An improved method compatible with metagenomic analyses to extract genomic DNA from soils in Tuber melanosporum orchards



Stéphane Uroz, UMR 1136 INRA-Université de Lorraine, Interactions Arbres Micro-organismes, 54280 Champenoux, France.




The development of high-throughput methods such as pyrosequencing and microarrays has greatly improved our understanding of the microbial diversity in complex environments such as soils. Nevertheless, albeit advancements in such techniques, the first major step is to obtain high quantity and good quality genomic DNA (gDNA). The work presented here aims to present an inherent problem with 260 : 230 nm ratio of extracted gDNA from calcareous soils of Tuber melanosporum orchards and a protocol to overcome this problem.

Methods and Results

Using two commercial gDNA extraction kits on spatially distant truffle orchards, we demonstrated that the 260 : 230 nm ratio was very low, consequentially yielding gDNA incompatible with microarray analyses. In order to solve this problem, optimization steps were tested including several wash steps performed before and/or after lysis. These washes significantly improved the gDNA quality (ratio 260 : 230 nm >1·7) without modification of the structure of the bacterial communities as stated by temporal temperature gradient gel electrophoresis analysis. A final re-extraction with phenol/chloroform was required for one of the soil samples.


A combination of wash steps included into the extraction protocol followed by phenol: chloroform re-extraction is recommended to obtain high-quality gDNA from calcareous soils of T. melanosporum orchards.

Significance and Impact of the Study

The method recommended here significantly improves gDNA quality obtained from T. melanosporum orchards to make it acceptable for highly sensitive methods such as microarray.


Advent of cheaper next-generation sequencing, methods (Neafsey and Haas 2011; Temperton and Giovannoni 2012) and advancements in microbial diagnostic microarray technologies (Streit and Schmitz 2004; He et al. 2007) have increased culture-independent studies in both numbers and resolution. As most prokaryotes remain recalcitrant to cultivation in laboratory conditions, these tools have greatly improved our understanding of microbial diversity and its functioning (Temperton and Giovannoni 2012), giving an unprecedented access to the rare phylogenetic groups, potentially not detected by the other alternative methods. Such molecular techniques require high quantity of quality genomic DNA (gDNA; Thakuria et al. 2008). In this context, the extraction of nucleic acids from soils is a central issue in microbial ecology (Robe et al. 2003). However, the extractability of high-standard gDNA is strongly dependent on soil physico-chemical properties (i.e. humic acids: Robe et al. 2003; Feinstein et al. 2009; pH or carbonate: Barton et al. 2006), plant cover (Robe et al. 2003) as well as the density and the composition of soil microbial communities (Zhou et al. 1996; Robe et al. 2003; Suz et al. 2006).

Various protocols have been developed and published to solve these potential problems (Zhou et al. 1996; Barton et al. 2006; Thakuria et al. 2008; Hu et al. 2010). However, among the terrestrial environments, calcareous soils remain the most challenging due to the high contents of very stable humic acids and polyphenol compounds. Indeed, the presence of exchangeable calcium and calcium carbonate induce the protection of humic acids and an increase in stable soil organic matter (Oades 1984; Olk et al. 1995). These specificities make difficult the extraction of high-quality gDNA compatible with the microarray requirements, which need gDNA with 260 : 230 nm ratio >1·7 to allow the downstream labelling process (Kashork et al. 2012). The ratio 260 : 230 nm ratio denotes the organic contaminants that are co-extracted during nucleic acid extractions (Yeates et al. 1998; Bürgmann et al. 2001). Such low 260 : 230 nm ratios have been encountered in the analysis of soil microbial communities inhabiting truffle orchards, which are developed on calcareous soils (Suz et al. 2006; Zampieri et al. 2012). Apart from the soil characteristics, it was reported that Tuber melanosporum produces dark melanin pigment (Harki et al. 1997; Martin et al. 2010), which could also interfere with gDNA extraction and with the downstream processes succeeding gDNA extraction, such as PCR, DNA–DNA hybridization, sequencing and microarray (Tebbe and Vahjen 1993; Price and Linge 1999; Ning et al. 2009).

The Perigord black truffle, T. melanosporum, is an ectomycorrhizal fungus, producing ascocarpic fruiting bodies (truffles) that are exquisite culinary delights (Martin et al. 2010). This fungus is known to cause a burnt zone (called ‘brulé’) where development of nonhost plants is inhibited compared to nonburnt zones (Streiblová et al. 2012). It seems that in these burnt zones, soil characteristics are modified (i.e. increase in active carbonate) (García-Montero et al. 2006). Hence, due to their economic significance and their role in nutrient cycling and tree health, researchers are engaged to better understand truffle development (Courty et al. 2010). They especially aim at identifying the interactions established by the truffle with other soil partners and at determining the relative role of these partners in the truffles’ lifecycle. Such investigations involve utilization of high-throughput technologies such as pyrosequencing and microarray.

In this study, gDNA was extracted from seven spatially distant T. melanosporum orchards with different tree species (Corylus sp. and Quercus sp.) to generate gDNA compatible with microarray (geochip) analyses. Two commercial kits, usually applied efficiently on forest soil samples rich in humic acids, were tested on soil samples collected in the burnt area in each truffle orchard. However, the gDNA extracted from each truffle orchard site reached 260 : 230 nm values <1, which is incompatible with microarray analyses. Thus, the objective of this study was to evaluate different steps of washing and purification to improve the quality of extracted gDNA. Quantity and quality of the gDNA obtained following each of these steps were assessed as well as the structure of the bacterial communities using temporal temperature gradient gel electrophoresis (TTGE). Based on our results, we further propose an optimized method for metagenomic works carried out with soils of T. melanosporum orchards.

Materials and methods

Soil samples

A total of eight experimental sites have been considered in this study. Seven are productive truffle orchards (T. melanosporum Vittad; black truffle) under oak (Quercus pubescens) and hazel tree (Corylus avellana L.) trees. All these truffle orchads correspond to open woodland where plants inoculated in nursery with T. melanosporum were implanted, except for Visan where the implanted oaks were not previously inoculated with T. melanosporum. Each sample corresponds to four soil cores (minimum of 5 g per soil core; c. 1 m distance between each soil sample) harvested in the burnt zone of one productive tree and then pooled before gDNA extraction. The acid soil of the forest experimental site of Breuil-Chenue was included as a comparative control and due to its richness in humic acids and the absence of T. melanasporum (Buée et al. 2009). The geographical locations of the different experimental sites are shown in Fig. 1. Physico-chemical properties and details of host trees are shown in Table 1. All extractions and additional purification steps mentioned in this study were carried out as triplicates.

Table 1. Soil characteristics of the different truffle orchards considered in the study. The sites are denoted according to their nearest town or village with the French department within brackets
SitesSample numbersg/kg
Water pHC/NClay (<2 μm)Silt (2–50 μm)Sand (50–2000 μm)Organic matterTotal limestonePhosphorus (P205)
Margès (Drôme)18·110·8102117781301970·009
Visan (Vaucluse)28·2812·224224651246·23510·017
Uzès (Gard)38·2110·515811872423·738·70·058
Aujols (Lot)48·269·56470366164491900·013
Sorges (Dordogne)58·1610·940535923663·71780·016
Cognac (Charente)68·239·9330536732840·73870·058
Rollainville (Vosges)77·9710·239350210554·288·50·02
Figure 1.

Map of France showing the geographical distribution of the sampling sites. The sites that are marked 1 to 7 were truffle (Tuber melanosporum, black truffle) orchards. The forest experimental site of Breuil-Chenue (c) was included as a comparative control. 1: Margès, 2: Visan, 3: Uzès, 4: Aujols, 5: Sorges, 6: Cognac, 7: Rollainville, and C: Breuil-Chenue.

Genomic DNA extraction

The gDNA from all soil samples were initially extracted using FastDNA Spin kit for soil (MP Biomedicals, Illkirch, France), according to the manufacturer's instructions. The presence of T. melanosporum in each soil was evaluated in another study by 454 pyrosequencing with fungal-specific primers (M. Claude, M. Emmanuelle, B. Marc and M. Francis, unpublished data). As the 260 : 230 nm ratios were consistently low, the same soil samples were taken further for extraction using Powersoil® DNA isolation kit (MO Bio, Carlsbad, CA, USA). When the expected quality of gDNA (260 : 230 ratio >1·7) was not reached, different alternative and sequential wash steps were tested to improve this quality. Whatever the wash step, all the gDNA were extracted from the same amount of soil (250 mg) and were eluted in the same final volume of 100 μl in molecular grade water (5 PRIME®; Gaithersburg, Deutschland).

Prelysis wash

The prelysis wash step was based on He et al. (2005). Soil samples were taken in individual 2-ml screw cap tubes. The soil samples were suspended in 1 ml of prelysis wash buffer (20 mmol l−1 EDTA; pH 7·5) at room temperature. The suspension was vortexed at speed mark 4 for 15 min (Vortex-Genie 2; Scientific industries, USA). Following this mixing step, the tubes were centrifuged 5 min at maximum (12 000 g; Eppendorf 5804R) speed. The supernatant was removed, and the soil pellet retained. The wash steps were repeated for a total of five times. The washed soils were used for gDNA extraction.

Postlysis wash

The gDNA was extracted using the Powersoil® DNA isolation kit according to the manufacturer's instruction until before the final elution step. Instead of direct elution with water or buffer, the columns were washed for 3, 5 or 10 times with postlysis wash solution (5·5 mol l−1 guanidine thiocyanate; pH 7; Luis et al. 2004). After the wash procedure, the gDNA in columns was eluted in molecular-grade water (5 PRIME®; Deutschland).

Inclusion of two wash steps

A method including both prelysis and postlysis wash steps was tested. The soil samples were washed using the prelysis wash buffer for five times prior to DNA extraction using Powersoil® DNA isolation kit. The gDNA thus extracted was washed using the postlysis wash solution for 3, 5 or 10 times prior to elution from the column.

Inclusion of an additional purification by phenol/choloform extraction

As one of the soils tested (sample 3) did not yield gDNA of required quality after including two wash steps (260 : 230 nm >1·7), extracts from this samples were further purified using the classical phenol/chloroform extraction method (Sambrook et al. 1989). The gDNA extract volumes were first increased to 200 μl (this was done to enable efficient separation of aqueous layer in the later step). This entailed mixing equal volume of phenol/chloroform/isoamylalcohol (25 : 24 : 1) with that of the extracted gDNA. The mixture was spun 5 min at 14 000 g. To the supernatant, 1/10th volume of sodium acetate (3 mol l−1, pH 5·2) was added. To the mixture, approximately 3 volumes of precooled ethanol was added and incubated 15 min on ice for before being spun 20 min at 14 000 g. The resultant pellet was rinsed with 70% ethanol and left to air-dry to dispose the residual ethanol. The resultant pellet was resuspended in molecular-grade water.

Evaluation of extraction and purification methods

The quality of the gDNA obtained was first checked by visualization under UV gel documentation system (Quantityone; Bio-Rad, Hercules, CA, USA). Quantification and estimation of gDNA purity were carried out using a NanoDrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA). A total of 2 μl was loaded to determine the concentration and the quality of gDNA. As the target was to achieve a yield of 100 ng of gDNA with a good 260 : 280 nm ratio and a 260 : 230 nm ratio >1·7, the methods that did not result in extracts of such quantity and quality were considered as failures.

TTGE community fingerprint

A GC-clamped PCR amplicons of approximately 450-bp fragment of the 16S rRNA gene covering V6 to V8 regions of the Escherichia coli SSU-rRNA sequence was amplified using GC-954f and 1369r, as described by Yu and Morrison (2004). PCR amplifications were carried out with the following conditions: 94°C for 10 min, 10 touchdown-cycles of 30 s at 94°C (denaturation) and annealing temperature starting at 57°C for 1 min and per cycle decrement of 0·5°C followed by 72°C for 90 s (extension). The reactions concluded with final extension for 10 min at 72°C and held at 4°C until use. A 30 ml of 6% (wt/vol) acrylamide gel was prepared in 1·5× TAE with 7 mol l−1 urea, 0·2% glycerol (for flexibility), 30 μl tetramethyl-ethylenediamine (TEMED) and 300 μl of 10% ammonium per sulphate (APS). After casting the gels in the sandwich plates, the wells were briefly washed with 1× TAE buffer and fixed to a DCode™ Universal Mutation Detection System (Bio-Rad Laboratories, München, Germany). PCR products were separated during the running in 1× TAE buffer (40 mmol l−1 Tris–acetate, 1 mmol l−1 EDTA, pH 8·0) at 110 volts with a temperature gradient increasing from 38 to 58°C with a temperature increment of 4°C per hour. After completion of the run, the gels were stained with SYBR Gold staining (1/10 000 final; Molecular Probes, Eugene, OR, USA) and analysed on a GelDoc transilluminator (Bio-Rad) coupled to the QuantityOne software for band pattern analysis.

The TTGE patterns were analysed as a matrix based on presence (1) or the absence (0) of each band detected. The binary matrix obtained from TTGE data was used to determine the proximity between samples extracted with the different wash steps using principal co-ordinate analyses. Multivariate analysis was carried out on the binary matrix using ADE-4 (Thioulouse et al. 1997).

Statistical evaluation

The effect of the different steps of the gDNA extraction methods tested on the ratios 260 : 280 nm, 260 : 230 nm and the yield of gDNA extracted were determined by one-factor (method) anova at a threshold level of P = 0·05 and a Bonferroni–Dunn test using XLstat2011 (Addinsoft, Paris, France).


Quantity of DNA

All extraction methods tested produced high concentration of gDNA (>1000 ng g−1 of soil), except in two wash conditions. The first one was when postlysis wash was carried out without a preceding prelysis wash step with EDTA. The other instance was when the postlysis wash was carried out for 10 times along with prelysis washes. In both cases, no gDNA was recorded on gel or by nanodrop estimation (<1 ng μl−1). The gDNA yields obtained from each soil samples by the remaining four methods are represented in Fig. 2a. No significant difference was noted between most of the extractions. The only exception was sample 2 when no prelysis or pre-elution wash steps were included. Inclusion of wash steps did not significantly affect the gDNA yields.

Figure 2.

Histograms representing the gDNA yield, quality and quantity. The values used were means of three replicates. Error bars denote standard error. (a) gDNA yields per gram of the five soil samples using different modifications to the extraction protocol. (b) the 260 : 230 nm ratios and (c) the 260 : 280 nm ratios. For all figures, statistical significance is represented by letters above each column, with different letters signifying distinct statistical groups, according to a one-way anova (P < 0·05). (The direct DNA extraction from the forest soil sample using the Powersoil DNA extraction kit gave the following results: yield: 883 ng ± 206; 260 : 280 nm ratio: 1·78 ± 0·01 and 260 : 230 nm ratio: 1·87 ± 0·05). Subsections: A: Powersoil DNA extraction kit with no modification to the manufacturer's protocol; B: kit extraction preceded by a prelysis wash step using 20 mmol l−1 EDTA solution; C and D: kit extraction preceded by prelysis wash step and included with an additional postlysis washing with 5·5 mol l−1 guanidine thiocynate (C with three postlysis washes and D with five postlysis washes). (image_n/jam12205-gra-0003.png) A; (image_n/jam12205-gra-0002.png) B; (image_n/jam12205-gra-0001.png) C and (■) D.

Extraction quality by 260 : 280 nm ratio

The purity of gDNA from protein contamination was estimated using the 260 : 280 nm ratios. Figure 2b shows histograms representing the values obtained by the four methods ranging from a minimum of 1·57 to 1·93. Sample 2 was significantly improved in quality when the wash steps were included. In samples 5 and 6, the ratios were improved as well with prelysis wash. The other samples produced gDNA with limited protein contamination even without washing. The analyses revealed that the different washing steps tested did not significantly affect the 260 : 280 nm ratio values for most of the samples tested.

Extraction quality by 260 : 230 nm ratio

The 260 : 230 nm ratios estimated the organic molecules that were co-extracted with the gDNA. This ratio was found to be low (<0·8) in all samples whatever the truffle orchards. In contrast, the gDNA from the control soil sample from the Breuil-Chenue forest experimental site, extracted using the same protocol, showed significantly higher values in this ratio (>1·8). Inclusion of wash steps gradually and significantly improved the 260 : 230 nm ratios of the calcareous samples from the truffle orchards (Fig. 2c). In samples 3, 4, 5, 6 and 7, extending the postlysis wash to fvie times did not significantly improve the quality. However, in samples 1 and 2, the postlysis wash had to be carried out for five times to achieve the desired value of 1·7. Nevertheless, although the quality of the gDNA extracted from the sample 3 (Uzès truffle orchard) increased significantly with the different optimization steps tested, it did not yield a 260 : 230 nm value of 1·7. When an additional phenol/chloroform extraction was added to the pre- and postlysis washed samples, the 260 : 230 nm ratios increased to >1·8 for this sample.

Community profile analyses

The 16S rRNA PCR products obtained for each optimization steps were compared by TTGE. Based on the numbers of detected bands, a multivariate analysis was carried out to determine the proximity of samples extracted using the different wash steps. This analysis revealed that the samples from the same truffle orchard but extracted using the different wash steps clustered together (Fig. 3).

Figure 3.

Multifactorial analysis generated from the temporal temperature gradient gel electrophoresis fingerprint analysis. In this analysis, axis 1 and 2 explain most of the variance in the data cumulatively (F1 = 32·8% and F2 = 24·1%). The analysis was based on presence and absence of each band. The clustering is marked with the source of soil samples. The smaller case alphabets denote the method used for extraction of gDNA. a: Powersoil DNA extraction kit with no modification to the manufacturer's protocol; b: kit extraction preceded by prelysis wash step and included with an additional postlysis washing with 5·5 mol l−1 guanidine for five times; and c: Combination of kit extraction, two wash steps and phenol/chloroform re-extraction.


In the past decades, several methods for isolating nucleic acids from complex environmental samples such as soils or sediments have been proposed (Robe et al. 2003). Unfortunately, no single method is found applicable for all kinds of samples (Zhou et al. 1996; Krsek and Wellington 1999; Thakuria et al. 2008). This is mostly because soils are heterogeneous and have a wide range of edaphic characteristics (pH, content of silt, clay, carbonates or presence of humic compounds), which can interfere with gDNA extraction. Apart from their direct effect of the gDNA extraction, several components of the soil matrix can also be co-extracted with the nucleic acids yielding in low-quality nucleic acids (Robe et al. 2003; Barton et al. 2006; Thakuria et al. 2008).

In the study, we encountered problems to get a good 260 : 230 nm ratio for the gDNA extracted from truffle orchards developed on calcareous soils using commercial kits. When compared with gDNA previously extracted from forest soil samples, it was seen that the problems were much reduced when soil was not calcareous and that the issue was only with truffle orchard soils. Such low-quality extracts were also encountered by Suz et al. (2006), where gDNA were extracted from nursery pots inoculated with truffle, from a wild truffle bed and from productive truffle orchards. These authors reported 260 : 230 nm ratio values ranging from 0·4 to 1. Similarly, Zampieri et al. (2012) reported the same low 260 : 230 nm ratio for gDNA extracted from T. melanosporum orchards, in and out of the burnt zone. The 230 nm spectrum denotes the presence of organic contaminants that are co-extracted with gDNA. Although it is not known what these organic compounds might be, it can be envisaged that the pigments (Harki et al. 1997; Martin et al. 2010) and the humic acid stabilized in calcareous soils (Oades 1984; Olk et al. 1995) are the contaminants. Indeed, it is widely reported that such molecules could lead to reduction in 260 : 230 nm ratios (Tebbe and Vahjen 1993). Interestingly, such problems were not encountered with soil samples collected in four white truffle Tuber magnatum orchards with host plants such as poplar, linden or oak. Considering these truffle orchards of T. magnatum, Iotti et al. (2012) obtained gDNA of good quality although all these truffle orchards were developed on calcareous soils. Altogether, these results suggest that the soil conditions (i.e. calcareous), the tree species and maybe the truffle type (i.e. white vs black truffle) can impact on the quality of the soil gDNA extracted from truffle orchards.

In order to obtain gDNA compatible with microarray analyses, we performed various optimization steps. The first optimization was based on He et al. (2005). These authors demonstrated that prelysis wash step could increase gDNA quality. In our study, this procedure based on the addition of 20 mmol l−1 EDTA did not significantly affect the yield of gDNA extracted in most of the soil samples tested but produced better 260 : 280 nm ratios, confirming the study of He et al. (2005). Notably, the 260 : 230 nm ratio also improved significantly, albeit not reaching the desired proportion (>1·7). We demonstrated that this first step was crucial for the efficiency of the final step of postlysis wash. Indeed, elimination of this step, the postlysis wash, which consisted in the addition of 5·5 mol l−1 guanidine thiocyanate, leads to the loss of gDNA. Nevertheless, when combined with the prelysis wash, it significantly improved the 260 : 230 nm ratio. The number of times this wash step was repeated also mattered. Apart from these aspects related to the quality of the gDNA extracted, the TTGE patterns generated for each soil samples treated with the different optimization steps revealed no visible impact on the taxonomic structure observed as stated on the multivariate analysis.

To conclude, our sequential analysis lead to the development of gDNA extraction protocol optimized for soil samples collected in T. melanosporum orchards (Fig. 4). We demonstrated that the inclusion of 20 mmol l−1 EDTA prelysis wash before the Powersoil® DNA isolation and five times postlysis wash with 5·5 mol l−1 guanidine thiocyanate produced gDNA compatible with microarray analyses. An extra step, based on a re-extraction with phenol/chloroform, was necessary for the soil sample 3 (Uzes) to reach a 260 : 230 nm ratio >1·7. Such protocol opens new perspectives for the functional and taxonomic analyses of the T. melanosporum-associated microbial communities, which can be applied on brûlé soils but also on ectomycorrhizosphere soil samples.

Figure 4.

Schematic summary of the optimal protocol for recovery of quality gDNA from the truffle orchards soil samples.


This work was funded by the French National Research Agency (ANR Programme Systerra: SysTruf, Project ANR-09-STRA-10) and the INRA (Institut National de la Recherche Agronomique) Center of Nancy. The authors would like to thank Bruno Chatron, Christian Tortel, Michel Tournayre, Pierre Sourzat, Pierre Fourès, Patrick Rejou, Jean Marie Doublet and Christophe Robin who allow us sampling in their truffle orchads. The UMR1136 is supported by the French Research Agency through the Laboratory of Excellence ARBRE (ANR-12-LABXARBRE-01).