Extraction of DNA from soil

Extraction of DNA from soilPatrick Robe a,*,Renaud Nalin a ,Carmela Capellano b ,Timothy M.V ogel c ,Pascal Simonet caLibraGen,11,bd Einstein,69100Villeurbanne,FrancebAventis Pharma,centre de recherche de Vitry-Alfortville,13,quai Jules Guesde B.P .14,94403Vitry sur Seine cedex,France cCenter for Microbial Ecology,UMR CNRS 5557,University Claude Bernard LYON 1,43,boulevard 11Novembre 1918,69622Villeurbanne cedex,FranceReceived 21March 2003;accepted 20May 2003AbstractThere is an increased interest in the extraction of nucleic acids from various environmental samples,since molecular techniques allow less biased access to a greater portion of uncultivable microorganisms.Two strategies have been developed to improve DNA recovery in terms of yield,purity and unbiased representation of the microbial diversity.The first approach consists of the direct extraction of nucleic acids from soil through in situ cell lysis followed by DNA purification.The alternative approach is based on the separation of bacteria from the soil particles followed by cell lysis and then DNA purification.Several published methods describe the recovery of highly purified nucleic acids that are well-suited for molecular purposes even though a new challenge concerns the recovery of large bacterial DNAs essential for functional investigation of gene clusters and biosynthetic pathways.This review presents an overview of the available methods to achieve this challenging objective.©2003Éditions scientifiques et médicales Elsevier SAS.All rights reserved.Keywords:Soils;DNA extraction;Direct lysis;Indirect extraction;Metagenome;Bacteria1.IntroductionProkaryotes are the most ubiquitous organisms on earth,represented in all habitats,including soil,sediment,marine and terrestrial subsurface,animals and plant tissues and play a key role in the biogeochemical cycles of the biosphere and represent an enormous reservoir of novel valuable molecules for health or industry.The vast majority of enzymes and antimicrobial products have been isolated from microorgan-isms cultivated on artificial media in the laboratory.Soil appears to be a major reservoir of microbial genetic diversity [11]and may be considered as a complex environ-ment.This extreme complexity results from multiple inter-acting parameters including soil texture and structure,water content,pH,climatic variations and biotic activity.Soil tex-ture and structure are mainly determined by sand,silt,clayand organic matter content through the organization of micro-and macro-aggregates.Microorganisms are heteroge-neously distributed inside microaggregates and in macro-porosities outside microaggregates [19,54].Moreover,mi-croorganisms strongly bind with soil particles (including clay)through a variety of binding mechanisms [4]that re-duce access to the whole bacteria.Microbial ecology studies have provided a better under-standing of (i)the structure of microbial communities and the evolution of these communities under various climatic,biotic and xenobiotic conditions and (ii)the activities of microor-ganisms.For several years,investigations of microbial diver-sity were mainly based on isolation and laboratory cultiva-tion of bacteria which lead to underestimation of the true boratory cultivation relied on a succession of critical steps that increase the bias:(i)soil dispersion in order to dislodge cells from inner soil compartments and from soil surface particles;(ii)centrifugation-based separation of the bacteria from soil and organic particles;(iii)plating of bac-*Corresponding author.E-mail address:pk.robe@ (P.Robe).European Journal of Soil Biology 39(2003)183–190/locate/ejsobi©2003Éditions scientifiques et médicales Elsevier SAS.All rights reserved.doi:10.1016/S1164-5563(03)00033-5terial suspension on various solid nutrient media;(iv)isola-tion and characterization of boratory cultiva-tion approaches remain of great interest when a selective enrichment targets isolates for their ability to metabolize specific substrates or to degrade deleterious xenobiotic com-pounds.However,only bacterial isolates for which optimal cultivation criteria have been determined can be recovered by this approach.The emerging use of molecular biology in microbial ecol-ogy has revealed that less than1%of the microorganisms present in numerous environments are readily cultivable[2]. Molecular analysis of bacterial diversity in complex environ-ments has relied on DNA/DNA hybridization,reassociation kinetics studies[66]and mainly on16S r DNA amplification, (for review see Ref.[74]).However,these methods require DNA extracts free of the numerous inhibitory contaminants commonly found in environmental samples(for review see Ref.[75])and must exhibit an unbiased sampling of the investigated bacterial community.The aim of this review is to present an overview of soil DNA extraction methods.Two approaches have been devel-oped for extracting nucleic acids from soil samples.One is the direct extraction of nucleic acids from soil samples after in situ cell lysis which is then followed by DNA purification [46].The second approach separates the cell fraction from soil particlesfirst.Then,this bacterial fraction is lysed and the nucleic acids purified[10,23,67].Both approaches have advantages and disadvantages related to DNA yields,DNA purity for molecular purposes,and the unbiased representa-tion of the entire microbial diversity[10,58].2.Direct lysis extraction methodThe direct in situ lysis extraction method has been widely used during the last decade.This method,which assumes complete in situ lysis of all microorganisms,generally pro-vides the highest DNA yields within acceptable processing times.All published techniques derived from the original proce-dure of Ogram et al.[46]are summarized in two critical steps.The disruption of the microbial cell wall is thefirst step leading to the release of all nucleic acids from all bacteria, theoretically independent of the cell wall sensitivity to lysis, the location of bacteria in microstructures,and their interac-tions with soil particles.In the second step,nucleic acids are separated from soil particles.2.1.Cell lysisCurrently,three types of cell lysis(or membrane disrup-tion)are used alone or in combination:(i)physical-,(ii) chemical-and(iii)enzymatic-disruption.Physical treatments,which destroy soil structure,tend to have the greatest access to the whole bacterial community, including bacteria deep within soil microaggregates.The most commonly used physical disruption methods are freez-ing–thawing or freezing–boiling[12,43,71]and bead-mill homogenization[37,41,42,63].When testing a“bead beat-ing”method,Bürgmann et al.[7]established that DNA yields increased with longer beating times,higher speed and reduced extraction buffer volumes but at the cost of DNA shearing.Other methods,which use mortar mill grinding[64], grinding under liquid nitrogen[73,77],ultrasonication [48,51]and microwave thermal shock[47]have also been reported.Frostegard et al.[17]observed that drying soil before grinding greatly improved the lysis efficiency.In general,physical methods have shown efficiencies for disruption of vegetative forms,small cells and spores[43] but they often result in significant DNA shearing[35,62].Chemical lysis either alone or in association with physical methods have been used extensively.Probably the most com-mon chemical is the detergent sodium dodecyl sulfate(SDS) which dissolves the hydrophobic material of cell mem-branes.Detergents have often been used in combination with heat-treatment and with chelating agents such as EDTA, Chelex100[21,28]and diverse Tris buffer or sodium phos-phate buffers[29].In several studies,increasing the EDTA concentration increased the strength of extraction and lysis buffers resulting in higher yields,but lower purity of isolated nucleic acids.This result leads to the conclusion that the choice of buffer is a compromise between the expected DNA quantity and the required DNA purity.In a comparative study,Miller et al.[42]reported that the Chelex100treat-ment had detrimental effect by decreasing the DNA yield.A rapid protocol was reported by Selenska and Klingmüller [60]based only on gentle shaking of soil slurry in a SDS–sodium phosphate buffer at70°C.Nucleic acids recovered from this“gentle procedure”were on average approximately 25kbp.Addition of cetyltrimethyl-ammonium bromide(CTAB) and polyvinylpolypyrrolidone(PVPP)has also been reported [29,44,77].Both CTAB and PVPP can partially remove hu-mic compounds,but(contrary to CTAB)PVPP resulted in DNA loss[77].CTAB forms insoluble complexes with dena-tured proteins,polysaccharides and cell debris[59].PVPP was shown to be ineffective during cell lysis,but efficient when used as a spin column during the nucleic acids purifi-cation step[29].Another denaturing agent,guanidine isothiocyanate has been used for mRNA extraction from seeded soils[47,71]but has[71]not proven valuable when extracting DNA from soil samples[37,42].Many protocols including an enzymatic lysis have been developed.Lysozyme treatment is one of the most common [6,37,55,64,70].Improvement in DNA purity may result from the lysozyme hydrolytic action on glycosidic or other humic components bonds.Another enzyme,achromopepti-dase,was used to improve the lysis of the recalcitrant Gram positive Frankia[61]and another,proteinase K,was used to digest contaminating proteins[37,50,77].184P.Robe et al./European Journal of Soil Biology39(2003)183–1902.2.Nucleic acid extraction and purificationSeveral methods for separating and purifying nucleic ac-ids from soil components have been investigated.A major soil component,humic acid,inhibits restriction enzyme di-gestion of DNA,the polymerase chain reaction[64]and also alters the results of quantitative membrane hybridizations by lowering the expected hybridization signal[1].The phenolic groups in humic acids denature biological molecules by bonding to amides or are oxidized to form a quinone which covalently bonds to DNA[76].In most studies,following cell lysis DNA afirst purification step of DNA is achieved by organic solvent extraction(either phenol or chloroform), followed by ethanol,isopropanol or polyethylene glycol-assisted precipitation[63].These crude extracts have been treated by numerous methods based on DNA physical and chemical properties.Nucleic acid binding on hydroxyapatite columns was successful for extracting DNA[46,63]and both DNA and rRNA[52]from soil and sediment samples.Ce-sium chloride(CsCl)density gradient centrifugation was also shown to allow further enzymatic restriction of purified nucleic acids[46,50].In contrast,Steffan et al.[63]noted that even extensive purification with CsCl gradient centrifu-gation and hydroxyapatite chromatography resulted in DNA loss and did not remove all humic acids.Since heterogeneous humic acids were dispersed throughout the CsCl gradient, some humic contaminants were coextracted with the DNA band,requiring a second CsCl gradient centrifugation to dilute these residual humic acids.Other purification proto-cols that are less time-consuming and adaptable to the pro-cessing of a higher number of samples have been used indi-vidually or in combination[18]:1.Agarose gel electrophoresis[18,43,56,77].DNA frag-ments may be recovered from gel using electroelution[34].Gel electrophoresis assisted by polyvinylpyrroli-done(PVP)was reported to facilitate separation of DNA from humic acids[24,31]since PVP retards the phenolic compounds that usually co-migrate with nucleic acids[32,76].2.Gelfiltration resins including sephadex G200[30]andG150,Sepharose2B,4B and6B,Biogel P100and P200have been evaluated in comparative studies [26,37,41].Sepharose resins demonstrated a better separation of DNA from humic acids.Microspin Sephacryl S-300and S-400columns(Pharmacia Bio-tech)were also used to purify crude DNA extracts contaminated with humic acids[15,17].3.Other commercial purification products were also in-vestigated such as Wizard DNA clean-up system (Promega)[20]and Centricon™50and Microcon™100concentrators(Amicon)[77],Elutip™D column from Schleicher&Schuell[12,17],silica-based DNA binding SpinBind Columns from FMC BioProducts[42]and Tip-100and Tip-500columns from Qiagen[25,64],respectively.Tebbe and Vahjen[64]showedthat97%of the initially coextracted humic acids could be removed through ion-exchange chromatography.Ecological and molecular PCR-based studies require ex-tensive purification of nucleic acids extracts in order to re-move humic acids and other contaminants.DNA losses dur-ing extensive purification may compromise the detection of rare DNA sequences.Jacobsen[27]successfully removed the PCR-inhibitory effect of humic acids through an alterna-tive magnetic capture hybridization(MCH)approach.MCH separates a specific DNA target from all other DNA and contaminants,including humic acids.A specific single-stranded DNA,labelled with biotin,is used to hybridize the unpurified nucleic acids of soil sample.After recovery of specific DNA–DNA hybrids by a magnetic extraction and washing,PCR amplification can be performed.Chandler et al.[8]similarly investigated affinity magnetic capture using peptide nucleic acids(PNA)clamps and oligomers as spe-cific hybridization probes.These affinity capture approaches are,therefore,particularly promising when specific DNA sequences are targeted in very small soil samples or in low biomass soil samples.2.3.Toward a universal lysis extraction method?Although numerous direct lysis extraction protocols re-sulting from complex combinations of physical,chemical, and/or enzymatic methods to purify nucleic acids have been published application has been limited to only a few samples. Therefore,the individual contribution of each step to thefinal success of DNA recovery cannot be clearly evaluated.Even when evaluation is attempted,the positive or negative effects of extraction parameters is difficult to extrapolate to other samples.However,some comparative studies have contrib-uted to better understanding the value of extraction and/or purification steps when applied to previously characterized soil samples[10,17,25,33,42,77].Unbiased nucleic acid-based access to bacterial diversity in environmental samples requires several conditions:1.All bacteria must be released from soil compartments.This includes those bacteria which are strongly ad-sorbed on soil colloids or those located within the inner microporosity of soil aggregates[54].Hattori[19]ob-served that culturable Gram-positive bacteria were more abundant in soil macropores whereas Gram-negative bacteria predominated in micropores.Physi-cal disruption methods such as bead-mill homogeniza-tion and freezing–thawing helped homogenize soil rendering confined bacteria available for lysis treat-ments[17],and thus lead to better DNA yields.2.Lysis of recalcitrant organisms such as Gram-positivecells,spores and small bacteria requires harsh treat-ments that may result in DNA shearing of lysis-sensitive bacteria.3.Nucleic acid extraction should be performed as soon aspossible after sample collection since storing samples at4°C for several weeks can result in the degradation of the large molecular-weight DNA fraction[65].185P.Robe et al./European Journal of Soil Biology39(2003)183–190Although recovery of large DNA fragments(40–90kb) using gentle lysozyme–SDS-based methods was reported [29],satisfactory DNA yields and unbiased DNA extraction require the introduction of mechanical lysis treatments which commonly result in DNA shearing.In most studies, direct extraction did not lead to the recovery of DNA frag-ments larger than20kb.Some applications,such as construc-tion of metagenomic libraries,require the extraction of large DNA fragments that are rarely obtained by direct extraction methods.3.Bacteria extraction methodBacterial extraction wasfirst reported by Faegri et al.[16] and Torsvik and Goksoyr[69]and is based on the following sequential steps:(i)dispersion of soil particles,(ii)separa-tion of the cells from soil particles by centrifugation accord-ing to sedimentation velocities,buoyant density or both,(iii) lysis of extracted cells and(iv)DNA purification[4].3.1.Soil dispersion methodsAs for the direct lysis extraction approach,dispersion of the soil sample by physical and/or chemical methods is a critical step because of the location of microorganisms within soil aggregates and because of bacterial adhesion to soil particles.Waring blender dispersion has been widely used[4,16].Other methods have been tested,such as sonica-tion[53],mild dispersal by shaking[72],and rotating pestle [36].When compared,waring blender and rotating pestle proved to be the most efficient dispersion methods for large and small soil sample volumes,respectively[36].Chemical dispersal has often been used in combination with mechanical methods[36].Cation exchange resins (Chelex100)have proven efficient for soil dispersion[28]. Bakken[3]did notfind a significant dispersion difference between hexametaphosphate and distilled water.Many other chemical agents have been evaluated:specific detergents (sodium cholate and sodium deoxycholate),which interact with bacterial lipopolysaccharides[40],polyethylene glycol (PEG)and SDS which dissolve hydrophobic material[63], PVPP which removes humic acids[63].However,chemical agents may have adverse side effects.For example,the meta-bolic state of soil bacteria as measured by cellular adenosine triphosphate levels was negatively affected when a combined SDS–Chelex chemical treatment was applied[4].Preserva-tion of bacterial integrity during cell separation seems to be essential in order to prevent released DNA to be degraded by physical,chemical and/or enzymatic processes.3.2.Centrifugation-based cell separationSeparation of bacteria from soil particles according to sedimentation velocities wasfirst described by Faegri et al.[16].This method is based on two successive centrifugations:first a low speed centrifugation ranging from500×g to 1000×g lasting2–15min,respectively,in order to remove (sediment)soil debris,fungal mycelia and heavy soil par-ticles[16].Optimal conditions for low-speed centrifugation is a compromise between cell recovery efficiency and elimi-nation of contaminating material[63].A subsequent high-speed centrifugation of the cell-containing supernatant pro-duces the bacterial fraction.Chains and clusters of bacteria and bacteria attached to soil particles might be easily lost during thefirst centrifugation.Holben et al.[23]established that approximately10%of the total bacteria present in the soil were released per round of homogenization-separation. According to these authors,the bacterial fraction released after a single round is apparently as representative of the total bacterial population as after multiple rounds.Several rounds of extraction may,however,greatly improve total cell recov-ery.Supernatants resulting from the low speed centrifugation might still contain non-cellular material and soil contami-nants such as humic acids.Flocculation of the cell debris and clay particles using CaCl2resulted in a considerable decrease in cell recovery[28].An alternative high-speed centrifuga-tion method based on density gradient centrifugation,was developed[3]in order to separate bacteria according to their buoyant density.Several multi-gradient media have been tested such as Percoll,metrizamide and Nycodenz[4]but the use of Nycodenz provided the best results.The dispersed soil suspension is deposited onto Nycodenz cushion (p=1.3g ml–1)and then centrifuged at high speed,com-monly10,000×g.Bacteria will settle on top of the Nycodenz “cushion”,and organic and mineral particles of greater den-sity will sediment to the bottom of the tube.Extraction efficiency(number of extracted bacteria as a percentage of total bacteria in the soil suspension),was between6%and 25%[39]and from20%to50%[4].The organic matter and clay content of the soil will affect the cell extraction effi-ciency[36]confirming that considerable attention needs to be paid to soil dispersion procedures(as described above). Even if most bacteria have densities below1.12g ml–1and should be recovered on the Nycodenz cushion,many bacteria are strongly attached to dense soils particles,such as clays and will sediment through the gradient.Although the bacte-rial fraction will still contain humic material and soil par-ticles having the same density as the bacteria[68],centrifu-gation on Nycodenz gradient allows recovery of a relatively clean bacterial fraction as compared to low speed centrifuga-tion methods.A similar method described by Pillai et al.[49] was based on sucrose density gradient centrifugation.3.3.Nucleic acid separation and purificationBacterial DNA can be extracted from the purified cellular fraction using numerous physical,chemical and enzymatic procedures as mentioned above.Cesium chloride–Ethidium bromide equilibrium density centrifugation has been suc-cessfully used to recover pure DNA[28,65]of large size(at least48kb)[9,23].186P.Robe et al./European Journal of Soil Biology39(2003)183–190Torsvik[67]developed a protocol in which the bacterial lysate was purified by passage through a hydroxyapatite column.A high urea concentration was used to disrupt hy-drogen bonds between DNA and humic acids.Steffan et al.[63]observed that the removal of humic acids by PVPP greatly improved the DNA purity but lowered the DNA recovery slightly.CsCl density centrifugation and hydroxya-patite column chromatographic purification improved the DNA purity,as measured by the spectrophotometric A260/280 and A260/230ratio,but both caused the loss of DNA[63].Gentle lysis methods combined with CsCl density purifi-cation recovered highly purified nucleic acids with improved sizes reaching more than100kb(Robe,unpublished data). The DNA size remains limitative for functional genomic approaches where the exploration of gene clusters and bio-synthetic pathways through cosmid and bacterial artificial chromosome(BAC)cloning requires DNA greater than 200kb.Embedding bacteria in agarose plugs before performing a gentle bacterial lysis recovered DNA fragments of a few hundred kilobases with limited mechanical shearing.This approach was used to establish libraries of BAC clones con-taining large DNA fragments from several microorganisms including Methanosarcina thermophila[14],Bacillus cereus [57],Mycobacterium tuberculosis[5]and Pseudomonas aeruginosa[13].An integrated approach combining centri-fugation-based cell separation from soil particles,in plug lysis and pulsedfield gel electrophoresis(PFGE)has been successfully applied to non-culturable bacteria from environ-mental samples[45].DNA fragments recovered using this method are more than300kbp in size and of adequate purity for further molecular cloning procedures[45].This method has already lead to the construction of a soil fosmid metage-nomic library containing more than3.5Gbp providing access to metagenomic functional genetics(Nalin and Robe,unpub-lished results).parative evaluation of direct and indirect methodsIn a comparative study between several extraction meth-ods,Leff et al.[33]established that the bead-beating-SDS method of Ogram et al.[46]allowed the greatest DNA yield but resulted in increased DNA shearing when compared to the comparatively gentle freezing-thawing lysozyme method described by Tsai and Olson[70].Compared to these two direct extraction methods,the indirect extraction method of Jacobsen and Rasmussen[28]based on centrifugation recov-ery of cell fraction before lysis and CsCl DNA purification, gave the lowest DNA yield but showed the highest DNA purity with a low degree of DNA shearing.Tien et al.[65] reported that the direct lysis methods produced similarly DNA yield at least a10-fold greater than the indirect extrac-tion methods.The main disadvantage of cell fractionation-based meth-ods is that the recovered bacterial fraction represents only 25–50%of the total endogenous bacterial community[4].In return direct extraction lysis has been assumed to recover more than60%of the total theoretical bacterial DNA[43]. Miller et al.[42]reported DNA recovery efficiencies ranging from86%to more than100%,based on direct counts of endogenous microorganisms[42].These values probably overestimate the recovery efficiencies,resulting from the eukaryotic and extracellular DNA recovery and the underes-timation of bacteria byfluorescent direct counts.Even though indirect extraction methods typically result in lower humic acid contamination than direct extraction methods,both approaches rely on subsequent purification steps in order to produce nucleic acids usable for molecular purposes.When comparing both direct and indirect methods, Steffan et al.[63]established that the degree of purity re-quired for restriction endonuclease digestion of recovered DNA varied with the enzyme used.While Pvu II did not require further hydroxyapatite purification after CsCl purifi-cation,target DNA restriction by Eco RI was only observed with DNA recovered by cell fractionation method using the most extensive CsCl-hydroxyapatite purification.Endonu-clease restriction by Sal I failed for both extraction methods. Clearly,comprehensive knowledge of the sensitivity of mo-lecular methods to inhibitory effect of contaminants coex-tracted with environmental DNA is a prerequisite when es-tablishing an extraction and purification strategy.Denaturing gradient gel electrophoresis(DGGE)was used by several authors to detect differences between extrac-tion methods[22,29,37].The DGGE band pattern was af-fected by the extraction procedure as indicated by the varia-tion of band intensity[29].Such sources of bias were also observed by Maarit Niemi et al.[37]when comparing DGGE community profiles in rhizosphere soils samples.Ribosomal intergenic spacer analysis(RISA)of soil mi-crobial diversity also revealed that both abundance and ap-parent members of the bacterial community are affected by the extraction method[38].By using hybridization of the PCR amplified16S rDNA gene Courtois et al.[10]did not find significant difference in the spectrum of diversity result-ing from the two extraction strategies,the only exception being the gamma-subclass of Proteobacteria.5.ConclusionBoth direct and indirect extraction methods can recover purified nucleic acids for molecular biology purposes.Direct extraction methods have been developed for several reasons; they are assumed to provide a less biased extraction of microbial DNA,with less laborious methods and higher nucleic acid yields.Direct lysis procedures are preferred when large quantities of nucleic acids are required for DNA-consuming methods,statistically significant detection of non-abundant microorganisms,and when the entire diversity of an environmental sample must be investigated with mini-mum bias.However,the resulting nucleic acid extracts are commonly sheared and contaminated with humic acids.Fur-187P.Robe et al./European Journal of Soil Biology39(2003)183–190thermore,extracts often contain unknown amounts of extra-cellular and/or eukaryotic DNA.Although time-consuming,indirect extraction methods are preferred for targeting prokaryotic communities,when high DNA purity is required for inhibitor-sensitive methods, and when recovery of high molecular weight DNA is neces-sary.For both direct and indirect extraction methods,several parameters may be varied to optimize soil dispersion,cell disruption,and separation and purification of nucleic acids. Evidently,there is no single ideal extraction method and the definition of an optimal strategy will rely on several ele-ments:1.A comprehensive description of soil is necessary.2.Is co-extraction of extracellular and eukaryotic nucleicacids acceptable and to which limit?3.Targeted organisms need to be defined(the whole com-munity or a specific taxon or group of taxons).4.Whether the targeted microorganisms can be easilyrecovered or whether they are confined inside inner sample aggregates or strongly adsorbed on particle surfaces needs to be considered.5.The targeted microorganisms sensitivity to lysis proce-dures needs to be determined.6.Constraints caused by non-satisfactory DNA quantityand purity will greatly influence the choice of extrac-tion and purification methods,as previously discussed.Undoubtedly,an unbiased description of microbial diver-sity depends on the optimization of nucleic acid extraction methods.In addition,considerable attention will need to be paid to other major sources of bias introduced by environ-mental sampling protocols and molecular biology methods. AcknowledgementsThis work wasfinancially supported by the Euro-pean project TRANSBAC No.QLK3-CT-2001-02242. References[1] E.W.Alm,D.Zheng,L.Raskin,The presence of humic sub-stances and DNA in RNA extracts affects hybridization results,Appl.Environ.Microbiol.66(2000)4547–4554. [2]R.I.Amann,W.Ludwig,K.H.Schleifer,Phylogenetic identi-fication and in situ detection of individual microbial cells without cultivation,Microbiol.Rev.59(1995)143–169. [3]L.R.Bakken,Separation and purification of bacteria from soil,Appl.Environ.Microbiol.49(1985)1482–1487.[4]L.R.Bakken,V.Lindahl,Recovery of bacterial cells from soil,in:J.D.Van Elsas,J.T.Trevors(Eds.),Nucleic Acids in the Environment:Methods and Applications,Springer-Verlag, Heidelberg,Germany,1995,pp.9–27.[5]R.Brosch,S.V.Gordon,A.Billault,T.Garnier,K.Eiglmeier,C.Soravito,B.G.Barrell,S.T.Cole,Use of a Mycobacteriumtuberculosis H37Rv bacterial artificial chromosome library for genome mapping,sequencing,and comparative genomics, Infect.Immun.66(1998)2221–2229.[6]K.D.Bruce,W.D.Hiorns,J.L.Hobman, A.M.Osborn,P.Strike,D.A.Ritchie,Amplification of DNA from native populations of soil bacteria by using the polymerase chain reaction,Appl.Environ.Microbiol.58(1992)3413–3416. [7]H.Bürgmann,M.Pesaro,F.Widmer,J.Zeyer,A strategy foroptimizing quality and quantity of DNA extracted from soil, J.Microbiol.Methods45(2001)7–20.[8] D.P.Chandler,J.R.Stults,S.Cebula, B.L.Schuck,D.W.Weaver,K.K.Anderson,M.Egholm,F.J.Brockman,Affinity purification of DNA and RNA from environmental samples with peptide nucleic acid clamps,Appl.Environ.Microbiol.66(2000)3438–3445.[9]S.Courtois,C.M.Cappellano,M.Ball,F.X.Francou,P.Nor-mand,G.Helynck,A.Martinez,S.J.Kolvek,J.Hopke, M.S.Osburne,P.R.August,R.Nalin,M.Guerineau,P.Jean-nin,P.Simonet,J.L.Pernodet,Recombinant environmental libraries provide access to microbial diversity for drug discov-ery from natural products,Appl.Environ.Microbiol.69 (2003)49–55.[10]S.Courtois,A.Frostegard,P.Goransson,G.Depret,P.Jean-nin,P.Simonet,Quantification of bacterial subgroups in soil: comparison of DNA extracted directly from soil or from cells previously released by density gradient centrifugation,Envi-ron.Microbiol.3(2001)431–439.[11]T.P.Curtis,W.T.Sloan,J.W.Scannell,Estimating prokaryoticdiversity and its limits,A99(2002) 10494–10499.[12]V.Degrange,R.Bardin,Detection and counting of nitrobacterpopulations in soil by PCR,Appl.Environ.Microbiol.61 (1995)2093–2098.[13]K.Dewar,L.Sabbagh,G.Cardinal,F.Veilleux,F.Sanscha-grin,B.Birren,R.C.Levesque,Pseudomonas aeruginosa PAO1bacterial artificial chromosomes:strategies for map-ping,screening,and sequencing100kb loci of the5.9Mb genome,p.Genomics3(1998)105–117.[14]S.V.Diaz-Perez, F.Alatriste-Mondragon,R.Hernandez,B.Birren,R.P.Gunsalus,Bacterial artificial chromosome(BAC)library as a tool for physical mapping of the archaeon Methanosarcina thermophila TM-1,p.Genom-ics2(1997)275–286.[15]V.P.Edgcomb,J.H.McDonald,R.Devereux,D.W.Smith,Estimation of bacterial cell numbers in humic acid-rich salt marsh sediments with probes directed to16S ribosomal DNA, Appl.Environ.Microbiol.65(1999)1516–1523.[16] A.Faegri,V.L.Torsvik,J.Goksoyr,Bacterial and fungalactivities in soil:separation of bacteria and fungi by a rapid fractionated centrifugation technique,Soil Biol.Biochem.9 (1977)105–112.[17] A.Frostegard,S.Courtois,V.Ramisse,S.Clerc,D.Bernillon,F.Le Gall,P.Jeannin,X.Nesme,P.Simonet,Quantification ofbias related to the extraction of DNA directly from soils,Appl.Environ.Microbiol.65(1999)5409–5420.[18]M.Harry,B.Gambier,E.Garnier Sillam,Soil conservation forDNA preservation for bacterial molecular studies,Eur.J.Soil Biol.36(2000)51–55.[19]T.Hattori,Soil aggregates as microhabitats of microorgan-isms,Biol.Fertil.Soils6(1988)189–203.188P.Robe et al./European Journal of Soil Biology39(2003)183–190。