A genomic region harboring the Pl1 allele from the Peruvian cultivar JC072A confers purple cob on Japanese flint corn (Zea mays L.)

Abstract

Purple corn is a maize variety (Zea mays L.) with high anthocyanin content. When purple corn is used as forage, its anthocyanins may mitigate oxidative stresses causing lower milk production in dairy cows. In this study, we analyzed quantitative trait loci (QTLs) for anthocyanin pigmentation of maize organs in an F₂ population derived from a cross between the Peruvian cultivar ‘JC072A’ (purple) and the inbred line ‘Ki68’ (yellowish) belonged to Japanese flint. We detected 17 significant QTLs on chromosomes 1–3, 6, and 10. Because the cob accounts for most of the fresh weight of the plant ear, we focused on a significant QTL for purple cob on chromosome 6. This QTL also conferred pigmentation of anther, spikelet, leaf sheath, culm, and bract leaf, and was confirmed by using two F₃ populations. The gene Pl1 (purple plant 1) is the most likely candidate gene in this QTL region because the amino acid sequence encoded by Pl1-JC072A is similar to that of an Andean allele, Pl-bol3, which is responsible for anthocyanin production. The markers designed for the Pl1 alleles will be useful for the breeding of F₁ lines with anthocyanin pigmentation in cobs.

Introduction

Maize (Zea mays L.) is grown worldwide as a cereal and fodder crop, and grain color variants (white, yellow, and purple) are also used to extract colorants in food and beverages. Purple corn is originated from the species “Kculli”, still cultivated in Peru (Grobman 1961). Its color is determined by anthocyanins, which are used to color jellies and syrups and as a health supplement in Japan (Aoki et al. 2002). Anthocyanins from purple corn have high antioxidant activity (Cevallos-Casals and Cisneros-Zevallos 2003), inhibit colorectal carcinogenesis in male rats (Hagiwara et al. 2001), reduce blood pressure in male spontaneously hypertensive rats (Shindo et al. 2007) and reduce AST (aspartate aminotransferase) activity and an increase SOD (superoxide dismutase) activity in lactating dairy cows (Hosoda et al. 2012). Then, anthocyanins from purple corn may mitigate oxidative stresses in dairy cows.

In purple cobs, the anthocyanin content ranges from 30% to 47% of total phenolic components (Jing et al. 2007), and the accumulation of anthocyanins are increased by drought, low temperature, visible light, and UV radiation (Chalker-Scott 1999). The regulatory genes R, B, C1, and Pl are responsible for anthocyanin synthesis; by sequence similarity, they are grouped into two families, R/B and C1/Pl (Chandler et al. 1989, Cone et al. 1986). The interaction of each gene of these families changes the temporal and spatial pattern of anthocyanin pigmentation.

The R and B genes encode proteins homologous to the basic-helix-loop-helix (bHLH) DNA-binding–protein dimerization domain of Myc oncoproteins (Chandler et al. 1989, Consonni et al. 1993, Ludwig et al. 1989). The C1 and Pl genes encode R2R3 MYB transcription factors (Cone et al. 1986, 1993a, 1993b, Paz-Ares et al. 1986). Anthocyanin pigmentation of maize organs is controlled by at least 20 loci, comprising genes encoding DNA-bindingprotein and transcriptional factors (Mol et al. 1998). It is necessary to understand the genetic architecture responsible for anthocyanin synthesis in breeding materials.

A program to breed Japanese elite flint cultivars with anthocyanin pigmentation has been designed to mitigate the oxidative stress causing the lower milk production in dairy cows. We used an unpigmented inbred line, ‘Ki68’ belongs to Japanese flint (Tamaki et al. 2014), and an anthocyanincontaining cultivar, ‘JC072A’, in the breeding program. ‘Ki68’ has green bract leaf, culm, and leaf sheath and white cobs, whereas ‘JC072A’ shows them with dark purple. Using F₂ progeny of a cross between ‘Ki68’ and ‘JC072A’, we identified QTLs for anthocyanin synthesis and elucidated the genetic mechanisms that contribute to anthocyanin pigmentation.

Materials and Methods

Plant materials and pigmentation grading

The F₂ and F₃ populations were established from a cross between the inbred line ‘Ki68’ belonged to Japanese flint and the Peruvian cultivar ‘JC072A’. Anthocyanin pigmentation was assessed at the time of self-pollination (anther, spikelet, and silk), at the kernel dent stage (leaf sheath, culm, midrib of leaf blade, prop root, and bract leaf), and at the physiological matured stage (top and dorsal sides of grain, and glumes of cob). Anthocyanin pigmentation was graded from 1 (absent or very weak) to 9 (very strong).

Genotyping of F₂ and F₃ populations

Genomic DNA of 116 F₂ individuals was extracted by CTAB method and PCR was carried out for indel amplification at 148 loci with (1) SSR markers available at MaizeGDB (http://www.maizegdb.org/), (2) markers designed on the basis of BAC sequences with linkage mapping information (Supplemental Table 1), and (3) a Pl1 marker designed as described below. PCR was performed in a 384-well thermal cycler (Applied Biosystems, Foster City, CA, USA) with KAPA2G Fast Ready Mix (Kapabiosystems, Boston, MA, USA) as follows: 95°C for 1 min; 35 cycles at 95°C for 10 s, 55°C for 10 s, and 72°C for 1 s; and a final 72°C for 30 s. PCR products were electrophoresed in 3% agarose gels in TBE (Tris–borate–EDTA) buffer. Gels were stained with ethidium bromide and the banding patterns were photographed under ultraviolet light for genotyping. Linkage mapping and QTL analysis were performed as described by Takai et al. (2012).

To validate the QTL for cob anthocyanin pigmentation on chromosome 6, we selected two F₃ populations: F3-46 (67 plants) and F3-82 (70 plants). These populations were derived from F₂ individuals with heterozygous alleles at the significant QTL region on chromosome 6 and homozygous ‘JC072A’ alleles at the QTL region with a high LOD peak on chromosome 10 (Supplemental Fig. 1). Genotypes and anthocyanin pigmentation of F3-46 and F3-82 were compared.

Isolation of Pl1 and marker development

Genomic DNA was extracted from ‘Ki68’ and ‘JC072A’ seedlings. Pl1 was isolated using the primers pl1myb_s_86 and pl1myb_as_87 (Supplemental Table 2). The Pl1 fragments amplified from ‘Ki68’ and ‘JC072A’ were purified with ExoSAP-IT (Affymetrix, Santa Clara, CA, USA) according to the manufacturer’s instructions. The purified PCR products were directly sequenced on an ABI PRIME 3130xl genetic analyzer (Applied Biosystems) with primers newly designed on the basis of the determined nucleotide sequences: pl1_389-s, pl1_34-s, pl1myb_696-s, pl1_355-s, pl1myb_361-as, pl1_720-as, pl1_407-as, and pl1_495-as (Supplemental Table 2). The nucleotide sequences of the Pl1 alleles of ‘JC072A’ and ‘Ki68’ were assembled from the sequenced fragments and were registered in DDBJ (LC384358 and LC384359, respectively). To distinguish between the Pl1 alleles from ‘Ki68’ and ‘JC072A’, we performed fragment analysis with pl1_389-s and pl1_720-as primers as described by Yonemaru et al. (2009).

Phylogenetic analysis of Pl1

Known Pl1 (Pl) sequences (Cocciolone and Cone 1993, Cone et al. 1993b, Pilu et al. 2003) were obtained from GenBank/EMBL/DDBJ/NCBI: L19496 (Pl-Tx303), L19494 (Pl-Rh), Q41842 (Pl-Bh), AY135018 (Pl-bol3), AY135019 (Pl-W22), and L19495 (Pl-McC). Amino acid sequences were aligned with CLUSTALW 2.1 software (Larkin et al. 2007). Phylogenetic analysis was conducted by the Neighbor Joining method in MEGA 7.0 software (Kumar et al. 2016).

Results

Detection of QTLs for anthocyanin pigmentation in maize organs

Significant QTLs were detected in 5 regions of chromosomes 1–3, 6, and 10; they affected anthocyanin pigmentation in 10 of the organs assayed (Table 1). No significant QTLs were found for pigmentation of the dorsal side of the grain. The effect of the QTL on chromosome 1 (qACprr1) was detected only in prop roots [LOD score (LOD) 4.2; additive effect (AE) 0.6]. The effects of significant QTLs on chromosome 2 (48.9–61.6 cM) were detected in spikelets (qACspk2), leaf sheaths (qAClsh2), culms (qACcul2), and bract leaves (qACbrl2). The values of LOD (10.5–13.6) and AE (2.0–2.3) were similar in these 4 organs. We only found the QTL on chromosome 3 for spikelets (qACspk3); it had the lowest values of phenotypic variance explained among all detected QTLs (Table 1).

Table 1.

QTLs for anthocyanin pigmentation of maize organs

Organ	QTL Name	Chr.	Marker interval		LOD score	Additive effecta	Dominant effecta	% PVEb	Bin	No. of listed genes per Bin	Candidate genes
Anther	qACant6	6	92.7	107.8	4.9	0.8	0.5	3.4	6.04	40	pl1, sm1
Anther	qACant10	10	73.7	76.1	19.8	−1.9	0.1	52.0	10.06	51	lc1, lcm1, Nc1, Nc2, Nc3, p, r1, S, Sc, sn1
Spikelet	qACspk2	2	49.9	59.7	13.6	2.1	1.0	16.8	2.02–2.04	42, 35, 53	b1
Spikelet	qACspk3	3	19.5	30.7	4.9	−0.5	−1.0	0.0	3.01–3.04	25, 21, 16, 62	no candidates
Spikelet	qACspk6	6	93.7	101.8	12.4	1.5	0.7	14.6	6.04	40	pl1, sm1
Silk	qACsil10	10	73.4	80.0	18.4	−0.6	0.0	40.6	10.06	51	lc1, lcm1, Nc1, Nc2, Nc3, p, r1, S, Sc, sn1
Leaf sheath	qAClsh2	2	49.4	60.2	10.5	2.3	0.3	23.8	2.02–2.04	42, 35, 53	b1
Leaf sheath	qAClsh6	6	96.0	102.5	7.3	1.4	0.4	10.1	6.04	40	pl1, sm1
Culm	qACcul2	2	48.9	59.5	11.0	2.0	0.8	9.6	2.02–2.04	42, 35, 53	b1
Culm	qACcul6	6	93.2	103.8	8.1	1.6	0.3	12.5	6.04	40	pl1, sm1
Midrib of leaf blade	qACmlb10	10	73.5	81.5	5.3	0.4	0.0	15.0	10.06	51	lc1, lcm1, Nc1, Nc2, Nc3, p, r1, S, Sc, sn1
Prop root	qACprr1	1	181.3	193.2	4.2	0.6	−0.2	12.7	1.07–1.09	79, 69, 69	bz2
Prop root	qACprr6	6	92.9	101.4	7.8	0.9	0.4	9.9	6.04	40	pl1, sm1
Bract leaf	qACbrl2	2	49.9	61.6	12.1	2.2	1.4	12.3	2.02–2.04	42, 35, 53	b1
Bract leaf	qACbrl6	6	94.1	102.7	10.8	1.6	1.1	7.9	6.04	40	pl1, sm1
Glumes of cob	qACglc6	6	94.3	100.2	22.2	2.7	2.8	7.1	6.04	40	pl1, sm1
Top side of grain	qACtsg6	6	95.7	100.2	30.8	2.6	2.4	11.7	6.04	40	pl1, sm1

^aPositive values show that the JC072A allele increases values.

^bPercentage of phenotypic variance explained.

The QTL region on chromosome 6 (92.7–107.8 cM) significantly affected most organs: anther (qACant6), spikelet (qACspk6), leaf sheath (qAClsh6), culm (qACcul6), prop root (qACprr6), bract leaf (qACbrl6), glumes of cob (qACglc6), and the top side of grain (qACtsg6). Among all QTLs, qACtsg6 had the highest LOD (30.8) and qACglc6 had the highest AE (2.7).

The QTL region on chromosome 10 (73.4–81.5 cM) significantly affected pigmentation of anthers (qACant10), silks (qACsil10), and midrib of leaf blade (qACmlb10). The AE values of qACant10 (–1.9) and qACsil10 (–0.6) and that of qACmlb10 had opposite signs (0.4). In this region, no significant LOD peaks were observed for glumes of cob or grains (Supplemental Fig. 1).

Development of the Pl1 marker and association between Pl1 alleles and cob anthocyanin pigmentation

Because the cob accounts for most of the fresh weight of the plant ear, we focused on cob pigmentation. To avoid the interaction of the QTL on chromosome 10 with qACglc6, we selected two F₃ populations (F3-46 and F3-82) for the validation of qACglc6 with the same alleles in the QTL region on chromosome 10 as in ‘JC072A’. In maize, Pl1 on chromosome 6 is required for anthocyanin synthesis in dark purple cobs (Hollick et al. 1995), implying that the functional Pl1 allele is responsible for anthocyanin pigmentation in ‘JC072A’. To detect the association between qACglc6 and Pl1, we developed a PCR marker (primers pl1_389-s and pl1_720-as). The segregation of the Pl1 genotype coincided with that of cob anthocyanin pigmentation at a 3:1 ratio in both F₃ populations (consistency of 0.94 and 0.91, respectively; Fig. 1). These results confirm that the Pl1 gene on chromosome 6 is responsible for anthocyanin synthesis in cobs.

Fig. 1.

Relationship between cob anthocyanin score and Pl1 genotype in two F₃ populations. The degree of anthocyanin pigmentation of cobs in F₃-46 (67 plants) and F₃-82 (70 plants) ranged from 1 (absent or weak) to 9 (very strong). Chi-squared tests assuming a 3:1 ratio were conducted on segregation of the phenotype (x ≤ 3, non-colored; X ≥ 6, colored) and Pl1 genotype (Ki68-type allele, non-colored; JC072A-type and heterozygous, colored). χ² (P), chi-squared P value of the phenotype; χ² (G), chi-squared P value of the genotype; consistency (P/G), consistency ratio between phenotype and genotype.

Comparison of Pl1 sequences between ‘JC072A’ and ‘Ki68’

We used the reported nucleotide sequences of Pl1 (Cone et al. 1993b) to obtain the sequences of Pl1 of ‘Ki68’ and ‘JC072A’ and designated the alleles Pl1-Ki68 and Pl1-JC072A, respectively. The comparison between the amplified fragments of Pl1-Ki68 (1014 bp) and Pl1-JC072A (1030 bp) showed that Pl1-Ki68 contained 2 SNPs in exon 1, 6 SNPs, and a 18-bp deletion in exon 3, 3-bp and 2-bp insertions in intron 1, and a 3-bp deletion in intron 2 (Fig. 2, Supplemental Fig. 2).

Fig. 2.

Comparison of Pl1 gene structures between ‘JC072A’ and ‘Ki68’. Boxes, exons; horizontal bars, introns. The positions of insertions (▽), deletions (△), and single nucleotide polymorphisms in ‘Ki68’ are shown. The sequence alignment is shown in Supplemental Fig. 2.

We aligned the deduced Pl1 amino sequences from ‘Ki68’ and ‘JC072A’ with those encoded by Pl-Rh, Pl-McC, Pl-Tx303, Pl-Bh, Pl-bol3, and Pl-W22 (Fig. 3). Pl1-Ki68 and Pl-Tx303 differed by substitutions of 2 amino acid residues and deletion of 1 residue in Pl-Tx303 (Supplemental Fig. 3). Pl1-JC072A and Pl-bol3 were of the same length with an amino acid substitution (Supplemental Fig. 3). In Pl-McC and Pl-W22, 3 residues around amino acid 190 were deleted and a residue was inserted relative to those of Pl-TX303 and Pl1-Ki68; the cobs of these cultivars are white. To check whether the identified Pl1 alleles mapped to chromosome 6, we designed a pair of primers for the Pl1 marker based on the identified nucleotide sequences of ‘Ki68’ and ‘JC072A’. Mapping and QTL analysis showed that Pl1 is located on chromosome 6, as previously reported, and detected a large-effect QTL at the Pl1 locus (Supplemental Fig. 4).

Fig. 3.

Phylogenetic analysis of Pl1 (Pl) based on amino acid sequences. Numbers indicate the bootstrap confidence values from 1000 replicates.

Discussion

In the F₂ population, we detected 17 QTLs for anthocyanin pigmentation on chromosomes 1–3, 6, and 10 (Table 1). For 6 out of 10 organs, at least 2 QTLs were found, one on chromosome 6 (92.7–107.8 cM) and the other(s) on other chromosomes. These 6 combinations confirmed that anthocyanin pigmentation is controlled by the interaction of two different members, R/B and C1/Pl (Table 1).

Four QTLs (qACspk2, qAClsh2, qACcul2, and qACbl2) detected on chromosome 2 (48.9–61.6 cM) were located in bins 2.02–2.04. Out of 130 genes listed in these bins, only b1 is associated with anthocyanin synthesis (Table 1); dominant B1 alleles determine anthocyanin synthesis in most aboveground plant parts (Chandler et al. 1989). Two major dominant B1 alleles, B-I and B-Peru, have been reported; B-I determines pigmentation of spikelet, leaf sheath, culm, and bract leaf (Chandler et al. 1989, Radicella et al. 1992), whereas B-Peru determines pale color in these organs but strong purple aleurone as well as a transcriptional regulator of R1 gene encoding bHLH protein. Although not significant, a QTL for anthocyanin pigmentation in both cob and grain was detected at a region that included R1 on chromosome 10 (Supplemental Fig. 1). The significant QTLs (qACant10, qACsil10, qACmlb10) on chromosome 10 (73.4–81.5 cM) were located in bin 10.06. Among 51 genes listed in this bin, 10 are associated with anthocyanin synthesis (Table 1) and are related to the R1 gene members of which occur as at least two tightly linked genes (Eggleston et al. 1995, Ludwig and Wessler 1990). Four R1 gene family haplotypes—R-g, R-st, R-sc and lc—result in colored seed, but our QTLs coincided with colorless anthers (Petroni et al. 2000, Tonelli et al. 1994). The R1 allele in the ‘JC072A’ haplotype may be any one of these four alleles.

Purple corn has a high content of anthocyanins, which are important natural colorants. Whole cobs contain about 55% of the anthocyanins and whole kernels contain about 45% (Cevallos-Casals and Cisneros-Zevallos 2003). Purple cobs have the highest contents of monomeric and acylated anthocyanins, which are considered natural food colorants. Only the QTL region on chromosome 6 (92.7–102.7 cM) located in bin 6.04 could result in cob and grain pigmentation. Among 40 genes listed in bin 6.04, pl1 and sm1 are functional genes for anthocyanin pigmentation (Table 1), but sm1 does not confer cob and grain pigmentation (McMullen et al. 2004). Several alleles of pl (pl1) genes have been reported. The light-independent dominant allele Pl-Rhoades (Pl-Rh) contains a transposable CACTA element in the promoter and results in anthocyanin-rich seedlings with dark purple husks, culms, sheaths, tassels, and cobs (Cone et al. 1993b). A cultivar with unpigmented seedlings derived from McClintock’s stock (McC) and inbred line Tx303 respectively carry the light-dependent recessive alleles Pl-McC and Pl-Tx303. The expression of the Pl-bol3 allele is partially light independent, and Pl-bol3 is active mainly at the juvenile phase (Pilu et al. 2003). The results of Pl marker mapping and analysis using two F₃ populations confirmed that the Pl allele of ‘JC072A’ (Pl1-JC072A) confers the anthocyanin pigmentation of cobs (Fig. 1, Supplemental Fig. 4). Sequence analysis of Pl genes showed that Pl1-JC072A is similar to the functional allele Pl-bol3 (Fig. 3, Supplemental Fig. 3). Therefore, Pl1 is the most likely candidate gene responsible for anthocyanin production in cobs.

Although cobs with high anthocyanin content are useful for anthocyanin production, they tend to be injured by mold, which hampers harvesting F₁ hybrid seeds. Each of the parents of an F₁ hybrid with a different gene of two genes could sustain seed production of F₁ hybrids. Our markers designed for the Pl1 genes will be useful for the breeding of F₁ parents with anthocyanin coloration in cobs.