Cloning of the Pleiotropic T Locus in Soybean and Two Recessive Alleles That Differentially Affect Structure and Expression of the Encoded Flavonoid 3' Hydroxylase

Corresponding author: Lila Vodkin, 1201 W. Gregory, Urbana, IL 61801., l-vodkin{at}uiuc.edu (E-mail)

Communicating editor: J. A. BIRCHLER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Three loci (I, R, and T) control pigmentation of the seed coats in Glycine max and are genetically distinct from those controlling flower color. The T locus also controls color of the trichome hairs. We report the identification and isolation of a flavonoid 3' hydroxylase gene from G. max (GmF3'H) and the linkage of this gene to the T locus. This GmF3'H gene was highly expressed in early stages of seed coat development and was expressed at very low levels or not at all in other tissues. Evidence that the GmF3'H gene is linked to the T locus came from the occurrence of multiple RFLPs in lines with varying alleles of the T locus, as well as in a population of plants segregating at that locus. GmF3'H genomic and cDNA sequence analysis of color mutant lines with varying t alleles revealed a frameshift mutation in one of the alleles. In another line derived from a mutable genetic stock, the abundance of the mRNAs for GmF3'H was dramatically reduced. Isolation of the GmF3'H gene and its identification as the T locus will enable investigation of the pleiotropic effects of the T locus on cell wall integrity and its involvement in the regulation of the multiple branches of the flavonoid pathway in soybean.

SECONDARY metabolites derived from the flavonoid pathway such as proanthocyanidins and anthocyanins play a relevant role in plant pathogen defense and protection from UV light exposure in addition to their nutritional value due to their antioxidant properties. In soybean (Glycine max) three independent loci (I, R, and T) control pigmentation of the seed coats and are distinct from those controlling flower color (reviewed in BERNARD and WEISS 1973 ; PALMER and KILEN 1987 ; NICHOLAS et al. 1993 ). The I locus controls distribution of anthocyanin and proanthocyanidin pigments and corresponds to the multigenic region containing the chalcone synthase (CHS) gene family (CHS1, CHS3, and CHS4), which in its dominant form exhibits homology-dependent gene silencing leading to a colorless seed phenotype (TODD and VODKIN 1996 ). The recessive i allele allows coloration of the entire seed coat while two other alleles, iⁱ and i^k, restrict pigment distribution to specific regions of the seed coat.

The R and T genes determine the anthocyanin and proanthocyanidin products and specific seed coat color. Thus, the self-colored seed coats are black (i,R,T), imperfect black (i,R,t), brown (i,r,T), or buff (i,r,t; Fig 1). No definitive function has been demonstrated to date for the R locus. Because R genotypes contain proanthocyanidins and anthocyanins and r genotypes contain only proanthocyanidins, there is speculation that it may encode an enzyme that acts after the formation of leucoanthocyanidin but previous to the formation of anthocyanins (TODD and VODKIN 1993 ). An epistatic effect of t results in damaged seed coat structure. Imperfect black (i,R,t) and buff (i,r,t) seed coats manifest splits and cracks (Fig 1). WOODWORTH 1921 originally describe the genetics of the T locus on the color of trichome hairs (pubescence). Plants with the dominant T allele have tawny or brown pubescence on the leaves, stems, and pods while those with the homozygous, recessive t genotype have gray pubescence (Fig 2).

View larger version (63K): In this window In a new window Download PPT slide   Figure 1. Effects of the I, R, and T loci on patterns and timing of seed coat pigmentation in Clark isolines during seed maturation. The approximate fresh weights of the seeds are indicated. DES, desiccating seed; DRY, mature harvested seed. The brown (i,r,T) or buff (i,r,t) pigment does not form until very late in seed coat development as compared to black (i,R,T) or imperfect black (i,R,t). The reason for the defective cracking in the pigmented t isoline is unknown. The recessive i allele allows coloration of the entire seed coat, the ii allele restricts pigment synthesis to the hilum of the seed coat (not shown), and the ik allele distributes it in a saddle pattern.

View larger version (68K): In this window In a new window Download PPT slide   Figure 2. Effect of the T locus on pubescence color of soybean pods. The dominant T allele controls the synthesis of pigments that accumulate and give the tawny or brown color to the trichomes seen on the pods on the left. Plants with the homozygous recessive t alleles have gray pubescence as shown on the two pods to the right.

Both anthocyanin and proanthocyanidin pigments are found in black (i,R,T) and imperfect black (i,R,t) seed coats (BUZZELL et al. 1987 ; TODD and VODKIN 1993 ) while brown (i,r,T) and buff (i,r,t) seed coats synthesize only proanthocyanidins (TODD and VODKIN 1993 ). The anthocyanins of the mature, black seed coat (i,R,T) have been identified as delphinidin-3-monoglucoside and cyanidin-3-monoglucoside (BUZZELL et al. 1987 ) that have 3', 4', 5' and 3', 4' B-ring hydroxylation patterns, respectively (Fig 3). In mature, imperfect-black seed coats (i,R,t), delphinidin is the primary anthocyanin (BUZZELL et al. 1987 ). In addition, immature, black (i,R,T) and brown (i,r,T) seed coats contain significant amounts of procyanidin, a 3', 4'-hydroxylated proanthocyanidin that has an affinity to bind proteins and RNA (TODD and VODKIN 1993 ). In contrast, imperfect black (i,R,t) or buff (i,r,t) seed coats contain propelargonidin, a 4'-hydroxylated proanthocyanidin, and no procyanidin. Therefore, in homozygous recessive i genotype, the T-t pair of alleles determines the type of anthocyanin and proanthocyanidin pigments present in mature seed coats. These findings led BUZZELL et al. 1987 to hypothesize and TODD and VODKIN 1993 to support that the T locus encodes a microsomal flavonoid 3' hydroxylase (F3'H) involved in the hydroxylation of the B ring in the formation of cyanidin-3-glucoside (Fig 3).

View larger version (25K): In this window In a new window Download PPT slide   Figure 3. Schematic of the anthocyanin biosynthesis pathway in G. max. Enzymes are indicated in uppercase letters. I, R, and T affect seed coat pigmentation. R is hypothesized to act after the formation of leucoanthocyanidins. PAL, phenylalanine ammonia-lyase; C4H, cinnamate 4-hydroxylase; 4CL, 4-coumarate: CoA ligase; CHS, chalcone synthase; CHI, chalcone isomerase; F3'H, flavonoid 3' hydroxylase; F3'5'H, flavonoid 3', 5' hydroxylase; F3H, flavanone 3 hydroxylase; DFR, dihydroflavonol-4-reductase; UFGT, UDP-flavonoid glucosyltransferase.

The F3'H enzyme is a cytochrome P450-dependent monooxygenase and these membrane-bound proteins are difficult to isolate. Cloning and identification of these genes by homology to the highly conserved regions shared by the P450 family in plants is complicated by the large number of P450-like sequences (SCHULER 1996 ; CHAPPLE 1998 ). BRUGLIERA et al. 1999 were the first to report the isolation of an F3'H cDNA (EMBL/GenBank accession no. AF155332) from Petunia hybrida and its linkage to the Ht1 locus. SCHOENBOHM et al. 2000 identified the gene TT7 encoding the F3'H in Arabidopsis thaliana (AtF3'H) in two different ecotypes and characterized the mutation of tt7 plants with pale brown seeds and reduced anthocyanin content of the whole plant body.

Here we report restriction fragment length polymorphism (RFLP) and expression data that identify GmF3'H as the T locus. The mRNA was found in highest levels in the developing seed coats; none was found in the cotyledons. Molecular characterization of cDNAs and genomic sequences from six soybean lines with varying T (t) alleles revealed that there are two types of t alleles in the lines examined. The stable t allele present in many soybean varieties results from a single-base deletion in the 3' half of the coding region leading to a truncated reading frame and gray pubescence. In contrast, the level of mRNA expression of the flavonoid 3' hydroxylase locus appears to be affected in a stable gray line with the t* allele that is derived from a mutable T locus.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant materials and genotypes:
The G. max cultivars and isolines used for this research are described in Table 1. Except for XB22A, 37609, and 37643 that were provided by Pioneer Hi-Bred International, all other cultivars and isolines were obtained from the United States Department of Agriculture (USDA) Soybean Germplasm Collections (Department of Crop Sciences, USDA Agricultural Research Service, University of Illinois, Urbana, IL). The genotypes and phenotypes of the lines used are shown in Table 1. All lines are homozygous and only one of the alleles at each locus is shown for brevity in the tables and text.

Plants were grown in the field or greenhouse. Seed coats dissected from seeds at varying stages of development, shoot tips, stems, mature leaves, and roots were frozen in liquid nitrogen, freeze dried (Multi-dry lyophilazer; FTS systems), and stored at -20°. For seed coat developmental studies, seeds were divided into the following groups according to the fresh weight of the entire seed: 25–50 mg, 50–75 mg, 75–100 mg, and 100–200 mg.

DNA isolation and DNA gel-blot analysis:
Genomic DNA was isolated from soybean freeze-dried shoot tips using the methods of DELLAPORTA 1993 with minor modifications. The nuclease inhibitor O-phenanthroline (10 mM) was added to the extraction buffer and the hexadecyltrimethylammonium bromide step was omitted. Genomic DNA (10 µg) was digested with restriction endonucleases BamHI, BclI, DraI, EcoRI, PstI, SacI, and XhoI for 2 hr at 37° and electrophoresed in a 0.7% agarose gel (SAMBROOK et al. 1989 ).

Size-fractionated DNAs were transferred to Optitran-supported nitrocellulose membrane (Midwest Scientific, Valley Park, MO) by capillary action in 10x SSC (0.15 M NaCl, 0.015 M sodium citrate) as described in SAMBROOK et al. 1989 and crosslinked in a UV stratalinker (Stratagene, La Jolla, CA). Nitrocellulose DNA blots were prehybridized, hybridized, washed, and exposed to Hyperfilm (Amersham, Arlington Heights, IL) as described by TODD and VODKIN 1996 .

RNA extraction and RNA gel-blot analysis:
Total RNA was isolated from seed coats and other soybean tissues using phenol-chloroform and lithium chloride precipitation methods (MCCARTY 1986 ; WANG et al. 1994 ). RNA was stored at -70° until used. RNA (10 µg) was electrophoresed in a 1.2% agarose-3% formaldehyde gel (SAMBROOK et al. 1989 ). Transfer of fractionated RNAs to supported nitrocellulose membrane was carried out as described for DNA gel blots.

cDNA synthesis:
Complete cDNA copies of the GmF3'H genes from cultivars Richland, Harosoy, XB22A, and 37609 were amplified from a first-strand cDNA pool synthesized using 1 µg of seed coat total RNA and the SuperScript first-strand synthesis system for reverse transcriptase (RT)-PCR (Invitrogen, San Diego). The sequences of the two primers used were 5'-CCTAGTCTGAAACCATAGCACAAAATCAACC-3' and 5'-AATCATTGAATCCCATCCATATAGCATATA-3'.

Screening of soybean cDNA on high-density nylon filters:
The NSF-soy Gm-c1019 high-density filter containing cDNAs from an immature seed coat library (A. Khanna and L. Vodkin) was prepared by Incyte Genomics (St. Louis). Prior to hybridization, the filter was washed as suggested by the supplier (0.5% SDS in H₂O at 60°). Prehybridization, hybridization, and washes were carried out as described for DNA gel blots.

Probes for DNA and RNA gel blots and high-density cDNA nylon filters:
Cloned DNAs used as probes were digested from their vectors, electrophoresed, and purified from the agarose using the QIAquick gel extraction kit (QIAGEN, Valencia, CA). DNA concentration of the final eluate was determined by comparison to a known amount of DNA standard upon gel electrophoresis. Purified DNA fragments (25–250 ng) were labeled [-³²P]dATP or [-³³P]dATP (for the high-density nylon filter) by random primer reaction (FEINBERG and VOGELSTEIN 1983 ). The unincorporated nucleotides were removed using Bio-Spin 30 chromatography columns (Bio-Rad, Richmond, CA).

Primer design, PCR reaction conditions, and DNA sequencing:
The P. hybrida flavonoid 3' 5' hydroxylases (Z22544.1 and Z22545.1) and flavonoid 3' hydroxylase (AF155332.1) DNA sequences were obtained from GenBank and aligned using BCM Search Launcher multiple sequence alignment (http://searchlauncher.bcm.tmc.edu/multi-align/multi-align.html). Degenerate primers representing regions with the highest homology were synthesized on an Applied Biosystems (Foster City, CA) model 394A DNA synthesizer at the Keck Center, a unit of the University of Illinois Biotechnology Center. The forward 5'-GTTTTTGCACCTTATGGWCC-3' and reverse 5'-CCATATGCTTCTTCCATATTMA-3' primer pair was used successfully to amplify a GmF3'H genomic clone (1A). Multiple primer pairs were synthesized to complete the GmF3'H cDNA sequence as well as to amplify and sequence the GmF3'H genomic DNA from seven different soybean lines.

Soybean genomic DNA fragments encoding the GmF3'H gene were obtained via PCR from several lines that are homozygous for the dominant T allele (Williams and XB22A with tawny pubescence) and from lines that are homozygous for the recessive t or t* alleles (Harosoy, Richland, T157, 37609, and 37643 with gray pubescence). Most PCR reactions were performed by an initial denaturation step at 96° for 2 min followed by 39 cycles of denaturing at 96° for 20 sec, annealing at 36° for 1 min, and polymerization at 72° for 2 min, to end with a 7-min extension at 72°. To amplify the 5' end of the gene, a higher denaturation temperature of 98° was required due to the higher GC content of this DNA region. High-fidelity and -efficiency polymerases Takara ExTaq and LA-Taq (Panvera) were used for these PCR reactions.

Genomic DNA fragments resulting from amplification with the degenerate primers were fractionated in a 0.7% agarose gel, purified with a QIAquick gel extraction kit (QIAGEN), and cloned into a pGEM-T-Easy vector (Promega, Madison, WI) in preparation for sequencing. Those genomic DNA fragments as well as the cDNAs generated via RT-PCR were sequenced directly after gel fractionation and purification with a QIAquick gel extraction kit.

Sequencing of cDNA, genomic clones, and purified PCR fragments was carried out at the Keck Center.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Cloning of a soybean flavonoid 3' hydroxylase using degenerate primers:
Three distinct classes of flavonoid hydroxylases are required for the synthesis of the three anthocyanin types: delphinidin-3-glycoside, pelargonidin-3-glycoside, and cyanidin-3-glycoside. Flavonoid 3 hydroxylase (F3H) adds a hydroxyl group to the C ring's carbon 3 position of naringenin, eriodictoyl, and 5' OH eriodictoyl, while flavonoid 3' hydroxylase (F3'H) and flavonoid 3', 5' hydroxylase (F3'5'H) hydroxylate the 3' or 3' and 5' positions of the naringenin B ring, respectively, and determine the type of anthocyanin synthesized (Fig 3). At the time we set out to identify and clone the genes encoding these proteins in soybeans, one F3'H (AF155332) and two F3'5'H (Z22544 and Z22545) cDNAs from P. hybrida had been cloned and sequenced (HOLTON et al. 1993 ; BRUGLIERA et al. 1999 ). The two PhF3'5'H sequences are 94% identical at the amino acid level (SCHULER 1996 ) while the PhF3'H is 65% similar to the PhF3'5'H sequences (BRUGLIERA et al. 1999 ). On the basis of a ClustalW alignment of the three petunia cDNAs, four portions of DNA sequence with the highest level of identity were selected. We designed degenerate oligonucleotide primers with homology to those regions and used them to amplify the corresponding sequence from genomic DNA of G. max cultivar Williams, via the polymerase chain reaction (PCR). The resulting 1-kb soybean amplicon (designated 1A) was sequenced and compared to the PhF3'H and PhF3'5'H cDNAs. It had 76% identity to the second half of the PhF3'H cDNA.

Isolation and characterization of a full-length flavonoid 3' hydroxylase cDNA using soybean genomics resources:
Using the DNA sequence of the 1A soybean genomic clone in a BLAST search, we then identified a dbEST clone Gm-c1025-1209 from a soybean expressed sequence tag (EST) project (SHOEMAKER et al. 2002 ) that matched a 564-bp region of the 1A soybean genomic clone. Using the Gm-c1025-1029 as a probe, we then screened a high-density nylon filter containing 18,000 cDNAs from an immature seed coat soybean library. Four additional cDNAs (Gm-c1019-10961, -7295, -17665, and -5252) were identified and sequenced. Fig 4 is a map of the cDNAs isolated from the filter screening as well as all cDNAs from the soybean EST project as compared to the soybean genomic clone 1A.

View larger version (12K): In this window In a new window Download PPT slide   Figure 4. G. max flavonoid 3' hydroxylase cDNA sequence contig. The cDNA clones Gm-1019-10961, -7295, -5252, and -17665 were identified as F3'H sequences upon hybridization of a cDNA high-density filter of library Gm-1019 to the Gm-c1025-1209 clone. GenBank accession numbers of the cDNA clones and the tissue and library of origin are given in Table 2 and Table 3. All the clones were isolated from cDNA libraries synthesized with RNA from Williams cultivar with T genotype except for clone Gm-c1053-348 that was isolated from a cDNA library made with RNA from Harosoy with a mutant t genotype. The sequence of this clone contains intron sequence at the 3' end where it terminates unexpectedly. The map location of the 1A genomic clone is shown with a dashed line and the position of an intron is shown with a black triangle. The consensus sequence for GmF3'H is represented by a thicker line at the bottom and corresponds to the longest cDNA, that of clone Gm-1019-10961.

All six cDNA clones were sequenced in their entirety and their DNA sequences are available from GenBank with the accession numbers listed in Table 2 along with information about the library and cultivars of origin. With the exception of clone Gm-c1053-348, they all are tawny pubescent lines. The sizes of the cDNA clones are also indicated in Fig 4. The cDNA clone Gm-c1019-10961 is a full-length clone of 1840 nucleotides and differs from clone Gm-c1019-7297 in that it contains 14 additional nucleotides at the 5' end. With the exception of cDNA clone Gm-1053-348 (which is discussed later), the others were shorter cDNA clones representing priming of the mRNA at the poly(A) tail and premature termination of the reverse transcriptase reaction. An additional cDNA sequence (AB061212) with high identity to that of the Gm-c1019-10961 clone except for a base-pair substitution at position 382 is also present in GenBank (Table 2 and Table 3).

Sequence analysis predicts a putative open reading frame of 1539 bp that encodes a polypeptide of 513 amino acids (Fig 5). Fig 5 shows the ClustalW1.8, multiple sequence alignment resulting from comparing GmF3'H deduced amino acid sequence to those of PhF3'H and AtF3'H. A high degree of identity exists among the three sequences except at the first 30 amino acids and at a second stretch, between positions 261 and 289, where all three molecules diverge. The soybean amino acid sequence had 67 and 66% identical residues to the petunia and Arabidopsis F3'H proteins, respectively. Overall, it has 75% similar amino acids to each of these proteins, taking into account the conserved amino acid substitutions (Fig 5). In contrast, the soybean F3'H had only 47 and 48% identical amino acids to the petunia Hf1 and Hf2 deduced amino acid sequences (CAA80265 and CAA80266) encoding the P. hybrida flavonoid 3', 5' hydroxylases. In addition, the GmF3'H deduced amino acid sequence contains the "GGEK" motif (423–426) that distinguishes F3'H from F3'5'H enzymes and is characteristic of all F3'H genes sequenced to date (BRUGLIERA et al. 1999 ; Fig 5). These results indicate that the Gm-c1019-10961 cDNA clone is a full-length cDNA that encodes the flavonoid 3' hydroxylase enzyme in soybean.

View larger version (107K): In this window In a new window Download PPT slide   Figure 5. Flavonoid 3' hydroxylase amino acid sequence alignment. Comparison of G. max flavonoid 3' hydroxylase-deduced amino acid sequence (Gm) with those of putative orthologous proteins from A. thaliana (At) and P. hybrida (Ph) is shown. The BOXSHADE server of BCM Search Launcher (http://www.ch.embnet.org/software/BOX_form.html) was used to highlight amino acids from the GmF3'H sequence with identity to either of the two sequences, PhF3'H and AtF3'H. Black boxes indicate identical residues and gray boxes indicate conserved residues between the three proteins. Amino acid numbering is shown at left. Four black dots indicate the location of the GGEK motif.

Restriction site polymorphisms associated with the T locus:
To determine if the GmF3'H gene is indeed linked to the T locus, we compared RFLPs of 10 soybean lines (Table 1) varying at the T locus. The recessive gray pubescence is a stable trait that is prevalent in many soybean lines. Line L68-2056 is an isoline created by backcrossing a line having gray pubescence (tt genotype) to the recurrent Clark parent (TT, tawny pubescence) for more than six generations. On the other hand, the 37609 and 37643 lines also have gray pubescence but their pedigree traces to an unstable line. In the mid-1980s, soybean breeders at Pioneer Hi-Bred found a single rogue plant that exhibited branches of both tawny and gray pubescence in the F₆ generation of an unreleased breeding line, XB22A (TT, tawny pubescence; TODD 1992 ). Multiple lines were derived from that original plant and some continued to exhibit the chimeric phenotype in variable patterns, while others bred true for the stable gray or stable tawny phenotype. The 37609 and 37643 lines with gray pubescence are two of the early descendants from the mutable plant and they produce progeny with gray pubescence in a stable manner. To differentiate the nature of the allele derived from the unstable plant, we designate it as t*, and thus the inbred stable gray lines 37609 and 37643 have genotype t*t*.

Fig 6 shows that multiple DNA polymorphisms were found between T and t genotypes as illustrated for DNAs cleaved with BclI (Fig 6A), BamHI (Fig 6B), and PstI (Fig 6C). There appeared to be three distinct patterns. For example, two BclI fragments of 11 and 1.6 kb that do not appear in genotypes containing the recessive t allele were found in genotypes with the dominant T allele (Fig 6). Instead, two fragments of 9 and 8 kb hybridized to the GmF3'H probe in the t lines. In addition, a shift of a 1.8-kb band to one of 1.9 kb was also manifested in the t lines (Fig 6A, lanes 4, 7, and 8). Similar results were obtained with BamHI restriction patterns in which 5- and 12.2-kb fragments found in genotypes carrying the T allele are replaced by 5.4- and 17.9-kb fragments in isolines with the recessive t allele (Fig 6B). Similarly, restriction fragment length polymorphisms using PstI showed substitution of a 9-kb fragment in lines having the T allele for a slightly larger fragment of 9.2 kb (Fig 6C). Additional polymorphisms were found in blots of DraI, EcoRI, SacI, and XhoI restriction digests (data not shown). In a separate experiment, hybridization of the GmF3'H probe to BamHI digests of DNAs from Williams (T), Richland (t), and T157 (t) lines showed the same differences in hybridization pattern between T and t lines described above (data not shown). A tight linkage of the GmF3'H probe with the T locus is strongly indicated given that the polymorphisms associated with the t allele (Fig 6, lane 4) are introgressed during six generations of backcrosssing to the recurrent Clark parent (TT genotype, tawny) to create isoline L68-2056 with gray pubescence (tt genotype).

View larger version (84K): In this window In a new window Download PPT slide   Figure 6. RFLP analysis of soybean lines varying at the T locus. DNA blot analysis of genomic DNA from nine color mutant soybean lines digested with three different restriction enzymes is shown. (A) BclI; (B) BamHI; (C) PstI. The genotype and phenotype of each line are described in Table 1. The type of T allele of each line is indicated at the top. The sizes of marker fragments are indicated at the right in kilobases. The Gm-c1025-1209 (AF499735) cDNA clone was used as probe (Fig 4).

In contrast to the results obtained with those t lines, no polymorphisms between XB22A (T) and 37609 (t*) were found using BclI or BamHI. However, a difference was observed with PstI where a 9-kb fragment was replaced with a 15-kb one (Fig 6C, lane 9). Differences between the 37609 mutant line (t*) and those carrying the T allele were also detected with SacI and EcoRI (data not shown). But once again, these changes differed from those found with the other three lines with t genotype. These polymorphic differences between the lines with the t allele are consistent with the independent origin of the t* mutation.

Restriction site polymorphisms of the GmF3'H gene cosegregate with the T locus:
An F₂ population of plants segregating pubescence color was created from a cross between L66-14 (T, tawny pubescence) and L68-2056 (t, gray pubescence) and analyzed to determine whether or not the GmF3'H polymorphisms cosegregated with the T locus.

The phenotype of the segregating plants was scored in the field and leaf tissue harvested and lyophilized. Genomic DNAs purified from portions of mature leaves of 42 segregating plants were digested with BamHI or BclI and DNA blots containing those DNAs were hybridized to the GmF3'H probe. The plants were expected to be genetically homozygous (TT) or heterozygous (Tt) for the tawny pubescence phenotype or to have gray pubescence (tt). In blots using BamHI-digested DNA, three GmF3'H polymorphisms corresponding to the two parent types, A (5-kb DNA fragment) and B (5.4-kb fragment), and the heterozygous pattern H (5- and 5.4-kb bands) were found in the segregating population. A summary of the polymorphism and phenotype of the segregating plants is shown in Table 4. The GmF3'H polymorphism segregated in a 1:2:1 ratio, representing a single locus with codominant alleles. Segregation ratio for pubescence color was 3:1 as expected for a single gene with dominant-recessive inheritance. Recombination analysis showed a clear cosegregation of the GmF3'H polymorphism with pubescence color (Table 4). The absence of recombinant types between the T locus and the GmF3'H polymorphism indicates either that T encodes GmF3'H or that there is a tight linkage of the 5-kb GmF3'H DNA fragment with the T allele. According to HANSON's (1959) equation, the maximum recombination value that might exist between the two loci is 0.069 assuming a 95% probability of not observing a recombinant among 42 individuals in the F₂ population. Coupled with the expression data in mutant lines described below, we conclude that the GmF3'H is the T locus.

G. max flavonoid 3' hydroxylase tissue-specific expression:
The T locus controls the color of the seed coats and pubescence hair on stems, leaves, and pods of soybean plants. To ascertain where in the plant the GmF3'H gene is expressed, RNA blots containing total RNA from several plant parts were hybridized to the GmF3'H probe (Gm-c1019-7295). The results shown in Fig 7 revealed that the GmF3'H probe hybridized with an 1.8-kb transcript and that the highest expression of this transcript occurred in the seed coat (25- to 50-mg seeds) of cultivar Williams (iⁱ,R,T; Fig 7A, lane 5; Table 1). Much weaker hybridization to the 1.8-kb RNA was detected in stems and shoot tips of 3-week-old plants. Shoot tips contain the meristem and a few, very young, developing leaves. This apparent low level of expression may be compartmentalized in the developing trichomes of young stems and developing leaves. No hybridization to RNAs from fully expanded mature leaves and roots of 3-week-old plants or to RNAs from developing cotyledons (25- to 50-mg seeds; Fig 7A, lanes 1, 3, and 6) was observed. Low levels of expression, similar to those detected for young stems and shoot tips, were found in flower buds of this Williams cultivar that produce white flowers. RNA of mature flowers did not hybridize to the probe (data not shown).

View larger version (74K): In this window In a new window Download PPT slide   Figure 7. G. max flavonoid 3' hydroxylase tissue-specific expression. (A) RNA blot containing 10 µg of total RNA samples purified from mature leaves, stems, roots, shoot tips, seed coats, and cotyledons of soybean plants, cv. Williams (ii,R,T). Seed coat total RNA from XB22A cv. (ii,r,T) and a mutant isoline, 37609 (ii,r,t*), are also included. RNA molecular markers are indicated in kilobases. The Gm-c1019-7297 cDNA clone was used as the probe (Fig 4). (B) Ethidium bromide-stained gel prior to membrane RNA transfer.

The high level of GmF3'H gene expression found in developing seed coat tapers off at latter stages of development (75- to 100-mg seeds) and it is barely detectable in seed coats of seeds 100–200 mg fresh weight (Fig 8A, lanes 17–20). The developmental decline observed in F3'H seed coat expression may also take place in developing trichomes and thus explain the lack of GmF3'H hybridization to RNA from mature leaves and flowers (Fig 7A, lane 1 and data not shown).

View larger version (67K): In this window In a new window Download PPT slide   Figure 8. G. max flavonoid 3' hydroxylase expression during seed coat development in soybean lines varying at the T locus. (A) RNA blot containing 10 µg of total RNA from four seed coat developmental stages in five color mutant soybean lines: Richland (I,R,t), T157 (i,R,t), XB22A (ii,r,T), 37609 (ii,r,t*), and Williams (ii,R,T). Seed fresh weight of each seed coat grouping in milligrams is shown at bottom. RNA molecular markers are indicated at right in kilobases. The Gm-c1019-7297 cDNA clone was used as the probe (Fig 4). (B) Ethidium bromide-stained gel prior to membrane RNA transfer.

The GmF3'H differential tissue expression correlates with the tissue-specific and developmental synthesis of anthocyanin and proanthocyanidin pigments, such as cyanidin and proanthocyanidin. The synthesis of the latter two compounds was shown to be controlled by the T locus (TODD and VODKIN 1993 ). These results represent additional evidence linking GmF3'H to the T locus.

G. max flavonoid 3' hydroxylase expression during seed coat development in soybean lines varying at the T locus:
Once it was determined that the GmF3'H gene expresses strongly in seed coats of the soybean variety Williams (iⁱ,R,T), with black hilum and tawny pubescence, it was relevant to analyze the expression of this gene in other soybean varieties with wild-type and mutant alleles of the T locus. RNA blots containing RNAs from seed coats at different stages of development from Williams (iⁱ,R,T) and two other cultivars, Richland (I,R,t) and XB22A (iⁱ,r,T), as well as their respective mutant lines, T157 (i,R,t) and 37609 (iⁱ,r,t*; Table 1), were hybridized to the GmF3'H probe (Gm-c1019-7295). The hybridization results shown in Fig 8A revealed a high level of expression of the 1.8-kb transcript at early stages of seed coat development (25- to 75-mg seed fresh weight) in three soybean lines, Richland, T157, and Williams. This high level of expression dropped sharply when the seed reached 100–200 mg fresh weight. The GmF3'H gene is also expressed, although at a lower level, in the XB22A variety. In contrast, mutant isoline 37609 carrying the t* allele appeared to completely lack the 1.8-kb transcript that hybridizes to the GmF3'H. In a separate hybridization experiment, using two different seed coat RNA batches (50- to 100-mg seed fresh weight) from XB22A and its mutant isoline 37609, similar results were obtained. The 1.8-kb transcript is clearly synthesized in the XB22A line carrying the T dominant allele (Fig 7A, lane 7) but is not detectable in the 37609 mutant (t*) line (Fig 7A, lane 8).

Richland and its mutant isoline T157 both are homozygous for the recessive t allele but the mutation in the allele does not appear to affect GmF3'H transcription considerably. High levels of the 1.8-kb transcript were detected at early stages of seed coat development in both sets of lines (Fig 8A).

A parallel study to the one described for the developing seed coats was carried out with corresponding developing cotyledons of the same five soybean lines. As shown for the Williams cotyledon RNA sample in Fig 7, lane 6, no expression of the GmF3'H gene was detected in any of the lines at any stage of cotyledon development (data not shown). This complete lack of expression in the cotyledons suggests a tight tissue-specific regulation of the GmF3'H gene in soybeans.

G. max flavonoid 3' hydroxylase genomic DNA sequences from soybean lines with varying T genotypes:
To further characterize the mutations at the T locus, genomic DNAs were amplified from Williams (T), Richland (t), T157 (t), Harosoy (t), XB22A (T), 37609 (t*), and 37643 (t*) soybean lines (Table 1), using four pairs of primers derived from the GmF3'H cDNA sequence. A schematic representation of the resulting amplified regions is shown in Fig 9. The two fragments of genomic sequence per each soybean line have been entered in GenBank as segments 1 (S1) and 2 (S2) and their accession numbers are listed in Table 3.

View larger version (10K): In this window In a new window Download PPT slide   Figure 9. Schematic representation of the GmF3'H genomic sequence segments featuring differences between T alleles. The genomic sequence of the GmF3'H gene from seven different soybean lines, Richland (I,R,t), T157 (i,R,t), Harosoy (I,r,t), XB22A (ii,r,T), 37609 (ii,r,t*), 37649 (ii,r,t*), and Williams (ii,R,T), was determined from four PCR fragments each. The resulting two segments of genomic sequence per each line, except for Harosoy where only segment 2 was sequenced, are indicated as S1 and S2. The S1 sequences were identical in the six soybean lines except for a base substitution at 414 in the t lines. The largest S2 segments contained an intron (Intron II) indicated with dotted lines, relative to their position in the cDNA depicted at the bottom. The 262-bp extra insertions in the introns of Richland, T157, and Harosoy lines are shown with dashed lines. The gap between the S1 and S2 segments corresponds to the 117 bp of cDNA sequence that we were unable to amplify and possibly contains a large intron (Large intron I ?). The sizes of the S1 and S2 segments and the cDNA are given in base pairs.

The largest of the two segments, S2, sequenced from Williams (T), XB22A (T), 37609 (t*), and 37643 (t*) lines was 2110 bp and contained an intron of 902 bp (designated intron II) located at position 958 in the cDNA clone Gm-c1019-10961. The S2 fragment from Richland (t), T157 (t), and Harosoy (t) lines was larger at 2372 bp. The difference between the two S2 fragments, 262 bp, was due to an insertion near the middle of Williams intron II (Fig 9).

Two sets of primers amplified the 5' end of the gene, called S1 in six of the lines. However, the small portion of the cDNA sequence of 117 bp in the 5' half of the gene shown in Fig 9 consistently failed to be amplified after many attempts using multiple primer pairs (up to 16 different pairs) at many different map locations and under various PCR conditions including use of high-fidelity and efficient polymerases (Takara ExTaq and LA-Taq, Panvera). The multiple failures of these PCR reactions likely indicate the presence of a very large intron (intron I in Fig 9) in this 5' region of the genomic sequence.

The genomic sequences of GmF3'H in Richland (t), T157 (t), and Harosoy (t) were identical and differed from those of Williams (T) and XB22A (T) lines. In addition to the 262-bp insertion into the intron, seven single-base changes were found in the intron sequences of those three lines when compared to those of Williams and XB22A varieties. Three other base differences were found outside the intron: a T substituting a C at base pair 414 of segment 1 (S1), an A substituting a G 58 bp downstream from the stop codon, and a base-pair deletion, C, in the coding region, 279 bp from the end of intron II (Fig 10). The T in place of C at base pair 414 results in a conserved amino acid change not having an effect on the translation product. Therefore, the only change that could account for the gray pubescence mutant phenotype of Richland, T157, and Harosoy (t) is the one-base deletion of a C at position 1498 denoted by an asterisk in Fig 10. This mutation creates a frameshift that could terminate the open reading frame prematurely, most likely resulting in a nonfunctional peptide.

View larger version (72K): In this window In a new window Download PPT slide   Figure 10. Mutations in Richland and Harosoy GmF3'H genomic sequences. Genomic sequences from seven soybean lines, Williams (ii,R,T), XB22A (ii,r,T), 37609 (ii,r,t*), 37649 (ii,r,t*), Richland (I,R,t), T157 (i,R,t), and Harosoy (I,r,t), were compared. The complete sequences are available from GenBank with accession numbers as listed in Table 2 and Table 3. Only three small portions of those sequences are shown here to point out two base substitutions (# and +) and the base deletion (*) that characterize the t allele found in cv. Harosoy and Richland and its T157 isoline.

No differences were detected among the genomic sequences of Williams (T), XB22A (T), and its mutant isolines, 37609 (t*) and 37643 (t*). The mutation in 37609 (t*) and 37643 (t*) could be located in the promoter region since no hybridizing transcripts were detected in RNA blots (Fig 7A, lane 8; Fig 8A, lanes 13–16) or, alternatively, in the putative large intron I that could not be amplified.

G. max flavonoid 3' hydroxylase RT-PCR from mutant lines:
To further characterize the nature of the mutation in Richland (t), Harosoy (t), and 37609 (t*) lines, we analyzed the sequences of GmF3'H cDNAs generated via RT-PCR from those lines and from XB22A (T). Using total seed coat RNAs from those four lines and 5' and 3' terminal primers designed to match the ends of Gm-c1019-10961 cDNA sequence, full-length cDNAs were obtained for Richland (t) and Harosoy (t) and entered into GenBank as listed in Table 2. The lack of any intron II sequences in these PCR products proves that they were amplified from mRNA and not genomic DNA contamination. These two cDNA sequences from the t lines were identical and differed from that of Williams (T) cDNA clone Gm-c1019-10961 in the bp substitutions and the base-pair deletion found when comparing their respective genomic sequences as discussed earlier. Thus, the base deletion that causes the frameshift mutation is present in the RNA population that was amplified as well as in the genomic DNA.

Fig 11 shows the alignment of Williams, Richland, and Harosoy cDNA amino acid-deduced sequences. The deletion of the C at 1498 of Fig 10 results in a codon change (CCC to CCA) that has no effect on translation; both codons translate into proline (P). However, the GmF3'H open reading frame in Richland and Harosoy terminates seven codons downstream from the base deletion, truncating the protein prematurely. In addition, the last five amino acids of the Richland and Harosoy-deduced protein sequence are different from those at the same position in the Williams sequence. The end result is a polypeptide lacking 124 amino acids at the 3' end. The signature sequence for the heme-binding domain (FxxGxxxCxG) of P450 enzymes (VON WACHENFELDT and JONSON 1995 ) lies on the carboxy terminus of the wild-type polypeptide derived from the Williams cDNA sequence, which is lacking in the Richland and Harosoy truncated polypeptides (Fig 11).

View larger version (100K): In this window In a new window Download PPT slide   Figure 11. Mutant and wild-type flavonoid 3' hydroxylase amino acid sequence alignment. Comparison of GmF3'H deduced amino acid sequences from Richland (t), Harosoy (t), and Williams (T) cDNAs. The proline (P) at position 388 where a 1-bp deletion in Richland and Harosoy shifts the frame of the wild-type open reading frame (Williams) is indicated with an arrow. That frameshift terminates the protein prematurely, eliminating 124 amino acids that contain the heme-binding domain FxxGxxxCxG. This is indicated with an underline.

The dbEST clone Gm-c1053-348 (Table 2; Fig 4) was isolated from a cDNA library constructed with RNA from Harosoy 3-week-old seedlings and therefore its cDNA sequence should match that of the RT-PCR-derived cDNA from Harosoy. However, clone Gm-c1053-348 cDNA sequence contains a portion of the intron sequence at the 3' end and terminates at a string of A's, 122 bp from the beginning of intron II (schematic in Fig 4, sequence not shown). This result suggests that EST clone Gm-c1053-348 must have been derived through false priming of contaminating genomic DNA at an intron region with multiple A's. Therefore, this EST from library Gm-c1053 does not reflect the true expression of the GmF3'H gene in the seedlings of 3-week-old Harosoy plants.

In the case of XB22A an 1840-bp cDNA with sequence identical to the one from cultivars Williams, cDNA clone Gm-c1019-10961, was synthesized through RT-PCR. Even though there was no detectable RNA from the 37609 (t*) line as determined by RNA blotting (Fig 8 and Fig 9), two RT-PCR fragments were amplified by the sensitive RT-PCR technique. In addition to the 1840-bp fragment, a larger cDNA (2029 bp) was reverse transcribed from the 37609 (t*) total RNA samples (Fig 12). The sequence of this larger cDNA was identical to that of the smaller one except for 189 extra base pairs at position 509 of the Gm-c1019-10961 cDNA (data not shown). The site of this cDNA insertion falls 8-bp from the location of the putative large intron I in the 5' half of the gene. Amplification of contaminating genomic DNA can be ruled out because the 902-bp intron II present in the 3' region of the GmF3'H gene was not present in the sequence of this RT-PCR product. This larger cDNA synthesized from the 37609 (t*) RNA samples could be the result of faulty editing in this mutant isoline if a very large intron exists in that region of the gene. Since this larger cDNA was not obtained in RT-PCR reactions with Richland (t), Horosoy (t), and XB22A (T) RNAs, its presence suggests that it may be a consequence of a specific defect in the 37609 (t*) isoline.

View larger version (61K): In this window In a new window Download PPT slide   Figure 12. A variant flavonoid 3' hydroxylase cDNA from the XB22A mutant isoline 37609. Ethidium bromide gel showing the 2.0-kb novel cDNA synthesized in two independent RT-PCR reactions with mRNA from the mutant isoline 37609 (t*) but not in XB22A (T genotype). Both XB22A and 37609 lines synthesized an 1.8-kb cDNA with identical sequences.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Biochemical studies had indicated that the soybean T locus that affects both trichome color and seed coat color would be a 3' flavonoid hydroxylase based on the type of flavonoids, anthocyanins, and proanthocyanidins synthesized and present in colored seed coats of various genotypes. Both cyanidin-3-monoglucoside and procyanidin, a 3', 4' hydroxylated proanthocyanidin, are found in black (i,R,T) seed coats while only procyanidin accumulates in brown (i,r,T) seed coats. However, imperfect-black (i,R,t) and buff (i,r,t) seed coats synthesize neither cyanidin nor procyanidin, suggesting the lack or inactivity of the F3'H enzyme in seed coats of plants with the t allele (BUZZELL et al. 1987 ; TODD and VODKIN 1993 ).

In this study, we confirmed this assignment at the molecular level by investigating the genetic structure and expression of the GmF3'H in a series of genetic lines with mutations in the T locus. We isolated a full-length flavonoid 3' hydroxylase cDNA from G. max (GmF3'H) with a high degree of similarity (75%) between the deduced amino acid sequences of G. max flavonoid 3' hydroxylase (GmF3'H) and those of the petunia (PhF3'H) and Arabidopsis (AtF3'H; Fig 5) using homology-based cloning with conserved primers and the cDNA resources from a soybean EST project (SHOEMAKER et al. 2002 ). The soybean F3'H also contains the GGEK motif characteristic of all F3'H isolated to date that distinguishes them from F3'5'H sequences (Fig 5; BRUGLIERA et al. 1999 ). Furthermore, restriction fragment length polymorphisms were found to correlate with well-characterized mutations in the T locus (Fig 6) and in a population of plants segregating for alleles of the locus (Table 4).

Proof that the GmF3'H gene is the T locus was obtained from analyzing the GmF3'H genomic sequences, RT-PCR-generated cDNAs, and GmF3'H gene expression in several soybean lines carrying variant alleles of T. The genomic sequences of Harosoy (I,r,t), Richland (I,R,t), and the mutant isoline T157 (i,R,t) were identical and differed from those of wild-type Williams (iⁱ,R,T) and XB22A (iⁱ,r,T) in an intron insertion, 8-bp substitutions, a 1-bp addition, and a 1-bp deletion. Only the 1-bp deletion and a 1-bp substitution took place within the open reading frame and, of these, only the 1-bp deletion affects the translation product. It creates a frameshift resulting in a truncated polypeptide lacking 124 amino acids at the carboxy terminus (Fig 11). This will render the enzyme inactive because it deletes the heme-binding domain required for this P450 monooxygenase enzyme to function. The putative inactivity of this mutant enzyme would explain the Richland and T157 mutant phenotype despite the high levels of GmF3'H transcripts synthesized in their seed coats (Fig 8). The extra 262 bp within the intron of all three varieties with the tt genotypes apparently does not negatively affect transcription from the gene.

Harosoy is a modern domestic soybean variety and Richland (PI 70.502-2) is a Chinese cultivar from Changling, Jilin, China (1926). Harosoy resulted from crossing Mandarin (Ottawa)(2) by A.K. (Arrow). Mandarin (Ottawa) was derived from Mandarin, a variety from Sui Hua, Heilongjiang, China (1913). A.K. (Arrow) was selected from A.K. that originated also from the Northeast of China. The fact that the GmF3'H genomic sequence is identical in Harosoy and Richland cultivars strongly suggests that either Richland and Mandarin had a common ancestor or Richland and A.K. (Arrow) did. Most modern soybean varieties are derived from these and a small number of other plant introductions to the U.S. in the early 1900s.

In contrast to the single-base deletion that causes a frameshift in the t allele, the molecular defect of the t* allele is different and results in very low levels of cytoplasmic mRNA transcripts. The genomic sequences obtained from XB22A (T) and mutant isolines, 37609 (t*) and 37643 (t*), were identical to that of Williams (T). Also, the sequences of XB22A (T) and 37609 (t*) cDNAs generated through RT-PCR were identical to the Williams (T) Gm-F3'H cDNA sequence. However, the larger cDNA molecule generated through RT-PCR with the 37609 (t*) RNA samples contained an extra 189 bp near the region corresponding to the genomic DNA of intron I that could not be amplified. It could be argued that the 37609 (t*) mutation has an effect on RNA processing. Perhaps it is due to an additional DNA insertion near or in the large putative intron I in the 5' half of the gene. The low level of cytoplasmic mRNAs observed in RNA blots is consistent also with a possible mutation in the promoter region. Such a defect reduces F3'H transcript abundance more severely than the one affecting the XB22A (T) gene. The latter results in lower transcript abundance and a lack of temporal expression when compared to the F3'H mRNA levels in Williams (T) (Fig 8); however, this change in the XB22A T locus does not result in gray pubescence as it does in 37609 (t*).

The mutable nature of the chimeric tawny/gray progenitor line is similar to the genetic behavior of a mutable allele (r^m) of the R locus in soybean that produces plants with both black and brown color seeds in the same plant. The r^m allele switches between its dominant and recessive forms both somatically and germinally at a high rate (CHANDLEE and VODKIN 1989 ). Likely, the mutable nature of the chimeric tawny/gray parent led to the genetic changes causing the abnormal expression of the flavonoid 3' hydroxylase gene in both the XB22A (T) and the 37609 (t*) stable isolines. Although not precisely identified, the mutation in the 37609 (t*) allele is clearly different from that of Richland (t) and Harosoy (t).

The GmF3'H gene is expressed mostly in pigmented tissues (seed coats and pubescence hair) with the highest levels of transcription found in seed coats of immature seeds (25–75 mg fresh weight) in those lines expressing the gene. Very low transcript levels were detected in shoot tips and young stems of the cultivar Williams with tawny pubescence (Fig 7). We propose that the transcription observed in young stems and shoot tips occurs in the pubescence hairs at early stages of development. The transcript concentration in the hairs will be diluted when combining the hair's RNA with that of other tissues in the stem or developing leaves of the shoot tip.

The complete lack of expression of the GmF3'H gene in cotyledons is noteworthy given the fact that it is expressed in high levels in the seed coats of plants with the dominant I allele [Richland (I,R,t); Fig 8] that inhibits chalcone synthase transcription (WANG et al. 1994 ). WANG et al. 1994 observed expression of the dihydroflavonol-4-reductase (DFR) gene that functions downstream of CHS in the anthocyanin pathway (Fig 3), in seed coats of a Clark isoline (I,R,t) that did not express CHS. The expression of GmF3'H and DFR at very early stages of seed development even in the absence of CHS transcription suggests that these genes and, therefore, the anthocyanin pathway are programmed to be turned on in the seed coats. On the other hand, evidence that isoflavones accumulate predominantly in the cotyledons but not in the seed coats suggests that these two branches of the flavonoid pathway (isoflavonoid and anthocyanin) are expressed differentially in cotyledons and seed coats. Consequently, the cloning of the GmF3'H gene becomes a necessary marker to study the regulation of this differential tissue expression as well as the channeling through the multiple branches of the flavonoid pathway in soybeans. Additionally, it will allow further characterization of the pleiotropic effects associated with the T locus, including how it affects structural integrity of the seed coat as well as pigmentation (NICHOLAS et al. 1993 ). It may have a role in determining the types of phenolic compounds associated with lignins and cell walls as well as with anthocyanin pigments in the vacuoles and trichome hairs.

	ACKNOWLEDGMENTS

We thank Anupama Khanna for advice on the cDNA library filter hybridizations. We gratefully acknowledge support from grants from the Illinois Soybean Program Operating Board, United Soybean Board (Public EST Project), and National Science Foundation grant DBI9872565.

Manuscript received June 29, 2002; Accepted for publication September 30, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BERNARD, R. L., and M. G. WEISS, 1973 Qualitative genetics, pp. 117–149 in Soybean: Improvement, Production and Uses, Ed. 1, edited by B. E. CALDWELL. American Society of Agronomy, Madison, WI.

BRUGLIERA, F., G. BARRI-REWELL, T. A. HOLTON, and J. G. MASON, 1999  Isolation and characterization of a flavonoid 3'-hydroxylase cDNA clone corresponding to the Ht1 locus of Petunia hybrida.. Plant J. 19:441-451.[Medline]

BUZZELL, R. I., B. R. BUTTERY, and D. C. MACTAVISH, 1987  Biochemical genetics of black pigmentation of soybean seed. J. Hered. 78:53-54.[Abstract/Free Full Text]

CHANDLEE, J. M. and L. O. VODKIN, 1989  Unstable expression of a soybean gene during seed coat development. Theor. Appl. Genet. 77:587-594.

CHAPPLE, C., 1998  Molecular-genetic analysis of plant cytochrome p450-dependent monooxygenases. Annu. Rev. Plant Physiol. 49:311-343.[Medline]

DELLAPORTA, S. L., 1993 Plant DNA miniprep version 2.1–2.3, pp. 522–525 in The Maize Handbook, edited by M. FREELING and V. WALBOT. Springer-Verlag, New York.

FEINBERG, A. P. and B. VOGELSTEIN, 1983  A technique for radiolabeling DNA restriction fragments to high specific activity. Anal. Biochem. 132:6-13.[Medline]

HANSON, W. D., 1959  Minimum family sizes for the planning of genetic experiments. Agron. J. 51:711-715.[Free Full Text]

HOLTON, A. H., F. BRUGLIERA, D. R. LESTER, Y. TANAKA, and C. D. HYLAND et al., 1993  Cloning and expression of cytochrome P450 genes controlling flower color. Nature 366:276-279.[Medline]

MCCARTY, D., 1986  A simple method for extraction of RNA from maize tissue. Maize Genet. Coop. Newsl. 60:61.

NICHOLAS, C. D., J. T. LINDSTROM, and L. O. VODKIN, 1993  Variation of proline rich cell wall proteins in soybean lines with anthocyanin mutations. Plant Mol. Biol. 21:145-156.[Medline]

PALMER, R. G., and T. C. KILEN, 1987 Qualitative genetics and cytogenetics, pp. 135–209 in Soybeans: Improvement, Production and Uses, Ed. 2, edited by J. R. WILCOX. American Society of Agronomy, Madison, WI.

TODD, J. J., 1992 Biochemical genetic analysis of flavonoid compounds in soybean. M.S. Thesis, University of Illinois, Urbana-Champaign, IL.

TODD, J. J. and L. O. VODKIN, 1993  Pigmented soybean (Glycine max) seed coats accumulate proanthocyanidins during development. Plant Physiol. 102:663-670.[Abstract]

TODD, J. J. and L. O. VODKIN, 1996  Duplications that suppress and deletions that restore expression from a chalcone synthase multigene family. Plant Cell 8:687-699.[Abstract]

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual, Ed. 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SCHOENBOHM, C., S. MARTENS, C. EDER, G. FORKMANN, and B. WEISSHAAR, 2000  Identification of the Arabidopsis thaliana flavonoid 3'-hydroxylase gene and functional expression of the encoded P450 enzyme. Biol. Chem. 381:749-753.[Medline]

SCHULER, M. A., 1996  Plant cytochrome P450 monooxygenases. Crit. Rev. Plant Sci. 15(3):235-284.

SHOEMAKER, R., P. KEIM, L. VODKIN, E. RETZEL, and S. W. CLIFTON et al., 2002  A compilation of soybean ESTs: generation and analysis. Genome 45:329-338.[Medline]

VON WACHENFELDT, C., and E. F. JONSON, 1995 Structures of eukaryotic cytochrome P450 enzymes, pp. 183–223 in Cytochrome P450: Structure, Mechanism, and Biochemistry, Vol. 2. Plenum, New York.

WANG, C., J. TODD, and L. O. VODKIN, 1994  Chalcone synthase mRNA and activity are reduced in yellow soybean seed coats with dominant I alleles. Plant Physiol. 105:739-748.[Abstract]

WOODWORTH, C. M., 1921  Inheritance of cotyledon, seed-coat, hilum, and pubescence colors in soy-beans. Genetics 6:487-553.[Free Full Text]

This article has been cited by other articles:

				G. Zabala and L. O. Vodkin A Rearrangement Resulting in Small Tandem Repeats in the F3'5'H Gene of White Flower Genotypes Is Associated with the Soybean W1 Locus Crop Sci., July 16, 2007; 47(S2): S-113 - S-124. [Abstract] [Full Text] [PDF]

				T. Iwashina, S. M. Githiri, E. R. Benitez, T. Takemura, J. Kitajima, and R. Takahashi Analysis of Flavonoids in Flower Petals of Soybean Near-isogenic Lines for Flower and Pubescence Color Genes J. Hered., May 1, 2007; 98(3): 250 - 257. [Abstract] [Full Text] [PDF]

				T. Iwashina, E. R. Benitez, and R. Takahashi Analysis of Flavonoids in Pubescence of Soybean Near-isogenic Lines for Pubescence Color Loci J. Hered., September 1, 2006; 97(5): 438 - 443. [Abstract] [Full Text] [PDF]

				K. Toda, M. Akasaka, E. G. Dubouzet, S. Kawasaki, and R. Takahashi Structure of Flavonoid 3'-Hydroxylase Gene for Pubescence Color in Soybean Crop Sci., September 23, 2005; 45(6): 2212 - 2217. [Abstract] [Full Text] [PDF]

				M. Senda, C. Masuta, S. Ohnishi, K. Goto, A. Kasai, T. Sano, J.-S. Hong, and S. MacFarlane Patterning of Virus-Infected Glycine max Seed Coat Is Associated with Suppression of Endogenous Silencing of Chalcone Synthase Genes PLANT CELL, April 1, 2004; 16(4): 807 - 818. [Abstract] [Full Text] [PDF]

				A. Hoshino, Y. Morita, J.-D. Choi, N. Saito, K. Toki, Y. Tanaka, and S. Iida Spontaneous Mutations of the Flavonoid 3'-hydroxylase Gene Conferring Reddish Flowers in the Three Morning Glory Species Plant Cell Physiol., October 15, 2003; 44(10): 990 - 1001. [Abstract] [Full Text] [PDF]