Optimization of Nitrogen Demand in Vegetables by Different Impacts on Autotrophic and Heterotrophic Nitrication

Aims The understanding of the interactions between N transformations and N uptake by plants in greenhouse soils with large N accumulation is still not clear. The aim is to understand the plant- soil interactions (vegetables) on N transformations with respect to N supply. Methods 15 N tracing studies were conducted in two greenhouse soils to simultaneously quantify soil gross N transformation and plant N uptake rates using the Ntrace plant tool. Results There were signicant feedbacks between vegetable N uptake and soil gross N transformation rates, whether soil N accumulation occurred or not. Plant NO 3– uptake rates (U NO3 ) were higher than the NH 4+ uptake rates (U NH4 ), which is consistent with the NO 3– -preference of the vegetable plants studied. While U NH4 was still responsible for 6-49% of total N uptake rates, signicantly negative relationships between U NH4 and NH 4+ immobilization rate and autotrophic nitrication rate (O NH4 ) were observed. O NH4 was signicantly inhibited in the presence of plants and decreased with time. O NH4 (1.11 mg N kg -1 d -1 ) was much lower than U NO3 (8.29 mg N kg -1 d -1 ) in the presence of plants. However, heterotrophic nitrication rate (O Nrec ), which ranged from 0.10 to 8.11 mg N kg -1 d -1 was signicantly stimulated and was responsible for 5-97% of NO 3– production in all plant treatments, providing additional NO 3– to meet N requirements of plants and microorganisms. nitrication M Nlab and M Nrec of D15 were none signicantly lower compared to CK. In LST, no signicant differences among CK, D15 and D26 in M Nlab were found, while the M Nrec of D26 was signicantly higher than in CK and D15. In HSC and LSC, M Nlab was 5-128 times lower than in CK at D15 and D26, while M Nrec was 2-47 times higher than in CK at D15 and D26. Therefore, M (M Nlab + M Nrec ) was comparable between soils with (D15 and D26) and without (CK) planting in all treatments.


Introduction
Greenhouse vegetable cultivation (GVC) is increasingly used to meet the demand for vegetables and improve farmers' income (Min et  Previous studies have focused on the feedbacks between plant N uptake and soil gross N transformations in soils with low N content. We hypothesized that 1) Vegetables depend mainly on the amount of available N in the soil, which is re ected in weak feedbacks on the processes of mineral N production in the soil, e.g. N mineralization and heterotrophic nitri cation; 2) Vegetables generally prefer NO 3 − , thus they should stimulate autotrophic nitri cation to meet their NO 3 − demand. In this study, two main varieties of vegetables under GVC (i.e. tomato (Lycopersicon esculentum L.) and cucumber (Cucumis sativus L.)) and two greenhouse soils with different N accumulation were selected. To test our hypotheses, a series of 15 N-tracing studies were carried out to quantify soil gross N transformation rates and plant N uptake rates.

Soil samples
Two greenhouse soil samples were collected from Suzhou, Jiangsu Province, China. On one soil, tomato and strawberry were grown continuously for more than 10 years, with a high accumulation of NO 3 − (on average 82.83 mg N kg −1 ) (HS). The other was previously planted with tomato and cucumber and lay fallow for several years, resulting in a low NO 3 − concentration (on average 8.97 mg N kg −1 ) (LS). Four plots (1 m ⋅ 1 m) were randomly selected at each sampling site. The surface soil (0-20 cm) was collected and pooled together in equal amounts. Then soil samples were immediately passed through a 2 mm sieve and roots were removed. Each sample was divided into three sub-samples. One was air-dried to determine soil chemical properties (Table 1), one was stored in a refrigerator at -80 o C for the measurement of soil microbial properties, and another sub-sample was stored at 4 o C to perform 15 N tracing studies. In addition, soil pH, dissolved organic C (DOC) concentration, the abundances of bacteria, fungi, ammonia-oxidizing archaea (AOA) and ammonia-oxidizing bacteria (AOB) were also determined.

Analytical methods
Soil pH was measured in a 2.5:1(v/w) water/soil ratio using a DMP-2mV/pH detector (Quark Ltd., Nanjing, China). Soil electrical conductivity (EC) was determined by a conductivity detector (Kang Yi Crops., Nanjing, China) at a 5:1 (v/w) water/soil ratio.  Due to the faster growth rate of cucumber than tomato, the N uptake rates of cucumber were higher than that of tomato, and thus, NH 4 + and NO 3 − in soil planted with cucumber could not be detected at 72 h after addition of 15  The T-test analysis was used to examine the differences in soil properties between studied two soils. The relationships among N transformation rates were examined using Spearman correlation analyses with two-sided tests.

Soil properties
The two soils had comparable concentrations of DOC and NH 4 + , as well as abundances of bacteria before the start of the study (Table   1) The DOC concentrations signi cantly increased at D26 in HSC and LSC treatments (cucumber), while, decreased in HST and LST treatments (tomato) (Fig. 2a). Soil pH of CK was not different from D26, while it was slightly lower than in D15 in all treatments (Fig. 2b).
There was no signi cant change in the abundances of bacteria and AOA in all treatments (Fig. 2c, e). The abundance of fungi was signi cantly higher in D15 and D26 than CK (Fig. 2d). Moreover, the abundance of fungi in D26 was higher than in D15, particularly in soils planted with cucumber. The abundance of AOB signi cantly decreased by the presence of plants, and decreased with planting time (Fig. 2f).

NH 4 + dynamics
Modeled and measured values of the concentrations and 15 N abundances of soil NH 4 + , NO 3 − and plant N pools matched well (Fig. S1-S4). The Ntrace analysis showed that soil gross N transformation rates were signi cantly affected by the presence of plants (Fig. 3), except for dissimilatory NO 3 − reduction to NH 4 + , release of adsorbed NH 4 + and adsorption of NH 4 + on cation exchange sites rates (data not shown, because they were close to zero).
Generally, there were no signi cant differences in total N mineralization rates (M, mineralization rate of labile (M Nlab ) and recalcitrant (M Nrec ) organic N pool to NH 4 + ) between HS and LS in all treatments (Fig. 3a) For CK, the total gross NH 4 + immobilization rate (I NH4 ) in HS was signi cantly higher than in LS (Fig. 3b). I NH4 was negligible at D15 and D26, and signi cantly lower than CK (0.39 mg N kg −1 d −1 ) in HST and HSC treatments. In LST and LSC treatments, the I NH4 of D15 was 3.35 mg N kg −1 d −1 and 1.56 mg N kg −1 d −1 , respectively, and signi cantly higher than in CK and D26 where I NH4 was less than 0.04 mg N kg −1 d −1 ; there were no signi cant differences in I NH4 between CK and D26.

NO 3 − dynamics
In the CK treatments, the rate of NH 4 + oxidation to NO 3 − (autotrophic nitri cation, O NH4 ) was signi cantly higher in HS than in LS, and the rate of organic N oxidation to NO 3 − (heterotrophic nitri cation, O Nrec ) was negligible in both soils (Fig. 3c). A signi cant decrease in O NH4 was found at D15 and D26 compared with CK in all treatments. O NH4 in D15 were 2, 2, 11, and 6 times lower than CK in the HST, LST, HSC and LSC treatments, respectively. At the same time, O NH4 decreased signi cantly in D26 compared to D15 in all treatments.
Interestingly, the other NO 3 − production pathway, oxidation of organic N to NO (Fig. 4e, f).
The NO 3 − immobilization rate (I NO3 ) was negligible in CK of both HS and LS (Fig. 3b). In HST and LST treatments, there were no signi cant differences between soil without (CK) and with planting (D15 and D26) in I NO3 . However, in HSC, I NO3 was signi cantly higher at D15 and D26 compared to CK, and signi cantly increased from 5.56 mg N kg −1 d −1 at D15 to 10.84 mg N kg −1 d −1 at D26. In the LSC treatment, I NO3 was signi cantly higher in D15 than in CK and D26, and no signi cant difference was observed in I NO3 between CK and D26.

Plant N uptake
The plant NH 4 + uptake rate (U NH4 ) increased with planting duration, except in the HSC treatment where U NH4 was signi cantly higher in D15 than in D26 (Fig. 3d) Although, U NH4 was responsible for 12%, 6%, 49% and 23% of total N uptake in D15, and 34%, 26%, 41% and 17% of total N uptake in D26 in the HST, LST, HSC and LSC treatments, respectively.
The U NH4 was signi cantly and positively related to M Nrec (p < 0.05) (Fig. 4a), while negatively related to I NH4 (p < 0.05) and O NH4 (p < 0.05) (Fig. 4b, c). A signi cantly positive relationship exists between the U NO3 + I NO3 and O Nrec (p<0.01), and between U TN +I TN and M + O Nrec (Fig. 4g, i). The rate of I NO3 and U NO3 were negatively correlated with each other (i.e. an increasing I NO3 and decreasing U NO3 ) (p<0.05) (Fig. 4h).

Discussion
Consistent with our hypothesis we observed signi cant feedbacks between vegetable N uptake and soil N transformations, irrespectively whether or not N accumulation occurred. The studied vegetables with NO 3 − preference also assimilated a considerable quantity of NH 4 + , while soil did not supply additional NH 4 + via N mineralization, resulting in a reduction of NH 4 + immobilization and NH 4 + oxidation (autotrophic nitri cation rate). Under conditions where NO 3 − production via autotrophic nitri cation alone did not meet vegetables NO 3 − demand a stimulation of NO 3 − production via heterotrophic nitri cation occurred, which was not consistent with our hypothesis that vegetables with a NO 3 − preference would stimulate autotrophic nitri cation to meet their NO 3 − demand. However it was in line with results by (He et al. 2022).
Soil N mineralization is the primary source of soil available N. In this study, there were no signi cant differences in total gross N mineralization rates with respect to the presence or absence of plants, because gross mineralization rates of speci c organic N pools (i.e. labile and recalcitrant organic N pool) responded in an opposite way to plants (Fig. 3a) there was a signi cant and positive relationship between M Nrec and plant NH 4 + uptake (p<0.05) (Fig. 4a) In the present study, autotrophic nitri cation rates signi cantly decreased with time of plant growth (Fig. 3c), indicating that presence of vegetables (even NO 3 − -preference) did inhibit autotrophic nitri cation rates. Our results found that there were no signi cant differences in gross N mineralization rate between presence and absence of plants, indicating that soil in the presence of plants failed to supply more NH 4 + via gross N mineralization to autotrophic nitri cation. In the plant-soil system, NH 4 + is not only the substrate of autotrophic We observed that autotrophic nitri cation rate signi cantly decreased with increasing plant NH 4 + uptake rate (p<0.05) (Fig. 4c). In spite of the NO 3 − -preference nature of studied plants (tomato and cucumber) (Al-Harbi 1995), the plant NH 4 + uptake rate was responsible for 6%-49% of total N uptake rates in all treatments. One of possible reasons were to maintain a suitable ionic balance in plant, therefore also NH 4 + is required for NO 3 − preference plants. To keep NH 4 + at a certain level it makes then sense that NH 4 + oxidation is inhibited, while NO 3 − is generated via organic N oxidation rather than NH 4 + oxidation. The assimilation of NH 4 + by plants could reduce the NH 4 + substrates, and further inhibited autotrophic nitri cation. Furthermore, plant NH 4 + uptake rate (average 3.13 mg N kg −1 d −1 ) was higher than microbial NH 4 + immobilization (average 0.62 mg N kg −1 d −1 ), indicating that plants are stronger competitors for NH 4 + than microorganisms (Schimel and Bennett 2004). This likely lead to a decrease in the abundance of microorganisms, as observed in our study in the presence of plants (Fig. 2f). AOB rather than the AOA carried out autotrophic nitri cation, despite a higher AOA than AOB abundance in greenhouse soil (Di et al. 2009), being in line with our results. Thus, the reduction of abundance of AOB was likely be responsible for lower autotrophic nitri cation rates in the presence of plants. This was strongly supported by a signi cant positive relationship between autotrophic nitri cation rate and abundance of AOB (p<0.01) (Fig. 4d). The autotrophic nitri cation rate (1.11 mg N kg −1 d −1 ) was much lower than plant NO 3 − uptake rate (8.29 mg N kg −1 d −1 ) in the presence of plants, indicating that NO 3 − production via autotrophic nitri cation alone failed to meet plants NO 3 − requirement.
Our results showed that the rate of heterotrophic nitri cation was signi cantly enhanced by the presence of plants (Fig. 4c), accounting for 5%-97% of total NO 3 − yield. We found that plant NO 3 − uptake rate and microbial NO 3 − immobilization rate were signi cantly positively related to heterotrophic nitri cation rate (p<0.01) (Fig. 4g). Furthermore, plant N uptake rate and microbial N immobilization rate were not signi cantly correlated with gross N mineralization rates, while they were signi cantly and positively related to total inorganic N production rates (i.e. gross N mineralization rate + heterotrophic nitri cation rate) (p<0.05) (Fig. 4i). . In our study, soil SOC, C/N ratio, and pH did not change signi cantly during the experimental duration. We found that DOC concentrations and the abundance of fungi signi cantly increased during plant growth and were related to heterotrophic nitri cation rates (Fig. 4e, f) Lang and Jagnow 1986). We also found a signi cantly positive relationship between O Nrec and the abundance of fungi (p < 0.05). Thus, plant activity and heterotrophic microorganisms in rhizosphere rather than soil reaction affected heterotrophic nitri cation in studied soils.
Another observation was that microbial N immobilization rates were signi cantly affected by the presence of plants, whether N accumulation occurred or not. Our results showed that the microbial NH 4 + immobilization rates were reduced by the presence of plants, except in D15 of LST and LSC treatments (Fig. 3b). Plant NH 4 + uptake rates were negatively correlated with microbial NH 4 + immobilization rates (p<0.05) (Fig. 4b) (Fig. 4h).

Conclusions
Our results highlighted that there were signi cant feedbacks between vegetable N uptake and soil gross N transformation rates, whether or not N accumulation occurred in soils. Despite that the studied vegetables preferred NO 3 − , they still possessed a stronger competitive capacity for NH 4 + than microorganisms, resulting in the reduction of NH 4 + immobilization rate and autotrophic nitri cation rate. Due to the low autotrophic nitri cation rate in the presence of plant, heterotrophic nitri cation was stimulated by NO 3 − -preference plants to become an important supply pathway of NO 3 − . Therefore, the management of organic N fertilizer for greenhouse cultivation with NO 3 −preference plants (such as tomato and cucumber) should take these results into account to enhance the capacity of NO 3 − production via heterotrophic nitri cation. Whether strong feedbacks of plant N acquisition and soil N transformations in soils exist under high N accumulation should be con rmed in further studies when different plants and soils are studied.