Effects of Cultivation Years in Protected Vegetable Crop Fields On The Structure and Abundance of Soil nosZ Denitrifying Microbial Community

The study of the changes in the structure and quantity of the denitrifying microbial community in the protected vegetable crop eld is crucial for evaluating the soil quality of long-term protected cultivation and improving the understanding of nitrogen conversion. In this study, quantitative uorescent PCR was used, with the help of the Illumina Miseq high-throughput sequencing platform, aiming at the nosZ gene, to study the effects of protected cultivation plots 3 a, 5 a, and 7 a, and the outdoor cultivation plot (ck) on the structure and quantity of soil denitrifying microbial community. The results demonstrated that the abundance of nosZ gene in ck was signicantly higher than the other treatments, and was 1.32, 1.45 and 1.69 times those of Group 3 a, 5 a and 7 a, respectively. As the cultivation year increased, the abundance of nosZ gene was decreased, and the α-diversity index was decreased. The Chao1 index and ACE index of ck were the highest. At the phylum level, the dominant group was Proteobacteria. While at the genus level, the dominant groups were Bradyrhizobium and Achromobacter. The relative abundances of Proteobacteria and Bradyrhizobium were decreased with the increase of protected cultivation time. The principal component analysis (PCA) results showed that the community structure of nosZ varied greatly with the increase of planting time. The community structures of Group 3 a and 5 a were similar, but the community structure of Group 7 a was very different from Group 3 a and 5 a. The soil available potassium (AK), ammonium nitrogen and nitrate nitrogen were the primary factors affecting the copy number, α-diversity and community structure of nosZ denitrifying microorganisms. As a conclusion, it can be seen that long-term cultivation of vegetable crops in a protected environment signicantly reduced the copy number of nosZ denitrifying microorganisms and had a signicant impact on the community structure.


Introduction
Protected cultivation of vegetable crops is an important branch of the modern intensive agriculture, and is greatly signi cant for improving natural resource utilization e ciency and boosting farmers'economic bene ts. With the growing population and increasing demand for vegetable consumption in China, as well as limited arable land resources, protected cultivation has been dramatically developed in the past 40 years 1 . As of 2016, the protected cultivation area in China had exceeded 58 million mu (1 mu = 0.165 acre), ranking rst in the world 2 . Currently, protected cultivation production has become a pillar industry in various regions of China. However, protected cultivation has the characteristics of high cropping index, great agricultural investment (such as fungicides, fertilizers, and agricultural lms), closed or semi-closed environments with high temperature and high humidity, and no rain leaching. Therefore, the long-term protected cultivation has caused concerns of soil degradation, potential soil and vegetable contamination, and adverse health effects 3 .
Soil microorganisms are one of the essential components of the soil ecosystem. They contribute to energy circulating, nutrient cycling, organic matter conversion, pollutant degradation, and prevention and control of soilborne diseases in the soil ecosystem 4 . Related studies believed that long-term protected cultivation had a negative impact on the structure and function of soil microbial community, primarily referring to the community succession from bacteria-dominant to fungi-dominant, and the reduction of the interaction between function-related microbial community members and species, resulting in the replacement of ecologically bene cial microbial community by soilborne pathogens [5][6][7][8] . Currently, scholars are more concerned about the overall changes of soil microbial community or targeted microorganisms related to soil-borne diseases. Due to the high degree of structural and functional diversity of soil microorganisms, various types of soil biological processes are driven by the functional groups of microorganisms involved. So far, the response of soil functional microorganisms to long-term vegetable crop production remains poorly understood, especially that of the microorganisms related to nutrient cycling. Nitrogen (N) is a primary factor limiting plant growth, and soil nitrogen availability plays a crucial role in determining the nitrogen absorption and yield of plants. Nitrogen conversion mainly includes the processes of N xation, mineralization, nitri cation and denitri cation 9 . Among them, denitri cation driven by soil microorganisms is an important part of the natural nitrogen cycle and the primary driving force for maintaining the global nitrogen balance. Nitrous oxide (N 2 O) produced by denitri cation is a powerful greenhouse gas and the primary ozone-depleting substance in the atmosphere 10 . It has a global warming potential (GWP) 298 times that of carbon dioxide (CO 2 ) 11 . Due to excessive organic fertilizer, nitrogen fertilizer input and ood irrigation, the denitri cation in soil was strong, and resulted in a N 2 O emission ux 1.41 times higher than that of the eld 12 . Therefore, how to regulate N2O conversion through microorganisms has become the focus of many scholars 13,14 . Currently, the only known biological process in denitri cation is the reduction of N 2 O by the N 2 O reductase encoded by the nosZ gene 15 . The nosZ gene has been used to study the community structure and diversity of denitrifying microorganisms in forests, grasslands and farmlands 16,17 . However, no reports of the impact of longterm protected cultivation on the structure and quantity of nosZ-type denitri er community in soil are available yet.
Therefore, this study used Illumina MiSeq high-throughput sequencing and real-time PCR to study the structure and abundance of nosZ-type denitri er communities in soil for various years of protected cultivation. The primary aims of this study were: (1) How will the structure and diversity of the denitri er community change as the planting time increases in protected cultivation plots? Which dominant community members will change? (2) For the long-term protected cultivation, which soil environmental factors have big impacts on the community structure of denitrifying microorganisms? Our study tried to reveal the denitri cation of the protected cultivation in the long-term and provided a scienti c basis for indepth understanding of the soil nitrogen cycle of protected cultivation.

Soil chemical properties
Long-term protected cultivation of vegetable crops had a signi cant impact on the chemical properties of the soil ( Table 1). The soil pH, soil organic carbon (SOC) and available potassium (AK) contents of all tested cultivation years were all signi cantly lower than those of the control, decreasing as the cultivation time increasing. The total nitrogen (TN), available phosphorus (AP), ammonium nitrogen (NH 4 + -N) and nitrate nitrogen (NO 3 − -N) of Group 3 a, 5 a and 7 a were all higher than those of ck. The total nitrogen contents of Group 5a and 7a were signi cantly increased by 38.26% and 51.30% compared with ck, whereas the difference between Group 3 a and ck was not signi cant. The AP contents of Group 3 a, 5 a and 7 a were respective 2.09, 1.62 and 2.27 times that of AP in ck. The soil ammonium and nitrate nitrogen contents increased by cultivation years, in an order of 7 a > 5 a > 3 a > ck. The range of nosZ gene copy number with different years of cultivation was 5.19 × 10 6 ~ 8.78 × 10 6 copies/g soil (Fig. 1). The copy numbers of nosZ gene in groups with various years of protected cultivation were signi cantly less than ck, whose number was respective 1.32, 1.45 and 1.69 times that of Group 3 a, 5 a and 7 a. There was no signi cant difference in the copy number of nosZ gene between Group 3 a, 5 a and 7 a, and the copy number was decreasing as cultivation time increased. The correlation analysis of nosZ gene abundance and soil chemical properties revealed that the nosZ gene abundance had signi cant positive correlations with soil pH and SOC, highly signi cant negative correlations with soil TN, ammonium nitrogen and nitrate nitrogen, a signi cant negative correlation with AP, and a highly signi cant positive correlation with AK ( Fig. 2).
α-diversity index of nosZ type denitri ers The cultivation year had a signi cant impact on the Chao1 index and ACE index of the nosZ community, and it had a highly signi cant impact on the Shannon's index and the Simpson's index ( treatments. The ranges of the Shannon's index and Simpson's index were 6.2 ~ 7.02 and 0.82 ~ 0.86, respectively, with Group 7 a signi cantly smaller than the other treatments. The Shannon's index was the highest in Group 3 a, in an order of 3 a > ck > 5 a > 7 a. The Simpson's index was the highest in Group 5a, in an order of 5 a > 3 a > ck > 7 a. The Chao1 index and the ACE index were both highly signi cantly positively correlated with soil pH, SOC and AK, and signi cantly negatively correlated with TN, AP, ammonium nitrogen and nitrate nitrogen. The Shannon's index was signi cantly positively correlated with SOC, highly signi cantly negatively correlated with TN, and signi cantly negatively correlated with ammonium nitrogen and nitrate nitrogen (Fig. 2).
nosZ community composition By classifying the nosZ community OTUs obtained from the samples, a total of six classes were obtained at the phylum level, including Proteobacteria, Bacteria_unidenti ed, Gemmatimonadetes, Bacteroidetes, Acidobacteria and Chloro exi (Fig. 3). The relative abundances of soil Proteobacteria, Bacteria_unidenti ed, Acidobacteria and Chloro exi were signi cantly different for treatments with different years of protected cultivation. The Proteobacteria, Bacteria_unidenti ed and Gemmatimonadetes were the dominant phyla, with relative abundance ranges of 62.27 ~ 70.42%, 16.66 25.19% and 2.88 ~ 3.82%, respectively. The relative abundances of Proteobacteria in Group 5a and 7a were less than that of ck, lowered by 2.69% and 10.34%, respectively. The relative abundance of Bacteria_unidenti ed was the highest in Group 7 a, signi cantly higher than that of Group 3 a and 5 a. The relative abundance of Gemmatimonadetes was the highest in Group 3a and the lowest in Group 5 a, in an order of 3 a > 7 a > ck > 5 a. The protected cultivation exerted an inhibitory effect on the Proteobacteria phylum, and the inhibition was increased with the increase of the cultivation time.
A total of ten classes were obtained at the genus level with those with an average relative abundance of < 1% classi ed into one group (Fig. 4), including Bradyrhizobium, Bacteria_unidenti ed, Achromobacter, Aromatoleum, Gemmatimonas, Ochrobactrum, Azoarcus, Azospirillum, Mesorhizobium and Rhodopseudomonas. Among them, signi cant differences in the relative abundance of Bradyrhizobium, Bacteria_unidenti ed, Achromobacter and Rhodopseudomonas were seen in the protected cultivations for different years. The Bradyrhizobium, Bacteria_unidenti ed and Achromobacter were the dominant genera, with relative abundance ranges of 39.01 ~ 47.68%, 18.48 ~ 27.68% and 5.28 ~ 8.57%, respectively. The relative abundance of Bradyrhizobium in various protected cultivation groups was higher than that of ck, respective 1.23, 1.12 and 1.07 times that of ck, in an order of 3 a > 5 a > 7 a > ck. The relative abundance of Achromobacter was signi cantly lower in Group 7a than the other treatments, in an order of 5 a > ck > 3 a > 7 a. The relative abundance of Aromatoleum was increased with the protected cultivation time increase, and the values of both Group 3 a and 5 a were lower than ck.
nosZ community structure and its relationship with soil chemical properties Clustering analysisBy the PCA conducted for the nosZ community, it can be seen that the structure of the nosZ community in the soils with various protected cultivation years was signi cantly different (Fig. 5). The rst and second principal components interpreted 48.41% and 21.53% of the bacterial community structure variation, respectively, giving a cumulative contribution rate of 69.94%. Group 7 a was far from Group 3 a and 5 a on PC1 and PC2, indicating a large change of nosZ community structure as the cultivation time extended. Group 3 a and 5 a were close, indicating a high community similarity.
The redundancy analysis of nosZ community structure and soil chemical indicators were performed (Fig. 6). The physical and chemical properties of soil could explain 46.61% of all the information on the rst axis and 22.44% on the second axis for the nosZ community, giving a number of 69.05% total information explained by the two axes. The SOC and nitrate nitrogen had signi cant impacts on the nosZ community structure. The soil AP, AK and ammonium nitrogen had highly signi cant effects on nosZ community structure.

Discussion
Long-term protected cultivation of vegetable crops had a signi cant impact on soil chemical indicators. The TN, ammonium nitrogen and nitrate nitrogen were all increased signi cantly with the increase of protected cultivation time, and all of them were higher than ck. This may be due to that the intensive planting, continuous high fertilizer input and long-term continuous cropping of protected cultivation restricted the growth of vegetable crops, hindered their nutrient absorption, and thus most of the nutrients accumulated in the soil over time 18,19 . This study demonstrated that the SOC contents of various protected cultivation years were all lower than ck, decreasing year by year. This was contrary to the results of previous studies 20, 21 , but similar to the results of Song et al 22 . It may be due to the small amount of organic fertilizer applied and a small return of plant residues to soil in this area. In this study, the soil pH value was decreased as the cultivation year increased, consistent with the results of various studies 23,24 . Presumably, it is because of the soil acidi cation caused by the accumulation of SO 4 2− and NO 3 − from long-term excessive application of (nitrogen) fertilizers.
Denitri cation is an important part responsible for soil nitrogen loss, greenhouse gas N 2 O emission and N 2 conversion. Denitri cation is primarily driven by microorganisms and is closely related to soil fertilization 25 . In this study, the copy numbers of nosZ in the soil for various protected cultivation years were all signi cantly less than that of ck, and the copy number of nosZ was decreasing with the extension of planting time. It may be due to the high water and fertilizer input of protected cultivation that the reduction of N 2 O by soil microorganisms was declined in greenhouses 26 . Bergaust et al.found that the N 2 O reductase of Paracoccus was di cult to assemble at a lower pH, ultimately leading to a decrease in the enzyme activity. 27 Hallin et al. believed that the copy number of nosZ gene dropped rapidly with increased nitrogen fertilizer application rate 25 . In this study, the nosZ gene abundance had signi cant positive correlations with soil pH and SOC. Because denitrifying microorganisms are generally chemical heterotrophs, they need substances like organic matter as energy sources 28 . The study conducted by Matlou and Haynes showed that SOC was closely related to soil microbial community structure, and organic carbon was the necessary substrate and energy source for soil microbial activities 29 . In our study, the SOC content was decreasing with the increase of protected cultivation time, and thus it could not provide a su cient energy source for denitri ers. Bowden et al. believed that the accumulation of SOC was due to the reduction of microbial biomass 30 . Soil pH is the primary environmental factor responsible for variations in the number of denitrifying microorganisms 31 . As the protected cultivation time increased, the soil pH kept decreasing. Soil acidi cation causes the accumulation of iron and aluminum. Iron and aluminum hydroxides can reduce the utilization of carbon sources by microorganisms via adsorbing soil soluble carbons 32 . The stronger the soil acidity, the greater the accumulation of Al 3+ , and the severer the damage to the soil bacterial cell membrane 33 . The research by Bauhus and Khanna showed that the addition of phosphorus promoted soil denitri cation in acid forest soils 34 . In this study, the copy number of nosZ gene was signi cantly correlated with AP, further supporting the regulation of soil phosphorus on denitri ers. Nevertheless, the mechanism of how phosphorus regulates the growth of denitrifying microorganisms remains elusive. Nitrate nitrogen can serve as the substrate of denitri ers and the electron acceptor in denitri cation, thus affects the growth of denitrifying microorganisms 35 . The results of this study showed that the abundance of nosZ gene was highly signi cantly negatively correlated with soil TN, ammonium nitrogen and nitrate nitrogen. Comprehensive analysis suggested that changes in soil pH and nitrogen content were the primary factors driving the changes in nosZ gene copy number.
In this study, the -diversity index of nosZ denitrifying microbial community was signi cantly different among different treatments. Both Chao1 and ACE indexes were the highest in ck, and gradually decreased with the extension of planting time. It may be due to the long-term protected cultivation that the soil acidi cation and salinization became severer, resulting in decreased abundance of soil denitri er species. The Shannon's and Simpson's indices were higher in Group 3a and lower in Group 7a, indicating decreased soil denitrifying microbial species diversity with the planting time extension. It may be that the long-term protected cultivation led to a decrease in SOC content, which then could not provide the necessary carbon source for the growth of denitri ers.
The dominant group of nosZ denitri ers was exclusively Proteobacteria, consistent with the results of many studies 36 . The abundance of soil Proteobacteria can represent the nutrient level including soil organic matter to a certain extent, and Proteobacteria favor a higher pH soil environment 37 . In this study, the abundance of Proteobacteria showed a decreasing trend with the extension of planting time, and was the highest in ck and Group 3a. It may be related to the high soil nutrient content and high pH that are bene cial to the growth of Proteobacteria. At the genus level, the common dominant group was Brachyrhizobium, whose relative abundance accounted for 43% of the nosZ gene sequences on average. The relative abundances of Brachyrhizobium in Group 3a, 5a and 7a were all higher than ck, and it was decreasing with the extension of planting time. According to the PCA results, the nosZ community was clustered clearly, and Group 7a was well separated from Group 3a and 5a, while the latter two were clustered relatively closely. It indicated that the nosZ community structure varied greatly with the planting time increase for the protected cultivation eld.
Related researchers believe that the difference in the response of denitri ers to fertilization is shown in soil pH, TN, ammonium nitrogen and nitrate nitrogen 25 . Soil pH is considered to be the primary factor affecting the community structure of denitrifying microorganisms. pH can exert a selective effect on denitri ers, and its variation can affect the community structure of denitrifying microorganisms and further affect their response to environmental changes 31 . In this study, soil pH was signi cantly different in the protected cultivations for different years. In addition, soil pH was signi cantly correlated with the relative abundance of nosZ, indicating soil pH as the primary cause of changes in denitrifying microbial community structure. Usually, the SOC, ammonium nitrogen and nitrate nitrogen serve as direct or indirect substrates and energy sources for denitrifying microorganisms, and they may affect the community structure of soil denitri ers. Nitrate nitrogen is the substrate of denitri cation, and its content directly affects the utilization of nitrogen by denitrifying microorganisms 38 .
In conclusion, the nosZ gene abundance, Chao1 index and ACE index demonstrated the same trends that decreased as the protected cultivation time increased. The abundance of nosZ gene was not signi cantly different in different protected cultivation years. At the phylum level, Proteobacteria was the dominant group, and the most abundant in Group 3 a. At the genus level, Bradyrhizobium and Achromobacter were dominant groups. The relative abundance of Bradyrhizobium was decreased as the protected cultivation time increased, but was all higher than that of ck. The relative abundance of Achromobacter was the highest in Group 5 a, showing a "V" trend as the protected cultivation time increased. The total nitrogen, available potassium, ammonium nitrogen and nitrate nitrogen had great impacts on nosZ gene abundance, α-diversity index and community structure. It can be seen that nitrogen fertilizer was an important factor limiting the fast-growing and high-yield of vegetables in protected cultivation elds. Soil nitrogen is conducive to vegetable crop absorption and utilization, and can directly or indirectly regulate the denitrifying microbial community.

Site Description and Soil Sampling
The study area was located in Heping Town, Tongren City, Guizhou Province, China. It has a subtropical monsoon climate with an average annual temperature of 18 ℃, an annual precipitation of 1,313 mm, and a yellow soil type. Three greenhouses with different planting years were selected as the experimental groups, and the nearby outdoor plot was used as a control (ck). For greenhouse plots, the fertilizers The main planted vegetables were eggplants, cucumbers, zucchini, and cowpeas. The samples were collected in July 2017. The 0-10 cm top-soils were sampled following an "S" shape using the ve-point sampling method in each plot, mixed as one soil sample, stored in an ice box and transported to the laboratory as soon as possible. After removing stones and roots, the soil sample was passed through a 2-mm sieve.
The prepared soil sample was then divided into three parts, one part of fresh soil was used for the determination of ammonium nitrogen and nitrate nitrogen, one part was stored in a -80 ℃ freezer for the analysis of nosZ community structure and abundance, and one part was used for the determination of soil chemical indicators after air-dried indoors and sieved.

Soil Physicochemical Analysis
Soil pH was measured by the potentiometry method. Soil organic carbon content (SOC) was measured via the potassium dichromate oxidation method. Soil total nitrogen (TN) was determined via Kjeldahl method; Soil available phosphorus (AP) was measured via sodium bicarbonate extraction-molybdenum antimony colorimetric. Soil available potassium (AK) was measured via ammonium acetate extractioname photometry. Soil ammonium nitrogen (NH 4 + -N) was measured via indophenol blue colorimetry. Soil nitrate nitrogen (N0 3 − -N) was measured via disulfonic acid colorimetry 39 .
Soil DNA extraction, nosZ gene ampli cation and quantitative uorescence PCR (qfPCR) Soil DNA extraction was conducted using the E.Z.N.A.® soil DNA kit (OMEGA, GA, USA) according to the manufacturer's instruction, using approximately 0.5 g of fresh soil. The purity and integrity of the extracted DNA were assessed via 1% agarose gel electrophoresis, and the concentration and purity of the extracted DNA were evaluated with the help of a nucleic acid quantometer (NanoDrop ND-1000). The nosZ gene were ampli ed using a primer set: nosZ-F(5′-GGGCTBGGGCCRTTGCA-3′) and nosZ-R(5′-GAAGCGRTCCTTSGARAACTTG-3′) 40 . The volume of the PCR reaction mixture was 25 µL, containing 5.0 µL10 × ExTaq, 2.0 µL dNPT (2.5 µmol·L − 1 ), 1.0 µLforward and reverse primer (10 µmol·L − 1 ), 2.0 µLDNA template (1-10 ng), 0.25 µL ExTaq (5U·µL − 1 ) and balanced ultrapure water (ddH 2 O). The PCR ampli cation conditions were as follows: predegeneration for 5 min at 94°C, 32 cycles of ampli cation at 94°C for 30 s, 57°C for 45 s, and 72°C for 55 s, and extension at 72°C for 10 min 41 . After the PCR product was puri ed and recovered, it was ligated to the pMD8-T vector, and transformed into E. coli DH5α competent cells. After cultivation, the positive clones were picked and the nosZ gene recombinant plasmid was extracted. The plasmid concentration was determined by a nucleic acid quanti er, and the copy number of the gene was calculated. The ten-fold serial dilutions were prepared to a gene copy number of 10 3~1 0 8 , followed by standard curve preparation.

High-throughput Sequencing
After the ampli cation of the PCR products, the samples were sent to Shanghai Personalbio Technology Co., Ltd. for sequencing with an Illumina MiSeq System. After the raw reads were available, according to the barcodes sequences and primers sequences, the effective reads were screened. Then the barcodes and primers sequences were trimmed off. FLASH was used to merge the paired-end reads that passed the quality preliminary screening based upon their overlapping regions. At the same time USEARCH software was used to check and remove chimeras to obtain high-quality effective sequences. The uclust tool of QIIME was applied to align sequences, the operational taxonomic units (OTUs) were clustered at a similarity threshold of 97%, and the most abundant sequence would serve as the representative sequence of the corresponding OTU. QIIME was further used to compare the representative sequences of OUT with the Functional Gene Pipeline (FunGene) to obtain taxonomic information of each OTU.
Data analysis SPSS21.0 was used to conduct the statistical signi cance analysis and correlation analysis of soil chemical properties, α-diversity index and abundance of denitri ers, and relative abundance of community composition. R was used to perform principal component analysis (PCA) and redundancy analysis. Declarations