APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1995, p. 3972–3976 Vol. 61, No. 110099-2240/95/$04.0010Copyright q 1995, American Society for Microbiology

Direct DNA Extraction for PCR-Mediated Assaysof Soil Organisms

TATIANA VOLOSSIOUK, E. JANE ROBB, AND ROSS N. NAZAR*

Department of Molecular Biology and Genetics, University of Guelph,Guelph, Ontario, Canada N1G 2W1

Received 6 June 1995/Accepted 12 September 1995

By using the rDNA of a plant wilt pathogen (Verticillium dahliae) as the target sequence, a direct method forthe extraction of DNA from soil samples which can be used for PCR-mediated diagnostics without a need forfurther DNA purification has been developed. The soil organisms are disrupted by grinding in liquid nitrogenwith the natural abrasives in soil, and losses due to degradation and adsorption are largely eliminated by theaddition of skim milk powder. The DNA from disrupted cells is extracted with sodium dodecyl sulfate-phenoland collected by ethanol precipitation. After suitable dilution, this DNA extract can be assayed directly by PCRamplification technologies. The method is rapid, cost efficient, and when combined with suitable internalcontrols can be applied to the detection and quantification of specific soil organisms or pathogens on alarge-scale basis.

PCR is being applied more often to the assay of microor-ganisms in the environment, including soils (for a review, seereference 17). The simplicity of this technology, together withits potential to detect small numbers of target organisms with-out a need for the culturing of cells, easily makes it an impor-tant method for monitoring pathogens and indicator bacteria.Despite this potential, technical limitations have continued tolimit the large-scale use of PCR with soil samples primarilybecause extraction techniques have been labor intensive andoften unreliable. While debate regarding the potential for ge-netic exchange in soils has continued for more than two de-cades and genetically engineered organisms are being releasedever more frequently, the quantification of genetic transfer andour knowledge of the fate of genetic materials in soils remainsurprisingly limited. Such questions underline the need to de-velop more-effective large-scale methods which can be effi-ciently applied when many samples must be evaluated.Over the last two decades, methods for the extraction of

DNA from soil samples for DNA analyses of all types havebeen markedly improved. The target organisms almost alwayshave been separated from soil samples and sometimes evencultured. For example, Faegri and colleagues (2) used differ-ential centrifugation followed by lysis of the cells, extraction ofthe nucleic acid, and purification of the DNA by hydroxyapa-tite column chromatography. Holben et al. (4) subsequentlymodified the procedure by using polyvinylpyrrolidone to re-move soil organic matter from the cell preparations and repet-itive cesium chloride density gradient centrifugation to purifythe DNA. While these approaches have been effective, theyremain very labor-intensive. Methods also have been devel-oped specifically for use with PCR amplification. For example,Pillai and coworkers (10) developed a method to separatebacterial cells by modified sucrose gradient centrifugation, butagain the approach is too labor-intensive for wide-scale appli-cation.The direct extraction of DNA from soil organisms without

prior purification or culturing clearly would provide an attrac-

tive alternative. In this instance, contaminants in the DNAextracts which inhibit PCR amplification and losses of DNA bydegradation or adsorption have proven to be the major limi-tations. Nevertheless, such approaches also have begun to beutilized with various degrees of success. For example, Picardand coworkers (9) used sonication, microwave heating, andthermal shocks to disrupt the cells in situ, but three steps ofchromatography were required to purify the DNA from con-taminants which severely inhibit the PCR. In a similar study,Smalla et al. (15) used cesium chloride for DNA purification,with even better results but again with a protocol that is rela-tively costly and labor-intensive.In our own efforts to apply PCR diagnostics on a large scale

to the detection of the Verticillium wilt pathogen in economi-cally important plants, we have exploited the advantages ofdirect extraction without a need for DNA purification (12).Tissue grinding in liquid nitrogen was combined with a simplesodium dodecyl sulfate (SDS) buffer-phenol extraction methodto provide extracts which could be subjected directly to PCRamplification. In the present study, we have adapted this ap-proach to soil samples, providing for a simple extraction pro-tocol which can be used directly with PCR amplification with-out additional DNA purification. The simplicity, speed, andlow cost of this approach make it especially attractive forlarge-scale studies.

MATERIALS AND METHODS

Organisms and plasmids. Three types of DNA were used as targets for PCRamplification: purified Verticillium dahliae genomic DNA, an internal controltemplate cloned in pTZ19R (6), and purified V. dahliae microsclerotia, kindlyprovided by G. Lazarovits (Agriculture and Agri-Food Canada). For genomicDNA, mycelia were grown without light in Czapek’s broth (19) at 228C withshaking, and the DNA was extracted by the hexadecyl trimethylammoniumbromide (CTAB) method of Rogers and Bendich (13), as previously described(6). The plasmid control template DNA was prepared as described by Holmesand Quigley (5). Both types of DNA were purified further by CsCl densitygradient centrifugation (11), and the amount of DNA was determined at A260with the assumption that 1 unit of double-stranded DNA at A260 is equivalent to50 mg/ml.Extraction of DNA from soil. Typical farm soils were taken from six diverse

regions in Canada, and sand, clay, and fine gravel were taken from a southernOntario shoreline. The optimized protocol developed in this study was based onpreviously described direct extraction methods for plant tissues containing Ver-ticillium pathogens (8). As indicated in Fig. 1, in the basic procedure, 0.25 g ofsoil sample is ground with liquid nitrogen by using a mortar and pestle for about

* Corresponding author. Phone: (519) 824-4120, ext. 3004. Fax:(519) 837-2075. Electronic mail address: RNNAZAR@UoGUELPH.CA.

3972

on April 16, 2015 by guest

http://aem.asm

.org/D

ownloaded from

5 min or until a fine powder remains. The powdered soil is suspended in 0.5 mlof skim milk powder solution (0.1 g of milk powder in 25 ml of H2O) by vigorousvortexing; for quantitative assays, internal control template DNA (usually 500pg) is also added at this time. The soil and debris are removed by centrifugationat 48C (12,000 3 g, 10 min), and the supernatant is mixed with 2 ml of SDSextraction buffer (0.3% SDS in 0.14 M NaCl, 50 mM sodium acetate [pH 5.1]) byvortexing. An equal volume of water-saturated phenol solution (16) is added; thephases are mixed by intermittent vortexing for 2 min at room temperature andthen separated by centrifugation (12,000 3 g, 10 min). The nucleic acid in theaqueous phase is precipitated with 2.5 volumes of ethanol at 2208C for severalhours or overnight when convenient. The precipitate is collected by centrifuga-tion at 48C, and the pellet is washed twice with ethanol with centrifugationbetween each rinse, and dried. The dry pellet is dissolved in 250 ml of water andstored at 2208C until it is assayed.PCR amplification of soil DNA extracts. Five microliters of DNA extract was

assayed; usually, the extract was first diluted 50-fold to reduce or avoid inhibitingsubstances. PCR amplification normally was conducted by using 50 ml of thePCR reaction mixture containing PCR buffer (normally, 50 mM KCl, 1.5 mMMgCl2, 10 mM Tris-HCl [pH 9.0], 0.1% Triton X-100), 0.1 mg of bovine serumalbumin (BSA) per ml, 0.2 mM each deoxyribonucleotide triphosphate, 12.5pmol of each V. dahliae-specific oligonucleotide primer (8), 2 U of Taq DNApolymerase (Promega Corp., Madison, Wis.), and the DNA extract. The primerswere synthesized by using a Cyclone Plus automated oligonucleotide synthesizer(Milligen/Biosearch, Milford, Mass.). The amplification was performed in aprogrammable block (Pharmacia Biotech, Uppsala, Sweden) by using 30 reactioncycles, each consisting of a 1-min denaturation step at 948C, a 1-min annealingstep at 608C, and a 2-min elongation step at 728C. For nested PCR amplifica-tions, the first amplification was carried out with a second set of oligonucleotideprimers (CTCATAACCCTTTGTGAACC and CCGAGGTCAACCGTTGCCG), with target sites external to the standardized V. dahliae-specific primerswhich are used in the second amplification phase.The products of PCR amplification were analyzed after fractionation by aga-

rose gel electrophoresis. Usually, 5 ml of the PCR reaction mixture was mixedwith 2 ml of loading dye (5% SDS, 25% glycerol, 0.025% bromophenol blue),heated to 658C for 1 to 3 min, and loaded on a 2% horizontal slab gel (7). Whenthe dye marker was approaching the bottom of the gel slab, the gel was stainedfor 40 min with ethidium bromide (0.5 mg/ml), rinsed with water (14), andvisualized with a UV transilluminator (300 nm). For quantitative measurements,

a charge-coupled device camera imaging system and Molecular Analyst/PC soft-ware (Bio-Rad Laboratories, Hercules, Calif.) were used to capture the imageand to calculate the band intensities.

RESULTS AND DISCUSSION

In this study, three potential problems were considered inthe application of PCR amplification to soil samples: DNAlosses due to degradation and adsorption as well as reaction-inhibiting contaminants. In parallel, an attempt was made tomaximize the simplicity of any extraction procedure. Becausecell or DNA purification steps are especially labor-intensive,direct extraction after cell grinding in liquid nitrogen wasadapted as a minimal method for cell disruption and DNAextraction. Previous experience had indicated that grinding inliquid nitrogen was entirely sufficient to disrupt both plant andfungal tissues (8). In the present study, the soil actually pro-vided additional abrasion in the cell disruption process and theuse of liquid nitrogen allowed cell disruption under tempera-ture conditions which minimized nucleic acid degradation.Usually the nucleic acid was extracted with SDS buffer-

phenol (16), a very common nucleic acid extraction procedurefor biochemical or genetic analyses which also had proven tobe effective with plant and fungal tissues (8). As illustrated bythe examples shown in Fig. 2 (gel A), when a target organismor an internal control template was added to soil samples suchbasic extracts often could not be adequately amplified directly(lane a), but some signal was occasionally detected even with-out dilution (lane b). In many cases the signal strength could beincreased significantly by further dilution of the extract (Fig. 2,gel B) presumably because levels of inhibiting substances arereduced and PCR amplification remains sufficiently sensitive topermit the detection of target DNA.Because more-drastic disruption methods or conditions have

previously been used for soil extracts, additional treatmentsalso were examined; they included the use of an alkaline SDSextraction buffer, often used for DNA preparations (18); vor-texing with glass beads; microwave heating or freeze-thawingto disrupt the cells (9, 15); and additional extraction withacetone or acetonitrile to remove substances which may inter-fere with PCR amplification. As illustrated in the examples

FIG. 1. Outline of direct DNA extraction protocol for soil organisms.

FIG. 2. PCR amplification of direct soil extracts. (A) Two farm soil samplesfrom different regions containing 2 pg/g of control template DNA were extractedas described in Fig. 1 without skim milk powder, PCR amplified, and fractionatedby agarose gel electrophoresis (lanes a and b). A reaction mixture containing anequivalent amount of purified template DNA is included in lane c. (B) A thirdsoil sample containing control template DNA also was extracted, and bothundiluted (lane a) and 50-fold diluted (lane b) extracts were PCR amplified andfractionated. A reaction mixture containing an equivalent amount of purifiedtemplate DNA is included in lane c as an uninhibited control reaction andmarker for the 294-bp product.

VOL. 61, 1995 DIRECT DNA EXTRACTION FOR PCR ASSAYS OF SOIL ORGANISMS 3973

on April 16, 2015 by guest

http://aem.asm

.org/D

ownloaded from

shown in Fig. 3, none of these treatments was found to bebeneficial and most actually reduced the signal or even elimi-nated it entirely. For example, an alkaline SDS buffer (gel B,lane b) resulted in much higher levels of inhibition, presumablybecause additional inhibitors were extracted under alkalineconditions, and boiling (gel A, lane b) led to much higherlosses presumably because the DNA was degraded or ad-sorbed. Whatever the mechanism, these methods were nothelpful and were not incorporated in the standardized proto-col.Because soil sample signals often remained lower than those

of equivalent DNA controls even when inhibiting substanceswere not detected, further efforts were made to eliminatelosses due to absorption or degradation. In many biochemicalstudies of various nucleic acids, adsorption and degradationare often minimized through the addition of nucleic acid car-rier or other polyvalent polymers. In hybridization analyses, forexample, Denhardt (1) addressed this problem by incorporat-ing a mixture of three carrier macromolecules: 1% BSA, 1%Ficoll (Pharmacia Biotech Inc., Uppsala, Sweden), and 1%polyvinylpyrrolidone, commonly referred to as Denhardt’s so-lution. To evaluate the possibility that such a solution or one ofthe constituents might significantly reduce or eliminate lossesdue to degradation or adsorption, the three components, bothas a complete mixture and as individual components, wereadded to the soil immediately prior to the extraction buffer. Asillustrated in Fig. 4, all were often found to significantly im-prove the signal strength, and when the PCR product yield wascompared with the yield of control reactions containing onlyequivalent amounts of target DNA (gel B, lane e), the recoverywas clearly high, with little loss of target DNA. In fact, a slightincrease in signal strength was often observed (e.g., gel B, laneb), possibly because the carriers further stabilize the Taq DNApolymerase or enhance the reaction by some other means.Because the constituents of Denhardt’s solution are rela-

tively expensive and not always readily available, a more com-mon carrier was examined, namely, skim milk powder. This

substance has also been reported to be effective as a carrier inreducing background signals and clearly would be inexpensiveand readily available. As shown in Fig. 5, with the same soilsample as used in Fig. 4, the results were again very satisfac-tory, with 0.1 g of milk powder per 25 ml of H2O being anoptimized concentration (lane b). Without milk powder virtu-ally no signal was observed (Fig. 4). Lower concentrationsoften resulted in a reduced signal strength (e.g., in lane a with0.01 g of milk powder the signal is reduced by 42%), and higherconcentrations resulted in streaking (e.g., lane c). Further-more, as shown in Fig. 6, when applied to typical farm soilsfrom six diverse regions of Canada, a signal was sometimesdetectable without dilution, but the signal strength was always

FIG. 3. Effect of additional treatments on the extraction and PCR amplifi-cation of soil DNA. (A) A liquid nitrogen-ground uninfected farm soil samplecontaining control template DNA (2 pg/g of soil) was treated further by vortexingwith glass beads (lane a), by microwave heating (lane b), or by brief boiling (3times) (lane c) before extraction and PCR amplification as described in thelegend to Fig. 2. A reaction mixture containing an equivalent amount of purifiedtemplate DNA is included in lane d. (B) A V. dahliae-infected farm soil samplecontaining control template DNA (4 mg/g of soil) was ground with liquid nitro-gen and extracted with SDS-phenol (lane a) or alkaline SDS-phenol (lane b) andPCR amplified. A reaction mixture containing an equivalent amount of purifiedtemplate DNA is included in lane d.

FIG. 4. Effect of Denhardt’s solution on extraction and PCR amplification ofsoil DNA. Denhardt’s solution (A) or the individual constituents (B) were addedto 0.25 g of liquid nitrogen-ground soil containing 0.5 mg of target DNA, and themixture was extracted with SDS-phenol as described in the legend to Fig. 2. Theextracted DNA was dissolved in 250 ml of water, and 5-ml aliquots of 50-fold-diluted extract were PCR amplified before fractionation by agarose gel electro-phoresis. (A) For the complete Denhardt’s solution, lanes a and b representextracts without and with Denhardt’s solution, respectively. (B) For the individ-ual components, lanes a to c represent extracts with 1% bovine serum albumin,1% Ficoll, or 1% polyvinylpyrrolidone, respectively. Lanes d and e representreaction mixtures with no macromolecular carrier added and with only an equiv-alent amount of purified target DNA, respectively.

FIG. 5. Effect of skim milk powder on extraction and PCR amplification ofsoil DNA. Skim milk powder solution was added to 0.25 g of liquid nitrogen-ground soil containing 0.5 mg of target DNA, and the mixture was extracted withSDS-phenol as described in Fig. 1. The extracted DNA was dissolved in 250 mlof water, and 5-ml aliquots of 50-fold-diluted extract were PCR amplified beforeagarose gel fractionation. Lanes a to c represent extractions with 0.01, 0.1, and 1g of milk powder per 25 ml of water, respectively, and lane d contains a reactionmixture with an equivalent amount of purified target DNA.

3974 VOLOSSIOUK ET AL. APPL. ENVIRON. MICROBIOL.

on April 16, 2015 by guest

http://aem.asm

.org/D

ownloaded from

strong when the extracts were diluted 50-fold prior to PCRamplification. As shown in Fig. 7, the standard protocol (gel A)was equally successful with sand and fine gravel (lanes a, c, e,and g), but only trace or no signals were observed with clay(lanes b and f). As also shown in Fig. 7, this problem could bepartially overcome with the use of higher milk powder concen-trations (gel B). Although quantitative analyses could remain aproblem with clay samples, the target DNA was detectable.Because control template DNA was used in developing the

extraction protocol, key experiments were also repeated withmicrosclerotia, a highly resistant storage form of V. dahliaewhich is commonly found in soils. As illustrated in Fig. 8, theconclusions were the same for both standard and nested PCRamplification. The genomic DNA signal was relatively weakwith the standard amplification protocol (gel A, lane a) butmuch stronger after nested PCR amplification (gel B, lane a).

As previously noted by other investigators (3), the applicationof nested PCR provides for a more dramatic level of sensitivityand permits much higher levels of dilution and diagnosticswhich should be able to detect almost any level of microbeactivity in soil samples.In summary, therefore, a rapid and cost-effective method to

extract DNA directly from soil samples which can be utilizedwith PCR amplification to effectively detect specific soil organ-isms has been developed. Many PCR-based assays for specificorganisms have already been developed and many more arecertain to follow. The extraction procedure which is defined bythis study should be applicable to many, if not all, of thesespecific assays, providing for accurate and efficient monitoringof these target organisms in soil. Extracts from samples con-taining large amounts of clay are less effective, but qualitativeanalyses are possible and the use of internal control templatesshould permit quantitative analyses as well.

ACKNOWLEDGMENTS

This work was supported by research contracts from Agriculture andAgri-Food Canada and Health Canada.

REFERENCES

1. Denhardt, D. 1966. A membrane filter technique for the detection of com-plementary DNA. Biochem. Biophys. Res. Commun. 23:641–646.

2. Faegri, A., V. L. Torsvik, and J. Goksoyr. 1977. Bacterial and fungal activitiesin soil: separation of bacteria by a rapid centrifugation technique. Soil. Biol.Biochem. 9:105–112.

3. Haqqi, T. M., G. Sarkar, C. S. David, and S. S. Sommer. 1988. Specificamplification with PCR of a refractory segment of genomic DNA. NucleicAcids Res. 16:11844–11850.

4. Holben, W. E., J. K. Jansson, B. K. Chelm, and J. M. Tiedje. 1983. DNAprobe method for the detection of specific microorganisms in the soil bac-terial community. Appl. Environ. Microbiol. 54:703–711.

5. Holmes, D. S., and M. Quigley. 1981. A rapid method of preparation ofbacterial plasmids. Anal. Biochem. 114:193–197.

6. Hu, X., R. N. Nazar, and J. Robb. 1993. Quantification of Verticillium bio-

FIG. 6. Extraction of DNA from soils of diverse origins. Target DNA wasadded to six different farm soil samples (lanes b to g) from diverse areas inCanada, and 0.25-g samples containing 0.5 mg of target DNA were extracted withSDS-phenol as described in Fig. 1. Undiluted (A) and 50-fold-diluted (B) ex-tracts were PCR amplified, and the reaction products were fractionated byagarose gel electrophoresis. Reaction mixtures with no extract (lanes a) and anequivalent aliquot of target DNA (lanes h) are included.

FIG. 7. Extraction and PCR amplification of DNA from shoreline samples.(A) Control template DNA was added to sand, clay, or gravel taken from a lakeshoreline, and 0.25-g samples containing 0.5 mg of control template DNA wereextracted with SDS-phenol as described in Fig. 1. Undiluted extracts (lanes a toc) and 50-fold-diluted (lanes e to g) extracts were PCR amplified before frac-tionation by agarose gel electrophoresis. Reaction mixtures containing equiva-lent amounts of purified target DNA are included (lanes d and h, respectively).(B) A clay sample of milk powder containing target DNA was further extractedby using 1 g/25 ml of water, and undiluted (lane b) or 50-fold-diluted (lane d)aliquots were PCR amplified. Reaction mixtures containing equivalent amountsof purified target DNA are included (lanes c and e, respectively), and a reactionmixture without extract is included (lane a).

FIG. 8. Extraction and PCR amplification of DNA from soil containing mi-crosclerotia of V. dahliae. An internal control template was added to farm soil(0.25 g) containing 1 mg of microsclerotia which was extracted as described inFig. 1 and PCR amplified by using a standard (A, lane a) or nested (B, lane b)PCR protocol before fractionation by agarose gel electrophoresis. With 30 cyclesof V. dahliae-specific amplification (lane a), a soil extract without microsclerotiaand a reaction mixture with an equivalent aliquot of control template are in-cluded (lanes b and c, respectively). By nested PCR with two 30-cycle amplifi-cations (lane a), a reaction mixture with only the second phase of V. dahliae-specific amplification and one containing an equivalent aliquot of controltemplate are included (lanes b and c, respectively). Lane d contains a reactionmixture without extract.

VOL. 61, 1995 DIRECT DNA EXTRACTION FOR PCR ASSAYS OF SOIL ORGANISMS 3975

on April 16, 2015 by guest

http://aem.asm

.org/D

ownloaded from

mass in wilt disease development. Physiol. Mol. Plant Pathol. 42:23–36.7. McDonnell, M. W., M. N. Simon, and F. W. Studier. 1977. Analysis ofrestriction fragments of T7DNA and determination of molecular weights byelectrophoresis in neutral and alkaline gels. J. Mol. Biol. 110:119–124.

8. Nazar, R. N., X. Hu, J. Schmidt, D. Culham, and E. J. Robb. 1991. Potentialuse of PCR-amplified ribosomal intergenic sequences in the detection anddifferentiation of verticillium wilt pathogens. Physiol. Mol. Plant Pathol. 39:1–11.

9. Picard, C., C. Ponsonnet, E. Paget, X. Nesme, and P. Simonet. 1992. Detec-tion and enumeration of bacteria in soil by direct DNA extraction andpolymerase chain reaction. Appl. Environ. Microbiol. 58:2717–2722.

10. Pillai, S. D., K. L. Josephenson, R. L. Bailey, C. P. Gerba, and I. L. Pepper.1991. Rapid method for processing soil samples for polymerase chain reac-tion amplification of specific gene sequences. Appl. Environ. Microbiol. 57:2283–2286.

11. Radloff, R., W. Bauer, and J. Vinograd. 1967. A dye-buoyant-density methodfor the detection and isolation of closed circular duplex DNA: the closedcircular DNA in HeLa cells. Proc. Natl. Acad. Sci. USA 57:1514–1520.

12. Robb, J., X. Hu, H. Platt, and R. N. Nazar. 1994. PCR assays for thedetection and quantification of Verticillium species in potato, p. 83–90. InM.

Dewey, R. Oliver, and A. Schots (ed.), Modern detection assays for plantpathogenic fungi. CAB International, The Netherlands.

13. Rogers, S. O., and A. J. Bendich. 1985. Extraction of DNA from milligramamounts of fresh, herbarium and mummified plant tissues. Plant Mol. Biol.5:69–76.

14. Sharpe, P. A., B. Sugden, and J. Sambrook. 1973. Detection of two restric-tion endonuclease activities in Haemophilus parainfluenzae using analyticalagarose. Biochemistry 12:3055–3058.

15. Smalla, K., N. Cresswell, L. C. Mendonca-Hagler, A. Wolters, and J. D. vanElsas. 1993. Rapid DNA extraction protocol from soil for polymerase chainreaction-mediated amplification. J. Appl. Bacteriol. 74:78–85.

16. Steele, W. S., N. Okamura, and H. Busch. 1965. Effects of thioacetamide onthe composition and biosynthesis of nucleolar and nuclear ribonucleic acid inrat liver. J. Biol. Chem. 240:1742–1749.

17. Steffan, R. J., and R. M. Atlas. 1991. Polymerase chain reaction: applicationin environmental microbiology. Annu. Rev. Microbiol. 45:137–161.

18. Steffan, R. J., J. Goksoyr, A. R. Bej, and R. M. Atlas. 1988. Recovery of DNAfrom soils and sediments. Appl. Environ. Microbiol. 54:2908–2915.

19. Tuite, J. 1969. Plant pathological methods: fungi and bacteria. Burgess Pub-lishing Co., Minneapolis, Minn.

3976 VOLOSSIOUK ET AL. APPL. ENVIRON. MICROBIOL.

on April 16, 2015 by guest

http://aem.asm

.org/D

ownloaded from