Microbial Diversity during Biodegradation of Crude Oil ... - Springer Link

2 downloads 0 Views 295KB Size Report
Microbial Diversity during Biodegradation of Crude Oil in Seawater from the North Sea. O.G. Brakstad and A.G.G. Lødeng. SINTEF Applied Chemistry ...
Microbial Diversity during Biodegradation of Crude Oil in Seawater from the North Sea O.G. Brakstad and A.G.G. Lødeng SINTEF Applied Chemistry, Department of Marine Environmental Engineering, N-7010 Trondheim, Norway Received: 7 October 2003 / Accepted: 16 February 2004 / Online publication: 23 September 2004

belonged to the family Rhodobacteraceae, with the closest relationship to the genera Sulfitobacter and Roseobacter.

Abstract

Microbial communities were characterized during biodegradation of immobilized oil in seawater from the Statfjord field and the German Bight in the North Sea. Seawater samples were collected at different distances from pollution sources at the two locations. A Statfjord oil was immobilized on hydrophobic synthetic Fluortex fabrics and submerged in closed flasks (no headspace) with natural or sterile seawater and incubated at 13C for 56 days. Biodegradation of immobilized n-alkanes was measured by gas chromatography, total microbes were enumerated by epifluorescence microscopy, and culturable heterotrophic and oil-degrading microorganisms were quantified by most probable number (MPN) analysis. Polymerase chain reaction (PCR) amplification of bacterial 16S rDNA in water samples was conducted during biodegradation experiments. The amplified 16S rDNA fragments were characterized by denaturing gradient gel electrophoresis (DGGE), and by sequence analysis of cloned inserts. Biodegradation rates of alkanes in seawater collected at different distances from the pollution sources did not differ significantly (P > 0.05). Concentrations of oil-degrading microorganisms showed a temporary peak after 7 days of degradation, with a subsequent decline later in the period. DGGE analysis of 16S rRNA genes showed that community diversity decreased during the first 2–3 weeks of biodegradation, with the emergence of a few dominant bands. Cloning, restriction analysis, and sequence analysis of the 16S rDNA fragments revealed >30 different phylotypes. Abundant types during biodegradation belonged to the a-Proteobacteria, in waters from both Statfjord and the German Bight. Cloning and sequencing studies indicated that the most abundant bacteria during biodegradation

Introduction

During biodegradation of oil hydrocarbons (HC) microbial cells adhere to the oil films and droplets. The contact between oil droplets and surfactant-producing microbial cells is associated with increased surface-tovolume ratios of the droplets by droplet size reductions [11], and this may result in increased biodegradation. The partitioning of many HC compounds is important for the biodegradability of these compounds. This was demonstrated in seawater with selected naphthalene compounds, in which the transformation rates were 4–8 times higher for water-soluble fractions than for mechanically generated dispersions or immobilized thin oil films [5, 6]. Catabolic mechanisms during degradation of alkanes and aromatic hydrocarbons have been investigated in a variety of marine bacteria, including Pseudomonas putida [32], Rhodococcus sp. [23], Burkholderia cepacia [20], and mycobacteria and Sphingomonas [14]. Further, bacteria such as Alcanivorax are predominant in oil-contaminated seawater [15, 19]. HC degradation is usually oxygen dependent and catalyzed by di- or monooxygenases [8, 15]. However, HC degradation has also been reported under anoxic conditions [2], using nitrate, iron(III), or sulfate as electron acceptors [29]. We have recently established a method for the immobilization of thin oil films onto hydrophobic surfaces, with subsequent film submersion into seawater. This method allowed studies of HC partitioning between water and oil–water interphases, as well as the possibility of distinguishing degradation of oil compounds in each face [6]. In the current study we compared the changes in microbial communities in seawater from different origins during crude oil biodegradation. Seawater was collected

Correspondence to: O.G. Brakstad; E-mail: [email protected]

94

DOI: 10.1007/s00248-003-0225-6

d

Volume 49, 94–103 (2005)

d

 Springer Science+Business Media, Inc. 2004

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

95

from two different North Sea localities, the Statfjord field and the German Bight. The Statfjord water was influenced by discharges of oil-containing produced water. Produced water discharges have gradually increased in the Norwegian sector of the North Sea, reaching a volume of 116.1 million m3 per year in 2001, and with an average oil content of 24.5 mg/L [24]. The German Bight experience a strong interaction with the Elbe estuary [4]. The German Bight receives a daily flow of 63 million m3 water from the Elbe and the Weser rivers containing considerable concentrations of heavy metals, herbicides, N/P-compounds, and suspended matter [17].

Materials and Methods Water Sampling. Seawater samples were collected near the oil production platform Statfjord B in the Norwegian sector of the North Sea and in the German Bight (Fig. 1) during two cruises with the German research vessel Walther Herwig III. Winter samples were obtained in the period 23 February to 10 March 2001, and summer samples in a cruise from 24 August to 9 September 2001. Statfjord waters were sampled 500 (ST1) and 2000 (ST2) m from the production platform (6112¢N; 150¢E). Reference samples (ST4) were collected 10 km southeast of the platform (607¢N; 33¢E). German Bight samples (GB1, GB3, GB4) were collected along stations more or less affected by the plumes of the rivers Elbe and Weser [4] along the coordinates 54.084N, 7.833E (GB1), 55.000N, 6.325E (GB3), and 55.38N, 4.50E (GB4). For studies of microbial communities seawater was collected in Niskin bottles from a water depth of 14–15 m. The seawater samples (4 to 5.5 L) were filtered through 0.2-lm Sterivex filters (Millipore, Bedford, MA, USA) after prefiltration through 2-lm Millex AP filters (Millipore) to separate free-living microbes from larger organisms and particles. The Sterivex filters were filled with 1.8 mL of lysis buffer (50 mM Tris-HCl, pH 8.0, 40 mM EDTA, 0.75 M sucrose) and stored at )20C until nucleic acid extraction was started. Sterilized 2-L glass flasks were completely filled with seawater, capped, and stored at 4C until start of biodegradation experiments. Picoplankton organisms were enumerated by epifluorescence microscopy with 4,6-diamino-2-phenylindole (DAPI). Most-probable-number (MPN) analysis [26] of culturable aerobic heterotrophic microbes were performed in Marine Broth 2216 with synthetic seawater (Difco Labs., Detroit, MI, USA) at 13C for 7 days. Oildegrading microbes were enriched in Bushnell-Haas Broth (Difco), supplemented with NaCl (33 g/L) and 0.1% (vol/vol) crude Statfjord oil [7] at 13C for 14 days. After incubation fluorescein diacetate (FDA; 0.1 mg/mL final concentration) was added to the media for 60 min

Figure 1. Statfjord (ST) and German Bight (GB) seawater sampling locations.

before results were recorded. FDA was used to visually detect viable cultures by generation of fluorescence signals due to the hydrolysis by intracellular esterases [10]. Nucleic Acid Extraction and PCR. Frozen lysis buffer in the Sterivex filters was thawed, and 2 lg lysozyme was applied to the buffer in the filter cartridge and incubated at 37C for 30 min, followed by incubation of 1 lg proteinase K and SDS [1% (w/v) final concentration] at 55C for 2 h. The lysates were recovered from the filter cartridges, and the Sterivex filters were washed once more with lysis buffer (55C, 10 min). The lysates from each filter were pooled and extracted with hot phenol– chloroform–isoamyl alcohol (25:24:1; pH 8), according to standard procedures [28]. Water with extracted nucleic acids was precipitated by 2.5 volumes of ethanol (96%) at )20C for 3 h, and centrifuged (4000g; 5 min at 4C). The pellets were washed (75% ethanol), recentrifuged, dried (nitrogen), and dissolved in sterile ultrapure water (Biochrom, Berlin, Germany). Recovered nucleic acids were quantified by an ethidium bromide method [28]. Nucleic acid extracts were stored at )20C. Seawater samples from biodegradation experiments were filtered through a 0.22-lm Durapore filter with diameter 25 mm (Millipore), and nucleic acids recovered from the filters by the Genomic Prep Cells and Tissue DNA Isolation kit (Amersham Pharmacia, Piscataway, NJ). Filters were placed in 600 lL of cell lysing reagent from the kit and further procedures performed as described by the manufacturer.

96

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

PCR amplification of 16S rDNA was performed with domain-specific primers for Bacteria. The primers Bac341f (5¢-CCT ACG GGA GGC AGC AG-3¢) and Bac907r (5¢-CCC CGT CAA TTC CTT TGA GTT-3¢) yielded a 550-bp fragment. The amplification was conducted as a ‘‘touchdown’’ PCR to reduce formation of spurious by-products [12], as previously described [30]. Annealing temperature was initially set at 65C, then decreased by 1C every second cycle until 55C, at which point 25 additional cycles were carried out. Denaturing Gradient Gel Electrophoresis DGGE was performed essentially as de(DGGE).

scribed by Teske et al. [30]. PCR was run with the Bac341f and Bac907r primers (see above). A GC-clamp (5¢-CGC-CCG-CCG-CGC-GCG-GCG-GGC-GGG-GCGGGG-GCA-CGG-GGG-G-3¢) was added to the 5¢ end of the Bac341f primer [22]. PCR fragments prepared with a GC clamp were separated in a 6% polyacrylamide gel with a continuous gradient of 20–70% of the denaturing agents urea and formamide (100% denaturants correspond to 7 M urea and 40% deionized formamide). Each well contained 10 lL from the completed PCR assay (0.5–1.0 lg DNA). DGGE was run at 60C in a DCode Universal Mutation Detection System (Bio-Rad, Hercules, CA, USA) at 150 V for 4 h. Gels were stained for 20– 30 min with SYBR Gold (Molecular Probes BV, Leiden, The Netherlands) according to descriptions from the manufacturer. Stained gels were scanned in a GelDoc system (Bio-Rad), and diversity indices determined as Ward’s Dice similarity coefficients with the Quantity One software of the Discovery Series (Bio-Rad). rDNA

Clone

Libraries,

RFLP,

and

Sequenc-

Nucleic acids extracted from seawater at days 0 and 21 of the biodegradation process were amplified by domain-specific PCR for Bacteria. PCR fragments were eluted (Perfectprep Gel cleanup kit; Eppendorf AG, Hamburg, Germany) from standard agarose gels after electrophoresis (150 V, 1.5–2 h) and cloned with a TOPO TA cloning kit (Invitrogen, Carlsbad, CA, USA), using One Shot TOP10 chemically competent Escherichia coli cells. Cells were spread on LB agar plates pre-treated with X-Gal (1.5 mg/plate) and supplemented with kanamycin (50 lg/mL medium). Putative positive clones were transferred to multiwell plates (6 · 4 wells) containing LB broth with kanamycin (50 lg/mL), and cells were stored at )80C. Plasmids were prepared with GeneElute Plasmid Miniprep kit (Sigma, St. Louis, MO, USA), and isolated plasmids stored at )20C. PCR product inserts were amplified using primers for the M13 sites on the plasmid. The primers flank the inserts [28], generating a 770-bp fragment. The fragment was analyzed by restriction fragment length polymorphism (RFLP) with digestion with HaeIII or RsaI ing.

(Sigma) at 37C for 2.5 h. Fragments shorter than 100 bp were not considered in the analysis of the RFLP profiles. Selected M13 PCR products were submitted for partial sequence analysis (Eurogentec, Ivoz-Ramet, Belgium), using a single-pass primer extension reaction, which typically gave 300- to 500-bp sequences. Doublestranded DNA sequences were determined by the Kodon version 1.0 software (Applied Maths, St. Martens-Latem, Belgium). Sequence alignments were performed by the BLAST program of the National Centre for Biotechnology Information [3], and phylogenetic trees were generated with the Kodon software. The Kimura twoparameter model was used to estimate evolutionary distance, and 100 bootstraps were performed to assign confidence levels to the nodes of the trees. Biodegradation

Experiments

with

Immobilized

Seawater received at the laboratory was aerated with sterile air, supplemented with inorganic nutrients (8.5 mg/L KH2PO4, 21.8 mg/L K2HPO4, 33.3 mg/L Na2HPO4 Æ 2H2O, 0.5 mg/L NH4Cl, 27.5 mg/L CaCl2, 22.5 mg/L MgSO4 Æ 7H2O, 0.25 mg/L FeCl3 Æ 6H2O) as recommended [25] and dispensed into 275-mL BOD bottles. Sterile controls of Statfjord water (ST1) were achieved by adding 50 mg/L HgCl2. A crude North Sea paraffinic oil (Statfjord fresh, batch no. 97-0264) with density 0.850 g/cm3 was applied to the surface of sterile seawater by introducing 50 lg of oil to a beaker containing 500 mL of water (temperature 13C). The oil distributed immediately across the water surfaces, generating even surface films. Hydrophobic synthetic Fluortex fabrics (Sefar Inc., Thal, Switzerland) were prewashed in dichloromethane (DCM), rinsed in sterile seawater, and air-dried. The wire area of the fabrics represented 64% of the total surface area, according to information from the manufacturer. The fabrics were cut in pieces of 1 · 1 cm and carefully applied to the seawater surfaces 2 min after oil film generation. The fabrics were incubated for 60 min (13C) floating on the oil surfaces, then removed and carefully washed in two separate baths of sterile seawater. Thin fishing lines (thickness 0.30 mm according to information from the manufacturer) with knots in the ends (prewashed in DCM, rinsed in sterile seawater and dried) were carefully forced through the fabrics and used for submerging the fabrics in the seawater in the BOD bottles. BOD bottles completely filled with enriched seawater (no headspace) were incubated at 13C for 0–56 days. Triplicate bottles with Fluortex fabrics in nutrient-supplemented seawater were withdrawn for chemical and microbiological analysis at days 0, 7, 14, 21, 28, and 56 of the biodegradation period. Seawater samples were immediately processed for microbiological analysis and for nucleic acid extraction (see above), while Fluortex fabrics were withdrawn for chemical analysis. Dissolved oxygen was measured in all Oil.

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

97

Figure 2. DGGE results for PCR-amplified 16S rRNA genes of bacteria seawater sampled from Statfjord (ST1, ST2, ST4) of German Bight (GB1, GB3, GB4) during winter (WIN) and summer (SUM) cruises. The stained gel is shown in A; diversity indices are shown in B, using Ward’s clustering method.

bottles with a sterilized BOD probe washed with 70% ethanol and rinsed with sterile seawater (YSI, Yellow Springs, OH, USA). Fabrics from sterile seawater controls (one parallel) were also withdrawn for chemical analysis. Fabrics were extracted with DCM (50 mL) to dissolve immobilized hydrocarbons, and the solvents were dried with solid Na2SO4, filtered through glass wool, and concentrated to 0.5–1.0 mL. Determination of C10–C36 n-alkanes in total extractable organic compounds (C10–C36 TEOC) was performed by GC-FID analysis as described previously [16]. Nonlinear regression analysis for determination of transformation rates, Pearson two-tailed analysis of corChemical Analysis and Transformation.

relation, and pairwise comparison of data by t-test, were performed by the GraphPad Prism 3.01 software (GraphPad Software, San Diego, CA, USA). Results Microbial Diversity in Collected Seawater. Nucleic acids from microbial communities in situ were collected from seawater during summer and winter cruises in the German Bight and the Statfjord field. DNA fragments were amplified with 16S rDNA domain-specific primers for Bacteria. DGGE analysis of the PCR products showed four to seven visible bands (Fig. 2A), representing abundant seawater phylotypes. Visual comparison of DGGE patterns from the two fields indicated that the

98

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

differences were higher between profiles from different fields than from the same field. DGGE patterns were compared by calculation of dendrograms, based on Ward’s Dice coefficient (Fig. 2B). Similarity coefficients were higher between Statfjord samples (mean values 73.4 with standard deviation 16.3) than between German Bight samples (mean values and standard deviations 57.5 and 31.1, respectively). Comparison of similarity coefficients between the Statfjord and German Bight groups by paired t-test analysis showed significant differences (P < 0.05). A separation between profiles from summer and winter cruises within each cluster was not clear in this study. Biodegradation Study. All seawater samples were used for biodegradation studies with fresh Statfjord oil. GC-FID analysis of immobilized oil in sterile seawater showed that nC12–nC36 alkanes were absorbed, in accordance with the GC profile of the original oil (results not shown). However, nC10–nC11 alkanes present in the original oil were not immobilized on the fabrics, probably because of evaporation. The seawater dissolved oxygen concentrations varied between 3.0 and 5.1 mg O2/L, demonstrating that oxic conditions were maintained during the experiments. The fabrics with immobilized oil were submerged in nutrient-supplemented or sterile seawater from Statfjord and German Bight. The amounts of immobilized hydrocarbons on the fabrics were examined by weekly measurements of the nC12–nC36 TEOC, and depletion curves with seawater from the winter cruise (February– March) are shown in Fig. 3. No significant reductions in n-alkane amounts were calculated with the sterile seawaters by correlation analysis (P > 0.05), and no corrections for hydrocarbon depletion in abiotic samples were required during the biodegradation. In the nutrientsupplemented seawaters TEOC depletion differed by 86 to 91% (average 90.1%) after 2 months of incubation. Regression analysis (Table 1) showed that n-alkane halflives varied between 14 and 27 days (mean values). By pairwise comparison of individual data sets (paired t-test analysis) no significant differences in transformation rates were determined (P > 0.05). This indicated that hydrocarbon biotransformation rates were not related to

Figure 3. Amounts of immobilized n-alkanes (C10–C36 TEOC) on the Fluortex fabrics during the biodegradation period in Statfjord (A) or German Bight (B) seawater from the winter cruise.

the distance of the original seawater samples from the pollution sources (Statfjord B production platform or the rivers Elbe and Weser), although this could be suggested by the results in Table 1. The biodegradation experiment was repeated with the seawater samples from the summer cruise, but the transformation results with waters from the two cruises did not differ significantly (P > 0.05; data not shown). Total and culturable concentrations of microbes in the seawater during the biodegradation periods are illustrated with the ST1 water from the winter cruise (Fig. 4). The concentrations of total microbes did not vary significantly during the experiment. Heterotrophic microbes showed increased concentrations up to a plateau, representing 1–6% of the total cells. The concen-

Table 1. Transformation rate constants, half-lives (t1/2) and goodness of fits of TEOC n-alkanes (C10–C36) of crude oil in Statfjord (ST1-ST4) or German Bight (GB1-GB4) seawater from the winter cruise

Source

Rate constanta (k1d)1)

t1/2a (d)

ST1 ST2 ST4 GB1 GB3 GB4

0.050 0.045 0.035 0.036 0.029 0.025

14 16 20 19 24 27

a

(0.032–0.067) (0.035–0.054) (0.026–0.044) (0.027–0.046) (0.020–0.037) (0.017–0.034)

(10–21) (13–20) (16–27) (15–26) (19–34) (21–41)

Rate constants and half-lives as means of three replicates with 95% confidence intervals in parentheses.

Goodness of fit (R2) 0.88 0.96 0.93 0.93 0.91 0.89

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

Figure 4. Enumeration of total microorganisms by epifluorescence microscopy and viable counts by MPN determinations (culturable heterotrophic and oil-degrading bacteria) in ST1 from winter cruise during oil biodegradation. The error bars show 95% confidence intervals of triplicate samples.

trations of oil-degrading microbes usually showed a temporary peak early in the degradation period, with a subsequent decline compared to those at the start of the experiments. Community Structures during Biodegradation. The changes in seawater bacterial community structures during the biodegradation experiments were examined by DGGE analysis of 16S rRNA gene fragments. DGGE results of bacterial fragments are shown in Fig. 5 with seawater DNA originating from winter (ST1, GB1) or

Figure 5. DGGE results for PCR amplified 16S rRNA genes of bacteria in seawater samples from Statfjord and German Bight during biodegradation of crude oil. Results are shown after 0, 7, 14, 21, 28, and 56 days of biodegradation in ST1 winter cruise (A), ST1 summer cruise (B), GB1 winter cruise (C), GB3 summer cruise (D), and GB4 summer cruise (E). Results for days 0 and 21 were not determined with GB3 water.

99

Figure 6. Similarities (Dice coefficient) between DGGE patterns by comparison of community profiles during oil biodegradation in each series of seawater. The 0-day community patterns in each series were used as references. The profile of each reference water (0-day) was pairwise compared to replicates from the same water retrieved at day 7, 14, 21, 28 or 56 of the biodegradation period. The results are shown as mean values for the Statfjord (n = 3) and German Bight (n = 3) series with standard deviations as error values.

summer (ST1, GB3, GB4) cruises. In general, the communities in the seawater samples from the German Bight contained higher numbers of visible bands (3–12; average 7) than the seawater samples from Statfjord (2–8; average 4) during the experiments. The results showed unambiguous changes in DGGE profiles during biodegradation. Temporary reductions of bands were observed after 7 to 21 days of biodegradation, and the emergence of a few strong bands (Fig. 5) indicated that this period was dominated by a few bacterial types. Toward the end of the biodegradation period (28–56 days) the number of DGGE bands increased. Thus, continuous changes in the community structures seemed to occur during biodegradation. The community DGGE profiles from each seawater sample during the oil biodegradation were compared to the profiles at the start of the degradation period (day 0) by calculation of Dice coefficients. The compiled results are shown in Fig. 6 for profile changes in Statfjord and German Bight samples, respectively. The similarities to 0day Dice coefficients were reduced early in the biodegradation period (days 7–21). This was the result of band reductions and the emergence of new bands. After 28 and 56 days of the period the similarity to the day 0 profiles increased, indicating that several of the re-emerging bands in Statfjord and German Bight matched DGGE waters bands from 0-day samples (Fig. 5 and Fig. 6). Phylogenetic Analysis. Nucleic acids extracted from seawater samples at days 0 and 21 of the biodegradation period were amplified by PCR and cloned. A library of 93 clones was generated from Statfjord (n = 68) and German Bight (n = 25) seawater. A total of 32 clones contained inserts originating from seawater

100

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

Figure 7. Phylogenetic tree based on partial sequences of 16S rRNA genes. An UPMGA analysis was performed, based on a pairwise distance method and with Kimura 2 parameter corrections. The sequences in the tree included clones from Statfjord (ST) or the German Bight (GB) at day 0 or 21 of the biodegradation period.

before start of biodegradation (19 from Statfjord and 13 from German Bight), while the rest of the clone inserts (n = 61) were extracted from water after 21 days of degradation (49 clones originating from Statfjord and 12 from German Bight samples). The clones were screened with the restriction enzymes HaeIII and RsaI, resulting in 11 HaeIII types and 13 RsaI types. A total of 30 RFLP phylotypes were found when the results with the two restriction enzymes were combined. Of these, 23 types were retrieved from clones at day 0, and 7 from day 21. Several RFLP types (n = 16) were unique, containing only one clone, and 8 of these originated from seawater before biodegradation (day 0), while five types included ‡5 clones, 81% originating from day 21 of biodegradation. Thus, clones from day 21 were restricted to a few RFLP types compared to the clones from day 0.

Partial 16S rRNA genes of 23 clones from Statfjord (seven clones from day 0 and three from day 21 of the biodegradation period) and 12 clones from German Bight (eight clones from day 0 and four from day 21) seawater samples were sequenced. The selected clones represented dominant RFLP types before and during biodegradation. A phylogenetic tree (Fig. 7) showed that the sequences from clones originating from seawater samples at the start of biodegradation (n = 15) were distributed within proteobacteria, Actinobacteria and Bacteroidetes. The clones from the Statfjord and German Bight waters belonged mainly to the Bacteroidetes (four of seven Statfjord clones) or the a-, b-, or c-subgroups of Proteobacteria (six of nine German Bight clones), respectively. Sequences from day 21 of the biodegradation period grouped within the Rhodobacteriales of the

O.G. BRAKSTAD, A.G.G. LØDENG: OIL BIODEGRADATION IN SEAWATER: MICROBIAL DIVERSITY

Figure 8. Percentage distribution of RFLP types related to different phyla at the start and after 21 days of oil biodegradation in seawater samples from Statfjord and from the German Bight.

a-proteobacteria, with one exception (clone 22-GB-Day 21). The Statfjord clones were closely related to Sulfitobacter sp. EE-36, while the German Bight clones resembled the sequences of Roseobacter sp. KT0202a (Fig. 7). The Statfjord and German Bight clones from biodegradation day 21 showed an overall similarity of 96.3%, demonstrating the close relatedness between the bacteria abundant during biodegradation, although originating from different sources. The distribution of clones within the library was grouped by phylotype based on RFLP typing and sequencing of the selected RFLP clones. Clone distribution changed over time, with a few types dominating strongly after 21 days of biodegradation (Fig. 8). The distribution of the clones belonging to the RFLP types similar to the Rhodobacteriales clones increased from 21% at day 0 to 89% at day 21 in the Statfjord water, while the corresponding increase in the German Bight water was from