OUP user menu

Assessment of the diversity, and antagonism towards Rhizoctonia solani AG3, of Pseudomonas species in soil from different agricultural regimes

Paolina Garbeva, Johannes Antonie van Veen, Jan Dirk van Elsas
DOI: http://dx.doi.org/10.1016/S0168-6496(03)00234-4 51-64 First published online: 1 January 2004


The genus Pseudomonas is one of the best-studied bacterial groups in soil, and includes numerous species of environmental interest. Pseudomonas species play key roles in soil, for instance in biological control of soil-borne plant pathogens and in bioremediation of pollutants. A polymerase chain reaction-denaturing gradient gel electrophoresis system that specifically describes the diversity of Pseudomonas spp. in soil was developed. On the basis of this molecular method as well as cultivation-based approaches, the diversity of Pseudomonas species in soil under different agricultural regimes (permanent grassland, arable land either under rotation or under monoculture of maize) was studied. Both types of approaches revealed differences in the composition of Pseudomonas populations between the treatments. Differences between the treatments were also found based on the frequency of isolation of Pseudomonas strains with antagonistic properties against the soil-borne pathogen Rhizoctonia solani AG3. Higher relative numbers of isolates either with antagonistic activity toward this pathogen or with chitinolytic activity were obtained from permanent grassland or from the short-term arable land than from the arable land. The results obtained in this study strongly indicate that agricultural regimes influence the structure of Pseudomonas populations in soil, with specific antagonistic subpopulations being stimulated in grassland as compared to arable land.

  • Pseudomonas
  • Bacterial diversity
  • Soil
  • Polymerase chain reaction-denaturing gradient gel electrophoresis
  • Antifungal activity

1 Introduction

As a result of an increasing interest in environmentally friendly agricultural practices, it has become necessary to assess how different cropping regimes affect microbial diversity in soil. In fact, our knowledge on how plant communities and their management influence microbial communities in soil is still limited [13] even though exudation from roots is known to be a key factor in these interactions [48].

One of the most important and best-studied bacterial taxa in soil is the genus Pseudomonas. This genus includes several functional groups of environmental interest, such as plant growth promoters [9], plant pathogens [10] and xenobiotic degraders [11]. Moreover, Pseudomonas species can also play important roles as biological control agents against soil-borne plant pathogens. Different mechanisms may be involved, such as the production of secondary metabolites (antibiotics, Fe-chelating siderophores), cellulolytic and chitinolytic activity, and the induction of systemic resistance against phytopathogens in the host plant [1214]. For example, antibiotic-producing Pseudomonas species have been isolated from soil that was naturally suppressive to different plant diseases, including take-all disease of wheat, black rot of tobacco and fusarium wilt [1517]. Root-associated fluorescent Pseudomonas spp. producing the antibiotic 2,4-diacetyl phloroglucinol, the key component of the specific suppression of the take-all disease agent, were shown to be enriched in numbers in take-all-suppressive soils [17,18]. This suppression was lost when these Pseudomonas spp. were eliminated. Conversely, conducive soil regained its suppressiveness to take-all disease when antibiotic-producing Pseudomonas strains were introduced. The realization that the soil microflora may be responsible for soil suppressiveness led to the idea that it should be possible to make a conducive soil suppressive by manipulating the soil microbial balance and diversity [19].

The aim of this study was to gain a better understanding of the influence of different agricultural management regimes applied at an experimental field site in The Netherlands on the diversity of Pseudomonas spp. in soil, as determined by both culture-dependent and culture-independent methods. We based our study on previous work that assessed pseudomonads in environmental settings either by cultivation [20] or by cultivation-independent methods [2124]. In particular, a newly adapted polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE) system was tested and used for the assessment of Pseudomonas diversity directly from soil. Moreover, we studied the influence of the different agricultural regimes on the prevalence of Pseudomonas antagonistic (antifungal and chitinolytic) activity against the potato pathogen Rhizoctonia solani AG3.

2 Materials and methods

2.1 Bacterial strains

The strains and isolates used in this study and their origins are listed in Tables 1 and 2. All strains were stored at −80°C in 20% glycerol.

View this table:

Bacterial strains tested with Pseudomonas-specific primers PsF and PsR

Strains and soil isolatesPCR product with PsF and PsR primers
1. Pseudomonas strains from collection at PRI
P. aeruginosa PA1+
P. aeruginosa PAO5+
P. aeruginosa PU21+
P. corrugata PD704+
P. corrugata MS1+
P. syringae Ps6+
P. cichorii Pc4+
P. chlororaphis Pc8+
P. glycinea Pg1+
P. putida Pp1+
P. putida Pp2+
P. stutzeri Ps1+
P. fluorescens Pf2+
Pseudomonas sp. PRI-PCA1+
Pseudomonas sp. PRI-DAPG1+
2. Pseudomonas soil isolates from this work
Presented in Table 2+
3. Non-Pseudomonas strains
Burkholderia cepacia LMG1222 and 18941
Ralstonia solanacearum IPO267 and IPO1609
Acinetobacter calcoaceticus DSM586
Agrobacterium radiobacter IPO-At2
Agrobacterium tumefaciens Gmi9023
Alcaligenes eutrophus IPO2
Alcaligenes faecalis a1501R
Bacillus sp. Sal1
Bacillus subtilis Bs2
Enterobacter agglomerans PS23S
Erwinia amylovora P4-17-8
Flavobacterium sp. ATCC39723
Klebsiella aerogenes 418
Listeria innocua ALM105
Mycobacterium chubuensis ATCC33609
Xanthomonas maltophilia PD1484
Streptomyces griseus ISP5236
Staphylococcus aureus 6538P
Serratia plymuthica PRI-2ca
Serratia liquefaciens PRI-K1.5
  • aPositive at low annealing temperature.

View this table:

Bacterial isolates obtained from Gould's S1 medium identified by partial 16S rRNA gene sequence

Origin (and numbers) of isolates aClosest hits to sequences in NCBI database% similarityAnt.bChit.cAntbdPCRe
Group I
G (2), A-R (4), A-M (1)P. syringae AF511511 (P. orientalis AF064457)100++
AM (1)P. syringae pv. mori AB00144699+
A-R (2)P. syringae pv. savastori AB021402 (P. chlororaphis AJ492826)99+
G (1), GA-M (31), A-M (1)P. veronii strain CA-4 AY081814 (P. brennerii AF268968)98–99+
G (2), GA-M (2), GA-R (2), A-R (6)P. rhodesiae AY043360 (P. lini AY035996)99–100+++
G (1), GA-R (1)P. tolaasii NCPPB 741 AF320992 (Pseudomonas sp. SMCC B0361 AF500621)100++
Group II
GA-M (4), GA-R (4) G (1), A-M (3)P. libaniensis AF057645 (P. marginalis AF364098)100+++
G (1), GA-M (2), GA-R (3), A-M (1)P. marginalis strain NZCX27 AF364098 (P. syringae AF511511)98–100++
Group III
G (5), GA-M (3), GA-R (2), A-R (4)Pseudomonas sp. NZ031 AY014807 (P. tolaasii AF094750)99–100+DAPG*+
G (5), GA-M (3), GA-R (2), A-M (2), A-R (1)Pseudomonas sp. Sau7 AF511510 (P. putida AB016428)100+++PRN/PCA+
G (5), GA-M (3), GA-R (6), A-M (2)Pseudomonas sp. E102 AF451270 (P. fluorescens AJ308303; P. gessardii AF074384)98–100+++PRN+
G (2), A-R (1)Pseudomonas sp. NZ065 AY014815 (Pseudomonas sp. NZ124 AY014829)100++
G (1), GA-R (2), A-R (2)Pseudomonas sp. NZ081 AF388206 (P. migulae AY047218)99++
G (4), GA-M (3), GA-R (3)Pseudomonas sp. FSL-D1-045 AF20513 (P. corrugata AF348508)99–100++PRN/PCA+
G (2), GA-R (1)Pseudomonas sp. A-07-10 AY136523 (P. fluorescens AY014829)99+++
GA-R (3), A-R (6)Pseudomonas sp 9-1 AF52165 (P. veronii AY144583)100++
Group IV
G (6), GA-M (4), GA-R (2), A-R (1)P. fluorescens LCSA0TU2 AF506042 (P. jessenii AF068259)99–100+++PRN/PCA+
G (1), A-M (2), A-R (3)P. fluorescens ATCC13525 AJ308308 (P. fluorescens AJ308303)99–100++PRN/PCA+
G (1), GA-M (2), A-M (2)P. fluorescens ATCC49642 AF094732 (P. fluorescens AF094732)99–100++
Group V
G(3), A-R (2)P. migulae AY047218 (P. putida AF094743)98–99+++
Group VI
G(1), GA-R (3), GA-M (2)P. orientalis AF064457 (P. syringae AF511511)98–99+
A-R (3), A-M (2)P. chlororaphis LMG 5004T Z76657 (P. chlororaphis AJ308301)98–100+
  • aCodes indicate from which treatment the isolates originate (see Section 2.3). Grouping is in accordance with groups defined in Fig. 6.

  • bAnt.: presence of antagonistic activity toward R. solani AG3. ++, halo >3 cm (diameter); +, halo 1–3 cm; −, halo <1 cm.

  • cChit.: presence or absence of chitinolytic activity.

  • dAntb: PCR-based detection of antibiotic synthesis genes. *DAPG was detected in only two isolates (G2, G3).

  • ePCR: Positive PCR signal obtained with Pseudomonas-specific primers PsF and PsR.

2.2 Fungal pathogen

R. solani AG3 (basidiomycete with a chitin/glucan-containing cell wall) was originally isolated from potato plants exhibiting symptoms of potato rot. The culture was kept on potato dextrose agar (PDA; Oxoid, Hampshire, UK) at 4°C and subcultured every second month.

2.3 Field treatments, soil and soil sampling

Soil samples were collected from a long-term ecological site at the Wildekamp field, located in Bennekom, The Netherlands. The soil in this field is a loamy sand rich in organic matter (2.5%) with slightly acidic pH (5.5–6.5). The site has been under permanent grassland (treatment G) for approximately 50 years, but part of it was turned into agricultural land about 20 years ago (long-term arable land – A). The long-term arable land was divided into arable land under common agricultural rotation (A-R; including oats, maize, barley and potato) and arable land under continuous maize (A-M). The site also contained short-term arable land (GA), which consisted of plots turned from permanent grassland into arable land in 2000. This short-term arable land was either kept under maize monoculture (GA-M), or placed under the same rotation as in the A-R plots (GA-R). In the year 2000, oats were grown in the arable land under rotation, followed by maize in 2001. In the first year, samples were taken from triplicate (10×10-m) plots per treatment at four points in time (February, May, September and November). Two treatments, A-R and G, were studied. In the second year, samples were taken at the end of the growing season, i.e. in September, from five different treatments (A-R, A-M, GA-R, GA-M and G), in triplicate.

Sampling from all plots proceeded as follows. About 100 (3–5-g) samples from the soil top layer (0–10 cm) per plot were taken randomly throughout that plot, using a sterilized auger (2 cm diameter) and mixed thoroughly in a plastic bag, to yield one composite sample per plot. The soil samples were used for analysis within 24 h after sampling. Rhizospheres of grass (G plots), oats and maize (A-R, GA-R, A-M and GA-M plots) were also sampled and processed as described previously [19].

2.4 Bacterial isolations, media and growth conditions

For isolation of bacterial cells, 10 g of soil was suspended in 95 ml of 0.1% tetrasodium pyrophosphate (Na4P2O7·12H2O, Merck) containing 10 g common aquarium gravel (2–4 mm diameter), and shaken for 25 min at 250 rpm. One ml of this soil suspension was used to prepare serial 10-fold dilutions in 0.8% NaCl. For the enumeration of Pseudomonas spp., 100 μl from the 10−2 and 10−3 dilutions were used for plating on Gould's S1 agar [20]. Plates were incubated at 27°C and enumerations of total as well as fluorescent (under UV light) Pseudomonas colonies were done after 48 h of incubation. Results are presented as the log numbers of CFU per g (dry weight) soil.

2.5 Screening of Pseudomonas isolates for antagonistic activity toward R. solani AG3

Antagonistic activity of 500 Pseudomonas isolates against R. solani AG3 was tested by dual culturing on 0.1×PDA. Using sterile toothpicks, four Pseudomonas sp. isolates were placed approximately 3 mm from the edges of the 0.1×PDA plates and a 6-mm agar disk containing grown fungal mycelium was placed in the center of each plate. The plates were analyzed after 7 days of incubation at 25°C, by measuring the extension of mycelia and zones of inhibition (haloes without mycelial growth) around the isolates. A strain with known antagonistic activity (P. syringae AM20) served as the positive control (halo >3 cm), whereas a strain without that activity (Agrobacterium radiobacter IPO-At2, no halo) was the negative control. Isolates that formed halo zones over 2 mm were accepted as antagonists to R. solani AG3. The scoring was performed as follows: halo ≥3 cm (++); halo 1–3 cm (+), and halo <1 cm or no halo (−).

2.6 Screening of Pseudomonas isolates for chitinolytic activity

Chitinase activity was measured by modification of the Schales procedure with colloidal chitin as an assay substrate [25]. Colonies were screened for chitinolytic activity by plating on two media: CY (chitin yeast extract) medium containing 2.5 g NaCl, 0.02 g MgSO4·7H2O, 0.02 g CaCl2·2H2O, 0.05 g yeast extract, 50 mg Delvocid (contains 50% natamycin; DSM, Delft, The Netherlands), 40 ml chitin stock (2.5%) and 10 g agar, and CTSB (chitin–tryptone soy broth) medium containing 2.5 g NaCl, 1.5 g TSB (tryptone soy broth), 50 mg Delvocid and 40 ml chitin stock (2.5%) plus 10 g agar. Clearance haloes indicating the enzymatic degradation of chitin were measured after 7 days of incubation at 27°C.

2.7 Soil DNA extraction

For soil DNA extraction, a cell homogenizer (Bead beater; B. Braun, Melsungen, Germany) was used prior to extraction with the MoBio Ultraclean soil DNA extraction kit (Biozym TC, Landgraaf, The Netherlands). Glass beads (1.5 g, 0.11 mm diameter) were added to 0.25-g soil samples in 0.5 ml buffer and the mixture was bead-beaten four times for 90 s each time. After bead beating, DNA extraction proceeded in accordance with the protocol furnished by the manufacturer (MoBio extraction protocol). A final DNA purification step was performed using the Wizard DNA cleanup kit (Promega, Leiden, The Netherlands). DNA purity and quality were assessed after electrophoresis of subsamples in 0.8% agarose gels in 0.5×TBE buffer [26]. DNA quality was checked by electrophoresis in 0.8% agarose gels and DNA was quantified by comparison to a standard (1-kb ladder, Invitrogen, Cat. no. 15615-024). DNA yields were, on average, about 25 μg g−1 soil.

2.8 Pure culture DNA extraction

DNA extraction from pure cultures (Tables 1 and 2) grown in 0.1×TSB was commenced by harvesting cells from 1.5-ml overnight cultures into 1 ml of 0.8% NaCl. Lysis was performed by bead beating (1 g of 0.11 mm diameter beads in 1 ml, four times 90 s). The lysate was extracted with phenol–Tris–HCl (pH 8.0) and chloroform/isoamyl alcohol (24:1) followed by precipitation with 96% ethanol in the presence of 5 M NaCl [26]. The DNA pellets were washed with 70% ethanol, vacuum-dried and dissolved in 50 μl sterile Milli-Q water.

2.9 PCR amplification of 16S rRNA gene fragments

A semi-nested system was used for amplification of the V6/V7 region of the 16S ribosomal RNA gene. The first PCR reaction was performed by applying the Pseudomonas-specific primers PsF and PsR (Escherichia coli positions 298 and 1258, respectively; Table 3) described by Widmer et al. [21]. PCR was performed in an MJ Research PT-200 thermal cycler in 50-μl reaction volumes containing 0.2 μM of each primer, 3.75 mM MgCl2 (Perkin-Elmer, Nieuwersluis, The Netherlands), 200 μM of each dNTP (Boehringer, Almere, The Netherlands) and 0.25 μg T4 gene 32 protein (Boehringer, Mannheim, Germany) using 5 U AmpliTaq Stoffel fragment in 1×Stoffel buffer. The thermal cycling was as follows: denaturation at 94°C for 5 min, followed by 35 cycles of 94°C for 1 min, 66 or 68°C (for soil DNA and pure culture, respectively) for 1 min, and 72°C for 2 min, and a final extension step at 72°C for 10 min. The PCR products (expected sizes 760 bp) were analyzed by running 5-μl aliquots of the reaction mixtures in 1.2% agarose gels. The remaining 45 μl of the PCR volumes were precipitated with 1/10 volume of 5 M NaCl and ice-cold 96% ethanol for 15 min at −20°C. After centrifugation, washing with 70% ethanol and air-drying, the pellets were resuspended in 100 μl sterile Milli-Q water.

View this table:

Primers used in this study

PrimerSequenceAnnealing temperature (°C)PCR product size (bp)Detection ofReference
PsRGGTCTGAGAGGATGATCAGT66960Pseudomonas spp.[21]
PsFTTAGCTCCACCTCGCGGC66960Pseudomonas spp.[21]
PRND1GGGGCGGGCCGTGGTGATGGA67786Pyrrolnitrin production locus (PRN)[28]
Phl2aGAGGACGTCGAAGACCACCA677452,4-Diacetylphloroglucinol production locus (DAPG)[18]
PCA1aTGCCAAGCCTCGCTCCAAC671150Phenazine 2-carboxylic acid production locus (PCA)[18]

The precipitated PCR products served as templates for a second PCR with conserved bacterial forward primer F968 with attached GC clamp (F968-GC [27]), and the Pseudomonas-specific primer PsR. The program used for the second PCR was as follows: initial denaturation at 94°C for 4 min, one cycle of 1 min at 94°C, 1 min at 60°C, and 2 min at 72°C, followed by 10 times the same cycle with every subsequent one using a 0.5°C lower annealing temperature (until 55°C), 20 cycles of 94°C (1 min), 55°C (1 min) and 72°C (2 min), and final extension at 72°C (10 min). The PCR products (expected sizes 290 bp) were first analyzed by running 5–10-μl aliquots of the reaction mixtures in 1.2% agarose gels and secondly by running 15 μl on 45%–65% DGGE gels.

For comparison, selected soil-derived DNA samples and Pseudomonas pure cultures were also analyzed by applying a PCR-DGGE system recently described by Gyamfi et al. [23].

2.10 DGGE analysis

DGGE was performed using 6% polyacrylamide gels (ratio of acrylamide to bis-acrylamide 37:1) with a gradient of 45–65% denaturant (100% denaturant was defined as 7 M urea plus 40% formamide). The gels were electrophoresed at 60°C at 100 V for 15 h in a PhorU2 apparatus (Ingeny, Goes, The Netherlands) and stained with SYBR gold (Molecular Probes, Leiden, The Netherlands). For analysis of the molecular community profiles, the Molecular Analyst Fingerprinting software (version 1.61, Bio-Rad, Veenendaal, The Netherlands) was used. Clustering was determined by the unweighted pair group with mathematical averages (UPGMA) method.

2.11 Cloning and sequencing of amplicons

For direct cloning of the PCR products obtained with primers F968 and PsR, the pGEM-T Easy vector system (Promega, Leiden, The Netherlands) was used. Prior to cloning, PCR products were purified with High-Pure PCR product purification kit (Boehringer Mannheim, Almere, The Netherlands) and cloned into the pGEM-T vector according to the manufacturer's instructions. Plasmid extraction was performed using the Wizard Plus SV miniprep DNA purification kit (Promega Benelux). Clones with the correct insert (as judged by size) were subjected to (one strand) sequencing with universal M13 primers using the services of BaseClear, The Netherlands (http://www.baseclear.nl).

For sequencing of DGGE bands, bands were first excised from the gel, re-amplified and analyzed on DGGE for purity and correct migration behavior, after which PCR products were sequenced. Two clones per band were sequenced, which gave identical sequences throughout.

2.12 PCR with primers for genes encoding the synthesis of antibiotics

Three sets of primers received from Dr. J. Raaijmakers (Phytopathology Department, Wageningen University, The Netherlands) were used for the detection of genes encoding the production of pyrrolnitrin (PRN) [28], 2,4-diacetylphloroglucinol (DAPG) [18] and phenazine-carboxylic acid (PCA) [18] in isolates. Primer sequences, annealing temperatures and references are listed in Table 3. Approximately 5–10 ng of genomic DNA was used per strain. The PCR products were analyzed by running 25-μl aliquots of the reaction mixtures in 1.2% agarose gels and by hybridization under high-stringency conditions with specific probes prepared by PCR on the basis of the reference strains Burkholderia cepacia LMG 1222 PRN (for PRN), Pseudomonas spp. PRI-DAPG1 (for DAPG) and Pseudomonas sp. PRI-PCA1 (for PCA). Probe preparation, hybridization and signal detection were performed as described by the PCR DIG Probe Synthesis Kit (Roche, Cat. no. 1636090).

2.13 DNA sequence analysis

The partial 16S rDNA sequences (length about 290 bp) obtained from 175 isolates and 25 DGGE bands were compared against those available in the database using BLAST-N provided by the Plant Research International server (http://lx10003.plant.dlo.nl), NCBI (http://www.ncbi.nlm.nih.gov) or the Ribosomal Database Project (http://rdp.cme.msu.edu/cgis/seq_match.cgi).

The sequences were further aligned and clustered using ClustalW provided by the Institut Pasteur (http://bioweb.pasteur.fr). A phylogenetic tree was constructed from these aligned sequences by neighbor joining, using Treecon (version 1.3b, Yves van de Peer, Ghent, Belgium). Bootstrapping was performed using the bootstrap modus of the program and values above 50% are reported. Sequences obtained were deposited in GenBank under numbers AY365075 to AY365106.

2.14 Statistics

For each treatment, samples were obtained from three replicate plots per treatment in the completely randomized block design used in the field. Bacterial counts (cfu g−1 soil) were analyzed after logarithmic transformation. The data were statistically analyzed using Genstat 5, Release 4.2 (Rothamsted Exp. Stat., UK). Data were considered to be significantly different at P<0.05.

3 Results

3.1 Number of total and fluorescent Pseudomonas spp. on Gould's S1 agar

The numbers of total and fluorescent Pseudomonas spp. recovered on Gould's S1 agar in the first sampling year are presented in Table 4, A. The total counts from bulk soil did not show large variations over the season or between the treatments (arable land versus grassland), ranging from log 5.5 to log 6.1 cfu g−1 of dry soil. On the other hand, the numbers of fluorescent pseudomonads increased significantly over the growing season in all treatments (P<0.05), ranging from log 4.1 (A-R, bulk) to log 5.5 (G, rhizosphere) cfu g−1 of dry soil. Both counts were generally higher, albeit not significantly, in the bulk soil of the grassland than in that of the arable land. In the November samples, higher numbers of total and fluorescent Pseudomonas spp. were detected in the rhizospheres than in the corresponding bulk soils. However, this was only significant (P<0.05) in the A samples.

View this table:

Pseudomonas populations (log CFU g−1 of dry soil) in bulk and rhizosphere soils under different management regimes

Soil treatmentaFebruaryMaySeptemberNovember
A: First sampling year
A-R, Bulk5.6 (0.2)5.5 (0.2)4.1 (0.1)5.5 (0.3)4.3 (0.1)5.4 (0.2)5.0 (0.4)
A-R, Rhizb6.1 (0.3)5.5 (0.2)
G, Bulk6.1 (0.1)5.8 (0.3)4.6 (0.2)5.9 (0.1)4.7 (0.2)6.1 (0.2)5.4 (0.1)
G, Rhiz6.0 (0.1)5.6 (0.1)4.8 (0.2)5.9 (0.3)4.8 (0.2)6.3 (0.1)5.8 (0.2)
B: Second sampling year
G, Bulk6.0 (0.3)5.0 (0.1)
G, Rhiz6.2 (0.1)5.7 (0.5)
GA-R, Bulk5.8 (0.2)4.6 (0.4)
GA-R, Rhizc5.9 (0.2)5.2 (0.2)
GA-M, Bulk5.9 (0.2)4.8 (0.1)
GA-M, Rhizc5.9 (0.3)5.3 (0.2)
A-R, Bulk5.1 (0.1)4.1 (0.3)
A-R, Rhizc5.3 (0.3)4.7 (0.2)
A-M, Bulk5.2 (0.2)4.2 (0.3)
A-M, Rhizc5.4 (0.2)4.4 (0.3)
  • a=number of total cfu; b=number of fluorescent Pseudomonas cfu. The numbers in parentheses represent the standard deviation. Total cfu: G, GA-R, GA-M>A-R, A-M (P<0.05). Fluorescent cfu: Bulk: G, GA>A (P<0.05). Rhiz: G, GA>A (not significant).

  • aFor soil treatments see see Section 2.3.

  • bRhizosphere of oat.

  • cRhizosphere of maize.

The results obtained during the second sampling year (one sampling) are presented in Table 4, B. Again, differences in Pseudomonas numbers between the treatments were observed. Significantly higher numbers of total and fluorescent pseudomonads (P<0.05) were measured in the permanent grassland (log 6.0 and log 5.0 cfu g−1 of dry soil) and short-term arable land (log 5.8 and log 4.6 cfu g−1 of dry soil) than in the long-term arable land (log 5.1 and log 4.1 cfu g−1 of dry soil). In all treatments, rhizosphere effects were observed. In particular, the numbers of fluorescent pseudomonads were significantly higher in the rhizosphere than in the bulk soil samples (P<0.05).

3.2 Screening for Pseudomonas isolates with antagonistic activity toward R. solani AG3, chitinolytic and/or antibiotic production capacity

In total, 500 Pseudomonas isolates originating from five different treatments (G, A-R, A-M, GA-M, GA-R; 100 isolates per treatment) were screened in an in vitro assay for their ability to suppress R. solani AG3. Approximately 17.4% (87 of 500) of all isolates showed this antagonistic capacity. The highest isolation frequencies were found in the short-term arable land, i.e. GA-M (25±3%) and GA-R (22±2%), which was followed by permanent grassland (19±4%) and long-term arable land, i.e. A-M (15±2%) and A-R (6±1%). These isolation frequencies were statistically similar between the samples derived from grassland (G, GA), whereas those from (long-term) arable land were lower, albeit only significantly for A-R (P<0.05). All 87 isolates that showed antagonistic activity were identified by partial 16S rRNA gene sequencing (Table 2). The most frequently found antagonists had 16S rRNA gene sequences that were highly (98–100% similarity) related to sequences of Pseudomonas sp. SaU7, P. fluorescens LCSA0TU2, P. migulae, Pseudomonas sp. E102, P. rhodesiae, P. syringae and P. libaniensis (Table 2).

To assess their potential action on the R. solani AG3 cell wall by chitinolysis, all 500 Pseudomonas isolates were also screened for chitinolytic activity. Chitinolysis was detected in 79 strains, or 15.8% of the total. Several, but not all, of the chitinolytic isolates also had shown anti-AG3 activity. Chitinolytic isolates were most prevalent in short-term arable land, i.e. GA-M (22±1%) and GA-R (20±3%), and permanent grassland (17±2%). In the long-term arable land, lower percentages (A-M: 14±4%; A-R: 6±2%) of the tested isolates were chitinolytic. Thus, the isolation frequencies of chitinolytic strains were significantly higher in the grassland-derived samples (G, GA) than in those from arable land (P<0.05), except for those from A-M. All 79 chitinolytic Pseudomonas spp. were identified by partial sequencing of their 16S rRNA genes (Table 2). The data showed that the isolates were distributed, at high (98–100%) similarity values, among a range of closest database hits, notably with (numbers of isolates in parentheses): Pseudomonas sp. SaU7 (13), Pseudomonas sp. E102 (15), Pseudomonas sp. FSL-D1-045 (10), Pseudomonas sp. A-07-10 (3), P. fluorescens LCSA0TU2 (9), P. fluorescens ATCC13525 (3), P. rhodesiae (7), P. libaniensis (10), P. marginalis (7) and P. migulae (2).

A selection (210) of the 500 Pseudomonas isolates was screened for evidence of the presence of the PRN, DAPG and PCA synthetic operons by PCR followed by hybridization with the appropriate probes. About 18% (38/210) of the tested isolates showed positive signals with the PRN primers, 1% (2/210) with the DAPG primers, and 12% (25/210) with the PCA primers (Table 2). Some isolates of the latter group also showed a positive signal with the PRN detection system.

3.3 Pseudomonas-specific PCR-DGGE analysis

3.3.1 Validation of the PCR system

Specific PCR amplification of Pseudomonas 16S rDNA genes directly from soil samples was performed by applying the highly selective PCR system described by Widmer et al. [21]. First, the specificity of the PCR system was re-checked with a range of Pseudomonas and non-Pseudomonas strains (Table 1). All Pseudomonas strains used produced a PCR product with this system. Among the 22 non-Pseudomonas strains tested, including representatives of Burkholderia sp., Ralstonia solanacearum, Agrobacterium sp., Alcaligenes sp., Bacillus sp. and Xanthomonas sp., only one non-Pseudomonas strain (identified as Serratia plymuthica) showed a positive PCR signal. This aspecific PCR amplification was avoided when the annealing temperature was increased, from 66 to 68°C. The Pseudomonas-specific PCR products were used as targets for a second, semi-nested, PCR performed with the forward bacterial primer F968 (with a GC-clamp) and the Pseudomonas-specific primer PsR. The amplified 290-bp products were successfully separated on 45–65% DGGE gels, yielding separated bands from different selected Pseudomonas strains (Fig. 1A, see below).


Evaluation of two Pseudomonas-specific PCR-DGGE systems using as targets pure culture and soil DNA. A: Pure cultures. Lanes 1–5: amplified with the primers located in the V6/V7 region of the 16S rRNA gene; 6–10: amplified with the primers located in the V1–V3 region of the 16S rRNA gene. Lanes 1, 6: Pseudomonas sp. E102; 2, 7: P. libaniensis; 3, 8: Pseudomonas sp. SaU7; 4, 9: P. fluorescens LCSA0UT2; 5, 10: Pseudomonas sp. Fa8. B: Five soil DNA samples. Lanes 1–5: soil DNA amplified with primers located in the V6/V7 region of the 16S rRNA gene; 6–10: same samples amplified with primers located in the V1–V3 region of the 16S rRNA gene. M: marker (from top to bottom): amplicons of Enterobacter cloacae BE1, Listeria innocua ALM105, Rhizobium leguminosarum bv. trifolii R62, Arthrobacter sp., Burkholderia cepacia P2.

The detection limit of the Pseudomonas-specific PCR, assessed by performing PCR with 10-fold dilutions of Pseudomonas sp. DNA (containing from 1 ng to 1 fg DNA) was 100 fg DNA, estimated to represent 15–20 genome equivalents per PCR. Inhibition of the PCR reaction by soil DNA was not observed, as evidenced by performing PCR with a mixed template containing 10 ng of soil DNA and 100 fg of Pseudomonas sp. DNA (not shown).

3.3.2 Evaluation of two Pseudomonas-specific PCR-DGGE systems

We compared the PCR-DGGE system used in this study, which was based on amplification of the highly variable V6 and V7 regions of 16S rDNA, with the one recently published by Gyamfi et al. based on the V1–V3 16S rDNA regions [23]. The results of the DGGE analysis based on amplification of both regions obtained with pure cultures as well as with soil DNA by both systems are presented in Fig. 1A,B.

First, when pure culture DNA of different pseudomonads was used, a better separation by DGGE was observed with the system amplifying the V6/V7 regions of the 16S rDNA than with that based on the V1–V3 region (Fig. 1A). In fact, some Pseudomonas isolates, which were differentiated by DGGE of the V6/V7 region products, were not distinguished by DGGE of those of the V1–V3 region. Specifically, isolates identified as P. libaniensis and Pseudomonas sp. E102 and those related to Pseudomonas sp. SaU7 and P. fluorescens LCSA0TU2 gave bands at the same position when analyzed by V1–V3 region-based DGGE, whereas they were separated by the DGGE system based on the V6/V7 regions (Fig. 1A).

In addition, the DGGE patterns obtained from soil DNA with the primers amplifying the V6/V7 regions showed higher numbers of bands than those obtained with the primers amplifying the V1–V3 regions (Fig. 1B). The maximum number of bands detected by DGGE of the V1–V3 regions was five, while, on the basis of the V6/V7 regions, maximally eight bands were detected in the profiles obtained from the same soil samples (Fig. 1B).

3.3.3 PCR-DGGE of soil DNA

The numbers of dominant DGGE bands obtained from soil samples of the first sampling year varied from three to nine, and depended mainly on soil treatment (Fig. 2A,B). Based on the DGGE banding patterns, clear differences were observed between the treatments, notably permanent grassland vs. arable land. Per treatment, very little variation was found between the DGGE profiles obtained from the three replicate plots. However, one plot of the permanent grassland showed slightly different DGGE patterns in comparison with the other two plots of that treatment (Fig. 2B). Clustering of the DGGE patterns using UPGMA confirmed the clear separation of all patterns in two main clusters (Fig. 3). All profiles from the permanent grassland formed one cluster, which grouped together with the profiles obtained from arable land of November, at 40% similarity. A second main cluster was formed by all other arable land samples, at 21% similarity to the former cluster. Whereas there was no clear effect of sampling time in the grassland samples, those from arable land showed a subclustering in accordance with sampling date.


DGGE analysis of PCR products from DNA extracted from soil samples under different treatments (first year sampling): A: long-term arable land under rotation; B: permanent grassland; F: February; M: May; S: September; N: November. For every sampling three replicate plots are presented (1, 2 and 3). m: marker (from top to bottom, amplicons of Enterobacter cloacae BE1, Listeria innocua ALM105, Rhizobium leguminosarum bv. trifolii R62, Arthrobacter sp., Burkholderia cepacia P2). DGGE bands (closest hits in database): a: P. rhodesiae (99%, AY043360); b: P. fluorescens (100%, AY092072); c: P. migulae (100%, AY047218); d: P. syringae (99%, AF511511); e: P. orientalis (99%, AF064457); f: Pseudomonas sp. E102 (100%, AF451270); g: Pseudomonas sp. SaU7 (100%, AF511510); h: P. rhodesiae (98%, AY043360); i: P. putida (97%, AF094743); j: P. tolaasii (98%, AF320986).


Dendrogram constructed with UPGMA representing the similarity between Pseudomonas-specific PCR-DGGE patterns obtained from soil samples under different management regimes over one growing season (2000). Codes: F: February; M: May; S: September; N: November; A: arable land under rotation; G: permanent grassland; 1, 2, 3: replicates 1, 2 and 3.

In order to identify the most dominant Pseudomonas species that made up the DGGE patterns, selected dominant DGGE bands were excised, re-amplified and sequenced. The single strong band apparent in the samples from the arable land of February and May 2000 (Fig. 2A, band a) was affiliated with a 16S rRNA gene sequence of P. rhodesiae (99%, accession number AY043360). The bands detected in the arable land later in the same sampling year, i.e. in September and November 2000 (Fig. 2A, bands b–d), were affiliated with sequences of P. fluorescens (100%, AY092072), P. migulae (100%, AY047218) and P. syringae (99%, AF511511). The numbers of DGGE bands detected in the permanent grassland were higher than those in the arable land in every sample during the season (Fig. 2B), and the banding patterns were relatively stable. The most dominant DGGE bands (Fig. 2B, bands e–j) were affiliated with sequences from P. orientalis (99%, AF064457), Pseudomonas sp. E102 (100%, AF451270), Pseudomonas sp. SaU7 (100%, AF5111510), P. rhodesiae (98%, AY043360), P. putida (97%, AF094743) and P. tolaasii (91%, AF320986).

In the second sampling year, the effects of five treatments, i.e. G, A-R, A-M, GA-R and GA-M, were analyzed. The DGGE patterns were consistent between the replicates per treatment, based on both the intensities and numbers of bands. The numbers of bands ranged from five to 12, depending on the treatment (Fig. 4). In the long-term arable land, only five dominant bands were detected, while in the short-term arable land and the permanent grassland, the numbers of dominant DGGE bands were nine and 12, respectively. Analysis of the DGGE patterns by UPGMA showed a clear clustering along treatment, with 40% similarity between all profiles (Fig. 5). The DGGE patterns obtained from the short-term arable land were most similar to those obtained from the permanent grassland (56% similarity), whereas those from arable land under maize monoculture clustered with these at 50% similarity.


PCR-DGGE banding patterns representing the dominant Pseudomonas populations in soil samples under different treatments during the second sampling year. For soil treatments see Section 2.3. Two or three replicates per treatment, numbered 1, 2 and 3, are shown. Where only two are shown, the third replicate was equal. M: Marker (from top to bottom, amplicons of Enterobacter cloacae BE1, Listeria innocua ALM105, Rhizobium leguminosarum bv. trifolii R62, Arthrobacter sp., Burkholderia cepacia P2). DGGE bands (closest hits in database shown): a: P. orientalis (99%, AF064457); b: Pseudomonas sp. E102 (100%, AF451270); c: P. libaniensis (100%, AF057645); d: Pseudomonas sp. SaU7 (100%, AF511510); e: P. fluorescens LCSA0TU2 (100%, AF506042); f: P. rhodesiae (98%, AY043360); g: P. tolaasii (98%, AF320986).


Dendrogram constructed with UPGMA representing the similarity between Pseudomonas-specific PCR-DGGE patterns obtained from soil samples under different management regimes (as shown in Section 2.3). Numbers 1, 2 and 3 indicate replicates.

Three DGGE bands present in all treatments (Fig. 4, bands b, d and e) were homologous to sequences of Pseudomonas sp. E102 (100%, AF451270), Pseudomonas fluorescens LCSA0TU2 (100%, AF506042) and Pseudomonas sp. SaU7 (100%, AF5111510). One other band, not detected in the first sampling year, was detected in the second year in the samples from A-M. This band (Fig. 4, band c) was found to be related to P. libaniensis (100%, AF057645). In the second sampling year, no significant change in the DGGE patterns was observed from the first year for the permanent grassland. The most dominant DGGE bands making up these patterns (Fig. 4, bands a, b, e–g) had sequences affiliated with P. orientalis (99%, AF064457), Pseudomonas sp. E102 (100%, AF451270), P. fluorescens LCSA0TU2 (100%, AF506042), P. rhodesiae (98%, AY043360) and P. tolaasii (99%, AF320992).

3.4 Phylogenetic and DGGE-based analysis of selected Pseudomonas isolates

About 165 Pseudomonas isolates selected among the 500 (33 per treatment; five treatments) obtained from Gould's S1 agar were analyzed by DGGE and partial sequencing of their 16S rRNA genes. All sequences showed between 96 and 100% similarity with 16S rRNA gene sequences of Pseudomonas spp. from the database. The sequences were pre-grouped on the basis of their affiliations, as well as on internal alignments. Then, a selection (minimally one sequence per pre-group) was used to generate a phylogenetic tree. This analysis showed that the sequences clustered into six different groups, with very high levels of relatedness between them (Fig. 6). The first group (I) included sequences that showed affiliation to the highly related species P. veronii, P. syringae, P. rhodesiae and P. tolaasii. The second group (II) included all sequences that were affiliated with P. marginalis and P. libaniensis. Group III was the largest one, including most of the sequences affiliated with organisms denoted Pseudomonas sp. (NZ031, Sau7, E102, NZ065, NZ081 and Fa2). All isolates identified as P. fluorescens formed group IV. Groups V and VI were formed by sequences affiliated with organisms denoted P. migulae and P. orientalis, respectively.


Neighbor-joining phylogenetic tree based on partial 16S rDNA sequences from soil-derived Pseudomonas isolates. In parentheses, the closest hit from the database is shown (>97–100% similarity). Representatives of each group of sequences shown here: group I (4 representatives), group II (3), group III (8), group IV (3), group V (1) and group VI (2). The sequence of Serratia plymuthica was used as an outgroup. A bootstrap analysis was performed with 100 repetitions and values greater than 50% are indicated. Codes of the isolates correspond with the codes of Table 2.

All isolates of the six sequence groups produced bands on DGGE that migrated to a limited number of different positions. The isolates generally produced single bands, with the exception of two isolates identified as P. tolaasii (100%, AF057645) and P. syringae (100%, AF511511), which each produced two DGGE bands. Specifically, about seven migration positions were found. There was no absolute match between sequence group and migration position in DGGE, as fragments representing some groups, e.g. V and VI, migrated to the same position, whereas within other groups, fragments were at different banding positions (e.g. groups I, three positions, and III, two positions). Among the most dominant Pseudomonas strains detected in this study was a group of isolates related to Pseudomonas sp. (NZ031, E102, FSL-D1, SaU7, NZ065, NZ081). Except for one (Pseudomonas sp. Sau7), all of these isolates produced DGGE bands at the same position (corresponding to band f in Fig. 2B). Other isolates, related to P. fluorescens strains ATCC13525, ATCC17556, ATCC49642 and LCSA0TU2, also produced single DGGE bands at a different migration position. All other isolates, related to P. rhodesiae, P. tolaasii, P. orientalis and P. libaniensis, produced different single DGGE bands.

Some isolates apparently became dominant as a result of the different soil treatments. For example, during the time when no crop was present in the arable land under rotation, only one dominant Pseudomonas type (related to P. rhodesiae, 99%, AY043360) was detected. Later, when oats were grown in the same plots, three more types were detected, i.e. P. migulae (100%, AY047218), P. fluorescens (99%, AF094732) and P. syringae (100%, AF511511). In the second sampling year in the same plots (maize as the crop), the most dominant types were related to Pseudomonas sp. E102 (99%, AF541270), P. fluorescens (99%, AJ308308), and P. libaniensis (100%, AF057645). Several types were detected in every treatment, but with large differences in the isolation frequencies (Table 2). For example, 50% of the isolates related to P. fluorescens LCSA0TU2 (99–100%, AF506042) originated from the permanent grassland, 41.6% from the short-term arable land and 8.4% from the arable land. One particular isolate, related to Pseudomonas sp. E102 (98–100%, AF451270), which possessed strong antifungal and chitinolytic activity and produced PRN, was found mainly in the permanent grassland (33%) or short-term arable land (GA-M 20%; GA-R 33%) and less in the arable land (13%). Conversely, 54% of isolates identified as P. rhodesiae (99–100%, AY043360) originated from the arable land under rotation and only 15% from the permanent grassland.

4 Discussion

Due to the wide distribution of different Pseudomonas species in the environment, and as a result of the relatively easy cultivation of the strains, the genus Pseudomonas is among the best-studied bacterial groups worldwide. In our study, we selected Gould's S1 as the selective medium for studying culturable Pseudomonas populations, as several previous studies [20,29] had proved this medium to be most selective, least susceptible to overgrowth by non-pseudomonads and most reproducible, allowing the growth of a broad diversity of Pseudomonas spp. Using this medium, differences in the numbers of total and fluorescent Pseudomonas cfu were found between the different soil treatments. These differences had most likely been induced by the differences in the soil management regimes, although a range of other factors may also have affected them. Most strikingly, grassland and arable land recently derived from grassland appeared to maintain a larger population density of pseudomonads than long-term arable land. However, the fact that the numbers of pseudomonads in the short-term arable land were similar to those in the permanent grassland may be due to remnants of grass roots that were still present in the short-term arable land. A similar observation was recently made for Bacillus populations studied in the same treatments [30]. More importantly, in the current study, the treatments were found to have different effects with respect to the frequencies of isolation of Pseudomonas strains with R. solani AG3 antagonistic properties. Hence, the abundances of the anti-AG3 types, within Pseudomonas populations of varying sizes per agricultural regime, were affected by treatment, clearly indicating that agricultural management influenced both the total and the specific (antagonistic) pseudomonads. Most of the isolates with either direct (unspecified) antagonistic activitynads. Most of the isolates with either direct (unspecified) antagonistic activity toward the soil-borne plant pathogen R. solani AG3 or with chitinolytic or antibiotic production capacity originated from the permanent grassland or from the short-term arable land, whereas isolation of such strains from long-term arable land was more rare. Hence, the plant (grass), by the action of its roots, most likely influenced the Pseudomonas community structure in soil, in the sense that Pseudomonas types with antimicrobial activity were enriched in numbers. Similar effects of plants on the abundance of antagonists of Verticillium dahliae were found in a recent study by Berg et al. [31]. These authors observed an effect of different plants on P. putida populations, which were enriched, in particular by strawberry. In our study, the differences in the sizes of the Pseudomonas spp. populations between different treatments were observed on the basis of not only culture-dependent, but also culture-independent analyses.

In this study, the discriminative value of a PCR-DGGE system based on the Pseudomonas-specific primers described by Widmer et al. [21] was evaluated. On the basis of the clear patterns obtained, the system proved to be very useful for studying the Pseudomonas diversity directly on the basis of soil-derived DNA. Thus, discrimination between Pseudomonas populations in fields under different agricultural management regimes became possible. The PCR-DGGE system, which includes the most variable region of 16S rDNA, i.e. V6, was compared to the one recently proposed by Gyamfi et al. on the basis of the V1–V3 regions of 16S rDNA [23]. The banding patterns obtained by the two systems showed clear differences. Since the Pseudomonas-specific PCR-DGGE system proposed in this work showed higher diversity and better separation of bands than the PCR-DGGE system of Gyamfi et al. [23], we recommend it for further work on the diversity of Pseudomonas spp. in soil. On the other hand, we found that there was no clear match between sequence type and band (migration) position, and, hence, we cannot make robust inferences about the total numbers of Pseudomonas types present solely on the basis of the patterns obtained. A combination of PCR-DGGE analysis with 16S rRNA gene sequencing on the basis of clone libraries and/or excised DGGE bands, such as applied in this study, is actually recommended, as sequencing should clearly complement the limited discrimination in different migration-based Pseudomonas‘types’ achieved by PCR-DGGE. Such a combined analysis will, thus, provide a more complete picture of the diversity of pseudomonads in soil than what can be achieved by fingerprinting alone. One striking observation was the high level of relatedness between all Pseudomonas sequences found from DGGE bands, from the clone library analyzed or from the isolates on Gould’s S1 agar. Apparently, the pseudomonads, targeted by these methods, are all taxonomically tightly knit, although numerous different types with clearly different ecological roles can be shown to be part of this tight group. This close relatedness, at the 16S rRNA gene level, is also supported by the literature [24].

The results obtained with the direct molecular analyses in this study supported the general view that agricultural management and crop type strongly influence the structure of Pseudomonas populations in soil. For example, permanent grassland consistently showed the presence of more diverse Pseudomonas populations than arable land under rotation (with initially no rhizosphere present, and a developing rhizosphere in the course of the season). Whereas this enhanced diversity in grassland may have been a long-term effect, also in the arable land itself the presence of a crop had clear effects on the Pseudomonas populations. During the time when no crop was present in the arable land under rotation, only one dominant Pseudomonas type, affiliated with P. rhodesiae, was detected by sequence analysis, but the number of different sequences clearly increased as soon as a developing crop was present. Hence, plants are indeed major factors that can quickly, i.e. within sub-periods of crop growth over one growing season, influence the structure of Pseudomonas communities in soil. Such plant effects on Pseudomonas populations corroborate results obtained before in several studies in other soil or agricultural systems. Latour et al. [32] showed that plant, as well as soil type, had a selective influence on the populations of fluorescent Pseudomonas spp. Clays-Josserand et al. [33] found that P. fluorescens was the numerically most abundant species in the rhizosphere of flax and tomato plants, followed by P. putida and P. chlororaphis. Lottmann and Berg [34] found that the most dominant pseudomonads associated with the potato rhizosphere were P. putida, P. fluorescens, P. chlororaphis and P. syringae. According to Misko and Germida [35], P. putida and P. chlororaphis were the predominant bacteria associated with the roots of field-grown canola plants in anada. As virtually all Pseudomonas spp. are known to be r-strategists, copiotrophic rganisms which are able to quickly respond to enhanced nutrient availability in their environment, it is expected that plants with roots exuding different compounds will indeed quickly select different Pseudomonas communities in their rhizospheres.

Taken together, the results obtained in the present study support the hypothesis that plants are major factors determining the Pseudomonas community structure and composition in soil, at both the cultivation-based and cultivation-independent levels. In particular, the selection in grassland of the group of pseudomonads defined by Pseudomonas sp. E102 and SaU7 was striking. These isolates were strongly antagonistic to the target phytopathogen R. solani AG3, produced chitinases and possessed the genetic machinery for the production of PRN and PCA. We suggest that the abundance and activity of these organisms, stimulated by grass roots, is one of the underlying causes of the suppression of R. solani AG3-induced potato rot. Further detailed analyses into the ecological role of these organisms are certainly warranted.


We thank Krysta Voesenek for excellent technical assistance. J. Raaijmakers is acknowledged for providing primer sequences for detection of the PRN, PCA and DAPG production loci, and relevant reference strains, and W. de Boer for his pertinent advice. We further thank the Dutch Organization for Pure Scientific Research (NWO), the Dutch Ministry of Agriculture, Directorate DWK, Agrobiodiversity Program (352) and the EU POTATOCONTROL project for financial support.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
View Abstract