1 INTRODUCTION

The nitrogen (N) cycle is one of the most important biogeochemical cycles on earth, and biogeochemical processes involving N control its form and availability (Lehnert et al., 2021; Zhang et al., 2020). The N cycle is a collection of N-transforming reactions that are controlled primarily by micro-organisms which carry the appropriate enzymes (Kuypers et al., 2018). Biogeochemical N cycle is often attributed to the following seven distinct N-transforming processes: assimilatory N reduction, nitrification, denitrification, anaerobic ammonium oxidation (Anammox), N fixation, dissimilatory N reduction to ammonia/ammonium (DNRA) and ammonification (Figure S1) (Wang et al., 2019). Although there are a number of genes encoding the enzymes involved in these seven processes, only a selected set, that are essential, has been commonly adapted to detect these processes in environmental studies (Berges & Mulholland, 2008; Jetten, 2008; Kuypers et al., 2018; Stein, 2019; Zhang et al., 2020), including nasA/narB (assimilatory nitrate reductase) and nirA/nirB (assimilatory nitrite reductase) for assimilatory N reduction; amoA (ammonia/ammonium monooxygenase), hao (hydroxylamine oxidoreductase), ncyA (nitric oxide oxidoreductase), nxrA (nitrite oxidoreductase) for nitrification; narG (membrane-bound nitrate reductase), nirS/nirK (haem-containing/copper-containing nitrite reductases), norB (cytochrome c-dependent nitric oxide reductases), nosZ (quinol-dependent nitrous oxide reductase) for denitrification; hzsA (hydrazine synthase), hzo (hydrazine oxidoreductase) and hdh (dehydrogenase) for Anammox; nifH (molybdenum-iron nitrogenase) for N fixation; napA (periplasmic dissimilatory nitrate reductase), nrfA (dissimilatory periplasmic cytochrome c nitrite reductase) for DNRA; ureC (urease) and gdhA (glutamate dehydrogenase) for ammonification. In addition, the recent discovery of complete ammonia oxidizers, comammox Nitrospira which perform both steps of nitrification, has significantly changed our perspective on the nitrification. The complete genomes of comammox organisms possess the full set of genes encoding ammonia monooxygenase (AMO) and hydroxylamine dehydrogenase for ammonia oxidation, and genes encoding nitrite oxidoreductase for nitrite oxidation, indicating their genetic potential to utilize both ammonia and nitrite as energy sources (Daims et al., 2015; Li et al., 2022; van Kessel et al., 2015).

A better understanding of how nitrogen cycling genes are involved in ecological processes is one of the crucial areas of microbial ecological research (Tu et al., 2019). A number of previous research studies have attempted to investigate the processes of N cycling through molecular biological methods based on primers/probes of the related functional genes in environmental samples. For example, quantitative polymerase chain reaction (qPCR) has been used to investigate the amoA gene in wastewater (Gao et al., 2013; Zhang et al., 2015), and the nirK and nirS genes (Lee & Francis, 2017) and hzsA gene (Bale et al., 2014) in marine environments. High-throughput amplicon sequencing was used to investigate the amoA gene in crop soil (Yang et al., 2017), and nirk and nirS genes in a reservoir (Zhou et al., 2016). Moreover, droplet digital PCR (ddPCR) was used for quantification of the nifH gene from the surface ocean (Delmont et al., 2018). Many specific primer sets were designed and evaluated for detecting the amoA gene of comammox in the environments. Several published primers generated non-specific PCR products or did not amplify target genes from lake water and other habitats. Quantification based on ddPCR with a newly designed comammox-specific primer set revealed very low abundances of comammox amoA genes in lake water samples (Harringer & Alfreider, 2021). Therefore, the accuracy of the utilized primers/probes is important for all of these methods. PCR primers are designed based on the sequence information available at the time. Gene databases are constantly being enriched, particularly with huge genome and metagenome information, so the PCR primers should be validated accordingly and updated as needed. The evaluation of primer/probe accuracy is tightly tied to the quality and completeness of an N cycling functional gene database.

Over the past decades, many databases related to N cycling functional genes have been developed. Among these, databases such as the nifH gene database (Gaby & Buckley, 2014) and another nifH sequence collection that is available from the Zehr research group (https://wwwzehr.pmc.ucsc.edu/nifH_Database_Public/), FunGene (Fish et al., 2013) and NCycDB (Tu et al., 2019) have been used for N cycling studies. However, these databases have their own advantages and limitations owing to the different design concepts (Tu et al., 2019). The nifH gene database is for phylogenetic and evolutionary analyses, the design and assessment of primers/probes, and the evaluation of nitrogenase sequence diversity (Gaby & Buckley, 2014), while Gaby's and Zehr's nifH sequence collection did not include other N cycling genes. FunGene offers databases of many common eco-functional genes and proteins and includes eight categories of functional genes (e.g. biodegradation, antibiotic resistances) (Fish et al., 2013). However, FunGen contains only partial coverage of the gene families involved in the N cycle (21 gene [sub]families involved in five N cycling processes). NCycDB contains a total of 68 gene (sub)families, including those for seven N cycle processes, and can be used to annotate genes in metagenomic research (Tu et al., 2019). Since using high-quality sequences from a small number of sequences may lead to a large bias in coverage and specificity checking, the primer evaluation and design procedures usually require all currently available sequences to be comprehensively collected. Therefore, none of the above databases are capable of achieving primer design and evaluation.

In this study, we first constructed a comprehensive sequence database for N cycling, called NcycFunGen, which contained 607,359 paired nucleotide and protein sequences with their taxonomic information. Thereafter, 608 published N cycling gene primers were assessed with the associated pipeline, and the primer quality report of each gene was integrated as a primer database. Users are able to compare their own primers against alternative primers in the primer database, allowing for the selection of high-quality primers for environmental investigation. We also designed new primer pairs for ureC, bacterial and archaeal amoA, and nifH genes with higher coverage based on NcycFunGen. To verify the effectiveness of the new primer based on the difference in outcome from the existing primer, the new and previous ureC and nifH gene primer pair was applied to the urea amendment samples by ddPCR and/or high-throughput amplicon sequencing. Our database and evaluated primers are expected to facilitate the detection and characterization of the N cycle in diverse environments.

中文翻译：

---

环境中氮循环基因检测引物的评价与重新设计

1 简介

氮 (N) 循环是地球上最重要的生物地球化学循环之一，涉及 N 的生物地球化学过程控制着其形式和有效性（Lehnert 等，  2021；Zhang 等，  2020）。N 循环是 N 转化反应的集合，主要由携带适当酶的微生物控制（Kuypers 等人，  2018 年）。生物地球化学 N 循环通常归因于以下七个不同的 N 转化过程：同化 N 还原、硝化、反硝化、厌氧氨氧化（Anammox）、N 固定、异化 N 还原为氨/铵（DNRA）和氨化（图 S1） （王等人，2019）。尽管有许多基因编码参与这七个过程的酶，但在环境研究中，只有一组必要的选定基因通常适用于检测这些过程（Berges & Mulholland,  2008 ; Jetten,  2008 ; Kuypers et al. .,  2018 ; Stein,  2019 ; Zhang et al.,  2020 )，包括用于同化 N 还原的nasA / narB（同化硝酸盐还原酶）和nirA / nirB （同化亚硝酸盐还原酶）；amoA (氨/铵单加氧酶), hao (羟胺氧化还原酶), ncyA（一氧化氮氧化还原酶），nxrA（亚硝酸盐氧化还原酶）用于硝化；narG（膜结合硝酸盐还原酶）、nirS / nirK（含血红素/含铜亚硝酸盐还原酶）、norB（细胞色素c依赖性一氧化氮还原酶）、nosZ（喹啉依赖性一氧化二氮还原酶）用于反硝化；Anammox 的 hzsA（肼合酶）、hzo（肼氧化还原酶）和hdh（脱氢酶）；nifH（钼铁固氮酶）用于固氮；napA（周质异化硝酸盐还原酶），nrfA（异化周质细胞色素c亚硝酸还原酶）用于 DNRA；ureC（脲酶）和gdhA（谷氨酸脱氢酶）用于氨化。此外，最近发现的完全氨氧化剂 comammox Nitrospira可同时执行两个硝化步骤，这极大地改变了我们对硝化的看法。comammox 生物的完整基因组具有编码氨单加氧酶 (AMO) 和用于氨氧化的羟胺脱氢酶的全套基因，以及编码用于亚硝酸盐氧化的亚硝酸氧化还原酶的基因，表明它们具有利用氨和亚硝酸盐作为能源的遗传潜力（Daims 等人）。等人，  2015 年；李等人，  2022 年; 范凯塞尔等人，  2015 年）。

更好地了解氮循环基因如何参与生态过程是微生物生态研究的关键领域之一（Tu et al.,  2019）。先前的一些研究试图通过基于环境样品中相关功能基因的引物/探针的分子生物学方法来研究N循环的过程。例如，定量聚合酶链反应 (qPCR) 已用于研究废水中的amoA基因（Gao 等人，  2013 年；Zhang 等人，  2015 年），以及nirK和nirS基因（Lee & Francis，  2017 年）和hzsA基因 (Bale et al.,  2014) 在海洋环境中。高通量扩增子测序用于研究作物土壤中的amoA基因（Yang et al.,  2017），以及水库中的nirk和nirS基因（Zhou et al.,  2016）。此外，液滴数字 PCR (ddPCR) 被用于从表层海洋中定量nifH基因 (Delmont et al.,  2018 )。设计和评估了许多特定的引物组来检测amoA环境中的comammox基因。一些已发表的引物产生了非特异性 PCR 产物，或者没有扩增来自湖水和其他栖息地的靶基因。使用新设计的 comammox 特异性引物组基于 ddPCR 的定量显示湖水样品 中 comammox amoA基因的丰度非常低（Harringer & Alfreider， 2021）。因此，所用引物/探针的准确性对于所有这些方法都很重要。PCR引物是根据当时可用的序列信息设计的。基因数据库不断丰富，特别是包含大量基因组和宏基因组信息，因此应相应地验证 PCR 引物并根据需要进行更新。引物/探针准确性的评估与 N 循环功能基因数据库的质量和完整性密切相关。

在过去的几十年中，已经开发了许多与 N 循环功能基因相关的数据库。其中，数据库如nifH基因数据库 (Gaby & Buckley,  2014 ) 和Zehr 研究组提供的另一个nifH序列集合 (https://wwwzehr.pmc.ucsc.edu/nifH_Database_Public/)、FunGene (Fish et al.,  2013 ) 和 NCycDB (Tu et al.,  2019 ) 已被用于 N 循环研究。然而，由于设计理念不同，这些数据库各有优势和局限性（Tu et al.,  2019）。尼夫H基因数据库用于系统发育和进化分析、引物/探针的设计和评估，以及固氮酶序列多样性的评估（Gaby 和 Buckley，  2014 年），而 Gaby 和 Zehr 的nifH序列集合不包括其他 N 循环基因。FunGene 提供许多常见生态功能基因和蛋白质的数据库，包括八类功能基因（例如生物降解、抗生素抗性）（Fish 等，  2013）。然而，FunGen 仅包含参与 N 循环的基因家族的部分覆盖（涉及五个 N 循环过程的 21 个基因 [亚] 家族）。NCycDB共包含68个基因（亚）家族，包括7个N循环过程的家族，可用于宏基因组研究中的基因注释（Tu et al.,  2019）。由于从少量序列中使用高质量序列可能会导致覆盖率和特异性检查的较大偏差，因此引物评估和设计程序通常需要全面收集所有当前可用的序列。因此，上述数据库均无法实现引物设计和评估。

在这项研究中，我们首先构建了一个用于 N 循环的综合序列数据库，称为 NcycFunGen，其中包含 607,359 对核苷酸和蛋白质序列及其分类信息。此后，608条已发表的N循环基因引物与相关管道进行评估，并将每个基因的引物质量报告整合为引物数据库。用户可以将自己的引物与引物数据库中的替代引物进行比较，从而选择高质量的引物进行环境调查。我们还为ureC、细菌和古细菌amoA和nifH设计了新的引物对基于 NcycFunGen 的具有更高覆盖率的基因。为了根据与现有引物的结果差异验证新引物的有效性，通过 ddPCR 和/或高通量扩增子测序将新的和以前的ureC和nifH基因引物对应用于尿素修正样品。我们的数据库和评估的引物有望促进不同环境中 N 循环的检测和表征。