Abstract

   Large amounts of nitrogen fertilizers applied during lettuce (Lactuca
   sativa L.) production are lost due to leaching or volatilization,
   causing severe environmental pollution and increased costs of
   production. Developing lettuce varieties with high nitrogen use
   efficiency (NUE) is the eco-friendly solution to reduce nitrogen
   pollution. Hence, in-depth knowledge of nitrogen metabolism and
   assimilation genes and their regulation is critical for developing high
   NUE varieties. In this study, we performed comparative transcriptomic
   analysis of the cultivated lettuce (L. sativa L.) and its wild
   progenitor (L. serriola) under high and low nitrogen conditions. A
   total of 2,704 differentially expressed genes were identified. Key
   enriched biological processes included photosynthesis,
   oxidation–reduction process, chlorophyll biosynthetic process, and cell
   redox homeostasis. The transcription factors (TFs) belonging to the
   ethylene responsive factor family and basic helix-loop-helix family
   were among the top differentially expressed TFs. Using weighted gene
   co-expression network analysis we constructed nine co-expression
   modules. Among these, two modules were further investigated because of
   their significant association with total nitrogen content and
   photosynthetic efficiency of photosystem II. Three highly correlated
   clusters were identified which included hub genes for nitrogen
   metabolism, secondary metabolites, and carbon assimilation, and were
   regulated by cluster specific TFs. We found that the expression of
   nitrogen transportation and assimilation genes varied significantly
   between the two lettuce species thereby providing the opportunity of
   introgressing wild alleles into the cultivated germplasm for developing
   lettuce cultivars with more efficient use of nitrogen.

   Subject terms: Plant sciences, Plant genetics

Introduction

   Nitrogen (N) is one of the most essential plant nutrients required by
   the plants for growth and development. N is a key constituent of
   nucleic acids, proteins, amino acids, many cofactors, and secondary
   metabolites. It also plays a pivotal role in regulation of several
   biological process like carbon metabolism, amino acid metabolism and
   protein synthesis^[32]1. N starvation impacts several major
   physiological and biological processes in plants affecting growth and
   development^[33]2, root architecture^[34]3, lignin content^[35]4,
   anthocyanin, phosphorus, and potassium content^[36]5, chlorophyll
   synthesis^[37]6, senescence, photosynthesis^[38]7, and CO[2]
   assimilation^[39]1. Therefore, to increase crop yields, application of
   N fertilizers has dramatically increased in recent decades^[40]2.
   However, up to half of the applied N remains unused causing water
   eutrophication and air pollution by release of nitrous oxide thereby
   contributing to global climate change^[41]8,[42]9. In fields, nitrate
   (NO[3]^-) is the major source of N for plants which is reduced to
   ammonium (NH[4]^+) ions and is incorporated into amino acids that are
   vital for plant growth^[43]10. Nitrate is also involved in signaling
   pathways regulating expression of genes in plant system^[44]11,[45]12.
   Several genes control N metabolism pathways (N-uptake, N-translocation,
   N-assimilation and N-remobilization) in plants. Major genes involved in
   N metabolism are nitrate transporters (nitrate, ammonium), reductase
   (nitrite, nitrate), synthase (glutamine, glutamate), aminotransferase
   (aspartate, alanine), and glutamate dehydrogenase and the expression of
   these genes varies under different N regimes^[46]13.

   Lettuce, Lactuca sativa L., cultivation is resource intensive; it is
   reliant on high amounts of N fertilizer application for its
   productivity and quality^[47]14. It is a shallow rooted crop, and
   therefore requires frequent irrigation which often leads to leaching of
   nitrate into groundwater. Over 75% of the total lettuce produced in USA
   comes from California, predominantly from the Salinas Valley, where the
   groundwater in agricultural region is adversely impacted due to nitrate
   leaching^[48]15. The amount of nitrate in a majority of wells in the
   Salinas Valley area exceeds the federal drinking water standard of
   10 mg/L^[49]16,[50]17. Therefore, it is important to determine methods
   to reduce N fertilizer application and improve nitrogen use efficiency
   (NUE) of lettuce without affecting its yield and quality.

   NUE is a complex trait that involves several molecular, biochemical,
   and physiological processes and have two major components, N uptake
   efficiency (NUpE) and N utilization efficiency (NUtE)^[51]18. Improved
   NUE often leads to increase in above-ground biomass, seed production,
   protein content and overall economic yields^[52]19. Currently, research
   efforts are underway to improve NUE through a variety of approaches
   including mode and timing of fertilizer application^[53]20, crop
   rotation^[54]21, and management practices^[55]22,[56]23. From the
   breeding perspective, two approaches to improve NUE may be undertaken.
   The first approach involves traditional breeding strategies in
   conjunction with quantitative trail loci (QTL) mapping and marker
   assisted selection. Several studies were conducted to improve NUE in
   lettuce using this approach. For example, genetic variation in NUpE was
   explored among lettuce cultivars which revealed high diversity in
   lettuce germplasm for root growth, resource capture and
   NUE^[57]24,[58]25. Molecular breeding efforts to improve NUE in lettuce
   by selecting for robustness trait associated plants with deeper root
   systems and efficient resource capture resulted in identification of
   QTLs for below ground traits and shoot traits^[59]26.

   The second strategy to improve NUE involves a targeted approach of
   identifying specific genes associated with N metabolism pathways and
   developing engineered plants with modified gene expression. Several
   genome-wide transcriptomics have been conducted to identify N
   metabolism genes and investigate responses to N stress in various
   plants such as Arabidopsis^[60]27–[61]29, rice^[62]30–[63]33,
   maize^[64]34–[65]36, wheat^[66]37,[67]38, brassica^[68]10,[69]39,
   spinach^[70]40, potato^[71]41. However, the efforts to identify
   candidate genes associated with N metabolism and transcriptomic
   response of different lettuce species to N stress is lacking, though
   QTLs related to NUE has been recently mapped^[72]42.

   To tackle this issue, here we performed an RNA-sequencing based
   comparative transcriptomic analysis of cultivated lettuce L. sativa cv
   Salinas and its wild progenitor L. serriola acc. US96UC23^[73]43 grown
   under high and low nitrogen conditions. L. sativa is a widely
   cultivated, high biomass producing species but is extremely susceptible
   to water and nutrient stress conditions. In contrast, the wild
   progenitor L. serriola L., often referred to as ‘prickly lettuce’, is
   more drought tolerant with an ability to survive under adverse
   conditions^[74]44 and is considered as invasive weed in several parts
   of Australia and the US ([75]http://www.weedscience.org). We
   hypothesized that the L. serriola may carry genes or allele variants
   that enable it to survive under water and nutrient stress conditions
   while the L. sativa carry genes for efficient nutrient metabolism under
   favorable conditions. Therefore, the objective of this study was to
   identify key N metabolism-related genes in the two lettuce species
   under favorable and N stress conditions and further evaluate
   co-expressing genes modules affecting important traits such as
   photosynthetic efficiency of photosystem-II (PSII), chlorophyll content
   and cell detoxification. The genes identified in this study can be used
   for improving NUE of cultivated lettuce varieties either by
   introgression breeding or by genetic engineering.

Results

Phenotypic measurements of nitrogen related traits

   N supply had significant effects on chlorophyll fluorescence and
   vegetative indices (Fig. [76]1) indicating changes in the chlorophyll
   content, chlorophyll composition and photosynthetic efficiency due to
   limited availability of N. Overall, the effects of low N (LN)
   conditions were more prominent in Salinas than UC. Under LN conditions
   the total N accumulation and the photosynthetic efficiency of PSII
   (QY_max) was reduced by 67% and 16% respectively in Salinas compared to
   63% and 5% reduction in the UC genotype. The percent reduction in SPAD
   was 21.3% in the UC but was significantly lower in Salinas at 14.1%.
   The non-photochemical quenching (NPQ) increased in both genotypes under
   N limited conditions (Fig. [77]1).

Figure 1.

   [78]Figure 1
   [79]Open in a new tab

   Phenotypic reaction of two lettuce genotypes to high nitrogen (HN) and
   low nitrogen (LN) treatments.

Transcriptome sequencing and differential gene expression analysis under
nitrogen stress

   A total of 12 libraries from three biological replicates per genotype
   per treatment were developed. After trimming adapters and low-quality
   reads, a total of 300.84 million clean reads (150 bp long) were
   generated with an average of 25 million reads per sample. A total of
   273.21 (90.81%) million reads were mapped on to the lettuce reference
   genome (V8) using STAR aligner software with an average of 22.76
   million reads per sample (Table [80]1). There was a stronger response
   to N starvation in Salinas compared to UC. A total of 1,999 DEGs among
   high N (HN) and LN treatments were identified in the Salinas genotype
   of which 828 (41.42%) genes were upregulated and 1,171 (58.58%) genes
   were downregulated under LN stress condition (Fig. [81]2, Table
   [82]S1). In UC genotypes, a total of 1,235 differentially expressed
   genes (DEGs) were identified of which 514 (41.62%) were upregulated
   while 721 (58.38%) genes were downregulated in response to N stress
   (Table [83]S1). These differences in the number of DEGs suggest that
   the response to N stress varied between the wild and cultivated
   species. To validate these results, we performed RT-PCR analysis in
   both genotypes using key genes involved in N metabolism (Table [84]S2).
   Results from RT-PCR show that the expression trends of the selected
   genes were in accordance with the expression detected by the RNA-seq
   analysis.

Table 1.

   Summary of RNA-Seq performed on two lettuce genotypes under high
   nitrogen (HN) and low nitrogen (LN) treatments.
   Genotype Treatment Replicate Library name Total clean Reads Mapped
   reads Mapped reads (%) Multiple Loci reads Multiple Loci reads (%)
   Unmapped reads Unmapped reads (%)
   Salinas HN R1 Salinas_HN1 25,413,058 22,885,448 90.05 1,056,980 4.16
   1,470,630 5.79
   Salinas HN R2 Salinas_HN2 26,614,697 23,869,149 89.68 1,284,629 4.83
   1,460,919 5.49
   Salinas HN R3 Salinas_HN3 27,246,323 24,925,925 91.48 1,123,729 4.12
   1,196,669 4.39
   US96UC23 HN R1 UC_HN1 24,527,750 22,232,436 90.64 934,275 3.81
   1,361,039 5.55
   US96UC23 HN R2 UC_HN2 21,366,114 19,456,364 91.06 728,930 3.41
   1,180,820 5.53
   US96UC23 HN R3 UC_HN3 27,134,208 24,689,908 90.99 1,012,409 3.73
   1,431,891 5.28
   Salinas LN R1 Salinas_LN1 28,084,512 25,711,442 91.55 822,793 2.93
   1,550,277 5.52
   Salinas LN R2 Salinas_LN2 21,345,239 19,213,005 90.01 826,202 3.87
   1,306,032 6.12
   Salinas LN R3 Salinas_LN3 21,576,629 19,867,359 92.08 556,433 2.58
   1,152,837 5.34
   US96UC23 LN R1 UC_LN1 27,719,243 25,049,883 90.37 941,785 3.4 1,727,575
   6.23
   US96UC23 LN R2 UC_LN2 23,762,434 21,834,284 91.89 763,931 3.21
   1,164,219 4.9
   US96UC23 LN R3 UC_LN3 26,052,847 23,469,949 90.09 907,243 3.48
   1,675,655 6.43
   [85]Open in a new tab

Figure 2.

   Figure 2
   [86]Open in a new tab

   Venn diagram showing the number of up- and down regulated genes in
   Salinas and UC in response to nitrogen stress.

Functional classification of N-deficiency-responsive genes

   GO enrichment analysis of the annotated DEGs revealed several enriched
   biological processes, molecular functions and cellular components
   (Fig. [87]3). The key enriched biological processes include
   photosynthesis (GO:0,015,979), oxidation–reduction process
   (GO:0,055,114), response to water deprivation (GO:0,009,414),
   chlorophyll biosynthetic process (GO:0,015,995), response to cytokinin
   (GO:0,009,735) and cell redox homeostasis (GO:0,045,454). The key
   molecular functions enriched included oxidoreductase activity
   (GO:0,016,491), amino acid transmembrane transporter activity
   (GO:0,015,171), chlorophyll binding (GO:0,016,168), and protein binding
   (GO:0,005,515). The enriched cellular components included chloroplast
   stroma (GO:0,009,570), chloroplast envelope (GO:0,009,941), chloroplast
   (GO:0,009,507), plasma membrane (GO:0,005,886), and cytosol
   (GO:0,005,829). Some of the key KEGG pathways enriched are metabolic
   pathways (ath01100), N metabolism (ath00910), carbon metabolism
   (ath01200) and photosynthesis (ath00195). These results indicate that
   genes in these pathways play important role in host plant response to N
   stress.

Figure 3.

   [88]Figure 3
   [89]Open in a new tab

   Enriched GO terms according to biological processes (red bars),
   cellular components (green bars) and molecular functions (orange bars).
   The x-axis represents log2 of the gene counts.

Differentially expressing transcription factors under nitrogen stress

   Transcription factors (TFs) play key roles in plant adaptation to N
   stress by regulating plant signal transduction pathways. We identified
   total of 155 transcription factors (TFs) related genes that were
   differentially expressed in response to N stress. The most
   differentially expressed TFs (Fig. [90]4) belong to ethylene responsive
   factor (ERF) family with 24 (15%) DEGs followed by the TFs belonging to
   bHLH family with 17 (11%) DEGs. Other key TFs identified in this study
   are MYB (10%), bZIP (8%), and NAC (8%) (Fig. [91]3). The differential
   expression of TF under N stress indicate different mechanisms in the
   two genotypes under N stress.

Figure 4.

   [92]Figure 4
   [93]Open in a new tab

   Differentially expressed transcription factors identified by RNA-Seq
   analysis.

Photosynthesis and carbon assimilation related responses

   Photosynthetic efficiency of Salinas and UC was reduced due to N
   stress. The expression of 42 photosynthesis-related genes was down
   regulated in Salinas which included genes encoding components of the
   photosystem-II light harvesting (LHC) complex, photosystem-I reaction
   center (PSAE, PSAF) and electron carriers (DRT112, NDF4) (Fig. [94]5).
   Major photorespiration genes (Ls0_14580.1, Ls3_16080.1, Ls5_136860.1,
   Ls7_16480.1) and 18 genes related to the Calvin cycle were also
   repressed under N stress including 9 RuBisCO genes encoding for
   ribulose bisphosphate carboxylase small chain 2B family protein.
   Similarly, the relative expression of photosynthesis genes in UC were
   also repressed, however the number of genes affected by N stress were
   fewer compared to Salinas. The expression levels of five genes
   (Ls5_107040.1, Ls9_104661.1, Ls9_104700.1, Ls9_104780.3, Ls9_89920.1)
   encoding for photosystem-II light harvesting (LHC) complex, a gene
   (Ls1_35241.1) encoding for photosystem-I reaction center and a RuBisCO
   gene (Ls4_31161.1) was downregulated in response to nitrogen stress in
   the UC genotype. These results indicate that the UC genotypes is able
   to maintain photosynthesis under N stress compared to the Salinas
   genotype.

Figure 5.

   [95]Figure 5
   [96]Open in a new tab

   Mapman analysis of differentially expressed genes related to various
   metabolic processes in (a) Salinas, (b) UC genotypes. Red squares
   represent genes that were significantly down-regulated; blue squares
   represent genes that were significantly up-regulated.

Cell wall associated responses

   Effect of N stress on cell wall components was analyzed by evaluating a
   change in the relative expression of the cell wall precursor-related
   genes. Expression of 21 genes involved in biosynthesis of cell wall
   precursors was affected by N stress in Salinas genotype. Expression
   levels of only three genes (Ls8_38141.1, Ls3_10960.1, Ls4_96981.1) were
   upregulated while the expression of the remaining 18 genes was
   downregulated. In the UC genotype, expression of 22 cell wall
   precursor-related genes was affected. The relative expression of only
   one gene (Ls7_110780.6) involved in plastidial glycolytic pathway was
   upregulated in response to N stress (Fig. [97]5). We detected the
   expression of 13 cell wall precursor’s related genes involved in the
   glycolysis process, were downregulated in both genotypes. Expression of
   inositol oxygenase 1 (MIOX1) [[98]Q8L799] gene (Ls4_96981.1) was
   upregulated only in Salinas genotype. This gene is involved in the
   biosynthesis of UDP-glucuronic acid (UDP-GlcA) thereby providing
   nucleotide sugars for cell-wall polymers may be crucial for cell wall
   associated changes under N stress.

Secondary metabolism under nitrogen stress

   Based on transcriptome analysis, response to secondary metabolite
   production under N stress was found to be different between the two
   genotypes. In Salinas, 48 secondary metabolism genes showed
   differential expression including 25 genes for terpenes biosynthesis,
   16 genes coding for flavonoids, and 7 phenylpropanoids and phenolics
   biosynthesis genes (Fig. [99]5). Among the terpenes biosynthesis genes,
   the genes (Ls5_42681.1, Ls5_100341.2) involved in sesquiterpene (C15)
   biosynthesis and synthesis of tocopherol (vitamin E) that protect
   thylakoid membrane lipids from photooxidation were significantly
   downregulated (LFC > − 6.0) in Salinas genotype while the gene
   (Ls1_127360.1) involved in the biosynthesis of homoterpenes and a gene
   (Ls6_52200.1) involved in phytoene biosynthesis pathway, which is part
   of carotenoid biosynthesis, were significantly upregulated (LFC > 3.0)
   under N stress in Salinas genotype. Among the flavonoid genes, the
   expression of two detoxifying genes (Ls5_149321.1, Ls9_61600.2)
   involved in anthocyanin and protoanthocyanidin biosynthesis was
   significantly upregulated (LFC > 3.5) while the genes involved in
   phenylpropanoids and phenolics biosynthesis such as the
   Four-Coumarate:CoA ligase (4CL) encoding lettuce gene (Ls1_18940.1) and
   a gene (Ls4_118261.1) involved in the oxylipin biosynthetic process was
   the most significantly downregulated (LFC > − 7.0) gene in Salinas
   genotype.

   Similarly, in the UC genotype, 13 genes each for terpenes biosynthesis
   and flavonoids and seven genes involved in phenolics biosynthesis were
   differentially expressing in response to N limitation. The terpenes
   gene (Ls5_71000.1) involved in lignin degradation and detoxification of
   lignin-derived products was significantly downregulated (LFC > − 4.0)
   while the cyto-detoxifying gene (Ls9_61600.2) was the most
   significantly upregulated (LFC > 5.0) gene.

Variation in nitrogen assimilation-related genes in response to N stress

   The expression of 12 N assimilation genes were upregulated in both
   genotypes while the expression of six genes was repressed under N
   stress. There were seven genes with higher expression in UC. Expression
   of genes encoding nitrate transmembrane transporters (NRT3.1;
   Ls4_3961.1)) and glutamate dehydrogenase 2 (GDH2; Ls8_11521.1) was
   higher in the Salinas genotype compared to the UC genotype. Under high
   N conditions, the expression of cyanase (CYN; Ls1_75161.1) and
   glutamine synthetase 2 (GLN2; Ls5_77601.1) genes was greater in the
   Salinas parent while the expression of four genes involved in nitrate
   uptake and transportation displayed higher expression in UC genotype
   under high N conditions suggesting that UC is more efficient in nitrate
   uptake and transfer from stored pools to cytoplasm (Fig. [100]6).

Figure 6.

   Figure 6
   [101]Open in a new tab

   Differential expression of nitrogen metabolism genes Salinas and UC
   genotypes when grown at high nitrogen (HN) and low nitrogen (LN)
   treatments.

Gene co-expression and network analysis

   To get further insight into the N stress-related genes regulatory
   network of lettuce we used WGCNA to identify co-expression modules
   using expression profile of all genes having significant differential
   expression in Salinas and UC genotypes. The dissimilarity of the
   modules was set as 0.2, and 9 co-expression modules were generated by
   DisTOM based hierarchical clustering post dynamic tree cutting method
   (Fig. [102]7a, b). Weighted Pearson correlation coefficients calculated
   between eigengenes (modules) and traits (phenotypic measurements) are
   presented in the Fig. [103]8 as the module trait relationship. The
   light-green module was positively correlated with QY_max, NPQ, NDVI,
   SPAD and total N content with correlation coefficient ranging from 0.73
   to 0.89. The blue module was negatively correlated with all traits with
   correlation coefficient ranging from − 0.67 for CRI to -0.94 for
   chlorophyll content (SPAD). High correlation between the module
   membership (MM) and gene significance (GS) for total N content was
   observed in light-green module (r = 0.7, p-value = 3.8e^−29) and in the
   blue module (r = 0.68, p-value = 1.4e^−37) suggesting that these
   modules are suitable for identifying hub genes associated with N stress
   associated response in lettuce (Fig. [104]S2). The eigengene adjacency
   heatmap shows that N content was more adjacent to the light-green
   module than the blue module (Fig. [105]S3).

Figure 7.

   [106]Figure 7
   [107]Open in a new tab

   (a) Gene modules identified by WGCNA. Dendrogram was obtained by
   clustering dissimilarity with the modules represented by the colored
   block based on the topological overlap. A total of nine module (colored
   blocks) were identified where each module represents a set of highly
   connected genes. (b) Heatmap plot of the gene network. The heatmap
   depicts Topological Overlap Matrix (TOM) among all genes in the
   analysis. Light color represents overlap of highly expressed genes and
   the darker color represent low overlap of genes. Blocks of yellow
   colors along the diagonal are the modules.

Figure 8.

   [108]Figure 8
   [109]Open in a new tab

   Heat map of relationship between modules and phenotypic traits. The
   color indicated direction of correlation, values represent Pearson
   correlation coefficients, and P-values are in parenthesis.

   Significantly correlated edges in the light-green and blue modules were
   further filtered by condition of weight value being greater than 0.25.
   A total of 424 edges and 95 nodes were identified in the light-green
   module while in the blue module a total of 64 edges and 48 nodes were
   detected. The most significant hub clusters were identified by applying
   the Molecular Complex Detection (MCODE) to the networks in the
   light-green and blue modules (Fig. [110]9). A single hub cluster of
   highly correlated gene was identified in the light-green module
   (Fig. [111]9a). This cluster contains 32 genes including 2
   transcription factors (Ls6_52740.1, Ls9_8621.1), 4 nitrogen
   metabolism-related genes (Ls1_51401.1, Ls1_62361.1, Ls_62501,
   Ls1_63521.1, Ls6_45380.1), 3 secondary metabolite genes (Ls2_83840.1,
   Ls5_162100.1, Ls8_129120.1) and a carbon assimilation gene
   (Ls7_110780.6). Two hub clusters were identified in the blue module
   which included 29 and 23 genes respectively (Fig. [112]9b, c). The
   first cluster appeared to be regulated by a single bHLH transcription
   factor (Ls4_164941.1) while the second cluster appeared to be regulated
   by a GRAS family TF (Ls4_92980.1) and Ethylene-responsive transcription
   factor RAP2-11 (Ls5_31460.1).

Figure 9.

   [113]Figure 9
   [114]Open in a new tab

   The co-expression network of highly correlated (r^2 > .90) hub genes
   and associated TFs identified by applying the Molecular Complex
   Detection (MCODE) to the networks. (a) The co-expression network of the
   significant genes in the lightgreen module. (b, c). Two co-expression
   clusters of the significant genes in the blue module. The TFs in each
   cluster are red colored.

Discussion

   Transcriptome profiling using RNA-Seq enabled comparing transcriptional
   response of two lettuce genotypes, Salinas and UC, grown in two
   contrasting N regimes. The results demonstrate the common and unique
   differential response of these genotypes. Although some stress
   responsive pathways were similar in both genotypes there were
   significant differences at transcriptional levels in the cultivated
   (Salinas) and the wild (UC) lettuce to the N stress conditions.

   We found that N stress adversely affected several photosynthesis
   related genes in lettuce such as those encoding for the light
   harvesting complex (LHC), the photosystem-I (PS-I) reaction center and
   RuBisCO. The LHC is associated with photosystem I and II which is made
   up of chlorophyll-a, chlorophyll-b, and binding proteins that act as
   light receptor to capture and deliver the excitation energy between
   PS-I and PS-II and also regulate the distribution of this energy under
   changing light conditions^[115]45. Similar results were reported in
   bread wheat^[116]46 and in rice^[117]47 where N stress significantly
   decreased expression of components of LHCs of both PSII and PSI.
   Earlier studies have documented the effect of low N on physiological
   processes including low photosynthetic CO[2] assimilation and decreased
   photosynthetic rate^[118]48–[119]50, and also reduced N^[120]51 and
   phosphorus^[121]5 content in leaves. In addition, insufficient N can
   affect photosynthetic pigments^[122]52, activity of Rubisco
   proteins^[123]50,[124]53, and chlorophyll fluorescence^[125]50.

   The photosynthetic response of the plant under abiotic stress can be
   quantified using chlorophyll fluorescence measurements. Our results
   showed that the photosynthetic efficiency (QY_max) and photosynthetic
   pigment concentration (SPAD values) were significantly lower in both
   genotypes under N stress conditions due to downregulation of genes
   involved in chlorophyll biosynthesis. Concomitantly, we observed a
   significant increase in non-photochemical quenching (NPQ) under N
   stress (Fig. [126]1). Thermal dissipation measured as NPQ, is a
   photoprotective mechanism that can eliminate excess irradiated energy
   absorbed by the plant. When plants absorb more light energy than they
   can utilize during abiotic stress, excess energy leads to the
   production of reactive oxygen species (ROS) which can cause severe
   damage to the plant’s photosynthetic apparatus and at higher levels can
   cause cell death. NPQ has been shown to increase under stresses
   conditions and plays an important role in the reduction in electron
   transport, increases in heat dissipation, and resistance to abiotic
   stresses^[127]54. Increases in NPQ in Salinas and UC genotypes suggest
   that dissipating the excess of excitation energy absorbed by PSII as
   heat is the principal pathway to reduce ROS formation under N stress
   conditions.

   N uptake, assimilation and metabolism and its effect on growth and
   biomass production have been studied in several agriculturally
   important plant^[128]19,[129]55. The process of NO[3]¯ uptake,
   translocation and storage in plant is a complex process and is often
   controlled by low-affinity transporter (LAT) and high-affinity
   transporter (HAT) genes^[130]13. Most nitrate transporter 1 genes act
   as nitrate sensors and function as low-affinity transporter genes for
   NO[3]^− at high concentrations, with the exception of AtNRT1.1 in
   Arabidopsis and MtNRT1.1 in Medicago truncatula that serve as
   dual-affinity transporter involved in LAT and HAT systems^[131]56.
   Previous studies demonstrated upregulation of NRT1.1 upon addition of
   nitrate to N starved plants. For example, cultivar specific differences
   were observed in Brassica juncea for the expression of NRT1.1 and
   NRT1.8 that were highly induced as early as 20 min after exogenous
   supply of nitrate^[132]10. In this study we did not find significant
   differences for Lettuce NRT1.1 gene expression under HN or LN condition
   confirming the role of NRT1.1 gene in nitrate perception. In contrast
   to our finding^[133]46, found that in wheat the expression of most of
   the dual affinity nitrate transporters (like NRT1.1) decreased under
   nitrogen stress and suggested that the reduced expression of NRT1.1
   genes may lead to retarded growth, low grain protein content and low
   grain yield of N stressed plants. We observed that the expression of
   many high affinity nitrate transporter NRT2 genes was altered under low
   N conditions. The NRT2 genes family belong to HAT system and are
   responsible for transporting nitrate at low concentrations, therefore
   they play an important role during nitrogen stress
   conditions^[134]13,[135]57,[136]58. Many NRT2 family members require
   NAR2 (nitrate assimilation related protein) for nitrate transportation
   and have different spatio-temporal distribution in
   roots^[137]57,[138]59,[139]60. In Arabidopsis, NRT2.1 mediates
   apoplastic nitrate absorption and affects root system architecture
   under low nitrogen conditions^[140]58,[141]61. Unlike NRT2.1, NRT2.4
   and NRT2.5 are expressed in epidermal cells of roots under long term
   nitrogen starvation and responsible for nitrate uptake from soil into
   the plant^[142]57,[143]60. Compared to Salians genotype, we found
   induced expression of several high affinity nitrate transporter genes
   (NRT2.1, NRT2.4 and NRT2.5) in UC genotype under low N condition which
   may explain higher N content observed in UC than Salinas genotype thus
   indicating that UC genotype could have higher N uptake efficiency under
   low N conditions.

   The process of converting inorganic nitrogen from soil into organic N
   is carried out by two key enzymes, glutamine synthase (GS) and
   glutamate synthase (GOGAT). GS catalyzes adenosine triphosphate
   (ATP)-dependent fixation of ammonium to form glutamine while the GOGAT
   enzyme catalyzes the conversion of glutamine to two molecules of
   glutamate^[144]62. Glutamine serves as a main nitrogen donor for
   biosynthesis of amino acids, nucleotides, and chlorophyll. Therefore
   the GS enzyme is the major factor affecting nitrogen assimilation in
   plants^[145]63. There are two isoenzyme forms of the GS enzyme, one
   located in the cystosol (GS1 or GLN1) and the other in
   chloroplast/plastids (GS2 or GLN2). The cytosolic GLN1 plays an
   important role in N assimilation in roots while the chloroplastic GLN2
   is expressed primarily in leaves and is responsible for the
   reassimilation of the ammonia generated by photorespiration^[146]63. In
   this study, we found that the expression of gene for GLN1 was
   up-regulated under N stress in both genotypes while the expression of
   GLN2 was higher in Salinas under high N but was down-regulated under N
   stress. There was no change in the expression levels of the GLN2 in UC
   under high or low N conditions. Together, these results suggest that
   the UC can sustain N assimilation under low N condition and may account
   for higher N content and biomass produced during nitrogen stress.

   TFs recognize specific motifs and function as a switch that can turn on
   or off a particular gene. In this study we found that the expression of
   TFs like bHLH, bZIP, ERF, MYB, NAC, and WRKY was altered in response to
   N stress (Supplementary_Fig. [147]4). The Apetala2/ethylene responsive
   factors (AP2/ERF) family TFs are the key regulators of numerous abiotic
   stresses and respond to several plant hormones^[148]64,[149]65. These
   TFs are induced upon specific stresses and have diverse DNA binding
   preferences, enabling these TFs to integrate responses of multiple