SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (37)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cattelan, A.J.
Right arrow Articles by Fuhrmann, J.J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Cattelan, A.J.
Right arrow Articles by Fuhrmann, J.J.
Agricola
Right arrow Articles by Cattelan, A.J.
Right arrow Articles by Fuhrmann, J.J.
Soil Science Society of America Journal 63:1670-1680 (1999)
© 1999 Soil Science Society of America

DIVISION S-3-SOIL BIOLOGY & BIOCHEMISTRY

Screening for Plant Growth–Promoting Rhizobacteria to Promote Early Soybean Growth

A.J. Cattelana, P.G. Hartela and J.J. Fuhrmannb

a Dep. of Crop & Soil Sciences, 3111 Plant Sciences Bldg., Univ. of Georgia, Athens, GA 30602-7272 USA
b Dep. of Plant and Soil Sciences, Univ. of Delaware, Newark, DE 19717-1303 USA

pghartel{at}arches.uga.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and methods
 Results
 Discussion
 REFERENCES
 
Although many studies have been conducted to identify the specific traits by which plant growth–promoting rhizobacteria (PGPR) promote plant growth, usually they were limited to studying just one or two of these traits. We selected 116 isolates from bulk soil and the rhizosphere of soybean [Glycine max (L.) Merr.] and examined them for a wide array of traits that might increase early soybean growth in nonsterile soil (PGPR traits). A subsample of 23 isolates, all but one of which tested positive for one or more of these PGPR traits, was further screened for traits associated with biocontrol, (brady)rhizobial inhibition, and rhizosphere competence. Six of eight isolates positive for 1-aminocyclopropane-1-carboxylate (ACC, a precursor of ethylene) deaminase production, four of seven isolates positive for siderophore production, three of four isolates positive for ß-1,3-glucanase production, and two of five isolates positive for P solubilization increased at least one aspect of early soybean growth. One isolate, which did not share any of the PGPR traits tested in vitro except antagonism to Sclerotium rolfsii and Sclerotinia sclerotiorum, also promoted soybean growth. One of the 23 isolates changed bradyrhizobial nodule occupancy. Although the presence of a PGPR trait in vitro does not guarantee that a particular isolate is a PGPR, the results suggest that rhizobacteria able to produce ACC deaminase and, to a lesser extent, ß-1,3-glucanase or siderophores or those able to solubilize P in vitro may increase early soybean growth in nonsterile soil.

Abbreviations: ACC, 1-aminocyclopropane-1-carboxylate • ANOVA, analysis of variance • DAP, days after planting • IAA, indoleacetic acid • FAME, fatty acid methyl ester • PGPR, plant growth–promoting rhizobacteria • TSA, trypticase soy broth solidified with agar • TSB, trypticase soy broth


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and methods
 Results
 Discussion
 REFERENCES
 
THE MECHANISMS by which PGPR promote plant growth are not fully understood, but are thought to include: (i) the ability to produce or change the concentration of the plant hormones indoleacetic acid (IAA; Mordukhova et al., 1991), gibberellic acid (Mahmoud et al., 1984), cytokinins (Tien et al., 1979), and ethylene (Arshad and Frankenberger, 1991; Glick et al., 1995); (ii) asymbiotic N2 fixation (Boddey and Döbereiner, 1995; Kennedy et al., 1997); (iii) antagonism against phytopathogenic microorganisms (e.g., Fusarium spp.; Scher and Baker, 1982) by production of siderophores (Scher and Baker, 1982), ß-1,3-glucanase (Fridlender et al., 1993), chitinases (Renwick et al., 1991), antibiotics (Shanahan et al., 1992), and cyanide (Flaishman et al., 1996); and (iv) solubilization of mineral phosphates and other nutrients (Sperber 1958a, 1958b; De Freitas et al., 1997). Many of the studies with PGPR show plant growth promotion, but only under gnotobiotic conditions (Tien et al., 1979; Shenbagarathai, 1993; Glick et al., 1995) or in potting media (Polonenko et al., 1987; Fuhrmann and Wollum, 1989b) where these bacteria do not compete with the normal array of soil microorganisms.

There are some cases where PGPR may promote plant growth in nonsterile soil by controlling fungal diseases. The addition of a siderophore-producing Pseudomonas putida converted a Fusarium-conducive soil into a Fusarium-suppressive soil for the growth of three different plants (Scher and Baker, 1982). An isolate of Pseudomonas cepacia, positive for ß-1,3-glucanase production, decreased the incidence of diseases caused by Rhizoctonia solani, Sclerotium rolfsii, and Pythium ultimum (Fridlender et al., 1993). Similarly, five fluorescent Pseudomonas isolates, each positive for antibiotic production, promoted potato (Solanum tuberosum L.) growth in nonsterile soil (Kloepper and Schroth, 1981).

In addition to the previously described PGPR traits, some rhizobacteria can promote plant growth indirectly by affecting symbiotic N2 fixation, nodulation, or nodule occupancy. In a competition study where soybean was coinoculated with Bradyrhizobium japonicum USDA 110 and USDA 118 and one of 17 different isolates of rhizobacteria, nine rhizobacterial isolates increased the weights of nodules formed by B. japonicum USDA 110 and three rhizobacterial isolates increased the number of nodules (Polonenko et al., 1987). Similarly, Fuhrmann and Wollum (1989b) found three fluorescent pseudomonads that consistently increased nodule occupancy of the more efficient B. japonicum USDA 110 over the less efficient B. japonicum USDA 123 and USDA 31 (now B. elkanii) in soybean grown in a potting medium with low Fe availability. Presumably these pseudomonads increased nodule occupancy because of their siderophore production. Growth factors such as vitamins might also indirectly affect the growth of (brady)rhizobia in the rhizosphere (Derylo and Skorupska, 1993; Streit and Phillips, 1996). Because some strains of the genus Bradyrhizobium do not produce biotin (Holt et al., 1994), these bradyrhizobial strains may benefit when grown with biotin-producing bacteria.

In addition to these direct and indirect traits, PGPR must also be rhizosphere competent and able to survive in soil. Traits associated with rhizosphere competence and survival in soil include an ability to tolerate a reasonable range of abiotic factors including temperature, pH, and moisture (Sylvia et al., 1998). Furthermore, PGPR should not be associated with traits attributed to deleterious rhizobacteria. These include production of cellulase and pectinase (Ulrich, 1976; Chatterjee and Starr, 1980). Cyanide production is an ambiguous trait and is sometimes associated with deleterious as well as beneficial rhizobacteria (Bakker and Schippers, 1987; Alström and Burns, 1989).

We screened 116 bacteria isolated from bulk soil and soybean rhizosphere for putative PGPR traits that might promote some aspect of early soybean growth such as shoot height, root length, shoot and root dry weight, and nodule number and dry weight. The traits we tested were production of siderophores, IAA, chitinase, ß-1,3-glucanase, ACC deaminase, and cyanide, P solubilization, biotin prototrophy, asymbiotic N2 fixation, and inhibition of growth of three pathogenic soil fungi. A subsample of 23 isolates, all but one of which tested positive for one or more of the above traits, was further screened for production of volatile antifungal compounds, inhibition of growth of eight strains of (brady)rhizobia, production of cellulase and pectinase, and growth at three different temperatures and at pH 5.5 and 7.0. The 23 isolates were tested in nonsterile soil under greenhouse or lightroom conditions for promotion of early soybean growth. A further subsample of PGPR, which promoted some aspect of early soybean growth, was coinoculated with B. japonicum USDA 110 to determine their ability to alter nodule occupancy.


   Materials and methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
Results
 Discussion
 REFERENCES
 
Origin of Bacteria
A subsample of 116 bacterial isolates was selected from a sample of 1131 isolates from bulk soil and the rhizosphere of soybeans (Cattelan et al., 1998). Subsample isolates were initially selected on the basis of fatty acid methyl ester (FAME) identification. From a collection of 572 isolates with a similarity index >0.300 for FAME profile match, 60 isolates representing 15 different species from eight different genera were selected. The most common genera were Pseudomonas or Burkholderia (39 isolates), Bacillus (eight isolates), and Alcaligenes (five isolates). Whenever possible, different species within each genus were selected. In order to maximize diversity among the remaining 559 unknown isolates, the isolates were assessed for Gram reaction and differences in cellular and colony morphology. Because the rhizosphere tends to select for Gram-negative bacteria over Gram-positive bacteria (Kloepper et al., 1992; Gilbert et al., 1993), an additional Gram-positive and 55 Gram-negative isolates were selected. This yielded a total of 105 Gram-negative and 11 Gram-positive isolates. Pseudomonas putida PH6 (Fuhrmann and Wollum, 1989a, 1989b) was selected as a PGPR control. The 116 isolates were then screened for traits that might be associated with ability to function as PGPR. Each in vitro test was replicated three times.

In Vitro Tests
Siderophore production was determined in solution as described by Schwyn and Neilands (1987) except the medium was 0.1x trypticase soy broth (TSB; Difco Laboratories, Detroit, MI). Indoleacetic acid production was determined as described by Brick et al. (1991) except that the medium was 0.1x TSB solidified with agar (15 g L-1; TSA).

Chitinase production was determined as described by Renwick et al. (1991) in a defined medium composed of (g L-1): colloidal chitin (Reid and Ogrydziak, 1981), 8.0; NH4NO3, 0.78; K2HPO4, 0.80; KH2PO4, 0.20; MgSO4 · 7H2O, 0.20; CaCl2, 0.06; NaCl, 0.10; Na2MoO4 · 2H2O, 0.002, ZnSO4 · 7H2O, 0.00024; CuSO4 · 5H2O, 0.00004; CoSO4 · 7H2O, 0.010; MnSO4 · 4H2O, 0.003; Na2FeEDTA, 0.028; H3BO3, 0.005; and agar, 15. Magnesium sulfate and CaCl2 were autoclaved separately and added to the medium after autoclaving. Biotin (5 µg L-1) and p-aminobenzoic acid (10 µg L-1) were filter-sterilized (0.2-µm) and were added to the medium after autoclaving.

ß-1,3-glucanase production was determined as described by Renwick et al. (1991) in the previously defined medium, except the C source was ß-1,3-glucan (5 g L-1; laminarin, Sigma Chemical Co., St. Louis, MO). Production of ACC deaminase was determined as described by Glick et al. (1995) by the previously described defined medium, except the C sources were sucrose (5.0 g L-1), mannitol (5.0 g L-1), and sodium lactate (0.5 mL of a 5.4 M solution), and the N source was ACC (5.0 g L-1). Cyanide production was determined as described by Bakker and Schippers (1987) in 0.1x TSA amended with glycine (4.4 g L-1) and FeCl3 · 6H2O (0.3 mM; Castric, 1975). Solubilization of P was determined as described by Katznelson and Bose (1959), except the medium was 0.1x TSA. Biotin prototrophy was determined by growing the isolates in 0.25x Bacto Biotin Assay Medium (Difco Lab., Detroit, MI).

Putative, free-living, N2-fixing bacteria were screened in the ACC deaminase defined medium, except the N source was eliminated and the agar was reduced to 1.75 g L-1 (Day and Döbereiner, 1976). The isolates that grew after being sequentially transferred 10 times to the same medium were considered presumptive positive for N2 fixation. To confirm nitrogenase activity, positive isolates were tested for their ability to reduce acetylene to ethylene (Dart et al., 1972) after 24 h in the same defined medium amended with 100 mg of yeast extract L-1.

To assess the ability of the 116 isolates to inhibit fungi, each isolate was tested under conditions of low and high Fe (to suppress siderophore production; Scher and Baker, 1982; Fuhrmann and Wollum, 1989a) against three different fungi with the circle method (Da Luz, 1990). The fungi — Sclerotium rolfsii, Fusarium oxysporum, and Sclerotinia sclerotiorum — were obtained from the culture collection of R.W. Roncadori (University of Georgia, Athens). Briefly, bacterial isolates were seeded in a 5.0-cm-diameter circle on a 0.1x TSA plate that was either unamended or amended with 0.1 mM FeCl3. After 24 h at room temperature, a 7-mm plug of each fungus was placed on the center of the circle. Plates with S. rolfsii and F. oxysporum were incubated at 28°C for 6 and 8 d, respectively, and plates with S. sclerotiorum were incubated at 22°C for 5 d. Fungal growth inhibition was assessed by measuring the mycelial radial growth.

A subsample of 23 isolates, all but one of which displayed strong tendencies for one or more of the aforementioned traits, was selected from the 116 isolates and, with the PGPR control P. putida PH6, were further tested for production of volatile antifungal compounds, ability to inhibit growth of eight strains of (brady)rhizobia, production of cellulase and pectinase, and ability to grow at 18, 28, and 37°C and pH 5.5 and 7.0.

The ability to produce volatile compounds inhibiting S. rolfsii, F. oxysporum, and S. sclerotiorum was determined in two-compartment Petri plates containing 0.1x TSA inoculated with a 7-mm, 0.1x TSA disc with each fungus on one side and a lawn of bacterial cells from each isolate spread on the other side (Gagné et al., 1991). Fungal growth was assessed by measuring the radial growth of the mycelium.

The isolates were tested for their ability to inhibit growth of eight strains of (brady)rhizobia: Bradyrhizobium elkanii USDA 31, 76, and 94; B. elkanii SEMIA 587 and 5079; B. japonicum USDA 110 and 123; and Rhizobium fredii USDA 205. Bradyrhizobium japonicum USDA 110 was obtained from A.G. Wollum (North Carolina State Univ., Raleigh), B. elkanii SEMIA 587 and 5079 were obtained from Fundação Estadual De Pesquisa Agropecuária (FEPAGRO, Porto Alegre, Brazil), R. fredii USDA 205 was obtained from the USDA culture collection (Beltsville, MD), and B. elkanii USDA 31, 76, and 94 were from our own culture collection. Tests were conducted in 0.1x TSA plates inoculated with a lawn of bacterial cells from each of the 23 isolates in one half and perpendicular streaks of each (brady)rhizobial strain in the other half. One to four (brady)rhizobial strains were streaked from a lightly turbid cell suspension in 0.15 M NaCl. The (brady)rhizobial strains were separated from the cell lawn by a distance of 2 mm. The linear inhibition of the (brady)rhizobial growth was measured after 7 d at 28°C. The control treatment was (brady)rhizobial inoculation without bacterial inoculation.

Cellulase production was determined in M9 medium (Miller, 1974) amended with yeast extract (1.2 g L-1) and cellulose (10 g L-1; Sigmacell Type 101, Sigma Chemical; Samanta et al., 1989). After 8 d of incubation at 28°C, isolates surrounded by clear halos were considered positive for cellulase production. Pectinase production was determined in the same M9 medium except the cellulose was replaced with pectin (4.8 g L-1). After 2 d of incubation at 28°C, the plates were flooded with 2 M HCl (T. Denny, 1997, personal communication), and isolates surrounded by clear halos were considered positive for pectinase production (Andro et al., 1984).

The ability of each bacterial isolate to grow at 18, 28, and 37°C was determined in 0.1x TSB. A 100-µL sample of a cell suspension, adjusted to an absorbance of 0.030 at 600 nm, was transferred into 10 mL of 0.1x TSB contained in 50-mL Erlenmeyer flasks and the contents periodically measured spectrophotometrically at 600 nm for 24 h. The ability of the bacterial isolates to grow at two different pHs was determined in a similar manner. The isolates were grown in 0.1x TSB at pH 7.0 (natural pH of the medium) and pH 5.5 (adjusted with 1 M HCl). The temperature was kept at 28°C.

In addition to the PGPR control P. putida PH6, one isolate, LN1116, was consistently negative for all the PGPR traits in vitro except antagonism to S. sclerotiorum and S. rolfsii, and it was selected as the negative bacterial control for the short-term plant tests.

Short-Term Plant Tests
The Ap horizon of Appling sandy loam (clayey, kaolinitic, thermic, Typic Kanhapludults) was obtained from the University of Georgia Plant Sciences Farm near Watkinsville, GA. The Ap horizon of Dothan loamy sand (fine-loamy, kaolinitic, thermic Plinthic Kandiudults) was obtained from the Southeast Georgia Branch Experiment Station near Midville, GA. Both soils were passed through a 2-mm sieve and stored at room temperature until used. The soils were analyzed for pH, total organic C, and selected nutrients according to standard procedures (Plank, 1989) at the University of Georgia Soil Testing and Plant Analysis Laboratory (Table 1) . The gravimetric water contents (kg water kg-1 soil) of the soils at -0.01 and -0.03 MPa were determined by the pressure plate method (Klute, 1986). Soil water content was determined gravimetrically by oven drying the soil overnight at 105°C.


View this table:
[in this window]
[in a new window]
  		Table 1 Some chemical and physical characteristics of Appling and Dothan soils

 
In all experiments, seeds of soybean cv. Lee were wetted in 20.0 M ethanol (95%) for 5 s and surface-sterilized in 0.21 M NaOCl for 5 min. The seeds were washed once in sterile distilled water, soaked for 10 min in 0.01 M HCl (to remove traces of NaOCl; Abdul-Baki, 1974), and washed five times in sterile, distilled water to remove traces of HCl. On the last rinse, seeds were left to imbibe water for 1 h. Seeds were germinated between pieces of sterile, moist filter paper at 28°C for 2 d.

All 23 bacterial isolates and P. putida PH6 were grown on 0.1x TSA overnight at 28°C. The cells were suspended in 0.1 M MgSO4 (pH 7.0) to give an absorbance of 0.55 at 600 nm ({approx}109 cells mL-1). Except for the nodule occupancy experiment, germinated seeds were immersed in the cell suspension of each isolate or into sterile 0.1 M MgSO4 (nonbacterial control) for 5 min. In the case of the nodule occupancy experiment, germinated seeds were immersed in equal volumes of cell suspensions of the bacterial isolates and B. japonicum USDA 110. Three seeds were planted per pot, except for the nodule occupancy experiment where four seeds were planted. The soil surface was covered with 1 cm of sterile sand to reduce aerial contaminants. In the lightroom experiments, plants were grown in {approx}500 g (dry weight basis) of Appling sandy loam or Dothan loamy sand (P-solubilization experiment only) contained in 600-mL D-pots (Stuewe and Sons, Corvallis, OR) with a light intensity of 550 µM photon m-2 s-1, a 16/8-h light/dark cycle, and day/night temperatures of 26 ± 1 and 24 ± 1°C, respectively. In the single greenhouse experiment, plants were grown in 1740 g (dry weight basis) of Appling sandy loam contained in 2-L polyethylene pots. Water was added daily to each soil to attain a water content equivalent to -0.03 MPa for all experiments except for the Fusarium antagonism experiment, where the soil water content was equivalent to -0.01 MPa. After emergence, plants were thinned to one plant per pot. Each treatment was replicated four times in a completely randomized design. At harvest, the shoots were cut off and the roots were washed gently to minimize nodule loss. Nodules were separated from the roots and counted. Roots, shoots, and nodules were oven dried at 50°C and weighed after 2 d.

For bacterial isolates positive for IAA or ACC deaminase production, plants were sampled at 14 d after planting (DAP) and plant shoot height was measured. Roots were stored in formalin-acetic acid-alcohol (Postek et al., 1980, p. 296) and root length was estimated by the modified line intersect method (Tennant, 1975). Roots were oven dried at 50°C and weighed. For bacterial isolates positive for nitrogenase activity, plants were harvested at 20 DAP and total N of plant shoots was determined by the micro-Kjeldahl method (Plank, 1992).

Bacterial isolates positive for P solubilization were tested in Dothan loamy sand amended with 54 mg of K2SO4 and 92 mg of CaHPO4 kg-1 soil. Plants were sampled at 18 DAP. Total shoot P and N contents were determined by colorimetry and micro-Kjeldahl methods, respectively (Plank, 1992). Serology of the nodules from the control treatment was determined by an ELISA assay (Fuhrmann and Wollum, 1985) for the B. elkanii USDA 31, 76, and 94, and B. japonicum USDA 110 and 123 serogroups.

Bacterial isolates positive for ß-1,3-glucanase, cyanide, siderophore, or chitinase production or positive for antagonism to F. oxysporum were tested in Appling sandy loam that was kept cooler (day and night temperatures of 23 ± 1 and 21 ± 1°C, respectively) and at higher matric potential (-0.01 MPa) than the other experiments. These conditions are better suited for the development of root rot caused by Fusarium spp. than are warmer and drier environments (Sinclair and Backman, 1989). Because antagonism against F. oxysporum was the most variable of the three in vitro fungal antagonism tests, this fungus was chosen for the plant tests. To add F. oxysporum to the soil, the fungus was grown for 4 d on 0.1X TSA, scraped from the surface with an inoculating loop, and sonicated for 1 h in 200 mL of distilled water. After adding the suspension to air-dried soil, the fungal counts were determined by serially plating a portion of the solution on 0.1X TSA. The final count of F. oxysporum in soil was {approx}1.7 x 103 colony-forming units g-1. The light and dark cycle was 14 and 10 h, respectively. Isolates positive for chitinase production were also tested in the presence and absence of chitin. After inoculation, each seed was coated with {approx}11.4 µg of ground (250 µm), autoclaved chitin. Plants were sampled at 14 DAP.

Eight bacterial isolates that produced significant effect(s) in the previous plant-growth and nodulation experiments or exhibited in vitro antagonism against strains of Bradyrhizobium spp. were tested, along with P. putida PH6, for their ability to alter nodule occupancy. Plants were sampled at 28 DAP. Serology of the nodules from the tap root was determined by an ELISA assay for the B. elkanii USDA 31, 76, and 94 and B. japonicum USDA 110 and 123 serogroups for all treatments. Each treatment was replicated five times in a completely randomized design.

Clustering of bacterial FAME profiles was determined with CANOCO, a program for canonical community ordination (Microcomputer Power, Ithaca, NY). Because of the possibility that some unknown isolates that increased early soybean growth could be clones of each other, these isolates were also tentatively identified with Biolog (Biolog, Hayward, CA) and ribosomal typing (RiboPrinter Microbial Characterization system, Qualicon, Wilmington, DE). In the case of ribotyping, the restriction enzyme was EcoRI.

Statistical Analysis
All fungal and bradyrhizobial antagonism data were analyzed by analysis of variance (ANOVA) and treatment means were separated by the Tukey's test with the SAS statistical package (SAS Institute, 1988). All plant test data were also analyzed by ANOVA but treatment means were separated by the Fisher's least significant difference test.


   Results
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
Discussion
 REFERENCES
 
In Vitro Tests
Of the 116 isolates, 22 (19%) showed strong tendencies for one or more of the PGPR traits tested, and one, LN1116, was negative for all traits (Table 2) . The PGPR control P. putida PH6 was negative for all the traits except biotin prototrophy. A majority of the isolates (19 of 23, 83%) were Gram-negative rods. Four isolates produced IAA, eight produced ACC deaminase, five solubilized P, five exhibited low levels of nitrogenase activity, seven produced siderophores, two produced chitinase, four produced ß-1,3-glucanase, and five produced cyanide. Only isolate LN1116 was a biotin auxotroph. All isolates were negative for cellulase and pectinase production (data not shown). With regards to temperature, all isolates grew better at 28 than at 18°C. Most of the isolates also grew better at 28 than at 37°C, but at least four isolates grew at virtually the same rate at both temperatures. With regards to pH, only LN1116 grew better at pH 5.5 than at pH 7.0, and at least nine other isolates grew equally well at both pHs. Although five isolates apparently fixed N2, they exhibited low nitrogenase activities (0.05–0.11 nM of ethylene tube-1 h-1) compared with the positive control, Azospirillum brasilense ATCC 29145 (173 nM of ethylene tube-1 h-1).


View this table:
[in this window]
[in a new window]
  		Table 2 Selected physiological and biochemical traits of bacterial isolates from soil or the rhizosphere of soybean. Isolates were tentatively identified by fatty acid methyl ester (FAME) analysis. All isolates were negative for pectinase and cellulase production

 
All isolates, including P. putida PH6, inhibited the growth of S. rolfsii and S. sclerotiorum at both low and high Fe when the isolates and fungi were not physically separated (Table 3) . Ten isolates significantly inhibited the growth of F. oxysporum at low Fe and eight at high Fe. Five of the ten isolates were positive for siderophore production, four for ß-1,3-glucanase, and three for cyanide production. Four of the ten isolates were positive for two traits and two were negative for all traits (including chitinase). When the fungi and isolates were physically separated in two-compartment plates, 18 of 23 isolates inhibited S. rolfsii, 17 inhibited S. sclerotiorum, and two inhibited F. oxysporum.


View this table:
[in this window]
[in a new window]
  		Table 3 In vitro antagonism of three soil fungi by 24 bacterial isolates. Antagonism was measured by the inhibition of fungal radial growth on 0.1x trypticase soy agar under conditions of no physical separation (single compartment plates) with high and low Fe in the medium and physical separation (two-compartment plates)

 
With regards to (brady)rhizobial inhibition, 22 isolates inhibited at least one strain of (brady)rhizobia (Table 4) . Only one isolate, LC3116, did not inhibit any of the eight strains tested. Bradyrhizobium elkanii USDA 31, 76, and 94 were not inhibited by any of the isolates.


View this table:
[in this window]
[in a new window]
  		Table 4 In vitro antagonism of Bradyrhizobium elkanii, B. japonicum, and Rhizobium fredii by 24 bacterial isolates. Antagonism was measured by the inhibition of growth on 0.1x trypticase soy agar

 
Short-Term Plant Tests
Of the 11 isolates positive for hormone production, four isolates significantly increased soybean growth at 14 DAP: LN3212 and LW2301 increased shoot dry weight, root length, and root dry weight; GN1201 increased root length and dry weight; and GW2306 increased shoot and root dry weight (Table 5) . Interestingly, the negative bacterial control, LN1116, also increased root dry weight. Except for LN1116, all the isolates that affected plant growth were positive for ACC deaminase production; none of the isolates that were positive for IAA production affected plant growth. Three of the isolates that were positive for ACC deaminase were also positive for ß-1,3-glucanase or siderophore production or both.


View this table:
[in this window]
[in a new window]
  		Table 5 Effect of 11 bacterial isolates, each positive for the production of the plant-growth regulator indoleacetic acid (IAA) (GW2103, GW3202, GN2323, and LC1118) or 1-aminocyclopropane-1-carboxylate (ACC, a precursor of ethylene) deaminase (GW1101, GS1206, GN1201, GN1212, GW2306, LN3212, and LW2301) on growth and nodulation of soybean cv. Lee. Pseudomonas putida PH6 was the plant growth–promoting rhizobacteria control and LN1116 was the bacterial control negative for both traits tested

 
None of the five isolates positive for nitrogenase activity significantly promoted soybean growth at 20 DAP and, in fact, one isolate (LN1310) significantly decreased nodule dry weight (data not shown). Of the five isolates positive for P solubilization, two significantly affected some aspect of soybean growth in a P-deficient soil at 18 DAP (Table 6) . The PGPR control P. putida PH6 and isolate GN1212 showed significantly higher shoot P content, but this was offset in the case of GN1212 by a significant decrease in shoot dry weight. Isolate GN2214 significantly increased root dry weight, but this was offset by significantly lower nodule weight and number. However, it did not decrease shoot N content. The serological reaction of 35 nodules from the nonbacterial control treatment was tested to check for the distribution of bradyrhizobial strains in the Dothan loamy sand. Of the 35 nodules tested, seven (20%) aligned with B. elkanii USDA 31, six (17%) with B. elkanii USDA 76, 10 (29%) with B. elkanii USDA 94, and none with either B. japonicum USDA 110 or 123. Fourteen nodules (40%) did not react with any of the sera tested, and two (6%) reacted with more than one.


View this table:
[in this window]
[in a new window]
  		Table 6 Effect of five bacterial isolates, each positive for P solubilization, on growth, nodulation, and shoot P and N content of soybean cv. Lee. Pseudomonas putida PH6 was the plant growth–promoting rhizobacteria control and LN1116 was the bacterial control negative for the trait tested

 
When soybeans were inoculated with one of 12 isolates, each positive for ß-1,3-glucanase, cyanide, siderophore, or chitinase production or the ability to inhibit the growth of Fusarium in vitro, and grown in soil inoculated with F. oxysporum for 14 d, only isolate LN3212 increased all aspects of soybean growth. Isolates GW3205 and GN1212 also increased nodule dry weight (Table 7) . Again, the bacterial control isolate LN1116, negative for all traits, increased shoot dry weight. The remaining isolates had no effect on soybean growth.


View this table:
[in this window]
[in a new window]
  		Table 7 Effect of 12 bacterial isolates, each positive for the production of ß-1,3-glucanase, cyanide, siderophores, or chitinase, on growth and nodulation of soybean cv. Lee. Whether or not the isolate inhibited the growth of Fusarium oxysporum in vitro is also noted. Isolates positive for chitinase production were also tested with or without chitin. Pseudomonas putida PH6 was the plant growth–promoting rhizobacteria control and LN1116 was the bacterial control negative for all traits tested

 
Of the eight bacterial isolates tested in the nodule occupancy experiment plus the PGPR control P. putida PH6, only one isolate, LC3116, changed nodule occupancy at 28 DAP (Table 8) . In this case, LC3116 increased nodule occupancy with bradyrhizobial strains belonging to the serogroup B. elkanii USDA 31. However, this increased nodule occupancy was at the expense of other native bradyrhizobial serotypes, not serotype B. japonicum USDA 110. None of the isolates significantly altered nodule occupancy by B. japonicum USDA 110 and none promoted soybean growth (data not shown).


View this table:
[in this window]
[in a new window]
  		Table 8 Effect of nine bacterial isolates on nodule occupancy of soybean cv. Lee inoculated with B. japonicum USDA 110. Isolates were chosen based on a significant effect in previous plant growth and nodulation tests or in vitro antagonism to Bradyrhizobium strains. Pseudomonas putida PH6 was the plant growth–promoting rhizobacteria control

 
When the initial 116 soil and rhizobacterial isolates were compared by FAME profiles, distinct clustering was observed among the isolates (Fig. 1) . The seven isolates that increased some aspect of early soybean growth were closely associated with Pseudomonas spp. FAME analysis had previously identified three of the isolates as P. cepacia (GN1201), P. pickettii (GN2214), and P. chlororaphis (GN1212); the remaining four isolates were unknown. Biolog and ribotyping analyses were unable to identify any of these unknown isolates. However, ribotyping did show that the unknown isolates were different (data not shown).



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1 Two-dimensional plot of principal components analysis of fatty acid methyl ester (FAME) profiles of the bacterial isolates tested plus the plant growth–promoting rhizobacteria (PGPR) control, Pseudomonas putida PH6. The x- and y- axes are in Euclidean distance. Isolates are identified as (i) increasing some aspect of soybean growth (solid triangle), (ii) not increasing some aspect of soybean growth (open square), and (iii) not tested in plant (open circle)

 

   Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
 Discussion
REFERENCES
 
Seven of the 23 isolates selected for the plant-growth experiments significantly increased at least one aspect of early soybean growth, and five significantly increased at least two aspects. All seven isolates were originally from the soybean rhizosphere. Many of these seven isolates shared common characteristics: six were positive for ACC deaminase, four were positive for siderophore production, three were positive for ß-1,3-glucanase production, and two for P solubilization. Previous research suggests that bacteria possessing these traits can increase plant growth. Glick et al. (1995) observed that seven bacterial strains positive for ACC deaminase production promoted canola (Brassica napus L.) seedling root elongation under gnotobiotic conditions. 1-Aminocyclopropane-1-carboxylate deaminase might act to stimulate plant growth and root elongation in particular by sequestering and hydrolyzing ACC from germinating seeds. This would lower the level of ACC and consequently the level of ethylene in seeds. For many plant species, ethylene stimulates germination and breaks the dormancy of the seeds (Esashi, 1991). However, if the ethylene concentration remains high after germination, root elongation (as well as symbiotic N2 fixation in leguminous plants) is inhibited (Jackson, 1991). In the case of ß-1,3 glucanase production, Fridlender et al. (1993) observed that an isolate of P. cepacia with strong ß-1,3 glucanase production decreased the incidence of three fungal diseases in three different plants grown in nonsterile soil. More than 50% of strains that controlled take-all [Gaeumannomyces graminis (Sacc.) Arx & D. Olivier] of wheat (Triticum aestivum L.) produced ß-glucanases (Renwick et al., 1991). In the case of siderophore production, a siderophore-producing P. putida converted a Fusarium-conducive soil to a Fusarium-suppressive soil (Scher and Baker, 1982). Similarly, inoculation of chickpea (Cicer arietinum L.) and soybean seed with a siderophore-producing fluorescent pseudomonad resulted in increased seed germination, growth, and yield of the plants (Kumar and Dube, 1992). The only trait tested here that changed some aspect of soybean growth independently of ACC deaminase activity was P solubilization. Some P-solubilizing organisms have been reported as plant growth promoters, but rigorous proof is lacking (De Freitas et al., 1997; Whitelaw et al., 1997). Interestingly, isolate LC3116, which increased nodule occupancy with serogroup B. elkanii USDA 31, was also positive for P solubilization. The results suggest that rhizobacteria able to reduce the concentration of ethylene and to a lesser extent, produce ß-1,3 glucanase or siderophores or able to solubilize P, may be important in promoting early soybean growth.

Under our conditions, isolates positive only for the production of IAA, chitinase, or cyanide did not enhance soybean growth. The evidence that IAA, chitinase, or cyanide promotes plant growth in nonsterile soil is scanty. Abbass and Okon (1993) hypothesized that IAA and other plant hormones were responsible for increased growth of canola, tomato (Lycopersicon esculentum Mill.), and wheat (Triticum turgidum L.) in nonsterile soil inoculated with Azotobacter paspali, but they did not establish that A. paspali actually produced these hormones. Tien et al. (1979) did establish that Azospirillum brasilense produced IAA, gibberellins, and cytokinins and that this bacterium could increase the number of lateral roots and root hairs in pearl millet [Pennisetum glaucum (L.) R. Br.], but this was only under gnotobiotic conditions. There are some reports that rhizobacteria that overproduce IAA inhibit root elongation; this is attributed to the stimulation of ethylene synthesis by IAA (Xie et al., 1996; Glick et al., 1998). Although chitinase has been proposed as a mechanism of fungal antagonism (Potgieter and Alexander, 1966; Renwick et al., 1991; Kurek and Jaroszuk, 1997), this has only been tested in vitro. Chitinase may have been involved in increased nodulation and shoot dry weight of soybean when the seeds were coated with chitin and coinoculated with Streptomyces griseus and antibiotic-resistant B. japonicum, but the authors attributed the PGPR mechanism to antibiotic production (Li and Alexander, 1988). Overproduction of cyanide may control fungal diseases in wheat seedlings, but the experiments were only done in vitro (Flaishman et al., 1996).

It was unclear if asymbiotic N2 fixation was a PGPR trait in our experiments. The nitrogenase activity of our five N2-fixing isolates was very low when compared with the nitrogenase activity of the positive control, A. brasilense ATCC 29145. It may be that high nitrogenase activity is necessary for good plant growth promotion. However, low nitrogenase activity is not unusual for some N2-fixing bacteria (Lifshitz et al., 1986; Al-Mallah et al., 1989).

Surprisingly, the bacterial control LN1116, an isolate that did not share any of the in vitro PGPR traits tested except antagonism against S. rolfsii and S. sclerotiorum, also promoted at least two aspects of soybean growth. Why this isolate is also a PGPR is unclear. There may be some other PGPR trait(s) for which we did not test (e.g., gibberellic acid production). The PGPR control, P. putida PH6, was also negative for most of the in vitro traits tested, including siderophore production. This isolate has been selected as a PGPR for soybean in part because it is believed to produce siderophores (Fuhrmann and Wollum, 1989a, 1989b). Our inability to show siderophore production for P. putida PH6 may be because we used 0.1X TSB and the original authors (Fuhrmann and Wollum, 1989a, 1989b) used King's medium B (King et al., 1954). King's medium B is designed specifically for enhancing pigment production by pseudomonads. Pseudomonas putida PH6 and GN1212 increased the shoot P content in the P-solubilization experiment, but this seemed to be a concentration effect because the two isolates yielded the lowest shoot weight and P. putida PH6 did not solubilize P in vitro.

None of the nine isolates tested increased nodule occupancy by B. japonicum USDA 110. While others have observed beneficial changes in nodule occupancy with B. japonicum USDA 110 over other, less efficient bradyrhizobial strains in potting mixes (Polonenko et al., 1987; Fuhrmann and Wollum, 1989b), such changes have not been observed in nonsterile soil. One isolate, Acinetobacter baumannii LC3116, did increase the percentage of nodules containing strains of the serogroup B. elkanii USDA 31, but these bradyrhizobial strains are considered inefficient (Israel, 1981). It is important to note that B. elkanii USDA 31 is a poor competitor when coinoculated with B. japonicum USDA 110 (Fuhrmann and Wollum, 1986b; Weiser et al., 1990), and understanding how a poor competitor can become a strong one may be important for understanding rhizosphere competition. None of the isolates, including LC3116, significantly affected nodule occupancy with B. japonicum USDA 110.

All 23 isolates inhibited S. rolfsii and S. sclerotiorum in vitro when the bacteria and fungi were not physically separated. This was somewhat surprising for S. rolfsii because in a similar study (Cattelan, 1994), only six of 32 rhizosphere isolates (19%) inhibited this fungus. The inhibition we observed for F. oxysporum is similar to that observed by Kurek and Jaroszuk (1997). Because most of the fungal antagonism in vitro was similar between low- and high-Fe media for eight of 10 isolates, it is unlikely that siderophore production is involved in this antagonism. When the isolates were physically separated and they inhibited one or more of the fungi tested, the most likely mechanism is the production of one or more volatile compounds (Gagné et al., 1991; Flaishman et al., 1996). Based on our results, it is unclear whether fungal inhibition in vitro was associated with promoting soybean growth. In our plant test, only one of 10 isolates positive for Fusarium inhibition promoted soybean growth in nonsterile soil inoculated with F. oxysporum.

The presence of any of the PGPR in vitro traits tested here does not guarantee that a particular isolate is a PGPR nor does their absence guarantee that it is not. However, given that six of eight isolates positive for ACC deaminase, four of seven isolates positive for siderophore production, three of four isolates positive for ß-1,3-glucanase production, and two of five isolates positive for P solubilization in vitro increased some aspect of early soybean growth in nonsterile soil, our results suggest that these traits are more worthy of screening for plant growth promotion than isolates positive for biotin prototrophy or production of IAA, chitinase, and cyanide. Finally, there are other traits for which we did not test (e.g., gibberellic acid production) that may be associated with PGPR. Further screening is necessary to establish traits definitively associated with PGPR, especially in nonsterile soil.


   ACKNOWLEDGMENTS
 
We thank J.M. Bain, H.R. Boerma, S. Brooks, M.L. Cabrera, O. Finlay-Moore, C. Golt, J.W. Gray, T. Olexa, L.H. Pratt, J.A. Rema, N. Stern, and D.A. Zuberer for their technical assistance. Alexandre Cattelan was supported by a scholarship from the Brazilian Agricultural Research Corporation (EMBRAPA) and the Inter-American Development Bank (BID). The research was supported by EMBRAPA, BID, and Regional Research funds to Project S-262.

Received for publication June 19, 1998.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
 Discussion
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (37)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cattelan, A.J.
Right arrow Articles by Fuhrmann, J.J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Cattelan, A.J.
Right arrow Articles by Fuhrmann, J.J.
Agricola
Right arrow Articles by Cattelan, A.J.
Right arrow Articles by Fuhrmann, J.J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome