The majority of japonica rice cultivars currently grown in China are of the erect panicle (EP) type, which shows greater yield potential than the drooping panicle type. However, EP cultivars require higher nitrogen (N) input to maintain an acceptably high grain yield. Although a crucial quantitative trait locus (QTL), qPE9–1/DEP1, for the EP type has been identified, its genetic effects on grain yield remain controversial. ‘Wuyunjing 8′ (WYJ8) is a typical EP cultivar harboring the naturally mutated qpe9–1 allele. WYJ8_qPE9–1 is a transgenic line carrying the qPE9–1 allele in the WYJ8 background. To explore the genetic effects of the qPE9–1/DEP1 alleles, we conducted field experiments to compare the responses of WYJ8 and WYJ8_qPE9–1 to different N supply rates and planting densities in three consecutive years. WYJ8_qPE9–1 outperformed WYJ8 in grain yield and nitrogen use efficiency (NUE) under low N supply and low planting density, whereas WYJ8 excelled under excessive N supply and high planting density. The differential response in yield of the two genotypes to N supply was predominantly attributable to changes in panicle number and spikelet number. NUE-associated genes responded to N supply rates differently in the WYJ8_qPE9–1 and WYJ8, suggesting that a coordinated, complex network of genes is involved in NUE and subject to regulation by qPE9–1/qpe9–1. The results demonstrated that cultivars carrying qPE9–1 or qpe9–1 require different management practices to achieve yield potentials and high NUE. Adoption of qPE9–1 can enhance grain yield and NUE under low to moderate N supply and low planting density, whereas qpe9–1 can be utilized to exploit excessive N supply and high planting density.