Microbial community structure from southern High Plains beef cattle feedyard manure and relationship with nitrous oxide emissions

Abstract

Modern molecular techniques enable characterization of the microbial biome in livestock manure, from which there is particular concern over emission of greenhouse gases. This study evaluated how sampling depth, time, temperature, and artificial rainfall affected microbial community structure in feedyard manure, and relationships between the manure biome and known parameters related to nitrous oxide (N₂O) emissions. In three large incubation chambers, maintained at different temperatures that received two applications of artificial rainfall, we evaluated manure microbiome composition and abundance of N₂O-producing enzymes (nirK and nirS) using quantitative polymerase chain reaction (qPCR). These data were used with previously published data from the same study on N₂O emissions and assessment of manure physicochemical properties, denitrification enzyme activity (DEA), and nitrification activity (NA). Microbiome composition was Firmicutes (50%), followed by 32% Actinobacteria, 11% Proteobacteria, 5% Bacteroidetes, 1% Chloroflexi, and small populations (<0.5%) of Planctomycetes, Deinococcus-Thermus, Gemmatimonadetes, Verrucomicrobia, Tenericutes, and other organisms. Average bacterial populations varied largely as a function of sampling depth and time. Firmicutes increases tended to coincide with high N₂O emissions. Overall, the largest change observed was increased Proteobacteria at 5–10 cm, where relative abundance increased from 10% (17.2 °C) to 24% (46.2 °C) over time and with increased temperature. Firmicutes and Actinobacteria predominated the microbial community of manure, but favorable conditions may lead to increases in Bacteroidetes, Proteobacteria, and Chloroflexi, which could influence N cycling and N₂O emissions from feedyards. Copy numbers of nirS at the beginning of the experiment were higher than nirK. Differences in concentrations of nirK and nirS indicated that denitrifying enzymes in feedyard manure, particularly nirS, were sensitive to environmental changes.

Abbreviations

AOA

ammonium oxidizing archaea

AOB

ammonium oxidizing bacteria

DEA

denitrifier enzyme activity

DNRA

dissimilatory nitrate reduction to ammonium

NA

nitrification activity

OM

organic matter

qPCR

quantitative polymerase chain reaction

1 INTRODUCTION

The manure in beef cattle feedyards is a significant source of nitrous oxide (N₂O) (Parker, Casey, et al., 2017; Redding et al., 2015; Waldrip et al., 2017, 2020; Woodbury et al., 2018), a potent greenhouse gas (GHG) implicated in climate change due to a global warming potential (GWP) of 265–298 carbon dioxide equivalents (CO₂e). An estimated 25% of total GHG (N₂O and methane [CH₄]) from beef production originates from animal pens and housing (Costa et al., 2014). Studies have investigated the effects of temperature, rainfall, and activities of enzymes involved in feedyard N₂O emissions (Ayadi et al., 2015; Waldrip et al., 2020; Woodbury et al., 2001, 2018), but little work has characterized the manure microbiome or presence of specific genes responsible for N₂O production. In 2011, livestock manure contributed an estimated 4.3–5.8 Tg N₂O-N yr⁻¹ globally (Syakila & Kroeze, 2011), which has been gradually and steadily increasing. Denitrification and nitrification are the two primary modes of N₂O production in soil, manure, sediments, and other substrates (Ayadi et al., 2015; Redding et al., 2015; Waldrip et al., 2017, 2020). In a previous work, we proposed that the majority of feedyard N₂O was derived from the top 5 cm of a feedyard manure pack and that it was likely derived from denitrification after rainfall (Waldrip et al., 2020). In the current study, we take this a step further and examine changes in the manure microbial community structure and abundances of genes involved in denitrification and nitrification. This data presented in this paper is one of the first phyla-level microbiome analyses of beef cattle manure in the Texas High Plains, where open-lot feedyards containing 100,000 or more cattle are common. In addition, the dairy industry is thriving in the region, with many open-lot dairies containing 3,000 or more milking cows.

In previous studies on manure N₂O emissions, we identified that substrate availability for microbial metabolism was highly related to feedyard manure N₂O emissions, denitrifier enzyme activity (DEA) and nitrification activity (NA) (Waldrip et al., 2017, 2020). Waldrip et al. (2017) conducted 15 nonflow-through–nonsteady-state measurement campaigns on two commercial Texas feedyards and developed predictive empirical models, where temperature and manure nitrate/nitrite (NO_x) and H₂O contents were positively related to measured N₂O emissions. In contrast, negative relationships were identified between N₂O and manure organic matter (OM), ammonia/ammonium (NH_x), dissolved organic carbon (DOC), and dissolved nitrogen (N) contents, as well as ultraviolet-visible (UV-vis) parameters related to OM complexity/availability. Differences in C and N availability for microbial growth and energy could impact microbial community structure, gene abundances, and activities of enzymes involved in denitrification and nitrification in feedyards. This study aimed to provide some answers to these questions.

Feedyard N₂O emissions are highly variable over both space and time (Aguilar et al., 2014; Parker, Casey, et al., 2017; Redding et al., 2015; Waldrip et al., 2016). This variability has been linked to temperature (Parker, Casey, et al., 2017, 2018; Waldrip et al., 2016, 2017, 2020), H₂O content (Aguilar et al., 2014; Parker, Casey, et al., 2017; Parker, Waldrip, et al., 2017, 2018), and manure characteristics (Ayadi et al., 2015; Liao et al., 2018; Waldrip et al., 2017, 2020; Woodbury et al., 2001). Both air temperature and precipitation range widely in the Texas Panhandle portion of the southern High Plains, where this study was conducted (−102.1, 35.2). This region is classified as semi-arid, with monthly average temperatures ranging from −7 °C in winter to 33 °C in summer. In addition to large seasonal changes, the region experiences extremes in daily temperatures, where it is not abnormal for summer temperature to change from 10 °C in the mornings, to above 39 °C by mid-afternoon. Precipitation in the region is sparse (average of 54 cm yr⁻¹) and occurs primarily as short, but often intense, rainfall episodes in the spring. Thus, the accumulated manure in open-lot cattle pens is subjected to both wet/dry cycling and temperature extremes.

Nitrous oxide is produced by numerous mechanisms, including nitrification, denitrification, coupled nitrification-denitrification, dissimilatory nitrate reduction to ammonium (DNRA), anaerobic ammonia oxidation (Anammox), and various forms of chemodenitrification (Butterbach-Bahl et al., 2013; Giles et al., 2012). However, attempts to discern the processes involved in feedyard N₂O production have been largely inconclusive. The complexity and heterogeneity of feedyard pen manure and changing weather in a system exposed to ambient conditions, makes it unlikely that one process would be responsible for all feedyard N₂O. Multiple processes could be working simultaneously or in tandem. This complicates development of effective mitigation strategies, which generally require process-level knowledge of N₂O production mechanisms. In commercial feedyards the OM-rich manure is a complex ecosystem containing microbes, extracellular enzymes, both older and newly excreted feces and urine, and varying H₂O content. Furthermore, as feedyard manure accumulates during the cattle finishing period, a layer structure forms with varied properties (e.g., density, porosity, OM complexity, NH_x and NO_x availability, microbial community composition, moisture content, and temperature) at different depths (Cole et al., 2009; Rice et al., 2007). It is unclear if N₂O production mechanisms differ with manure depth.

The fate of manure N depends on the consortia of microbial species present, activities of enzymes involved in OM degradation and N mineralization, manure chemistry, aeration (i.e., O₂ content), H₂O content, and substrate availability for microbial energy and respiration. Similar to soil observations, specific N₂O production pathways in manure vary with conditions and may occur consecutively or simultaneously in sequestered micro-sites within the manure pack (Azam et al., 2002; Lai and Denton, 2018; Wu et al., 2017). Some of the known N₂O production processes and genes involved are presented in Figure 1.

Core Ideas

• Relationships were established among the beef manure microbiome, nitrous oxide losses, and environmental conditions.
• The predominant manure microbiome was Firmicutes > Actinobacteria > Proteobacteria > Bacteroidetes > Chloroflexi > others.
• Average microbial populations varied with sampling depth and time. Increases in Firmicutes coincided with nitrous oxide emissions.
• Copy numbers of nirK and nirS differed with date and among chambers, indicating sensitivity of denitrifying enzymes to environmental changes.

FIGURE 1

Diagram of major processes, compounds and genes implicated in manure nitrous oxide (N₂O) emissions

In soils, denitrification and nitrification are the predominant pathways of N₂O production. Nitrification occurs under aerobic conditions and is the stepwise oxidation of NH₄⁺ or organic N to NO₃^– (Figure 1) by autotrophic bacteria and archaea, and heterotrophic fungi and bacteria. Nitrification requires enzymes for oxidation of NH₃ and hydroxylamine, which are encoded by amoABC and hao genes. Denitrification is a facultative respiratory pathway that occurs under anaerobic/anoxic conditions (or in anaerobic microsites) by a wide microbiome (Butterbach-Bahl et al., 2013; Jones et al., 2008; Shoun & Tanimoto, 1991; Zumft, 1997). In the absence of O₂, oxidized N species (i.e., NO_x, NO, and N₂O) are reduced and coupled to electron transport phosphorylation. Nitrous oxide is a frequent end-product if conditions are not optimal for nitrous oxide reductase (nos), which is the final enzymatic step of N₂O reduction to N₂ (Firestone et al., 1980). Complete denitrification generally requires a consortium of microbes. Denitrification steps and the genes responsible for enzyme synthesis have been well defined (Butterbach-Bahl et al., 2013; Ye et al., 1994). These are nitrate reductase (nar; NO₃^– → NO₂^–), nitrite reductase (nir; NO₂^– → NO), nitric oxide reductase (nor; NO → N₂O), and the previously mentioned nitrous oxide reductase (nos; N₂O → N₂) (Figure 1).

As N₂O production is largely biochemical, complex interactions exist among variables that affect microbial community structure and activity in manure. To date, no study has surveyed the microbiome of manure from open-lot beef cattle facilities and assessed the interaction among N₂O emissions, key variables, microbial populations, and enzyme activities. The objective of this study was to improve understanding of the composition of the microbial structure of beef manure and feedyard N cycling. We hypothesized that environmental conditions interact with manure properties dynamically, creating diverse microsites with differing N₂O-producing capacity. Once these interactions are better understood, targeted mitigation methods can be developed and evaluated to reduce feedyard N₂O-N losses and improve manure fertilizer N value.

2 MATERIALS AND METHODS

2.1 Large chamber incubation study

This work is a continuation of Parker et al. (2018) and Waldrip et al. (2020), who investigated N₂O emissions, manure physicochemical properties, and nitrification (NA) and denitrifier enzyme activities (DEA) in feedyard manure under different temperatures and following simulated rainfall. This report takes these works a step further to investigate how the manure microbiome responded and how specific microbial phyla and gene abundances are related to previously measured variables. For the basic chamber setup, manure that had accumulated during a typical 150-d finishing period was scraped from a pen at a commercial feedyard in the Texas Panhandle. The manure was a composite of unconsolidated, loose surface manure and deeper packed manure from cattle fed a steam-flaked corn (Zea mays L.)-based diet and housed at ∼15 m² per animal. Selected physicochemical parameters of the manure were: 55% OM; 341 mg NH_x kg⁻¹; 5.7 mg NO_x kg⁻¹; 2.6% Total N; 27.7% Total C; and C/N of 10.8. Feedyard manure in this region is typically dry and hard packed. The manure was 91% dry matter (DM; as collected) and was roughly ground (<0.64 mm) and stored indoors to avoid further exposure to rainfall, temperature changes, and loss of gaseous N. Manure (109-mm depth) was added to each of three 1 m² (surface area) chambers. Compacted native caliche (calcium carbonate soil; 89 mm) underlaid the manure in each chamber to simulate the relatively impermeable soil-manure interface under open-lot feedyard pens (Mielke et al., 1974; Miller et al., 2008). The manure was compacted to an approximate dry bulk density of 0.61 g cm⁻³ with a handheld tamper. The experimental period was 59 d in length and occurred in winter and spring of 2017. Chamber temperatures were controlled with 1.3 by 1.3 by 1.3 m “hotbox” material warmers (Model HB64-1440, Powerblanket) with digital temperature controllers and 12 V exhaust fans for venting. In addition, silicone heating pads with digital temperature controllers (ProTherm Industries, Inc.) were attached under each chamber. A single chamber (Chamber 1) did not have thermal regulation and was exposed to ambient temperatures, which varied diurnally and seasonally in an indoor facility without temperature control.

To simulate the periodic wet/dry cycling that occurs in the region, we applied two episodes of simulated rainfall by evenly applying distilled H₂O, equivalent to a 25-mm rainfall event, to the surface of the manure with a handheld watering can on 13 February (Day 1) and 17 March (Day 22) of 2017. Manures dried naturally between H₂O applications. From Day 1 to 21 the three chambers were maintained at: (a) 5.0^o (ambient), (b) 11.2^o, and (c) 17.2 °C to simulate winter and mornings when temperatures are low. At Day 22, the second 25 mm of simulated rainfall was applied, and the temperatures of the chambers were increased and maintained at: (a) 15.0^o (ambient, increased due to seasonal variation), (b) 38.1^o, and (c) 46.2 °C until Day 59. This change was intended to simulate normal and extreme regional temperatures.

2.2 Manure collection and analyses

Manure samples were collected in triplicate from chambers 2 h after H₂O application on Days 1 and 22, and then at 10:00 a.m. CST on Days 18 and 29. Manures were collected with a small trowel at depths of 0–5 cm and 5–10 cm by vertically inserting a 30 cm (height) by 7.6 cm (diameter) tin cylinder. Care was taken during sampling to avoid disturbance of the manure both inside (i.e., depth mixing) and near the cylinder (i.e., integrity of emitting surface). For each sample, manure from the top 5 cm was removed manually, placed in polyethylene bags, and then stored on dry ice. The 5-to-10-cm depth fraction was collected and stored in the same manner.

Basic manure properties were reported in detail in Waldrip et al. (2020). In brief, oxidation–reduction potential (Eh) was measured according to Brown et al. (2000) using a Pocket Pro ORP Tester (Hach Company). Manure pH was measured [1:10 (wt/wt) with deionized H₂O (pH 8.01)] with an Accumet XL250 pH meter and AccuCap combination pH electrode (Thermo Fisher Scientific). Manure DM was determined gravimetrically after drying overnight at 105 °C. Concentrations of OM were determined by Loss on Ignition at 500 °C. Total N and TC contents were measured with a varioMAX CN analyzer (Elementar Analysensyteme GmbH). The NH_x and NO_x were extracted from manure samples with 2.0 M potassium chloride (KCl) for 30 min and quantified colorimetrically with a SEAL Analytical AQ2 Discrete Analyzer (SEAL Analytical Inc.) and USEPA's Method 350.1 (NH_x) and Method 353.1 (NO_x). Analyses of DEA and NA were conducted and reported in Waldrip et al. (2020) based on the methods of Woodbury et al. (2001) and Ayadi et al. (2015).

2.3 Quantification of N cycling genes

Quantitative polymerase chain reaction (qPCR) was used to assess relationships between experimental parameters, targeted bacteria, and targeted bacterial functional genes. Standard qPCR was used to quantify amoA gene copy numbers for ammonium oxidizing archaea (AOA), ammonium oxidizing bacteria (AOB), and nirK and nirS (i.e., copper [Cu] and cytochrome cd1-type nitrite reductases, respectively). The qPCR assays were performed in triplicate, as described (Blaud et al., 2021).

3 RIBOSOMAL RNA SEQUENCING FOR MICROBIAL COMMUNITY ANALYSIS

For 16S rRNA sequencing, triplicate sampled manures from each chamber/depth/date were composited into one sequencing run for each chamber/depth/date. These days largely coincided with measured peaks or lags in N₂O activity (Figure 2). The DNA was extracted from 200 mg bulk manure (n = 25) with a Quick-DNA Fecal/Soil Microbe Miniprep Kit (Zymo Research) according to the manufacturer's protocol. Bead beating was done with a Precellys 24 (Bertin Instruments) for three 15 s cycles at 6,000 rpm. The DNA was eluted with 50 μl of DNA Elution Buffer and stored at −20 °C. Sequencing was conducted by MR DNA (Molecular Research LP). Bacterial 16S ribosomal RNA gene variable regions 1–3 were amplified using previously described primers (Dowd et al., 2008). An Ion Torrent PGM system (ThermoFisher Scientific) was used for sequencing according to manufacturer guidelines. Low quality reads were filtered from the dataset. Data were initially screened using Phred quality scores, which measure the identification quality of the nucleobases generated during sequencing. The Q25 sequences (indicating >99% base call accuracy) were retained for further processing using a proprietary analysis pipeline (www.mrdnalab.com). Following Q25 screening, data were further processed by removing barcodes, primer sequences, and sequences with fewer than 150 base pairs. Also removed were sequences with ambiguous base calls, homopolymer runs greater than 6 bp, chimeras (single DNA sequences originating from multiple transcripts or parent sequences), and singleton sequences. Operational taxonomic unit (OTU) clustering was performed at the 97% sequence similarity level (Capone et al., 2011; Eren et al., 2011; Swanson et al., 2011). Taxonomic assignments for each OTU were assigned using a BLASTn search against a curated database assembled from GreenGenes, RDPII, and NCBI (www.ncbi.nlm.nih.gov; DeSantis et al., 2006, http://rdp.cme.msu.edu).

FIGURE 2

Nitrous oxide-nitrogen (N₂O-N) fluxes from feedyard manure following simulated rainfall application to chambers held at differing temperatures. Simulated rainfall episodes occurred on Days 0 and 22. Arrows represent manure sampling points. Modified from Parker et al. (2018)

3.1 Nitrous oxide emission measurements

Headspace N₂O concentrations were measured from each chamber at 30-min intervals on the days of H₂O addition, and then daily for the remainder of the study. Details on the chamber system are available in Parker, Casey, et al. (2017); Parker, Waldrip, et al. (2017); and Parker et al. (2018). To measure emissions, chambers were fitted with a vented, portable lid that was moved among the chambers. Headspace air was recirculated from the sealed chamber with polyethylene tubing to a real-time N₂O analyzer (Model N2O/CO-30-EP Enhanced Performance, Los Gatos Research, Inc.). Concentrations of N₂O were recorded at 1-s intervals during 60-s measurement periods, with flux rates calculated from the slopes of N₂O concentrations vs. time using linear regression for 30-s periods. Emissions data were reported by Parker et al. (2018) and are presented in Figure 2. Chamber temperatures (± 0.1 °C) were monitored with thermistors (model no. ACC-SEN-SDIP, Acclima, Inc.) placed mid-depth in the manure of each chamber.

3.2 Statistical analyses

As the collected data were not sufficient for full statistical analyses, we present survey information from two manure depths (0–5 cm, 5–10 cm) in three chambers that were subjected to six temperatures (5.9, 11.2, 15.0, 17.2, 38.1, and 46.2 °C) and two H₂O applications (Days 0 and 22). There were four manure sampling days (Days 0, 18, 22, 29). Data for gene copy numbers were log-transformed prior to analyses due to non-normality of the data. All data were analyzed using PROC GLM (SAS Institute). Pearson's correlation coefficients were employed to determine positive and negative relationships among variables with Proc CORR (SAS Institute). Qualitative inferences were made among measured variables and changes in prevalent microbial phyla.

4 RESULTS AND DISCUSSION

4.1 Microbial community composition

This study evaluated effects of sampling depth, time, temperature, and H₂O content on microbial community structure in feedyard manure and relationships with N₂O emissions. Bacteria were grouped into 10 main phyla (including unknowns) (Table 1, Figure 3). Average bacterial populations varied largely as a function of sampling depth and time (Table 2). There were no chamber × time interactions for the majority of the manure bacterial community, with the exception of Bacteriodetes (P < .01): this was more pronounced at 5–10 cm than at 0–5 cm. Firmicutes tended to increase at Day 22 at both 0–5 and 5–10 cm in all chambers, which coincided with high N₂O fluxes following the second H₂O addition and increased temperatures (Figure 2). At the end of the sampling period (Day 29), there were increased proportions of Proteobacteria and Bacteroidetes at 5–10 cm in all chambers, from ∼7 to 20% (Proteobacteria) and ∼5 to 8% (Bacteroidetes), with a decrease in Firmicutes from 52 to 39%. In addition, there were increased proportions of Chloroflexi and decreases in Actinobacteria. There were also small increases in the proportion of “All Other” microbes over time at both depths. Overall, the largest change observed was increased Proteobacteria at 5–10 cm: the relative abundance of Proteobacteria increased from 10.4% (17.2 °C) to 24.1% (46.2 °C) over time and with increased temperature in Chamber 3. These data suggested that Firmicutes and Actinobacteria predominate the microbial community of manure; however, favorable conditions may lead to growth and expansion of Bacteroidetes, Proteobacteria, and Chloroflexi, which could influence N cycling and N₂O emissions from feedyards.

TABLE 1. Effect of sampling depth on bacterial phyla in feedyard manure. Population data were averaged across sampling times and chambers

	Proportion of total bacterial community
Phyla	0–5 cm	5–10 cm	Average	SEM	P value
Firmicutes	50.7	48.8	49.8	0.18	.18
Actinobacteria	32.3	31.9	32.1	0.49	.57
Proteobacteria	10.0	11.4	10.7	1.00	.32
Chloroflexi	1.10	1.20	1.10	0.07	.77
Bacteriodetes	4.60	5.30	4.90	0.24	.06
Planctomycetes	0.34	0.44	0.40	0.08	.38
Deinococcus thermus	0.10	0.11	0.10	0.02	.84
Gemmatimonadetes	0.09	0.11	0.10	0.02	.63
Verrucomicrobia	0.02	0.03	0.02	0.007	.17
Tenericutes	0.01	0.01	0.01	0.002	.92
Other	0.03	0.01	0.01	0.007	.05*
Firmicutes/Bacteriodetes ratio	11.0	9.20	–	–	–

• Note. SEM, standard error of the mean.
• ^*P value is statistically significant at P < .05.

FIGURE 3

Relative abundance of major bacterial phyla in feedyard manure (0–5 and 5–10 cm depths) incubated for 29 d. Artificial rainfall events were applied of Days 0 and 22. In addition, temperature increases occurred in all chambers on Day 22

TABLE 2. Effect of sampling depth and incubation time on bacterial phyla in feedyard manure. Population data were averaged across chambers. Days 0 and 22 were sampled 2 h after artificial rainfall

	Proportion of microbial community
	Depth										SEM
	0–5 cm					5–10 cm
	Day
Phyla	Initial	0	18	22	29	Initial	0	18	22	29
Firmicutes	49.5	54.8	47.3	55.2	46.7	49.5	52.6	50.8	51.2	40.4	2.05
Actinobacteria	32.9	32.2	33.9	32.7	29.9	32.9	33.3	32.4	32.9	27.8	1.04
Proteobacteria	9.7	6.8	10.2	7.4	15.6	9.7	7.6	9.6	9.6	20.4	2.13
Chloroflexi	1.1	1.1	1.2	1.1	1.1	1.1	1.0	1.1	1.1	1.6	0.16
Bacteriodetes	5.2	4.0	6.0	2.7	5.2	5.2	4.0	5.0	4.2	8.0	0.53
Planctomycetes	0.4	0.1	0.3	0.4	0.5	0.4	0.5	0.2	0.2	0.8	0.16
Deinococcus thermus	0.1	0.05	0.1	0.1	0.1	0.1	0.06	0.07	0.08	0.2	0.06
Gemmatimonadetes	0.1	0.04	0.07	0.1	0.1	0.1	0.07	0.06	0.06	0.2	0.05
Verrucomicrobia	0.02	0.05	0.01	0.01	0.04	0.02	0.02	0.01	0.01	0.09	0.001
Tenericutes	0.02	0.01	0.01	0.01	0.01	0.02	0.01	0.01	0.01	0.01	0.005
Other	0.01	0.04	0.02	0.01	0.08	0.01	0.01	0.01	0.01	0.02	0.01

The dominance of Firmicutes and Bacteriodetes in our collected samples was similar to that found in other ruminant manures (Bernhard & Field, 2000; Min, Solaiman, et al., 2014; Min, Wright, et al., 2014; Shanks et al., 2011). Both Firmicutes and Bacteriodetes have members that produce beneficial short-chain fatty acids (acetate, propionate, and butyrate) from relatively indigestible carbohydrates in the ruminant colon. Bacteroidetes isolated from agricultural soils, wastewater sludge, and feedyard manure express nosZ (Figure 1), although it is genetically divergent from that expressed by Proteobacteria (Jung et al., 2013). Bacteriodetes are of particular interest because the production of fatty acids in manure could provide a C source for nitrifiers and denitrifiers, leading to N₂O emissions. Proteobacteria, the third most common phyla observed in the manure at both depths, include pathogens as well as genera involved in N₂ fixation and transformation (e.g., Nitrosomonas, Bradyrhizobium, Nitrospira, Nitrosolobus, and Nitrobacter). Actinomycetes possess mycelia that grow in a web-like pattern similar to fungal hyphae, which increases direct contact with metabolic substrates. They are common in soil, sediments, and wastewater sludge and have denitrifiying capacity (Chèneby et al., 2000; Shoun & Tanimoto, 1991) and decompose complex OM, although they tend to exhibit slow growth and activity.

4.2 Nitrous oxide gene quantitative polymerase chain reaction results

Biological denitrification involves a stepwise reduction of N oxides associated with electron transport phosphorylation and production of NO, N₂O, and N₂ in most situations. The reduction of NO₂^– to NO by nitrite reductase (nirK and nirS) is the first step that distinguishes denitrifiers from nitrate-respiring bacteria, which do not reduce NO₂^– to a gaseous form. Results from the current study indicate that nirK and nirS genes might influence competition with aerobic microorganisms and emission rates of N₂O. It has been reported that different denitrifying organisms (Abou-Seada & Ottow, 1988; Tiedje, 1988) and denitrifying communities in soils (Coyne et al., 1989) show differences in parameters that may influence competition with aerobic heterotrophs and N₂O emission rates (e.g., oxygen threshold, C requirement, and enzyme kinetics). Hence, knowledge of the underlying composition and diversity of denitrifier communities may help us better understand and manage N cycling in manure.

In this study, log-transformed average copy numbers of 16S rRNA, amoA from AOA (AOA-amoA) and AOB (AOB-amoA), and nitrite reductase genes (nirK and nirS) from Chambers 1, 2, and 3 at several key periods of N₂O emissions are presented in Table 3. At Day 0, there were 10.26 copies of 16S rDNA, with no significant change with incubation time (P = .40), chamber (P = .62), or time × chamber (P = .34). The log-transformed copy numbers of AOA-amoA ranged from 7.69 (Day 0) to 8.08 (Chamber 3, Day 18) and did not change significantly during the incubation. Copy numbers of AOB-amoA were lower at 5.55 (Day 0), but similarly showed no significant change during incubation. However, average log-transformed AOB-amoA concentrations tended to increase significantly over time across all chambers (P = .07) and time periods (P = .05).

TABLE 3. Effect of sampling time and temperature on 16S rDNA and copy numbers of amoA from ammonia oxidizing archaea (AOA), ammonia oxidizing bacteria (AOB), and nitrite reductase enzymes (nirK and nirS) in feedyard manure. Model effects (time, chamber, and time × chamber) were significant at ^*P < .05 and ^**P < .01)

Time	Chamber no. (temperature, °C)	16S rDNA	AOA-amoA	AOB-amoA	nirK	nirS
		Log copy no.^a
0	Initial	10.48	7.69	5.55	7.89	9.21
0 (2 h)^b	1 (5.9)	10.10	7.83	5.39	7.95	9.21
	2 (11.2)	10.32	7.80	5.25	8.06	9.22
	3 (17.2)	10.27	7.84	5.47	7.79	9.11
18	1 (5.9)	10.25	7.75	5.20	7.71	8.99
	2 (11.2)	10.49	7.98	5.72	7.71	9.46
	3 (17.2)	10.45	8.08	5.19	8.14	9.52
22 (2 h)	1 (15.0)^c	10.40	7.97	5.37	8.19	8.97
	2 (38.1)	10.29	7.96	5.79	7.74	9.34
	3 (46.2)	10.25	7.97	5.73	7.92	9.31
29	1 (15.0)	10.29	7.94	5.38	8.20	9.57
	2 (38.1)	10.30	7.98	5.90	8.30	9.76
	3 (46.2)	10.33	7.82	5.87	8.23	9.65
	SEM	0.018	0.023	0.191	0.122	0.294
ANOVA	Time	0.40	0.39	0.16	0.08	0.01^**
	Chamber	0.62	0.67	0.13	0.07	0.05^*
	Time × chamber	0.52	0.50	0.53	0.45	0.47
Averages
Time	0	10.26	7.82	5.37	7.94	9.17
	18	10.39	7.93	5.36	8.01	9.32
	22	10.31	7.97	5.63	7.94	9.20
	29	10.30	7.91	5.72	8.24	9.66
	SEM	0.054	0.061	0.125	0.085	0.082
	P value	.10	.11	.07	.02^*	.001^**
Chamber	1	10.28	7.87	5.33	7.90	9.18
	2	10.34	7.93	5.67	8.17	9.44
	3	10.32	7.93	5.56	8.03	9.40
	SEM	0.047	0.053	0.109	0.074	0.071
	P value	.34	.45	.05^*	.02^*	.02^*

• Note. AOA, ammonia oxidizing archaea; AOB, ammonia oxidizing bacteria; nirK, copper nitrite reductase; nirS, cadmium nitrite reductase; 16S, 16S rRNA amplicon pyrosequencing; SEM, standard error of the mean.
• ^a All values were log transformed prior to analysis. ^bDays 0 and 22 were sampled 2 h after artificial rainfall. ^cChanges in temperature are denoted in bold.

Copy numbers of nirS at the beginning of the experiment were higher (9.21) than nirK (7.89), indicating that the rate-limiting NO₂^– reduction step was dominated by bacteria, which produce both nirK and nirS, vs. fungi which only possess nirK. Soares et al. (2016) found no relationship between N₂O losses from a tropical soil and abundance of denitrification genes (nirK, nirS, nosZ); however, there was a positive correlation with amoA that suggested that nitrification by AOB was the main contributors to N₂O production. Differences in concentrations of nirK and nirS were slightly to highly significant (P ≤ .08–.01) with time and chamber, indicating that denitrifying enzymes in feedyard manure, particularly nirS, were sensitive to environmental changes. Results indicated that nirK and nirS copy numbers might influence competition with aerobic microorganisms and N₂O emission rates. In soils, it has been reported that different denitrifiers and denitrifying communities have dissimilarities in oxygen threshold, C requirements, and kinetic parameters (Coyne et al., 1989; Tiedje, 1988). These preferential differences may influence competition with aerobic heterotrophs and N₂O emissions. Understanding the underlying composition and diversity of denitrifier communities may improve understanding and management of feedyard N cycling.

In a review on the soil N cycle, Hanke and Strous (2010) identified relevant bacterial phyla and processes involved in N cycling. The denitrifiers identified included Alpha-, Beta-, Gamma-, Delta-, and Epsilon-Proteobacteria, and Nitrospira. In addition, Bacillus can contain nosZ and have been classified as denitrifiers (Jones et al., 2008). Jones et al. (2008) examined pH effects on denitrification kinetics in soil Bacillus isolates and found that N₂O production tended to begin after 10 h of incubation with some, but not all, Bacillus strains if pH was 6.0, but N₂ was the predominant end-product at pH 7.0. These researchers speculated that an additional nosZ gene copy, expressed under different conditions, provided a competitive advantage and/or both genes were constitutively expressed under similar dentitrifying conditions.

The growth of NH_x oxidizers and denitrifiers are inhibited by temperatures >40 °C, particularly in bacterial and archaeal communities (Xu et al., 2017). In contrast, fungal communities, including nirK-producers, are more resistant to high temperatures and may proliferate at temperatures ranging from 10 to 40 °C in manure-amended soils. Thus, the N₂O-producing microbial community changes depending on environmental temperature. Lin et al. (2018) found that fertilization with swine manure slightly reduced AOA-amoA copy numbers compared with unfertilized control, but increased AOB. In addition, they found no relationship between AOA abundance and NA in swine manure-amended agricultural soils. In contrast, AOB, particularly Nitrospira Cluster 8a, had a positive relationship and was important in nitrification following fertilization with both organic and inorganic sources. Lin et al. (2018) also noted that AOB (primarily Proteobacteria including Betaproteobacteria and Gammaproteobacteria) abundance were strongly related to NO₃–N concentration and soil pH. Under aerobic conditions, biological NH_x oxidation is well documented in both natural and engineered systems. Both AOA and AOB have been found to be involved in composting cattle manure (Yamamoto et al., 2010, 2012; Zeng et al., 2011); however, some studies (Posmanik et al., 2014; Yamada et al., 2013) identified that NH_x oxidation and nitrification in manure composts was related to AOB (Nitrospira and Nitrosomonas), but not AOA. Yamada et al. (2013) attributed the difference from other studies to high NH_x concentrations (>500 mg N kg⁻¹), as well as high NaCl and H₂O contents at the beginning of composting. This points to the importance of archaea in feedyard N₂O emissions, as AOA were identified as important NH_x oxidizers during the thermophilic phase (>50 °C) of composting (Jarvis et al., 2009; Yamamoto et al., 2012; Zeng et al., 2011) and in the digestion of poultry manure (Posmanik et al., 2014).

Although beyond the scope of the current study, another anaerobic N₂O-producing pathway is DNRA, which is carried out by obligate and facultative anaerobes, including Clostridium, Desulfovibrio, and Enterobacteridae, aerobes (Bacillus, Pseudomonas, Arthrobacter, Nitrobacter), and Ascomycota fungi (Rütting et al., 2011; Tiedje, 1988; Zhou et al., 2002). In soils, DNRA is regulated by NH₄⁺ and organic N concentrations and occurs under the same conditions as respiratory denitrification (Rütting et al., 2011). Typically, N₂O from DNRA is distinguished from denitrification-derived N₂O by isotope tracer studies. Although DNRA has not been found to be a primary process for N₂O in soils, it is an important pathway in situations where the ratio of available C to electron acceptors (i.e., NO₃^–) is high, such as sediments, sludge digesters, and the bovine rumen (Kaspar & Tiedje, 1981; Shu et al., 2016). Thus, DNRA could be involved in N₂O production from anoxic zones deeper in the manure pack, where older OM has undergone more decomposition/fermentation to form available C and NO₃^–. Further study is required to explore the role of DNRA in feedyard emissions.

4.3 Correlation coefficient (R) matrix

To further understand the effects of manure properties on N₂O emissions, microbiome changes, and nitrification and denitrification activities, measured parameter values were correlated against N₂O production (data not shown). Manure depth, DEA activity, and AOA and AOB copy numbers were not associated with N₂O production. However, manure temperature (R² = .63; P < .001), NH_x (R² = .66; P < .001), NO_x (R² = .73; P < .001), and nitrite reductase genes (nirK and nirS; R² = .57; P < .01) were positively related, and Eh (R² = −.39; P < .05), pH (R² = −0.34; P < .05), DM (R² = −.58; P < .01), and NA (R² = −.57; P < .01) were negatively related to N₂O production. Firmicutes, Actinobacteria, Proteobacteria, Bacteroidetes, Chloroflexi, Planctomycetes, Deinococcus, and Gemmatimonadetes had strong correlations with nirK and nirS gene copy numbers (R = .50–.79; P < .01–.001). The AOA-amoA and bacterial phylum (Chloroflexi, Planctomycetes, Deinococcus; R= .40–.68; P < .05–.01) and AOB-amoA and bacterial phylum (R = .41–.90; P < .05–.001; Firmicutes, Actinobacteria, Proteobacteria, Bacteroidetes, Chloroflexi, Planctomycetes, Deinococcus, and Gemmatimonadetes) were highly correlated. In addition, ratios of dissolved C/N in manure were negatively associated with NA (R = −.51; P < .05), Firmicutes (R = .49; P < .01), Actinobacteria (R = .40; P < .05), Proteobacteria (R = .41; P < .01), Bacteroidetes (R= −.46; P < .05), AOB-amoA (R = .47; P < .01), and nitrite reductase gene copy numbers (nirK and nirS; R = .44–.49; P < .05–.01). These results are consistent with other data (Waldrip et al., 2017), where an empirical model was developed to predict N₂O emissions based on temperature and manure NO_x and H₂O contents. In contrast, negative relationships were identified between N₂O and manure OM, NH_x, DOC, and dissolved N, as well as UV-vis parameters related to OM complexity/availability. Differences in C and N availability for microbial growth and energy could be involved in microbial community structure, gene abundances, and activities of enzymes related to N₂O emissions from feedyards. It is still unclear if N₂O production mechanisms differ with manure depth, but manure depth in the present study was negatively correlated (R² = −.75–.97; P < .001) with denitrification.

5 CONCLUSIONS

This study evaluated how sampling depth, time, temperature, and H₂O content affected microbial community structure in feedyard manure, and relationships between the manure biome and known parameters related to N₂O emissions. The 16s ribosomal RNA gene sequencing was used to assess microbial community structure. The majority of the microbial community was comprised of Firmicutes (50%), followed by 32% Actinobacteria, 11% Proteobacteria, 5% Bacteroidetes, 1% Chloroflexi, and small populations (<0.5%) of Planctomycetes, Deinococcus-Thermus, Gemmatimonadetes, Verrucomicrobia, Tenericutes, and other organisms. Average bacterial populations varied largely as a function of sampling depth and time. There were no chamber × time interactions for the majority of the manure bacterial community, with the exception of Bacteriodetes (P < .01): this was more pronounced at 5–10 cm than at 0–5 cm. Firmicutes tended to increase at Day 22 at both 0–5 and 5–10 cm in all chambers, which coincided with high N₂O fluxes following the second H₂O addition and increased temperatures. Towards the end of the experiment, there were increased proportions of Proteobacteria and Bacteroidetes at 5–10 cm in all chambers, from ∼7 to 20% (Proteobacteria) and ∼5 to 8% (Bacteroidetes), with a decrease in Firmicutes from 52 to 39%. In addition, there were increased proportions of Chloroflexi and decreases in Actinobacteria. There were also small increases in the proportion of “All Other” microbes over time at both depths. Overall, the largest change observed was increased Proteobacteria at 5–10 cm: the relative abundance of Proteobacteria increased from 10.4% (17.2 °C) to 24.1% (46.2 °C) over time and with increased temperature in Chamber 3. These data suggested that Firmicutes and Actinobacteria predominate the manure biome; however, favorable conditions may lead to growth and expansion of Bacteroidetes, Proteobacteria, and Chloroflexi, which could influence N cycling and N₂O emissions from feedyards and other open-lot cattle systems.

ACKNOWLEDGMENTS

The authors wish to thank Heather Robbe, Lana Castleberry, Kathleen Webb, and Will Willis for technical assistance for the current study. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

AUTHOR CONTRIBUTIONS

Heidi Waldrip: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Supervision; Visualization; Writing-original draft; Writing-review & editing. David Parker: Conceptualization; Investigation; Methodology; Project administration; Supervision; Writing-review & editing. Sierra Miller: Data curation; Formal analysis; Investigation; Writing-review & editing. Lisa M. Durso: Data curation; Formal analysis; Investigation; Validation; Writing-review & editing. Daniel N. Miller: Conceptualization; Formal analysis; Investigation; Methodology; Resources; Writing-review & editing. Byeng R Min: Formal analysis; Writing-original draft; Writing-review & editing. Kenneth Casey: Investigation; Methodology; Project administration; Validation; Writing-review & editing. Mindy J. Spiehs: Validation; Writing-review & editing. Bryan Woodbury: Conceptualization; Investigation; Methodology; Validation.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.