Method optimization for denaturing gradient gel electrophoresis ( DGGE ) analysis of microflora from Eucalyptus sp . wood chips intended for pulping

Eucalyptus is the predominant exotic wood species used in South African pulp and paper industry. Once chipped and stored in piles, the wood becomes vulnerable to microbial degradation and spontaneous combustion. The denaturing gradient gel electrophoresis (DGGE) technique was optimized for the detection of microbial diversity in the wood. Wood chips were collected and milled to different specifications. The 16S and 18S rRNA genes were amplified using 338F-GC/518R and 933FGC/1387R for bacteria and NS26/518R-GC and EF4F/518R for fungi. Several gel gradients were examined to determine optimal separation. A comparison of DGGE profiles revealed greater diversity in the milled wood chips amplified using primer sets of 338F-GC/518R (16S) and NS26/518R-GC (18S) with gradients of 30/60% (16S) and 25/50% (18S), respectively. Once optimized, this protocol was tested against five samples to assess its applicability to wood chip samples. Profiles were generated and amplicons excised from gels, re-amplified and sequenced to determine origin of DNA. Using this technique, 18 bacterial and 12 fungal species were identified, compared to ten bacterial and nine fungal isolates which were identified using the culturing technique and standard rRNA gene sequence analysis. The optimised DGGE is an appropriate tool for microbial community studies of Eucalyptus wood chips.


INTRODUCTION
Eucalyptus sp. is the predominant hardwood in commercial plantations for the production of pulp and paper (FAO, 2005).Within a few days following harvesting, the wood is debarked, chipped at the mill and stored in piles, and thus remains wet and may hold over 50% of their weight as water (Brown et al., 1994).
The availability of water initiates the growth of bacteria and promotes fermentation (Li et al., 2006).Storage of wood chips in piles leads to redistribution of the moisture resulting in a wet outer surface and drier inner part (Bedane et al., 2011;Noll and Jirjis, 2012) and may influence microbial succession within the chip piles based on moisture and temperature levels within the piles (Novinscak et al., 2009).Favourable conditions for bacteria are provided by the release of heat by the respiring parenchyma cells in the wood chips.Erosion and tunneling bacteria are common wood-inhabiting microorganisms and initial colonizers of wood (Clausen, 1996).Subsequent colonization by fungi is in response to the abundant wood extractives produced by bacterial degradation (Fuller, 1985).The presence of thermophilic and thermotolerant bacteria is dominant in wood chip piles as temperatures reach up to 45-50°C, thus preventing the growth of basidiomycetes (Clausen, 1996).The source of these microorganisms are different sections of the tree (leaves, bark, needles), colonization of the inner wood by tunneling bacteria and wood-rotting fungi, transportation into wood chip piles by wind and rain and microbial migration from the soil (Noll and Jirjis, 2012).
During the pulping process, if microbially degraded or physically and chemically damaged wood chips are used, they may become overcooked, leading to poor pulp quality.While physical and chemical properties of wood chips have been extensively analysed, there are limited reports of microbial community within chip piles (Noll and Jirjis, 2012).It is therefore of interest to establish both fungal and bacterial populations in freshly chipped hardwoods and monitor changes in populations over time to prevent the production of low quality pulp.Previous studies have focused either on bacteria (Kallioinen et al., 2003;Gelbrich et al., 2008;Nilsson et al., 2008), fungi (Schabereiter-Gurtner et al., 2001;Adair et al., 2002;Rajala et al., 2010;Rajala et al., 2011) or the isolation of a particular microorganism (often pathogenic) from wood (Roux et al., 2004;Kluczek-Turpeinen et al., 2007;Grobbelaar et al., 2010).Many of these studies relied on traditional culturing techniques to study microflora.Molecular methods relying on DNA or RNA extracted directly from environmental samples such as polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE), automated ribosomal intergenic spacer analysis (ARISA) and terminal restriction fragment length polymorphism (T-RFLP) are increasingly being used, particularly in analyzing community structure of microorganisms from different environments, for example, soil (Gelsomino et al., 1999;Maarit-Niemi et al., 2001;Hoshino and Matsumoto, 2007;Xue et al., 2008), compost (Adams and Frostick, 2009;Novinscak et al., 2009), wastewater (Moura et al., 2009;Liu et al., 2010;Chen et al., 2013), food (Greaves, 1975;Miller et al., 1999;Ji et al., 2004;Handschur et al., 2005) and decaying chip piles intended for fuel (Rajala et al., 2010;Rajala et al., 2011).Several recent reports focused on understanding the microbial effects on storage of woody biomass intended for biofuel production, as well as composting (Ashraf et al., 2007;Adams and Frostick, 2009;Novinscak et al., 2009).A recent review highlighted the critical factors influencing storage of woody biomass intended for biofuel production but also stated that most reports on microbial communities from wood chips and logs can be regarded as case studies rather than statistical analyses (Noll and Jirjis, 2012).The lack of data on wood chips for pulping emphasizes the need for this study.The first step in acquiring data for Eucalyptus species and examining correlations between microflora, seasons and the chemical and physical properties of wood chips and the final dissolving pulp quality, requires optimization and standardization of the PCR-DGGE technique so that shifts in populations could be attributed to wood species and season and not due to the variation in the analytical techniques.This paper reports for the first time the optimization of PCR-DGGE and preliminary application of this standardized protocol to assess its applicability in analyzing the spectrum of microbial species (fungal and bacterial) present in hardwood chips intended for pulping.

Sample collection and preparation
Five samples were collected from different areas/points of a commercial wood chip pile in an open woodyard at a pulping mill in Umkomaas located on the south coast of South Africa.Samples were collected from a pile that was largely sheltered by two larger piles located on either side (Figure 1).The pile height was approximately 2 m on the day of sampling.Five samples were collected from different areas (S1=north, S2=middle, S3=east, S4=west and S5=south) from the bottom section of the pile (0.5 m below the surface).The middle and north areas were located closest to a separating wall which provided some shelter from the elements, whilst the remaining areas were mostly exposed.Samples were collected using latex gloves and sterile zip-lock bags (22 x 34 cm) and transported back to the laboratory and temporarily stored at 4°C (~5 days) until the samples could be milled and stored at -20°C.In accordance with TAPPI T257, air-dried wood chips were ground in a Wiley-type mill to coarse sawdust.The saw dust was then passed through a 0.40 mm (40 mesh) screen.Two milled samples at sizes of >40 and <40 mm were obtained (TAPPI, 2012).

DNA extraction methods
Genomic DNA was extracted from 0.2 g milled and un-milled chips using the Soil DNA extraction kit (Zymo Research, United Sates).A modification was necessary however, as the milled wood chips absorbed the lysis buffer which had to be increased to 1200 µl per extraction.Genomic DNA was isolated from the pure bacterial and fungal isolates using the ZR fungal/bacterial DNA Kit (Zymo Research, United Sates) per manufacturer's instructions.

PCR reactions
Ribosomal genes were amplified from microbial genomic DNA from milled chips, un-milled chips and purified cultures.Universal primer sets for 16S and 18S rRNA were used and amplification conditions are listed in Table 1.Amplification reactions (50 µl) contained 1.25 mM MgCl 2 , 0.125 µM forward and reverse primers, 0.2 mM dNTPs, 0.25 U SuperTherm Taq DNA polymerase (Southern Cross Biotech, South Africa), and approximately 20-200 ng of template DNA (measured with a NanoDrop 1000 Spectrophotometer, Thermo Scientific, USA).The volume of DNA was maintained constant in order to establish a standardised method to monitor community changes over time and season in a later study.PCR was performed using the GeneAmp PCR System 9700 (Applied Biosystems, United States).The amplicons were analyzed by electrophoresis on 1% agarose gels (SeaKem, United States) in 1× Tris-Acetate EDTA running buffer at 90 V for 45 min.After electrophoresis, the gels were stained in 0.5 µg/ml ethidium bromide and visualized using a Chemi-Genius 2 BioImaging System (Syngene, United States).Following PCR, the amplicons were sequenced (Inqaba Biotech, South Africa), and the sequences edited and entered in the Basic Alignment Search Tool (BLAST) algorithm (Altschul et al., 1990) for identification of microorganisms.
Upon the confirmation of 16S and 18S amplicons, products were purified using a GeneJET™ PCR purification kit (Fermentas, Lithuania) and re-amplified in a touchdown thermal profile program using nested PCR and primers with GC-clamps (Table 1).The composition of the reaction mixture was the same as that used for the first PCR.
After electrophoresis, gels were stained in 0.5 μg/ml ethidium bromide for 60 min, destained in the same volume of 1× TAE buffer for 30 min and visualized using a Chemi-Genius 2 BioImaging System (Syngene, United States).Bacterial and fungal DNA ladders for DGGE were constructed.They consisted of known microbial species which were isolated, cultured and identified together with bands that were excised, re-amplified and sequenced from other DGGE gels.

Isolation of bacterial and fungal cultures
Five grams of wood chips were thoroughly washed by vortexing with 5 ml of phosphate buffer (pH 8.0) for 5 min.The washings were serially diluted and spread onto nutrient agar (Merck, South Africa) incubated at 37°C for 36 h (bacteria) and potato dextrose agar (PDA) (Merck, South Africa) incubated at 30°C for four to six days (fungi).Colonies were selected from the spread plates based on size, shape, pigmentation, margin, consistency and elevation and purified on appropriate agar plates.

DGGE analysis and identification of bacteria and fungi
Three parameters were optimized for DGGE analysis; gradient choice, primer and sample processing.The optimal gradients for resolution of bacterial and fungal amplicons were 30/60 and 25/50%, respectively.The DGGE profiles with greatest variety and visual clarity were produced using primer sets 338F-GC/518R and NS26/518R-GC.A greater variety of species were observed for bacteria (Figure 2) and fungi (not shown) from the milled chips compared to the fine-milled and unmilled chip samples.Bacterial and fungal DGGE community profiles of the different areas sampled are shown in Figure 3 and the identities summarized in Tables 2 and 3.The DGGE profiles clearly indicated a diversity of bacterial and fungal species.The varying intensity of the bands was an indication of varying population densities/abundance of the species in different sampling areas (Figure 3).Twenty-six and 14 distinct amplicons were visualized for bacteria and fungi, respectively, using the optimized DGGE method.At least eight bands (representing Leclercia sp., Prauserella sp., Pseudomonas stutzeri, Uncultured Klebsiella sp., Klebsiella sp., and Saccharomonospora sp.) were common in the five samples.Klebsiella pneumoniae was found only in the north (S1) and east (S3), Bacillus thuringiensis in the north (S1), middle (S2) and east (S3) areas, Inquilinus limosus in the east (S3), west (S4) and south (S5), and Pantoea sp. in the east area (S3) only (Figure 3A).
The DGGE profiles for fungi showed greater variability (Figure 3B).The greatest variability with the appearance and disappearance of bands were evident for samples from the east (S3) and middle (S2).Torrendiella eucalypti, Paecilomyces variotii, Basidiomycota sp. and Lodderomyces sp.appeared as bright bands.Basidiomycota sp. was present in samples from the north (S1), middle (S2), east (S3) and less abundant in the south (S5) but absent in the west (S4) area.Similarly, T. eucalypti were detected in the middle (S2) and east (S3) samples, but faint in the south (S5) sample.It was not possible to sequence nine bands as they were too faint and/or co-migrated.The west (S4) sample did not display any bands, which may be due to insufficient amounts of eukaryotic DNA extracted from this sample.

Identification of isolated bacteria and fungi
Ten (10) bacterial and nine fungal species were isolated using the traditional culture and identification method using   3).Bacillus spp.(33%) and A. fumigatus (29%) were the predominant bacterial and fungal species, respectively.

DISCUSSION
The aim of this investigation was to optimize the PCR-DGGE method to obtain a standardized method to assess the microbial community present in hardwood chips intended for pulping.This was achieved by testing different parameters such as sample specifications (milled), primers (16S-338f-GC/518r and 18S-NS26f/518r-GC) and gradients of DGGE gels (16S-30/60% and 18S-25/50%).
Several DNA isolation techniques were attempted.Initially, genomic DNA was extracted directly from the wood chips using the supernatant from the wood chip washings and manual methods of extraction for bacteria [modified from La Montagne et al. ( 2002)], fungi [modified from Miller et al. (1999)] and a combination of bacteria/fungi (modified from Zhou et al. [1996]).The DNA was quantified, however, when PCR was conducted, no bands were observed on a 1% agarose gel.Several adjustments to optimize extraction were made to the protocols, however, this did not have any effect on PCR amplification although satisfactory amounts of genomic DNA were obtained.It was suspected that plant phenols in the samples might have inhibited the PCR reaction, therefore a commercial kit comprising purification columns was used.
Other studies have reported optimal gradients of 18/58 and 45/60% for fungi (Rajala et al., 2010, Rajala et al., 2011) and 20/60% for bacteria (Li et al., 2011).Various 18S primer sets have been applied in other DGGE analysis studies, namely NS1-GC/NS2+10 (566 bp) (Kowalchuk et al., 1997), NS1/FR1-GC (1647 bp) (Kowalchuk et al., 1997;Li et al., 2011), FR1-GC/FF700 (700 bp) and FR1-GC/FF1100 (1100 bp) (Vainio and Hantula, 2000).Other bacterial community studies applying the DGGE technique used similar primer sets such as 968F-GC/1401R (433 bp) (Gelsomino et al., 1999).Optimum DGGE separation patterns were reported when short fragments in the range of 200 bp were applied to the gel (Muyzer et al., 1993;Muyzer and Smalla, 1998;Schabereiter-Gurtner et al., 2001).This was the basis for primer selection in this study as the 16S and 18S primers generated 237 and 316 bp amplicons, respectively.Therefore, this selection significantly influenced the determination of the gradients applied and optimal separation.Eighteen (18) bacterial and 12 fungal species were identified by sequencing of DGGE bands.Some species remained unidentified due to poor visibility and close proximity of bands leading to underreporting of the full microbial spectrum.These challenges were also encountered by other authors (Gelsomino et al., 1999;Maarit-Niemi et al., 2001;Xue et al., 2008).Sample overloading and variation of DGGE gradients may resolve this issue.Another drawback to this technique is that a simple relationship of one band representing one genus/species is not always true because a single point mutation may sometimes result in two bands.This was evident in the DGGE profile of the pure fungal and bacterial isolates.The microheterogeneity in the different rRNA-operons present in different species may be responsible for this (Muyzer and Smalla, 1998).Since DGGE detects most single-base substitutions, this may explain the variety of Klebsiella spp. that was observed at different migration points in the gel (Miller et al., 1999).The majority of fungal bands with inferred identities are known to be either thermotolerant or thermophilic, therefore, it is expected that these isolates would be found in the middle and back areas of the pile which were sheltered from cross winds, creating an environment which retains heat generated by microorganisms.Eleven ascomycetes and one basidiomycete were identified from the excised bands.The majority of endophytic fungi which dominate healthy tissue of almost all wood species are ascomycetes (Noll and Jirjis, 2012).A number of ascomycetes (soft-rot fungi) are capable of growing at higher levels of heat, moisture and pH than are tolerated by wood-decaying basidiomycetes (Lindgren and Eslyn, 1961).In a study by Rajala et al. (2011), basidiomycetes dominated decaying spruce logs, however, ascomycetes formed the majority of the inhabiting community in slightly decayed logs.Thus, the dominance of ascomycetes may be attributed to the state of the wood.The universal primer set used (ITS5F/ITS4R) may have also been inadequate in detecting the basidiomycetes due to lack of primer binding to DNA isolated from this group of microorganisms.Adair et al. (2002) utilized primer set ITS1F/NL2R which they reported as specific for the identification of basidiomycetes.ITS5F and ITS1F have binding sites in close proximity to each other whereas NL2 primer set was selected as it covered a larger area of 18S rDNA gene (ITS 1 and 2 region, 5.8S and part of the small ribosome subunit).Limitations such as primer bias and specificity have previously been reported for primer sets (Anderson and Cairney, 2004).*Accession numbers could not be assigned to excised bands that were less than 150 bp after sequencing.
The DGGE technique provides an insight to the relative abundance of microorganisms within a sample as well as variations in populations from different sections of a chip pile.One of the drawbacks to this technique was that less abundant species represented by very faint bands are difficult to excise and sequence.However, by overloading the gels with high concentrations of amplified DNA, bands may become more visible and available for excision and sequencing.Also bands may migrate very close together and make excision difficult.Optimising gradients for regions not well-separated will allow better separation.These two strategies will ensure that a greater variety and number of microorganisms may be identified compared to the culturing technique.
Culturing techniques allowed the isolation of ten bacterial and nine fungal species (Tables 2 and 3).Based on morphological differences, more isolates were selected for identification, however, sequencing data revealed a few of them to be identical to either A. fumigatus or Penicillium spinulosum.The microbial genera identified in this study have been reported in other studies on hard and softwood chips in piles and compost, including Paecilomyces sp., Phialophora sp., Penicillium sp., Curvularia sp., Streptomyces sp., Bacillus sp., Pseudomonas sp. and Micrococcus sp.(Lindgren and Eslyn, 1961;Greaves, 1975;Esyln and Davidson, 1976;de Koker et al., 2000;Ashraf et al., 2007;Adams and Frostick, 2009;Rajala et al., 2010).Basic culture media and standard temperatures were used to isolate bacterial and fungal species.If the aim was to isolate as many microbes as possible, then it is recommended that a range of enrichment media and incubation at various conditions (temperature, pH, aeration) be applied (Shields, 1969;Esyln and Davidson, 1976;Flannigan and Sagoo, 1977;de Koker et al., 2000).For enrichment and isolation of thermophilic species, Emerson YpSs Agar (Difco) or 2% malt agar and incubation in damp chambers may be used (Esyln and Davidson, 1976).Four of the fungal isolates were identified as A. fumigatus.This is expected as A. fumigatus is one of the most common fungi isolated from wood chip piles (Tansey, 1971;Greaves, 1975;Flannigan and Sagoo, 1977).Most isolates of A. fumigatus show some shade of green, but some also exhibit buff-coloured colonies (Flannigan and Sagoo, 1977).This type of variation explained the number of A. fumigatus colonies selected for purification and identification.

Conclusion
DGGE enabled identification of a greater number of isolates making it a favourable culture-independent method compared to the basic culture-dependent technique tested.Our results suggest that the optimized PCR-DGGE parameters developed in this study will be suitable for establishing and monitoring changes in microbial communities within wood chip piles in the follow-up study.This paper lays the groundwork for microbial community studies of industrial scale hardwood chip piles intended for dissolving pulp production.Analysis of lignocellulolytic enzyme production, wood and pulp chemistry, relationships between the natural microflora present and their combined effects on the final pulp yield can assist in determining the potential effect of microflora on dissolving pulp quality.

Figure 1 .
Figure 1.Schematic layout of the wood chip piles at Sappi-Chemical Cellulose and sampling points (1 st July 2010).

Table 1 .
Primers and PCR reaction conditions for amplification of 16S and 18S ribosomal genes.

Table 2 .
Bacteria identified by sequencing of 16S rRNA amplicons of cultured pure isolates and excised bands from PCR-DGGE gels.Accession numbers could not be assigned to excised bands that were less than 150 bp after sequencing. *

Table 3 .
Fungi identified by sequencing of 18S rRNA amplicons of cultured pure isolates and excised bands from PCR-DGGE gels.