Soil properties associated with organic matter-mediated suppression of bean root rot in field soil amended with fresh and composted paper mill residuals

ARTICLE IN PRESS

0038-0717/$ - se

doi:10.1016/j.so

�CorrespondManhattan, KS

E-mail addr1Present addr2Present addr

08901, USA.3Present addr

Soil Biology & Biochemistry 39 (2007) 2936–2948

www.elsevier.com/locate/soilbio

Soil properties associated with organic matter-mediated suppression ofbean root rot in field soil amended with fresh and composted

paper mill residuals

Dorith Rotenberga,�, Ana Jimenez Wellsa, Elisabeth J. Chapmana,1, Anna E. Whitfieldc,Robert M. Goodmanb,2, Leslie R. Cooperbanda,3

aDepartment of Soil Science, University of Wisconsin-Madison, 1525 Observatory Drive, Madison, WI 53706, USAbDepartment of Plant Pathology, University of Wisconsin-Madison, 1630 Linden Drive, Madison, WI 53706, USA

cDepartment of Plant Pathology, Kansas State University, 4024 Throckmorton Plant Sciences Center, Manhattan, KS 66506, USA

Received 18 March 2007; received in revised form 18 June 2007; accepted 22 June 2007

Available online 16 July 2007

Abstract

The ability of an organic amendment to suppress soil-borne disease is mediated by the complex interactions between biotic and abiotic

soil factors. Various microbiological and physicochemical soil properties were measured in field soils with histories of receiving 4 or 5

years of spring additions of paper mill residuals (PMR), PMR composted alone (PMRC), PMR composted with bark (PMRB), or no

amendment under a conventionally managed vegetable crop rotation. The objectives of this study were to (i) determine the residual and

re-amendment effects of the organic materials on root rot disease severity; (ii) determine the influence of amendment type on the

structure of bacterial communities associated with snap bean roots grown in these soils; and (iii) quantify the relative contributions of

microbiological and physicochemical properties to root rot suppression in the field and greenhouse. While all amendment types

significantly suppressed root rot disease compared to non-amended soils in both environments, only soils amended with PMR or PMRB

sustained suppressive conditions 1 year after the most recent amendment event. Disease severity was inversely related to microbial

activity (fluorescein diacetate assay) in recently amended soils only. Terminal restriction fragment length polymorphism (T-RFLP)

analysis of the 16s rRNA gene was performed to obtain bacterial profiles. Principal component analysis (PCA) of terminal restriction

fragments (TRFs) revealed general differences in bacterial community composition (PC1) among amendment types, and specific TRFs

contributed to these differences. Correlation and multiple regression analyses of the measured soil variables revealed that the

composition of root-associated bacterial communities and the amount of particulate organic matter—carbon in bulk soils imparted

independent and relatively equal contributions to the variation in disease severity documented in the field and greenhouse. Together, our

findings provide evidence that disease suppression induced by annual PMR inputs was mediated by their differential effects on bacterial

communities and the amount and quality of organic matter in these soils.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Paper mill residuals; Compost; Terminal restriction fragment length polymorphism (T-RFLP); Particulate organic matter; Microbial

community; Disease suppression; Common root rot; Rhizosphere

e front matter r 2007 Elsevier Ltd. All rights reserved.

ilbio.2007.06.011

ing author. Present address: Department of Plant Pathology, Kansas State University, 4024 Throckmorton Plant Sciences Center,

66506, USA. Tel.: +1 785 532 1376; fax: +1 785 532 5692.

ess: [email protected] (D. Rotenberg).

ess: Molecular and Cellular Biology Program, Oregon State University, ALS 3021, Corvallis, OR 97331, USA.

ess: Rutgers’ School of Environmental and Biological Sciences/NJAES Rutgers, The State University of New Jersey, New Brunswick, NJ

ess: University of Illinois, Department of Human and Community Development, 905 S. Goodwin Avenue, Urbana, IL 61801, USA.

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–2948 2937

1. Introduction

Soil organic matter (SOM) plays a major role in thefunctioning and sustainability of agricultural soils. Theamount and composition of SOM significantly impactsphysical, chemical, and biological properties of soils(reviewed in Weil and Magdoff, 2004), which togetherinfluence crop productivity and health. One way growerscan manipulate SOM content is through the addition oforganic amendments to croplands. For example, applica-tions of composted manures in agronomic crop productionsystems were shown to significantly enrich the ‘young’labile fraction of SOM (Fortuna et al., 2003; Griffin andPorter, 2004; Darby et al., 2006), a component of the soilcarbon pool that responds to various farm managementpractices (Sikora et al., 1996; Wander and Bollero, 1999),and has been associated with suppression of soil-bornediseases (Stone et al., 2001; Darby et al., 2006).

Numerous studies document the utility of incorporatingorganic amendments to container mixes to suppress soil-borne diseases of greenhouse-grown plants. The nature ofthis suppression is associated with high total microbialactivity (Zhang et al., 1998), organic matter composition(Boehm et al., 1993; Stone et al., 2001), competition for C-substrates by seed-colonizing bacteria (McKellar andNelson, 2003), and the availability of carbon substratesthat sustain high microbial activity (Chen et al., 1988; Huet al., 1997; Hoitink and Boehm, 1999). Recent findings areadvancing our understanding of the possible mechanismsunderlying the effect of organic amendments on crophealth and disease suppression in the field (Vallad et al.,2003; Rotenberg et al., 2005; Darby et al., 2006; Perez-Piqueres et al., 2006). The emerging picture indicates thatcomplex interactions between soil physicochemical andmicrobiological properties modulate disease suppression infield soils amended with organic materials.

Applying large C-inputs to agricultural soil is one way tomanage SOM to enhance the collective activities of nativesoil- and rhizosphere-associated microbes that contributeto general suppression (van Bruggen and Semenov, 2000).General suppression is defined as reduced disease severityor incidence due, in part, to the combined activities ofdiverse microbes, including bacteria and fungi (Whipps,2001; Yin et al., 2003), that directly or indirectly inhibitpathogen infection and survival (Cook and Baker, 1983).Of these microbes, native populations of rhizosphere-colonizing bacteria have been shown to play significantroles in crop health and protection from root pathogens(Weller et al., 2002; McSpadden Gardener, 2007) andto be sensitive to various farm management practicesincluding compost amendment (Rotenberg et al., 2007).This apparent link between root disease suppression androot-colonizing bacteria represented the basis of thepresent study.

In recent years, scientists have been interested inunraveling the connection between soil microbial commu-nity structure and soil-borne disease suppression using

DNA-based techniques (Yin et al., 2003; Perez-Piquereset al., 2006; Benıtez et al., 2007). One such techniqueis terminal restriction fragment length polymorphism(T-RFLP) analysis, a culture-independent method forresolving differences among complex microbial commu-nities (Liu et al., 1997; Clement et al., 1998; Lukow et al.,2000; Blackwood et al., 2003). In the present study,T-RFLP analysis of the 16s rRNA gene was performedto assess bacterial community structure in the rhizosphereand root surfaces of snap bean grown in sandy field soilsamended with paper mill residuals (PMR) and PMRcomposts.The context of the current study was a 5-year,

conventionally managed vegetable cropping systems trialinitiated in April 1998 in the Central Sands region ofWisconsin, USA. The study was designed to investigate theshort- and long-term effects of annual spring additions ofPMR and two forms of PMR composts on soil physico-chemical properties and crop health. In these sandy soils, 4years of repeated additions of fresh and composted PMRto the field soils significantly altered soil physical andchemical properties and related soil functions (Foley andCooperband, 2002; Newman et al., 2005). By the 4th yearof amendment, the soils had reached an elevatednew steady-state equilibrium between annual C-input andC-decay (Newman et al., 2005).Given the link between physicochemical and biological

properties of soil, we conjectured that biological properties,particularly microbial community structure and activity,were altered by amendment. During the first snap beancycle (2 years of amendment), common root rot incidence(Pythium spp. and Aphanomyces euteiches) in the field waslow at this site (Stone et al., 2003). However, after the 4thyear of amendment, root rot severity was significantlyhigher in non-amended field soils compared to amendedsoils in growth chamber bioassays (unpublished data). Wehypothesized that reduced root rot severity was due, inpart, to the incidence and abundance of particularmembers of the root-associated bacterial communitiesand their collective activities. It was also hypothesized thatthe amount and quality of soil C in these soils contributedto the magnitude of root rot suppression. The objectives ofthis study were to (i) determine the residual and re-amendment effects of the PMR amendments on root rotseverity (4 vs. 5 years of additions); (ii) determine theinfluence of amendment on the structure of bacterialcommunities (T-RFLP analysis) associated with snap beanroots grown in these soils; and (iii) quantify the relativecontributions of biological and physicochemical soilproperties to root rot suppression.

2. Materials and methods

2.1. Field site description and design

The vegetable cropping systems trial consisted of tworotation cycles of potato: snap bean: cucumber. The

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–29482938

predominant soil type at this site is plainfield loamy sand(sandy, mixed, mesic, Typic Udipsamment). In years 1–4 ofthe study, the experiment design was a randomizedcomplete block design with seven amendment treatments(two rates of PMR, PMRC, PMRB and a non-amendedcontrol) and five blocks. Each treatment was assignedrandomly to a 4.6m� 7.7m plot in each block. Organicamendments were applied annually over each plot andincorporated to a depth of 15 cm each April. Prior toamending in the 5th year, each plot was divided in halflength-wise to achieve two split-plots of 2.3m� 7.7mareas. One of the split-plots for each treatment receivednew amendment (A), while its companion half remainednon-amended (N) that year. Splitting of each plot provideda means for evaluating the residual effects (4 years ofadditions) of amendment on soil physical and chemicalproperties, crop productivity, and foliar disease during the5th year (Rotenberg et al., 2005). A detailed description ofamendment rates and chemical characteristics, supplemen-tal fertilizer applications, and plot management werepublished previously for the 1st–4th year (Foley andCooperband, 2002; Stone et al., 2003; Newman et al.,2005) and for the present study (Rotenberg et al., 2005).

2.2. Soil measurements

Seven soil physical and chemical properties werequantified each amendment year. During the 5th year,total soil carbon (TC) and nitrogen (TN), particulateorganic matter carbon (POM-C) and nitrogen (POM-N),bulk density, and volumetric moisture (VM) were deter-mined at three critical time points: 1 week beforeamendment (April 8), 1 month after amendment (May13), and 1 week prior to harvest (July 15). Plant availablenitrogen (NO3-N and NH4-N) was measured midseason(July 2). A composite soil sample consisting of 15 soil cores(3.5 cm-diam.� 15 cm-depth) was removed from each split-plot treatment and bulk density, TC, TN, POM-C and N,and VM were determined as described previously(Newman et al., 2005). In situ amounts of NO3-N andNH4-N were measured across each split-plot as describedby Foley and Cooperband (2002).

2.3. In-field disease assessment

Common root rot severity was determined for each fieldsplit-plot. Plants were sampled from a 1-m section of linearrow (10–18 plants) within the designated yield row on July17 (54 days after planting) and each root system was scoredfor root rot disease using the disease scale described byKobriger et al. (1998) (0 ¼ healthy firm roots; 1 ¼ slightdiscoloration of hypocotyl; firm hypocotyls; 2 ¼ hypocotyldark brown, collapsible under pressure; extensive rootpruning; 3 ¼ necrotic hypocotyls and collapsible undergentle pressure; severe root pruning; 4 ¼ dead). An averagescore was calculated for each treatment split-plot (A vs. N).One representative root sample for each treatment was

submitted to the University of Wisconsin-Madison PlantDisease Diagnostics Clinic to determine the presence ofroot-invading pathogens. Using standard microscopic andculturing techniques, it was confirmed that the hypocotylsof all symptomatic samples were infected with Pythium, acommon causal agent of root rot in Wisconsin sandy soils(Reeleder and Hagedorn, 1981; Kobriger et al., 1998).Crop yield was determined for each split-plot. A 3-m linearrow of beans was hand-harvested on July 22 from thedisease assessment row. Beans were graded with a rollinggrader into sieve size classes 1–3 (5–10mm-diameter), 4(10mm-diam.), and 5 (11mm-diam.). The beans wereharvested from each treatment split-plot when it wasdetermined that roughly 50% of the beans were sieve size 5.

2.4. Greenhouse root rot experiments

Experiments were conducted with soils collected fromthe field site in May, 4 weeks after the 5th year of organicamendment. Soil was systematically sampled across eachplot to a depth of 15 cm and bulked to achieve 3785 cm3 ofsoil for each amendment type/split-plot application com-bination of four field blocks (32 soils total). Fouramendment types (no amendment and the high rates ofPMR, PMRC, and PMRB), two split-plot applications(A and N), four replicates representing the field blocks, andtwo subsamples per amendment/split-plot treatment wereincluded for each experiment. At the time of planting, theaverage number of Pythium propagules in these soilsranged from 128 to 240 colony-forming units g�1 of soil asdetermined by plating 10-fold serial dilutions of 1-g soilsamples in sterile water on a selective corn meal agarmedium for Pythium and Phytopthora (Kannwischer andMitchell, 1978). One day before planting, 11.43-cm-diameter plastic pots were filled with soil and watered with80ml of tap water. Four snap bean seeds (same variety andseed treatment used in the field experiment) were planted ineach pot which received water daily beginning 48 h afterseeding. Pots were arranged in a randomized completeblock design with two subsamples per treatment per block(64 pots total). Once the majority of plants reached the V2stage of development (12 days after planting), soils weresaturated daily with water to promote root rot diseasedevelopment. Root systems were removed from pots 17–20days after planting (V4 stage) when the 1st above-groundsymptoms, notably stunting, were apparent. Each rootsystem was scored for root rot disease (severity scale usedin field experiment) and an average score was calculatedper pot. The greenhouse experiment was conducted twice.

2.5. Total soil microbial activity (FDA)

Bulk soil from each greenhouse pot was collected andair-dried at the time of root rot disease assessment. Totalmicrobial activity in the soil was estimated using thefluoroscein diacetate hydrolysis assay (FDA) (Darby et al.,2006), with modifications. Dried bulk samples were mixed

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–2948 2939

thoroughly and 3-g subsamples were placed into 250mlNalgene bottles. Three bottles were prepared for eachtreatment (i.e., three subsamples) to which 50ml of 60mMfluorescein diacetate (30, 60-diacetylfluorescein lipase sub-strate in sodium phosphate buffer [pH 7.8]) was added totwo of the bottles, while the third received 50ml of bufferwithout substrate, which served as a procedural controlused to correct for background. The soil solutions wereshaken for 2.5 h at 240 strokesmin�1 at room temperature(25–26 1C) after which 2ml of acetone was added to eachbottle to stop enzymatic activity. The bottles werecentrifuged at 2192g for 10min and 10ml of the resultingsupernatant was sampled, filtered through a 45-mm nylonsyringe filter, and diluted two-fold prior to takingabsorbance readings at 490 nm with the use of a spectro-photometer. To accommodate the large number ofsamples, the assay was performed in batches, with eachbatch representing treatments for one greenhouse block. Astandard fluorescein concentration curve, ranging from 0to 4mg l�1, was prepared fresh for each batch of samplesusing a stock solution of 60mM of fluorescein in sodiumphosphate buffer. The rate of fluorescein diacetate hydro-lysis (mg of product corrected for background fluorescenceper minute per gram of soil) was calculated to estimatetotal microbial activity for each experimental unit.

2.6. Bacterial community profiling (T-RFLP)

One intact seedling was removed from each pot, 17 daysafter planting at the first tri-foliate stage of development(V4). Seedlings were shaken gently to remove looselyadhering soil, placed in plastic bags, and processedimmediately. Root systems were chopped into 1.5-cmpieces and 0.2 g of root with rhizosphere soil was arbitrarilysubsampled. DNA extraction was performed with theUltraClean Soil DNA Kit (MoBio) using the manufac-turer’s protocol and the BIO 101 Fast Prep bead beater for30 s at 5.5ms�1 for the cell lysis step. Each DNA samplerepresented organisms colonizing the rhizosphere, rhizo-plane, and endophytic environments of the root system;therefore, herein, we use the term ‘root-associated’ todescribe the resulting consortia of bacteria. Purified DNAwas stored at �20 1C prior to PCR amplification.

Stock DNA samples were diluted 1:50 in sterile distilledwater and a volume of 1 ml was used as template for PCRamplification of the SSU (16s) rDNA sequences with thebacteria-specific primers 1492R (50-ACG GCT ACC TTGTTA CGA CTT) and HEX-labeled 27F-NotI (50-GCGGAT CCG CGG CCG CAG AGT TTG ATC MTG GCTCAG). Independent PCR reactions were performed two orthree times for each template. Each 25-ml PCR reactionvolume consisted of 1�Taq DNA polymerase buffer A(Promega), 200 mM dNTP mix (Promega), 200 mM eachprimer, and 1U of Taq polymerase (Promega). Reactionswere performed in a Robocycler Gradient 96 Thermocycler(Stratagene) using the temperature cycling protocol of 1cycle of 94 1C for 1min and 35 cycles of 94 1C for 30 s,

55 1C for 1.5min, 72 1C for 2.5min, and 72 1C for 5min.Negative controls containing sterile distilled water insteadof template were included in each run and produced nodetectable product. Gel electrophoresis was performedwith 5 ml of product on a 0.8% agarose gel, followed bystaining with ethidium bromide, resulting in the expectedsingle band of 1465 bp. Desalting and primer removal ofPCR reactions were performed by size-exclusion chroma-tography with Sephadex G50 microspin columns (Ausubelet al., 2003) packed in a MADV N65 10 Durapore 96-wellplate (Millipore). Replicate cleaned PCR products werepooled for each sample.Purified PCR products were digested with RsaI (Prome-

ga) in 20-ml reaction volumes containing 1�Promegabuffer C, 1�Promega BSA (10 mg ml�1), 5U of restrictionenzyme RsaI (Promega), and 50 ng PCR product in 10 ml ofsterile distilled water. Digestions were performed in a 37 1Cwater bath for 3 h, followed by incubation at 65 1C for15min to inactivate the enzyme. Digestion products weredesalted using the procedure described above for PCRproduct purification. To confirm digestion, 5 ml of thedigestion product was loaded on a 1% Meta-Phor agarosegel, followed by ethidium bromide staining. Two microliterof the digestion product was mixed with 15 ml of formamideand 0.3 ml of the ROX-labeled GeneFlo 625 internalstandard (Chimerix, Milwaukee, WI), then heat-denaturedprior to separation of restriction fragments by multi-capillary electrophoresis (Trotha et al., 2002) with an ABI3700 automated sequencer (Applied Biosystems). Thelength (bp) and signal intensity of the HEX-labeled TRFswere automatically calculated by GeneScan Analysis Soft-ware v3.1 (Applied Biosystems, Foster City, CA). Sampleswere aligned with the internal size standard present in thesample and peak heights greater than 50 fluorescent units(above background fluorescence) represented unique TRFswithin a community profile for each sample. TRF profileswere converted to binary data (presence or absence of aTRF) and a data matrix consisting of TRF lengths incolumns and samples for each treatment in rows wascreated in Excel. The relative abundance of TRFs in eachprofile was estimated by the area under the peak calculatedby the GeneScan software. A matrix containing theseparametric data were also prepared.

2.7. Statistical analyses

2.7.1. Disease severity and soil properties

Analysis of variance (ANOVA) was performed onaverage root rot severity per pot (greenhouse) and plot(field) to determine the main effects of organic amendmenttype and split-plot application on amount of disease. Asplit-plot factorial model was tested with amendment typewithin block as the whole-plot unit, split-applicationtreatment as the split-plot unit, and included the interac-tion term. The model was tested using SAS v9.1 (SASInstitute Inc., Cary, NC) with the PROC MIXEDprocedure and the RANDOM statement to include the

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–29482940

whole-plot and split-plot error terms. The LSMEANSstatement was used to calculate least square differences andP-values for making treatment mean comparisons withinmain effect variables and the interaction.

To test the hypothesis that the outcomes of the twogreenhouse experiments were similar, a model includingthree main effects (amendment type, split-plot application,and experiment) and all two and three-way interactionterms was tested using the SAS PROC MIXED procedure.All terms in the model including ‘‘experiment’’ were notsignificant (P40.1) and therefore justified pooling thegreenhouse severity data. With these compiled data,severity values were averaged for each treatment. A similarfactorial analysis was performed to determine if the generaltrends documented in the greenhouse and field experimentswere similar.

ANOVA was performed on TC, TN, POM-C, POM-N,VM, and log10-transformed NO3-N or NH4-N values.ANOVA was also performed on the rate of FDAhydrolysis to determine the effect of organic amendmentand split-plot application on total microbial activity inthe soil at time of disease assessment. The same split-plot statistical models were tested as described for thedisease data.

2.7.2. Bacterial community profiles (T-RFLP)

Community profiles generated for each treatment werecompared using principal components analysis (PCA).PCA was performed on the incidence and relativeabundance data matrices to (i) explore the associationsbetween TRFs and each treatment; (ii) to generatesummary variables (principal components PCs) thatrepresented the community profile for each treatment;and (iii) to use each community summary variable and soilvariable as independent variables in multiple regressionanalysis of root rot severity. PCA was performed with JMPsoftware (SAS Institute, Cary, NC). ANOVA wasperformed on PCs representing TRF incidence using theSAS procedure PROC MIXED to determine if there was ageneral difference in community profiles among treat-ments. Welch’s ANOVA was performed on PCs represent-ing TRF abundance to adjust for the unequal varianceamong treatment groups (Levene’s test of equal variance,P ¼ 0.03). The Mann–Whitney test statistic (Po0.05) wascalculated to determine treatment effects on summaryvariables that represent abundance (PCs) and for abun-dance of specific TRFs.

To determine if the occurrence of a particular TRF wasassociated with a particular treatment, the Fisher’s exactprobability test was used to test the null hypothesis thattreatment and incidence of a TRF are independent. Toperform this analysis, a two-way contingency table offrequency data for each treatment and incidence status wasprepared. The Fisher’s exact statistic was calculated usingthe SAS procedure PROC FREQ with the subcommandEXACT FISHER. The two-sided P-value was used to testthe null hypothesis of independence.

2.7.3. Quantitative relationships between soil properties

and disease

Statistical correlation and multiple regression analyseswere performed on the soil and disease severity measures todetermine if a single or a combination of soil variables wasassociated with the variation in root rot severity docu-mented in the field and greenhouse experiments. Pearsoncorrelations (r) were calculated in Minitab v10 between soilproperties, disease severity (under field and greenhouseconditions), and crop yield across organic amendment typefor each split-plot application. Analysis of the data in thismanner provided a range of values for each measurement(property and severity) and a suitable number of indepen-dent measurements (n ¼ 16 per split-plot application).Using this dataset, best subsets regression and backwardselection step-wise regression was performed in SAS.Regression models that produced the lowest variance,highest R2, and where all of the independent variables (Xi)significantly (Po0.05) contributed to the variation indisease (R2) were selected. The relative importance of asoil variable, independent of other soil variables in themodel, was calculated by partitioning the variance in R2

(Selvin, 1995) and expressing this partial correlation as aproportion of R2 by the equation: % of model R2

¼

[Absolute value (single correlation coefficient (r)�standardized b-coefficient of Xi in the model)/(S absolutevalue (r� standardized b-coefficient of Xi)]� 100%; wherer ¼ correlation coefficient calculated for Xi and Y (disease)for each X-variable separately.

3. Results

3.1. Amendment effects on root rot disease in the field

and greenhouse

The split-plot design of the experiment provided a meansto compare common root rot severity in split-plots thatreceived 4 years (residual organic matter) or 5 years(residual plus new additions) of PMR amendments duringthe 5th year of study. Snap bean roots were sampled fromthe field and greenhouse experiments to estimate rootdisease severity (Fig. 1). Averaged across all amendmenttypes (including the non-amended soils) and split-plotapplications, disease severity approximated by a severityindex, was lower under field compared to greenhouseconditions (P ¼ 0.0005; disease index ¼ 0.78 and 1.42,respectively). Excessive watering to promote disease devel-opment in the greenhouse experiments likely contributed tothis general difference between environments. Nonetheless,the effect of amendment type, regardless of split-plotapplication, was apparently similar (P ¼ 0.19) between thetwo environments. Five years of amendment with fresh orcomposted PMR produced significantly less disease(P ¼ 0.001–0.004) than non-amended control soils underthe two environments. With the exception of the PMRCtreatment under field conditions, disease severity wassignificantly lower (P ¼ 0.001–0.02) in split-plots amended

ARTICLE IN PRESS

PMR PMRC PMRB NA

Roo

t Rot

Sev

erity

0.0

0.5

1.0

1.5

2.0

2.5

PMR PMRC PMRB NA

A-half split-plotN-half split-plot

*

Fig. 1. Effect of paper mill residuals and composts on severity of common root rot disease of snap bean grown under field (A) and greenhouse conditions

(B). Field plots historically amended for four consecutive growing seasons with fresh paper mill residuals (PMR), PMR composted without bark (PMRC),

PMR composted with bark (PMRB), and non-amended plots (NA) were divided in half length-wise to achieve two split-plots. One of the split-plots

received new amendment during the fifth growing season (A-half), while its companion half remained non-amended during that year (N-half). Non-

amended (NA) plots were split in the same manner to mimic amendment. The NA bar represents the mean of the two split-plots of this treatment.

* ¼ statistical significant difference between pairs of bars within a treatment at Po0.05 level. Each bar represents a mean of four replicate plots and the

standard deviation of the mean.

Table 1

Re-amendment (A) and residual (N) effects of organic amendmentsa on

soil properties in field plots

Soil propertyc Split-plot application treatmentsb

PMR PMRC PMRB NA

A N A N A N

TC (Mgha�1) 21.6 19.4 30.0* 24.2 38.2* 28.0 9.5

TN (Mgha�1) 1.4 1.3 1.9* 1.7 1.8* 1.5 0.8

POM-C (Mgha�1) 7.4 6.6 13.4 10.4 27.0*

21.9 5.9

POM-N (Mgha�1) 0.4 0.4 0.7 0.6 1.1 1.0 0.4

VM (mlL�1 soil) 17.0 13.6 19.1* 13.6 23.3* 18.5 12.7

NO3-N (mgL�1 soil)d 9.0 0.5 4.2 1.4 0.8 0.4 2.9

Values in bold indicate split-plot treatment means (n ¼ 4) that are

significantly different (Po0.05) from the non-amended control treatment

for a given soil property.

Values followed by an asterisk (*) indicate a significant difference

(Po0.05) between the two split-plot applications (A and N) of a given

D. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–2948 2941

for 4 years, but left un-amended in the 5th year (N),compared to the non-amended soils. On average, there wasa marginal difference between split-plot applications(P ¼ 0.07) regardless of amendment type; however, thisdifference was attributed to the difference between splitapplications of PMRC in the field (P ¼ 0.02).

Root rot reduced snap bean yields in the non-amendedcontrol plots (data not shown). In split-plots that receivednew material in the 5th year, yields increased 56–83%(Po0.05) over the control treatment. Re-application of thefresh PMR produced the greatest yields. Both compostsproduced yield-enhancing effects 15 months after the soilreceived the amendment from the previous season. Class 5bean yields (canning beans) were greatest in the plots thatreceived 5 years of fresh PMR, while the lowest yieldingtreatments were the non-amended controls. Re-applicationof the two composts increased the proportion of beans inthe canning size category of bean.

amendment type.aAmendments were fresh paper mill residuals (PMR), PMR composted

alone (PMRC), PMR composted with bark (PMRB); NA ¼ non-amended

control (NA).bFollowing 4 years of spring additions of organic amendment, field

plots were divided in half length-wise to achieve two split-plots. One of the

split-plots for each amendment treatment received fresh amendment in the

5th year of study (A), while its companion half remained non-amended (N)

that year. NA plots were split in the same manner to mimic amendment

plots.cSoils were retrieved from split-plots 1-month (May) following

amendment to conduct greenhouse root rot experiments and to measure

soil properties, with the exception of NO3-N, which was measured during

the middle of the snap bean season (July). TC ¼ total soil carbon;

TN ¼ total soil N; POM-C ¼ particulate organic matter carbon; POM-

N ¼ particulate organic matter nitrogen; VM ¼ volumetric moisture;

NO3-N ¼ nitrate-nitrogen in the soil solution surrounding plant roots.dPlant available nitrogen (NO3-N) was measured in situ with the use of

five pairs of ion exchange membranes (cation/anion) placed adjacent to

and between plants within two 7.7-m rows per split-plot.

3.2. Soil properties in split field plots

Soil properties were determined for individual split-plots1 month after amendment to approximate the baselinesoil conditions for the greenhouse root rot experiments(Table 1). Regardless of re-amendment for the 5th year,split-plots with a history of PMRB-amendment containedsignificantly more (P ¼ 0.01) total and POM fractioncarbon compared to the non-amended control soils. A5th year of amendment of either compost significantlyincreased carbon stocks in the soil from the previousyear (Po0.01; 24% and 36 % increase in TC for PMRC-Aand PMRB-A compared to PMRC-N and PMRB-N,respectively). These increases in the amount of TC inthe soil were associated with 40% and 23% increases

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–29482942

(Po0.05) in soil moisture for the PMRC-A and PMRB-Atreatments over their non-amended counterparts (PMRC-N and PMRB-N), respectively. The average amount ofPOM-C in soils re-amended with fresh PMR was margin-ally higher compared to the amount in non-amended soils,indicating that the carbon in the POM-fraction of PMR-amended soils was fairly labile compared to the compost-amended soils (supported by previous findings—Newmanet al., 2005).

3.3. Total microbial activity (FDA)

Total microbial activity, estimated by the rate ofFDA hydrolysis, was measured for the amended andnon-amended field soils to determine the quantitativerelationship between activity and root rot severity. In re-amended split-plots (A), activity was three fold greater(Po0.01) compared to non-amended control soils, regard-less of amendment type (0.9 vs. 0.3 mgFDAmin�1 g�1 soil).Activity was significantly lower (P ¼ 0.02) in residual-plots(N-half) compared to their re-amended counterparts(A-half); however, the average activity in these plotswas significantly higher (P ¼ 0.05) than the activity in thenon-amended plots (0.6 mg compared to 0.3 mgFDAmin�1 g�1 soil). There was a significant negativecorrelation (r ¼ �0.74, P ¼ 0.001) between total microbialactivity and root rot severity documented in the green-house for a dataset including re-amended treatments andcontrol soils (Fig. 2A). The linear relationship was drivenby the difference between non-amended and amended soils,which was evident by the apparent lack of a significantcorrelation (P40.25) upon removal of the non-amendedtreatment from the analysis. In the split-plots that receivedamendment the previous year, there was no apparentcorrelation (r ¼ �0.19, P ¼ 0.47) between activity anddisease severity (Fig. 2B).

Total Microbial Activity (ug of hydro

0.0 0.2 0.4 0.6 0.8 1.0 1.

Roo

t Rot

Sev

erity

0.5

1.0

1.5

2.0

2.5

3.0

PMRBPMRPMRCNo Amendment

Fig. 2. Relationship between soil microbial activity and severity of common ro

plots historically amended for four consecutive growing seasons with fresh pa

composted with bark (PMRB) were divided in half length-wise to achieve two

the spring of the fifth growing season (A), while its companion half remained

3.4. Effect of organic amendment on bacterial community

profiles associated with roots

T-RFLP analysis of PCR-amplified 16s rDNA sequencesrevealed a suite of 21 possible TRFs. PCA of the binarydata (presence or absence of a TRF) generated two PCs(PC1 and PC2) which collectively explained 45% of thevariation among samples; however, only PC1 was sig-nificantly influenced by treatments (Fig. 3A). Whilecommunity profiles described by PC1 varied with locationin the field (i.e., blocks) for any given treatment in the field,the variation between treatments exceeded the variationwithin treatments. There was a significant interactionbetween amendment type and split-plot application(P ¼ 0.05) on community composition described by theincidence of TRFs, allowing for pairwise comparisonsbetween all treatments. Non-amended control soils sig-nificantly varied in community composition from PMR(Po0.001) or PMRC-amended soils (Po0.02), regardlessof re-amendment for a 5th year. Community profiles weresimilar between non-amended control soils and plotshistorically amended with bark compost (PMRB-N,P ¼ 0.15) or freshly amended with this material (PMRB-A,P ¼ 0.43). Addition of new PMR or PMRC amendmentapparently did not alter community composition from theirnon-amended counterparts (PMR-N or PMRC-N). Basedon their loadings (correlation coefficients), nine of the 21possible TRFs significantly contributed to the differencesin the community profiles among organic amendmenttreatments (Table 2). Pairwise comparisons (Fisher’s exacttest of independence) revealed significant differences(Po0.01) between treatments for all of the TRFs thatcontributed to the distribution of treatments along PC1.Terminal fragment 613 was associated with both split-plotapplications (A and N) of PMR and PMRC only, while theremaining eight TRFs were associated with roots growing

lyzed fluorescen diactetate min-1 g-1 soil)

2 0.0 0.2 0.4 0.6 0.8 1.0 1.2

ot rot of snap bean grown in field soils under greenhouse conditions. Field

per mill residuals (PMR), PMR composted without bark (PMRC), PMR

split-plots. One of the split-plots received new amendment (re-amended) in

non-amended during that year (B).

ARTICLE IN PRESS

PC1 (34%)

PC

2 (

11

%)

0.8

0.8

-0.8

-0.8

0.6

0.6

-0.6

-0.6

0.4

0.4

-0.4

-0.4

0.2

0.2

-0.2

-0.2

0.0

0.0

1.0

1.0 1.2 1.4

-1.0

0.8

-0.8

0.6

-0.6

0.4

-0.4

0.2

-0.2

0.0

1.0

-1.0

-1.0-1.2-1.4

0.8-0.8 0.6-0.6 0.4-0.4 0.2-0.2 0.0 1.0 1.2 1.4-1.0-1.2-1.4

123124

128

323324

426

427428

613

PC1 (40%)

PC

2 (

16

%)

123

426427

613

PMR-A

PMRC-A

PMR-N

PMRC-N

PMRB-NPMRB-A

NA

PMR-A

PMRC-A

PMR-N

PMRC-N

PMRB-NNA

PMRB-A

Fig. 3. Re-amendment and residual effects of organic amendment on bacterial community structure described by the presence/absence (A) and relative

abundance (B) of terminal restriction fragments (TRFs) in the rhizosphere of snap bean grown in field soils. Amendment types included paper mill

residuals (PMR), PMR composted without bark (PMRC), PMR composted with bark (PMRB), and no amendment. Field plots were divided in half

length-wise to achieve two split-plots. One of the split-plots received new amendment in the spring of the fifth growing season (designated—A), while its

companion half remained non-amended during that year (designated—N). These bi-plots simultaneously project PC scores (PC1 vs. PC2) and loadings

(represented by TRF lengths 123–613). The proximity of an amendment treatment and a TRF indicates their level of association. Each point represents the

mean (n ¼ 4) and standard error of the mean.

D. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–2948 2943

in non-amended and PMRB-amended soils. PCA of areaunder the peak for each TRF (estimate of relativeabundance) generated two PCs that described 56% of thevariation among treatments (Fig. 3B); PC1 was onlymodestly influenced by treatments (Po0.09), likely dueto the large variation within treatment groups. Terminalfragment 613 significantly contributed to the separationamong treatments along PC1 (Table 2), where it was mostabundant in PMR-amended soils (Po0.05).

3.5. Univariate relationships between soil properties

and disease

Statistical correlations between soil properties and rootrot severity measured in the field and companion green-house experiments were determined. Correlation analysisof the field and greenhouse data revealed three interestingpatterns (Table 3). First, the direction and statisticalsignificance of the correlation between any soil variable

ARTICLE IN PRESS

Table 2

Principal component loadingsa of terminal restriction fragments (TRFs)

that significantly contribute to differences in root-associated bacterial

community profiles among treatmentsb

Loading variable

(TRF)

PC1 (34%)c

(TRF incidenced)

PC1 (40%)

(TRF abundancee)

123f 0.70g 0.47

124 0.77 ns

128 0.48 ns

323 0.44 ns

324 0.57 ns

426 0.84 0.42

427 0.86 ns

428 0.54 ns

613 �0.50 �0.98

ns ¼ not significant (P40.05).aA loading is the correlation between the original loading variable

(incidence or abundance of a TRF) and the principal component score

(PC1) of a treatment sample.bTreatments included fresh paper mill residuals (PMR), PMR

composted with bark (PMRB), and PMR composted without bark

(PMRC); materials were incorporated annually to field plots for 5 years

(A split-plot) or 4 years and left un-amended during the 5th year of study

(N split-plot). Non-amended controls were included in the analysis.cProportion of total variation in the data sets explained by principal

component 1 (PC1).dPrincipal components analysis (PCA) was conducted on a binary

matrix representing the presence or absence of each TRF in each sample

profile (i.e. community composition).ePCA was conducted on a data matrix consisting of the relative

abundance of each TRF in sample profiles. Relative abundance of a

particular TRF was approximated by the proportion of total area under

the TRF peaks of a given profile.fTerminal restriction fragment lengths (base pairs) generated by RsaI

digestion of the PCR-amplified 16s rDNA sequences associated with snap

bean roots.gA large negative or positive correlation coefficient indicates the

importance of a TRF variable on the differences among treatments.

Table 3

Correlations between soil properties and severity of common root rot of

snap bean grown in the field and in field soil under greenhouse conditions

Soil propertya Field Greenhouse

Ab N A N

TC �0.57 �0.67 �0.71 �0.66

TN �0.59 �0.63 �0.72 �0.78

TC:N �0.55* �0.63 �0.68 �0.49*

POM-C �0.34 �0.39 �0.39 �0.36

POM-N �0.32 �0.39 �0.34 �0.32

POMC:N �0.40* �0.61 �0.57 �0.59

VM �0.56* �0.33 �0.64 �0.10

NO3-N �0.12 0.77 �0.10 0.69

FDA �0.69 �0.55 �0.54 �0.12

PC1 0.57 0.77 0.56* 0.55*

PC2 �0.27 �0.29 �0.41 0.02

Pearson’s correlation coefficient calculated for datasets include all

amendment types (PMR, PMRC, PMRB, and NA) (n ¼ 16 plots).

Coefficients (r) in bold indicate significant correlations between disease

severity and a soil property at Po0.05 level.*Significant correlations at Po0.1 level.

aTC ¼ total soil carbon; TN ¼ total soil nitrogen; TC:N ¼ total soil

C:N ratio; POM-C ¼ particulate organic matter carbon; POM-

N ¼ particulate organic matter nitrogen; POMC:N ¼ particulate organic

matter C:N ratio; VM ¼ volumetric moisture; NO3-N ¼ nitrate nitrogen

in the soil solution surrounding plant roots; FDA ¼ rate of hydrolysis of

fluorescein diacetate (measure of total soil microbial activity); PC1

(principal component 1) and PC2 (principal component 2) ¼ summary

variables representing bacterial community composition associated with

roots as determined by T-RFLP analysis.bField plots were divided in half length-wise to achieve two split-plots.

One of the split-plots for each amendment type (fresh paper mill residuals

(PMR), paper mill residuals composted without bark (PMRC), and PMR

composted with bark (PMRB)) received fresh amendment in the 5th year

of study (A), while its companion half remained non-amended (N) that

year (4 years of amendment). Non-amended (NA) plots were included in

the analyses.

D. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–29482944

and disease severity measured under field conditions wasmirrored in the companion greenhouse experiments,regardless of amending the soil with 5 (A) or 4 (N)consecutive years. Out of 44 correlations, only one pairwisecomparison (disease and microbial activity in (N) plots)produced a different outcome between the two environ-ments. Second, the community composition summaryvariable that explained the majority of the variation amongtreatments (PC1) was correlated positively (Po0.01–0.06)with severity (low PC1 scores tended to be related to moredisease-suppressive soils (Fig. 3)). Third, the C:N ratio inthe POM fraction (an indicator of the biochemical qualityof POM), and not POM-C alone, was significantlycorrelated with disease (P ¼ 0.01–0.05). The significantvariation in POM-C (coefficient of variation (cv) ¼3–69%) and POM-N (cv ¼ 2–57%) within treatmentsmay have masked the relationship. Calculating the ratioof POM-C to POM-N reduced this intra-treatment varia-tion (cv ¼ 6–20%).

Several correlations between the measured variableswere also notable. Regardless of split-plot application,there was a negative relationship between root rot severity

and snap bean yield (r ¼ �0.62 to �0.71; Po0.05).Bacterial community profiles described by PC1 weresignificantly correlated (r ¼ �0.656, P ¼ 0.02 for re-amended split-plot; r ¼ 0.519, P ¼ 0.08 for un-amendedsplit-plot) with the amount of NO3-N in the soil solutionssurrounding plant roots. In freshly amended soils, theposition of the treatment along PC1 correlated withthe amount of NO3-N associated with each treatment(high N, negative PC score; low N, positive PC score(Fig. 3, Table 1)).

3.6. Multivariate relationship between soil properties and

disease severity

Multiple regression analyses were performed to identifya suite of soil properties that collectively explained thevariation in root rot severity documented in the field andgreenhouse experiments. Regression models that bestdescribed the variation in disease included rhizobacterialcommunity composition (PC1 for incidence of TRFs) andC:N ratio in the POM-fraction of the bulk soil. This findingwas consistent between the two split-plot applications

ARTICLE IN PRESS

Table 4

Multiple regression (MR) modelsa that best explain root rot severity and the relative importance of soil properties that contribute significantly (Po0.05) to

disease in each model

Split-plotb MR model R2 (%) Xic VIFd % of model R2e

A 3.47+0.48 (PC1)–0.11 (POMC:N) 80 PC1 1.0 49

POMC:N 1.0 51

N 2.56+0.16 (PC1)–0.05 (POMC:N) 43 PC1 1.1 46

POMC:N 1.1 54

aMultiple regression analysis was performed to determine the best model for disease (highest R2 and lowest variance) by using best subsets regression

combined with step-wise backward elimination regression analysis.bField plots were divided in half length-wise to achieve two split-plots. One of the split-plots for each amendment treatment received fresh amendment in

the 5th year of study (A), while its companion half remained non-amended (N) that year (4 years of amendment). NA plots were split in the same manner

to mimic amendment plots.cIndependent variables (Xi) that significantly explain the variability in disease in the multiple regression models. PC1 (principal component 1) ¼

summary variable representing bacterial community composition associated with roots as determined by T-RFLP analysis; POMC:N ¼ carbon to

nitrogen ratio in the particulate organic matter (POM) fraction of soil.dVariance inflation factor. Values closest to 1 signify low multicollinearity of an X-variable in the multiple regression model.eThe proportion of the variation (% of R2) in Y (disease) explained by each X-variable independent of all other X-variables in the multiple regression

models. % of R2¼ {Absolute value [single correlation coefficient (r)� standardized b-coefficient of X in multiple regression model]� 100%} divided by

the sum of the absolute value [r�b-coefficient] of each Xi variable in the model.

D. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–2948 2945

(Table 4) and the two growth environments (data notshown). In general, the combined effects of communitycomposition and POM C:N in the soil influenced theamount of disease (or disease suppression) to a greaterextent than any single soil variable alone (Table 3). Therelative contributions of the two soil variables on diseaseseverity were closely matched (% of model R2, Table 4).

4. Discussion

Various farm management practices can influence theincidence and severity of soil-borne diseases. In this study,conventionally managed vegetable rotation plots withhistories of receiving five consecutive spring additions offresh or composted PMR maintained low root rot diseasepressure as compared to non-amended field plots, and as aresult, produced higher snap bean yields during the 5thyear. The disease-suppressive condition of the soils at thissite appeared to be initiated between the second and thirdannual amendment event as documented in greenhousebioassays (Leon et al., 2006), much like the PMR-mediatedroot rot suppression documented after two annual inputsat a field site continuously cropped to snap bean (Stoneet al., 2003). In the present study, re-amendment in the 5thyear was not required to achieve root rot suppression(N-half, Fig. 1). This finding provides evidence thatsuppressive soil conditions were maintained in the absenceof freshly applied organic inputs. Interestingly, by the endof the third amendment year, labile carbon pools(estimated by the POM-C fraction) in PMR and PMRC-amended soils had reached elevated steady states, indicat-ing a net balance between C-input and C-decay (Newmanet al., 2005). This new equilibrium supports the hypothesisthat populations of microflora in these soils adapted to theform of C-inputs (high cellulose, wood fibers) and theirdecomposition products after repeated exposure to the

organic materials, resulting in a shift in communitycomposition (Sugai and Schimel, 1993; Balser and Fire-stone, 2002). In addition, retention of carbon in the POMfraction of these soils may have provided the ‘continuous’substrate required to sustain root rot suppression (Hoitinkand Boehm, 1999; Stone et al., 2001) in the absence ofnewly incorporated amendment.Suppression of root rot in the present study was not

universally linked to microbial activity in these soils. Wholesoil microbial activity in the present study was approxi-mated by the rate of hydrolysis of FDA incorporated intothe field soils. This enzymatic assay is used to quantify thecollective extracellular activities of diverse soil microflorathat use lipase, protease, or esterase to hydrolyze organiccompounds. Because the rate of FDA hydrolysis has beenassociated with general suppression of soil-borne diseases(Hoitink and Boehm, 1999; van Bruggen and Semenov,2000; Darby et al., 2006), we hypothesized that there is aquantitative, inverse relationship between root rot severityand microbial activity in these sandy soils. Soil microbialactivity was boosted in plots following a recent incorpora-tion of PMR amendments (Fig. 2A) and the relationshipbetween root rot severity and activity was inverse, yetqualitative in nature (high activity/low disease, lowactivity/high disease) due to the binary grouping of non-amended plots vs. amended plots. In the absence of newlyadded amendments (Figs. 2B, 1 year after the 4thamendment), soils obtained from half of the field plotswere root rot suppressive, yet microbial activities in thesesoils were comparable to those measured in soils that hadnever received organic amendments. Other researchershave documented a disconnect between suppressive soilsand microbial activity (van Bruggen and Grunwald, 1996;Leon et al., 2006; Perez-Piqueres et al., 2006). For example,Perez-Piqueres et al. (2006) determined that only 50% ofthe soil/compost combinations were both suppressive to

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–29482946

root disease caused by Rhizoctonia solani (AG2-2) andfunctioning at high respiration rates (measure of soilmicrobial activity). It is apparent that root rot suppressionin the present study was mediated by factors other than thelevel of whole soil microbial activities.

One primary objective of the present study was toexamine the influence of fresh and composted PMRamendments on composition and structure of root andrhizobacterial communities of snap bean in a vegetablerotation. The rhizosphere of crop plants is a dynamicenvironment and root-colonizing bacteria play significantroles in plant health, growth, and protection againstinvading pathogens (Weller, 1988; McSpadden Gardener,2007). Soil microbial communities have been shown to bealtered by various farm management strategies (Blackwoodand Paul, 2003; Perez-Piqueres et al, 2006; Widmer et al.,2006; Benıtez et al., 2007), and changes in soil populationsmight reasonably be expected to alter rhizosphere commu-nity structure to some degree. In fact, agriculturalstrategies known to alter SOM pools, such as crop rotation(Peters et al., 2003), organic mulch applications (Tiquiaet al, 2002), and compost amendments (Perez-Piquereset al., 2006), have been shown to modify bacterialcommunity structure in the rhizosphere of agronomiccrops. In the present study, consecutive additions of twoforms of the organic amendment (fresh PMR and PMRC)to their respective soils altered rhizobacterial communitiesin general (T-RFLP profiles) compared to non-amendedand PMRB-amended soils. With the use of PCA, asummary variable (PC1) representing the communityfingerprint (composition or abundance of TRFs associatedwith roots) was generated; statistical analysis of PC1confirmed that amendment type significantly influencedcommunity composition (presence/absence of TRFs), buthad no significant effect on the abundance of subcompo-nents in the community (Fig. 3B). The abundance ofparticular TRFs associated with roots varied widely in soilsrecovered across the 0.6 acre field site, and withinparticular treatments, which may have contributed to thelack of power to resolve differences due to treatments.

The composition of the organic amendments (C-quality)applied in this study differentially influenced generalcommunity profiles (Fig. 3A). This finding is consistentwith the previous reports that document the effect ofdecomposition level of composts in potting mixes (Boehmet al., 1993) or mulches applied to soil surfaces (Tiquiaet al., 2002) on rhizosphere bacterial community structure.In the present study, the composition of rhizosphere androot-colonizing bacterial communities in soils amendedwith PMR and PMRC differed from communities asso-ciated with roots grown in PMRB-amended soils. FreshPMR is composed of primary and secondary by-product ofthe paper-manufacturing process. Compared to its com-posted form, it is relatively labile, decaying at a more rapidrate (Foley and Cooperband, 2002) in the soil. Paper millresiduals composted with bark contain large wood fibersthat partially degrade during the composting process; as

such, these lignin-rich components are more recalcitrantand have a longer resident time in sandy soils. Compostingof PMR without a bulking agent (PMRC) results inanaerobic decomposition and produces a less stabilizedform of compost. In a previous report, decomposition ofPMR and PMRC supplied significantly more NO3-N overthe growing season to the snap bean crop compared toPMRB (Rotenberg et al., 2005). Because each amendmenttype was composed of variable amounts of mineral andorganic N at the time of application (Rotenberg et al.,2005) and produced different effects on decomposition ofOM in these soils, we hypothesize that variable Nmineralization influenced the microbial community com-position in the rhizosphere/rhizoplane of snap bean.Indirect support of this hypothesis is the finding thatthe only single soil variable that correlated with PC1 wasNO3-N (univariate analyses).Two statistical approaches (PCA and Fisher’s exact test)

revealed the presence of a subset of TRFs that contributedmost to the community compositional differences amongPMR treatments (Table 2). For the majority of theseinfluential TRFs, the proportion of snap bean root samplesharboring a TRF was greater in soils amended with PMRBand non-amended soils compared to the other two PMRforms, illustrated by the proximity of these TRFs (PCAloadings) to the treatments (PC scores) in the bi-plots(Fig. 3). Since abundance of these TRFs was unaffected bythe bark compost, our findings indicate that the suppres-sive nature of the PMRB-amended soils was due, in part, toanother soil factor. Perez-Piqueres et al. (2006) reportedchanges in fungal communities in soil following amend-ment with spent mushroom compost. Perhaps wood-colonizing fungal communities and their byproductscontributed to disease suppression in the present study.One TRF (613) was found exclusively associated withPMR and PMRC, regardless of a recent re-amendmentevent. The bacteria, represented by this TRF, may haveoriginated from the fresh PMR amendment itself. Underanaerobic conditions, decomposition of PMR (i.e. PMRC)likely proceeded under mesophilic conditions, and thus it ispossible that a subset of amendment-associated organismssurvived the composting process. Another possible scenar-io is consistent with substrate-dependent disease suppres-sion (Boehm et al., 1993), a case where PMR amendmentsdifferentially altered SOM quality to select for bacteria bestequipped to utilize the chemical constituents in thesematerials.Each TRF in the community profile is an operational

taxonomic unit, and an individual TRF often representsmultiple bacterial species that produce the same restrictionfragment size of a specific region of DNA. While the nextlogical step would be to sequence the TRFs obtained in thisstudy, we used an on-line tool called Microbial CommunityAnalysis III (MiCA3) (http://mica.ibest.uidaho.edu; Shyuet al., 2007) to perform a virtual digest with RsaI of thepredicted PCR products generated from sequences foundon the Ribosomal Database Project II (RDPII, Release 9,

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–2948 2947

Update 37, bacterial SSU 16s rRNA) with the primer pairused in our study. As expected, we found multiple matchesfor each TRF, ranging from 2 to 49 different possiblecandidates, further justification for cloning and sequencingfor correct identification of the putative member of thecommunity.

Biotic and abiotic components of soil mediate soilfunction and contribute to crop health. The challenge liesin our ability to differentiate between the effects ofbiological and physicochemical properties on diseasesuppression. In this study, multiple regression analyseswere performed to determine the relative contributions ofsoil properties to the variation in disease severity in fieldsoils annually amended with three forms of PMR amend-ments. We found that rhizobacterial community composi-tion of snap bean and the biochemical quality of POM(C:N ratio) in bulk soils imparted independent andrelatively equal contributions to the amount of root rotdisease experienced in these plots; together these variablesexplained more of the variation in disease than any one ofthe nine soil variables measured. The two-componentmodel adequately described the variation in root rotseverity in the greenhouse and field, and in soils amendedfor 4 or 5 consecutive years, underscoring the robust natureof the relationship under different environmental condi-tions. In these sandy soils, the POM C:N ratio representsthe amount of carbon in the POM fraction and thebiochemical composition of the POM fraction as affectedby the different PMR forms (Newman et al., 2005).Collectively, our findings provide evidence that annualPMR inputs facilitated the development of soil-bornedisease suppression in these soils, which was linked to thedifferential effects of PMR amendments on both root-associated bacterial communities and the amount andquality of organic matter in these soils.

Acknowledgments

This project was supported by the National ResearchInitiative of the USDA Cooperative State Research,Education and Extension Service, Grant no. 2001-35316-11026, and the Wisconsin Potato Vegetable Growers’Association. We thank the University of Wisconsin’sHancock Agricultural Research Station (ARS) and theWest Madison ARS crews for their assistance with cropplanting and harvesting, irrigation scheduling, and com-posting. We also thank Heather Darby for providing uswith the FDA protocol.

References

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, R.E., Seidman, J.G.,

Smith, J.A., Struhl, K., 2003. Current Protocols in Molecular Biology,

vol. 1. Wiley, Inc., New York.

Balser, T.C., Firestone, M.K., 2002. Linking soil microbial communities

and ecosystem functioning. In: Kinzig, A., Pacala, S., Tilman, D.

(Eds.), Biodiversity and Ecosystem Functioning. Princeton University

Press, Princeton, NJ, pp. 265–293.

Benıtez, M., Tustas, F.B., Rotenberg, D., Kleinhenz, M.D., Cardina, J.,

Stinner, D., Miller, S.A., McSpadden Gardener, B.B., 2007. Multiple

statistical approaches of community fingerprint data reveal bacterial

populations associated with general disease suppression arising from

the application of different organic field management strategies. Soil

Biology & Biochemistry 39, 2289–2301.

Blackwood, C.B., Marsh, T., Sang-Hoon, K., Paul, E.A., 2003. Terminal

restriction fragment length polymorphism data analysis for quantita-

tive comparison of microbial communities. Applied Environmental

Microbiology 69, 926–932.

Boehm, M.J., Madden, L.V., Hoitink, H.A.J., 1993. Effect of organic

matter decomposition level on bacterial species diversity and

composition in relationship to Pythium damping-off severity. Applied

Environmental Microbiology 59, 4171–4179.

Chen, W., Hoitink, H.A.J., Madden, L.V., 1988. Microbial activity and

biomass in container media for predicting suppressiveness to damping-

off caused by Pythium ultimum. Phytopathology 78, 1447–1450.

Clement, B.G., Kehl, L.E., DeBord, K.L., Kitts, C.L., 1998. Terminal

restriction fragment patterns (TRFP’s), a rapid PCR based method for

the comparison of complex bacterial communities. Journal of

Microbiological Methods 31, 135–142.

Cook, R.J., Baker, K.F., 1983. The Nature and Practice of Biological

Control of Plant Pathogens. The American Phytopathological Society,

St. Paul, MN, 539pp.

Darby, H.M., Stone, A.G., Dick, R.P., 2006. Compost and manure

mediated impacts on soilborne pathogens and soil quality. Soil Science

Society of America Journal 70, 347–358.

Foley, B.J., Cooperband, L.R., 2002. Paper mill residuals and compost

effects on soil physical properties, soil fertility, and crop production.

Journal of Environmental Quality 31, 2086–2095.

Fortuna, A., Harwood, R.R., Paul, E.A., 2003. The effects of compost

and crop rotations on carbon turnover and the particulate organic

matter fraction. Soil Science 168, 434–444.

Griffin, T.S., Porter, G.A., 2004. Altering soil carbon and nitrogen stocks

in intensively tilled two-year rotations. Biology and Fertility of Soils

39, 366–374.

Hoitink, H.A.J., Boehm, M.J., 1999. Biocontrol within the context of soil

microbial communites: a substrate-dependent phenomenon. Annual

Review of Phytopathology 37, 427–446.

Hu, S., van Bruggen, A.H.C., Wakeman, R.J., Grunwald, N.J., 1997.

Microbial suppression of in vitro growth of Pythium ultimum and

disease incidence in relation to soil C and N availability. Plant and Soil

195, 43–52.

Kannwischer, M.E., Mitchell, D.J., 1978. The influence of a fungicide on

the epidemiology of black shank of tobacco. Phytopathology 68,

1760–1765.

Kobriger, K.M., Hagedorn, D.J., Stevenson, W.R., 1998. Analysis of the

snap bean root rot potential of Wisconsin fields. Cooperative

Extension Programs, University of Wisconsin, A3242 Bulletin, 4pp.

Leon, M.C.C., Stone, A., Dick, R.P., 2006. Organic soil amendments:

impacts on snap bean common root rot (Aphanomyces euteiches) and

soil quality. Applied Soil Ecology 31, 199–210.

Liu, L., Marsh, T.L., Cheng, H., Forney, L.J., 1997. Characterization of

microbial diversity by determining terminal restriction fragment length

polymorphism of genes encoding 16S rRNA. Applied and Environ-

mental Microbiology 63, 4516–4522.

Lukow, T., Dunfield, P.F., Liesack, W., 2000. Use of the T-RFLP

technique to assess spatial and temporal changes in the bacterial

community structure within an agricultural soil planted with

transgenic and non-transgenic potato plants. FEMS Microbiology

Ecology 32, 241–247.

McKellar, M.E., Nelson, E.B., 2003. Compost-induced suppression of

Pythium damping-off is mediated by fatty-acid-metabolizing seed-

colonizing microbial communities. Applied and Environmental Micro-

biology 69, 452–460.

McSpadden Gardener, B.B., 2007. Diversity and ecology of biocontrol

Pseudomonas spp. in agricultural systems. Phytopathology 97,

221–226.

ARTICLE IN PRESSD. Rotenberg et al. / Soil Biology & Biochemistry 39 (2007) 2936–29482948

Newman, C.M., Rotenberg, D., Cooperband, L.R., 2005. Paper mill

residuals and compost effects on particulate organic matter and related

soil functions in a sandy soil. Soil Science 170, 788–801.

Perez-Piqueres, A., Edel-Hermann, V., Alabouvette, C., Steinberg, C.,

2006. Response of soil microbial communities to compost amend-

ments. Soil Biology & Biochemistry 38, 460–470.

Peters, R.D., Sturz, A.V., Carter, M.R., Sanderson, J.B., 2003. Develop-

ing disease-suppressive soils through crop rotation and tillage manage-

ment practices. Soil and Tillage Research 72, 181–192.

Reeleder, R.D., Hagedorn, D.J., 1981. Pythium populations in Wisconsin

bean fields. Canadian Journal of Plant Pathology 3, 90–93.

Rotenberg, D., Joshi, R., Benıtez, M., Chapin, L.G., Camp, A., Zumpetta, C.,

Osborne, A., Dick, W.A., McSpadden Gardener, B.B., 2007. Farm

management effects on rhizosphere colonization by native populations of

2,4-diacetylphloroglucinol-producing Pseudomonas spp. and their con-

tributions to crop health. Phytopathology 97, 756–766.

Rotenberg, D., Cooperband, L.R., Stone, A., 2005. Dynamic relationships

between soil properties and foliar disease as affected by annual

additions of organic amendment to a sandy-soil vegetable production

system. Soil Biology & Biochemistry 37, 1343–1357.

Shyu, C., Soule, T., Bent, S.J., Foster, J.A., Forney, L.J., 2007. MiCA: a

web-based tool for the analysis of microbial communities based on

terminal-restriction fragment length polymorphisms of 16s and 18s

rRNA genes. Microbial Ecology 53, 562–570.

Selvin, S., 1995. Practical Biostatistical Methods. Duxbury Press,

Belmont, CA, 528pp.

Sikora, L.J., Yakovchenko, V., Cambardella, C.A., Doran, J.W., 1996.

Assessing soil quality by testing organic matter. In: Magdoff, F.R.,

Tabatabai, M.A., Hanlon, Jr., E.A. (Eds.), Soil Organic Matter:

Analysis and Interpretation. Soil Science Society of America,

Madison, WI, pp. 41–50.

Stone, A., Traina, S., Hoitink, H., 2001. Particulate organic matter

composition and Pythium damping-off of cucumber. Soil Science

Society of America Journal 65, 761–770.

Stone, A.G., Vallad, G.E., Cooperband, L.R., Rotenberg, D., Darby, H.R.,

James, R.V., Stevenson, W.R., Goodman, R.M., 2003. Effect of

organic amendments on soilborne and foliar diseases in field-grown

snap bean and cucumber. Plant Disease 87, 1037–1042.

Sugai, S.F., Schimel, J.P., 1993. Decomposition and biomass incorpora-

tion of 14C-labeled glucose and phenolics in taiga forest floor effect of

substrate quality, successional state, and season. Soil Biology &

Biochemistry 25, 1379–1389.

Tiquia, S.M., Lloyd, J., Herms, D.A., Hoitink, H.A.J., Michel, F.C., 2002.

Effects of mulching and fertilization on soil nutrients, microbial

activity and rhizosphere bacterial community structure determined by

analysis of TRFLPs of PCR-amplified 16S rRNA genes. Applied Soil

Ecology 21, 31–48.

Trotha, R., Reichl, U., Thies, F.L., Sperling, D., Konig, W., Konig, B.,

2002. Adaption of a fragment analysis technique to an automated

high-throughput multicapillary electrophoresis device for the precise

qualitative and quantitative characterization of microbial commu-

nities. Electrophoresis 23, 1070–1079.

Vallad, G.E., Cooperband, L.R., Goodman, R.M., 2003. Plant foliar

disease suppression mediated by composted forms of paper mill

residuals exhibits molecular features of induced resistance. Physiolo-

gical and Molecular Plant Pathology 63, 65–77.

van Bruggen, A.H.C., Grunwald, N.J., 1996. Test for risk assessment of

root infections by plant pathogens. In: Doran, J.W., Jones, A.J.

(Eds.), Methods for Assessing Soil Quality. SSSA, Madison, WI,

pp. 293–310.

van Bruggen, A.H.C., Semenov, A.M., 2000. In search of biological

indicators for soil health and disease suppression. Applied Soil

Ecology 15, 13–24.

Wander, M.M., Bollero, G.A., 1999. Soil quality assessment of tillage

impacts in Illinois. Soil Science Society of America Journal 63,

961–971.

Weil, R.R., Magdoff, F., 2004. Significance of soil organic matter to soil

quality and health. In: Magdoff, F., Weil, R. (Eds.), Soil Organic

Matter in Sustainable Agriculture. CRC Press, Boca Raton, USA,

pp. 1–44.

Weller, D.M., 1988. Biological control of soilborne plant pathogens in the

rhizosphere with bacteria. Annual Review of Phytopathology 26,

379–407.

Weller, D.M., Raaijmakers, J.M., McSpadden Gardener, B.B.,

Thomashow, L.S., 2002. Microbial populations responsible for specific

soil suppressiveness to plant pathogens. Annual Review of Phyto-

pathology 40, 309–348.

Whipps, J., 2001. Microbial interactions and biocontrol in the rhizo-

sphere. Journal of Experimental Botany 52, 487–511.

Widmer, F., Rasche, F., Hartmann, M., Fliessbach, A., 2006. Community

structures and substrate utilization of bacteria in soils from organic

and conventional farming systems of the DOK long-term field

experiment. Applied Soil Ecology 33, 294–307.

Yin, B., Valinsky, L., Gao, X.B., Becker, J.O., Borneman, J., 2003.

Identification of fungal rDNA associated with soil suppressiveness

against Heterodera schachtii using oligonucleotide fingerprinting.

Phytopathology 93, 1006–1013.

Zhang, W., Han, D.Y., Dick, W.A., Davis, K.R., Hoitink, H.A.J., 1998.

Compost and compost water extract-induced systemic acquired

resistance in cucumber and Arabidopsis. Phytopathology 88, 450–455.