**3. Discussion**

SAM differentiates all organs, and initiates new branches/tillers for more panicles. Understanding the molecular mechanism regulating SAM is of great importance to improve N use efficiency in many crop species, especially in rice. There are so many transcriptomic researches in rice plant. However, surprisingly, none is dealing with this specific important tissue. NPB and 9311 (sister line of YD6) are the two representative *japonica* and *indica* rice varieties being deeply sequenced and several mutant collections are derived from them. In this study, we utilized the transcriptome and co-expression network to analyze the different response to N rate between *indica* and *japonica* rice (YD6 and NPB) at the early stage of tillering occurrence. This study also reveals that the hub genes, tiller genes, NAM genes, and N related genes in each module are playing specific roles.

#### *3.1. Reducing N Input to Low or Moderate Rate Is Still Good to Promote Enough Tillers*

*Indica* and *japonica* rice subspecies possess much differences in morphological, physiological and cultivation characteristics especially in some important agronomic traits such as tillering. Tiller feature of a specific variety is the product of its genetic background, cultivation practices and environmental conditions. Among the later factors, N availability acts a key role [25,26]. Tiller is the basis for achieving panicle number which plays as a major factor in yield determination. In actual production, excessive N is usually applied aiming to produce enough number of tillers in the early growth stages. Therefore, to reduce N without yield penalty lies much in achieving enough tiller number at a minimal N requirement. Our results suggest that, even for varieties like NPB and YD6 being very different in their sensitivity to N enrichment, they both consistently show that a mild N enrichment can enhance tillering; however, excessive N enrichment will not promote more tillers as wished, instead it even suppresses tillering for a short term effect. Yet, for a more N tolerant variety like YD6, a frequent mild N enrichment may be necessary to boost tillering. Therefore, for reducing N input to a paddy field, cutting N rate to avoid heavy topdressing and switching to a low to moderate N enrichment can serve the same effect on promoting tillers. Meanwhile, at a reduced N rate, it can significantly cut off N run off as well as raise N use efficiency [25].

Meanwhile, we found the genes related to N uptake, transport, and N metabolism mainly fall in the blue, brown and turquoise modules through co-expression network in this study (Figure S12). Among them, the brown module correlated positively with the dry weight accumulation. *OsNRT1* encodes a low-affinity nitrate transporter and belongs to the constitutive expression of the outermost layer of roots, epidermis and root hairs [27]. This gene is not only homologous to the *CHL1* (*AtNRT1*) gene in Arabidopsis, but also to the polypeptide transporter that widely present in plants, animals, fungi, bacteria etc. *OsNRT2.3a* plays an important role in the long-distance transport of nitrate from root to shoot in the low-nitrate supply condition [28]. *OsNRT2.3a* is a rice vascular-specific, NRT2 family high-affinity nitrate transporter. *OsNAR2.1* can interact with *OsNRT2.1*/*2.2* and *OsNRT2.3a* to promote nitrate uptake by rice roots at different nitrate supply levels [29]. *OsNRT2.3b* can enhance the buffering capacity of rice to pH status, increase the absorption of N, iron and phosphorus, improve the effective utilization of N, and is very important for plants to adapt to different forms of N sources [30]. In our study, though *OsNRT1*, *OsNRT2.3* and *OsNPF2.4* showed similar response pattern to N rate, *OsAMT1.2* showed an even different response pattern to N rate between varieties. In previous transcriptome and co-expression network analysis, *NRT* and *AMT* play an important role in N uptake and utilization efficiency in *Brassica juncea* cultivars and N transporters regulate some aspects (shoot or leaf) of the coordination of N and C metabolism in *Arabidopsis* [31,32]. Interestingly, these genes have a close positive regulation of dry weight of stem and leaf in this study, which suggests that these genes may share common roles in regulating N transportation and biomass. These results indicate that these genes have a promoting effect on the growth and development of the above-ground organs. Although it does not belong to hub genes, it plays a role in the transport and absorption of N under the control of upstream hub genes.
