1 and b1 genes have arisen from a common ancestral gene through a polyploidization event, which presumably occurred at the time of divergence of Zea mays from Sorghum (Purugganan and Wessler 1994). The proliferation of R sequences at and near the r1 locus on the long arm of chromosome 10 is probably the result of intrachromosomal duplicative events of independent origin (Robbinset al. 1988). The view that the complex r1 alleles and the displaced r1 genes on chromosome 10 represent more recent duplication events is supported by the occurrence of the variable r1 constitution of different geographic alleles as well as by the presence of Lc or Sn confined to specific accessions (Coeet al. 1988). In this scenario Hopi might be considered a novel allele of the same displaced r1 homologous locus. In accordance with this view is the finding that Sn, Lc, and Hopi cDNA sequences are more closely related to r1 than r1 to b1 (Figure 4).
The functional interchangeability of the R1 proteins has led to the hypothesis that molecular divergence, following the duplication events, more significantly affected the gene promoters than coding regions, so that the diverse pigmentation pattern controlled by each r1 gene reflects differences in regulatory sequences rather than their gene products (Ludwiget al. 1990; Radicellaet al. 1992). A series of genetic tests has revealed that the wild progenitor of maize, teosinte, possesses functional alleles at both r1 and c1 regulatory loci, but predominantly alleles not expressed during kernel development (Hansonet al. 1996). A dominant functional allele of the r1 locus able to develop pigment during seed maturation was not found in teosinte populations, although the possibility that it is present, at very low frequency, cannot be excluded. Similarly, at the c1 locus, a dominant C1 allele was found in a single teosinte population, perhaps representing a more recent introgression from maize. It has been proposed that purple kernels evolved by changes in cis-regulatory elements involving both regulatory loci, followed by human selection so that seed pigmentation became predominant (Hansonet al. 1996). The origin of the Hopi gene can presumably be traced back to the Indian Hopi population. The Hopi planted their corn deep in the soil to avoid loss when rainfall from the mountains hit the high plain during cultivation (Weatherwax 1954). Germinating seeds had a long phase of early growth under the soil. It is likely that these conditions led to the selection of strains accumulating flavonoids in their seedling tissues, since these pigments provide important defensive agents against insects and other phytopathogens (Reidet al. 1992; Byrneet al. 1996).
We thank Cathie Martin and Chris Bowler for critically reviewing the manuscript, Jerry Kermicle for providing the r-Δ902 stock, and Domenico Allegra for computer graphics. This work was supported by Ministero delle Politiche Agricole e Forestali Piano Nazionale “Biotecnologie Vegetali” (Area 1, Progetto N. 2) to C.T.
Communicating editor: V. Sundaresan
- Received July 22, 1999.
- Accepted January 4, 2000.
- Copyright © 2000 by the Genetics Society of America