lable at ScienceDirect

Soil Biology & Biochemistry 68 (2014) 44e51

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lbio

Effects of sample storage on microbial Fe-reduction in tropicalrainforest soils

Brian R. Ginn a, Mussie Y. Habteselassie b, Christof Meile c, Aaron Thompson a, *

a Department of Crop and Soil Sciences, University of Georgia, Athens, GA 30602, USAb Department of Crop and Soil Sciences, University of Georgia, Griffin, GA 30223, USAc Department of Marine Sciences, University of Georgia, Athens, GA 30602, USA

a r t i c l e i n f o

Article history:Received 12 December 2012Received in revised form4 September 2013Accepted 7 September 2013Available online 20 September 2013

Keywords:Iron reductionStorage protocolStorage timeLuquillo Critical Zone ObservatorySoil homogenizationRISA

* Corresponding author. Tel.: þ1 706 410 1293.E-mail address: AaronT@uga.edu (A. Thompson).

0038-0717/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.soilbio.2013.09.012

a b s t r a c t

Many biogeochemical investigations of soil processes require temporary and/or long-term samplestorage, which may alter soil biogeochemistry and impact the outcome of laboratory experiments. Thisstudy examines the influence of storage conditions on soil iron (Fe) reduction. We subjected soil samplesfrom the Luquillo Experimental Forest, Puerto Rico to: (a) 2-mm sieving/homogenization within ananoxic environment, and storage at 4 �C or (b) 22 �C; (c) air-drying at 30 �C followed by 2-mm sieving/homogenization and storage at 22 �C; and (d) storage at 22 �C with no sieving/homogenization (control).We assessed changes in Fe-reduction after one week, six months and 12 months of storage by incubatingsoils in a 95% N2:5% H2 headspace for 7e15 days. After one week of storage, FeII production was linearwith time yielding similar Fe-reduction rates (ave. 2.8 mmol kg�1 d�1) in all treatments except theunhomogenized controls, which exhibited only slightly lower rates (2.2 mmol kg�1 d�1). Storage for aduration of 12 months at 4 �C decreased FeII production following 7 days of anoxic incubation, while air-drying soils maintained similar FeII production levels. Most probable number counts of Fe-reducingmicrobes were comparable in the air-dried and 4 �C stored soils, but bacterial isolates from the air-dried soils exhibited higher net FeII production over a one week incubation in pure culture mediathan strains isolated from soils stored at 4 �C. When microbial activity was not limited by low tem-perature (e.g., during the 22 �C storage treatment), homogenization slowed FeII production. The observedFeII production and cell counts in these tropical soils suggest that storage at cold temperatures, as well ashomogenization when microbial activity is not suppressed during storage, reduced the resilience of theresident Fe reducing microbial communities, a finding consistent with studies on other microbial pro-cesses in tropical soils. Overall our results suggest air-drying tropical soils maintains their capacity forFeIII reduction during long-term sample storage more effectively than cold-storage.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Investigations of soil biogeochemistry often require analyticaland experimental procedures that cannot be performed in situ inthe field. For these ex situ analyses and experiments it is preferableto use freshly collected soil. But very commonly there is some delaybetween sample collection and experiment, or analysis, which re-quires the soils be stored for some period of time.

Cold storage at 4e5 �C is usually recommended for microbialexperiments (Brohon et al., 2001; González et al., 2009) becausethis reduces biomass growth (Anderson, 1987; Bloem et al., 2006)and substrate depletion (Coxson and Parkinson, 1987) relative to

All rights reserved.

storage at higher temperatures. However, cold storage has beenshown to permanently decrease soil nitrate reduction potential andmicrobial enzyme activity (Verchot, 1999; Turner and Romero,2010). Conversely, samples are often dried to extend storagetimes, but this too can alter the microbial and chemical composi-tion of the soil (Bartlett and James, 1980; Sheppard and Addison,2007). Air-dried soils have been shown to exhibit reducedenzyme activity (Turner and Romero, 2010) and altered rates of CrIII

oxidation (Bartlett and James, 1980), although the ability to reduceand oxidize sulfur can be retained for many years (Bollen, 1977).One approach to minimize changes in soil metabolism duringstorage, supported by a literature review by Bloem et al. (2006), isto impose the conditions that slow microbial populations in thenative environment. Specifically, Bloem et al. (2006) found thatstorage at the in situ soil temperature best preserved microbialactivities. Following USDA temperature classifications (Soil Survey

Table 1Storage procedure.

Temperature Sieved/homogenized

Moist 4 �C 4 �C <2 mmMoist 22 �C 22 �C <2 mmAir-dried 22 �C <2 mmNon-slurried control 22 �C None

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e51 45

Staff, 2010), the optimal storage method may thus be freezing forfrigid soils; 4 �C storage for mesic soils; and air-drying for thermicsoils.

Iron cycling is an important mediator of carbon oxidation inmany soils (e.g. Dubinsky et al., 2010), and can control the avail-ability of limiting nutrients such as phosphorous (Parfitt, 1989;Guzman et al., 1994; Liptzin and Silver, 2009; Chacon et al.,2006). It has been studied extensively under laboratory condi-tions following a range of storage treatments including air-drying(e.g. Frenzal et al., 1999; Peretyazhko and Sposito, 2005; Valencia-Cantero et al., 2007; Hansen et al., 2011) or after cold-storage (e.g.Kappler and Brune, 2002; Bennett and Dudas, 2003; Komlos andJaffé, 2004; Wildung et al., 2004; Weiss et al., 2004; Valencia-Cantero et al., 2007). Qu et al. (2005) and He and Qu (2008) sug-gest air-drying delays the onset of FeIII reduction but otherwiseminimally affects FeIII reduction rates. However, the impact ofdifferent storage conditions on soil FeIII reduction has not beencomparatively assessed. This is surprising given the large body ofwork on the effects of storage conditions on metal partitioning andspeciation in soils, including iron (e.g., Bartlett and James, 1980;Thomson et al., 1980; Rapin et al., 1986; Pezzolesi et al., 2000).

Additionally, the common practice of mixing the soil to mini-mize the impact of spatial variability, often termed homogeniza-tion, can also impact the experimental results (e.g., Sheppard andAddison, 2007; Quantin et al., 2002; Thompson et al., 2006;Dubinsky et al., 2010). Homogenization is distinct from making asoil slurry by suspending the soil in awater solution, althoughmostslurried soils have been homogenized previously. Homogenizationcauses a loss of soil structure that can alter bacterial activity (Tehand Silver, 2006; Teh et al., 2008). As a consequence, some re-searchers prefer not to alter their samples prior to experimentation,but to perform the experiment directly on the moist field sample(e.g., Liptzin and Silver, 2009).

In this paper, we examine the influence of storage conditionsand sample handling on soil iron reduction and microbial compo-sition across one year of storage. We aimed to answer two ques-tions: (1) How do different storage protocol preparations influencethe soil’s Fe reduction capacity? (2) Which storage condition(s)maintain Fe reduction capacity over 12 months of storage? We dothis by applying four common storage protocols (homogenized soilair-dried or storedmoist at 4 �C or 22 �C and a no-treatment controlstored at 22 �C) to soil from the Luquillo Experimental Forest inPuerto Rico. To answer our first question, we subjected the soils toincubation under anoxic conditions and measured the productionof 0.5 M HCl extractable FeII after one week of storage. To answerour second question we repeated the anoxic incubation after sixand 12 months storage.

2. Methods

2.1. Field site

Soils were collected from a single location in Puerto Rico at theBisley Research Watersheds, Luquillo Experimental Forest (LEF),which is part of the NSF-funded Luquillo Long-term EcologicalResearch (LTER) site and the NSF-funded Critical Zone Observatory(CZO). Soils in the LEF are classified predominately as Ultisolsformed fromvolcanic parent material with quartz diorite intrusions(Beinroth, 1982; Frizano et al., 2002). The major primary mineralsconsist of quartz and plagioclase feldspars with lesser amounts ofbiotite, hornblende, K-feldspar, and accessory magnetite, sphene,apatite, and zircon (White et al., 1998). Prior work by Silver et al.(1999) has identified periodic fluctuations in pO2 sufficient togenerate anoxic conditions within the soil. Peretyazhko and Sposito(2005) identified these soils as Fe-rich, mildly acidic (pH w5), and

nutrient poor. We collected the upper 10 cm of soil from the Bisleywatershed valley as an intact soil block at a similar site to thatdescribed by Peretyazhko and Sposito (2005). Small portions of adiscrete overlying organic horizon were included in the soil block,but their contribution was minimal relative to the total mass of soilcollected. Total Fe and P concentrations in the soilweremeasured byICP-MS following Li-metaborate fusion (Hossner, 1996) by AcmeLabs, Vancouver, BCCanada. Total Cwasmeasuredbygrindingoven-dried samples to less than 250 mm in a ball-mill and combustingthem in 5�5mmtin capsules on a Carlo Erba Elemental Analyzer. Inaddition, we performed a citrate/ascorbate extraction to assess theabundance of short-range-ordered Fe (Reyes and Torrent,1997). Thesamples were suspended in a 0.2 M Na-citrate/0.05 M ascorbic acidsolution (pH ¼ 6) at 60:1 reagent:sample ratio and shaken for 16 hon a horizontal shaker, and then centrifuged at 11,000 rcf for 30min.Then the supernatants were analyzed with via atomic absorption(AA) spectroscopy (Perkin Elmer AAnalyst 200).

2.2. Storage protocols

Samples were placed in 4-mil polyethylene ziplock bags andtransported without cooling inside an insulated container to theUniversity of Georgia. Within 24 h of sampling all moist soils wereprocessed into four storage treatments (Table 1) by (a) 2-mmsieving/homogenization (to remove large roots and any gravel-sized rocks) under field soil moisture conditions within a 95%:5%(N2:H2) glovebox followed by storage at 4 �C or (b) 22 �C; (c) air-drying at 30 �C for 24 h followed by 2-mm dry-sieving/homogenization and storage at 22 �C; and (d) storage at 22 �Cwith no sieving/homogenization. For the rest of this article, thesetreatments will be referred to as moist 4 �C, moist 22 �C, air-dried,and controls, respectively. All soils were stored under oxic condi-tions in the dark.

2.3. Iron reduction experiments

We performed iron reduction experiments on each storagetreatment after one week, six months, and 12 months. All solutionsand soil samples were stored and prepared under oxic conditions.For the moist 4 �C, moist 22 �C, and air-dried treatments, 4.5 g (dry-weight) of soil was suspended in a 2 mM KCl solution with a 45 gfinal mass and supplemented with 3 mmol L�1 of Na-lactate toeliminate any effects of storage on carbon substrate availability.Triplicate slurries were incubated for 7 d or 14 d in a Coy anaerobicchamber (95% N2, 5% H2) on an orbital shaker (Eberbach 6130, AnnArbor, MI) at 250 rpm. We prepared two different control treat-ments for our experiments. In the first set of controls, 4.5 g of (dry-weight equivalent) soil was placed into 33 separate centrifugetubes each at oneweek, six months and 12months, sufficient for 11triplicate time points during each sampling period. An equivalentaliquot of Na-lactate (0.67 mmol g�1 soil) was added to eliminateany effects of storage on initial carbon substrate availability, and thetubes were sealed in the anoxic chamber. However, we foundsubstantial drying of the control samples occurred after six monthsand 12 months of storage under oxic conditionsddespite being

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e5146

sealed in ziploc bags (4-mil polyethylene)dsuch that samples notsuspended in slurries produced negligible iron reduction (<5% in-crease in Fe2þ; data not shown). Thus, we prepared a second controltreatment for the six-month and 12-month experiments in whichthe controls were suspended in slurries (triplicates with 11 timepoints each) as described above for the other storage treatments.

Samples were extracted at 12 h to 3 d intervals using wide-orifice pipette tips to prevent size exclusion of the soil particles.Acid-extractable FeII was solubilized by suspending 0.5 mL of eachsoil slurry in 1.0 mL of 0.75 M HCl or suspending 0.05 g (dry weightequivalent) of the non-slurried samples in 1.5 mL of 0.5 M HCl andshaking for 2 h on a horizontal shaker. Extracts were then centri-fuged at 11,000 rcf (relative centrifugal force) for 10 min, afterwhich the supernatant was collected and analyzed for FeII using amodified ferrozine protocol (Thompson et al., 2006) on aShimadzu-1700 spectrophotometer at l ¼ 562 nm. The dry weightof each pellet from the centrifugation step was determined tocorrect for minor differences in suspension density betweentreatments and allow accurate dry-weight normalization. All valuesare reported as averages of triplicates.

2.4. Most probable number (MPN)

To assess differences in microbial population across our treat-ments, the abundance of Fe-reducing microorganisms wasmeasured following one year of storage using a triplicate series ofMPN tubes (Woomer, 1994). The MPN tubes contained thefollowing growthmedia (per liter): 0.5 g KH2PO4,1.0 g Na2SO4, 2.0 gNH4Cl, 1.0 g yeast extract, 0.5 mM CaCl2, 0.1 mMMgSO4, 10 mMNa-lactate, and 50 mM Fe-citrate. A gram of soil (dry-weight) wassuspended in 10 mL of growth media, diluted 10 times sequentiallyto create a logarithmic dilution series. The MPN tubes were incu-bated in the same anaerobic atmosphere described above (95% N2,5% H2). Samples were scored positive for growth of Fe-reducers ifthe initial reddish-brown color of growth media turned clearconsistent with the reduction of the FeIII-citrate complex withintwo months.

2.5. Identification of the moist 4 �C and air-dried isolates

Strains from soils stored air-dried and moist 4 �C for 12 monthswere isolated from the last MPN tube exhibiting growth andsampled at the termination of the Fe reduction experiment. Thesetreatments were chosen because they are the most commonly usedstoragemethods and they exhibited the greatest differences in FeIII-reduction potential. Even though the MPN tubes were incubatedunder anaerobic anoxic conditions, we observed that the isolatedstrains were capable of both aerobic and anaerobic growth.Consequently, the strains were spread onto a 1.5% tryptic soy broth(TSB) agar plate, and grown under oxygen. Single colonies werepicked from these plates and grown on separate plates. Each iso-lated strain was placed in long-term storage by freezing them in a12.5% glycerol solution at �80 �C.

Genomic DNA from the isolates was extracted using the DNAzolGenomic DNA Isolation Kit as per the manufacturer’s instructions(Molecular Research Center Inc., Cincinnati, Ohio). The DNA ex-tracts were subjected to polymerase chain reactions (PCR) using theprimer pair 968f (50 eAAC GCG AAG AAC CTT AC-30) and 1401r (50-CGG TGT GTA CAA GAC CC-30) (Teixeira et al., 2010) to amplify asection of 16S rDNA. The PCR consisted of 1 mL template volume,12.5 mL 2X GoTaq Hot Start Colorless Master Mix (Promega, Madi-son, WI), 2 mL of each primer pair (0.4 mM final conc.) and 7.5 mLnuclease free water. The PCR conditions were as follows: 2 min ofinitial denaturation at 94 �C followed by 35 cycles of denaturationat 94 �C for 1 min, annealing at 58.5 �C for 1 min and extension

72 �C for 2 min. The PCR was concluded with a final extension stepat 72 �C for 10 min. The amplification products were run on agarosegel (1.5%) to confirm the expected size of the amplicons (432 bp).

The PCR amplicons were ligated into the pGEM-T Easy Vectorplasmids, followed by transformation with competent E. coli cellsaccording to the manufacturer’s instructions (Promega, Madison,WI). The transformation products were plated on Lysogeny Brothagar containing kanamycin (50 mg/L). After a 24 h incubationperiod, five colonies were randomly picked for each isolate andgrown for 12 h in LB broth containing kanamycin. The plasmidswere purified from the overnight cultures using the PureYieldPlasmid Miniprep System from Promega (Madison, WI). The pres-ence of inserts in the plasmids was confirmed using restrictiondigestion reaction with EcoRI before they were sequenced withboth M13R and M13F primers at UGA’s Georgia Genomics Facility.Sequencing was performed with an ABI PRISM 3730xl DNAanalyzer (Applied Biosystems, Grand Island, NY).

The nucleotide sequences were cleaned of vector contaminationand investigated for sequence identity and similarity using theNational Center for Biotechnology Information’s (NCBI) VecSreenand Basic Local Alignment Search Tool (BLAST) tools, respectively(Altschul et al., 1997).

2.6. Ribosomal Intergenic Spacer analysis (RISA)

We performed RISA community analysis on DNA extracted fromarchived (frozen and stored at �20 �C) samples of the six-monthstored air-dried, moist 22 �C and moist 4 �C soils collected at thetermination of the Fe reduction experiment and on genomic DNAsamples taken from the isolates described in Section 2.5. We alsoperformed RISA community analysis (Ranjard et al., 2000) on DNAextracted from air-dried and moist 4 �C soils after 18 months ofstorage. The PCR for RISA was performed with S-D-Bact-1522-b-S-20 (50- TGCGGCTGGATCCCCTCCTT-30) and L-D-Bact-132-a-A-18 (50-CCGGGTTTCCCCATTCGG -30) primer pair. We extracted genomicDNA from soil samples using the PowerSoil DNA Isolation Kit (MOBIO Laboratories, Inc. Carlsbad, CA) as described by the manufac-turer. The PCR was performed as above, but PCR conditions were3 min initial denaturation at 94 �C followed by 25 cycles of dena-turation at 94 �C for 1min, annealing at 55 �C for 30 s and extension72 �C for 1 min. The PCR was concluded with a final extension stepat 72 �C for 5 min and a hold at 4 �C.

The PCR amplicons were run on 5% (w/v) acrylamide/bis-acrylamide (37.5:1; 40%) gel of non-denaturing gradient (1 mm,16 cm long) using the Dcode Universal Mutation Detection System(BioRad Laboratories, Hercules, CA) at constant temperature (20 �C)for 12 h at 60V. Gels were stained with SYBR Green I Nucleic AcidGel Stain (Lonza, Rockland, ME) according to the manufacturer’sinstructions. The banding patterns were photographed using theKodak Gel DOC 100 system (Kodak Co. Rochester, NY) under a302 nm UV source.

2.7. Pure-culture experiments

To examine differences in the FeIII-reducing capacity of theisolated strains, theywere grown on lactate with solid-phase FeIII asthe only available electron acceptor. Strains were rejuvenated byincubating frozen stocks for 12 h (overnight tubes) at 22 �C in theMPN growth media, with oxygen as an electron acceptor instead ofFe-citrate. These overnight tubes were centrifuged at 21,000 rcf for30 min and washed three times in fresh growth media without anyNa-lactate or yeast extract. Aliquots of the washed cells were sus-pended in fresh growth media to give all samples the same initialcell concentration (between 105 and 106 cells mL�1) and 50 mM ofhydrous ferric oxide (HFO) was added as an electron acceptor. The

0 1 2 3 4 5 6 7

-5

0

5

10

15

20

25

FeII (m

mol

kg-1

)

Days

FeII

Rate (mmol kg-1

d-1

)

Air-dry 2.7 (±0.2) 4C 2.7 (±0.1) 22C 2.9 (±0.1) HET 2.2 (±0.2)

Fig. 1. Extractable FeII (0.5 M HCl) in 1:10 soil slurry reactors (air-dry, moist 4 �C andmoist 22 �C) or as unaltered soil cores (non-slurried control) following storage for oneweek under the following conditions: Air-dried at 30 �C for 24 h, homogenized andstored moist at 22 �C (squares); homogenized and stored moist at 4 �C (circles); ho-mogenized and stored moist at 22 �C (triangles); and controls stored as collected at22 �C and incubated as unaltered soil cores (inverse triangles). 0.67 mmol g�1 of lactatewas added to each treatment. Error bars are standard deviations of three replicates.Errors on the rate of FeII production given in the legend are 2 sigma for a linearregression of slope of the line.

Table 2Comparison of least square means of HCl extractable FeII concentrations(mmol kg�1 soil) between storage regimes.

Storage regime Marginal mean

1 week 6 months 12 months

Moist 22 �C 6.3a (0.29) 15.3a (2.2) 10.5c (1.3)Air-dried 6.4a (0.29) 13.5a (2.2) 26.6a (1.3)Moist 4 �C 5.6ba (0.3) 2.8b (2.2) 4.9d (1.3)Slurried Control NA 11.0a (2.2) 18.5b (1.3)Non-slurried Control 4.9b (0.29) NA NA

Means with the same letters were not significantly different from each other(a ¼ 0.05).Comparisons are valid only within the same column.Standard errors are in parentheses.

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e51 47

HFO was prepared by precipitating Fe(NO3)3 with NaOH accordingto the protocol given in Kostka and Nealson (1998). The reactiontubes were shaken on an orbital shaker at 250 rpm in an anoxicatmosphere (95% N2: 5% H2). Samples were analyzed for 0.5 M HClextractable iron using a protocol given in Kostka and Nealson(1998). Briefly, 200 mL of sample was transferred to 1 mL of 0.5 MHCl and shaken for 1 h. The samples were then vortexed and 100 mLof the supernatant was mixed with 1.9 mL of ferrozine reagent(0.02% in 50 mM HEPES buffer). The samples were centrifuged at21,000 rcf for 2 min and analyzed for FeII.

2.8. Statistical analysis

To answer our first question, regarding the influence of storageprotocol preparations, we analyzed iron reduction rate for eachtreatment after one week of storage calculated as the slope of ironconcentration versus time. The rates were then subjected to a one-way ANOVA to test the statistical significance of the effect of thetreatments on the rates. After six and 12 months of storage theproduction of FeII was non-linear, so we used a repeated measuresanalysis of variance (RM-ANOVA) in SAS 9.3 (SAS Institute, Inc.,Cary, NC) using SAS Proc Glimmix to test the statistical significanceof storage regime on iron reduction with storage regime as a fixedeffect and anoxic incubation time as repeated effect factors. Thissame test was used to test differences in FeII production from ourbacteria isolates.

Testing differences in treatment across storage time (question#2) was complicated because our sampling time points were notprecisely the same, sowe first constructed a numerical model of theFeII production vs. time during the anoxic incubations assuming:

d½Fe�dt

¼ kP (1)

where [Fe] is the 0.5 M HCl extractable FeII concentration, t is time,P is the bacterial population size, and k is the cell-specific Fereduction rate. We then calculated FeII production over 153 h(denoted as FeII�h153, where 153 h is the length of the anoxic in-cubation following the one week storage event) by integratingunder the best-fit curves of each replicate anoxic incubationtreatment following oneweek, six months, or 12months of storage.These values were then subjected to a one-way ANOVA to test forthe statistical significance between storage timepoints (seeAppendix B for full details). All pair-wise mean comparisons werecarried out using Fisher’s least squares mean comparison. All testswere done at an a-value of 0.05.

3. Results

The soil used for the incubation experiments had a total ironcontent of 62.7 � 0.2 g kg�1, of which 10.2 � 0.6 g kg�1 wasextractable by citrate-ascorbate; the total P content was276 � 67 mg kg�1 soil, typical for iron-rich, nutrient-poor Ultisols(Peretyazhko and Sposito, 2005); and the total carbon content was25.5 � 0.1 g kg�1 soil (error is the standard deviation of 3 replicatesfor all values).

3.1. Comparison of iron reduction within storage time

After oneweek of storage net FeII production rates were linear at2.2 � 0.2 mmol kg�1 d�1 for the non-slurried control soils and at2.7 � 0.2, 2.7 � 0.1, and 2.9 � 0.1 mmol kg�1 d�1 for the homoge-nized soils stored air-dried, moist 4 �C, and moist 22 �C, respec-tively (Fig. 1). The three homogenized soil treatments exhibitedsimilar reduction rates after one week of storage (the means were

not statistically different at the 0.05 level using a one-way ANOVA).However, the non-slurried control soils, which were incubated asintact cores, exhibited significantly lower FeIII reduction rates thanair-dried and moist 22 �C stored samples incubated as soil slurries.The marginal mean test (Table 2) also supports these conclusions.

After six months of storage, incubations of the moist 22 �C, air-dried and slurried control soils exhibited similar, non-linear pro-duction of FeII as a function of incubation time (Fig. 2 and Table 2),while the soils stored moist at 4 �C exhibited FeII concentrationsthat were more than three times lower than the other treatmentsafter 7 days of incubation (Fig. 2c). The marginal mean FeII con-centrations for the moist 4 �C treatment were significantly lowerthan the other treatments, which were all similar to each other(Table 2).

After one year of storage the marginal mean FeII concentrationswere significantly different in all treatments (Table 2). The anoxicincubation was extended to 15 d for this storage time and thisresulted in HCl-extractable FeII concentration plateaus or nearplateaus for all treatments. The air-dried soils were the only

0 2 4 6 8 10 12 14 16

0

10

20

30

40

50

FeII (m

mol

kg-1

)

Days

storage

One week Six months One year

0

10

20

30

40

50

FeII (m

mol

kg-1

)

Air-dried storage

One week Six months One year

0 2 4

storage

One week Six months One year

0 2 4 6 8 10 12 14 16

Days

Control Storage

One week Six months One year

A

B

C

D

Fig. 2. Extractable FeII (0.5 M HCl) in 1:10 soil slurry reactor following storage for one week, six months and 12 months under the following conditions: (a) air-dried at 30 �C for24 h, homogenized and stored at 22 �C; (b) homogenized and stored moist at 4 �C; (c) homogenized and stored moist at 22 �C; (d) stored as collected at 22 �C (slurried control).0.67 mmol g�1 of lactate was added to each soil. Only the 12 month treatment was monitored past 7 days. Error bars are standard deviations of three replicates.

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e5148

treatment to achieve a plateau within one week, reaching approx.35 mmol kg�1 after 7 d of incubation. In contrast, the slurriedcontrol soils reached a plateau of similar concentration after 13 d(Fig. 2a,d), while the moist 4 �C and moist 22 �C stored soilsappeared to plateau at much lower FeII concentrations of ca.15 mmol kg�1 and 5 mmol kg�1, respectively (Fig. 2b,c).

3.2. Comparison of iron reduction across storage times

Fe reduction in each treatment was compared across storagetimes using a one-way ANOVA on time-integrated FeII concentra-tion (FeII�h153) as described in Appendix B (Table 3). FeII�h153 valuesfor the air-dried samples were similar following one week, sixmonths and 12 months of storage. In all other treatments, FeII�h153values exhibited significant changes with storage time. Forinstance, relative to the FeII�h153 value after one week of storage,the moist 4 �C treatment exhibited a decrease in its FeII�h153 valueafter 6 months of storage, which remained similar after 12 months.

Table 3Comparison of the ferrous�hours (FeII�h153 with units of mmol kg-1 hr) under thebest-fit curves for one-week, six-month, and 12-month storage times.

Treatment 1 Week 6 Month 12 Month

Air-dried 1320a (233) 1279a (356) 2003a (435)Moist 4 �C 1366a (37) 279b (37) 332b (18)Moist 22 �C 1512a (38) 1294ab (531) 486b (76)Slurried Control N/A 881a (7) 401b (108)Non-slurried Control 1146a (12) 69b (12) 79b (2)

Values with the same letters were not significantly different from each other(a ¼ 0.05).Comparisons are valid only within the same row.Standard deviations are in parentheses.

In contrast, the moist 22 �C treatment did not exhibit a significantdecrease in its FeII�h153 value (relative to the 1 week storage time)until after 12 months of storage. The slurried controls exhibitedsignificantly different FeII�h153 values between six and 12 monthsof storage.

3.3. Microbial community analysis

We assessed the microbial community in the moist 4 �C and air-dried soils using RISA profiles of extracted soil DNA and measuredthe numbers of Fe-reducers using Most Probable Number tech-niques.We also isolated the strains of Fe-reducing bacteria from themost diluted MPN tube exhibiting growth, identified them using16S rDNA sequences and then quantified FeIII reduction for thesestrains in a common liquid media. Finally, we compared RISAprofiles between the strains, the non-incubated soils and archivedsamples of the moist 4 �C and air-dried soils collected at thetermination of the 6-month Fe reduction experiment.

After one year of storage, all soils had similar numbers of Fe-reducing bacteria based on Most Probable Number (MPN) counts(moist 4 �C: 6.0 � 1; moist 22 �C, air-dried and control: all 5.4 � 1).Bacteria strains isolated from the moist 4 �C (GenBank accession#KF008238) and air-dried (GenBank accession #KF008237) MPNtubes had 16S rDNA nucleotide sequences with 99e100% similarityto Pseudomonas moraviensis strain CCM 7280 and 97e98% simi-larity to Bacillus thuringiensis strain IAM 12077, respectively. Whilethe RISA profile of the isolate from the air-dried soil containsmultiple bands (Fig. 3, lane AD18MO), our sequence data confirmedthe isolate was a single bacterial strain, consistent with findings byothers using RISA techniques (e.g. Nocker et al., 2007). Both isolatesreduced FeIII from hydrous ferric oxide (HFO) in liquid growth

Fig. 3. RISA profiles for pure moist 4 �C (4C) and air-dried (AD) MPN isolates. 4C18MOand AD18MO are the non-incubated moist 4 �C and air-dried soils stored for 18months. AD6MO, 22C6MO and 4C6MO are RISA profiles of air-dried, moist 22 �C, andmoist 4 �C soils, respectively, after the anoxic Fe reduction experiment shown in Fig. 2(after six months of storage). All anoxic incubations were performed at room tem-perature. The L and E lanes are a ladder and empty control, respectively. The arrowsindicate bands of the same size across the lanes.

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e51 49

media and rapidly achieved a FeII concentration peak within oneday of the reduction experiment (Fig. 4). The FeII concentration forboth isolates declined over time, perhaps reflecting the formationof stable iron minerals not extractable by 0.5 M HCl (Zachara et al.,2002; Hansel et al., 2003). However, the strain from the moist 4 �Cstored soil (Pseudomonas sp.) reached stationary phase earlier andexhibited a significantly lower FeII concentration plateau than thestrain from the air-dried soil (Bacillus sp.) (Fig. 4). This is reflected inthe significantly lower (a ¼ 0.05) marginal least square mean[mean � standard error] of the FeII concentration for the moist 4 �Cisolate (136 � 5 mg L�1) than for the air-dried isolate(317 � 5 mg L�1).

We extracted DNA from our soils after 6 months of storagefollowing the Fe-reduction experiment (Fig. 2) and from the non-incubated soils after 18 months of storage. We then compared theextracted DNA from our soils with the DNA extracted from ourMPNisolates using RISA fingerprinting. The isolated strains did not

0 2 4 6 8

0

200

400

600

FeII (m

mol

L-1)

Days

Air-dried

Fig. 4. Extractable FeII (0.5 M HCl) from incubations of pure culture strains from soilair-dried at 30 �C for 24 h (solid squares) or soil homogenized and stored moist at 4 �C(open circles). Error bars are standard deviations of three replicates.

comprise a substantial portion of the extracted DNA from thestored soils prior to incubation under anoxic conditions (Fig. 3;lanes 4C18MO and AD18MO). We detected a weak band corre-sponding to the 4 �C isolate (Pseudomonas sp.) in the moist 4 �C soil(Fig. 3, lane 4C18MO), but the fingerprint bands for the air-driedisolate (Bacillus sp.) were below detection (Fig. 3, lane AD18MO).However, following the anoxic Fe reduction experiment (Fig. 2) ourisolates represented the brightest bands in the respective RISAprofiles (Fig. 3; lanes AD6MO and 4C6MO). This dominance of theisolates in the six-month whole soil RISA profiles is consistent withour isolation of the strains from the most dilute FeIII reducing MPNtube, which should select for themost abundantmicrobe capable ofFe reduction.

4. Discussion

4.1. Short-term storage

Our observations suggest experiments performed within oneweek of soil collection will yield similar iron reduction ratesregardless of storage protocol. The slightly lower Fe reduction ratein the non-slurried control soils likely reflects a lower mass of soilexposed to optimal FeIII-reducing conditions compared to theslurried treatments. Suspending the homogenized soils in slurriescould affect microbial metabolism and increase their capacity toreduce Fe by bringing them together with labile-carbon and easily-reducible Fe. Substrate/electron acceptor availability and accessi-bility, as well as microbial composition and population, may limitFeIII reduction across all treatments. We maintained consistentinitial carbon availability in the one-week, six-month and 12-month experiments through additions of lactate to all treatments,and also observed similar FeII production in our unamended con-trols (data not shown), so it is unlikely that the systems were car-bon limited. We did not make measurements of microbialpopulations at the one week mark, however all treatmentsexhibited linear FeII production curves after one week of storage,with no discernable lag or exponential increase in rates that wouldbe expected during microbial growth (Fig. 1). Thus, the microbialcommunity appears equally capable of FeIII reduction across thetreatments after one week of storage. Therefore, we surmise that,after storing our soils for one week, FeIII reduction is most likelyconstrained by the quantity of immediately reducible Fe.

4.2. Air-dried vs. refrigerated

Our findings are consistent with previous studies suggestingcold storage eventually reduces the activity of microbes in tropicalsoils (Verchot, 1999; Turner and Romero, 2010). This effect isevident in our experiments after six months of storage with lowerFeII production (i.e., FeII�h153, see Table 3 and Appendix B) andlower FeII concentrations in themoist 4 �C soils relative to oneweektreatments and soils stored at 22 �C (Fig. 2b,c; Table 3). Turner andRomero (2010) also observed a delay in the effects of cold storageon the hydrolytic enzyme activity of tropical soils, with enzymeactivities similar between soils stored at 4 �C and room tempera-ture during the first two weeks of storage followed by a decrease inthe 4 �C treatment after two weeks. We did not detect lowernumbers of Fe-reducing organisms in ourmoist 4 �C soils relative tothe other treatments (Blodgett, 2005), but the Fe-reducing bacterialstrain we isolated from the moist 4 �C samples exhibited a lowerplateau in FeII concentration than the strain isolated from the air-dried samples (Fig. 4). This suggests cold storage selects for Fe-reducing bacteria with either a lower stationary phase or a lowercell specific Fe-reduction rate than air-dried storage. Soil samplesstored at 4 �C for six months also had substantially different RISA

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e5150

profiles than soils with different soil treatments. Together, theseresults may indicate that growth of psychrotrophic bacteria duringstorage may have shifted the population in the moist 4 �C soilstoward Fe-reducing organisms with lower temperature growthoptimums or lower overall growth rates.

In contrast, even after one year of storage, the air-dried samplesproduced greater quantities of FeII than all other treatments,including the controls (Fig. 2a; Table 2). Desiccation impacts mi-crobial communities as a whole (Potts, 1994), and many soil bac-teria are capable of withstanding desiccation stress, e.g. throughsporulation. This is especially true for gram-positive organismssuch as Bacillus and Clostridium (Potts, 1994), which are common insoils with variable water contents. Clostridium species are wellknown to facilitate iron reduction during their fermentation ofglucose (Lovley, 1995; Dobbin et al., 1999), and the most abundantFe-reducer in our air-dried soils after a 14-day anoxic incubationwas a Bacillus sp. (Fig. 4) that could grow under both oxic andanoxic conditions. Drying the soils at 30 �C could have selected forthe Bacillus sp., since they are spore-forming bacteria and also growoptimally at warmer temperatures (Logan and De Vos, 2009). Pre-vious work shows that when air-dried soils are rehydrated inanoxic solutions, they can recover their original (pre-dried) Fereduction capacity after an initial lag phase (Qu et al., 2005; He andQu, 2008). Thus, while air-drying tropical soils has a significant andimmediate impact on overall enzyme activity (Turner and Romero,2010) and microbial communities, our work suggests it may notdecrease overall Fe reduction capacity.

4.3. The impact of homogenization

For the first six months of storage, the production of FeII in thesoil slurries was similar for both homogenized and unhomogenizedsoils stored at 22 �C (Table 2). However, after 12 months of storage,slurries of the pre-homogenized soil exhibited much lower pro-duction of FeII than slurries of the unhomogenized soil (Fig. 2c andd;Table 2). Both treatments were stored near the native field tem-peratures under oxic conditions, so we assume microbial activitycontinued throughout storage. While our data set does not allow usto conclusively identify the mechanism underlying this decrease inFeII production in our moist 22 �C soils between six and 12 months(Table 3), we hypothesize that it is a result of sieving/homogeniza-tion. We suspect this decrease in FeII production after storing soilspre-homogenized for 12 months reflects the destruction of low-redox microsites that likely protected Fe-reducing organisms fromcompetitionwith other organisms. Both Teh et al. (2008) and Liptzinand Silver (2009) observed redox potential variation at differentmicrosites within soils. Indeed, they found that Fe reduction occursevenwhen soils are incubated under aerobic conditions, suggestingthat storing soils in the presence of oxygen may preserve anaerobicpopulations as long as soil structure is maintained.

4.4. Preservation of FeIII-reducing capacity

After one year of storage, all treatments developed a lag in theinitiation of FeII production during the anoxic incubations relative tothe trend in FeII production after one week of storage (Fig. 2;Appendix B). This lag was w3 d for the air-dried and w7 d for thecontrol soils, after which FeII concentrations were similar in thesetwo treatments for soils stored for 12months and 1week. However,FeII concentrations during FeIII reduction for the moist 22 �C soilsstored for 12 months never returned to the FeII concentrationsobserved after one week of storage (Fig. 2b; Table 3). Furthermore,by sixmonths of storage, the air-dried andmoist 4 �C soils showed asignificantly different banding pattern in our RISA analysis, indi-cating that different bacterial species are responsible for Fe

reduction in soils exposed to these different soil treatments.Collectively, these results suggest the capacity to reduce FeIII is bestpreserved in our soils by techniques that prevent competition be-tweenmicrobial groups, either by uniformly reducing themicrobialpopulation and preventing growth during storage in the case of air-drying, or by not homogenizing the soil and thereby facilitating awide distribution of micro-niches. Treatments that accentuated theinteraction of microbes through homogenization and allowedcontinued microbial growth under oxic conditions (moist 4 �C andmoist 22 �C treatments) exhibited the lowest FeII�h153 values after12months of storage (Table 2). Presumably, oneweek of storagewasinsufficient to substantially alter the microbial composition of thesoils, so we suspect that the pool of immediately reducible FeIII

constrained reduction rates for all treatments. However, we suggestthat after six months of storage, microbial composition changedenough to generate significant difference in FeII concentrations be-tween the moist 4 �C and other storage treatments (Fig. 2 andTable 2).

5. Conclusion

Soil storage methods differ considerably with regard to tem-perature, humidity, and homogenization and thus can be expectedto yield different FeIII reduction rates. Sheppard and Addison (2007)highlight the importance of coupling storage methods with thephysico-chemical conditions of the native soil environment. How-ever, our experiments show that during the first week, storagemethod had a negligible impact on the rates of iron reduction inslurried soils when compared to the rates of FeIII reduction in non-slurried soils. As storage time increased to six and 12 months, allbut the air-dried soils exhibited significant decreases in the amountof FeII produced during the first week (153 h) of anoxic incubation,i.e., ferrous�hours (FeII�h153). Microbial fingerprinting analysis(RISA) showed that many bacterial strains are amplified duringincubation of soil slurries and that air-dried, moist 22 �C, or moist4 �C stored soils only differed by a few strains. Homogenizing soilsprior to storage seemed to negatively impact microbial FeIII

reduction, potentially by destroying microsites within the soils,thereby forcing anaerobic species to compete with aerobic speciesin a homogenous environment. When subject to longer storage, ourdata (from a warm, wet climate) suggests FeIII reduction capacity ismaintained more effectively by air-drying tropical soils than bystoring them field-moist at 4 �C.

Acknowledgments

We thank Whendee Silver for providing access to the field siteand her assistance sampling soils and Jerry Davis at UGA-Griffin forhis help with the statistical analysis. We thank two anonymousreviewers and editor Josh Schimel for suggestions that greatlyimproved the manuscript. Funding was provided by the UnitedStates Department of Agriculture (USDA), Soil Processes ProgramAFRI Grant (2009-65107-05830) to AT and CM.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.soilbio.2013.09.012.

References

Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J.,1997. Grapped BLAST and PSI-BLAST: a new generation of protein databasesearch programs. Nucleic Acids Res. 25, 3389e3402.

B.R. Ginn et al. / Soil Biology & Biochemistry 68 (2014) 44e51 51

Anderson, J.P.E., 1987. Handling and storage of soils for pesticide experiments. In:Somerville, L., Greaves, M.P. (Eds.), Pesticide Effects on Soil Microflora. Taylorand Francis, London, pp. 45e60.

Bartlett, R., James, B., 1980. Studying dried, stored soil samples e some pitfalls. SoilSci. Soc. Am. J. 44, 721e724.

Beinroth, F.H., 1982. Some highly weathered soils of Puerto Rico: 1. Morphology,formation and classification. Geoderma 27, 1e73.

Bennett, B., Dudas, M.J., 2003. Release of arsenic and molybdenum by reductivedissolution of iron oxides in a soil with enriched levels of native arsenic.J. Environ. Eng. Sci. 2, 265e272.

Blodgett, R.J., 2005. Upper and lower bounds for a serial dilution test. J. AOACInternat. 88, 1227e1230.

Bloem, J., Schouten, A.J., Sørensen, S.J., Rutgers, M., van der Werf, A., Breure, A.M.,2006. Monitoring and evaluating soil quality. In: Bloem, J., Hopkins, D.W.,Benedetti, A. (Eds.), Microbiological Methods for Assessing Soil Quality. CABIPublishing, London, pp. 23e49.

Bollen, W.B., 1977. Sulfur oxidation and respiration in 54-year-old soil samples. SoilBiol. Biochem. 9, 405e410.

Brohon, B., Delolme, C., Gourdon, R., 2001. Complementarity of bioassays and mi-crobial activity measurements for the evaluation of hydrocarbon-contaminatedsoils quality. Soil Biol. Biochem. 33, 883e891.

Chacon, N., Silver, W.L., Dubinsky, E.A., Cusack, D.F., 2006. Iron reduction and soilphosphorous solubilization in humid tropical forest soils: the roles of labilecarbon pools and an electron shuttle compound. Biogeochemistry 78, 67e84.

Coxson, D.S., Parkinson, D., 1987. Winter respiratory activity in aspen woodlandforest floor litter and soils. Soil Biol. Biochem. 19, 49e59.

Dobbin, P.S., Carter, J.P., San Juan, C.G.-S., von Hobe, M., Powell, A.K., Richardson, D.J.,1999. Dissimilatory Fe(III) reduction by Clostridium beijerinckii isolated fromfreshwater sediment using Fe(III) maltol enrichment. FEMS Microbiol. Lett. 176,131e138.

Dubinsky, E.A., Silver, W.L., Firestone, M.K., 2010. Tropical forest soil microbialcommunities couple iron and carbon biogeochemistry. Ecology 91, 2604e2612.

Frenzal, P., Bosse, U., Janssen, P.H., 1999. Rice roots and methanogenesis in a paddysoil: ferric iron as an alternative electron acceptor in the rooted soil. Soil Biol.Biochem. 31, 421e430.

Frizano, J., Johnson, A.H., Vann, D.R., Scatena, F.N., 2002. Soil phosphorus fraction-ation during forest development on landslide scars in the Luquillo Mountains,Puerto Rico. Biotropica 34, 17e26.

González, A.M., Pineiro, A.L., Fernández, M.R., 2009. Viability of culturable soilmicroorganisms during freeze storage. In: Mendez-Vilas, A. (Ed.), CurrentResearch Topics in Applied Microbiology and Microbial Technology. WorldScientific Publishing Co., Toh Tuck Link, pp. 114e118.

Guzman, G., Alcantara, E., Barron, V., Torrent, J., 1994. Phytoavailability of phosphateadsorbed onto ferrihydrite, hematite, and goethite. Plant Soil 159, 219e225.

Hansel, C.M., Benner, S.G., Neiss, J., Dohnalkova, A., Kukkadapu, R.K., Fendorf, S.,2003. Secondary mineralization pathways induced by dissimilatory ironreduction of ferrihydrite under advective flow. Geochim. Cosmochim. Acta 67,2977e2992.

Hansen, D.J., McGuire, J.T., Mohanty, B.P., 2011. Enhanced biogeochemical cyclingand subsequent reduction of hydraulic conductivity associated with soil-layerinterfaces in the vadose zone. J. Environ. Qual. 40, 1941e1954.

He, J., Qu, D., 2008. Dissimilatory Fe(III) reduction characteristics of paddy soilextract cultures treated with glucose or fatty acids. J. Env. Sci. 20, 1103e1108.

Hossner, L.R., 1996. Dissolution for total elemental analysis. In: Sparks, D.L. (Ed.),Methods of Soil Analysis: Part 3-Chemical Methods, second ed. Soil ScienceSociety of America, Madison, pp. 49e64.

Kappler, A., Brune, A., 2002. Dynamics of redox potential and changes in redox stateof iron and humic acids during gut passage in soil-feeding termites (Cubitermesspp.). Soil Biol. Biochem. 34, 221e227.

Komlos, J., Jaffé, P.R., 2004. Effect of iron bioavailability on dissolved hydrogenconcentrations during microbial iron reduction. Biodegradation 15, 315e325.

Kostka, J., Nealson, K.H., 1998. Isolation, cultivation, and characterization of iron-and manganese- reducing bacteria. In: Burlage, R.S., Atlas, R., Stahl, D.,Geesey, G., Sayler, G. (Eds.), Techniques in Microbial Ecology. Oxford UniversityPress, Oxford, pp. 58e78.

Liptzin, D., Silver, W.L., 2009. Effects of carbon additions on iron reduction andphosphorous availability in a humid tropical forest soil. Soil Biol. Biochem. 41,1696e1702.

Logan, N.A., De Vos, P., 2009. Genus I. Bacillus Cohn 1872, 174AL. In: Boone, D.R.,Garrity, G.M., Castenholz, R.W., Brenner, D.J., Krieg, N.R., Staley, J.T. (Eds.), Ber-gey’s Manual of Systematic Bacteriology: the Firmicutes. Springer, New York,pp. 31e34.

Lovley, D.R., 1995. Microbial reduction of iron, manganese, and other metals. Adv.Agron. 54, 175e231.

Nocker, A., Burr, M., Camper, A.K., 2007. Genotypic microbial community profiling: acritical technical review. Microb. Ecol. 54, 276e289.

Parfitt, R.L., 1989. Phosphate reactions with natural allophane, ferrihydrite andgoethite. J. Soil Sci. 40, 359e369.

Peretyazhko, T., Sposito, G., 2005. Iron(III) reduction and phosphorous solubilizationin humid tropical forest soils. Geochim. Cosmochim. Acta 69, 3643e3652.

Pezzolesi, T.P., Zartman, R.E., Hickey, M.G., 2000. Effects of storage methods onchemical values of waterlogged soils. Wetlands 20, 189e193.

Potts, M., 1994. Desiccation tolerance of prokaryotes. Microbiol. Mol. Biol. Rev. 58,755e805.

Qu, D., He, J.Z., Sun, L.R., 2005. Microbial reducing characteristics of iron ox-ides in different paddy slurries. J. Northwest A. F. Univ. (Nat Sci. Ed.) 33,97e101.

Quantin, C., Becquer, T., Rouiller, J.H., Berthelin, J., 2002. Redistribution of metals in anew caledonia ferralsol after microbial weathering. Soil Sci. Soc. Am. J. 66,1797e1804.

Ranjard, L., Poly, F., Combrisson, J., Richaume, A., Gourbiere, F., Thioulouse, J.,Nazaret, S., 2000. Heterogeneous cell density and genetic structure of bac-terial pools associated with various soil microenvironments as determinedby enumeration and DNA fingerprinting approach (RISA). Microb. Ecol. 39,263e272.

Rapin, F., Tessier, T., Campbell, P.G.C., Carignan, R., 1986. Potential artifacts in thedetermination of metal partitioning in sediments by a sequential extractionprocedure. Environ. Sci. Technol. 20, 836e840.

Reyes, I., Torrent, J., 1997. Citrate-ascorbate as a highly selective extractant forpoorly crystalline iron oxides. Soil Sci. Soc. Am. J. 61, 1647e1654.

Sheppard, S.C., Addison, J.A., 2007. Soil sample handling and storage. In:Carter, M.R., Gregorich, E.G. (Eds.), Soil Sampling and Methods Analysis, seconded. CRC Press, Boca Raton, pp. 39e49.

Silver, W.L., Lugo, A.E., Keller, M., 1999. Soil oxygen availability and biogeochemistryalong rainfall and topographic gradients in upland wet tropical forest soils.Biogeochemistry 44, 301e328.

Soil Survey Staff, 2010. Keys to Soil Taxonomy, eleventh ed. USDA-Natural ResourcesConservation Service, Washington, DC.

Teh, Y.A., Silver, W.L., 2006. Effects of soil structure destruction on methane pro-duction and carbon partitioning between methanogenic pathways in tropicalrain forest soils. J. Geophys. Res. Biogeosci. 111, 2005e2012.

Teh, Y.A., Dubinsky, E.A., Silver, W.L., Carlson, C.M., 2008. Suppression of meth-anogenesis by dissimilatory Fe(III)-reducing bacteria in tropical rain forestsoils: implications for ecosystem methane flux. Glob. Change Biol. 14, 413e422.

Teixeira, L.C.R.S., Peixoto, R.S., Cury, J.C., Sul, W.J., Pellizari, V.H., Tiedje, J.,Rosado, A.S., 2010. Bacterial diversity in rhizosphere soil from Antarctic vascularplants of Admiralty Bay, maritime Antarctica. ISME J. 4, 989e1001.

Thompson, A., Chadwick, O.A., Rancourt, D.G., Chorover, J., 2006. Iron-oxide crys-tallinity increases during soil redox oscillations. Geochim. Cosmochim. Acta 70,1710e1727.

Thomson, E.A., Luoma, S.M., Cain, D.J., Johansson, C., 1980. The effect of samplestorage on the extraction of Cu, Zn, Fe, Mn and organic material from oxidizedestuarine sediments. Water Air Soil Pollut. 14, 215e233.

Turner, B.L., Romero, T.E., 2010. Stability of hydrolytic enzyme activity and microbialphosphorus during storage of tropical rain forest soils. Soil Biol. Biochem. 42,459e465.

Valencia-Cantero, E., Hernández-Calderón, E., Velázquez-Becerra, C., López-Meza, J.E., Alfaro-Cuevas, R., López-Bucio, J., 2007. Role of dissimilatory fermen-tative iron-reducing bacteria in Fe uptake by common bean (Phaseolus vulgarisL.) plants grown in alkaline soil. Plant Soil 291, 263e273.

Verchot, L.V., 1999. Cold storage of a tropical soil decreases nitrification potential.Soil Sci. Soc. Am. J. 63, 1942e1944.

Weiss, J.V., Emerson, D., Megonigal, J.P., 2004. Geochemical control of microbialFe(III) reduction potential in wetlands: comparison of the rhizosphere to non-rhizosphere soil. FEMS Microbiol. Ecol. 48, 89e100.

White, A.F., Blum, A.E., Schulz, M.S., Vivit, D.V., Stonestrom, D.A., Larsen, M.,Murphy, S.F., Eberl, D., 1998. Chemical weathering in a tropical watershed,Luquillo Mountains, Puerto Rico. I. Longterm versus short-term weatheringfluxes. Geochim. Cosmochim. Acta 62, 209e226.

Wildung, R.E., Li, S.W., Murray, C.J., Krupka, K.M., Xie, Y., Hess, N.J., Roden, E.E., 2004.Technetium reduction in sediments of a shallow aquifer exhibiting dissimila-tory iron reduction potential. FEMS Microbiol. Ecol. 49, 151e162.

Woomer, P.L., 1994. Most Probable Number Counts. In: Sssa Book Series, vol. 5,pp. 59e79.

Zachara, J.M., Kukkadapu, R.K., Fredrickson, J.K., Gorby, Y.A., Smith, S.C., 2002. Bio-mineralization of poorly crystalline Fe(III) oxides by dissimilatory metalreducing bacteria (DMRB). Geomicrobiol. J. 19, 179e207.