Genetic improvement of grain yield is important for ensuring food security. Many important agronomic traits, including yield, show continuous phenotypic variation . Grain yield is a complex quantitative trait, and is determined by several components including kernel number and kernel weight . To date, more than 185 quantitative trait loci (QTL) underlying yield components including kernel row number (KRN), kernel number per row (KNPR), and hundred kernel weight (HKW) have been identified across the maize genome using various mapping populations (2012 November update to Gramene database), which has improved our understanding of the genetic basis of maize yield. Therefore, grain size and kernel weight are important targets for artificial selection for high grain yield. For example, QTL associated with GS3  and GW5/SW5 [4, 5] have been selected for grain size in rice. Also, QTL associated with TaGW2-6A Hap-6A-A for grain size  and TaCKX6-D1  for grain weight have been identified and artificially selected from within Chinese wheat germplasm. Markers derived from QTL controlling similar components of grain yield in maize will be very useful for marker-assisted selection.
Several genes that control kernel weight or kernel number have been cloned in crops, and some of these genes were found to be involved in carbon and nitrogen metabolism [8–10], protein degradation [11, 12], and hormone metabolism [13, 14]. All of these processes affect the production and export of C- and N-assimilates to the seed, thereby increasing crop yield. C- and N-metabolism are essential to all processes in plants, including reproductive development, but are especially so during grain filling . Among the genes involved in carbon metabolism, rice grain incomplete filling 1, which functions similarly to the invertase encoded by miniature 1 in maize , also encodes a cell-wall invertase required for carbon partitioning during early grain-filling . A gene involved in nitrogen metabolism in rice, OsARG, encodes an arginine hydrolase that, when overexpressed, increases grain number per plant under nitrogen-limited conditions, due to increased nitrogen remobilization at the reproductive stage .
In addition, degradation of proteins via the ubiquitin/proteasome pathway negatively regulates cell division and grain yield. Another gene, Grain Weight 2 (GW2), which encodes a RING-type E3 ubiquitin ligase, functions in ubiquitin-mediated degradation, and the loss of GW2 function can enhance rice grain width, weight, and yield . Furthermore, ZmGW2-CHR4, which functions in the same manner as GW2, is also associated with kernel width and kernel weight in maize . Although they have potentially positive impacts on yield, alleles at these loci also should be monitored to avoid possible negative effects of introduced germplasm on yield.
As to the influence of plant hormones, brassinosteroids (BRs) and cytokinins (CKs) are useful for controlling grain yield in crops. BRs stimulate the transport of sucrose and other sugars to the endosperm and embryo. Expression of the maize, rice, or Arabidopsis thaliana C-22 hydroxylase that is involved in BR synthesis in stems, leaves, and roots had clear effects on seed weight . In rice, the accumulation of CK in the inflorescence meristems increases the number of grains per panicle. The rice gene Gn1a encodes an oxidase/dehydrogenase (OsCKX2) that functions in the degradation of CKs. An 11-bp deletion in its coding region created a premature stop codon that reduced expression of OsCKX2 and resulted in increased grain number . In wheat, TaCKX6-D1, an ortholog of rice OsCKX2, was significantly associated with 1000-grain weight by linkage mapping and association analysis . Thus, the CK regulatory pathway likely plays a crucial role in grain yield.
There is also evidence for a role of CKs in balancing the source and sink relationship . Accordingly, exogenous applications of CKs increased seed set and yield stability under heat stress in maize . CKs are co-regulated by isopentenyl transferase (IPT) and oxidase/dehydrogenase in plants. The discovery of biosynthetic enzymes IPT1 through IPT10 [17–19] and degradative enzymes CKX1 through CKX12 [18, 20–23] has led to a better understanding of the role of CKs in maize kernel development. Among the IPTs, IPT1 and IPT10 are highly abundant and are constitutively expressed in all organs, but other IPT transcripts show distinct spatial and temporal patterns of expression . ZmIPT2 is specifically expressed in the pedicel, endosperm, and embryo [18, 19]. ZmIPT2 consists of an intronless 966 bp coding region for dimethylallyl diphosphate (DMAPP): adenosine triphosphate (ATP)/adenosine diphosphate (ADP) IPT, which preferred ATP and ADP over adenosine monophosphate (AMP) as substrates for IPT activity .
To date, the allelic diversity of ZmIPT2 and the most favorable allele(s) affecting kernel weight have not yet been reported for maize. The objectives of this study were to (1) examine the sequence diversity of ZmIPT2 in 175 Chinese maize inbred lines; (2) test associations between nucleotide polymorphisms in ZmIPT2 and various yield components including KRN, KNPR, and HKW, and identify further favorable allele(s) for grain yield components; (3) characterize the IPT activity of the protein products of different alleles in vitro; and (4) investigate selection for favorable ZmIPT2 alleles for kernel weight during maize breeding in China.