Root RNA was extracted using Trizol reagent according to the manufacturer’s guidelines followed by DNase digestion using RQ1 RNase-free DNase . Total RNA was purified using the RNeasy Plant Mini Kit . RNA concentrations and quality were assessed using the Agilent Nanodrop and the RNA 6000 Nano Assay . Only RNA samples with RNA integrity numbers of at least 7.0 were used for subsequent analyses. These RNA were used for cDNA synthesis for qRT-PCR analysis. cDNA was synthesized from 0.5 μg DNase-treated total RNA using the Superscript III kit .The 13 organically-managed Roma-type tomato fields spanned a three-fold range of soil C and N and had similar soil texture, soil types, and soil pH . Field numbers are in order of increasing total soil C. Variation in nutrient inputs, including highly-labile secondary inputs indicated diverse and intensive organic management strategies across these farms . The 13 fields encompassed the majority of the variation in the focal landscape , i.e. all but one of the five clusters, or landscape types, identified by GIS and multivariate analysis were represented . Thus, a range of soil characteristics representative of this region was accounted for by the fields sampled. Other characteristics, such as the low number of crop rotation types differed little across the clusters and reflect the intensity of agricultural management in this region.Measures of tomato N sufficiency varied widely across the 13 organic fields, ranging from deficient to luxury N levels. Total above ground N concentration at midseason overlapped or fell slightly below the critical N concentration for processing tomatoes in most fields , with N concentrations between 2.5 and 3.5% . Exceptions were fields 1 and 2 which were markedly lower and field 4,maceta 25 litros which was higher . The same general pattern occurred for the harvest sampling. Petiole NO3 – concentration in four fields overlapped the sufficient concentration, based on published guidelines , while five fell below it, and four rose above it .

Petiole NO3 – was especially high in field 4. Petiole NO3 – -N showed a broadly similar pattern to total above ground N concentration, as reflected in the strong linear relationship between them . At the mid-season sampling, shoot δ15N ranged from 4.22 ± 0.65‰ in field 10 to 13.29 ± 1.18‰ in field 6. Fields 3, 4, 6, and 9 had the highest shoot δ15N, all above 12‰, and all but field 3 used seabird guano. Fields 8, 10, 11, and 13 had the lowest shoot δ15N, close to 4‰, and all but field 8 used Chilean nitrate. Mean harvestable fruit yield across all 13 fields was 86.7 ± 7.2 Mg ha-1 and was similar to the overall Yolo County mean 2011 tomato yield , which included both conventional and organic fields . Field 4 had the highest yield overall followed closely by field 9 , and field 1 had the lowest . Nine of 13 fields had means higher than the county average, and six of these fields were significantly higher. There was also substantial variability in tomato above ground biomass and N content at harvest across fields , which largely reflected the pattern of fresh weight yields. For instance, total above ground N ranged from 64 kg N ha-1 in field 1 to 243 kg N ha-1 in field 4 with a mean across all fields of 154 kg ha-1.Expression of cytosolic glutamine synthetase GS1 in roots was more strongly related to indicators of plant-soil N cycling than were the other six key genes involved in root N metabolism . Of the soil variables, GS1 was more strongly related to soil bioassays for N availability than to inorganic N pools . Microbial biomass N and PMN were most strongly associated with expression of GS1 in roots, followed by soil NO3 – . Permanganate oxidizable C and MBC, both indicators of labile soil C pools, also had significant associations with GS1 expression in roots, but soil NH4 + did not. Expression of GS1 also was positively associated with shoot N and petiole NO3 – , as was glutamate synthase NADH-GOGAT. Inclusion of GWC as a covariate in multiple linear regression models improved the proportion of explained variation in GS1 expression .PCA of 28 indicators of yield and plant nutrient status, root N metabolism, and soil C and N cycling showed strong relationships among suites of variables, which clearly differentiated fields along the first two principal components .

The first principal component explained 28.3% of the variation; on the left side of the biplot are higher values of most variables, including yield, soil bioassays, expression of root GS1 and NADH-GOGAT, and labile and total soil C and N pools . Soil NH4 + and NO3 – concentrations from all three sampling times as well as AMT1.2 were associated with one another and with positive values along principal component 2, which explained 19.4% of the variation. Total soil C and N were strongly associated with EOC and EON, the soil C:N ratio, and POXC. These variables had negative values along axis 2 and thus contrasted with the pattern of soil inorganic N. Weak loading of AMT1.1, NRT2.1, Nii, and GS2 on the first two principal components reflects the lack of association of expression levels of these genes with biogeochemical and plant variables. Non-overlapping confidence ellipses for seven out of 13 fields on the PCA biplot indicated distinct N cycling patterns . Fields 1 and 2, with the highest values along axis 1, had low values of all variables included in the analysis. Field 4 had the highest values along axis 2 corresponding with higher soil NH4 + and NO3 – . Fields 10, 11, 12, and 13 were associated with high values of labile and total soil C and N. Overlapping confidence ellipses of fields 3, 5, 6, 7, 8, and 9 close to the origin indicate similar, moderate values of this suite of variables for these fields. Three groups of fields were identified by k-means cluster analysis of the same 28 variables included in the PCA . Group 1 included fields 1 and 2, which had low mean values for yield , the lowest mean soil C and N and soil inorganic N pools , and the lowest mean value of GS1 relative expression in roots. Groups 2 and 3 had similarly higher mean yield , shoot N,maceta redonda and petiole NO3 – than group 1, but these two groups differed substantially in their soil C and N pools. Group 2 had higher soil NH4 + and NO3 – pools as well as root expression of AMT1.2 while group 3 had higher total and labile soil C pools. Expression of GS1 was similar in both groups. Based on the relative magnitude of F-statistics calculated for each variable, soil C and N, EOC, EON, shoot N, and soil NO3 – at transplant and anthesis were most strongly differentiated across the three groups. The high F-statistics of AMT1.2 and GS1 relative to other N metabolism genes indicate that root expression of these genes are most responsive to soil N cycling.

This study confirms that working organic farms can produce high yields with tightly-coupled N cycling that minimizes the potential for N losses. Such farms had the highest soil C and N and used high C:N organic matter inputs coupled with labile N inputs that resulted in high soil biological activity, low soil inorganic N pools, high expression for a root N assimilation gene, adequate plant N, and high yields. Organic systems trials have previously shown crop N deficiencies that lead to less-than-ideal crop productivity; losses of N when NTo characterize the substantial variation in crop yield, plant-soil N cycling, and root gene expression across 13 fields growing the same crop on similar soil types, we propose three N cycling scenarios: “tightly-coupled N cycling”, “N surplus”, and “N deficient”. Values of indicator variables suggest differing levels of provisioning, regulating, and supporting ecosystem services in each scenario . Fields in group 3 show evidence of tightly-coupled plant-soil N cycling, a desirable scenario in which crop productivity is supported by adequate N availability but low potential for N loss. Despite consistently low soil NO3 – pools in these fields, well below the critical mid-season level for conventional processing tomatoes in California, total above ground N concentrations were very close to or only slightly below the critical N concentration for processing tomatoes. Tomato yields were also above the county average . This discrepancy between low soil inorganic N pool sizes and adequate tomato N status is due to N pools that were turning over rapidly as a result of efficient N management, high soil microbial activity, and rapid plant N uptake. Composted yard waste inputs with relatively high C:N ratios in concert with limited use of labile organic fertilizers applied during peak plant N demand provided organic matter inputs with a range of N availability. A companion study showed how high potential activities of N-cycling soil enzymes but lower activities of C-cycling enzymes in this set of fields reflect an abundant supply of C but N limitation for the microbial community, thus stimulating production of microbial enzymes to mineralize N. Plant roots can effectively compete with microbes for this mineralized N, especially over time and when plant N demand is high. High root expression of GS1 in these fields indicates that root N assimilation was elevated and thus actual plant N availability and uptake was higher than low inorganic N pools would suggest .

Fields from group 2 demonstrated N surplus, showing similar yields to group 3 but with lower total and labile soil C and N and a higher potential for N losses, given much higher soil inorganic N . While actual N losses depend on a host of factors , high soil NO3 – is considered an indicator for N loss potential. Results from a companion study support the idea that soil microbes were C rather than N-limited in these fields, showing higher potential activities of C-cycling soil enzymes but low activities of N-cycling soil enzymes, the inverse of group 3 . An alternative multivariate clustering approach based on an artificial neural network suggests multiple potential drivers of higher inorganic N pools in these fields, including both management factors and soil characteristics . For instance, field 4 had strong indications of surplus N driven at least in part by a large application of seabird guano , a readily-mineralizable organic N fertilizer, at tomato transplanting when plant N demand is low. In contrast, higher inorganic N in field 8 was likely driven by low plant N demand based on very low soil P availability, which resulted in plant P limitation. These site-specific problems were identifiable due to the focus on variability across similar organic fields and illustrate the need for site-specific approaches to reduce N losses. Finally, the two fields included in group 1 were exemplary of N deficiency, in which low N availability compromises crop productivity but also likely limits N losses within the growing season. While low soil NH4 + and NO3 – concentrations were similar to group 3, low total and labile soil organic matter and poorly-timed organic matter inputs compromised microbial activity and likely limited N mineralization.Cytosolic glutamine synthetase GS1 encodes for the enzyme that catalyzes the addition of NH4 + to glutamate, the former resulting from either direct uptake of NH4 + from soil or reduction of NO3 – in roots. GS1 is thus the gateway for N assimilation in roots and is upregulated to increase root N assimilation capacity. Similar levels of GS1 expression in groups 2 and 3, in spite of large differences in soil NH4 + and NO3 – concentrations at the anthesis sampling, suggests that plant N availability is indeed higher in group 3 fields than would be expected based on measurement of inorganic N pools alone.These results complement recent experimental approaches that showed rapidly increased expression of GS1 in tomato roots in response to a pulse of 15NH4 + -N on an organic farm soil, which was linked to subsequent increases in root and shoot 15N content, even when this pulse did not significantly change soil inorganic N pools.