Dynamics of a Pasture Soil Microbial Community after Deposition of Cattle Urine Amended with [13C]Urea

Within grazed pastures, urine patches are hot spots of nitrogenturnover, since dietary N surpluses are excreted mainly as ureain the urine. This short-term experiment investigated ¹³C uptakein microbial lipids after simulated deposition of cattle urineat 10.0 and 17.1 g of urea C m^–2. Confined field plotswithout or with cattle urine amendment were sampled after 4and 14 days, and soil from 0- to 5-cm and 10- to 20-cm depthswas analyzed for content and composition of phospholipid fattyacids (PLFAs) and for the distribution of urea-derived ¹³C amongindividual PLFAs. Carbon dioxide emissions were quantified,and the contributions derived from urea were assessed. Initialchanges in PLFA composition were greater at the lower levelof urea, as revealed by a principal-component analysis. At thehigher urea level, osmotic stress was indicated by the dynamicsof cyclopropane fatty acids and branched-chain fatty acids.Incorporation of ¹³C from [¹³C]urea was low but significant,and the largest amounts of urea-derived C were found in commonfatty acids (i.e., 16:0, 16:1 ${omega}$ 7c, and 18:1 ${omega}$ 7) that would be consistentwith growth of typical NH₄⁺-oxidizing (Nitrosomonas) and NO₂^–-oxidizing(Nitrobacter) bacteria. Surprisingly, a 20 ${per thousand}$ depletion of ¹³Cin the cyclopropane fatty acid cy17:0 was observed after 4 days,which was replaced by a 10 to 20 ${per thousand}$ depletion of that in cy19:0after 14 days. Possible reasons for this pattern are discussed.Autotrophic nitrifiers could not be implicated in urea hydrolysisto any large extent, but PLFA dynamics and the incorporationof urea-derived ¹³C in PLFAs indicated a response of nitrifierswhich differed between the two urea concentrations.

INTRODUCTION

Top

Introduction

Urine deposition is an important source of N in grazed pastures.Nitrogen turnover in urine patches is intense, and the potentialfor environmental losses is significant, depending on urinecomposition (31, 49). Excess N in the diet of cattle is mainlyexcreted as urea in the urine (23), and understanding the regulationof urea turnover in urine patches is therefore of particularinterest. Urea typically disappears from the soil solution within24 to 48 h and is replaced by high concentrations of NH₄⁺ (18,36). The accompanying increase in osmotic pressure and freeNH₃ concentrations can lead to N₂O emissions and other microbialstress reactions (12, 40).

Urea hydrolysis to NH₄⁺ and CO₂ may occur both on plant surfacesand in the soil (20). Hydrolysis takes place through the actionof free or colloid-bound extracellular ureases or intracellularlyfollowing microbial uptake (20, 39). Klose and Tabatabai (26)determined extracellular and intracellular urease activitiesto be on average 54 and 46% of the total potential activityin various cropping systems. Nielsen et al. (33) found thaturea turnover was comparable in magnitude to gross N mineralizationand hypothesized that direct uptake and intracellular metabolismof urea are quantitatively important aspects of N cycling inagricultural soil. Different bacteria have shown a capacityfor urea uptake, including Klebsiella and Alcaligenes spp. (21)and autotrophic NH₄⁺-oxidizing bacteria (2).

Stable-isotope analysis of microbial lipids has recently beenintroduced in community-level studies. Even though isotope fractionationduring lipid biosynthesis is a matter of concern (5), this approachcan potentially link specific substrates with individual populationsin situ (6, 16, 22). Substrate-derived ¹³C has been traced inmembrane lipids of heterotrophic (1, 7) and autotrophic (27)organisms. Further, ¹³C enrichment of phospholipid fatty acid(PLFA) profiles has been used to investigate microbial communitydynamics (9).

Here we report on the turnover of [¹³C]urea in pasture soiland the associated incorporation of ¹³C in PLFAs after simulateddeposition of cattle urine at two different urea concentrations.It was hypothesized that the dynamics and labeling of PLFAscould provide information about the fate of urea and the responseof nitrifiers to the sudden increase in N availability.

MATERIALS AND METHODS

Top

Materials and Methods

Site information.
The field experiment was carried out between 20 September and4 October 2001. The site was an 8-year-old grass-clover pasturenear the Danish Institute of Agricultural Sciences, Tjele, Denmark(55°52'N, 9°34'E). The soil was a sandy loam and wasclassified as a Typic Hapludult. It contained 2.7% C and 0.18%N, the

was 5.5, and total cation exchange capacity was 8.7 cmol kg^–1. Air and soil temperatures(10-cm depth) averaged 12.8 and 13°C, respectively, duringthe experiment, and total precipitation was 56 mm (27 mm withinthe first 48 h).

Experimental design.
The experimental design contained three randomized blocks, eachwith three treatments, i.e., a control with no amendments (CTL),standard cattle urine (LU), and urine with approximately doublethe urea concentration (HU). Urine was collected from dairycows during milking 3 days before start of the experiment andwas stored at 2°C until used; the urea content was determinedprior to storage. On day 0 of the experiment, the cattle urinewas amended with [¹³C]urea (99 atom%; CK Gas Products, Hook,United Kingdom) and unlabeled urea (HU treatment only), to giveconcentrations of 12.0 and 20.4 g of urea liter^–1 with[¹³C]urea concentrations of 11.0 atom% (LU) and 18.3 atom% (HU).The labeled urine was applied to 25- by 35-cm frames, installedto a 15-cm depth, at rates corresponding to 10.0 and 17.1 gof urea C m^–2 in for the LU and HU treatments, respectively.

Sampling.
Emissions of CO₂ were determined after 0.2, 1, 2, 4, 6, and14 days. Insulated chambers, equipped as previously described(35), were mounted on top of the permanent frames. Gas samples(13 ml) were taken in preevacuated Exetainers (Labco Ltd., HighWycombe, United Kingdom) after 5, 15, and 45 min and subsequentlyanalyzed for total concentrations and isotopic composition ofCO₂.

Soil sampling took place after 4 and 14 days. Three soil cores(0 to 20 cm in depth; 2 cm in diameter) were collected and thedepth intervals 0 to 5, 5 to 10, and 10 to 20 cm were pooledin separate bags. The samples were stored at 2°C until sieved(<4 mm) and were processed within 2 days. Subsamples wereextracted in 1 M KCl and analyzed for urea, NH₄⁺, and NO₂^–plus NO₃^– (25, 32). From two of the three blocks, 3- to3.5-g subsamples from 0- to 5-cm and 10- to 20-cm depths (selectedsamples) were prepared for PLFA analysis by using a modifiedBligh-Dyer single-phase extraction, solid-phase extraction on100-mg SPE columns (Varian, Harbor City, Calif.), and mild alkalinetransesterification as previously described (37).

Soil was dried at 105°C overnight for determination of gravimetricsoil moisture. Dried subsamples were analyzed for total C and¹³C. The soil bulk density of each depth interval was determinedby the end of the experiment.

GC-IRMS analyses.
The concentrations and isotopic composition of CO₂ were analyzedwith a Europa (Crewe, United Kingdom) Scientific Tracermassisotope ratio mass spectrometer (IRMS) coupled to an automatedgas analysis system. The total C content and isotopic compositionof soil samples were determined by using an automated combustionelemental analyzer interfaced with a Europa ANCA-SL IRMS system.¹³C-labeled PLFA methyl esters were analyzed on a Finnigan DeltaPlus XL gas chromatograph (GC)-combustion IRMS (ThermoQuest,Pegnitz, Germany). The gas chromatograph (Hewlett-Packard 6890)was equipped with an HP-5MS column (60 m by 0.25 mm [inner diameter])and a GC/C III combustion interface. Helium was used as carriergas.

Fatty acids were tentatively identified from retention timesand cross-referencing with samples analyzed by GC-MS. The ${delta}$ ¹³Cvalues determined by GC-combustion IRMS, based on authenticstandards certified relative to PeeDee Belemnite, were correctedfor the isotope ratio of the methyl moiety of fatty acid methylesters (1), as follows: ${delta}$ ¹³C_FA = [(C_n + 1) x ${delta}$ ¹³C_FAME – ${delta}$ ¹³C_MeOH]/C_n, where ${delta}$ ¹³C_FA is the ${delta}$ ¹³C of the fatty acid, C_n isthe number of C atoms in the fatty acid, and ${delta}$ ¹³C_FAME is the ${delta}$ ¹³C of the fatty acid methyl ester. The fractions of ¹³C ineach fatty acid and the amounts of ¹³C incorporated were calculatedas outlined by Boschker and Middelburg (5).

Statistical analyses.
Concentrations of urea, NH₄⁺, and NO₂^– plus NO₃^–,respectively, were compared across treatment, sampling day anddepth by a linear mixed model, as follows: Y = µ + ${alpha}$ _S +ß_T + ( ${alpha}$ ß)_ST + ${gamma}$ _D + ( ${alpha}$ ${gamma}$ )_SD + (ß ${gamma}$ )_TD+ ( ${alpha}$ ß ${gamma}$ )_STD + ${epsilon}$ _B + E_TB + E'_STB + E''_DTB + E'''STDB, whereGreek letters represent treatment effects, S is sampling day,T is urine treatment, D is depth interval, B is block, and Eterms represent random errors. Nitrogen concentrations werelog transformed to reduce heteroscedacity.

The PLFA distributions were analyzed by a principal componentanalysis (PCA) after log(n + 1) transformation of moles percentagesand with the covariance matrix. The concentrations of individualfatty acids at the 0- to 5-cm depth were also compared acrosstreatments and sampling days by individual analyses of variance(ANOVAs), using the stepwise Bonferroni procedure to controlthe overall table-wise error rate (41). Statistical analyseswere carried out with SAS 8.2 (SAS Institute, Cary, N.C.).

RESULTS

Top

Results

CO₂ emissions and ¹³C recovery.
The cumulated emissions of CO₂ from soil and vegetation duringthe 14-day period (average ± standard error) amountedto 19.8 ± 2.9, 25.8 ± 6.0, and 22.6 ± 1.8g of C m^–2 in the CTL, LU, and HU treatments, respectively,and were not statistically different between treatments. Inboth the LU and HU treatments, the proportion derived from ureaconstituted around 35% of the total flux at the first samplingafter 3 h, a proportion which decreased to <1% by day 4 andto 0.1% by day 14. The cumulated recovery of ¹³C in CO₂ by day14 was 15 ± 9.2% in the LU treatment and 7.7 ±2.0% in the HU treatment. The recovery of ¹³C in the soil was37 ± 2.6% in the LU treatment and 16 ± 1.0% inthe HU treatment. Finally, the recovery of ¹³C in PLFAs wasmuch less than 1% in both treatments (see below).

Soil nitrogen dynamics.
Background concentrations of extractable N in the pasture soilwere very low, and N introduced via simulated urine depositionwas therefore readily detected (Table 1). There were strongvertical gradients, with more than half of the extractable Nin the top 5 cm. By day 4, urea concentrations were low, yetthey were elevated at the 0- to 5-cm depth in the HU treatmentrelative to the other treatments. The NH₄⁺ pool remaining inthe soil by day 14 was significantly higher in the HU treatmentthan in the LU treatment at all depths. Accumulation of NO₂^–plus NO₃^– occurred at similar rates in the LU and HU treatmentsand by day 14 corresponded to 23 and 17%, respectively, of theurea N added.

View this table:
[in this window]
[in a new window]
TABLE 1. Concentrations of extractable nitrogen at three soil depth intervals 4 and 14 days after simulated urine deposition^a

PLFA content and composition.
Only the 13 phospholipid fatty acids shown in Table 2 were consistentlydetected in all treatments and on both sampling days. Table2 presents the moles percent distributions and total yieldsat the 0- to 5-cm soil depth. Urine deposition resulted in PLFAconcentrations increasing from 26.6 to 36 to 39 nmol g^–1,an increase that was maintained in the HU treatment, but notin the LU treatment, by day 14. The need to control type 1 errorsmade the statistical tests of individual PLFAs relatively conservative.Accordingly, only a few significant effects were identified(Table 2), the most notable being the reduction of cy17:0/16:1 ${omega}$ 7ratios after 4 days. There was a net production of cyclopropanefatty acids (CFAs) in the HU treatment by day 4.

View this table:
[in this window]
[in a new window]
TABLE 2. Distribution of PLFAs, total yields of PLFAs, and cyclopropane fatty acid-to-precursor ratios at the 0- to 5-cm soil depth in the CTL, LU, and HU treatments

A PCA based on all treatments and both depths (selected samplesfrom 10- to 20-cm depth only) separated the two depth intervalsalong the first principal component (Fig. 1A). The samples fromthe 10- to 20-cm depth were enriched in cy17:0 and cy19:0, in10Me16:0, and in 18:0 (Fig. 1B). At the 0- to 5-cm depth, thePLFA profiles of the CTL and HU treatments were similar after4 days. In contrast, the PLFA composition of the LU treatmenthad changed dramatically, but it shifted towards the compositionof the CTL treatment after 14 days. The PLFA profile of theHU treatment changed in the same direction as that of the LUtreatment between day 4 and day 14.

View larger version (15K):
[in this window]
[in a new window]
FIG. 1. A. Score plot of the first two principal components of a PCA based on moles percent distributions of PLFAs in pasture soil from the 0- to 5-cm or 10- to 20-cm depth (selected samples only). The designations indicate treatment (CTL, LU, or HU) and sampling day (4 or 14). B. Component loadings of the same PCA analysis.

${delta}$ ¹³C signature of PLFAs.
The recovery of urea-derived ¹³C in PLFAs was small, resultingin ${delta}$ ¹³C changes of not more than 20 ${per thousand}$ (Fig. 2, top panels). Incorporationof ¹³C was observed in branched-chain fatty acids (i15:0, a15:0,and i16:0), in C₁₆ straight-chain fatty acids (mainly 16:0 and16:1 ${omega}$ 7), and in C₁₈ straight-chain fatty acids (18:0, 18:1 ${omega}$ 9,and 18:1 ${omega}$ 7). Interestingly, a decrease in ${delta}$ ¹³C was observed withboth cy17:0 (LU treatment by day 4) and cy19:0 (LU and HU treatmentsby day 14). Even though the changes in concentrations of CFAswere moderate, the ${delta}$ ¹³C depletion thus revealed that these CFAswere characterized by significant turnover. At the 0- to 5-cmsoil depth, the ${delta}$ ¹³Cs of CFAs in the LU and HU treatments wereon average 13 to 15 ${per thousand}$ lower than the average ${delta}$ ¹³C for all otherPLFAs by day 4 and were 7 to 12 ${per thousand}$ lower by day 14. Cyclopropanefatty acids in the CTL treatment were also depleted of ¹³C,with a ${delta}$ ¹³C of 2 to 5 ${per thousand}$ below the average for all other PLFAs (datanot shown).

View larger version (28K):
[in this window]
[in a new window]
FIG. 2. ${delta}$ ¹³C values (top panel) and incorporation of urea-derived C (bottom panel) in PLFAs extracted from pasture soil 4 and 14 days after simulated urine deposition (mean ± standard error; n = 2).

Recoveries of urea-derived C in PLFAs after 4 days were comparablein the HU and LU treatments, at 0.001 and 0.0009%, respectively.The concentrations of urea C incorporated in individual PLFAsare presented in Fig. 2, bottom panels. These data show a selectivityof ¹³C incorporation between and within treatments; six fattyacids (i15:0, a15:0, 16:1 ${omega}$ 7, 16:0, 18:1 ${omega}$ 9, and 18:1 ${omega}$ 7) accountedfor 85 to 95% of the total ¹³C incorporation. These fatty acidsalso accounted for most of the PLFA concentration changes inurine-amended soil relative to the CTL treatment. The LU andHU treatments differed with respect to the incorporation ofurea C into the branched-chain fatty acids i15:0 and a15:0,resulting in low i15:0/a15:0 ratios in the HU treatment by day4 (Table 1). The absence of excess ¹³C in the common fatty acid16:0 in the LU treatment after 4 days is notable, but it cannotbe explained at present.

There was a significant (P < 0.01 or better) positive relationshipbetween total concentration changes and the incorporation of¹³C in individual PLFAs. In Fig. 3, the average fractions ofPLFA carbon derived from urea in the LU and HU treatments after4 and 14 days are plotted against the fractions of CO₂ derivedfrom urea. On day 14, the incorporations of urea C in CO₂ andPLFAs were relatively similar in the LU and HU treatments (closeto the 1:1 line). In contrast, on day 4, the labeling of excessPLFA was greater than the labeling of CO₂ emitted, especiallyin the HU treatment.

View larger version (22K):
[in this window]
[in a new window]
FIG. 3. The fractions of CO₂ derived from urea on day 4 and day 14 were plotted against the fractions of urea C in excess PLFA, i.e., PLFA concentration changes in the LU and HU treatments relative to the CTL treatment (mean ± standard error; n = 2).

DISCUSSION
Urine patches in grazed pastures constitute a harsh environmentthat is potentially stressful for soil organisms. Scorchingof the vegetation is known to occur, depending on depositionrate and urea concentration (42), and it is a result of osmoticstress, NH₃ toxicity, or a combination of both (18, 31, 42).Plants and microbes respond similarly to hyperosmotic conditions(11), and adverse effects on microbial populations can thereforealso occur. Polonenko et al. (40) exposed soil columns to osmoticpotentials of –0.5 MPa and below and found a significantdecrease in the number of viable cells leached from the salt-stressedcolumns but a significant increase in viable cell numbers followinga period of stress relief. In the present study, extractableNH₄⁺ alone corresponded to osmotic potentials of down to –0.19and –0.33 MPa in the LU and HU treatments, respectively.

The urine patch environment is expected to stimulate nitrification,but high urinary N concentrations may also inhibit nitrifyingbacteria (12, 31, 38, 47). Nitrite oxidizers such as Nitrobacterare far more sensitive to adverse environmental conditions thanNitrosomonas, in particular with respect to concentrations offree NH₃ (46). Inhibitory NH₃ levels of 0.1 to 1 and 10 to 150mg liter^–1 have been reported for Nitrobacter and Nitrosomonas,respectively (3). For the HU and LU treatments, concentrationsof free NH₃ were estimated from NH₄⁺ concentrations and pH tobe 80 to 120 and 40 to 65 mg N liter^–1, respectively,suggesting that a selective inhibition of NO₂^– oxidationwas likely to occur, especially in the HU treatment. Monaghanand Barraclough (31) found that maximum NO₂^– concentrationsin urine-amended soil increased progressively from 5 to 160mg of N kg^–1 soil as urine N increased from 3.8 to 25g of N liter^–1. Nitrite accumulation has also been observedat a urea level corresponding to that in the HU treatment inthe pasture soil used in the present study (38).

A response to urine deposition was detectable in the PLFA profilesof the soil microbial community. Significant growth was indicatedby the increase in PLFA yields with both the LU and HU treatmentsrelative to the unamended CTL treatment after 4 days (Table2). This was in accordance with results from the related laboratorystudy with comparable urea amendments to pasture soil, wherea twofold increase in potential NH₄⁺ oxidation activity alsowas observed 3 days after urea amendment (38). In the LU treatment,vigorous growth was further indicated by low CFA-to-precursorratios (15). In the HU treatment, however, such a reductionin CFA-to-precursor ratios was not observed despite the increasein PLFA yield (Table 2). Instead, a net production of CFAs occurredin the HU treatment, which could be interpreted as a stressresponse, since cyclopropane fatty acids have been shown toappear in microbial cell membranes in connection with variousstresses, including hyperosmotic conditions (14, 29). Thereis evidence that lipid extractability or partitioning duringsample preparation increases with ionic strength (13), but theimportance of this factor is not known.

The LU and HU treatments also differed in the proportions ofthe branched-chain fatty acids i15:0 and a15:0 after 4 days(Fig. 2, bottom left panel). A similar shift towards productionof anteiso fatty acids has been observed in response to saltstress with Listeria monocygotenes and several halotolerantbacteria (reference 10 and references therein); the resultingincrease in membrane permeability was explained as a mechanismto facilitate adaptation to hyperosmotic conditions (10). Reinspectionof PLFA results from the laboratory study referred to above(38) showed a consistent reduction of i15:0/a15:0 ratios afterurea amendment corresponding to the HU treatment but not afterurea amendment corresponding to the LU treatment (data availableupon request).

As indicated in Fig. 3, the incorporation of urea C in PLFAafter 4 days was higher than the proportion of urea C in CO₂,indicating that part of the CO₂ originated from unlabeled substratesin the soil. Stress-induced microbial turnover or degradationof soil organic matter dissolved by the urine could have causedthis (24, 38). However, cattle urine contains organic componentsbesides urea, such as hippuric acid (8), and dilution of ¹³CO₂by degradation of these components could also have occurred.

The observed depletion of ¹³C in CFAs (Fig. 2A) was unexpectedbut testifies to the turnover of these pools during the experiment.Boschker et al. (6) found cy17:0 to be slightly depleted of¹³C after [¹³C]methane amendment to an intertidal sediment,but ¹³C depletion of CFAs on the order of 10 to 20 ${per thousand}$ has not previouslybeen reported for complex microbial communities. Bacterial CFAsare produced from the monoenes 16:1 ${omega}$ 7c and 18:1 ${omega}$ 7c exclusivelyby trans-methylation from S-adenosylmethionine (14). Isotopedepletion of the active methyl group in S-adenosylmethionineby up to 39 ${per thousand}$ has been observed in natural compounds other thanfatty acids (19, 49), suggesting that the potential for ¹³Cdepletion in CFAs is high.

The source of ¹³C introduced in the experimental system wasa C₁ compound, urea. Whereas the isotopic composition of heterotrophsis mostly close to that of the growth substrates (1, 30, 48),the pathways of C assimilation during autotrophic growth canresult in ¹³C depletion of cell material by up to 27 ${per thousand}$ (34). Further,lipids are generally depleted in ¹³C relative to the total biomass(17, 34, 45). Experiments with sulfate-reducing bacteria grownon acetate led to a ¹³C depletion of 12 and 13 ${per thousand}$ in the fattyacid 10Me16:0, whereas growth on CO₂ resulted in depletionsof 24 and 18 ${per thousand}$ (30). Also, as mentioned above, Boschker et al.(6) found a depletion of ¹³C in cy17:0 isolated from a sedimentmicrobial community after [¹³C]methane amendment, but this wasnot the case after [¹³C]acetate amendment. Finally, discriminationagainst ¹³C during algal lipid biosynthesis was shown to increasewith CO₂ availability (43), suggesting that a high soil CO₂availability in urine patches could also increase isotope fractionationduring autotrophic growth.

The actual contribution of nitrifiers to ¹³C incorporation inPLFAs is not known. The greatest amounts of ¹³C were found inthe common fatty acids 16:0, 16:1 ${omega}$ 7c, and 18:1 ${omega}$ 7c, which predominatein Nitrosomonas and Nitrobacter cell membranes (4, 27, 28) butare also present in many other organisms. Fixation of CO₂ byheterotrophic bacteria may have accounted for some incorporationof ¹³C into PLFAs, although a more uniform distribution of labelwithin a range of microbial PLFAs would have been expected ifthis had been the main mechanism for urea C incorporation (44).Both cy17:0 and cy19:0 can be synthesized by autotrophic NO₂^–oxidizers, including Nitrobacter winogradski (27, 28). If Nitrobacterwas the main source of de novo CFA synthesis, then this mightexplain why ¹³C depletion was so extreme in these compounds.Future work should address the response of Nitrobacter to urinedeposition in more detail.

In summary, carbon and nitrogen transformations are intensein pasture soil affected by urine. The response of the pasturesoil microbial community was complex. The HU treatment was characterizedby osmotic pressures and free NH₃ concentrations which probablycaused some stress-induced metabolism and a partial inhibitionof nitrification activity during the first few days after deposition(12, 38, 40, 47). The low recovery of urea-derived C in PLFAsdid not suggest intracellular urea hydrolysis as a major mechanismfor turnover of urinary urea, but still some information aboutthe microbial response to urine deposition was obtained thatwas not revealed by overall PLFA dynamics. A stress responseof nitrifiers to urine deposition was indicated, which differedbetween the two levels of urinary urea applied.

ACKNOWLEDGMENTS

This study was supported by the EU Framework Programme 5 projectMIDAIR (EVK2 CT-2000-00096) and by the Danish Research Centrefor Organic Farming.

FOOTNOTES

* Corresponding author. Mailing address: Danish Institute of Agricultural Sciences, P.O. Box 50, DK-8830 Tjele, Denmark. Phone: 45 8999 1723. Fax: 45 8999 1619. E-mail: soren.o.petersen{at}agrsci.dk.

REFERENCES

Top