Can Ghd7 and PCF3 improve rice yield and eating quality together?

A yield-quality tradeoff traced to two genes

Rice breeders have long faced a tradeoff: raising yield often drags eating quality down, and improving taste often costs yield. This study asks whether a single regulatory pathway can move both traits in the favorable direction, and identifies Ghd7 as the causal gene behind qAC7, a quantitative trait locus for grain amylose content.

The authors frame the problem in terms of grain composition rather than yield components alone. Amylose and total starch content, gel consistency, taste value, chalkiness, and grain protein content are measured as separate traits, each with its own unit and comparison baseline. That distinction matters because the paper's claim is not that one number improved, but that a specific combination of starch and protein changes shifted several quality measures together with yield.

The key terms are worth fixing early. Ghd7 is a known pleiotropic gene that the paper shows also affects grain quality; PCF3 is a plant-specific TCP-family transcription factor that Ghd7 activates; OsNF-YC4 is a gene the paper describes as positively regulating grain protein and negatively regulating starch. The proposed chain runs Ghd7 to PCF3 to OsNF-YC4, and the natural T/TA variant in PCF3 decides whether the last link is repressed.

Ghd7 shifts starch, protein, and taste value

The most direct evidence comes from near-isogenic lines that differ at Ghd7. Compared with NIL(KY131), NIL(DHX2) had 1.1% higher grain amylose content and 4.9% higher grain starch content, with gel consistency up 4 mm and taste value up 4.1. On the protein side, glutelin fell by 6 mg/g and total protein by 13.1 mg/g.

Complementation, knockout, and overexpression materials moved the same traits in the same directions, which is what supports calling Ghd7 a quality gene rather than a bystander. In the overexpression line Ghd7-OX, amylose rose to 22.4% versus 19.7% in ZH11, and taste value rose to 85.9 versus 82.6. The knockout line went the other way, with amylose at 15.94% and total protein at 115.9 mg/g versus 98.6 mg/g in NIP.

These are field-trial measurements with three randomized blocks and 20 plants per genotype, reported as means with standard errors. The paper also notes that Ghd7 raises yield mainly by increasing grains per panicle, consistent with earlier work, so the quality effect is not coming at the cost of the yield effect in these materials.

Two locations, same direction, different size

Because a single site can flatter a genotype, the authors grew the NILs in Wuchang, in Northeast China, and in Wuhan, in central China. In Wuchang, NIL(DHX2) gained 2.2% amylose, 1% total starch, 8.7 cm gel consistency, and 3.3 taste value over NIL(KY131), while protein fell 8.4 mg/g and chalkiness rate and area fell 89.2% and 95.5%. Yield per plant rose 17 g, with 33.7 more grains per panicle and heading delayed 11.5 days.

In Wuhan the direction held but the magnitudes differed: amylose up 0.6%, starch up 4.9%, gel consistency up 5 cm, taste value up 5.3, protein down 9.1 mg/g, chalkiness rate down 49.8% and area down 87.9%. Yield per plant rose 13.4 g with 58.8 more grains per panicle and heading 20.7 days later.

The authors attribute the gap between sites to photoperiod and temperature differences and to genotype-by-environment interaction, and they present it as an observation rather than a tested mechanism. For readers, the practical point is that the quality gain reproduced across two environments, but the size of the yield and heading-date change is environment-dependent, so a single-location number should not be treated as a fixed expectation.

Ghd7 activates PCF3 and dominates the interaction

The regulatory link was built with several independent assays rather than one. A yeast one-hybrid mutagenesis assay showed the CCACC motif is required for Ghd7 to associate with the PCF3 genomic fragment; EMSA and ChIP-qPCR confirmed binding in vitro and in vivo; and a dual-luciferase reporter assay in rice protoplasts showed Ghd7 enhances PCF3 transcriptional activity. Expression data agreed: PCF3 was lower in lines lacking functional Ghd7 and higher in Ghd7 overexpression lines.

The genetic relationship is described as dominant epistasis, meaning the effect of PCF3 depends on Ghd7 status. In the NIL(KY131) background, where Ghd7 function is lost, knocking out PCF3 did not change grain quality phenotypes, and PCF3 overexpression caused no obvious change in grain thickness or 1000-grain weight. In the NIL(DHX2) background with functional Ghd7, PCF3 overexpression markedly decreased grain thickness and 1000-grain weight, and altered quality traits including chalkiness, gel consistency, and taste value.

The two-factor interaction assays showed non-parallel phenotypic trends between the Ghd7 and ghd7 backgrounds, which the authors read as a significant non-additive interaction. This is the paper's explanation for why stacking PCF3 alleles without functional Ghd7 would not deliver the expected gain.

The TA variant breaks PCF3 DNA binding

PCF3 was identified as a TCP-family transcription factor localized mainly to the nucleus and highly expressed in stem and panicle. DAP-seq found that the top ten candidate binding motifs all contained the 5-base sequence CCCAC, and the OsNF-YC4 gene region carries nine such motifs, seven in the promoter and two in the 3'UTR.

EMSA showed PCF3 protein bound the CCCAC motifs at all sites except P4, P5, and P8, and a dual-luciferase assay showed PCF3 significantly inhibited OsNF-YC4 transcriptional activity. DAP-qPCR detected in-planta interaction with six fragments carrying the motif. The paper's model figure states that the functional PCF3-T protein binds six CCCAC motifs of OsNF-YC4, five in the promoter and one in the 3'UTR, repressing its transcription.

The T/TA variant sits in the third exon. The PCF3-TA allele carries an inserted A base that causes a frameshift, disrupting protein structure and abolishing binding to the OsNF-YC4 promoter; EMSA showed no specific DNA-protein complexes for this allele. With OsNF-YC4 left highly expressed, protein accumulates and starch synthesis is inhibited, which the authors link to lower taste value and higher chalkiness.

The favorable allele was never selected

The paper's population analysis is what makes PCF3 unusual. The T allele frequency was 0.54 in both the Aus and Ind subgroups, only slightly above the 0.37 found in wild rice, and significantly lower in the Jap subgroup. The authors call this pattern typical of subgroup differentiation under neutral evolution.

Selective sweep analysis placed PCF3 in a gray region rather than in a segment under strong positive selection. The θπWild/θπgroup ratios showed no significant reduction relative to genome-wide background, with values around 5 for Admix, 40 for Aus, 12 for Ind, 15 for Jap, and 50 for VI, and no conspicuous troughs that would indicate diversity loss from artificial selection.

The authors offer three possible explanations for this neutral pattern: strong functional dependence of PCF3-T on Ghd7, mismatched selection priorities during early domestication, and linkage drag. They present these as candidate explanations, not as demonstrated causes. The breeding implication they draw is that useful variation can persist outside the loci that domestication favored.

Combining Ghd7-1/4 with PCF3-T

The combination test is the paper's most decision-relevant result. Combining Ghd7-1/4 with PCF3-T further enhanced both yield and eating quality, and the discussion quantifies the gain as a 14.92% increase in yield per plant, a 7.45 increase in taste value, and a 28.99% decrease in chalkiness rate. The authors describe this as the combination that realizes the cryptic allele's breeding potential.

The comparison that matters is between allele classes, not between genes in isolation. PCF3-T supports the synergistic outcome because it retains DNA binding; PCF3-TA does not, so pairing it with Ghd7 would not be expected to reproduce the gain. Likewise, PCF3 overexpression partially restored phenotypes lost with Ghd7 loss of function, reaching levels comparable to NIL(DHX2) for total grain starch, total protein, taste value, panicle number, effective panicle number, seed setting rate, and grain size, but the effect still depended on Ghd7 status.

The paper also reports that both KY131 and DHX2 carry the Wxb allele yet differ substantially in eating quality, and it points to Ghd7 as likely one of the primary causes. That framing keeps the claim specific: the proposed targets are Ghd7-1/4 plus PCF3-T in materials that already carry functional Ghd7, not PCF3-T alone in any background.

What the evidence supports and where it stops

The strongest supported claim is mechanistic and comparative: Ghd7 activates PCF3, PCF3-T represses OsNF-YC4, and the resulting shift from protein toward starch accompanies higher taste value and lower chalkiness in the tested NILs and transgenics. The yield gain is documented through grains per panicle and yield per plant in those same materials, with the caveat that heading date and plant height also changed, and the size of the change varied by location.

Several boundaries sit next to the results they qualify. The two-location trial covers specific japonica backgrounds and two environments, and the authors attribute the differences to photoperiod, temperature, and genotype-by-environment interaction without isolating them. The population-genetics evidence shows the absence of a selection signature at PCF3, but the explanations offered for that absence are hypotheses. The 14.92% yield and 7.45 taste-value figures come from the combination analysis described in the discussion, so they should be read as the paper's reported combination outcome rather than as a general performance guarantee.

The priority the paper establishes is therefore a pairing rule rather than a single-gene recommendation. PCF3-T is worth pursuing only where functional Ghd7 is present, and the allele to avoid is PCF3-TA, which fails to bind OsNF-YC4. Among the paper's variables, the one that most changes a reader's judgment is the PCF3 T/TA genotype: it separates a combination that improved yield and eating quality together from one that cannot.

Sources used in this answer