**Transcriptomic and Co-Expression Network Profiling of Shoot Apical Meristem Reveal Contrasting Response to Nitrogen Rate between** *Indica* **and** *Japonica* **Rice Subspecies**

**Xiaoxiang Zhang 1,2,**†**, Juan Zhou 1,**†**, Niansheng Huang 2, Lanjing Mo 1, Minjia Lv 1, Yingbo Gao 1, Chen Chen 3, Shuangyi Yin 1, Jing Ju 4, Guichun Dong 1, Yong Zhou 1, Zefeng Yang 1, Aihong Li 2, Yulong Wang 1, Jianye Huang 1,\* and Youli Yao 1,\***


Received: 6 October 2019; Accepted: 23 November 2019; Published: 25 November 2019

**Abstract:** Reducing nitrogen (N) input is a key measure to achieve a sustainable rice production in China, especially in Jiangsu Province. Tiller is the basis for achieving panicle number that plays as a major factor in the yield determination. In actual production, excessive N is often applied in order to produce enough tillers in the early stages. Understanding how N regulates tillering in rice plants is critical to generate an integrative management to reduce N use and reaching tiller number target. Aiming at this objective, we utilized RNA sequencing and weighted gene co-expression network analysis (WGCNA) to compare the transcriptomes surrounding the shoot apical meristem of *indica* (Yangdao6, YD6) and *japonica* (Nipponbare, NPB) rice subspecies. Our results showed that N rate influenced tiller number in a different pattern between the two varieties, with NPB being more sensitive to N enrichment, and YD6 being more tolerant to high N rate. Tiller number was positively related to N content in leaf, culm and root tissue, but negatively related to the soluble carbohydrate content, regardless of variety. Transcriptomic comparisons revealed that for YD6 when N rate enrichment from low (LN) to medium (MN), it caused 115 DEGs (LN vs. MN), from MN to high level (HN) triggered 162 DEGs (MN vs. HN), but direct comparison of low with high N rate showed a 511 DEGs (LN vs. HN). These numbers of DEG in NPB were 87 (LN vs. MN), 40 (MN vs. HN), and 148 (LN vs. HN). These differences indicate that continual N enrichment led to a bumpy change at the transcription level. For the reported sixty-five genes which affect tillering, thirty-six showed decent expression in SAM at tiller starting phase, among them only nineteen being significantly influenced by N level, and two genes showed significant interaction between N rate and variety. Gene ontology analysis revealed that the majority of the common DEGs are involved in general stress responses, stimulus responses, and hormonal signaling process. WGCNA network identified twenty-two co-expressing gene modules and ten candidate hubgenes for each module.

Several genes associated with tillering and N rate fall on the related modules. These indicate that there are more genes participating in tillering regulation in response to N enrichment.

**Keywords:** shoot apical meristem; transcriptomic analysis; co-expression network; tiller; nitrogen rate; rice (*Oryza sativa* L.)

#### **1. Introduction**

As one of the three most important cereal crops cultivated for thousands of years, rice (*Oryza sativa* L.) provides staple foods for nearly half of the world's population. Keeping its production apace with the increasing demand is critical to food security. The most gain in rice yield is attributable to increased N fertilizer addition to the paddy field. N is an essential macronutrient for completing rice plant growth and development, therefore, usually a limit to its production [1]. Inarguably, N is the most effective fertilizer in promoting crop growth and increasing crop yield [1]. However, the magic effects of N fertilization deceptively lead to excessive application, which gravely contributes to more direct ammonium gas loss, N run-off, water eutrophication, nitrous oxide greenhouse gas emissions and soil acidification [1–3]. Consequentially, excessive N fertilization has become a fundamental environmental issue and a health problem. Improving N use efficiency is believed to be the ultimate solution to mitigate these problems. Ideal N application rate and proper timing are the keys to reaching a balance of yield gain and N use efficiency [2,3]. *Indica* and *japonica* rice are two major subspecies in Asia, and reportedly being different in response to N enrichment [4,5].

Rice tiller begins at the leaf sheath auxiliary of the bottom nodes of a host culm, close to shoot apical meristem (SAM) at the early growth stage. Usually, rice tiller first appears when the 4th true leaf emerges. Tiller number at 8-9th week after germination essentially represents a variety's tiller production feature [6]. Tillers provide more opportunities to the development of more panicles. Among the agronomic yield traits of rice, panicle number is the one that being determined at the earliest stage. More importantly, panicle number largely associates with the spikelet number and filled grain number per panicle. Therefore, reaching a suitable number of tillers timely is a benchmark in setting the cornerstone for achieving an ideal grain yield in rice production.

Meanwhile, as a monocot fully sequenced species, rice provides a great model to decode the molecular secrets of tiller regulation. Recent progress in rice molecular genetics reveals that more than 65 genes at different stages engage in tiller number regulation [7–14]. However, how these genes concert in tiller control remains to be an enigma. High throughput RNA sequencing (RNA-Seq) lends us an effective tool to discover the transcriptomic profile of every active gene [15]. Though much transcriptome data exists in rice, none targets at the tissues near the SAM region, especially in response to N enrichment.

In the present study, we validated the differential tillering responses to N rates of *indica* (Yangdao6, YD6) and *japonica* (Nipponbare, NPB) rice subspecies, performed RNA sequencing and weighted gene co-expression network analysis (WGCNA) to compare their transcriptomes in the surrounding tissues of SAM. The objectives of this study is to compare their differential transcriptomic responses to N rate, and to reveal the concerted molecular network in tiller regulation in rice. The results would expand our understanding of how complicated an important agronomic trait is controlled.

#### **2. Results**
