Characterization of Hydrocarbon-Degrading Microbial Populations in Contaminated and Pristine Alpine Soils

Abstract

Biodegradation of petroleum hydrocarbons in cold environments, including Alpine soils, is a result of indigenous cold-adapted microorganisms able to degrade these contaminants. In the present study, the prevalence of seven genotypes involved in the degradation of n-alkanes (Pseudomonas putida GPo1 alkB; Acinetobacter spp. alkM; Rhodococcus spp. alkB1, and Rhodococcus spp. alkB2), aromatic hydrocarbons (P. putida xylE), and polycyclic aromatic hydrocarbons (P. putida ndoB and Mycobacterium sp. strain PYR-1 nidA) was determined in 12 oil-contaminated (428 to 30,644 mg of total petroleum hydrocarbons [TPH]/kg of soil) and 8 pristine Alpine soils from Tyrol (Austria) by PCR hybridization analyses of total soil community DNA, using oligonucleotide primers and DNA probes specific for each genotype. The soils investigated were also analyzed for various physical, chemical, and microbiological parameters, and statistical correlations between all parameters were determined. Genotypes containing genes from gram-negative bacteria (P. putida alkB, xylE, and ndoB and Acinetobacter alkM) were detected to a significantly higher percentage in the contaminated (50 to 75%) than in the pristine (0 to 12.5%) soils, indicating that these organisms had been enriched in soils following contamination. There was a highly significant positive correlation (P < 0.001) between the level of contamination and the number of genotypes containing genes from P. putida and Acinetobacter sp. but no significant correlation between the TPH content and the number of genotypes containing genes from gram-positive bacteria (Rhodococcus alkB1 and alkB2 and Mycobacterium nidA). These genotypes were detected at a high frequency in both contaminated (41.7 to 75%) and pristine (37.5 to 50%) soils, indicating that they are already present in substantial numbers before a contamination event. No correlation was found between the prevalence of hydrocarbon-degradative genotypes and biological activities (respiration, fluorescein diacetate hydrolysis, lipase activity) or numbers of culturable hydrocarbon-degrading soil microorganisms; there also was no correlation between the numbers of hydrocarbon degraders and the contamination level. The measured biological activities showed significant positive correlation with each other, with the organic matter content, and partially with the TPH content and a significant negative correlation with the soil dry-mass content (P < 0.05 to 0.001).

Petroleum hydrocarbons are the most widespread contaminants in the environment. Biodegradation of many components of petroleum hydrocarbons at low temperatures in Arctic (11, 47, 48), Alpine (24, 26), and Antarctic soils (2, 8) has been reported and is a result of the degradation capacity of indigenous cold-adapted microorganisms. Cold-adapted psychrophilic and psychrotrophic microorganisms are able to grow at temperatures around 0°C and have adapted their metabolism to function optimally at low temperatures. Psychrophiles have an optimum growth temperature of ≤15°C and do not grow above 20°C, whereas psychrotrophs (or psychrotolerant organisms) have optimum and maximum growth temperatures above 15 and 20°C, respectively (33). Cold-adapted microorganisms play a significant role in the in situ biodegradation of hydrocarbons in cold environments, where ambient summer temperatures often coincide with their growth temperature range. A large number of degrading bacteria from contaminated cold soils have been identified, including representatives of gram-negative and gram-positive genera (1, 8, 9, 17, 49, 50).

Various methods are used to characterize hydrocarbon-degrading populations in soil. Soil biological investigations, such as measurements of soil respiration, enzyme activities, and microbial counts, can give information about the presence of viable microorganisms and on the impact of the effects of environmental stresses, such as hydrocarbon contamination, on the metabolic activity of soil (7, 19, 28, 31, 39). Direct, non-cultivation-based molecular techniques for detecting microbial pollutant-degrading genes in environmental samples are also powerful tools for studying the structure and functions of complex microbial communities. Catabolic gene probes that are designed from specific genes involved in key enzymatic steps in the microbial degradation pathways for environmental pollutants can be used to examine both pristine and contaminated environments to determine the presence of organisms having specific functional capacities (6, 16).

There is little information about the prevalence and geographical distribution of various hydrocarbon-degrading populations in soils. Catabolic genotypes involved in the degradation of representative fractions of petroleum hydrocarbons, including n-alkanes, and aromatic and polycyclic aromatic hydrocarbons (PAHs), appear to be widespread in Arctic soils (46, 47) and Alaskan sediments (42). In a recent study of the prevalence of various alkane monooxygenase genes in Arctic and Antarctic soils, Rhodococcus spp. were shown to be the most abundant alkane-degradative genotypes in pristine and contaminated soils while Pseudomonas spp. may become enriched following contamination events and Acinetobacter spp. were not found to be predominant members of polar alkane-degrading microbial communities (51). However, there is no information available on catabolic genotypes in the Alpine environment. There are various differences between Antarctic, Arctic, and Alpine soils. Besides the different geographical locations, European Alpine soils are subjected to regular freeze-thaw events, to large temperature fluctuations (the air temperatures can vary from −5 to +20°C), and to high precipitation (2,000 to 3,000 mm per year). The Alpine microbial communities may differ from those in Arctic and Antarctic habitats because strong valley winds from boreal and Mediterranean landscapes continuously transport microorganisms to the Alps.

In the present study, the prevalence of seven genotypes involved in the degradation of n-alkanes, aromatic hydrocarbons, and PAHs in oil-contaminated and pristine European Alpine soils was determined by culture-independent analyses (PCR and hybridization analyses). The soils were also analyzed for various physical, chemical, and microbiological parameters, including culture-dependent microbial enumeration, and statistical correlations were determined among all parameters. This is the first molecular analysis of hydrocarbon-degrading microbial populations in Alpine soils and the first attempt to correlate these genotypes with other parameters.

MATERIALS AND METHODS

Soils.

Twelve oil-contaminated and eight pristine (uncontaminated) soils were sampled in September 2001 from various Alpine sites in Tyrol (Austria) at altitudes ranging from 500 to 2,900 m above sea level. Samples were collected from multiple areas within a site and mixed to produce composite samples. The contaminated soils were taken from areas near diesel oil storage tanks, petrol stations, petrol pumps, or garages. The corresponding pristine soils were collected from the same or comparable locations without oil pollution (Table 1). The soils were collected from the surface to a depth of about 0 to 5 cm with sterile spatulas, transported in coolers to the laboratory, and stored at −20°C until they were analyzed. Subsamples were stored at −80°C for molecular biological analysis.

TABLE 1.

Physical, chemical, and biological characteristics of the pristine and contaminated Alpine soils^a

Soil	Location (Tyrol, Austria)	Altitude (m)b	TPH (mg/kg)	Dry mass (%)	Organic matter (%)	pH	Respiration (μg of CO2/g/24 h)	FDA hydrolysis (μg of fluorescein/ g/2 h)	Lipase activity (μg of pNBPc g/10 min)
Pristine soils
P1	Ötztaler Alps, Rettenbach glacier (south)	2,800	83	92	1.7	4.8	191	48	320
P2	Ötztaler Alps, Rettenbach glacier (east)	2,800	41	81	0.2	5.9	108	41	97
P3	Stubaier Alps, Kühtai	2,000	74	65	11.2	5.4	168	103	766
P4	Stubaier Alps, Eisgrat glacier (north)	2,900	87	89	0.6	5.1	99	69	100
P5	Ötztaler Alps, Weisssee glacier	2,750	20	89	1.0	5.4	99	41	174
P6	Hohe Tauem, Grossglockner (south)	2,550	89	71	8.1	6.6	465	60	521
P7	Kitzbühler Alps, Reith im Alpbachtal	800	54	92	0.6	7.5	144	10	91
P8	Inntal, Innsbruck	600	29	83	6.1	8.1	75	51	196
Contaminated soils
C1	Kitzbühler Alps, Kirchberg	700	428	97	0.8	8.2	160	42	138
C2	Stubaier Alps, Eisgrat glacier	2,900	743	80	2.5	6.5	274	82	721
C3	Lechtaler Alps, Hahntennjoch	1,715	756	99	1.3	7.7	392	52	242
C4	Stubaier Alps, Kühtai	2,000	1,052	98	1.7	7.7	138	45	154
C5	Stubaier Alps, Eisgrat glacier	2,875	2,085	94	0.8	8.0	168	29	43
C6	Ötztaler Alps, Tiefenbach glacier (east)	2,780	3,148	80	1.5	9.2	138	31	68
C7	Hohe Tauem, Grossglockner (south)	2,140	3,317	59	14.0	6.7	1,447	224	1,022
C8	Ötztaler Alps, Weisssee glacier	2,750	8,385	87	2.0	7.5	127	26	45
C9	Inntal, Kirchbichl	500	13,903	93	1.2	7.9	212	33	173
C10	Inntal, Innsbruck	600	22,288	89	2.0	8.5	108	12	286
C11	Kitzbühler Alps, Kirchberg	700	22,447	96	3.4	7.4	251	73	388
C12	Zillertaler Alps, Patscherkofel	2,000	30,644	63	15.1	6.8	1,816	150	2,565

^aAll values are expressed on a dm basis of soil.

^bMeters above sea level.

^cpNBP, p-nitrophenyl butyrate.

Chemical and physical analysis.

Total petroleum hydrocarbons (TPH) were extracted from the soils with 1,1,2-trichlorotrifluoroethane and quantified by infrared spectroscopy (13). Soil dry mass (dm) was determined from the weight loss after heat treatment (20 h at 105°C). Loss on ignition as a measure of soil organic matter content was calculated from the weight loss of oven-dried soils after 3 h at 500°C (37). To determine the soil pH, 1 part of soil was mixed with 2.5 parts of 10 mM CaCl₂ and the pH was measured with a glass electrode after 2 h (39). All analyses were carried with two replicates, and the mean values obtained are reported.

Microbiological analysis.

Three replicates were used for all analyses, and the mean values obtained are reported. The standard deviations obtained were ≤10%.

(i) Respiration and enzyme activities.

Soil respiration (CO₂ evolution) was determined by the Isermeyer technique, where CO₂ produced during 24 h at 10°C was quantified by titration (39). To measure lipase activity, the p-nitrophenol released from p-nitrophenyl butyrate after 10 min at 30°C and pH 7.25 was quantified colorimetrically at 400 nm (27). Fluorescein diacetate (FDA) hydrolysis was determined using a method adapted (28) from that of Schnürer and Rosswall (40); the amount of fluorescein released from FDA after 2 h at 25°C and pH 7.6 was quantified at 490 nm.

(ii) Microbial counts.

Soil microbial counts (culturable microorganisms) were determined by the plate count method for viable cells. Soil suspensions were prepared by shaking soil corresponding to 5 g of dm with 45 ml of 0.28% sodium pyrophosphate for 30 min at 10°C and 150 rpm. Appropriate dilutions, prepared in 0.9% NaCl, were surface spread onto agar plates. R₂A-agar plates (36) were used to enumerate aerobic heterotrophic microorganisms. CFU of heterotrophs were counted after 28, 14, 7, and 3 days at 2, 10, 25, and 37°C, respectively. Hydrocarbon-degrading populations were quantified on agar plates that contained purified agar and a phosphate-buffered pH-neutral mineral salts medium (24) supplemented with yeast extract (10 mg liter⁻¹); 20 μl of the carbon source (diesel oil or hexadecane) was dropped onto a small piece of filter paper placed on the lid of the petri dish. Hydrocarbon degraders were enumerated after 14 and 6 days at 10 and 37°C, respectively.

Molecular characterization.

To determine the prevalence of various hydrocarbon-degrading genotypes in the hydrocarbon-contaminated and pristine Alpine soils, total community DNA was extracted from the soils and screened by PCR using oligonucleotide primer sets specific for each degrader genotype.

(i) Total community DNA extraction from soil and DNA purification.

The total community DNA of each soil sample was extracted using a method adapted (15) from that of Flemming et al. (14). Prior to lysis treatment, 1 g of soil was mixed with 950 μl of sterile distilled water. Then 50 μl of 250 mM Tris-HCl (pH 8.0) containing 5 mg of lysozyme was added, and the samples were incubated for 30 min at 30°C and than for 30 min at 37°C with mixing by inversion every 10 min. After the addition of 5 μl of proteinase K (20 mg ml⁻¹), the samples were incubated for 1 h at 37°C. The lysis treatment was completed with the addition of 5 μl of 20% sodium dodecyl sulfate and incubation for 30 min at 85°C. Samples were centrifuged (13,600 × g) for 10 min at room temperature. Supernatants were treated with 0.5 volume of 7.5 M ammonium acetate, incubated on ice for 15 min to precipitate proteins and humic acids, and centrifuged for 5 min at 4°C. The DNA was precipitated with 1 volume of isopropanol at −20°C overnight. The pellets were washed with cold 70% ethanol and dried by speed vacuum. Each DNA sample was resuspended in 200 μl of 10 mM Tris-HCl (pH 8.0)-1 mM EDTA. To obtain a high-quality PCR-amplifiable DNA, all samples were purified using polyvinylpolypyrrolidone spin colums to remove PCR-inhibitory compounds (10). To confirm that DNA had been successfully extracted from the soils, the soil DNA extracts were analyzed by agarose gel electrophoresis (0.7% agarose) for observable chromosomal DNA bands. To verify the absence of any PCR inhibitory compounds in the DNA preparations and to confirm that DNA could be successful amplified by PCR, 16S rDNA universal eubacterial primers (12) were used as a positive PCR amplification control on appropriate dilutions (1:10, 1:50, and 1:100) of each soil DNA extract.

(ii) Detection of catabolic genes by PCR and hybridization analysis.

The 20 purified DNA extracts were subsequently screened by PCR to detect the following seven catabolic genes that encode enzymes involved in a a variety of known bacterial hydrocarbon degradative pathways: alkB, alkane monooxygenase from Pseudomonas putida GPo1 ATCC 29347, formerly designed P. oleovorans, C₅ to C₁₂ alkane degradation (20, 44, 45); alkM, alkane monooxygenase from Acinetobacter sp. strain ADP-1, C₁₀ to C₂₀ alkane degradation (35); alkB1 and alkB2 (C₁₂ to C₁₆ alkane degradation), alkane monooxygenases from Rhodococcus spp. (52); xylE, catechol-2,3-dioxygenase from P. putida ATCC 33015, xylene and toluene degradation (34); ndoB, naphthalene dioxygenase from P. putida ATCC 17484, PAH (naphthalene) degradation (21); and nidA, pyrene dioxygenase large subunit from Mycobacterium sp. strain PYR-1, PAH (pyrene) degradation (18). The oligonucleotide primer sets specific for these genotypes have been described previously: alkB, alkM, alkB1, and alkB2 in reference 51, and xylE and ndoB in reference 49. For the nidA gene, oligonucleotide primers (forward primer, 5′-ATCTTCGGGCGCGCCTGGGTGTTTCTCGG 3′; reverse primer, 5′-AATTGTCGGCGGCTGTCTTCCAGTTCGC-3′) were derived from regions of high DNA sequence identity from four dioxygenase large-subunit gene sequences from two Rhodococcus spp. (GenBank accession numbers AF121905, AF082663), one Nocardioides sp. (ABO17794), and one Mycobacterium sp. (AF249301) and resulted in the amplification of 323-bp PCR fragments. Since the strain ADP-1 alkM primers may have been too specific to amplify Acinetobacter alkM homologues other than ADP-1 alkM from the source strain (51), universal Acinetobacter alkM primers (forward primer [universal alkM-F], 5′-CGIGIIGCIACICCTGAAGATCCAGC-3′; reverse primer [universal alkM-R], 5′-ITTATTITTCCAICTATGCTCTGG-3′) were derived from regions of high DNA sequence identity from seven alkM genes from six Acinetobacter strains (Acinetobacter sp. strain ADP-1 [35]; Acinetobacter strains EB104, 69-V, NRRLB-2769A, and NCIB 8250 [41]; and Acinetobacter sp. strain M1 [alkMa and alkM] [43]).

All PCR amplifications were carried out as previously described (49) for 30 cycles of 1 min of denaturation at 94°C, 1 min of annealing at 60°C, and 1 min of extension at 72°C, with a final extension of 3 min at 72°C. The PCR fragments were analyzed by agarose gel elecrophoresis (1.2% agarose) and visualized by ethidium bromide staining (38). To verify amplification of the correct PCR fragment, PCR fragments were transferred from the agarose gels to nylon membranes and analyzed by Southern hybridization with DNA probes specific for alkB, alkM, alkB1, alkB2, xylE, ndoB, and nidA, using high-stringency prehybridization, hybridization, and washing conditions at 65°C, essentially as previously described (51). The alkM fragments produced from both the strain ADP-1 alkM and the universal Acinetobacter alkM primer sets were probed with a 499-bp probe (51) derived from the strain ADP-1 alkM PCR fragment. The probes were labeled with the digoxigenin DIG nonradioactive nucleic acid labeling and detection system, using the DIG DNA labeling and detection kit (Roche Molecular Biochemicals, Laval, Quebec, Canada).

To determine the specificity and the utility of the two alkM primer sets, five Acinetobacter strains (ADP-1, EB104, 69-V, NRRLB-2769A, and NCIB 8250) and two negative control strains (P. putida GPo1 and Rhodococcus sp. strain Q15) were tested using both alkM primer sets by the PCR hybridization procedure. For the ADP-1 alkM primer set, the expected PCR fragment size only was amplified and a hybridization signal was obtained from the positive control strain ADP-1 but not the other four strains, indicating that it was specific for this organism. With the universal Acinetobacter alkM primer set, the expected 372-bp PCR fragments of relatively equal intensity were amplified from all five Acinetobacter strains tested but the hybridization signal was stronger for ADP-1 than for the other four strains: the 370-bp fragments amplified by the universal alkM primer set were internal to the original 472-bp strain ADP-1 alkM PCR fragment in all of the Acinetobacter strains tested; the DNA sequence indentity within the 370-bp region ranged from 67 to 100%. For the negative control strains, very weak amplification and hybridization signals were observed with the Rhodococcus strain but not with P. putida GPo1.

Statistical data analysis.

Statistical calculations were done using Statistica 6.0 software. Normal distribution of the data was tested by the Kolmogorov-Smirnov test. According to the lack (TPH content, organic matter content, and soil respiration) or presence (all other parameters investigated) of normal distribution correlations between the investigated soil characteristics were analyzed by Spearman rank order correlation or regression analysis (Pearson product-moment correlation).

RESULTS AND DISCUSSION

Chemical and physical analysis, respiration, and enzyme activities of soils.

Most of the soils investigated in this study were characterized by a low organic matter content (<2%); only three soils had a high organic matter content content (>10%). pH values ranged from 4.8 to 9.2. The 12 oil-contaminated soils had a TPH content ranging from 428 to 30,644 mg/kg of soil dm (Table 1). Oil contamination can result in significant stimulation of the physiological activities of soil microorganisms (7, 19, 28, 31).

The biological activities measured in this study (respiration in terms of CO₂ evolution, FDA hydrolysis, and lipase activity [Table 1]) showed a statistically significant positive correlation with each other and with the organic matter content and a significant negative correlation (except for soil respiration) with the soil dm content (Table 2). Significant positive correlations were also found between the contamination level and soil respiration, which points to adaptation of the indigenous soil microorganisms to the contamination and to the microbial utilization of hydrocarbons as an additional carbon source. Soil respiration is a measure of the total biological activity in soil and results from the degradation of organic matter, where the formation of CO₂ is the last step of carbon mineralization. Respirometric measurements provide information on the biodegradability potential of hydrocarbons in soils and are often used as a relevant parameter during bioremediation treatments. The increase of respiration after oil application indicates successful hydrocarbon mineralization (7, 22, 24, 28).

TABLE 2.

Correlation matrix between physical and chemical parameters and biological activities of the Alpine soils

Parameter	Correlation coefficient and significance (n = 20) fora:
	TPH	dm	Organic matter	pH	Respiration	FDA hydrolysis
dm	NS
Organic matter	NS	−0.547*
pH	NS	NS	NS
Respiration	0.466*	NS	0.496*	NS
FDA hydrolysis	NS	−0.749***	0.624**	NS	0.538*
Lipase activity	NS	−0.692**	0.777***	NS	0.611**	0.730***

^aNS, not significant; ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001.

The lack of correlation between the TPH content and the measured hydrolytic enzyme activities (FDA hydrolysis and lipase activity) in this study may be attributed to interfering factors, such as the composition of the contaminant components due to the age of the contamination and/or the presence of inhibiting compounds, which may influence enzyme activity in contaminated soils. The rate of FDA hydrolysis in soil has been considered a suitable index of overall enzyme activity because FDA is hydrolyzed by a number of different enzymes such as proteases, lipases, and esterases (40). In a study of the impact of hydrocarbons on soil enzyme activities, the presence of diesel oil resulted in significantly increased FDA hydrolysis and lipase activity compared with those in uncontaminated controls (28). However, other authors found distinctly lower rates of FDA hydrolysis in oil-contaminated soils than in pristine soils (32). Soil lipase activity is a valuable tool to monitor oil biodegradation in freshly diesel oil-contaminated soils (30), most probably due to a high content of available aliphatic compounds.

Soil microbial counts.

Although the enumeration of microbial populations is typically performed to gain information on the biodegradation potential of the hydrocarbons and/or to test bioremediation efficiency, it is associated with a number of difficulties both from the methodological point of view and from the interpretation of the results. Only a small fraction of microorganisms (<1 to 10%) can be isolated and cultivated on laboratory media since the growth requirements for many strains are unknown. For this reason, plate counts underestimate the true viable population density (3-5).

Enumeration of microorganisms in pristine and contaminated soils demonstrated that significant microbial populations were present in all of the soils as shown by viable counts of heterotrophs and oil degraders (Table 3). The numbers of culturable heterotrophs were determined at 2, 10, 25, and 37°C to gain information about the growth temperature range of the indigenous soil microorganisms. The greatest numbers of viable microorganisms were observed at 25°C in 15 of the 20 investigated soils and at 10°C in 5 of the soils (P4, P6, C6, C8, and C10). In none of the soils was 2 or 37°C the optimum growth temperature. However, in the majority of the soils, populations growing at 2 and 10°C were larger (by 1 to 3 orders of magnitude at 2°C and by 2 to 4 orders of magnitude at 10°C) than those growing at 37°C, indicating the predominance of cold-adapted microorganisms. The 10 and 25°C populations were always larger than the 2 and 37°C populations and indicated that the cold-adapted culturable bacteria from the Alpine soils were psychrotrophic rather than psychrophilic in nature, similar to those found in Arctic soils (47, 48).

TABLE 3.

Enumeration of viable microorganisms in the pristine and contaminated Alpine soils^a

Soil sample	TPH (mg/kg)	No. of heterotrophs (106 CFU/g) at:				No. of hydrocarbon degraders (106 CFU/g) for:
		2°C	10°C	25°C	37°C	Hexadecane, 10°C	Diesel oil, 10°C	Diesel oil, 37°C
Pristine soils
P1	83	1	2	3	0.2	1.3	0.1	0.1
P2	41	3	2	7	0.05	0.9	0.5	0.0001
P3	74	6	30	40	0.1	20	15	0.04
P4	87	7	13	11	<0.001	7	4	<0.0001
P5	20	2	2	4	0.05	2	2	0.02
P6	89	11	31	21	0.02	13	18	0.02
P7	54	2	10	38	0.7	4	4	0.01
P8	29	2	29	93	8	6	5	1.4
Contaminated soils
C1	428	0	3	15	0.3	2	2	0.02
C2	743	1	8	10	0.04	0.9	1	0.002
C3	756	5	28	58	11	12	10	0.04
C4	1,052	32	67	144	0.3	64	47	0.03
C5	2,085	3	54	90	0.1	3	2	0.0004
C6	3,148	2	7	6	0.02	1	1	0.0004
C7	3,317	51	268	708	4	93	99	0.3
C8	8,385	24	127	70	0.03	11	11	0.01
C9	13,903	2	15	139	0.1	1	11	0.01
C10	22,288	35	70	66	4	17	15	0.4
C11	22,447	2	10	93	0.6	2	2	0.4
C12	30,644	22	62	286	0.6	32	40	0.2

^aGrowth after 28, 14, 7, and 3 to 6 days at 2, 10, 25 and 37°C, respectively. All values are expressed on a dm of soil basis.

Numbers of culturable diesel oil degraders were determined at 10°C (cold adapted; psychrophiles and psychrotrophs) and 37°C (mesophiles). In Alpine soils, temperatures above 10°C are reached only during the period of high solar irradiation and on hot summer days. Cold-adapted microorganisms (psychrotrophs) grow rarely at ≥37°C, while mesophiles grow rarely at ≤10°C. The numbers of cold-adapted oil degraders were in general greater than the numbers of mesophilic degraders. In most of the soils, the numbers of cold-adapted hydrocarbon-utilizing microorganisms were greater by 2 to 3 orders of magnitude (10 soils [P3, P5, P6, P8, C1 to C3, C7, C8, and C12]) or even >3 to 4 orders of magnitude (5 soils [P2, P4, C5, C6, and C9]) than those of the corresponding mesophilic populations (Table 3). This points to the important role of cold-adapted microbial communities in the bioremediation of contaminated soils in Alpine habitats.

Calculation of the ratios of hydrocarbon degraders and heterotrophs in the investigated soils showed that a significant portion of heterotrophs were able to utilize hexadecane and diesel oil. In 17 soils, 13 to 92% of the cold-adapted (10°C) heterotrophs were hexadecane and oil degraders; in the residual 3 soils, 4 to 9% of the heterotrophic population still utilized hydrocarbons. Among mesophilic populations (37°C), the fraction of oil degraders among heterotrophs tended to be lower, with 11, 7, and 2 soils that contained 10 to 97%, 1 to 8%, and 0.3 to 0.4% oil degraders respectively (Table 3). The above results were independent of the contamination level of the soils. This can be explained by the ubiquity of hydrocarbon-degrading microorganisms. Oil-degrading cold-adapted microorganisms have been found in both pristine and contaminated Alpine (25, 29) and Arctic (48, 53) environments. However, an increase in the hydrocarbon-degrading population after contamination has also been documented in various cold soils (2, 11, 24, 47).

The ubiquity of hydrocarbon degraders was also confirmed by correlation analysis (Table 4). There was no correlation between the TPH content and the number of culturable hydrocarbon degraders. The number of cold-adapted heterotrophs (grown at 10°C) showed a significant positive correlation with the number of cold-adapted diesel oil and hexadecane degraders, and the same result was obtained for mesophilic populations (grown at 37°C). As already observed for biological activities, the numbers of both heterotrophs and hydrocarbon degraders generally showed a significant positive correlation with the soil organic matter content, and there was a negative, although not significant, tendency to correlate with the soil dm. However, microbial counts did not correlate significantly with biological activities (with some exceptions for FDA hydrolysis). Similar results have been described previously (32). Biological activities were higher in pristine than in oil-contaminated soils, while microbial numbers were similar in pristine and contaminated soils (32). Positive correlations between microbial counts and soil biological activities have been reported mainly for bioremediation studies after the biostimulation of soil microorganisms (11, 31, 47).

TABLE 4.

Correlation matrix between microbial counts, biological activities, and physical and chemical parameters of the Alpine soils

Parameter	Correlation coefficient and significance (n = 20) fora:
	Heterotrophs at:				Hydrocarbon degraders
	2°C	10°C	25°C	37°C	Hexadacane, 10°C	Diesel oil, 10°C	Diesel oil, 37°C
TPH	NS	0.516*	0.614**	NS	NS	NS	NS
dm	NS	NS	NS	NS	NS	NS	NS
Organic matter	NS	0.525*	0.465*	NS	0.507*	0.519*	0.668**
pH	NS	NS	NS	NS	NS	NS	NS
Respiration	NS	NS	NS	NS	NS	NS	NS
FDA hydrolysis	NS	NS	0.477*	NS	0.527*	0.466*	NS
Lipase activity	NS	NS	NS	NS	NS	NS	NS
Heterotrophs (2°C)		0.828***	0.612**	NS	0.871***	0.789***	NS
Heterotrophs (10°C)			0.842***	NS	0.822***	0.841***	NS
Heterotrophs (25°C)				0.540*	0.674**	0.796***	0.478*
Heterotrophs (37°C)					NS	NS	0.777***
Hexadecane degraders (10°C)						0.866***	NS
Diesel oil degraders (10°C)							NS

^aNS, not significant; ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Molecular characterization.

(i) Prevalence of genotypes.

PCR amplification of the correct 16S rDNA fragment was obtained for all soil extracts (data not shown), indicating that the soil DNA had been successfully extracted and inhibition of the PCR amplification had not occurred. Subsequent PCR and hybridization analyses of the indigenous microbial populations indicated that the seven genotypes (alkB, alkM, alkB1, alkB2, xylE, ndoB, and nidA) were present in the pristine and/or contaminated Alpine soils (Table 5). In most cases, DNA fragments amplified from PCR of DNA extracts from pristine and contaminated Alpine soils hybridized to the corresponding gene probe, indicating that the amplified PCR fragments possessed a high level of DNA sequence identity (≥68%) to the corresponding regions of the target catabolic genes studied. A representative gel is shown in Fig. 1. The PCR fragment of the expected size did not hybridize to the respective gene probes in three cases; PCR fragments of the pristine soils P4 and P2 did not hybridze to alkB and xylE, respectively, and the universal alkM PCR fragment in the contaminated soil C5 did not hybridize to the ADP-1 alkM gene probe, indicating that distantly related genotypes or spurious PCR fragments had been amplified. In some cases, PCR fragments that were not or were only weakly visually detected after ethidium bromide staining were detected by subsequent hybridization analysis because of the greater sensitivity of the hybridization technique (51). In a previous study using a very similar soil DNA extraction technique with polar soils (51), the detection limits for alkB1, alkB, and ADP-1 alkM were reported as ≈10⁴ cells/g of soil for visually detecting PCR amplification products and ≈10² to 10³ cells/g of soil by hybridization of the PCR amplification products.

TABLE 5.

Screening for various hydrocarbon-degradative genes in pristine and contaminated Alpine soils^a

Soil	TPH (mg/kg)	n-Alkane degradation								Xylene degradation		Naphthalene degradation
		alkB		Universal alkM		alkB1		alkB2		xyIE		ndoB		nidA
		PCR	Probe	PCR	Probe	PCR	Probe	PCR	Probe	PCR	Probe	PCR	Probe	PCR	Probe
Pristine soils (n = 8)
P1	83	−	−	−	−	−	−	−	−	−	−	−	−	−	−
P2	41	−	−	−	−	+	+	+	+	+	w	w	−	+	+
P3	74	−	−	−	−	−	−	−	−	−	−	−	−	−	−
P4	87	+	−	−	−	+	+	+	+	+	+	−	−	+	+
P5	20	−	−	−	−	−	−	w	+	−	−	−	−	+	+
P6	89	−	−	−	−	−	−	−	−	−	−	−	−	−	−
P7	54	+	+	−	−	+	+	+	+	−	−	+	+	+	+
P8	29	−	−	−	−	−	−	−	−	−	−	−	−	−	−
Positive hybridization (% of soils)			12.5%		0%		37.5%		50.0%		12.5%		12.5%		50.0%
Contaminated soils (n = 12)
C1	428	+	+	+	+	−	−	w	+	w	+	+	+	+	+
C2	743	−	−	−	−	w	−	−	−	−	−	−	−	+	+
C3	756	+	+	+	+	+	+	+	+	−	−	−	−	−	−
C4	1,052	−	−	−	−	−	−	+	+	−	−	w	+	−	−
C5	2,085	+	+	+	−	+	+	+	+	w	+	+	+	−	−
C6	3,148	−	−	+	+	+	+	−	+	−	−	+	+	+	+
C7	3,317	+	+	+	+	+	+	w	+	−	+	−	−	−	+
C8	8,385	+	+	−	−	+	+	+	+	+	+	+	+	−	+
C9	13,903	+	+	+	+	−	−	+	+	+	+	+	+	−	−
C10	22,288	w	+	−	−	−	−	−	w	+	+	−	−	+	+
C11	22,447	+	+	+	+	w	−	w	−	+	+	+	+	−	−
C12	30,644	−	+	−	−	−	−	+	+	−	−	+	+	+	+
Positive hybridization (% of soils)			75.0%		50.0%		41.7%		75.0%		58.3%		66.7%		58.3%
Positive control		+	+	+	+	+	+	+	+	+	+	+	+	+	+
Negative control		−	−	−	−	−	−	−	−	−	−	−	−	−	−

^a+, positive; −, negative; w, weak reaction (counted as negative).

FIG. 1.

Detection of P. putida ndoB by PCR analysis and Southern hybridization in Alpine soils. (A) Agarose gel electrophoresis (1.2% agarose) showing the expected 642-bp ndoB fragment obtained by PCR analysis of DNA extracts from Alpine soils. (B) Southern hybridization analysis of ndoB PCR fragments shown in panel A, transferred to a nylon membrane and probed with the 642-bp ndoB gene probe derived from P. putida (ATCC 17848).

For the alkane-degradative genotypes, the alkB1 and alkB2 genotypes were detected in many of the pristine and contaminated Alpine soils. alkB1 could be found in 37.5% (percentage of soils that gave a positive hybridization signal from the PCR amplification products) of the pristine and 41.7% of the contaminated soils, while alkB2 was detected to a greater extent in both pristine (50%) and contaminated (75%) soils. On the other hand, alkB was detected in many of the contaminated soils (75%) but only in one of the eight pristine soils (12.5%). Using the alkM universal primers, alkM genotypes were detected in 50% of the contaminated soils but were not found in any of the eight pristine soils, similar to the alkB result (Table 5). However, it is interesting that the ADP-1 alkM PCR fragments were not detected in any of the Alpine soils when using the original ADP-1 alkM primer set by PCR and hybridization analysis (data not shown); ADP-1 alkM was also rarely detected in polar pristine contaminated soils (51). The 370-bp alkM fragments amplified by the universal alkM primer set obtained from the soils hybridized relatively weakly to the ADP-1 alkM gene probe (472 bp) (compared to the positive control, ADP-1 alkM amplified by the universal alkM primer set) (data not shown), although the intensity of the PCR fragments obtained from the soils and the control were roughly equal. Similar results were observed with PCR hybridization validation analyses of the Acinetobacter control strains (i.e., strong alkM PCR amplification of the five Acinetobacter strains tested and only relatively strong hybridization to the control strain ADP-1 alkM PCR amplification fragment). These results indicate that the ADP-1 alkM genotype and the organisms containing it are not present in significant amounts in polar or Alpine soils but other alkM genotypes are present in Alpine soils in sufficient amounts to be detected by PCR hybridization.

The aromatic degradative genotype P. putida xylE (BTEX degradation) was found in only one of the eight pristine soils (12.5%) but in 58.3% of the contaminated soils (Table 5). For the PAH-degradative genotypes, ndoB was present in only one of the pristine soils (12.5%) but became enriched in contaminated soils (66.7%). The mycobacterial nidA genotype was present in pristine (50%) and contaminated (58.3%) soils to similar extents (Table 5).

(ii) Effect of TPH content on the prevalence of genotypes.

There was a statistically significant correlation between the level of contamination and the prevalence of genotypes in the Alpine soils when all seven genotypes investigated were considered (coefficient = 0.508, P = 0.022*, n = 20). Genotypes containing genes from gram-negative bacteria (alkB, xylE, and ndoB from P. putida, and alkM from Acinetobacter sp.) were detected to a significantly greater extent in the contaminated (50 to 75%) than in the pristine (0 to 12.5%) soils (Table 5). This was confirmed by a highly significant positive correlation (P < 0.001) between the TPH content of the soils and the number of genotypes containing genes from P. putida and Acinetobacter sp. (coefficient = 0.692, P = 0.0007∗∗∗, n = 20). On the other hand, there was no significant correlation between the level of contamination and the number of genotypes containing genes from gram-positive bacteria (alkB1 and alkB2 from Rhodococcus sp., nidA from Mycobacterium sp.). These genotypes were detected with a high frequency in both contaminated (41.7 to 75%) and pristine (37.5 to 50%) soils (Table 5).

Overall, these results indicate that microorganisms containing hydrocarbon-degradative genotypes derived from pseudomonads and Acinetobacter are enriched following oil contamination but significant populations are rarely found in uncontaminated Alpine soils. In contrast, substantial numbers of bacteria containing genotypes derived from Rhodococcus (alkB1 and alkB2) Mycobacterium (nidA), and probably other closely related high-G + C, mycolic acid-containing actinomycetes are already present in soils before contamination events occur. Similar trends were observed with Arctic and Antarctic soils (51) and with cold-adapted populations from CFS-Alert (47) and from Eureka (48) in the Canadian high Arctic. MacNaughton et al. (23) also observed an increase of the gram-negative population with time in oiled plots compared to the unoiled plots.

The observed prevalence of certain genotypes in pristine or contaminated soils may be explained by the r-K scheme, which assumes that evolution favors either adaptation to high rates of reproduction (r strategists) or optimal utilization of environmental resources (K strategists) (5). Bacteria such as pseudomonads (in this study P. putida) and members of the genus Acinetobacter, which rapidly colonize and grow on nutrient-rich materials (in this study, the nutrients were represented by the hydrocarbon contamination), are r strategists. Others, such as streptomycetes, corynebacteria, and similar soil bacteria (in this study Rhodococcus and Mycobacterium spp.) tend to be more successful in resource-limited situations and are K strategists. Populations of K strategists are usually more stable and permanent members of the community (5).

No correlation was found between the prevalence of hydrocarbon-degradative genotypes and the biological activities investigated (respiration, FDA hydrolysis, and lipase activity) or the numbers of culturable hydrocarbon-degrading soil microorganisms. This may be explained by several facts. Quantification of viable cells gives no information on the activity of the populations. There are methodological differences between classical culture-dependent microbiological methods and culture-independent molecular biological methods, and the missing correlations between the prevalence of catabolic genotypes and the numbers of hydrocarbon degraders suggest that many unculturable, as yet unknown bacteria containing homologous genotypes exist in these Alpine soils. There may have been an increase in the number of hydrocarbon degraders in contaminated soils, but they may not have been culturable by the methods used. Also, there are no selective media to distinguish between gram-negative (K strategists) and gram-positive (r strategists) microorganisms. Additional studies are required to ascertain whether correlations can be obtained between measured activities and potential activities determined by culture-independent approaches.