Identification and gene duplication of NPF genes in cassava
To identify the entire NPF genes in cassava, we used a HMM profile (PTR2, PF00854) to search NPF proteins. Finally, we identified a total of 72 NPF genes in cassava genome. In order to clarify the phylogenetic relationship of NPF genes among Arabidopsis, cassava and rice, we performed multiple sequence alignment of NPF sequences using Clustal Omega, and a phylogenetic tree was constructed by IQ-TREE (Fig. 1a). According to the similarity of NPF sequences and the classification of members in Arabidopsis, the cassava NPF family was obviously divided into eight subfamilies. Among the cassava NPF gene family, NPF5 subfamily contained the most abundant NPF genes (20 members), followed by NPF2 and NPF4 (12 members of each subfamily), NPF6 (7 members), NPF3 (6 members), NPF1, NPF7 and NPF8 (5 members of each subfamily) (Fig. 1a; Supplementary Table S1-3).
Chromosomal localization analysis revealed that 72 MeNPF genes were widely distributed on 18 chromosomes, but not evenly distributed in the cassava genome (Fig. S1). Chromosome 1 (Chr1) had the highest density of NPF genes with 8 members, while Chr2, Chr3 and Chr12 only contained one gene, respectively. Due to the incomplete assembly and annotation of the current version of cassava genome, MeNPF5.14 was temporarily located at Scaffold01657. In addition, it was also found that the density of MeNPF genes at both ends of the chromosome was higher than that in the middle region.
Gene families are expanded mainly through whole genome duplication (WGD), tandem duplication (TD) and segmental duplication (SD)(Panchy et al. 2016). Collinearity analysis revealed that there were collinear relationships between some NPF genes. In this study, we found 21 pairs of segmental duplicated genes which were distributed on different chromosomes (Fig. 1b). There were four pairs of tandem duplicated genes. The above results indicate that segmental duplication events participate in the expansion of NPF gene family in cassava.
Expression patterns of MeNPF genes under different nitrogen treatments
To determine the expression of MeNPF genes under different nitrogen conditions, we explored the expression patterns of MeNPF gene family based on RNA-seq data. As shown in Fig. 2, some MeNPF genes, such as MeNPF1.1, MeNPF2.5, MeNPF2.10, MeNPF5.18 and MeNPF8.5, were highly expressed in all three tissues, regardless of exogenous NO3− addition. While MeNPF1.2, MeNPF1.3, MeNPF1.4, MeNPF1.5, MeNPF2.2, MeNPF2.3, MeNPF2.4, MeNPF2.8, MeNPF2.12, MeNPF3.6, MeNPF4.1, MeNPF4.2, MeNPF4.8, MeNPF4.10, MeNPF4.11, MeNPF5.13, MeNPF6.4 MeNPF7.4 and MeNPF8.3 was not detected in all tissues under different nitrogen conditions. Some MeNPF genes showed preferred expression in a specific tissue at different expression levels, including MeNPF2.1, MeNPF4.4, MeNPF5.11, MeNPF6.2, MeNPF7.2, MeNPF7.3, MeNPF7.5. Compared to stems and leaves, MeNPF2.1, MeNPF4.4, MeNPF4.7, MeNPF7.3 and MeNPF7.5 genes were preferentially expressed in roots. Despite this, we still found two genes, MeNPF5.4 and MeNPF6.2, were preferentially and highly expressed in stems. MeNPF5.4 was not induced by nitrogen treatment, whereas MeNPF6.2 was up-regulated in high nitrogen treatment, suggesting these two genes might have diverse roles in nitrogen transportation and remobilization.
MeNPF5.4 and MeNPF6.2 are localized in the plasma membrane
To further characterize the function of MeNPF5.4 and MeNPF6.2, we investigated the subcellular localization of MeNPF5.4 and MeNPF6.2 in rice mesophyll protoplast using the eYFP fusion constructs driven by the 35S promoter. The signal of eYFP alone as a control was distributed throughout the protoplast, whereas both of the MeNPF5.4-eYFP and MeNPF6.2-eYFP fusion protein signals were observed in the plasma membrane, a pattern similar to the membrane-localized protein maker ZmCDPK7-mCherry driven by 35S promoter (Fig. 3). These results indicated that MeNPF5.4 and MeNPF6.2 proteins functioned in the plasma membrane.
MeNPF6.2 mediated NO3− uptake in transgenic rice
To evaluate the potential value of MeNPF5.4 and MeNPF6.2 in crop genetic improvement, we respectively introduced MeNPF5.4 and MeNPF6.2 into rice (Nipponbare) to produce the MeNPF5.4 and MeNPF6.2 overexpression (OE) lines. From the rice transgenic lines, we selected three pure homozygous lines with variable expressions for MeNPF5.4 and MeNPF6.2, respectively (Fig. 4a-b). As shown in Fig. 4a-b, both of MeNPF5.4-OE and MeNPF6.2-OE showed significantly higher expression compared with the wild type (WT, Nipponbare). To test the effect of MeNPF5.4 and MeNPF6.2 on nitrogen utilization, we investigated the NO3− flux in wild-type and transgenic rice roots using NMT. As shown in Fig. 4c, we observed that the NO3− flux differed significantly between WT and transgenic lines. MeNPF6.2-overexpressing line exhibited more significant NO3− influxes, whereas MeNPF5.4-overexpressing line exhibited a marked decrease in NO3− effluxes compared with that of the control (Fig. 4c). These findings indicated that both of MeNPF5.4 and MeNPF6.2 can enhance nitrogen utilization by decreasing nitrogen effluxes and increasing nitrogen influxes in rice, respectively.
Overexpressions of MeNPF6.2 and MeNPF5.4 impact rice physiological characteristics and grain yield
In order to investigate the phenotypic and physiological impacts of transgenic rice, we grow plants hydroponically for two weeks under NN and LN conditions. We measured the plant height, root and shoot weight, chlorophyll contents and the activities of the antioxidant enzymes. As shown in Fig. 5, plant height of MeNPF6.2-OE was lower compared with WT under NN (Fig. 5a). The shoot to root fresh weight ratio was higher in MeNPF6.2-OE lines under LN (Fig. 5b). Both shoot and root biomass showed no significant difference between wild type and MeNPF5.4-OE lines (Fig. S3). Total root length and the number of root tips were decreased in transgenic plants (Fig. S3). The POD activity of MeNPF5.4-OE and MeNPF6.2-OE plants was higher than wild type under both N conditions (Fig. 5c). Moreover, MeNPF6.2-OE plants displayed higher CAT activity (Fig. 5d). There was no significant difference in plant nitrogen content between WT and OE lines under NN conditions, whereas MeNPF6.2-OE line exhibited higher nitrogen accumulation (Fig. 5e). The chlorophyll-b content of OE lines was lower under NN condition, while the OE lines showed higher chlorophyll-b content compared with WT under LN condition (Fig. S3). The OE lines showed higher total chlorophyll content than WT under LN condition (Fig. S3).
To investigate the role of MeNPF5.4 and MeNPF6.2 in rice NUE, we examined agronomic traits related to yield for WT and transgenic lines. Compared to the wild type plants, MeNPF5.4-OE plants had significant increase (P < 0.01) in grain length and thousand-grain weight, whereas no significant differences were observed between wild type and the MeNPF6.2 overexpressing lines (Fig. 6a-d). As shown in Fig. 6g and Fig. 6h, the grain number and grain weight per panicle of MeNPF5.4-OE plants were significantly increased relative to that of the wild type. In addition, the protein and nitrogen accumulation of grain were increased in MeNPF6.2-OE plants (Fig. 6i-j). Overall, our results indicated that elevated MeNPF5.4 expression level significantly enhanced the total grain number and weight per panicle.
Overexpressions of MeNPF5.4 and MeNPF6.2 improve salt stress tolerance
Previous studies have revealed that some nitrate transporters play critical roles in the response of plants to abiotic stress (Zhang et al. 2018). As described above, our RNA-seq identified DEGs that were related to oxidoreductase activity (Fig. S4). To examine whether overexpression of MeNPF5.4 and MeNPF6.2 can improve stress tolerance of transgenic rice, we treated 1-week-old seedlings with 150 mM NaCl and 20% PEG respectively. As shown in Fig. 7, when treated with NaCl, the OE lines mostly still remained green and were more vigorous than wild type under both N conditions. Strikingly, MeNPF6.2-OE showed better performance specifically (Fig. 7a-b). We counted the wilting degree of the latest leaf of wild type and transgenic plants, and found that the leaf wilting degree of transgenic plants was significantly lower than that of the wild type, especially MeNPF6.2-OE plants (Fig. 7d). Compared with NN, we also found that transgenic lines were more tolerant to salt stress under LN. There were no significant differences between WT and OE lines under PEG treatment (data not shown). These findings suggested that MeNPF5.4 and MeNPF6.2 can enhance rice tolerance to salt stress. In particular, MeNPF5.4 coordinates both grain yield and salt stress.