uLori A. Phillips , Cassandra R.Roger D. Armstrong d, Pauline Ma Department of Economic Development, Jobs, Transporb Department of Economic Development, Jobs, Transporc Department of Economic Development, Jobs, Transportd Department of Economic Development, Jobs, Transpore Centre for AgriBioscience, La Trobe University, Melbou

a r t i c l e i n f o

neutral peptidase and bacterial ammonia monooxygenase genes under moist conditions. Rotation his-unities to changingial community, withicrocosm study, ourbial organic-N pro-at sowing.. All rights reserved.

Sowing in the grain growing regions of south eastern Australiausually occurs between March and June, after the autumn rainfallbreak (Pook et al., 2009; Stokes and Howden, 2010). Unlike

ops are sown afterer daytime tem-

peratures in these regions often exceed 30 C, with a cumulativerainfall of less than 100 mm. Management strategies for thissummer fallow period have primarily focused on agronomic mea-sures to retain soil water and soil nitrogen (N) for subsequent crops(Kirkegaard et al., 2014 and references therein), often by controllingweeds (Haskins andMcMaster, 2012; Hunt et al., 2013). The amountof N remaining in the system at sowing, however, will also beinuenced by interactions between soil microbes and their envi-ronment during the fallow period. Previous assumptions that

* Corresponding author. Tel.: 61 3 9032 7141.E-mail address: lori.phillips@ecodev.vic.gov.au (L.A. Phillips).

1 Current address: Riverine Plains Inc., Mulwala, NSW, 2647, Australia.2 Current address: Victoria Institute of Strategic Economic Studies, Victoria Uni-

Contents lists availab

Soil Biology &

journal homepage: www.els

Soil Biology & Biochemistry 86 (2015) 201e211versity, Melbourne, 8001, Australiatory affected both the structural and functional resilience of the soil microbial commsoil moisture. Discing in green manure (vetch) residues promoted a resilient microba high organic-N cycling capacity in dry soils. Although this was a small-scale mresults suggest that management strategies could be developed to control microcessing during the summer fallow period and thus improve crop-available N levels

Crown Copyright 2015 Published by Elsevier Ltd

1. Introduction northern hemisphere cropping systems, grain cra typically hot and dry summer. Average summtrate levels were best predicted by laccase and neutral peptidase genes under drought conditions, but byArticle history:Received 26 November 2014Received in revised form25 March 2015Accepted 7 April 2015Available online 23 April 2015

Keywords:Organic nitrogen cyclingMicrobial functional genesQuantitative PCRAustralian soilsMineralizationNitricationhttp://dx.doi.org/10.1016/j.soilbio.2015.04.0040038-0717/Crown Copyright 2015 Published by ElsSchefe , Masha Fridman , Nicholas O'Halloran ,. Mele a, e

t and Resources, Bundoora, Victoria, 3083, Australiat and Resources, 1 Rutherglen, Victoria, 3685, Australiaand Resources, Tatura, Victoria, 3616, Australia

t and Resources, Horsham, Victoria, 3400, Australiarne, Victoria, 3086, Australia

a b s t r a c t

Some microbial nitrogen (N) cycling processes continue under low soil moisture levels in drought-adapted ecosystems. These processes are of particular importance in winter cropping systems, whereN availability during autumn sowing informs fertilizer practices and impacts crop productivity. Weevaluated the organic and inorganic N-cycling communities in a key cropping soil (Vertosol), using acontrolled-environment incubation study that was designed to model the autumn break in southAustralian grain growing regions. Soils from wheat, lucerne, and green manure (disced-in vetch) rota-tions of the Sustainable Cropping Rotations in Mediterranean Environments trial (Victoria, Australia)were collected during the summer when soil moisture was low. Microbial community structure andfunctional capacity were measured both before and after wetting (21, 49, and 77 days post-wetting)using terminal restriction fragment length polymorphism measures of bacterial and fungal commu-nities, and quantitative PCR of nitrogen cycling genes. Quantied genes included those associated withorganic matter decomposition (laccase, cellobiohydrolase), mineralization of N from organic matter(peptidases) and nitrication (bacterial and archaeal ammonia monooxygenase and nitrite oxidore-ductase). In general, the N cycling functional capacity decreased with soil wetting, and there was anapparent shift from organic-N cycling dominance to autotrophic mineral-N cycling dominance. Soil ni-a, * b, 1 a, 2 cAustralian agricultural soil

Organic nitrogen cycling microbial commevier Ltd. All rights reserved.nities are abundant in a dry

le at ScienceDirect

Biochemistry

evier .com/locate/soi lb io

& Bimicrobial activity during this period will be low due to low soilmoisture levels are being challenged (Alster et al., 2013; Barnardet al., 2013; Placella and Firestone, 2013), raising the question asto whether management strategies should also consider the activeN-cycling microbial communities that are likely to be present.

In drought-adapted ecosystems some functions, includingnitrication, may continue at a high level during the dry season(Parker and Schimel, 2011; Sullivan et al., 2012; Sher et al., 2013).Sullivan et al. (2012) found that N-uxes during the dry season inarid grassland soils in Arizona were either equal to or greater thanthose in the wet season. The authors suggest that a dominance ofarchaeal nitriers (AOA) over bacterial nitriers (AOB) in arid sitesmay drive this observed nitrication. AOA are generally detected ingreater numbers than AOB when soils are sampled during hot, dryperiods in Australia (O'Sullivan et al., 2013) and in other compa-rable drought-adapted environments (Sher et al., 2013). AlthoughAOA found in arid soils grow optimally at higher temperatures (ie.>35 C; Hatzenpichler, 2012), evidence from Western Australiasuggests that as soil dries nitrication becomes limited morequickly than other N-cycling processes, including mineralization(Hoyle and Murphy, 2011). The decomposition of incorporatedresidues and the subsequent mineralization or release of NH4

appears less affected by seasonal changes in soil moisture(Schomberg et al., 1994; Coppens et al., 2007; Hoyle and Murphy,2011), allowing mineral N to accumulate in dry soils. If rainfallevents do occur when no crops are present, this mineral-N poolmay be quickly nitried and lost from the system. A greater un-derstanding of the microbial communities driving this apparentlydrought-resilient organic-N cycling is necessary (Cabrera et al.,2005), and is particularly needed to inform management strate-gies for the Australian summer fallow period.

How soils respond to prolonged drought and subsequentrewetting events will depend on the organic matter levels (Yusteet al., 2011), the length of the drought period (Unger et al., 2010;Meisner et al., 2013), the frequency of the rewetting events(Xiang et al., 2008; Butterly et al., 2009), and, inherently, the natureof the microbial community present when the re-wetting eventoccurs. Most recent research has focused on microbial or chemistryresponses in native grassland soils, with a climate change focusregarding nutrient uxes (Huygens et al., 2011; Parker and Schimel,2011; Sullivan et al., 2012; Sher et al., 2013). Although the responseof N-cycling microbial communities to extreme changes in mois-ture is of particular importance in agricultural systems, such in-formation is still limited (for review see Borken andMatzner, 2009).Agronomic management practices, such as crop rotation (Yin et al.,2010) and tillage (Cookson et al., 2008; Hoyle and Murphy, 2011;Souza et al., 2013), are known to impact microbial communitycomposition and thus will also likely inuence how drought andrewetting inuences N-cycling.

In this controlled environment study, we used soil from a longterm eld trial in the Wimmera grain growing region of southernAustralia, to evaluate the N-cycling functional capacity of dry soilsfrom different agronomic treatments (rotation and residue man-agement). We examined how previous agronomic managementimpacted the resistance and resilience of these microbial commu-nities to increased soil moisture after the initial pulse of activity(Placella et al., 2012; Blazewicz et al., 2014) stabilized, and howmicrobe-environment interactions affected soil N pools. We hy-pothesized that microbes involved in organic-N transformationswould be key contributors of N release in dry soils, while microbesinvolved in mineral-N transformations would be more importantafter soil re-wetting occurred. Although studies on the N cycle havetypically focused on one step, these are not linear independentprocesses. Products of one step become substrates in another, and

L.A. Phillips et al. / Soil Biology202any of the downstream products may be re-bound to soil organicmatter via immobilisation processes. To assess the inter-relatedmicrobial functional capacity of this soil, we used qPCR methodsthat targeted genes associated with organic matter and residuedecomposition (laccase and cellobiohydrolase), mineralization(alkaline and neutral aminopeptidases), and nitrication (bacterialand archaeal amoA and nitrite oxidoreductase).

2. Materials and methods

2.1. Site and treatment characterisation

The soil was a self-mulching grey Vertosol (Isbell,1996) collectedfrom a long-term (est. 1998) eld experiment Sustainable CroppingRotations in Mediterranean Environments (SCRIME) located atLongerenong, Victoria, Australia (36.671 S; 142.290 E.110M asl).The SCRIME site investigates how rotation, cultivation and stubblemanagement affect soil chemistry, biology, and physics, all in rela-tion to crop yield. Soils were collected from the 0e10 cm depth ofthree cropping rotations (three replicates per treatment) withinSCRIME. The three rotations sampled were 1) continuous wheat(Triticum aestivum L.), 2) wheatebarleyepurple vetch (T. aestivumL. e Hordeum vulgare L. e Vicia benghalensis L.), and 3) canolaewheatepeaelucerneelucerne (Brassica napus L. e T. aestivum L. ePisum sativum subsp. arvense L. eMedicago sativa L. eM. sativa L.).

In the cropping season prior to sampling the plots were in thelast rotation of the above listed sequences and included (1) wheat(2) purple vetch, and (3) lucerne. The different crops were underdifferent management strategies. The wheat plots were under areduced tillage regime (chisel plough) and were sownwith a tynedseeder. The vetch was disced into the soil at owering (earlyOctober). The lucerne was slashed prior to owering and the resi-dues retained in situ. Soils from each of these treatments will bereferred to as Wheat, Green manure and Lucerne, respectively.Soils were collected in the post-harvest summer period when soilmoisture was low and passed through a 2-mm sieve. The soils werestored dry, in the dark, and at ambient temperatures until use, inorder to minimise pre-study changes to summer-adapted micro-bial communities.

2.2. Sample preparation and controlled environment studyconditions

A 40 g soil sample (2 mm sieved) was lightly packed into in-cubation containers consisting of a 40 mm diameter 40 mm highPVC tube, with a mesh bottom (aperture 0.06mm). Three replicatesof each treatment, corresponding to the original eld plots, wereestablished for each destructive sampling point. Soils were wet to60% eld capacity with sterile deionised water (dH2O) (the volumeof which was previously determined) and placed in 1 L air-tightincubation chambers. Prior to soil wetting, gravimetric moisturewas approximately 0.05 g/g; post soil wetting gravimetric moisturewas approximately 0.23 g/g for all samples. A beaker containing30 ml of water was placed in the incubation chamber to maintainhumidity. The chambers were incubated in the dark at 25 C foreleven weeks. A gas sample was taken periodically through an air-tight septum and analysed for CO2 using a Servomex 1450 gasanalyser (Servomex Pty Ltd, Crowborough, England). The incuba-tion samples were weighed regularly, with dH2O added as requiredto maintain the samples at 60% eld capacity.

Microbial activity (CO2 respiration) was monitored anddestructive samples (n 3) were taken at specic time points (0,21, 49, 77 days) for comprehensive chemical and microbial char-acterization. Time 0 samples (n 3) were dry soils, taken just priorto soil wetting. The next sampling date (day 21) was determined by

ochemistry 86 (2015) 201e211monitoring CO2, and coincided with the time after which the pulse

& Biof CO2 (the Birch effect; Birch, 1958) induced by soil wetting hadsubsided (Supplementary Figure 1). Sub-samples for molecularmicrobial analyses were frozen at 20 C until analysis.

2.3. Soil chemistry

The background chemical characteristics of the soil weredetermined at all designated time points. Electrical conductivity(EC; 1:5 soil/water), pH (1:5 soil/0.01 M CaCl2), nitrate- andammonium-nitrogen (NO3

, NH4; 1:10 soil/2 M KCl extract),sorbed phosphorus (Colwell-P; 1:100 soil/0.5 M NaHCO3-pH 8.5extract) and soil solution-P (CaCl2-P; 1:5 soil/0.005M CaCl2 extract)were determined as fully described in methods 3A1, 4B2, 7C2, 9B1and 9F1 respectively, of Rayment and Higginson (1992). A high-frequency induction furnace (LECO Pty Ltd) was used to measuretotal soil C and N. Organic matter content was determined using amultiplier of total C (1.85). Total soil P, Na, K, Ca, Mg, and S weredetermined by perchloric acid digestion before analysis by induc-tively coupled plasma atomic emission spectroscopy (SpectroAnalytical Instruments Pty Ltd, Kleve, Germany). Potentiallymineralizable N (PMN) was measured by the anaerobic incubationmethod, as described in Rayment and Lyons (2010). All assays wereperformed at the Department of Environment and Primary In-dustries State Chemistry Lab, Macleod, Victoria, Australia.

2.4. Microbial structural and functional analyses

2.4.1. Microbial community DNA extractionTotal microbial DNA was extracted from 0.25 g of each soil

sample using the MoBio PowerSoil DNA Isolation kit (MoBioLaboratories Inc., Carlsbad, CA, USA). Duplicate DNA extractionswere performed for each sample, DNA quality was visualized on0.7% agarose gels (SYBR Safe; Invitrogen, Mulgrave, Vic, Australia),replicate extractions were pooled, and DNA quantity was deter-mined spectrophotometrically (NanoDrop 2000, ThermoScientic).This total community DNA was used in all subsequent molecularanalyses.

2.4.2. Microbial community structureMicrobial community structure was assessed using terminal

restriction fragment length polymorphism (TRFLP) analyses of eu-bacterial 16S rRNA and fungal ribosomal ITS genes, with the primerpairs 63F/1087R (V1eV5 variable regions; Hauben et al., 1997;Marchesi et al., 1998) and ITS1F/ITS4R (ITS1, 5.8S, and ITS2 re-gions; White et al., 1990; Gardes and Bruns, 1993), respectively.Forward primers were labelled at the 50 end with 6-carboxy-uorescein uorescent dye. Each 50 mL amplication reactioncontained 20 ng of template DNA, 0.2 mM of each primer, 25 mLGoTaq master mix (Promega, Sydney, NSW, Australia) and 21 mLsterile H20. After an initial denaturation step of 95 C for 5 minamplication proceeded for 30 cycles of 95 C for 30s, 55 C(eubacteria) or 53 C (fungi) for 30s, and 72 C for 1min, with a nalextension step of 10 min at 72 C. PCR products were visualized on1.4% agarose gels with SYBR Safe DNA gel stain (Invitrogen, Mul-grave, Vic, Australia). Amplicons were puried using MilliporeMultiScreen lter plates (Merck Pty. Ltd, Kilsyth, Victoria, Australia)and then 200 ng of puried ampliconwas digested using restrictionendonuclease MspI (Promega, Sydney, NSW, Australia) for 4 h at37 C. Restriction digests were cleaned using 0.01 volume glycogen(20 mg/mL; Invitrogen, Mulgrave, Vic, Australia), 0.1 volume sodiumacetate (3 M), and 2.5 volume ethanol (95%), followed by twosequential washes with 70% ethanol. Cleaned digests (2 mL) wereresuspended in 10 mL formamide (Hi-Di, Applied Biosystems,Mulgrave, Vic, Australia) with 0.25 mL 500 LIZ size standard

L.A. Phillips et al. / Soil Biology(Applied Biosystems, Mulgrave, Vic, Australia), and fragmentanalysis was performed on an ABI 3730xl DNA analyser (AppliedBiosystems, Mulgrave, Vic, Australia).

TRFLP data was evaluated using GeneMapper software (v. 3.7;Applied Biosystems, Mulgrave, Vic, Australia), and proles weretrimmed to between 50 and 500 bp. After within-project normal-ization using the sum of signal algorithm, a bin width of 1.0 bp,maximum peak width of 1.5 bp, and a minimum peak height of50 rfu was used to determine the presence or absence of each TRFin each sample.

2.4.3. Microbial community functionMicrobial community functionwas assessed by quantitative PCR

of genes involved in soil organic matter cycling (high molecularweight organic matter and cellulose decomposition) and nitrogentransformations (mineralization and autotrophic nitrication)(Table 1). Assays were performed in a 25 mL reaction volume con-taining 20 ng DNA, with primers, reagents and cycling conditions aslisted in Table 1, using a Stratagene Mx3005P Q-PCR system (Agi-lent Technologies, La Jolla, California, USA). Duplicate assays foreach gene were assessed in a single run (Smith et al., 2006), on aplate that included a full range of the relevant standards. Absolutequantication was performed by comparison against calibrationstandards with a linear range from 2 to 2 107 copies.

Standard genes were amplied either from bacterial isolates(npr, apr) or from soil DNA (all other genes) using the relevantprimers (Table 1) in a 50 mL reaction volume (GoTaq Master Mix;Promega, Sydney, NSW, Australia), using a GeneAmp 9700 PCRsystem (Applied Biosystems, Mulgrave, Vic, Australia). After aninitial denaturation step of 95 C for 5min, amplication proceededfor 35 cycles of 95 C for 30s, 48 C (cbhI), 53 C (Cu, norA) or 55 C(apr, npr, bacterial amoA, archaeal amoA) for 45s, 72 C for 45s,followed by a nal extension step of 72 C for 10 min. PCR productswere visualized on 1.4% agarose gels with SYBR Safe DNA gel stain(Invitrogen, Mulgrave, Vic, Australia) to verify correct amplicationby size. Amplicons were cloned into pCRII-TOPO vectors in OneShot TOP100 chemically competent E. coli cells (Invitrogen, Mul-grave, Vic, Australia) and plasmids were extracted using a QIAprepSpin Miniprep Kit (Qiagen, Chadstone, VIC, Australia). The identityof the target gene was veried by sequencing with Big Dye v 3.1 kiton an ABI 3730xl DNA analyser (Applied Biosystems, Mulgrave, Vic,Australia). All procedures followed the manufacturers recom-mended protocols. Sequences were submitted for comparison tothe GenBank databases and a representative NCBI gene identity isgiven in Table 1.

2.5. Statistical analyses

Unless otherwise indicated, analyses were carried out usingStata12/SE (StataCorp LP, College Station, TX, USA). A critical valueof p 0.05 was taken to be signicant. No adjustment wasmade formultiple testing, however care was taken not to base conclusionson individual signicant p-values. Hierarchical cluster analysis ofpresence/absence TRFLP data was used to identify microbial com-munity patterns in soil samples, and furthest-neighbour distancemeasures were used to group the samples (SupplementaryFigure 2). Ubiquitous bacterial TRFLP fragments were excludedfrom analysis. All fungal TRFLPs were included. Signicant differ-ences in soil chemistry and microbial gene abundance levels overtime and between treatments were examined by ANOVA, followedby post hoc tests to determine where signicant differencesoccurred.

Relationships between measures were assessed by correlationanalysis, using Pearson's r if data normal and Spearman's rho if datanot normal (normality assessed by ShapiroeWilk's W), in both

ochemistry 86 (2015) 201e211 203Stata12/SE and the statistical software package PAST (v2.17;

community composition, shifting from clusters 1 or 2 to clusters 3 or

ditions and primer sources, and calibration standard identity.

rimer concentration;PCR reagents

Cycling conditionsa Best identity of calibrationstandard source

.0 mM; iTaqb 40 cycles of 95 C/5 s,58 C/30 s

Unknown fungus (FJ883810) andPseudomonas putida(CP005976)

.65 mM;ensiFASTc

40 cycles of 95 C/5 s,60 C/30 s

Unknown fungus (FN688723)

.5 mM;soAdvancedb

40 cycles of 95 C/5 s,60 C/30 s

Pseudomonas aeruginosa(AY973251)

.75 mM;soAdvancedb

45 cycles of 95 C/5 s,60 C/30 s

Bacillus cereus (CP001176)

.5 mM;ensiFASTc

35 cycles of 95 C/5 s,60 C/15 s

Nitrosospira multiformis(CP000103)

.5 mM;soAdvancedb

35 cycles of95 C/5sec, 60 C/30 s

Candidatus Nitrososphaeragargensis (EU281319)

.5 mM;soAdvancedb

40 cycles of 95 C/5 s,60 C/20 s

Nitrobacter vulgaris (AF344875)

d with a melt curve from 65 C to 95 C.

& Biochemistry 86 (2015) 201e211Hammer et al., 2001). All signicant correlations were assessedvisually to ensure outliers did not inuence results.

Distance based linear models (DistLM) were used to select thebiotic and abiotic variables that best predict the soil N pools (Se-lection criterion, AICc; Selection procedure, Best), using the Primer6and Permanova software package (v6.1.16 and v1.0.6 respectively;PRIMER-E Ltd, Ivybridge, United Kingdom). The overall signicanceof both treatment and time on biological and chemical patterns wasalso assessed by Permutational MANOVA (Permanova) using thePrimer6/Permanova software package. Biological analyses usedBray Curtis similarity resemblance matrices and soil chemistry an-alyses used an Euclidean Distance resemblance matrix, with 999permutations of residuals under a reduced model.

3. Results

3.1. Inuence of crop treatment and time on soil chemicalcharacteristics and nutrient status

The rotation history signicantly impacted some soil N poolsprior to soil wetting, but did not affect most other nutrient pools(Table 2). Soils from the green manure rotation had up to 6 times

Table 1Decomposition and nitrication steps evaluated, target organisms, amplication con

Process Target group andenzyme

Primer pair andreference

PQ

Polyphenolic soilorganic matterdecomposition

Fungi/bacteria:Laccase

cu1Af/2r(Kellner et al., 2007)

1

Cellulasedecomposition

Fungi:Cellobiohydrolase

cbhIf/r(Edwards et al., 2008)

0S

Soil organicnitrogen release

Bacteria: Alkalinemetallopeptidase

aprf/r(Bach et al., 2001)

0S

Bacteria: Neutralmetallopeptidase

nprf/r(Bach et al., 2001)

0S

Nitrication(NH4

to NO2- )Bacteria: ammoniamonoxygenase

amoA1f/2r(Rotthauwe et al., 1997)

0S

Archaea: ammoniamonoxygenase

Arch-amoAf/r (Wuchter et al.,2006)

0S

Nitrication(NO2e to NO3

)Bacteria: nitriteoxidoreductase

norAf1/r1 (Poly et al.,2008)

0S

a All assays started with an initial denaturation step of 95 C for 3 min and nisheb Bio-Rad, Gladesville, NSW, Australia.c Bioline, Alexandria, NSW, Australia.

L.A. Phillips et al. / Soil Biology204more nitrate (NO3; p < 0.001) than other treatments, and these

higher levels were maintained throughout the study (Fig. 1). Incontrast, both lucerne and wheat had low starting levels of NO3

,which then signicantly increased with time (p < 0.05). TheNO3

eN that accumulated in these soils was not simply derivedfrom the available NH4

N pool, which remained relatively stablein all treatments. Instead, small decreases in the C:N ratio of the soiland in the pool of potentially mineralizable N (PMN; p < 0.024 forlucerne and green manure, p 0.064 for wheat) suggest that thereleased NO3

eN was derived from soil organic matter. Althoughthere was no signicant difference in total organic matter betweenthe treatments (Supplementary Table 1), this source of organic Nalso gradually decreased over time in all treatments (p < 0.01).

3.2. Inuence of crop treatment and time on soil microbialcommunity structure

Eighty nine different bacterial and 132 different fungal terminalrestriction fragments (TRFs) were identied across the samples,with an average of 59 bacterial and 28 fungal TRFs per sample. Theprevious cropping treatment inuenced fungal and bacterialcommunity composition at the end of the dry period, and4 (Fig. 2A). These bacterial clusters differed between soil treatments.Cluster 4 occurred only once in green manure soils, cluster 1 wasabsent in wheat soils, and all clusters were detected in lucerne soils.

Soil fungal communities formed ve main clusters (clusters 1e5;Fig. 2B), based on hierarchical clustering of fungal TRFLP data. Thefungal communities did not follow similar clear treatment- and time-dependent shifts as bacterial communities. Instead, fungal commu-nities generally differed both with soil treatments and with time.

3.3. Inuence of crop treatment and time on soil microbialfunctional capacity

The soil microbial functional capacity (gene abundance pergram of soil) in dry soil at the start of the study was highly variable,continued to inuence how those communities changed once soilmoisture increased (Fig. 2; Permanova p for treatment or fortime 0.001).

Soil bacterial communities formed four major clusters (clusters1e4; Fig. 2A), based on a hierarchical cluster analysis of bacterialTRFLP data. Over time, there was a notable change in bacterialTable 2Chemistry of soils coming out of wheat, green manure, or lucerne rotations, prior tosoil wetting (t0).

Parameter Green manure Lucerne Wheat

EC1:5, dS/m** 0.28 (0.02) a 0.19 (0.03) b 0.18 (0.02) bpH(CaCl2) 7.77 (0.06) 7.83 (0.06) 7.80 (0.10)Organic matter, % 2.07 (0.12) 2.07 (0.32) 1.90 (0.26)Total C, % 1.13 (0.06) 1.11 (0.16) 1.02 (0.16)Total N, % 0.10 (0.01) 0.09 (0.01) 0.09 (0.01)C:N* 11.2 (0.3) b 12.1 (0.1) a 11.7 (0.5) abPMN, mg N kg1* 17.0 (5.3) ab 25.0 (3.6) a 14.0 (3.6) bNH4

-N, mg kg1 2.3 (0.3) 2.2 (0.4) 2.3 (0.2)NO3

-N, mg kg1*** 63.3 (8.6) a 10.0 (4.7) b 14.0 (2.0) bColwell P, mg kg1 39.3 (6.8) 34.7 (8.1) 33.3 (4.0)CaCl2 P, mg kg1 0.6 (0.1) 0.5 (0.2) 0.5 (0.0)Total P, mg kg1 237 (6) 213 (46) 243 (25)Sulphur, mg kg1 180 (0) 190 (26) 167 (21)Sodium, cmol kg1 1.3 (0.1) 1.3 (0.1) 1.2 (0.1)Calcium, cmol kg1 46.7 (2.0) 52.8 (6.2) 42.5 (6.9)Magnesium, cmol kg1 50.0 (0.4) 52.8 (2.1) 51.1 (1.3)Potassium, cmol kg1 19.5 (0.7) 18.5 (2.0) 18.6 (1.3)

Data are presented as means (n 3) with standard deviation in parentheses. Data inthe same row followed by different letters are signicantly different at p *0.05,**0.01, ***0.001. PMN: Potentially mineralizable N; CaCl2 P: soil solution P; ColwellP: sorbed P.

Fig. 1. Changes to soil nitrogen pools following soil wetting. Data points are means (n 3) with error bars representing 1 standard deviation. Signicant differences (p < 0.05) innitrogen pools between treatments at the different time points are indicated by different letters (a, b).

Fig. 2. Changes in soil microbial community patterns in different treatments over time, as assessed by terminal restriction fragment length polymorphism (TRFLP) assays. A)Bacterial TRFLP cluster classication; B) Fungal TRFLP cluster classication. The size of each circle indicates the number of samples represented in each cluster: large circlesrepresent 3 samples, medium circles represent 2 samples and small circles represent 1 sample. Complete linkage cluster analysis was used to classify soil samples on the basis oftheir bacterial and fungal TRF distribution.

L.A. Phillips et al. / Soil Biology & Biochemistry 86 (2015) 201e211 205

& Biboth between treatments and within a given treatment (t0; Fig. 3).Both cropping treatment and soil chemistry inuenced this mi-crobial functional capacity. For example, soils from green manuretreatments generally had a greater functional capacity than othersoils, and contained approximately three times more neutralpeptidase genes (p < 0.05) than soils from wheat treatments. Therelative abundance of these and other genes was signicantlyassociated with differences in soil chemistry (Table 3). Neutralpeptidase gene abundance was signicantly associated with bothsoil solution and sorbed P pools (CaCl2 and Colwell P, p < 0.05),while alkaline peptidase gene abundance was highly associatedwith organic matter, total C and N (p < 0.001 all), sulphur (p < 0.01)and all soil P measures (p < 0.05). Cellulase gene abundancessignicantly correlated with similar soil factors as alkaline pepti-dase. In contrast, bacterial and archaeal ammonium mono-oxygenase communities were not strongly associated with soilchemical pools in the dry soils, and only showed signicant nega-tive relationships (p < 0.05) with calcium and magnesium con-centrations (respectively).

After soil wetting, inter- and intra-treatment differences in mostmicrobial functional capacities generally decreased (signicantly atp 0.05 for apr, npr and nxr in greenmanure and for bacterial amoAin wheat) such that by day 77 there were no signicant differencesbetween the cropping soils (Fig. 3). However, inherent differencesin the nutrient pools in the different cropping soils continued to bestrongly associated with microbial functional potential. These re-lationships were particularly evident with respect to different soil Npools across all sampling points (Table 4). For example, in greenmanure soils, both organic and inorganic N-cycling microbialcommunities were highly associated with the PMN pool (p < 0.05).Although some similar relationships with PMN were found inlucerne soils, the microbial N-cycling communities in these treat-ments were more strongly associated with total soil N pools.

3.4. Biotic and abiotic interactions inuencing soil nitrogen pools

We used distance based linear models to predict the primarybiotic and abiotic interactions that were associated with threemeasured N-pools, PMN, NH4

, and NO3, both at sowing and

across all time points (Table 5). An X in Table 5 indicates a bio-logical or chemical factor that contributes to the best predictivemodel for these N pools. Soil solution P and Mg concentrationswere signicant contributors to all models. In combination withthese (and other) soil chemistry pools, laccase gene abundance wasthe best predictor of PMN, archaeal amoA abundance was the bestpredictor of NH4

, and neutral peptidase in combination withbacterial amoA were the best predictors of NO3

pools.

4. Discussion

Weused a controlled-environment incubation study, designed tomodel the autumn rainfall break in the grain growing regions ofsouth Australia, to evaluate the organic and inorganic N-cyclingmicrobial communities present in a key Australian grain croppingsoil. We further examined how previous rotation treatment inu-enced thosemicrobial communities, how this in turn inuenced theamount of plant-available N at sowing, and what the longer termimpacts of crop and tillagewere onmicrobial/chemical interactions.

4.1. Nitrogen-cycling microbial communities are abundant in drysoils

Our cropping soils maintained a high microbial N-cycling po-tential, even under dry soil conditions (i.e. 5% gravimetric soil

L.A. Phillips et al. / Soil Biology206moisture). Genes associated with organic matter breakdown(laccase, cellobiohydrolase), mineralization of N from organicmatter (peptidases) and nitrication (bacterial and archaealammonia monooxygenase and nitrite oxidoreductase) weremeasured in all dry soils at levels equivalent to or exceeding thosemeasured after soils were re-wetted (Fig. 3). Similar increases inmicrobial populations or biomass during dry periods have beenreported in other drought-adapted ecosystems (Parker andSchimel, 2011; Barnard et al., 2013), including a semi-arid Austra-lian agricultural sandy soil (Hoyle and Murphy, 2011). Barnard et al.(2013) measured microbial community changes during the sum-mer dry-down in California grasslands, and found that taxonomicgene copy numbers (bacterial and archaeal 16S, fungal ITS)increased as soils dried. While the authors hypothesized thatincreased gene abundance may have been due in part to reductivecellular division in response to desiccation stress (Alvarez et al.,2004; Nystrom, 2004), they also found that transcription levelswere comparable to moist soils. Although some of our quantiedgenes may be derived from dead or dehydrated microbial cells(Blazewicz et al., 2014), most are found in soil organisms that areknown to have facultative and/or constitutive resistance to droughtstress (Schimel et al., 2007). If these communities retain even a lowlevel of activity during the summer fallow period, they wouldcontribute to the N pools available at sowing.

The signicantly higher NO3eN levels in our green manure

soils (Table 2), where vetch residues were incorporated by chiselplough, suggests that microbial processing of the incorporatedresidues continued after incorporation and during the subsequentfallow period when soils are normally dry. The organic matterdecomposition genes we measured primarily belong to fungi(Kellner et al., 2007; Edwards et al., 2008), which are generallymore drought resilient than bacterial communities and continue togrow under water stress conditions that inhibit bacterial growth(Guenet et al., 2012; Alster et al., 2013; Barnard et al., 2013;Kaisermann et al., 2013). Soil bacteria however, are physiologicallydiverse and exhibit a wide range of tolerances and responses todrought stress (Placella et al., 2012). Some taxa, including Actino-bacteria (Zvyagintsev et al., 2007; Barnard et al., 2013) and Grampositive bacteria such as Bacillus spp. (Halverson et al., 2000),remain active under drought conditions. The mineralization genenpr (encoding neutral peptidase) that was signicantly higher inour dry green manure soil is primarily found in Bacillus spp. in soil(Bach et al., 2001). These neutral peptidases remain active at lowersoil moistures than alkaline peptidases (apr), which are primarilyfound in Gram-negative bacteria such as Pseudomonas spp.(Vranova et al., 2013). In our study, npr abundance was positivelyassociated with NO3

levels in the dry soil, and, together withlaccase and soil magnesium levels, was the best predictor of NO3

at that time (Table 5).Although autotrophic nitrication genes (both bacterial and

archaeal) were also abundant in our dry soil and are known toremain active at low soil moisture levels (Sullivan et al., 2012;Placella and Firestone, 2013; Sher et al., 2013), there was no sig-nicant relationship between these genes and NO3

levels. Whilethe generally positive relationship we observed may have beensignicant with a greater sample size, it is also possible that wetargeted the wrong nitriers. Previous research suggests that het-erotrophic nitrication of organic Nmay account for more than 50%of total nitrication in Australian semi-arid soils (Cookson et al.,2006). Heterotrophic nitriers are phylogenetically diverse (DeBoer and Kowalchuk, 2001; Hayatsu et al., 2008), and includefungal (De Boer and Kowalchuk, 2001; Huygens et al., 2011) andbacterial (Verstraete and Alexander, 1973) species that are likely toremain active at lower moisture levels. Although we did notquantify heterotrophic nitriers (ie. Arthrobacter spp.; Verstraete

ochemistry 86 (2015) 201e211and Alexander, 1972), taxa such as Actinobacteria are very

& BiL.A. Phillips et al. / Soil Biologyabundant in Australian soils (Mele et al., 2010; Hayden et al., 2012).It is highly probable that heterotrophic nitrication is the missinglink between organic-N mineralization and nitrite oxidization, andthe subsequent accumulation of NO3

eN in our dry soils.

4.2. Rotation history affects microbial community recovery aftersoil re-wetting

The pulse of CO2 (the Birch effect; Birch, 1958) that occurred inour soils immediately after wetting is consistent with the

Fig. 3. Temporal change in the carbon- and nitrogen-cycling functional capacity of soil micrlucerne, and wheat rotations. Data points are means (n 3) with error bars representing treatments at the different time points are indicated by different letters (a, b).ochemistry 86 (2015) 201e211 207osmoregulatory hypothesis. Many soil organisms capable of sur-viving a drought period do so by accumulating osmolytes. Whensoil rewetting occurs, these same organisms must immediatelyrespond to the increased soil water potential by disposing of thoseosmolytes, or suffer cell lysis (Fierer and Schimel, 2003; Schimelet al., 2007). Although there is still uncertainty regarding theextent to which osmoregulation occurs in different ecosystems(Borken and Matzner, 2009; Boot et al., 2013; Kakumanu et al.,2013), at least one study found that drought-stressed Australiansoils accumulate known microbial osmolytes (Warren, 2014). Some

obial communities (gene copy number per g soil) in cropping soils from green manure,1 standard deviation. Signicant differences (p < 0.05) in gene copy numbers between

and

l pe

3038561*08**0*531474**1

& BiTable 3Relationship between the soil microbial functional potential (gene copies g1 soil)increased (t0).

Laccase Cellulase Alkaline peptidase Neutra

OM 0.299 0.779* 0.909*** 0.42Total C 0.333 0.775* 0.891*** 0.47Total N 0.255 0.792* 0.915*** 0.59PMN 0.353 0.541 0.353 0.36C:N 0.176 0.157 0.181 0.40NH4

0.083 0.556 0.469 0.40NO3

0.014 0.469 0.441 0.76Total P 0.222 0.567 0.682* 0.25P (CaCl2) 0.301 0.831** 0.741* 0.85P (Colwell) 0.264 0.833** 0.701* 0.76Ca 0.011 0.471 0.500 0.26Mg 0.487 0.37 0.502 0.34K 0.696* 0.385 0.358 0.35Na 0.052 0.149 0.317 0.28S 0.464 0.801** 0.822** 0.44EC 0.254 0.598 0.553 0.80pH 0.083 0.493 0.670* 0.28

L.A. Phillips et al. / Soil Biology208CO2 may also have originated from the mineralization of soilorganic carbon, as has been reported for a range of Australian soiltypes (Butterly et al., 2010). Drying, particularly in cracking claysoils, can open aggregates, exposing previously protected organiccompounds to microbial decomposition (Fierer and Schimel, 2003;Miller et al., 2005). In our study however, the ush of CO2 was alsofollowed by a decrease in extractable DNA in both legume treat-ments several weeks later. Although this may have been partiallydue to the decomposition of dehydrated biomass (Blazewicz et al.,2014), the duration of this change suggests that a signicant pro-portion of the microbes in our soils were dry-adapted and lysed assoil moisture increased.

The presence of a dry-adapted microbial population is sup-ported by our nding that the communities present once the initialush of activity subsided (three weeks after wetting) were bothstructurally (Fig. 2) and functionally (Fig. 3) different than thosepresent in the dry soils. Microbial communities immediatelyrespond to soil wetting in a manner predicated by taxonomy(Placella et al., 2012; Blazewicz et al., 2014) combined withecological strategy (Lennon et al., 2012). How these communities

Relationship assessed by correlation analysis, using Pearson's r if data normal and Speacorrelation, n 9. All signicant correlations assessed visually to ensure outliers not inamoA: ammonium monoxygenase; OM: organic matter; PMN: potentially mineralizable

Table 4Relationship between the total soil functional potential (gene copies g1 soil) and organic(t0et77days).

N pools Soil Laccase Cellulase Alkaline peptidase Neutral

OM GM 0.347 0.513 0.694* 0.276L 0.479 0.697* 0.785** 0.646*

W 0.147 0.553 0.026 0.415Total N GM 0.305 0.738 0.802** 0.391

L 0.550 0.547 0.832*** 0.675*

W 0.034 0.463 0.012 0.279PMN GM 0.576* 0.782** 0.308 0.966*

L 0.361 0.575 0.673* 0.723*

W 0.195 0.149 0.022 0.305NH4

GM 0.163 0.282 0.361 0.539L 0.308 0.640* 0.683* 0.732*

W 0.241 0.290 0.148 0.059NO3

GM 0.193 0.104 0.303 0.060L 0.034 0.190 0.141 0.176W 0.127 0.393 0.030 0.088

Relationship assessed by correlation analysis, using Pearson's r if data normal and Speacorrelation n 12. All signicant correlations were assessed visually to ensure outliers noGM: green manure; L: lucerne; W: wheat; amoA: ammonium monooxygenase; OM: orgsoil chemistry, across all treatments at the start of the study before soil moisture

ptidase Archaeal amoA Bacterial amoA Nitrite oxidoreductase

0.026 0.205 0.6070.037 0.144 0.6320.115 0.011 0.732*

0.242 0.439 0.3430.176 0.497 0.3740.331 0.483 0.5020.477 0.371 0.689*

0.132 0.503 0.4510.208 0.432 0.889***0.062 0.324 0.759*

0.091 0.647* 0.3200.656* 0.564 0.5350.152 0.407 0.4820.194 0.04 0.127

0.257 0.251 0.6070.428 0.286 0.760*0.179 0.160 0.391

ochemistry 86 (2015) 201e211change or recover over the longer period, a critical factor fornutrient cycling in agricultural systems, is less clear. Microbialcommunities may take over a month to recover from drought andre-wetting stress (Pesaro et al., 2004) and specic functionalitymaytake longer to recover (Fierer et al., 2003), if this recovers at all(Evans et al., 2014). In our study, agronomic treatment impacted thetime it took for the microbial communities to reach a new stable-state (Shade et al., 2012) after wetting, as determined by TRFLP(Fig. 2). Bacterial communities in the green manure treatmentsreached awet-adapted alternative stable-state (Beisner et al., 2003)within three weeks, while those in lucerne and wheat treatmentswere still changing after seven weeks. Fungal communities weregenerally both more resistant and resilient to changing soil mois-ture than bacterial communities, but once again communities inthe tilled green manure treatments exhibited the highest level ofstability. Although tillage is often considered to reduce the capacityof soil microbial communities to resist drought stress (Steenwerthet al., 2005; Kaisermann et al., 2013), the incorporation of organicmatter increases that capacity (Yuste et al., 2011; Grifths andPhilippot, 2013), in part by increasing substrate availability. In our

rmans rho if data not normal (normality assessed by ShapiroeWilk's W). For eachuencing results. Bolded correlations are signicant at P *0.05, **0.01, ***0.001);nitrogen; EC: electrical conductivity.

and inorganic N pools in the different cropping treatments, across all sampling points

peptidase Archaeal amoA Bacterial amoA Nitrite oxidoreductase

0.633* 0.196 0.2870.145 0.546 0.779**

0.197 0.213 0.5340.489 0.376 0.4540.247 0.615** 0.857***

0.081 0.085 0.431** 0.594* 0.929*** 0.856**** 0.242 0.511 0.890***

0.172 0.072 0.1730.344 0.446 0.165

* 0.778** 0.383 0.611*

0.455 0.430 0.628*

0.105 0.112 0.0700.028 0.072 0.331

0.179 0.668* 0.421rman's rho if data not normal (normality assessed by ShapiroeWilk's W). For eacht inuencing results. Bolded correlations are signicant at P *0.05, **0.01, ***0.001;anic matter; PMN: potentially mineralizable nitrogen.

tha

t-we

83

N pN ha

& Bisoils, the incorporation of vetch residue was more signicant thandisturbance in determining overall community resilience tochanging soil moisture.

Community resilience does not, however, necessarily translateto functional resilience. In all our soils the overall microbial N-cycling capacity was sensitive to soil wetting, with the greatestdecreases observed in the legume treatments (Fig. 3). As discussedabove, the decomposition of previously dehydrated biomassimmediately post-wetting may account for some of the decrease inmeasurable functional capacity (Blazewicz et al., 2014). However,most of the functional changes were also evident as specic de-creases in functional gene abundance (i.e. on a per ng DNA basis,see Supplementary information). There are numerous likely rea-sons for this occurrence, including shifts in community structure asdry-adapted generalists decreased and wet-adapted specialistsincreased in abundance (Lennon et al., 2012). However, anotherfactor to consider is that other organisms, including bacterivores,would also have increased in abundance after soil wetting (Saetre

Table 5DistlM analysis of the biological (genes per gram of dry soil) and chemical interactionsbest predictive model for a given nitrogen pool.

Pre-wetting Pos

PMN NH4a NO3

b PMN

Laccase X X XCellulase X XNeutral peptidase X XAlkaline peptidase XArchaeal amoABacterial amoAOrganic matter XPMN N/A N/ATotal-P X XCaCl2-PMg XKCa X X XAICc 15.5 10.8 7.0 6.R2 0.973 0.970 0.961 0.39

Only biological and chemical variables that occur in the best t solutions are listed. PMhighly collinear with each other and with organic matter (r > 0.9); therefore C and

a PMN included in analysis.b PMN and NH4

included in analysis.

L.A. Phillips et al. / Soil Biologyand Stark, 2005; Eisenhauer et al., 2012). Thus, increased preda-tion and/or a dilution effect associated with increased non-microbial DNA being extracted (i.e. the total pool of extractedDNA post-wetting would have included protozoa and nematodeDNA) may also have contributed to the observed decreases infunctional gene abundance. Most of the assessed functional geneshowever, began to recover approximately one month post-wetting(Fig. 3), indicating that the overall N-cycling capacity had a highdegree of resilience.

Soil wetting is known to increase N mineralization and nitri-cation (Borken and Matzner, 2009), as some microbes shift from adormant to an active state (Saetre and Stark, 2005). After moistureincreases in grassland soils, nitrication transcriptional activityremains elevated until available NH4

pools are depleted (Placellaand Firestone, 2013). In our soils, increased moisture also stimu-lated N-cycling activity, with organic N pools (potentiallymineralizable-N) decreasing and NO3

accumulating (Fig. 1). Priorto soil wetting, decomposition (laccase) and mineralization (npr)genes, combined with Mg, were the best predictors of NO3

levels(R2 0.96, Table 5). After soil wetting, mineralization (npr) andautotrophic nitrication genes (bacterial amoA), combined withPMN and total P pools, were the best predictors of NO3

levels(R2 0.76, Table 5). This nding supports a switch fromheterotroph-dominated to autotroph-dominated nitrication(Cookson et al., 2006; Hoyle and Murphy, 2011) in a range ofAustralian soils when soil moisture increases after the autumnbreak.

4.3. Soil micronutrient levels and microbial functional capacity

The strong relationships between soil micronutrients and mi-crobial gene abundance in our soil are particularly interesting. Aspreviously noted, magnesium was a component of the best pre-dictive model for NO3- in dry soils and across all time points, inassociation with neutral peptidase and other genes (Table 5).Calcium was a similarly important component of predictivemodels for PMN and NH4

(Table 5). Although magnesium is aknown essential micronutrient for all life, both magnesium andcalcium have also been shown to stabilize the activity of bacterialextracellular peptidases in culture (Rahman et al., 2006;Jankiewicz and Frak, 2012). These divalent cations may be simi-larly assisting in maintaining extracellular peptidase functionality

t best predict the soil nitrogen pools. An X indicates the factors that contribute to the

tting Across all time points

NH4a NO3

b PMN NH4a NO3

b

XX

X X

X XX X

X XX N/A XX X

X X X XX X X

XX X0.9 43.6 14.3 5.2 45.20.346 0.757 0.505 0.412 0.806

otentially mineralizable N; CaCl2eP: soil solution P. Note that total N and total C areve been removed and organic matter is acting as a proxy for all.

ochemistry 86 (2015) 201e211 209in a soil environment. In addition, although not a component ofone of the predictive models, soil sulphur levels were stronglyassociated with both cellulase and alkaline peptidase geneabundance (Table 3). Cellulose breakdown by lamentous fungihas been shown to be connected to sulphur metabolism, withboth the amount and type of available sulphur regulating cellu-lase gene expression (Gremel et al., 2008). Although micro-nutrients may not have as direct an inuence on microbialextracellular enzyme functioning under eld conditions, at leastone study has shown that sulphur fertilizer enhances carbohy-drate metabolism in soil (Verdenelli et al., 2013). Future researchon management strategies to control microbial organic-N cyclingshould further investigate the importance of thesemicronutrients.

5. Conclusions

Nitrogen-cycling microbial communities are abundant in a drygrain cropping soil from south Australia. In this alkaline crackingclay, microbes with the capacity to process nitrogen from organicsources, via decomposition, mineralization, and, potentially, het-erotrophic nitrication pathways, remain abundant and apparentlyactive under conditions of very low soil moisture. These dry-adapted organic-N cycling communities were generally more

& Biabundant in soils from a green manure rotation, where vetch res-idues were disced in at the end of the previous cropping season.Once soil moisture increased, as it would during the autumn break,both the structure and the functional capacity of these commu-nities changed, in patterns inuenced by rotation history. Fungaland bacterial communities in the green manure soils were struc-turally the most resistant and resilient (respectively), with bacterialcommunities reaching a new alternate stable state within a fewweeks. In all soils, the functional capacity was sensitive to soilwetting, and there was an apparent shift from organic-N cyclingdominance to autotrophic mineral-N cycling dominance.Throughout the study, measurable links were found betweenagronomic treatment, microbial function, micronutrients, and therelease of mineral-N. Although this was only a small-scale micro-cosm study, our results do suggest that management strategiescould be developed to control microbial organic-N processingduring the summer fallow period and thus improve crop-availableN levels at sowing.

Acknowledgements

This research was funded by the Grains Research and Develop-ment Corporation (GRDC) through the Soil Biology Initiative IIProgram and by the Victorian Department of Environment andPrimary Industries (now The Department of Economic Develop-ment, Jobs, Transport, and Resources), through Project NumberDAV00106. Special thanks to Helen Hayden for assisting withexperimental design and to Damian Bougoure for providing T-RFLPprotocols.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.soilbio.2015.04.004.

References

Alster, C.J., German, D.P., Lu, Y., Allison, S.D., 2013. Microbial enzymatic responses todrought and to nitrogen addition in a southern California grassland. Soil Biologyand Biochemistry 64, 68e79.

Alvarez, H.M., Silva, R.A., Cesari, A.C., Zamit, A.L., Peressutti, S.R., Reichelt, R.,Keller, U., Malkus, U., Rasch, C., Maskow, T., Mayer, F., Steinbchel, A., 2004.Physiological and morphological responses of the soil bacterium Rhodococcusopacus strain PD630 to water stress. FEMS Microbiology Ecology 50, 75e86.

Bach, H.J., Hartmann, A., Schloter, M., Munch, J.C., 2001. PCR primers and functionalprobes for amplication and detection of bacterial genes for extracellularpeptidases in single strains and in soil. Journal of Microbiological Methods 44,173e182.

Barnard, R.L., Osborne, C.A., Firestone, M.K., 2013. Responses of soil bacterial andfungal communities to extreme desiccation and rewetting. ISME Journal 7,2229e2241.

Beisner, B.E., Haydon, D.T., Cuddington, K., 2003. Alternative stable states in ecology.Frontiers in Ecology and the Environment 1, 376e382.

Birch, H.F., 1958. The effect of soil drying on humus decomposition and nitrogenavailability. Plant and Soil 10, 9e31.

Blazewicz, S.J., Schwartz, E., Firestone, M.K., 2014. Growth and death of bacteria andfungi underlie rainfall-induced carbon dioxide pulses from seasonally dried soil.Ecology 95, 1162e1172.

Boot, C.M., Schaeffer, S.M., Schimel, J.P., 2013. Static osmolyte concentrations inmicrobial biomass during seasonal drought in a California grassland. SoilBiology and Biochemistry 57, 356e361.

Borken, W., Matzner, E., 2009. Reappraisal of drying and wetting effects on C and Nmineralization and uxes in soils. Global Change Biology 15, 808e824.

Butterly, C.R., Bnemann, E.K., McNeill, A.M., Baldock, J.A., Marschner, P., 2009.Carbon pulses but not phosphorus pulses are related to decreases in microbialbiomass during repeated drying and rewetting of soils. Soil Biology andBiochemistry 41, 1406e1416.

Butterly, C.R., Marschner, P., McNeill, A.M., Baldock, J.A., 2010. Rewetting CO2 pulsesin Australian agricultural soils and the inuence of soil properties. Biology andFertility of Soils 46, 739e753.

Cabrera, M.L., Kissel, D.E., Vigil, M.F., 2005. N mineralisation from organic

L.A. Phillips et al. / Soil Biology210residues: research opportunities. Journal of Environmental Quality 34,75e79.Cookson, W.R., Mller, C., O'Brien, P.A., Murphy, D.V., Grierson, P.F., 2006. Ni-trogen dynamics in an Australian semiarid grassland soil. Ecology 87,2047e2057.

Cookson, W.R., Murphy, D.V., Roper, M.M., 2008. Characterizing the relationshipsbetween soil organic matter components and microbial function and compo-sition along a tillage disturbance gradient. Soil Biology and Biochemistry 40,763e777.

Coppens, F., Garnier, P., Findeling, A., Merckx, R., Recous, S., 2007. Decomposition ofmulched versus incorporated crop residues: modelling with PASTIS clariesinteractions between residue quality and location. Soil Biology and Biochem-istry 39, 2339e2350.

De Boer, W., Kowalchuk, G.A., 2001. Nitrication in acid soils: micro-organisms andmechanisms. Soil Biology and Biochemistry 33, 853e866.

Edwards, I.P., Upchurch, R.A., Zak, D.R., 2008. Isolation of fungal cellobiohydrolase Igenes from sporocarps and forest soils by PCR. Applied and EnvironmentalMicrobiology 74, 3481e3489.

Eisenhauer, N., Cesarz, S., Koller, R., Worm, K., Reich, P.B., 2012. Global changebelowground: impacts of elevated CO2, nitrogen, and summer drought on soilfood webs and biodiversity. Global Change Biology 18, 435e447.

Evans, S.E., Wallenstein, M.D., Burke, I.C., 2014. Is bacterial moisture niche a goodpredictor of shifts in community composition under long-term drought?Ecology 95, 110e122.

Fierer, N., Schimel, J.P., 2003. A proposed mechanism for the pulse in carbon dioxideproduction commonly observed following the rapid rewetting of a dry soil. SoilScience Society of America Journal 67, 798e805.

Fierer, N., Schimel, J.P., Holden, P.A., 2003. Inuence of drying-rewetting frequencyon soil bacterial community structure. Microbial Ecology 45, 63e71.

Gardes, M., Bruns, T.D., 1993. ITS primers with enhanced specicity for basidio-myceteseapplication to the identication of mycorrhizae and rusts. MolecularEcology 2, 113e118.

Gremel, G., Dorrer, M., Schmoll, M., 2008. Sulphur metabolism and cellulase geneexpression are connected processes in the lamentous fungus Hypocrea jecorina(anamorph Trichoderma reesei). BMC Microbiology 8, 174.

Grifths, B.S., Philippot, L., 2013. Insights into the resistance and resilience of thesoil microbial community. FEMS Microbiology Reviews 37, 112e129.

Guenet, B., Lenhart, K., Leloup, J., Giusti-Miller, S., Pouteau, V., Mora, P., Nunan, N.,Abbadie, L., 2012. The impact of long-term CO2 enrichment and moisture levels onsoilmicrobial community structure and enzyme activities. Geoderma 170, 331e336.

Halverson, L.K., Jones, L.J., Firestone, M.K., 2000. Release of intracellular solutes byfour soil bacteria exposed to dilution stress. Soil Science Society of AmericaJournal 64, 1630e1637.

Haskins, B., McMaster, C., 2012. Summer fallow management in 2010 across centralwest NSW. In: Capturing Opportunities and Overcoming Obstacles in AustralianAgronomy. In: Yunusa, I.B., G.J. (Eds.), 16th Australian Agronomy Conference.Australian Society of Agronomy/The Regional Institute Ltd, Armidale, NSW,Australia.

Hatzenpichler, R., 2012. Diversity, physiology, and niche differentiation ofammonia-oxidizing archaea. Applied and Environmental Microbiology 78,7501e7510.

Hammer, ., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics soft-ware package for education and data analysis. Palaeontologia Electronica 4, 9.

Hauben, L., Vauterin, L., Swings, J., Moore, E.R.B., 1997. Comparison of 16S ribosomalDNA sequences of all Xanthomonas species. International Journal of SystematicBacteriology 47, 328e335.

Hayden, H.L., Mele, P.M., Bougoure, D.S., Allan, C.Y., Norng, S., Piceno, Y.M.,Brodie, E.L., Desantis, T.Z., Andersen, G.L., Williams, A.L., Hovenden, M.J., 2012.Changes in the microbial community structure of bacteria, archaea and fungi inresponse to elevated CO2 and warming in an Australian native grassland soil.Environmental Microbiology 14, 3081e3096.

Hayatsu, M., Tago, K., Saito, M., 2008. Various players in the nitrogen cycle: diversityand functions of the microorganisms involved in nitrication and denitrica-tion. Soil Science and Plant Nutrition 54, 33e45.

Hoyle, F.C., Murphy, D.V., 2011. Inuence of organic residues and soil incorporationon temporal measures of microbial biomass and plant available nitrogen. Plantand Soil 347, 53e64.

Hunt, J.R., Browne, C., McBeath, T.M., Verburg, K., Craig, S., Whitbread, A.M., 2013.Summer fallow weed control and residue management impacts on winter cropyield though soil water and N accumulation in a winter-dominant, low rainfallregion of southern Australia. Crop and Pasture Science 64, 922e934.

Huygens, D., Schouppe, J., Roobroeck, D., Alvarez, M., Balocchi, O., Valenzuela, E.,Pinochet, D., Boeckx, P., 2011. Dryingerewetting effects on N cycling in grass-land soils of varying microbial community composition and management in-tensity in south central Chile. Applied Soil Ecology 48, 270e279.

Isbell, R.F., 1996. The Australian Soil Classication. CSIRO Publishing, Melbourne,Australia.

Jankiewicz, U., Frak, M., 2012. Proteolytic activity of Pseudomonas chlororaphis 5Nisolated from rhizosphere soil. Electronic Journal of Polish Agricultural Uni-versities 15, 1.

Kaisermann, A., Roguet, A., Nunan, N., Maron, P.-A., Ostle, N., Lata, J.-C., 2013.Agricultural management affects the response of soil bacterial communitystructure and respiration to water-stress. Soil Biology and Biochemistry 66,69e77.

Kakumanu, M.L., Cantrell, C.L., Williams, M.A., 2013. Microbial community response

ochemistry 86 (2015) 201e211to varying magnitudes of desiccation in soil: a test of the osmolyte accumula-tion hypothesis. Soil Biology and Biochemistry 57, 644e653.

Kellner, H., Luis, P., Buscot, F., 2007. Diversity of laccase-like multicopper oxidasegenes in Morchellaceae: identication of genes potentially involved in extra-cellular activities related to plant litter decay. FEMS Microbiology Ecology 61,153e163.

Kirkegaard, J.A., Hunt, J.R., McBeath, T.M., Lilley, J.M., Moore, A., Verburg, K.,Robertson, M., Oliver, Y., Ward, P.R., Milroy, S., Whitbread, A.M., 2014. Improvingwater productivity in the Australian Grains industry-a nationally coordinatedapproach. Crop and Pasture Science 65, 583e601.

Lennon, J.T., Aanderud, Z.T., Lehmkuhl, B.K., Schoolmaster Jr., D.R., 2012. Mappingthe niche space of soil microorganisms using taxonomy and traits. Ecology 93,1867e1879.

Marchesi, J.R., Sato, T., Weightman, A.J., Martin, T.A., Fry, J.C., Hiom, S.J., Dymock, D.,Wade, W.G., 1998. Design and evaluation of useful bacterium-specic PCRprimers that amplify genes coding for bacterial 16S rRNA. Applied and Envi-ronmental Microbiology 64, 795e799.

Mele, P., Sawbridge, T., Hayden, H., Methe, B., Stockwell, T., Pfannkoch, C., Lewis, M.,Tanenbaum, D., Rusch, D., Heidelberg, K., 2010. How does surface soil geo-

Schomberg, H.H., Steiner, J.L., Unger, P.W., 1994. Decomposition and N dynamics ofcrop residues: residue quality and water effects. Soil Science Society of AmericaJournal 58, 372e381.

Shade, A., Peter, H., Allison, S.D., Baho, D.L., Berga, M., Burgmann, H., Huber, D.H.,Langenheder, S., Lennon, J.T., Martiny, J.B., Matulich, K.L., Schmidt, T.M.,Handelsman, J., 2012. Fundamentals of microbial community resistance andresilience. Frontiers in Microbiology 3, 417.

Sher, Y., Zaady, E., Nejidat, A., 2013. Spatial and temporal diversity and abundance ofammonia oxidizers in semi-arid and arid soils: indications for a differentialseasonal effect on archaeal and bacterial ammonia oxidizers. FEMS Microbi-ology Ecology 86, 544e556.

Smith, C.J., Nedwell, D.B., Dong, L.F., Osborn, A.M., 2006. Evaluation of quantitativepolymerase chain reaction-based approaches for determining gene copy andgene transcript numbers in environmental samples. Environmental Microbi-ology 8, 804e815.

Souza, R.C., Cant~ao, M.E., Vasconcelos, A.T.R., Nogueira, M.A., Hungria, M., 2013. Soilmetagenomics reveals differences under conventional and no-tillage with croprotation or succession. Applied Soil Ecology 72, 49e61.

L.A. Phillips et al. / Soil Biology & Biochemistry 86 (2015) 201e211 211morphology and land-use inuence the soil microbial ecosystems in southeastern Australia? Insights gained from DNA sequencing of the soil meta-genome. In: 19th World Congress of Soil Science, Soil Solutions for a ChangingWorld, Brisbane, Australia.

Meisner, A., Bth, E., Rousk, J., 2013. Microbial growth responses upon rewettingsoil dried for four days or one year. Soil Biology and Biochemistry 66, 188e192.

Miller, A.E., Schimel, J.P., Meixner, T., Sickman, J.O., Melack, J.M., 2005. Episodicrewetting enhances carbon and nitrogen release from chaparral soils. SoilBiology and Biochemistry 37, 2195e2204.

Nystrom, T., 2004. Stationary-phase physiology. Annual Review of Microbiology 58,161e181.

O'Sullivan, C.A., Wakelin, S.A., Fillery, I.R.P., Roper, M.M., 2013. Factors affectingammonia-oxidising microorganisms and potential nitrication rates in south-ern Australian agricultural soils. Soil Research 51, 240.

Parker, S.S., Schimel, J.P., 2011. Soil nitrogen availability and transformations differbetween the summer and the growing season in a California grassland. AppliedSoil Ecology 48, 185e192.

Pesaro, M., Nicollier, G., Zeyer, J., Widmer, F., 2004. Impact of soil drying-rewettingstress on microbial communities and activities and on degradation of two cropprotection products. Applied and Environmental Microbiology 70, 2577e2587.

Placella, S.A., Firestone, M.K., 2013. Transcriptional response of nitrifying commu-nities to wetting of dry soil. Applied and Environmental Microbiology 79,3294e3302.

Placella, S.A., Brodie, E.L., Firestone, M.K., 2012. Rainfall-induced carbon dioxidepulses result from sequential resuscitation of phylogenetically clustered mi-crobial groups. Proceedings of the National Academy of Sciences 109,10931e10936.

Poly, F., Wertz, S., Brothier, E., Degrange, V., 2008. First exploration of Nitrobacterdiversity in soils by a PCR cloning-sequencing approach targeting functionalgene nxrA. FEMS Microbiology Ecology 63, 132e140.

Pook, M., Lisson, S., Risbey, J., Ummenhofer, C.C., McIntosh, P., Rebbeck, M., 2009.The autumn break for cropping in southeast Australia: trends, synoptic in-uences and impacts on wheat yield. International Journal of Climatology 29,2012e2026.

Rahman, R.N.Z.R.A., Geok, L.P., Basri, M., Salleh, A.B., 2006. An organic solvent-stablealkaline protease from Pseudomonas aeruginosa strain K: enzyme puricationand characterization. Enzyme and Microbial Technology 39, 1484e1491.

Rayment, G.E., Higginson, F.R., 1992. Australian Laboratory Handbook of Soil andWater Chemical Methods. Inkata Press, Melbourne, Australia.

Rayment, G.E., Lyons, D.J., 2010. Soil Chemical Methods e Australasia. CSIRO Pub-lishing, Collingwood, Australia.

Rotthauwe, J.H., Witzel, K.P., Liesack, W., 1997. The ammonia monooxygenasestructural gene amoA as a functional marker. Applied and EnvironmentalMicrobiology 63, 4704e4712.

Saetre, P., Stark, J.M., 2005. Microbial dynamics and carbon and nitrogen cyclingfollowing re-wetting of soils beneath two semi-arid plant species. Oecologia142, 247e260.

Schimel, D.S., Balser, T.C., Wallenstein, M., 2007. Microbial stress-response physi-ology and its implications for ecosystem function. Ecology 88, 1386e1394.Steenwerth, K., Jackson, L., Calderon, F., Scow, K., Rolston, D., 2005. Response ofmicrobial community composition and activity in agricultural and grasslandsoils after a simulated rainfall. Soil Biology and Biochemistry 37, 2249e2262.

Stokes, C., Howden, M., 2010. Adapting Agriculture to Climate Change: PreparingAustralian Agriculture, Forestry, and Fisheries for the Future. CSIRO Publishing,Collingwood, Victoria, Australia.

Sullivan, B.W., Selmants, P.C., Hart, S.C., 2012. New evidence that high potentialnitrication rates occur in soils during dry seasons: are microbial communitiesmetabolically active during dry seasons? Soil Biology and Biochemistry 53,28e31.

Unger, S., Maguas, C., Pereira, J.S., David, T.S., Werner, C., 2010. The inuence ofprecipitation pulses on soil respiration e assessing the Birch effect by stablecarbon isotopes. Soil Biology and Biochemistry 42, 1800e1810.

Verdenelli, R.A., Conforto, C.B., Perez-Brandan, C., Chavarra, D., Rovea, A., Vargas-Gil, S., Meriles, J.M., 2013. Integrated multivariate analysis of selected soil mi-crobial properties and their relationships with mineral fertilization manage-ment in a conservation agriculture system. Acta Agriculturae Scandinavica,Section B e Soil & Plant Science 63, 623e632.

Verstraete, W., Alexander, M., 1972. Heterotrophic nitriction by Arthrobacter sp.Journal of Bacteriology 110, 955e961.

Verstraete, W., Alexander, M., 1973. Heterotrophic nitrication in samples of naturalecosystems. Environmental Science and Technology 7, 39e42.

Vranova, V., Rejsek, K., Formanek, P., 2013. Proteolytic activity in soil: a review.Applied Soil Ecology 70, 23e32.

Warren, C.R., 2014. Response of osmolytes in soil to drying and rewetting. SoilBiology and Biochemistry 70, 22e32.

White, T.J., Bruns, T.D., Lee, S., Taylor, J., 1990. Amplication and direct sequencing offungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H.(Eds.), PCR Protocols: A Guide to Methods and Applications. Academic Press,London, UK, pp. 315e322.

Wuchter, C., Abbas, B., Coolen, M.J., Herfort, L., van Bleijswijk, J., Timmers, P.,Strous, M., Teira, E., Herndl, G.J., Middelburg, J.J., Schouten, S., SinningheDamste, J.S., 2006. Archaeal nitrication in the ocean. Proceedings of the Na-tional Academy of Science United States of America 103, 12317e12322.

Xiang, S.R., Doyle, A., Holden, P.A., Schimel, J.P., 2008. Drying and rewetting effectson C and N mineralization and microbial activity in surface and subsurfaceCalifornia grassland soils. Soil Biology and Biochemistry 40, 2281e2289.

Yin, C., Jones, K.L., Peterson, D.E., Garrett, K.A., Hulbert, S.H., Paulitz, T.C., 2010.Members of soil bacterial communities sensitive to tillage and crop rotation.Soil Biology and Biochemistry 42, 2111e2118.

Yuste, J.C., Pe~NUelas, J., Estiarte, M., Garcia-Mas, J., Mattana, S., Ogaya, R., Pujol, M.,Sardans, J., 2011. Drought-resistant fungi control soil organic matter decom-position and its response to temperature. Global Change Biology 17,1475e1486.

Zvyagintsev, D.G., Zenova, G.M., Doroshenko, E.A., Gryadunova, A.A., Gracheva, T.A.,Sudnitsyn, I.I., 2007. Actinomycete growth in conditions of low moisture.Biology Bulletin 34, 242e247.

Organic nitrogen cycling microbial communities are abundant in a dry Australian agricultural soil1. Introduction2. Materials and methods2.1. Site and treatment characterisation2.2. Sample preparation and controlled environment study conditions2.3. Soil chemistry2.4. Microbial structural and functional analyses2.4.1. Microbial community DNA extraction2.4.2. Microbial community structure2.4.3. Microbial community function

2.5. Statistical analyses

3. Results3.1. Influence of crop treatment and time on soil chemical characteristics and nutrient status3.2. Influence of crop treatment and time on soil microbial community structure3.3. Influence of crop treatment and time on soil microbial functional capacity3.4. Biotic and abiotic interactions influencing soil nitrogen pools

4. Discussion4.1. Nitrogen-cycling microbial communities are abundant in dry soils4.2. Rotation history affects microbial community recovery after soil re-wetting4.3. Soil micronutrient levels and microbial functional capacity

5. ConclusionsAcknowledgementsAppendix A. Supplementary dataReferences