Microb Ecol (1995) 30:3-24

MICROBIAL ECOLOGY © 1995Springer-VerlagNewYorkInc.

Polyphasic Characterization of a Suite of Bacterial Isolates Capable of Degrading 2,4-D N.L. Tonso, ~ V.G. Matheson, 2 W.E. Holben, ~* ~Center for Microbial Ecology, Michigan State University, East Lansing, Michigan 48824, USA 2Department of Plant Pathology and Environmental Microbiology, West Virginia University, Morgantown, West Virginia 26506, USA Received: 25 April, 1994; Revised: 17 October, 1994

Abstract. To develop a better understanding of the ecological aspects of microbial biodegradation, it is important to assess the phenotypic and biochemical diversity of xenobiotic degrading organisms. Forty-six bacterial isolates capable of degrading 2,4-dichlorophenoxyacetic acid (2,4-D) and representing several geographically distinct locations were characterized and placed into taxonomic groups based on the results of several independent analyses. The isolates were characterized based on Gram's reaction, colony morphology, cell morphology, fatty acid methyl ester (FAME) fingerprints, carbon substrate oxidation patterns (BIOLOG), DNA homology to whole-plasmid probes and repetitive extragenic palindromic (REP) fingerprints. Attempts to group organisms taxonomically based on colony morphology and cell morphology were largely unsuccessful. Both FAME and BIOLOG analyses were generally unable to provide reliable genus or species identifications of these environmental isolates by comparison of fingerprints or substrate use patterns to existing data bases. Modification of the standard protocols for these analyses, however, allowed taxonomic grouping of the isolates and the construction of new data bases, comprised solely of 2,4-D-degrading organisms, against which future novel isolates can be compared. Independent cluster analysis of the FAME and BIOLOG data shows that the isolates can be segregated into five taxonomic classes. The collection of 2,4-D-degrading isolates was also separated into five classes based on DNA homology to whole-plasmid probes obtained from individual isolates. REP analysis allowed isolates that likely represent the same (or very similar) organism(s) to be identified and grouped. Each of the analyses used represents a mechanistically different means of classifying organisms, yet the taxonomic groupings obtained by several of the methods (FAME, BIOLOG, DNA homology, and to some degree, REP analysis) were in good

*Present address: The Agouron Institute, Department of Environmental Microbiology, 505 Coast Boulevard South, La Jolla, California 92037 USA. Correspondence to: W.E. Holben.

4

N.L. Tonsoet al. agreement. This indicates that the features discriminated by these different methods represent fundamental characteristics that determine phylogenetic groups of bacteria.

Introduction

The herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) has been widely used in agricultural and turf grass applications for more than 40 years. This compound has been shown to be degraded by microorganisms in numerous environmental samples [16, 34, 49]. To date, only one 2,4-D-degrading bacterial species (Alcaligenes eutrophus strain JMP134) has been studied in detail [14, 31, 41, 45]. Other genera of 2,4-D-degrading bacteria (and potentially other biochemical pathways for degradation) have been identified [2, 4, 10, 12, 48]. More recently, gene probe-based approaches have been developed and used with bacterial isolates and communities to detect, enumerate, and monitor indigenous 2,4-D-degrading bacterial populations in soils [8, 27-30]. These studies have indicated that substantial genetic diversity exists in genes encoding 2,4-D degradation. Still, the extent of phylogenetic and biochemical diversity that exists for this trait is not yet clear. For this study, a collection of forty-six 2,4-D-degrading isolates representing geographically separated environments and, apparently, several phylogenetic groups was assembled for diversity analyses. ". Determining the identity of each of a large group of novel isolates to the genus and species level can be both tedious and expensive using traditional biochemical assays and phenotypic characteristics. More recently, rapid bacterial identification strategies have been developed and marketed commercially. Among these, bacterial identification systems based on fatty acid methyl ester (FAME) fingerprints [43] and marketed by Microbial ID, Inc. (Newark, Del.), and carbon substrate oxidation patterns [6] marketed as the BIOLOG system by the Biolog Inc. (Hayward, Calif.), have received widespread attention. Both of these approaches have been successfully used for the identification of clinically important isolates [3, 7, 9, 23, 35, 36, 40]. FAME and BIOLOG have also been used with environmental bacterial isolates [2, 15, 17, 18, 32, 33, 37]. Perhaps because the culture collections used to develop the data bases on which these approaches rely for bacterial identification are heavily skewed toward clinical (vs. environmental) isolates, these rapid bacterial identification systems are generally less successful when used for environmental isolates [3, 32]. Recently, the potential for classification and identification of bacterial isolates based on fingerprints generated by polymerase chain reaction (PCR) amplification of repetitive DNA sequences present in all eubacterial genomes has been demonstrated [11, 50]. Plasmid-based analyses have previously been used to identify and characterize species or strains within a genus [21, 38, 47] but have not generally been used to characterize functional groups of isolates comprised of multiple taxonomic groups. No general data base specifically intended for the identification of environmental bacterial isolates yet exists for any of these approaches. Nonetheless, there has been increasing interest in the study and identification of environmental isolates, especially those encoding functions with potential application in the areas of biotechnology and bioremediation. Thus, it becomes necessary to develop new

Polyphasic Characterization of Bacterial Isolates

5

approaches that allow rapid screening and classification of relatively large numbers of isolates into useful taxonomic groups to reliably assess diversity and avoid redundancy of effort with unidentified but identical organisms. This allows screening of large numbers of isolates to identify those that warrant more detailed study and species identification. This study compiles data from several independent analyses and compares their ability to describe and classify a large group of 2,4-D-degrading bacterial isolates. Analyses based on Gram's reaction, colony and cell morphology, FAME fingerprints, carbon substrate oxidation (BIOLOG) pattems, DNA homology to wholeplasmid probes, and repetitive extragenic palindromic (REP) fingerprints were performed on the isolates. Modifications of the prescribed protocols for the preparation of bacterial cells for FAME and BIOLOG analysis that produced more reliable results with this collection of environmental isolates are also decribed. Using this polyphasic approach, it was possible to assess the diversity and redundancy that exists in the collection of isolates and to delineate five taxonomic classes of 2,4-D-degrading bacteria. Based on the results of this study, representatives of each of these groups were chosen as candidates for further characterization and detailed molecular analysis of their 2,4-D-degradative pathways. FAME or BIOLOG analysis in combination with REP analysis provides the ability to group similar organisms and to identify identical organisms in a collection of unknown bacterial isolates. Materials and Methods

Origin of Bacterial Isolates A total of 46 bacterial isolates capable of using 2,4,D as the sole source of carbon and energy were used in these experiments. Many of the isolates were obtained from soil samples from 2,4-D-treated, and untreated control research plots at the Kellogg Biological Station, Hickory Comers, Michigan, and the Experimental Farm, Indian Head, Saskatchewan, Canada. Isolates TFD3 (RASC), TFD17 (EML157), TFD19 (EML159), TFD26 (EML146), TFD32 (EML148), and TFD40 (EML157), isolated from water and sludge samples in Oregon, have been previously described [2] and were kindly provided by Penny Amy. Isolate TFD43 (JMP134) has been described [13, 14] and was obtained from R. Olsen. Isolate TFD44 was isolated from an industrial waste stream in Washington State and was provided by R. Sanford. A complete listing of the isolates and their origin is given in Table 1.

MeSa Bacterial isolates were grown on either 10% PTYG, which contains (per liter) 0.5g glucose, 0.5 g yeast extract, 0.25 g peptone,0.25 g tryptone, 3.5 mg CaC12, and 30 mg MgSO4; 10% PTYG supplemented with 500 ppm 2,4-D (from a 20-mg/ml stock in 0.1 M NaH2PO4, pH 7.0); or MMO basal salts medium (44) supplemented with 500 ppm 2,4-D. All bacterial cultures were incubated at 25°C. Solid media contained 1.5% Bacto agar (Difco Laboratories, Detroit, Mich.). Liquid cultures were incubated with shaking at 200 rpm on a rotary shaker.

Purification of Isolates and Verification of 2,4-D Degradation Where necessary, putative 2,4-D-degrading bacterial isolates were verified as pure cultures capable of metabolizing 2,4-D as follows: Bacterial cultures in MMO 2,4-D broth were streaked for single-

6 Table 1.

N.L. Tonso et al. Source and morphological characteristics of 2,4-D-degrading bacterial isolates Morphology

Isolate

Source

TFDlf TFD2f TFD3e TFD4f TFD5f TFD6f TFD7f TFDSf TFD9f TFD10f TFDllf TFD12f TFD13s TFD14f TFD15f TFD16f TFD 17f TFD 18 f TFD 19f TFD20f TFD21f TFD22f TFD23s TFD24~ TFD25f TFD26f TFD27s TFD28~ TFD29~ TFD30~ TFD3 l s TFD32f TFD33e TFD34f TFD35~ TFD36r

IH +a KBS -b OW ras d KBSKBSKBS +e K]3SKBSIH-S KBS + IHKBS + KBS + KBS + KBS + KBS + OR water IH + OR ras KBS + KBS + KBS + KBS + IH + KBS + OR river KBS + IH + IHIH + IHOR creek KBSKBS + KBS + KBS-

Gram Stain + ---+ --

Cell 1 3 3 3 2 2 5 2 8 2 1

Colony PTYG/MMO D/e D/c D/a G/e D/e L/c C/a E/e K/e K/g D/e

7

J/c

3 11 4 4 2 4 2 2 4 3 9 9 4 3 2 11 8 10 12 3 6 8 2 4

H/f Old D/c Q/e N/e D/a M/b N/h L/c D/e R/d T/e O/e M/b l/a X/e V/i V/e S/a M/b E/a M/e P/b C/a

colony isolation on rich medium (10% PTYG agar plates) to confirm that a single colony type was present. Single colonies from the 10% PTYG plates were than streaked on minimal medium (MMO 2,4-D agar) to confirm that single colonies could form with 2,4-D as sole carbon source. Individual colonies were then picked and transferred to MMO 2,4-D broth and incubated as described above. The spent MMO 2,4-D broth medium was analyzed by high-performance liquid chromatography (HPLC), as described [27] to verify the disappearance of 2,4-D. This HPLC analysis detects 2,4-D and degradation intermediates to the point of cleavage of the carbon ring and thus served to eliminate organisms capable of autotrophic growth, oligotrophic growth on trace contaminants in the medium, or growth on the glyoxylate side chain of 2,4-D (without ring cleavage). Permanent stocks of confirmed 2,4-D degraders were prepared from fresh cultures in MMO 2,4-D broth (by bringing to 15% glycerol) and stored at - 7 0 ° C for future analyses.

Polyphasic Characterization of Bacterial Isolates Table 1. (continued) Morphology Isolate

Source

Gram Stain

Cell

Colony PTYG/MMO

TFD37~ TFD38f TFD39f TFD40s TFD41s TFD42s TFD43f TFD44f TFD45s TFD46f

IH + KBSIH ÷ OR ras KBSIHAustralia Wash. stateg IHIH

-

10 2 2 7 6 5 2 3 2 8

T/i E/a W/k U/k F/d F/f A/a B/b R/e F/d

qH+; Indian Head Research Station soil, Saskatchewan, Canada; had prior 2,4-D treatment. bKBS-; Kellogg Biological Station soil, Hickory Comers, Mich.; no prior 2,4-D treatment. cOR; Corvallis, Ore. dras; return activated sludge. ~KBS+; Kellogg Biological Station soil, Hickory Comers, Mich.; had prior 2,4-D treatment. qH-; Indian Head Research Station soil, Saskatchewan, Canada; no prior 2,4-D treatment. gIsolated from an industrial waste stream in Washington state. fFast grower sSlow grower Key to morphotypes: Cell morphology: 1, large (lg), pinched rods; 2, medium (med) rods; 3, small (sm), stubby rods; 4, sm rods; 5, sm-med curved rods; 6, sm rods and filaments; 7, very sm stubby rods; 8, lg rods; 9, lg, fat rods; 10, lg rods, end-to-end; 11, med squarish rods, end-to-end; 12, pleomorphic rods. Colony morphology--PTYG: A, med-lg, white, spreading, wavy; B, med, bright yellow, round, entire, hilly, glossy; C, med, white, round, entire; D, lg, white, mucoidal, round, entire, biphasic; E, lg, white, round, wavy, radiating; F, lg, whitish-yellow, round, entire; G, lg, white, round, entire, flat; H, lg, bright yellow, whitish edge, round, entire; I, lg, whitish-yellow, entire, round, nippled; J, med-lg, white, round, entire; K, lg, yellow, clear, hilly, round, wavy; L, reed, white, round, entire, biphasic; M, sm-med, yellow, round, entire, biphasic; N, med, opalescent, round, wavy; O, med, whitish-yellow, round, entire concentric; P, med, yellow, round, wavy, fried-egg; Q, lg, opalescent, round, entire, biphasic; R, reed, white, irregular, wavy; S, med, yellow, round, entire; T, sin, whitishyellow, clear, round, wavy, speckled; U, lg, white, irregular, spreading, lobate, wavy; V, sin, white, clear, wrinkled, round, wavy; W, med-lg, yellow, clear, irregular; X, sm, white, round, entire. Colony morphologymMMO: a, reed, white, round, entire; b, sm, yellow, round, entire; c, sm-med, white, round, entire, biphasic; d, reed, whitish-yellow, round, entire; e, sm, white, round, entire; f, sm, yellow, round, entire, biphasic; g, lg, colorless, wavy; h, sm, white, round, wavy; i, sm, white, round, wavy; j, reed, clearish-white, round, lobate; k, sin, clearish-white, round, wavy.

Gram-Staining Reaction Gram-staining reactions were performed on fresh cultures in 10% PTYG broth medium incubated at 25°C with shaking. Cells were collected by centrifugation, resuspended in distilled water, then stained and analyzed as described [19] using a Leitz Orthoplan microscope (E. Leitz, Inc., Rockleigh, N.J.).

Cell and Colony Morphology Determination Cell morphology determinations were performed on fresh cultures of cells grown in 10% PTYG broth medium. Cells were collected by centrifugation and resuspended in distilled water. Cell morphology

8

N.L. Tonso et al. STREAK ON 1096 PTYG AGAR, INCUBATE AT 25"C FOR 72 h

CHECK GROWTH

INOCULATE I 0 0 m l 10% PTYG,

RESUSPEND IN MMO SALTS, STARVE 1 h

HARVEST CELLS FOR FAME ANALYSIS HARVEST CELLS, WASH WITH MMO SALTS

INCUBATE AT 25°C FOR 72 h

/\

RESUSPEND IN MMO SALTS, STARVE 1 h

FAME

INOCULATE BIOLOG PLATE, INCUBATE AT 25"C FOR 7 DAYS

ANALYSIS

Fig. 1 Strategy for growth and preparation of bacterial cultures for FAME and BIOLOG analysis illustrating the modifications to the manufacturer's recommended protocols for these bacterial identification systems.

was determined according to Harold [22] by phase-contrast microscopy using a Leitz Orthoplan microscope. Colony morphology was determined for colonies on both 10% PTYG agar and MMO 2,4-D agar after incubation at 25°C for 3 and 7 days, respectively. Colony morphological types are according to Harold [22].

FAME and BIOLOG Analysis Modifications to the prescribed protocols for preparing bacterial cells for FAME [39] and BIOLOG [5] analysis are shown schematically in Fig.1. FAME and BIOLOG analyses were performed at least in triplicate for each isolate from independent cultures as follows: Single colonies of each isolate (on 10% PTYG agar) were picked and streaked onto fresh 10% PTYG plates at a density sufficient to form a confluent lawn of cells in the secondary streak, then incubated at 25°C for 72 h. The plates were examined to assess whether each isolate had sufficient growth for FAME analysis, which requires 20 mg of cells per reaction for half-volume reactions [39]. ff sufficient biomass was present on each of the replicate plates (hereafter called fast growers), a small amount of ceils (from each plate) was transferred with a sterile, cotton-tipped swab to 20 ml of MMO basal salts medium (with no 2,4-D), starved for 1 h at room temperature, then used to inoculate separate BIOLOG GN plates as described [5]. The remaining biomass from each plate was harvested with a sterile inoculating loop and used for FAME analysis with half-volume reactions [39]. Cells from plates having insufficient biomass (i.e., slow growers) were suspended in 10 ml of 10% PTYG and used to inoculate 200 ml of 10% PTYG broth, then incubated for 7 days at 25°C with shaking. The cells were collected by centrifugation, washed with 15 ml of MMO basal salts medium, and again pelleted by centrifugation. A small amount of each pellet was transferred with a sterile cotton-tipped swab into 20 ml of MMO basal salts, starved, and used for BIOLOG analysis as above. The remainder of each pellet was resuspended and used for FAME analysis with half-reagents as

Polyphasic Characterization of Bacterial Isolates

9

above. Thus, FAME and BIOLOG analysis were performed on cells harvested from plates (for fast growers) and from broth cultures (for slow growers). All BIOLOG reaction plates were incubated at 25°C and analyzed, using the BIOLOG MicroStation (Biolog, Inc.) after 72 h (for fast growers) or 7 days (for slow growers). All FAME analyses were performed using the Microbial Identification System Software. Cluster analysis of data and generation of dendrograms for FAME and BIOLOG were performed using the manufacturer's software and recommendations.

DNA Homology to Whole-Plasmid Probes The homology of the total DNA of each isolate to various wbole-plasmid probes was determined by colony hybridization, essentially as described [42]. Briefly, broth cultures of each isolate in 10% PTYG supplemented with 250 ppm 2,4-D (10% PTYG/2,4-D) were used to inoculate several replicate sterile, gridded, nitrocellulose filters (Schleicher and Scbuell, Keene, N.H.) resting on the surface of 10% PTYG/2,4-D agar plates. The plates were incubated at 30°C until growth of colonies was observed at each inoculated position. The colonies were lysed by sequentially transferring the nitrocellulose fitlers (colony side up) to Whatman No. 1 filter paper soaked with 10% SDS; then 1.5 ra NaC1, 0.5 M NaOH; then 1.5 M NaC1, 0.5 M Tris (pH 7.4); then 2 × SSC [42] for 5 rain each. The liberated DNA was fixed to the nitrocellulose filters using an ultraviolet (UV) cross-linker (Stratagene Cloning Systems, La Jolla, Calif.) and the manufacturer's recommendations for exposure. Whole-plasmid probes were prepared from certain isolates (described elsewhere) as follows: Supercoiled plasmid DNA was purified by the alkaline lysis method of Hirsch et al. [24]. Purified plasmid was digested with HindIII (Gibco-BRL, Gaithersburg, Md.) according to the manufacturer's specifications, then labeled with 32p-dCTP (ICN Biomedicals, Irvine, Calif.) using a nick translation kit (Gibco-BRL) and the manufacturer's specifications. Prehybridization, hybridization, and washing conditions were as described previously [26], resulting in a stringency requiring approximately 95% homology between probe and target molecules for hybridization to occur. The washed hybridization filters were exposed to radiographic film as described [26] to obtain autoradiograms.

REP Fingerprint Analysis The primers (REP1R-I and REP2-I) and the PCR protocol used for REP analysis were as described [11], except that an inoculating loopful of the permanent glycerol stock of each isolate (rather than purified chromosomal DNA) was used to initiate the reactions. After the PCR reactions, 10-~1 subsamples of the REP PCR products were size fractionated on 0.7% agarose gels and stained with ethidium bromide, as described [42], then photographed using Polaroid type 55 film. The photographic negatives were analyzed using an AMBIS 9000 Optical Imaging System (Scanalytics, Billerica, Mass.). Cluster analysis of the data was performed using the simple matching algorithm and UPGMA (unweighted pair group method with arithmatic mean) provided in the AMBIS MicroPM version 2.12 software supplied with the instrument. The results of the cluster analysis and visual inspection of the photographs were used to determine identical and related fingerprints.

Results

Gram-Staining Reaction I s o l a t e s T F D 1 and T F D l l s t a i n e d g r a m - p o s i t i v e (Table 1). T h e r e m a i n i n g 4 4 i s o l a t e s stained g r a m - n e g a t i v e .

10

N.L. Tonso et al.

Cell and Colony Morphology Cell and colony morphologies of the isolates are given in Table 1. The morphology of the cells of most isolates was bacilloid, although some (e.g., TFD41 and TFD33) exhibited a biphasic cell shape, and one isolate (TFD31) was pleomorphic. Generally, cell morphologies were not sufficiently distinct to sort the collection of 2,4-D-degrading organisms into useful taxonomic groups representing the apparent phylogenetic diversity of the collection as determined by the other analyses. The colony morphology of each isolate was determined on both rich (10% PTYG) and minimal (MMO 2,4-D) medium. More distinct morphotypes were obtained on rich medium (23 morphotypes) than on minimal medium (11 morphotypes). As will be shown below, grouping of the isolates based on colony morphotype was in poor agreement with taxonomic groupings obtained from the other analyses.

FAME and BIOLOG Analyses Reliable genus and species identifications of these environmental isolates using FAME and BIOLOG analyses according to the manufacturer's suggested protocols for growth and preparation of cell biomass were generally not obtained. Only three of the 46 isolates analyzed were reliably identified to the species level by FAME (85% confidence interval), whereas no isolates were identified to the species level by BIOLOG (85% confidence interval). In most cases, the species designations obtained by FAME and BIOLOG for a particular isolate were different (data not shown). Often, the replicate samples for individual isolates generated different identifications of low reliability with both FAME and BIOLOG. Indeed, reproducible primary data (i.e., FAME fingerprints and BIOLOG carbon substrate oxidation patterns) with replicate samples of each isolate were difficult to obtain with the manufacturer's suggested protocols. Presumably, these difficulties reflect inherent features in the design of protocols and reference data bases primarily used for the identification of clinical isolates. The environmental isolates used in these experiments tended to grow poorly on the medium recommended for these analyses by the manufacturers tryptic soy broth (TSB) and generally did not achieve sufficient growth in the recommended incubation time to provide adequate biomass for the analyses. Several modifications to the recommended protocols for growth and preparation of bacterial biomass for FAME and BIOLOG analysis were made (Fig. 1). Most notably, 10% strength PTYG medium was used to culture cells because it was found (by screening several types) that this medium produced the most robust growth for the majority of the isolates to be tested. Because the BIOLOG system detection method is based on a general assay for cellular respiration, the 1-h starvation period was introduced to deplete cellular carbon reserves before inoculation of the BIOLOG assay wells containing specific carbon substrates. It was also determined that stable positive or negative reactions in the BIOLOG analyses were obtained only after 72 h (for fast growers) or 7 days (for slow growers) incubation at 25°C. Shorter or variable incubation times were deemed a contributing factor to the variability observed in the data for replicate samples. By employing these modifications, we obtained reliable FAME fingerprints and BIOLOG substrate use patterns for the replicate analyses of each isolate (data not shown).

Polyphasic Characterization of Bacterial Isolates

11 Slow-Growers

Fast-Growers

Euclidian Distance 10.63 31.88 i i 0~00 21°.25

Euclidian Distance

31.88 lO.63 [ 42~50 I 21.25 , I O.~l~

I

YFD3 rFDI9 TFD5 TFD8 TFD2 TFD36 TFD4 TFD7 TFDI4 TFDI8 TFD34 TFDI5 TFDI7 TFDI6 TFD21 TFD6 TFD22 TFD20 TFD25

Class

3

Class

2

Class

5

Classl

TFD35 TFD23 TFD28 TFD30 TFD37 TFD24 TFD31 TFD42 TFD45 TFD41• TFD29

Class 4

II~l~ ~.._]

, I 42.50

I

Fig. 2. Dendrograrns indicating phylogenetic relatedness and taxonomic class designations for fast- and slow-growing TFD isolates based on FAME data using MIDI cluster analysis software. Isolates that grouped together below a Euclidian distance of 25.00 we1~egiven the same class designation. The dendrograms represent the results of triplicate analysis of each isolate.

Because environmental isolates, in general, are poorly represented in the FAME and BIOLOG data bases, and comparisons of FAME fingerprints or BIOLOG substrate use patterns obtained from different media are probably not valid, new data bases (or libraries) were created for the 2,4-D-degrading isolates used in this study. The FAME and BIOLOG analyses both resulted in the TFD isolates being grouped into five taxonomic classes (Figs. 2, 3). To determine the relatedness between the taxonomic groups of slow- and fast-growing organisms, and to calibrate the classes obtained for each, several fast growers were analyzed by both FAME and BIOLOG using the protocol for slow growers, with the data subjected to cluster analysis using the software provided with each system. The results indicated that the two taxonomic groups designated for the slow growers by both FAME and BIOLOG correspond to classes 1 and 4 of the fast-growing organisms (Figs. 2, 3; some data not shown). Although it is probably reasonable, based on these results, to cluster the data for broth- and plate-grown isolates together, we chose to present separate dendrograms because it is possible that the two different methods of cultivation resulted in fingerprints or substrate use patterns that were not directly comparable (i.e., were not free of cultivation-induced artifacts). The FAME and BIOLOG analyses each produced five nearly identical classes

12

0~

N.L. Tonso et al. Fast-Growers

Slow-Growers

Similarity Index

Similarity Index

o.so

o.~5

1~o

~

110

0.;5

0~0

0.25

Class3 Class2

Class

1

Class

4

®

Fig. 3. Dendrogramsindicatingphylogenetic relatedness and taxonomicclass designationsfor fastand slow-growingTFD isolates based on BIOLOGdata using BIOLOGMicroStation Release 3.00 software. Isolates that grouped together above a similarityindex of 0.70 were given the same class designation.The Class5 dendrograms represent the results of triplicate analysis of each isolate. TM

~ ' ~

or subgroups of isolates (see summary, Table 2). The relative position of individual isolates within the five classes may vary when FAME and BIOLOG results are compared (Figs. 2 and 3), but in only three cases (of 42 isolates analyzed by both FAME and BIOLOG) were individual isolates placed in different classes by these alternate approaches (Table 2). It is interesting to note that all three of these exceptions group together by each of the methods (i.e., class 1 by FAME and class 4 by BIOLOG). We found no correlation between taxonomic groups (i.e., classes 1 to 5) and the geographic location and 2,4-D-treatment history of the sites from which the isolates were obtained (Table 2).

DNA Homology Analyses Naturally occurring plasmids purified from several of the isolates were labeled with 32p and used as probes against the total DNA of the isolates in the collection. Hybridization analyses were performed sequentially as follows: Total genomic DNA from the isolates was first probed with the pJP4 plasmid from A. eutrophus JMP134 (TFD43 in this collection), the canonical 2,4-D degrader first described by Don and Pemberton [13]. This probe hybridized significantly to the DNA of

Table 2. Summary of classification of isolates based on FAME, BIOLOG, plasmid homology, and REP fingerprint analyses

Isolate a

Source b

TFD33f TFD38r TFD4 l s TFD24~ TFD30s TFD37s TFD39r TFD29~ TFD42~ TFD45~ TFD46f TFD12f TFD28~ TFD3 l s TFD35~ TFD43f TFD23~ TFD9f

KBSKBSKBS IH ÷ IH + IH + IH ÷ IHIHIHIHKBS + IH + IHKBS + Australia KBS + IHKBS + Oregon Wash. state KBS + KBS + KBS + KBS + KBS + Oregon IH + KBS + KBS ÷ KBS + KBS + KBS + KBS + KBSKBSKBSKBSKBSKBSOregon Oregon Oregon Oregon IH + IH-

TFD10f TFD32f

TFD44f TFD6f TFD13s TFD14f TFD15f TFD16f TFD 17f TFD 18f TFD20f TFD21f TFD22f TFD25f TFD27s TFD34f TFD2f

TFD4f TFD5f TFD7f

TFD8f TFD36f TFD3~ TFD 19f TFD26f TFD40~ TFDlf TFD 11f

FAME c class

BIOLOG d class

Plasmid" class

REPi class

1

1

ct

C

1

1

c~

C

1 1 1 1 1 1 1 1 ND g 1 1 1 1 1 1 ND 2 ND 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 ND 4 5 5

1 1 1 1 4 1 4 1 1 1 4 1 1 1 1 2 2 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5

tx a 8 ct e ct tx ~ a 13 a e e tx tx c~ e 7 ~/ [3 "y [3 13 ND [3 13 13 [3 [3 [3 [3 [3 ~ ~ 8 ~ 8 ~ ~ c~ e e ct ND

C e e f f g g g g Unique Unique Unique Unique Unique ND Unique Unique Unique Unique A A A A A A A A A A A A A B B B B B B Unique Unique Unique Unique D D

% fast grower; s, slow grower. hAs described in Table 1. "As in Fig. 2. dAs in Fig. 3. eAs in Fig. 4B. SAs in Fig. 6; uppercase letters designate groups composed of isolates having identical REP patterns, i.e., the same organism at the subspecies level; lowercase letters designate groups composed of isolates that have similar but not identical REP patterns, i.e., related organisms. Unique, isolates with unique REP pattems. gND, not done.

14

N.L. Tonso et al.

Fig. 4. Data from colony hybridization experiments using whole plasmid probes. Colonies of 2,4-D degrading bacterial isolates and strain DBO1, a negative control, were grown directly on gridded hybridization membranes, lysed and hybridized to whole-plasmidprobes. (A) Representative autoradiogram obtained after hybridization to 32p-labeled pJP4 plasmid (isolated from TFD43). (B) Template showing the positions of individual isolates on the hybridization membranes and the Homology Class designations determined by hybridization to plasmids isolated from TFD43 (c0, TFD6 ([3), TFD44 (~/), and TFD3 (~). Homology Class e is defined as those isolates having no homology to any of the plasmid probes used. Note that isolates not having a TFD designation were not further characterized in this study. 14 of the TFD isolates but had little or no homology to the remainder of the TFD isolates (Fig. 4A).* These are hereafter called homology class o~. Plasmid D N A was then purified from isolate TFD6 (which had no detectable homology to the pJP4 plasmid) and used as probe with another of the replicate filters. This probe hybridized strongly to the D N A of 12 of the TFD isolates, hereafter called homology class [3, but not to the D N A of any of the isolates in homology class ~x. Plasmid DNA purified from isolate TFD44 (which had not hybridized to either of the previous plasmids) was then used as probe with another of the replicate filters. This probe hybridized strongly to the D N A of three TFD isolates, hereafter called homology class ~, but not the D N A of any of the isolates in homology classes ~x or [3. Purified plasmid D N A from isolate TFD3 (not in homology class

Polyphasic characterization of a suite of bacterial isolates capable of degrading 2,4-D.

To develop a better understanding of the ecological aspects of microbial biodegradation, it is important to assess the phenotypic and biochemical dive...
2MB Sizes 0 Downloads 0 Views