SOIL MICROBIOLOGY

Rice to Vegetables: Short- Versus Long-Term Impactof Land-Use Change on the Indigenous SoilMicrobial Community

Bo Sun & Zhi-Xing Dong & Xue-Xian Zhang & Yun Li &Hui Cao & Zong-Li Cui

Received: 10 November 2010 /Accepted: 15 January 2011 /Published online: 5 February 2011# Springer Science+Business Media, LLC 2011

Abstract Land-use change is known to have a significanteffect on the indigenous soil microbial community, but it isunknown if there are any general trends regarding how thiseffect varies over time. Here, we describe a comparativeanalysis of microbial communities from three adjacentagricultural fields: one-century-old paddy field (OP) and twovegetable fields (new vegetable field (NV) and old vegetablefield (OV)) that were established on traditional paddy fields10 and 100 years ago, respectively. Soil chemical and physicalanalysis showed that both vegetable fields were more nutrientrich than the paddy field in terms of organic C, total N, total P,and available K. The vegetable fields possessed relativelyhigher abundance of culturable bacteria, fungi, and specificgroups of bacteria (Actinomyces, nitrifying bacteria, andcellulose-decomposing bacteria) but lower levels of micro-bial biomass C and N. Notably, the decrease of biomass wasfurther confirmed by analysis of seven additional soils inchronosequence sampled from the same area. Next weexamined the metabolic diversity of the microbial communityusing the EcoPlateTM system from Biolog Inc. (Hayward,

CA, USA). The utilization patterns of 31 unique C substrates(i.e., community-level physiological profile) showed thatmicroorganisms in vegetable soil and paddy soil prefer touse different C substrates (polymeric compounds for NVandOV soils, phenolic acids for OP soil). Principal componentanalysis and the average well color development datashowed that the NV is metabolically more distinct from theOVand OP. The effect was likely attributable to the elevatedsoil pH in NV soil. Furthermore, we assessed the diversity ofsoil bacterial populations using the cultivation-independenttechnology of amplified ribosomal DNA restriction analysis(ARDRA). Results showed that levels of bacterial diversityin OP and NV soils were similar (Shannon’s diversity indexH=4.83 and 4.79, respectively), whereas bacteria in OV soilhave the lowest score of diversity (H=3.48). The low levelof bacterial diversity in OV soil was supported bysequencing of ten randomly selected 16S rDNA clonesfrom each of the three rDNA libraries. Phylogeneticanalysis showed that all the ten OV clones belonged toProteobacteria with eight in the gamma-subdivision andtwo in the alpha-subdivision. In contrast, the ten clonesfrom NV and OP soils were classified into four and eightbacterial classes or unclassified groups, respectively.Taken together, our data suggest that land-use changefrom rice to vegetables resulted in a decrease of bacterialdiversity and soil biomass despite an increase in theabundance of culturable microorganisms and, moreover,the decrease of bacterial diversity occurred during long-term rather than short-term vegetable cultivation.

Introduction

Soil microorganisms are active components of a terrestrialecosystem that contribute significantly to plant health and

Electronic supplementary material The online version of this article(doi:10.1007/s00248-011-9807-x) contains supplementary material,which is available to authorized users.

B. Sun (*) : Z.-X. Dong :Y. LiInstitute of Soil Science, Chinese Academy of Sciences,71 East Beijing Road,Nanjing 210008, Chinae-mail: bsun@issas.ac.cn

X.-X. ZhangInstitute for Advanced Study, Massey University,Auckland 0745, New Zealand

H. Cao : Z.-L. CuiKey Laboratory of Microbiological Engineering of AgriculturalEnvironment, MOA, Nanjing Agricultural University,Nanjing 210095, China

Microb Ecol (2011) 62:474–485DOI 10.1007/s00248-011-9807-x

soil productivity. Thus, there is growing interest inunderstanding how the indigenous soil microbial commu-nity is affected by agricultural practices, particularly whenland is converted to a different agricultural purpose. Changeof land use is common and is often associated with a shiftof land management strategies, e.g., the utilization oforganic and inorganic fertilizers, soil tillage, and irrigation.Many reports have shown that agricultural practices cause adecrease of microbial diversity [11, 22, 24, 32]. However, itis currently less clear whether the impact is greater overshorter or longer periods of time.

Some studies reveal a rapid response in the soilmicrobial community to land-use change. For example,using phospholipid fatty acid (PLFA) composition andcommunity-level physiological profile (CLPP) analysis,Yao et al. [34] examined the microbial community in achronosequence of turfgrass that had been established fromnative pines. Their results show that the primary factorresponsible for alteration of the microbial community wasthe initial land-use change rather than ongoing turfgrassmanagement over 95 years. By using the similar methods ofPLFA and CLPP analysis, Kulmatiski and Beard [17] alsofound that plants could change soil microbial communitycomposition within a few years.

Curiously, other studies have led to the oppositeconclusion that greater impact results from long-term ratherthan short-term agricultural practices. Buckley and Schmidt[4] examined microbial community structure, mainly bymeasuring relative abundance of ribosomal RNA, in fieldswith different cultivation histories. A significant differencewas observed between uncultivated fields and those havinga long-term history (>50 years) of cultivation [4]. Consis-tently, only fields that had been abandoned from cultivationfor >45 years had similar microbial community structure tofields that had never been cultivated [5]. Stark et al. [27]employed 16S rDNA polymerase chain reaction–denaturinggradient gel electrophoresis (PCR–DGGE) to examine theeffects of land management history on eubacterial andactinomycete communities. Their data revealed a close linkbetween microbial community structure and long-termmanagement history; no significant effect was observedfrom one-off amendment with organic matter [27].

The above contradicting evidence regarding short-versus long-term effects may represent fundamentaldifferences among the ecosystems analyzed. It is alsopossible that the differences were due to the differentanalytical techniques used. Therefore, further case studieswould help clarify whether there is a general trend thatthe response of the soil microbial community to land-usechange is either rapid or slow. In such studies, multiplemicrobiological techniques to assess the microbial com-munity should be employed in order to minimize thetechnical bias.

As urbanization increases the demand for fresh vegeta-bles, a lot of traditional rice-framing land has beenconverted to vegetable fields during the past two decadesin China. According to the China Agricultural Yearbook1981 and 2003 [21], the vegetable growing area in Chinaincreased from 3.5 million ha to 17.9 million ha; however,in the same period of time, rice cultivation decreased from33.3 million ha to 26.5 million ha. Management of paddyfields and vegetable fields is very different, particularlyregarding water supply and N fertilizer application. Adap-tation of the indigenous soil microbial community to landconversion from paddy fields to vegetable fields is thusinteresting but poorly understood. There have been fewreports on comparison of microbial communities in paddyand vegetable fields [7, 15, 18, 26], but a systematicanalysis is currently lacking.

The long-term goal of our research is to elucidate theresponse of the soil microbial community to land-useconversion from paddy to vegetable over short versus longperiods of time. Here we report an initial comparison ofindigenous microbial communities in three cultivated fields:new vegetable (NV), old vegetable (OV), and old paddy(OP). The two vegetable fields NV and OV have been usedfor horticultural purposes for 10 and 100 years, respective-ly. Both vegetable fields were established from traditionalpaddy fields. An adjacent OP was thus selected forcomparative analysis. This initial analysis was performedin three soils in chronosequence so that the microbialcommunity can be assessed by multiple techniques, whichimportance is highlighted above.

The area of investigation is located in the centralYangtze River Delta of southeast China, which has a longrecorded history of rice farming of more than 1,000 years[8]. This study was motivated by the dramatic difference insoil management between vegetable and paddy fields.Paddy fields in this region have a typical crop rotation ofrice (June to October) and wheat (November to May), andthe fields are subjected to flooding during the rice seasonbut are well drained during the wheat season. On thecontrary, crop rotation in vegetable fields is complex,irregular, and involves more than seven crops, e.g., Chinesecabbage, lettuce, leek, cucumber, tomato, garlic, andcowpea. The vegetable fields are managed under continu-ous drainage conditions. Another significant difference isthat vegetable fields have much higher application rates ofN fertilizer than paddy fields with an annual inputof ~580 kg ha−1 for paddy fields and ~1,900 kg ha−1 forvegetable fields.

We analyze the indigenous soil microbial communitiesin paddy soil and two vegetable soils in chronosequence,using both the classic plate counting method and therecently developed cultivation-independent technology ofamplified ribosomal DNA restriction analysis (ARDRA).

Microbial Adaptation in Paddy and Vegetable Soils 475

The ARDRA analysis was complemented by full-length 16SrDNA sequencing of ten clones from each soil. We examinethe functional diversity of soil microbial communities usingCLPP analysis, which assesses the utilization of 31 uniquecarbon substrates. The data are analyzed to test whether soilmicrobial communities differ significantly in paddy andvegetable fields and, moreover, whether the impact followingland use change is greater in the short term or long term.

Methods

Site Description and Soil Sampling

This study was carried out in the suburbs of Yixing citylocated in Jiangsu province, southeast China. This areabelongs to the northern subtropical monsoon region with amean annual rainfall of 1,158 mm and temperature of 15.7°C. Soils in this area have been classified as typicalFe-leaching Stagnic Anthrosols, which were derived fromfluviolimnic deposition. This study focused on three land-use types (OV, NV, and OP), and each has a unique historyof agricultural practice. OV soil has been used for growingvegetables for ~100 years; NV soil is a 10-year-oldvegetable soil; OP soil has grown rice–wheat in rotationfor a period of more than a century. Both vegetable soils(OV and NV) were converted from old paddy soils.

Geographic locations of the sampling sites are listed asbelow: OV, 31°23′7.00″~31°23′13.28″ N to 119°58′33.77″~119°58′41.73″ E; NV, 31°23′27.32″~31°23′35.50″ Nto 119°58′7.67″~119°58′17.29″ E; OP, 31°23′41.90″~31°23′48.73″ N to 119°58′34.65″~119°58′42.79″ E. The threefields are separated from each other by ~500 m. A surveyof local farmers and official records showed that thevegetable fields have much higher fertilizer input thanpaddy soils. While about 1,900 kg N, 250 kg P, and1,500 kg K fertilizers are generally applied annually tovegetable fields, the annual fertilizer use in paddy soil isabout 580 kg N, 68 kg P, and 124 kg K (per hectare). Thereis a regular rotation of rice and wheat in paddy soil.However, crop rotation is irregular and more frequent invegetable fields. When soils were sampled, rice was grownin the OP field and the same crop (mature Chinese cabbage)was grown in OV and NV fields.

For each land-use type (OV, NV, or OP), one plot of ~5 hawas selected and divided into three subplots (20×20 m), whichare separated from each other by ~100 m. The sampling siteswere subjected to a prior analysis of spatial variation in termsof soil fertility (organic carbon, pH, total N, total P, and totalK), and the results suggested that, with a coefficient varianceless than 5%, the OP, NV, and OV fields should be analyzedwith a minimal of 9, 13, and 16 soil samples, respectively (datanot shown). Based on the spatial variance analysis, each

subplot was grid-sampled at 5×5 m with 25 samples taken atthe surface layer (0–15 cm). Soil sampling was performedusing a hand auger (2 cm diameter×20 cm depth). The 25 drysamples from each subplot were then bulked to produce a totalof 3 samples for each land-use type. Soil subsamples wereeither stored at 4°C for microbial analysis or air-dried andsieved (<2 mm) for chemical analysis.

Seven additional surface soil samples (0–15 cm) weretaken from the same area of Yixing city (31°22′18″N, 119°49′45″ E). These vegetable fields (~0.03 ha) have been convertedfrom paddy fields and subjected to continuous vegetablefarming for 1, 5, 10, 15, 20, and 30 years. Every field hadthree replicate soil samples with each being a mixture of eightto ten subsamples. A nearby traditional paddy field of similarsize was used as a control (year 0). These samples weremicrobiologically analyzed in this work to further confirm theopposite trends of microbial abundance (i.e., microbialbiomass C (MBC) and culturable aerobic bacteria). Of noteis that five such soil samples (years 5, 10, 15, 20, and 30) wereused in an overlapping study, which focused on theaccumulation of organochlorine pesticides in these soils. Soilproperties of these five fields have been previously reported[15]. A list of main soil properties of all the seven samples isavailable in Table S1.

Soil Chemical and Physical Characterization

Soil electrical conductivity was measured with an electrolyticconductivity meter (soil/water=1:5). Soil pH was determinedwith a glass electrode (soil/water=1:2.5). Soil organic C andtotal N content was determined using the standard procedureof dichromate oxidation and Kjeldahl digestion, respectively.Mineral Nwas extracted with 2MKCl (soil/KCl=1:4) for 1 h.NH4-N and NO3-N in soil extracts were determined usingthe Segmented Flow Analyzer (Skalar SANplus, TheNetherlands). Total P and available P were extracted withHF–HNO3–HClO4 and sodium bicarbonate, respectively,and then determined by the molybdenum blue method. TotalK and available K were extracted with HF–HNO3–HClO4

and ammonium acetate, respectively, and then determined byflame photometry. A laser diffraction particle size analyzerpurchased from Beckman Coulter LS 230 (USA) was usedto determine the soil particle size distribution following themanufacturer’s recommendations. A field emission scanningelectron microscope (SEM) from FEITM was used todetermine the soil elemental composition, and the micro-scope is equipped with an energy-dispersed X-ray spectros-copy (20 kV).

Soil Microbiological Analysis

The classic dilution plating method was used to enumerateculturable microorganisms in soil samples. Colony forming

476 B. Sun et al.

units (CFU) of total bacteria, fungi, and Actinomyceswere determined in beef extract peptone medium,Czapek’s medium, and Gause’s no. 1 synthetic medium[3], respectively. Enumeration of nitrifying bacteria andcellulose-decomposing bacteria was performed as previ-ously described [33].

Soil MBC and microbial biomass N (MBN) weredetermined using the previously described fumigation-extraction method [31]. Extractable organic C in the extractwas measured using an automatic analyzer (modelTOC500, Japan) and N by steam distillation. MBC andMBN were calculated from: MBC=2.64EC and MBN=1.85EN, where EC and EN are the difference between theamount of organic C and N extracted from the fumigatedand non-fumigated soils, respectively.

Metabolic Fingerprinting of Soil Samples

CLPP were obtained using the Biolog EcoPlateTM systemrecently developed by Biolog, Inc. CA, USA. Each 96-well plate assesses the utilization of 31 unique carbonsources in triplicate [6]. To prepare the soil suspensions asinoculants, 10 g fresh soil was suspended into 100 mlsterile phosphate buffer (0.05 M, pH 7.0) in a 250-mlflask. After shaking at 200 rpm for 10 min, soil solution(i.e., 10−1 dilution) was then diluted to 10−3 and 100 μl perwell was inoculated into the microtiter plate. The plate wasincubated at 25°C and color development was monitoreddaily over a period of 6 days using an Emax PrecisionMicroplate Reader (Biolog Inc., Hayward, CA, USA).Absorbance at 590 nm was recorded and analyzed usingthe Microlog Rel. 4.2 software (Biolog Inc., Hayward,CA, USA). Microbial metabolic activity was expressed asaverage well color development (AWCD) and calculatedas described in Zak et al. [35].

Amplified Ribosomal DNA Restriction Analysis and 16SrDNA Sequencing

The soil bacterial community was analyzed at the molecularlevel using the cultivation-independent method of ARDRAfollowed by 16S rDNA sequencing. Total DNA wasprepared from a 50-g soil sample according to the protocoldescribed by Gabor et al. [12] with slight modifications.Soil extract solution was freeze-thawed twice before SDSand enzymatic treatment; the amount of lysozyme andproteinase K was increased 5-fold. The total soil DNA wasthen subjected to 16S rDNA amplification using theuniversal bacteria-specific primers 27f (5′-AGAGTTTGATCCTGGCTC AG-3′) and 1492r (5′-TACCTTGTTACGACTT-3′) [10]. PCR products were purified through 1%agarose gel and subsequently cloned into the pMD18-Tvector (TaKaRa, Dalian, China).

Restriction analysis of the cloned 16S rDNAs wasperformed by PCR amplification of the inserts using a pairof primers M13-47 (5′-CGCCAGGGTTTTCCCAGTCACGAC-3′) and RV-M (5′-GAGCGGATAACAATTTCACACAGG-3′), located within the vector, flanking thecloned DNA fragment. After digestion with HhaI or RsaI,the DNA fragments were separated in 2% agarose gels.Restriction profiles were documented in a photographsystem (Tanon GIS 2500, Shanghai, China) and analyzedusing the Tannon Digital Image analysis software. UniqueHhaI and RsaI restriction patterns were defined asoperational taxonomic units (OTUs), representing different16S rDNA genotypes. A total of 135, 159, and 174randomly selected 16S rDNA clones were analyzed fromOP, NV, and OV soils, respectively.

Full-length 16S rDNA sequencing was performed byInvitrogen (Shanghai, China). DNA sequences were assem-bled with the help of BioEdit v7.0.1.14. Multiple 16SrDNA sequences were aligned and neighbor-joining treesconstructed using CLUSTALX [28]. The tree was displayedand edited in TreeView [23]. The 30 unique 16S rDNAsequences reported in this work are available in theGenBank database under the accession numbersDQ011835, DQ011836, DQ011838–DQ011843,DQ086464–DQ086465, and DQ083100–DQ083119.

Data Analysis

Soil property data were subjected to the post hoc multiplecomparisons of ANOVA analysis with the help of theStatistical Package for Social Science (SPSS 13, SPSS Inc.,Chicago, IL, USA). Significance among means wasidentified using least significant difference calculated atthe level of P=0.05. Principal component analysis (PCA) ofthe Biolog data (A590) was performed according to apreviously described method [29]. Only components witheigenvalue >1 and explaining >10% of the variability wereretained. Varimax rotation with Kaiser normalization wasthen performed to enhance interpretability of the uncorre-lated components. All meaningful loadings (i.e., loadings>0.50) were included in the interpretation of principalcomponents. Graphic interpretation of PCA analysis wasdone by constructing a biplot of the first two principalcomponents for each soil sample. The values of first twoprincipal components were subject to multiple t test.

Shannon’s diversity index (H), Shannon’s index of even-ness (E), and Margalef richness index (M) were calculated

using the following equations: H ¼ �Psi¼1

Pi lnðPiÞ, Pi ¼ niN,

E ¼ HlnðSÞ, and M ¼ ðS�1Þ

lnN . For analysis based on the Biolog

data, Pi is the proportional color development of ith well (ni)relative to total color development (N) of all wells; theMargalef index is expressed as the number of wells with

Microbial Adaptation in Paddy and Vegetable Soils 477

positive reaction (S) divided by natural logarithm of totalcolor development (N, sum of A590 for all 31 wells after96 h). The S value is 31 in this work because all 31 wellsshowed positive color developments for each field. Substraterichness index (RI) refers to the number of wells or substratesthat have net absorbance values greater than 0.25 [35]. In theARDRA data analysis, Pi refers to the ratio of OTUi to thetotal number of clones (N). Similarity between the composi-tions of two microbial communities was estimated using theSorensen index (CZ): CZ ¼ 2c

ðaþbÞh i

� 100, where c is thenumber of OTUs in common and a and b are the totalnumbers of OTUs for each community.

Results

Soil Properties

A summary of soil properties is provided in Table 1. Theold paddy soil has significantly lower contents of organicC, total N, and total P with higher C/N and N/P ratioscompared to the vegetable soils (NV and OV). Interestingly,the three soils have similar levels of total K but differsignificantly in available K (P<0.05); available K in oldvegetable (OV) soil is ~5 and ~6.5 times higher than that ofNV and OP soil, respectively. Compared with the OV andOP soils, the NV soil has significantly higher pH andelectrical conductivity (Table 1).

Soil structure was examined under a SEM. Representa-tive images are shown in Fig. 1. The three soils exhibitsimilar patterns of soil texture. However, soil particles inOP soil are relatively smaller than those of vegetable soils(OV and NV). The difference was likely due to the lengthof the soil particles with all three soils displaying particles

of a similar diameter of ~20 μm (Fig. 1). Elementalcomposition in fine soil particles (n=8) was determinedby SEM analysis with the help of an energy-dispersiveX-ray spectrometer detector (20 kV). Among the sevenelements analyzed (C, O, Al, Si, K, Mn, Fe), a significantdifference was found only in C (mean ± standard error): OP,60.86±2.11%; NV, 63.79±1.30%; and OV, 67.83±1.76%.Notably, the C element data are consistent with the organicC contents of the soils (Table 1).

Abundance of Soil Microbial Communities

The classic plate counting method was employed todetermine the abundance of culturable soil microorganisms.Total number of bacteria, fungi, Actinomyces, and thespecific groups of nitrifying bacteria and cellulose-decomposing bacteria were enumerated under differentgrowth media (see details in “Methods”). Results aresummarized in Table 2. There are significant differencesbetween the vegetable soils and the control paddy soil inthe abundance of total bacteria, fungi, and Actinomyces.The relative lower abundance in paddy soil is consistentwith the observed difference of soil nutrients in terms oforganic C, total N, and total P (Table 1). The analysis alsorevealed a significant difference between the two vegetablesoils. The NV field contained a significantly higher numberof total bacteria as well as nitrifying bacteria, comparedwith the OV field. Interestingly, both the total number offungi and the cellulose-decomposing bacteria showed thesame trend in order from highest to lowest abundance ofOV>NV>OP.

To further estimate the general abundance of themicrobial population, the three soils were subjected toMBC and MBN analysis. Data shown in Table 2 indicate

Soil propertya OP NV OV

Organic C (g kg−1) 13.5±0.4 (a) 16.4±1.1 (b) 16.9 ±1.0 (b)

Total N (g kg−1) 1.43±0.11 (a) 1.94±0.13 (b) 1.96±0.14 (b)

Total P (g kg−1) 0.53±0.03 (a) 1.96±0.19 (b) 1.49±0.13 (c)

C/N 9.49±0.47 (a) 8.47±0.22 (b) 8.66±0.34 (b)

N/P 2.73±0.25 (a) 1.03±0.10 (b) 1.47±0.42 (b)

Total K (g kg−1) 10.42±018 (a) 9.78±0.11 (a) 10.37±0.31 (a)

NH4-N (mg kg−1) 10.20±0.43 (a) 9.17±0.58 (a) 9.78±0.92 (a)

Available K (mg kg−1) 48.5±2.0 (a) 257.1±15.0 (b) 314.9±10.6 (c)

pH 5.60±0.06 (a) 6.84±0.21 (b) 5.45±0.31 (a)

Electric conductivity (mS m−1) 147.5±2.4 (a) 243.2±2.3 (b) 106.3±3.2 (c)

Soil water content (g kg−1) 580.7±11.9 (a) 347.5±15.4 (b) 332.0±6.9 (b)

Clay (<0.002 mm; %) 9.0±0. 8 (a) 11.0±0.8 (b) 10.6±0.7 (b)

Fine silt (0.002–0.05 mm; %) 80.8±4.9 a 80.5±4.4 a 83.3±5.7 a

Coarse silt (0.05–0.2 mm; %) 6.5±0.5 (a) 7.7±0.5 (b) 5.6±0.4 (a)

Sand (0.2–2 mm; %) 3.7±0.3 (a) 0.8±0.2 (b) 0.5±0.1 (b)

Table 1 Chemical and physicalproperties of soil samples fromtop layer (0–15 cm) of paddyfield (OP) and two vegetablefields (NV and OV)

Scores identified by differentletters in parenthesis are signif-icantly different (P<0.05)among three soilsa Data are means and standarderrors of three repeats

478 B. Sun et al.

significant lower MBC and MBN values in OV soil thanOP soil with NV soil in the intermediate. This result issurprising given the observed increase of culturablemicrobes from OP, NV to OV soils.

The contradictory data for soil biomass and culturablemicrobes described above prompted a detailed microbio-logical analysis with more soil samples. In a separate study[15], which aimed to examine the dynamics of residualorganochlorine pesticides, soils samples were taken fromvegetable fields in the same area in a chronosequence of 0,1, 5, 10, 15, 20, and 30 years of land conversion frompaddy to vegetable. These soils were subjected to furthermicrobiological analysis here, and the data of MBC, MBN,and culturable bacteria and fungi are presented in Fig. 2.The results clearly indicate a decrease of soil MBC overtime (R2=0.96, Fig. 2a), and the MBC value is stabilizedfrom year 15 and remains at the level of ~200 μg g−1. TheMBN data displays the similar trend as MBC (Fig. 2b, R2=0.66), and there is significant positive correlation betweenthese values (R=0.782, P=0.002). MBC values are nega-tively correlated with those of SOC (R=−0.60, P=0.003)and total P (R=−0.93, P=0.002). Significantly, the totalnumber of bacteria and fungi are higher in vegetable soilsthan in paddy soil (year 0). Taken together, the analysis ofadditional soil samples further confirmed the observed

decrease of soil biomass (MBC and MBN) despite anincrease of culturable microbes.

Functional Diversity of Soil Microbial Communities

CLPPs were obtained using the Biolog EcoPlateTM

system, whereby the ability to utilize 31 carbon sourceswas assessed in triplicate. The resulting AWCD datashown in Fig. 3a indicate that microbial activity variedsignificantly in order from highest to lowest score of NV>OV>OP over a period of 2–5 days. The same trend (NV>OV>OP) was found when functional diversity wasexpressed as Shannon’s diversity index (H) and evennessindex (E; Table 3). However, Margalef richness index (M)and substrate RI resulted in an opposite trend OP>OV>NV (Table 3). A strong negative correlation betweenmicrobial richness and diversity (correlation coefficient(R) between H and M −0.997, P=0.003).

To further examine the extent of variations regardingcarbon source utilization, substrate utilization patterns ofthe three soils were subjected to PCA. Results are shown inFig. 3b. The first principal component exhibits 36.6% of thedata variance and produces good separation of NV soilfrom OV and OP soils at the level of significance P<0.05.All three soils were significantly different (P<0.05), when

OP NV OV

Figure 1 Representative images of soil particles under scanning electron microscope. OP old paddy soil, NV new vegetable soil, OV oldvegetable soil

Microorganism OP NV OV

Bacteria (107 CFU g−1) 3.07±0.37 (a) 8.58±0.88 (b) 5.26±0.67 (c)

Fungi (105 CFU g−1) 0.49±0.14 (a) 1.79±0.43 (b) 3.53±0.70 (c)

Actinomyces (107 CFU g−1) 3.27±0.56 (a) 14.80±2.27 (b) 14.22±1.19 (b)

Nitrification bacteria (103 CFU g−1) 5.40±2.78 (a) 42.12±9.10 (b) 4.68±1.04 (a)

Cellulose-decomposing bacteria (CFU g−1) 128±55 (a) 731±137 (b) 1,902±395 (c)

MBC (μg g−1) 625.1±106.1 (a) 250±3 (b) 193±6 (c)

MBN (μg g−1) 46.59±2.7 (a) 27.5±4.4 (b) 29.3±5.8 (b)

Table 2 Abundance ofmicroorganisms in paddy soil(OP) and vegetable soilsNV and OV

Data are means and standarderrors of three repeats. Scoresidentified by different letters inparenthesis are significantlydifferent (P<0.05) among threesoils

Microbial Adaptation in Paddy and Vegetable Soils 479

both the first and second principal components were takeninto account, representing 57.7% of the total variation(Fig. 3b).

Next we examined that the preference of C substrateutilization−substrate loading score was calculated accord-ing to their contribution to the separation of three microbialcommunities by the first principal component. As shown inTable 4, substrates with positive loading scores wereassociated with vegetable soils, whereas those with nega-tive scores were associated with the paddy soil. Carbohy-drates were the most preferred C substrates for all threesoils (Table 3). Interestingly, polymeric compounds weremore preferred by microorganisms in the vegetable soilsthan the paddy soil, whereas phenolic acid substrates weremore preferred by the microbial community in the paddysoil than the vegetable soils. Taken together, the substrateutilization data imply that microbial communities in thevegetable soils and paddy soil differ in their preferred Csubstrates.

Phylogenetic Diversity of Soil Bacterial Populations

A cultivation-independent approach (ARDRA) was used toassess the phylogenetic diversity of the bacterial popula-tions. To this end, total DNAs were prepared from soil

samples, from which 16S rDNAs were amplified andcloned into Escherichia coli. The cloned 16S rDNAfragment was then subjected to restriction analysis. Eachunique restriction profile is considered as an OTU. Theresulting number of OTUs was plotted against the numberof clones analyzed. Results shown in Fig. 4 indicated thatthe OTU numbers are almost saturated for all three soils.OV soil has the lowest bacterial diversity with only 63OTUs identified, whereas more than twice as many OTUswere identified in NV and OP soils with 133 and 128OTUs, respectively. Three biological indexes (Shannon’sdiversity index, evenness index, and the Margalef richnessindex) were then calculated. As shown in Table 5, while thescores for OP and NV are similar, there is an obviousdifference between OV and OP/NV soils (OP≈NV>OV).Taken together, the phylogenetic data clearly show that adecrease of bacterial diversity occurred after long-term (notshort-term) adaptation to land-use change from rice tovegetables.

Analysis of Bacterial Community Structures

To compare bacterial community structures from the threesoils, CZ was calculated based on the ARDRA datadescribed above. The 16S rDNA libraries shared only

100

200

300

400

500

600

700

800

MB

C (

mg

kg-1

)

Year

0

10

20

30

40

50

Year

MB

N (

mg

kg-1

)

0

2

4

6

8

10

Year

CF

U (

107

g-1 )

0

2

4

6

8

10

12

14

16

0 1 5 10 15 20 30 0 1 5 10 15 20 30

0 1 5 10 15 20 30 0 1 5 10 15 20 30

YearC

FU

(10

5g-

1 )

A (MBC) B (MBN)

C (Bacteria) D (Fungi)

Figure 2 Dynamic changes ofmicrobial abundance invegetable fields within 30 yearsof conversion from paddy fields.Soil samples were taken fromthe surface layer (0–15 cm) andone-century-old paddy field wasused as a control (year 0). Dataare means and standard errorsof three samples

480 B. Sun et al.

1.6% similarity between OP and NV and 1% between NVand OV. No similarity was revealed between OP and OV(CZ=0). This result suggests that the three soils areoccupied by different bacterial genotypes. To provide asupport of this finding and also get a sense of thetaxonomic identity of the dominant genotypes, we per-formed 16S rDNA sequencing of ten randomly selectedclones from each of the three rDNA libraries. A phyloge-netic tree was constructed based on bootstrap analysis offull-length 16S rDNA sequences. Results shown in Fig. 5indicate that the bacterial population in OV soil isdominated by members of the γ-Proteobacteria with eightout of ten clones classified in this bacterial class. Theremaining two OV clones belong to the α-Proteobacteria.In contrast, bacteria from NV and OP soils belong tovarious phyla. The ten NV clones are composed of threeγ-Proteobacteria, three α-Proteobacteria, three Firmicutesand one unclassified (Fig. 5). Interestingly, the ten OP

clones were classified into six classes and two unclassifiedgroups. Together the phylogenetic data suggest that thebacterial population in OV soil is less diverse than that ofNV soil and OP soil, which provides additional support tothe conclusion drawn from the ARDRA analysis (Fig. 5).

Discussion

The land-use change from rice to vegetable has been thefocus of our investigation, not only because of itsagricultural importance but also because of the ecologicaluniqueness involved in a dramatic change of land manage-ment from paddy fields to vegetable fields. Here, wecompared the microbial communities in a traditional paddyfield (OV) and two vegetable fields established 10 and100 years ago (NV and OV, respectively). The three fieldsunder investigation are ~5 ha in size and separated fromeach other by 500 m; each field was analyzed with threereplicates or subplots (20×20 m for each and ~100 m apartfrom each other). Moreover, the two vegetable soils (NVand NV) have been subjected to the similar vegetablerotation during the past 10 years. The shared plantationhistory was further confirmed by soil pollen compositionanalysis (data not shown). Together, both historical recordand prior experimental data show that soils from the threefields have the same origin—indeed the three soils sharedthe similar patterns of soil texture (Table 1 and Fig. 1).These ensure that the impact of land-use change over shortversus long periods of time can be appropriately assessed inthis work.

In a previous study [15], we examined the dynamics ofsoil properties in a chronosequence of 0, 5, 10, 15, 20, 30,and 50 years of land conversion from paddy to vegetable.The analysis indicated an increase of SOC from 14.1 to~16.9 gkg−1 and total N from 1.5 to 1.96 gkg−1 within thefirst 10 years of land-use change [15]. Soil analysis datapresented in this work showed that the NV and OV soilshad similar elevated levels of SOC (~16.6 gkg−1) and totalN (1.95 gkg−1). These results collectively suggest aninitial increase of soil nutrients (SOC and total N) within10 years, followed by stability in the nutritional statusduring the extended period of vegetable cultivationbetween 10 to 100 years. Notably, this finding is ingeneral agreement with other studies that have comparedthe soil properties of paddy and vegetable fields in thesame area of Yangtze River Delta [19, 26].

The increase of SOC is likely due to the decompositionof a large amount of disposed vegetable roots and remains.The annual organic manure input was estimated to be 84.9–111.3 tonnes/ha in vegetable fields, which was more than10-fold of the 7.5 tonnes input/ha for paddy fields [7].Interestingly, the aerobic cellulose-decomposing bacteria

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

-1.6

-1.2

-0.8

-0.4

0.0

0.4

0.8

1.2

1.6

24 48 72 96 120 144

-1.6 -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2 1.6

Time (hours)

AW

CD

A

B

Sec

ond

prin

cipa

l com

pone

nt (

21.1

%)

First principal component (36.6%)

OP

NV

OV

OP

NV

OV

Figure 3 Functional diversity of the indigenous microbial communityin OP, NV, and OV soils estimated by average well color development(AWCD) analysis (a) and principle component analysis (b). TheAWCD data are means and standard errors of three soil repeats.Average loading scores along the first and second principal compo-nents were shown with standard errors of nine repeats (three repeats ofsoil samples each with three repeats of Biolog EcoPlateTM)

Microbial Adaptation in Paddy and Vegetable Soils 481

were the highest in old vegetable soil (OV>NV>OP;Table 2), and the population size is correlated well withnitrogen availability of the soil (the correlation coefficientbetween cellulose-decomposing bacteria and ammoniumnitrogen is calculated as 0.946; P=0.013; Table 1 andTable S2). The higher abundance of aerobic cellulose-decomposing bacteria suggests that a significant proportionof organic carbon in vegetable soils had not been convertedinto forms that can support the growth of other saprophyticbacteria. This notion is consistent with the fact that paddysoils have the highest MBC compared with vegetable soils(Table 2 and Fig. 2).

The structure of bacterial population in paddy fielddiffers significantly under flooded than dry conditions [9].It has been well-known that anaerobes dominate the

population in flooded paddy soils [1, 25, 30]. However, inthis work, only aerobic bacteria were examined in thenumeration of culturable microorganisms and the assess-ment of functional diversity (i.e., the CLPP analysis). It isthus worthwhile to note that both the paddy field and thevegetable fields were dry when the soil samples were takenand, moreover, that the study has been focused on theoxygen-replete surface layer of the soil. These implicatethat the possible effect of active anaerobic bacteria isminimal. Indeed, the OP soil possesses the lowest abun-dance of aerobic bacteria and cellulose-decomposingbacteria (Table 2), which are consistent with the lowestSOC content in OP soil (Table 1).

Contrary to the decrease of soil biomass C and N, thetotal number of culturable aerobic bacteria and the specific

Substrate Substrate class Substrate loading score

D-Galacturonic acid Carboxylic acids −0.930β-Methyl-D-glucoside C Carbohydrates −0.785γ-Hydroxybutyric acid Carboxylic acids 0.759

L-Arginine Amino acids 0.763

α-Cyclodextrin Polymeric compounds 0.816

Tween 40 Polymeric compounds 0.856

Glycogen Polymeric compounds 0.870

4-Hydroxybenzoic acid Phenolic acids 0.893

D-Malic acid Carboxylic acids 0.903

Pyruvic acid methyl ester Carbohydrates 0.924

Table 4 List of ten carbonsubstrates that made the mostsignificant contribution to theseparation of microbial commu-nities from three soils (OP, NV,and OV)

Table 3 Summary of functional diversities of soil microbial communities in paddy field (OP) and the two vegetable fields (NV and OV) asdetermined by Biolog EcoPlateTM analysis

OP NV OV

Substrate classa

Carbohydrates 6 5 8

Carboxylic acids 3 3 2

Amino acids 3 2 4

Phenolic acids 3 1 1

Polymeric compounds 0 4 3

Amides 1 1 2

Phosphate compounds 1 1 1

Diversity indexb

Shannon’s diversity index (H) 3.08±0.03 (a) 3.31±0.02 (b) 3.19±0.04 (c)

Shannon’s evenness index (E) 0.898±0.008 (a) 0.963±0.005 (b) 0.930±0.011 (c)

Margalef richness index (M) 8.70±0.11 (a) 7.90±0.03 (b) 8.26±0.06 (c)

Substrate richness indexc 28.3±0.3 (a) 24.3±2.0 (a) 20.0±0.6 (b)

Scores identified by different letters in parenthesis are significantly different (P<0.05) among three soilsa Number refers to substrates within each class that made the most contributions to the separation of three fields (loading score>0.5) as revealed in theprincipal component analysisb Data are means and standard errors of three repeatsc Substrate richness refers to the number of wells that have net absorbance values greater than 0.25

482 B. Sun et al.

group of cellulose-decomposing bacteria showed theopposite trend with lower abundance in OP soils (Table 2and Fig. 2). Most soil-dwelling bacteria are uncultivableunder laboratory conditions. It is highly possible that theirresponse to land-use change in term of abundance isdifferent from the culturable ones and thus may contributeto the unexpected negative correlation between MBC andcultural aerobic bacteria.

While this work was not designed to elucidate theecological causes of biological and chemical dynamicchanges after land conversion, the analysis of sevenadditional soil samples can provide some insight. Thesesamples, which belong to the same soil type, were takenfrom the same geographic location, representing 0, 1, 5, 10,20, and 30 years of land conversion from paddy tovegetables. Significant positive correlations are foundbetween MBC and soil pH, water content, or total K(Table S2), suggesting a combined effect of tillage,irrigation, and fertilization. Regarding culturable aerobicbacteria, their abundance fluctuates over periods of timeduring vegetable cultivation (Fig. 2) and exhibits strongsignificant correlation with SOC, available K, and watercontent (Table S2). This suggests that abundance of culturalbacteria reflects the nutrient status of the soil and greatlyaffected by fertilization and irrigation.

We assessed the soil microbial diversity using CLPP andthe cultivation-independent method ARDRA. While CLPPreports functional capability of the whole soil microbialcommunity to metabolize 31 carbon sources, ARDRAprovides a glimpse of the soil bacterial communitycomposition. CLPP analysis revealed a significant differ-ence between paddy soil and vegetable soils—the paddysoil had the lowest metabolic activity as indicated byAWCD scores (Fig. 3a) and the Shannon’s diversity/evenness indexes (Table 3). This result is in agreementwith the poorest nutritional status of the paddy soil in termsof SOC, N, and P (Table 1); it is also consistent with thelowest abundance of culturable aerobic microorganisms inpaddy soil (Table 2). However, a previous CLPP analysis ofa 5-year-old vegetable soil in the same geographic region,which was established from traditional paddy field, showed

92 OV43 (DQ086464)

OV117 (DQ011838)

OV53 (DQ011841)100

82

85

100

99

97

100

100

100

100

0.1

100

92100

100

86

99

10097

85100

99

Pseudomonas fluorescens (GQ260082)OV66 (DQ011843)

OV135 (DQ011836)

OV49 (DQ011839)

OV50 (DQ011840)

OV65 (DQ011842)

OP138 (DQ083112)

NV14 (DQ083100)

NV174 (DQ083104)

NV35 (DQ083107)

OP7 (DQ083116)

NV84 (DQ083109)

NV164 (DQ083101)

Bradyrhizobium japonicum (AB195266)

OV23 (DQ011835)

OV11 (DQ086465)

NV33 (DQ083106)

OP27 (DQ083113)

NV20 (DQ083105)

NV172 (DQ083103)

Bacillus aestuarii (AB062696)

NV17 (DQ083102)

NV98 (DQ083108)

OP38 (DQ083115)

Holophaga foetida (X77215)

OP117 (DQ083110)Sphingoterrabacteriumcomposti (AB267719)

OP76 (DQ083118)

OP74 (DQ083117)

Magnetobacterium bavaricum (X71838)

OP137 (DQ083111)

OP34 (DQ083114)

Geobacter grbicium (AF335183)

Gamma-Proteobacteria

Alpha-Proteobacteria

Firmicutes

Unclassified

Acidobacteria

Bacteroidetes

Unclassified

Nitrospira

Delta-Proteobacteria

Figure 5 Phylogenetic tree showing the taxonomic positions of tenrandomly selected 16S rDNA clones for each of the three soils (OP,NV, and OV). Sequences are designated by prefixes OP, NV, or OVindicating from which soils they were isolated. GenBank accessionnumbers are given in parenthesis. Sequences showing less than 70%similarity at phylum level with sequences in the GenBank databasewere considered as “unclassified” groups. Sequence OP118(DQ083119) was phylogenetically very distinct from the rest of theenvironmental clones, and it was thus excluded from the tree

0

20

40

60

80

100

120

140

160

0 30 60 90 120 150 180

Number of clones

Num

ber

of O

TU

s OP

NV

OV

Figure 4 Rarefaction curve of the total number of OTUs against thetotal number of clones sampled from OP, NV, and OV soils

Table 5 Diversity of soil bacterial populations in paddy field (OP)and the two vegetable fields (NV and OV) as determined by ARDRAanalysis

Diversity index OP NV OV

Shannon’s diversity index (H) 4.83 4.79 3.48

Shannon’s evenness index (E) 0.993 0.941 0.675

Margalef richness index (M) 24.2 25.8 12.0

Total clones 135 159 173

Total OTUs 128 133 63

Microbial Adaptation in Paddy and Vegetable Soils 483

a slight decrease of metabolic function (AWCD anddiversity index) of soil microbial community after landconversion from paddy to vegetables [18]. The independent5-year data suggest possible small-scale fluctuations ofmicrobial community during the first 10 years of land-usechange.

The ability of CLPP to report microbial diversity islimited by the potential metabolic redundancy [16]. Otherfactors such as cultivability and relative abundance ofdifferent bacterial species also restrict the utility of CLPP inthe assessment of microbial diversity. It has thus been achallenging task to interpret the meaning of CLPP indexes[14]. The Margalef index was initially developed tomeasure species richness based on absolute numbers ofspecies and individuals [20], whereby the effect ofsampling size is minimized. The equation has also beenused to calculate species richness using density data such asbiomass, though caution should be taken [13]. Interestinglyin this work, when functional diversity was expressed asMargalef richness index, paddy soil had the highest“species” richness, which is contrary to the H and Eindexes (Table 3). The Margalef index of CLPP data iscorrelated with the substrate RI, which reports the number ofsubstrates that each field can significantly utilize (A590>0.25).Significantly, the ARDRA data clearly indicate that OP soilhas the highest bacterial diversity, which is consistent withthe Biolog Margalef richness index (M) but not the Biologdiversity indexes (H and E). Together, our data suggest theusefulness of richness index (M) in assessing functionalmicrobial diversity.

Bacterial community composition in paddy soils hasrecently been studied in Japan using DGGE analysiscombined with 16S rDNA sequencing [2]. Sequenceanalysis of 250 DGGE bands led to the identification of13 phyla with Proteobacteria and Bacteroidetes as domi-nant groups that were present in all tested habitats.Interestingly, in this work, eight of the ten 16S rDNAsclones are classified into four phyla (Proteobacteria (alpha,gamma, and delta subdivisions), Bacteroidetes, Nitrospira,Acidobacteria), which were also identified in Japanesepaddy soil [2].

Conclusion

Elucidating the short- verse long-term impact of land-usechange on soil microbial community poses special chal-lenges due to the ecological complexity of the agriculturalsystems. Current experimental evidence from differenttypes of land conversion using various microbiologicaltechnologies led to contradictory conclusions regardinghow the effect varies over time [4, 17, 27, 34]. More casestudies are thus desperately required to further our

understanding of the existence of a general trend formicrobial adaptation to land-use change, or if the trenddiffers among different agricultural systems. This reportdescribes the comparison of microbial communities in acenturies-old paddy field (OP) and two vegetable fields(NV and OV) that were established from traditional paddyfields 10 or 100 years ago. The analysis was performedwith three fields in chronosequence to ensure a compre-hensive assessment of microbial community using multiplemicrobiological techniques. Our data implicate that land-use change from rice to vegetables resulted in an increase ofabundance of culturable microorganisms but a decrease ofsoil biomass (MBC and MBN) and bacterial diversity.Change of microbial abundance occurred in the short term(<15 years) whereas the decrease of bacterial diversity isgreater following long-term rather than short-term vegeta-ble cultivation. The dynamic changes of culturablemicrobes and soil biomass were further confirmed byassessing seven additional soils in chronosequence fromthe same area. Further investigation will focus on the trendsof bacterial diversity with more soils in chronosequencewhile only utilizing a single technique (such as theARDRA).

Acknowledgments We thank Jonathan Gauntlett for critical readingof the manuscript and Hongtao Hao for help in soil sampling and soilchemical analysis and the two anonymous reviewers for feedback thathelped improve the manuscript. This work was supported by theNational Natural Science Foundation of China (40871123, 30590381)and Knowledge Innovation Program of Chinese Academy of Sciences(KZCX2-YW-407, ISSASIP0730).

References

1. Akasaka H, Izawa T, Ueki K, Ueki A (2003) Phylogeny ofnumerically abundant culturable anaerobic bacteria associatedwith degradation of rice plant residue in Japanese paddy field soil.FEMS Microbiol Ecol 43:149–161

2. Asakawa S, Kimura M (2008) Comparison of bacterial commu-nity structures at main habitats in paddy field ecosystem based onDGGE analysis. Soil Biol Biochem 40:1322–1329

3. Atlas RM (1993) Handbook of microbiological media. CRC,Boca Raton

4. Buckley DH, Schmidt TM (2001) The structure of microbialcommunities in soil and the lasting impact of cultivation. MicrobEcol 42:11–21

5. Buckley DH, Schmidt TM (2003) Diversity and dynamics ofmicrobial communities in soils from agro-ecosystems. EnvironMicrobiol 5:441–452

6. Campbell C, Grayston S, Hirst D (1997) Use of rhizospherecarbon sources in sole carbon source tests to discriminate soilmicrobial communities. J Microbiol Methods 30:33–41

7. Cao ZH, Huang JF, Zhang CS, Li AF (2004) Soil qualityevolution after land use change from paddy soil to vegetableland. Environ Geochem Health 26:97–103

8. Ding JL (2004) Origin of Neolithic paddy fields and rice farmingin the lower reaches of Yangtze River. Southeast Culture 2:19–31

484 B. Sun et al.

9. Drenovsky RE, Vo D, Graham KJ, Scow KM (2004) Soil watercontent and organic carbon availability are major determinants ofsoil microbial community composition. Microb Ecol 48:424–430

10. Edwards U, Rogall T, Blocker H, Emde M, Bottger EC (1989)Isolation and direct complete nucleotide determination of entiregenes. Characterization of a gene coding for 16S ribosomal RNA.Nucleic Acids Res 17:7843–7853

11. Fraterrigo JM, Balser TC, Turner MG (2006) Microbial commu-nity variation and its relationship with nitrogen mineralization inhistorically altered forests. Ecology 87:570–579

12. Gabor EM, Vries EJ, Janssen DB (2003) Efficient recovery ofenvironmental DNA for expression cloning by indirect extractionmethods. FEMS Microbiol Ecol 44:153–163

13. Gamito S (2010) Caution is needed when applying Margalefdiversity index. Ecol Indicators 10:550–551

14. Garland JL (1997) Analysis and interpretation of community-levelphysiological profiles in microbial ecology. FEMS Microbiol Ecol24:289–300

15. Hao H, Sun B, Zhao Z (2008) Effect of land use change frompaddy to vegetable field on the residues of organochlorinepesticides in soils. Environ Pollut 156:1046–1052

16. Konopka A, Oliver L, Turco RF Jr (1998) The use of carbonsubstrate utilization patterns in environmental and ecologicalmicrobiology. Microb Ecol 35:103–115

17. Kulmatiski A, Beard KH (2008) Decoupling plant-growth fromland-use legacies in soil microbial communities. Soil BiolBiochem 40:1059–1068

18. Li ZP, Wu XC, Chen BY (2007) Changes in transformation of soilorganic C and functional diversity of soil microbial communityunder different land uses. Agr Sci China 6:1235–1245

19. Lin XG, Yin R, Zhang HY, Huang JF, Chen RR, Cao ZH (2004)Changes of soil microbiological properties caused by land usechanging from rice–wheat rotation to vegetable cultivation.Environ Geochem Health 26:119–128

20. Margalef R (1958) Information theory in ecology. GeneralSystems 3:36–71

21. Ministry of Agriculture (1982–2008) China Agricultural Yearbook1982–2008. China Agricultural Press, Beijing

22. Nsabimana D, Haynes RJ, Wallis FM (2004) Size, activity andcatabolic diversity of the soil microbial biomass as affected byland use. Appl Soil Ecol 26:81–92

23. Page RD (1996) TreeView: an application to display phylogenetictrees on personal computers. Comput Appl Biosci 12:357–358

24. Potthoff M, Steenwerth KL, Jackson LE, Drenovsky RE, ScowKM, Joergensen RG (2006) Soil microbial community composi-tion as affected by restoration practices in California grassland.Soil Biol Biochem 38:1851–1860

25. Rui J, Peng J, Lu Y (2009) Succession of bacterial populationsduring plant residue decomposition in rice field soil. Appl EnvironMicrobiol 75:4879–4886

26. Shen W, Lin X, Gao N, Zhang H, Yin R, Shi W, Duan Z (2008)Land use intensification affects soil microbial populations,functional diversity and related suppressiveness of cucumberFusarium wilt in China’s Yangtze River Delta. Plant Soil306:117–127

27. Stark C, Condron L, O’Callaghan M, Stewart A, Di HJ (2008)Differences in soil enzyme activities, microbial communitystructure and short-term nitrogen mineralisation resulting fromfarm management history and organic matter amendments. SoilBiol Biochem 40:1352–1363

28. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, HigginsDG (1997) The CLUSTAL_X windows interface: flexiblestrategies for multiple sequence alignment aided by qualityanalysis tools. Nucleic Acids Res 25:4876–4882

29. Weber KP, Grove JA, Gehder M, Anderson WA, Legge RL (2007)Data transformations in the analysis of community-level substrateutilization data from microplates. J Microbiol Methods 69:461–469

30. Weber S, Stubner S, Conrad R (2001) Bacterial populationscolonizing and degrading rice straw in anoxic paddy soil. ApplEnviron Microbiol 67:1318–1327

31. Wu J, Joergensen RG, Pommerening B, Chaussod R, Brookes PC(1990) Measurement of soil microbial biomass C by fumigation-extraction—an automated procedure. Soil Biol Biochem 26:1167–1169

32. Wu T, Chellemi DO, Graham JH, Martin KJ, Rosskopf EN (2008)Comparison of soil bacterial communities under diverse agricul-tural land management and crop production practices. MicrobEcol 55:293–310

33. Xu GH, Zhang HY (1986) Handbook of analysis methods of soilmicrobiology. Agricultural, Beijing

34. Yao H, Bowman D, Shi W (2006) Soil microbial communitystructure and diversity in a turfgrass chronosequence: land-usechange versus turfgrass management. Appl Soil Ecol 34:209–218

35. Zak JC, Willig MR, Moorhead DL, Wildman HG (1994)Functional diversity of microbial communities: a quantitativeapproach. Soil Biol Biochem 26:1101–1108

Microbial Adaptation in Paddy and Vegetable Soils 485