Deepsea bacteria enriched by oil and dispersant from the Deepwater

Deep-sea bacteria enriched by oil and dispersant fromthe Deepwater Horizon spillemi_2780 1..12

Jacob Bælum,1,2 Sharon Borglin,1

Romy Chakraborty,1 Julian L. Fortney,1

Regina Lamendella,1 Olivia U. Mason,1

Manfred Auer,1 Marcin Zemla,1 Markus Bill,1

Mark E. Conrad,1 Stephanie A. Malfatti,3

Susannah G. Tringe,3 Hoi-Ying Holman,1

Terry C. Hazen1 and Janet K. Jansson1,3*1MS 70A-3317, 1 Cyclotron Rd., Lawrence BerkeleyNational Laboratory, Berkeley, CA 94720, USA.2The Geological Survey of Denmark and Greenland,Øster Voldgade 10, 1350 Copenhagen, Denmark.3Joint Genome Institute, 2800 Mitchell Drive, WalnutCreek, CA 94598, USA.

Summary

The Deepwater Horizon oil spill resulted in a mas-sive influx of hydrocarbons into the Gulf of Mexico(the Gulf). To better understand the fate of the oil,we enriched and isolated indigenous hydrocarbon-degrading bacteria from deep, uncontaminatedwaters from the Gulf with oil (Macondo MC252) anddispersant used during the spill (COREXIT 9500).During 20 days of incubation at 5°C, CO2 evolution,hydrocarbon concentrations and the microbial com-munity composition were determined. Approximately60% to 25% of the dissolved oil with or withoutCOREXIT, respectively, was degraded, in addition tosome hydrocarbons in the COREXIT. FeCl2 additioninitially increased respiration rates, but not the totalamount of hydrocarbons degraded. 16S rRNA genesequencing revealed a succession in the microbialcommunity over time, with an increase in abundanceof Colwellia and Oceanospirillales during the incu-bations. Flocs formed during incubations with oiland/or COREXIT in the absence of FeCl2. Synchrotronradiation-based Fourier transform infrared (SR-FTIR)spectromicroscopy revealed that the flocs were com-prised of oil, carbohydrates and biomass. Colwelliawere the dominant bacteria in the flocs. Colwellia sp.strain RC25 was isolated from one of the enrichmentsand confirmed to rapidly degrade high amounts

(approximately 75%) of the MC252 oil at 5°C. Togetherthese data highlight several features that provide Col-wellia with the capacity to degrade oil in cold, deepmarine habitats, including aggregation together withoil droplets into flocs and hydrocarbon degradationability.

Introduction

On 20 April 2010, high-pressure oil and gas caused theDeepwater Horizon drilling rig in the Gulf of Mexico toexplode resulting in the second largest marine oil spill inthe history of the petroleum industry. It has been esti-mated that ~ 4.1 million barrels of light crude oil andnatural gas leaked into the Gulf from the Macondo well(MC252) during the ~ 3 months before the well head wascapped (OSAT, 2010). Primarily in an attempt to improvesafety for surface operating vehicles, ~ 1.8 million gallons(~ 37 500 barrels) of the chemical dispersant COREXIT(mainly 9500, but also 9527 formulation) was applied tothe Gulf surface as well as directly into the wellhead at1500 m below surface level (mbsl; 9500 only). Very little isknown about the effects and persistence of COREXIT inthe environment, including its impact on indigenousmicrobes in the Gulf (Judson et al., 2010).

During the spill, hydrocarbon concentrations variedfrom 100% at the source to undetectable with increas-ing distance from the wellhead. During exploratoryand surveillance cruises conducted for the US CoastGuard, National Oceanic and Atmospheric Administration(NOAA), and the US Environmental Protection Agency(EPA), a plume or cloud of dispersed MC252 crude oil wasdetected at 1100–1220 mbsl at distances up to 35 kmfrom the wellhead (Camilli et al., 2010). The plume isthought to have formed due to application of COREXIT atthe wellhead and a combination of physical and chemicalproperties at 1100 mbsl depth [e.g. pressure, temperature(~ 5°C) and salinity (Yapa et al., 2008; Dasanayaka andYapa, 2009; Adcroft et al., 2010)], which resulted in the oilattaining neutral buoyancy at this depth. During theongoing release of oil several research groups reportedthe presence of the plume, but after the wellhead wascapped the plume was no longer detectable (OSAT,2010).

Several reports have provided strong evidence thatmicrobial processes were a key factor for degradation and

Received 20 February, 2012; revised 23 April, 2012; accepted 23April, 2012. *For correspondence. E-mail [email protected]; Tel. (+1)510 486 7487; Fax (+1) 510 486 7152.

bs_bs_banner

Environmental Microbiology (2012) doi:10.1111/j.1462-2920.2012.02780.x

Published 2012. This article is a U.S. Government work and is in the public domain in the USA.

mailto:[email protected]

removal of oil from the plume (OSAT, 2010). For example,Hazen and colleagues (2010) reported a shift in the com-position of the microbial community in the deep-seaplume within 1 month of the spill (May 2010). They foundthat sequences corresponding to Oceanospirillales domi-nated in the oil plume, but that they were rare in uncon-taminated water at the same depth. Later sampling of theplume in June 2010 reported that sequences representa-tive of Cycloclasticus and Colwellia were dominant,together accounting for more than 95% of the sequencedata (Valentine et al., 2010). Recently, Redmond andValentine (2011) reported that Colwellia sequencesincreased in abundance during enrichment on crude oil atcold temperatures (4°C) compared with warmer (20°C)incubation conditions and suggested that temperaturewas a major determinant in selection of this group ofmicroorganisms.

Here we aimed to study the succession of the indig-enous microbial community and the formation of microbialflocs in deep-sea water from the Gulf of Mexico duringlaboratory enrichments at cold temperature (5°C) and withhigh concentrations of MC252 oil and COREXIT. We alsoaimed to select and isolate specific members of the micro-bial community having the capability to degrade oil in thepresence or absence of dispersant. The source of inocu-lum was uncontaminated seawater collected from theGulf of Mexico during the Deepwater Horizon oil spill atthe depth of the reported hydrocarbon plume (1100 mbsl).We also looked at the impact of Fe2+ on the degradationrates because iron was reported to be a potentially limitingnutrient for microbial growth in the water column (Hazenet al., 2010; OSAT, 2010; Lu et al., 2012).

During the enrichments the community compositionwas monitored by 454 pyrotag sequencing of 16S rRNAgenes. We also monitored respiration and degradation ofoil and COREXIT during the incubations. These experi-ments enabled us to gain a better understanding of thepattern of microbial succession, and the response of spe-cific microbial populations when exposed to high concen-trations of hydrocarbons such as occurred during theDeepwater Horizon oil spill or that naturally occur from oilseeps on the ocean floor.

Results

Respiration during enrichments

Microbial populations in water collected from the deep-sea in the Gulf of Mexico were enriched to grow on anddegrade high concentrations of Macondo MC252 oil, inthe presence or absence of the dispersant, COREXIT9500 at low temperature (5°C). Additional treatmentsincluded Fe amendment and sterile controls. As a dis-solved oxygen concentration of 6.07 mg l-1 was measured

on site (Hazen et al., 2010), our experiments were per-formed under aerobic conditions. Dissolved oxygen con-centrations measured on 0, 5 and 20 days of incubationwere 8.4, 7.5 and 6.1 mg l-1 respectively.

During the enrichments we observed significant differ-ences (t-test, P < 0.05) in CO2 evolution patterns betweenthe different treatments (Fig. 1A). There was a substantialincrease in CO2 evolution in the enrichments containingoil and/or COREXIT compared with sterile controls. In theenrichments with oil alone and oil + Fe, the cumulativeCO2 evolution during the 20-day incubation period wasin the range of 2.8–3.0 mg of CO2, and in those withCOREXIT in the range 4.4–4.9 mg of CO2, regardless ofthe presence of oil (Fig. 1A). In the presence of FeCl2there was an immediate onset of CO2 evolution, while inthe enrichments without FeCl2, there was a lag phase of~ 6 days before CO2 evolution commenced. However,regardless of the addition of FeCl2, the total amount ofCO2 that accumulated after 20 days was similar.

Degradation of Macondo crude oil

During the incubations we determined the concentrationof dissolved crude oil (measured as total petroleum hydro-carbons, TPH) by extracting aliquots of the seawater andanalysing by GC-FID as described below. In the treat-ments with oil and no COREXIT, oil adhered to the sidesof the bottles, reducing the amount in the seawater toapproximately 30% of the initial concentration of 100 ppm.In samples with COREXIT more of the oil was dissolved;~70% of the initial concentration (Fig. 1B), as expectedbecause COREXIT is a dispersant. After 5 days of incu-bation 25% of the dissolved oil was degraded in sampleswith oil alone and 40% in samples with oil and COREXIT.After 20 days no additional oil was degraded in thesamples with oil alone, but 60% had been degraded insamples with oil and COREXIT (for mass balances seeFig. S2). While the oil degradation was reported as TPH,the majority of compounds lost were straight chainalkanes, which were predominant components of the MC252 source oil.

COREXIT 9500 degradation

The concentrations of COREXIT components were alsomeasured during the incubations. COREXIT 9500 is awater-soluble mixture containing hydrocarbons (50%),glycols (40%) and dioctylsulfosuccinate (DOSS) (10%)and these different classes of compounds were degradedat different rates (Fig. 1C). While degradation of thehydrocarbon compounds was initially fast, slowing downin the latter phase, the glycol compounds and DOSS weredegraded at a linear rate. Although the largest portion ofdegraded mass originated from the hydrocarbon fraction,

2 J. Bælum et al.

Published 2012. This article is a U.S. Government work and is in the public domain in the USA, Environmental Microbiology

still a significant amount of the DOSS was degraded,while the glycols were more recalcitrant.

Cell counts

There was an increase in microbial cell density in allof the enrichments with oil and/or COREXIT added(Fig. 1D). The bacterial counts for the source waterbefore incubation was in the range 2–4 ¥ 104 cells ml-1.Over time we observed a significant (t-test, P < 0.05)increase in the cell density in the oil and/or COREXIT-amended enrichments. The highest densities of 1–2 ¥107 cells ml-1 were obtained in enrichments with both oiland COREXIT, corresponding to the highest respirationrates also observed in those treatments. In enrichmentswith oil alone the growth was slower with ~ 1 ¥ 105 cellsml-1 on day 5 of incubation and ~ 2 ¥ 106 cells ml-1 onday 20 (Fig. 1D).

Microbial community analysis

One of our main goals was to determine which specificmembers of the indigenous microbiota in the deep sea of

the Gulf of Mexico were enriched in the presence of highconcentrations of oil. Therefore, DNA was extractedduring the incubations and 16S rRNA genes weresequenced using a 454 pyrotag sequencing approach. Atotal of 148 276 quality-filtered reads were clustered inoperational taxonomic units (OTUs) with 97% similarityenabling identification of OTUs down to genus level, butno differentiation on the species level was possible.During the enrichments there was a clear succession inthe bacterial community composition (Fig. 2). In particular,Colwelliaceae increased from barely detectable in thesource water to relative abundances of 15–30% in all ofthe enrichments, similar to the recent findings reported byRedmond and Valentine (2011). Also bacteria within theOceanospirillales, including Oleispira spp. that was previ-ously found to be abundant in the deep-sea oil plumeduring the Deepwater Horizon oil spill (Hazen et al.,2010), increased from being rare (i.e. < 1% relative abun-dance) to relative abundances of 5–10% as the hydrocar-bons were degraded. As these values are all relative,the overall microbial abundance increased substantially,such that even those groups that declined in relativeabundance may have increased in total numbers.

Fig. 1. Enrichments of deep-sea bacteria collected from the Gulf of Mexico during the Deepwater Horizon oil spill to high concentrations ofMacondo (MC252) crude oil (100 ppm) and the dispersant, COREXIT 9500 (60 ppm). (A) Measurement of CO2 evolution (microbialrespiration); (B) degradation of MC252 oil; (C) degradation of components in COREXIT 9500; (D) microbial cell counts. Treatments were asfollows: seawater control, sterile seawater control, sterile seawater with oil and FeCl2 added, oil alone,

oil + COREXIT, COREXIT, Oil + Fe, Oil + COREXIT + Fe, hydrocarbon fraction in COREXIT 9500, glycolfraction in COREXIT 9500, and dioctylsulfocussinate (DOSS) fraction in COREXIT 9500. indicates the time points that triplicateenrichments were sacrificed during the incubations. Error bars represent standard error of three replicate enrichments.

Enrichment of oil degraders from Gulf of Mexico 3


Microbial floc formation

Large flocs (Fig. 3A and C) were observed in all of theenrichments with oil and/or COREXIT and without FeCl2,

while in enrichments with FeCl2 there was no floc forma-tion (Fig. 3B). Typically, there was one large floc sus-pended in the water after 20 days of incubation, butalready after 2 days cell aggregates were detectable bylight microscopy (Fig. 3A). After 20 days of incubation atypical floc had a size of 1–4 cm, was white in colour, andsuspended in the water.

Synchrotron radiation-based Fourier transform infrared(SR-FTIR) spectromicroscopy at the LBNL AdvancedLight Source (ALS) was used to monitor the compositionof the flocs during their formation. The SR-FTIR spectrarevealed intense absorption peaks (Fig. 4) between 3100and 2800 cm-1, strong absorption peaks at ~ 1730 and~ 1545 cm-1, and absorption profiles between 1200 and900 cm-1 associated with the flocs (Fig. 4). These absorp-tion features are well described for the hydrocarbon (C-H)vibration modes of the MC252 oil, the carbonyl (C=O)vibrational modes of oil degradation products, amide II(N-H) vibrational modes of proteins, and ring vibrations ofdiverse polysaccharide and exopolysaccharide (EPS)groups (Naumann, 2000; Hazen et al., 2010). A time-course analysis of image and intensity of absorptionfeatures revealed that the chemical and biochemical com-position of the flocs evolved over time (Fig. 5). The inten-sity of the absorption bands corresponding to MC252 oilcomponents increased in intensity from 0.056 relativeabsorbance units (a.u.) to 1.11 a.u. during the initial 20days of incubation and subsequently declined to 0.98 a.u.for the remaining 20- to 40-day incubation period. This

Fig. 2. Relative abundance of different phylogenetic groups basedon operational taxonomic units (OTUs) obtained by pyrotagsequencing of amplified 16S rRNA genes. Each bar represents themean of three replicate enrichments. C, COREXIT.

Fig. 3. Microscopic images of cellaggregation and floc formation.A and B. Microscopic images of cellscollected on 0.2 mm filters and stained withacridine orange: (A) cells that are aggregatingin the presence of oil alone after 2 days ofincubation, (B) lack of cell aggregation in thepresence of oil and FeCl2 after 2 days ofincubation.C. A microscopic image of a portion of a floccollected from an enrichment after 40 days ofincubation with oil.

4 J. Bælum et al.


suggests that the oil tended to initially concentrate in thefloc material, but was subsequently degraded. Mean-while, the absorption intensity of the protein amide IIsignal increased from 0 (at t = 0 days) to 0.123 a.u. (t = 40days) throughout the incubation period indicating anincreasing abundance of biological material. The absorp-tion intensity and the absorption profile complexitythroughout the infrared fingerprint region (~ 1500 to700 cm-1) also increased during the incubation period(Fig. 4) and exhibited the spectral features typical ofmarine mucilages containing polysaccharides (Bertoet al., 2005) or marine snow (Mecozzi and Pietrantonio,2006).

The bacterial community composition in the flocs wasdetermined by 16S rRNA gene 454 pyrotag sequencing.The sequencing data revealed that Colwelliaceae werethe dominant taxa in the flocs at relative levels of 70% atall time points (Fig. 2). The microbial diversity was alsoconsiderably lower in the flocs compared with the bulkseawater (Fig. 2). Another group that was detected in theflocs corresponded to Methylococcaceae that increasedin relative abundance from less than 1% at 10 days ofincubation to 16% after 40 days of incubation. All of thebacteria that were detected in the flocs were also detectedin the bulk water, but their relative amounts differed in thetwo sample types, suggesting that only certain membersof the community were associated with the flocs and theirformation.

Isolation of hydrocarbon-degrading Colwellia

Because of the dominance of Colwelliaceae in the enrich-ments and in the flocs, a representative strain was iso-lated in order to determine whether it was capable of oil

degradation. Subsamples were collected from one of theenrichments on oil and COREXIT at the conclusion of the20-day incubation period and streaked onto Marine Brothagar plates. After incubation at 5°C for 1 week, individual,smooth, beige colonies appeared on the plates. Scanningelectron microscopy (SEM) imaging revealed that thecells were slightly curved rods (Fig. 6). The cells wereapproximately 1.5–2 mm long and approximately0.25–0.4 mm wide. 16S rRNA gene sequencing revealedthat the isolate was a Colwellia species (strain RC25) with98.6% sequence homology to the most abundant Col-wellia spp. observed by 16S pyrosequencing in the origi-nal enrichments. The 16S rRNA gene sequence ofColwellia sp. RC25 had 96% sequence similarity to thetype strain, Colwellia psychrerythraea 34H, isolated byMethé and colleagues (2005). Cells of strain RC25 wereGram-negative, non-spore forming, appeared motile andgrew at 5°C.

The isolate was transferred to liquid medium containing100 ppm MC252 oil and 60 ppm COREXIT. After 10 daysof incubation, approximately 75% of the initial amount ofMC252 oil was degraded by Colwellia strain RC25(Fig. 7), demonstrating that this organism is a likely can-didate for the observed hydrocarbon degradation in theenrichments.

Discussion

As a result of millions of years of natural oil seeps from theseafloor, oil residues in the Gulf of Mexico are a commonphenomenon (OSAT, 2010) and hence it is not surprisingthat the microbial potential for biodegradation of oilcompounds is present. Several studies have reporteddegradation of oil in surface seawater (Kasai et al., 2002;

Fig. 4. SR-FTIR spectra of floc materialcollected after 0 (no floc material), 5, 10, 20and 40 days of incubation with 100 mg l-1

MC252 oil. Tentative band assignment of theprotein amide I and II vibration modes at~ 1648 and ~ 1542 cm-1, of the carbohydratevibration modes at ~ 1000 cm-1, of alkaneC–H vibration modes in oil from MC252, andof carbonyl (C=O) vibration modes at~ 1730 cm-1 in oil oxidation products.



Gertler et al., 2009; Zahed et al., 2010), but to our knowl-edge no investigations of oil degradation potentials indeep-sea environments were reported prior to the Deep-water Horizon oil spill. Although the deep sea is notabledue to its low ambient temperature and high pressure, the

microbial responses to oil determined from our experi-ments are similar to those reported for several other envi-ronments (Atlas and Hazen, 2011).

Different hydrocarbon fractions of Macondo (MC252) oilin the deepwater plume resulting from the Deepwater

Fig. 5. SR-FTIR images of the formation and growth of a floc during enrichments with 100 mg l-1 MC252 oil at five different time points.A. An illustrative model of the evolution of floc formation over time. Colours in the illustration correspond to the following: yellow, oil; blue,EPS, with associated microbial cells. The scale for the first three time points is in tens of mm and for the last two time points in hundredsof mm.B. Microscopic images of a floc at t = 0, 5, 10, 20 and 40 days.C. Distribution heat map of alkane C–H vibration modes in MC252 oil, of protein amide II vibration modes at ~ 1542 cm-1, of thecarbohydrate/EPS vibration modes at ~ 1000 cm-1, and of the carbonyl (C=O) vibration modes at ~ 1730 cm-1 in oil oxidation products.The corresponding maximum absorbance intensity (the value given in white) is given in absorbance units (a.u.).

6 J. Bælum et al.


Horizon oil spill were previously found to be largelyremoved by a combination of dispersion and microbialdegradation by the indigenous microbes in the deep sea(Hazen et al., 2010; Valentine et al., 2010; Atlas andHazen, 2011; Kessler et al., 2011; Lu et al., 2012). Hereenrichments were performed with high concentrations ofMC252 oil to identify specific members of the indigenouscommunity that could respond to high inputs of hydrocar-bons. Uncontaminated seawater collected from a depth of1100 mbsl in the Gulf of Mexico during the time of theDeepwater Horizon oil spill was used as a source ofinoculum. The oil served as a carbon substrate andinduced a substantial microbial bloom over a short 20-dayincubation period. When COREXIT 9500 was added,more of the oil was in solution and approximately twice asmuch was potentially available for degradation by bacteriain the seawater. The hydrocarbon components of thedispersant, COREXIT 9500, were also mineralized andserved as a carbon source for microbial growth.

Amendment with limiting nutrients, nitrogen and phos-phorous, has sometimes been used to increase the rate ofdegradation in previous oil spills (Bragg et al., 1994).However, the Gulf samples were replete in N and P,although low in Fe (Hazen et al., 2010). A positive impactof iron addition on biodegradation rates of oil has beenreported previously in habitats with iron-limiting conditions(Teralmoto et al., 2009; Wang et al., 2010). When weadded FeCl2 to the enrichments, we observed an initialincrease in degradation and an increase in microbialgrowth rates, but after 20 days of incubation no difference

in terms of amount of oil degraded was detected. Anothereffect of the addition of FeCl2 was the absence offloc formation. Therefore, although Fe concentrationswere low the natural microbial community still had thecapacity to grow and degrade the hydrocarbons withoutFe supplementation.

In agreement with this enrichment study, flocs were alsoobserved in The Gulf of Mexico in situ after substantialamounts of oil had been degraded following the DeepwaterHorizon oil spill (Hazen et al., 2010). To our knowledge nosuch floc formation has previously been reported in asso-ciation with oil degradation. However, marine flocs, alsoknown as marine snow, are otherwise a common phenom-enon in marine ecosystems (Azam and Malfatti, 2007),where they represent hot spots for nutrients and therebymicrobial activity (Azam and Long, 2001). We found thatcells began to aggregate immediately after addition of oil tothe water (Fig. 3A) and our SR-FTIR data suggest that theinitial floc formation occurred on the surface of oil droplets(Fig. 5C). Multiple aggregates formed within days, but withtime they merged together into one large floc. Eventually,the flocs were comprised of a complex structure of EPS,protein, oil, and oil degradation products and bacteria. Amodel of proposed steps in the floc formation based on ourcumulative data is shown in Fig. 5.

Interestingly, the flocs were primarily comprised of aselect group of bacteria that were enriched from the origi-nal source water, dominated by an OTU with closestmatch to Colwellia. Members of the Colwelliaceae havepreviously been shown to produce EPS under extremeconditions (i.e. low temperature, high pressure and lowsalinity) (Marx et al., 2009), and some species containgenes for enhanced survival in cold environments(Methé et al., 2005). Species of Methylococcaceae and

Fig. 7. Total hydrocarbons recovered from Colwellia sp. RC25incubations with Macondo (MC252) crude oil (100 mg l-1) and thedispersant, COREXIT 9500 (60 mg l-1) after aerobic incubation at5°C for 37 days. The data represent the mean of triplicate samples.

Fig. 6. Scanning electron microscopic (SEM) image of a singleColwellia sp. RC25 cell. The cell is approximately 1.5–2 mm longand approximately 0.25–0.4 mm wide.



Rhodobacteriaceae were also found in the flocs and thesehave representatives that have previously been associ-ated with degradation of hydrocarbons (Brakstad andLodeng, 2005; Coulon et al., 2007; McKew et al., 2007;Redmond et al., 2010). We propose that the ability ofsome members of the community, to produce EPS andaggregate into flocs was a key physiological mechanismfor the oil degrading community to aggregate togetherwith oil droplets and to conserve nutrients.

The same Colwellia sequence that we found in the flocswas also enriched in the water phase during the incuba-tions. Valentine and colleagues (2010) also reported thatColwellia sequences dominated near-well plume samplesduring the Deepwater Horizon spill, and Redmond andValentine (2011) showed that Colwellia sequences wereparticularly enriched at low temperature (4°C) comparedwith a higher temperature (20°C). Using stable isotopeprobing (SIP) Colwellia sequences were found to be themost abundant sequences in heavy DNA from incubationswith ethane, propane and benzene, suggesting that theyhad the capability to oxidize these compounds (Redmondand Valentine, 2011). An OTU with closest match toOleispira, a member of the Oceanospirillales previouslysuggested to be involved in hydrocarbon degradation(Yakimov et al., 2003), was also initially enriched in thepresence of oil. Subsequently, Colwellia sequencesincreased in relative amounts, similar to the successionpattern observed in situ during the Deepwater Horizon oilspill (Hazen et al., 2010; Valentine et al., 2010; Redmondand Valentine, 2011). One hypothesis is that there was anatural succession of the indigenous deep-sea microbiotaover time in the deep-sea oil plume, depending on thehydrocarbon fractions available to the microbial commu-nity, as also recently suggested by Redmond and Valen-tine (2011). Collectively, these findings suggest thatrepresentatives of the Collwelliaceae and Oceanospirilla-les play a predominant role in hydrocarbon degradation inthe deep sea. In particular, Colwellia could have a com-petitive advantage in the presence of high concentrationsof oil and dispersant due to their ability to produce EPSand form flocs.

In order to better understand the degradation capabilityand physiology of Colwellia in the enrichments, we suc-cessfully obtained an isolate (Colwellia sp. strain RC25)and demonstrated that it is able to degrade hydrocarbonsin crude oil at low temperature. A psychrotrophic lifestyleis apparently a key feature of the Colwellia genus. Forexample, Colwellia sequences have also been detected inoil contaminated sea ice (Brakstad et al., 2008) and thetype species of the genus (C. psychrerythraea 34H) wasisolated from arctic marine sediments (Huston et al.,2000). C. psychrerythraea 34H was genome sequenced,revealing several features that may contribute towards apsychrophilic lifestyle (Methé et al., 2005). Interestingly,

the C. psychrerythraea 34H genome included genes forproduction of EPS that were postulated to be important forbiofilm formation. In our study, SR-FTIR spectromicros-copy revealed that EPS were abundant in the flocs thatwere dominated by Colwellia. Therefore, it would be inter-esting to determine if the genes for this process are simi-larly found in strain RC25 in future studies. In addition thesequenced genome of C. psychrerythraea 34H suggestthe presence of putative dioxygenases and monooxyge-nases critical to ring cleavage and aliphatic compounddegradation (Methé et al., 2005). Together with ourresults, this suggests that Colwellia have potential fordegradation of hydrocarbons in cold marine habitats.

Here we also provide the first data regarding COREXIT9500 degradation by indigenous microbes from the Gulf,although other laboratories (Garcia et al., 2009) havedone studies on biodegradability of comparable dispers-ant agents. In situ evidence suggests a slow degradationof DOSS at plume depth after the Deepwater Horizon oilspill (Kujawinski et al., 2011). We found that both DOSSand the glycol compounds were degraded more slowlythan the hydrocarbon fraction in COREXIT 9500. Althoughnone of the glycol compounds were detected in situ in theGulf during surveillance cruises (OSAT, 2010), they aredesigned to be easily soluble in water and could havebecome so diluted that their concentrations were belowthe detection limit. One serious concern for the massiveuse of COREXIT 9500 during the MC252 oil spill was itsunknown toxic effects, but in our enrichments we found nonegative effects of high amounts of COREXIT on growthof indigenous microorganisms from the site.

In conclusion we demonstrated a high potential formicrobial degradation of oil in the deep sea of the Gulf ofMexico. The bacterial communities adapted rapidly to theintroduction of hydrocarbons with sequences representa-tive of first Oleispira, then Colwellia becoming dominant.These data help to explain how the indigenous microbialcommunity in the Gulf evolves when exposed to highinputs of hydrocarbons. The combination of 16S sequenc-ing and the SR-FTIR approach that we used enabled us tomonitor the evolution of flocs during exposure to highconcentrations of oil for the first time; the ecologicalsignificance of which remains to be investigated.

Experimental procedures

Water, oil and COREXIT 9500 collection

Uncontaminated water was collected from the Gulf ofMexico on 6 June 2010 during the active phase of theDeepwater Horizon oil spill on a cruise aboard the R/VOcean Veritas as described in Hazen and colleagues(2010). Water collected for this study was sampled alongwith the general monitoring done by the ship using a CTDsampling rosette (Sea-Bird Electronics, Bellevue, WA). The

8 J. Bælum et al.


exact position for the sampling was 28.6746°N, 88.3298°Wat 1100 mbsl. No oil was detected at this position in situusing a dual coloured dissolved organic matter (CDOM)WETstar fluorometer (WET Labs, Philomath, OR). Tempera-ture at the depth of the sampling site was 4.8°C, dissolvedoxygen (DO) was 6.07 mg l-1, and Fe was below the detec-tion limit (< 50 ppb). The water was stored and shipped inthe dark at 4°C to Lawrence Berkeley National Laboratory(LBNL, CA, USA).

Macondo (MC252) oil was sampled on 22 May 2010directly from the Discovery Enterprise drillship located abovethe wellhead during the Deepwater Horizon oil spill. Adetailed analysis of the oil composition is given in Hazen andcolleagues (2010). COREXIT 9500 (Nalco, Sugar Land, TX),the dispersant used during the oil spill, was kindly provided byThomas Azwell (UC Berkeley).

Enrichment set-up

Enrichments were established in 125 ml serum bottles with100 ml of uncontaminated water collected from the Gulf asdescribed above, together with combinations of 100 mg l-1

MC252 oil, 60 mg l-1 COREXIT 9500 and 0.1 mM FeCl2. Fivedifferent conditions were investigated; oil, oil + COREXIT,COREXIT, oil + FeCl2 and oil + COREXIT + FeCl2. Controlsincluded water alone, water sterilized with 10 mM NaN3 andwater sterilized with 10 mM NaN3 and supplemented with100 mg l-1 oil and 0.1 mM FeCl2. Additional enrichments wereset up with 100 ml of uncontaminated water from the Gulf and100 mg l-1 oil in order to study floc formation. Bottles weresealed with Teflon coated rubber stoppers and attached to aMicro-oxymax respirometer (Columbus Instruments, Ohio,USA) to measure CO2 evolution. The respirometer attach-ment had a closed loop that cycled the headspace with peri-odic measurement of CO2 evolution every second hour. Thetotal headspace in the bottles was 150 ml, including the bottleheadspace and headspace in the tubing of the respirometer,enough to maintain aerobic conditions throughout the experi-ment. As CO2 is present in seawater as carbonate, thenumbers obtained from the respirometer were corrected bymultiplying the measured concentrations by 2.3. This multi-plication factor was determined in a separate experimentwhere we investigated the CO2 uptake of sterile seawater at5°C and found that the ratio between CO2 dissolved in sea-water (including total inorganic carbon) and CO2 in the head-space was 2.3 (for further information see Supportinginformation, Fig. S1).

Because of the need to destructively sacrifice samples forhydrocarbon and DNA extraction, as described below, eachenrichment was set up with three replicates (n = 3) for sam-pling points 0, 5 and 20 days and after 2, 5, 10, 20 and 50days of incubation for the set of enrichments used to studyfloc formation. In addition, at each sampling point cells wereenumerated using a standard acridine orange staining proto-col (Francisc et al., 1973) and imaged with a FITC filter on aZeiss Axioskop (Carl Zeiss, Germany). The enrichments wereincubated at 5°C in the dark. To verify that aerobic conditionswere present throughout the incubation period, bottles wereopened in an anaerobic chamber and dissolved oxygen wasmeasured using the AccuVac colorimetric method (HACH,Loveland, USA).

Microbial isolation and identification

Subsamples collected from one of the enrichments with oiland COREXIT after 20 days of incubation were streaked ontoMarine Broth agar plates. After incubation at 5°C for 1 week,individual colonies appeared on the plates. Colonies weresubsequently individually transferred to modified minimalmarine medium (Coates et al., 1995), supplemented with 100ppm MC252 oil, 60 ppm COREXIT and 1 g l-1 bactopeptone.The ability of the cultures to degrade hydrocarbons wasdetermined as described below.

For 16S rRNA-based identification of the isolate, DNA wasextracted using the MoBio UltraClean Microbial DNA IsolationKit (MoBio, Carlsbad, CA). PCR amplification was carried outusing universal bacterial 16S rRNA gene primers 1492R and27F in 50 ml of reactions. Verified 16S amplicons were puri-fied using the procedure provided in the MoBio UltracleanPCR Clean-up kit (MoBio, Carlsbad, CA). Samples weresubmitted to the UC Berkeley DNA Sequencing Facility for16S rRNA sequencing using the Applied Biosystems BigDyeTerminator V3.1 Cycle Sequencing protocol and ABISequencing Analysis Software Version 5.1.

Scanning electron microscopy was performed on 0.2-micron-pore-size Millipore filters (Millipore, Billerica, MA,USA). A 1.0 ml aliquot of the suspended sample solution waspushed through the Millipore filters using a syringe. The filterswere then washed several times with 0.1 M sodium cacody-late buffer (pH 7.4) to remove excess material followed byon-ice fixation with 2% glutaraldehyde for 1 h, and on-icepost-fixation with 1% OsO4 for 1 h. Fixation was followed bydehydration with a graded ethanol series (20%, 40%, 50%,70%, 90%, 100%, 100%), critical point drying using a Tousi-mis AutoSamdri 815 Critical Point Dryer (Tousimis, Rockville,MD, USA), and sputter coating with gold-palladium usinga Tousimis Sputter Coater (Tousimis, Rockville, MD, USA).Images were collected using a Hitachi S5000 ScanningElectron Microscope (Hitachi High Technologies America,Pleasanton, CA, USA).

Hydrocarbon measurements

Samples were taken from the enrichments for hydrocarbonmeasurements after 0, 1, 5 and 20 days of incubation. Todetermine hydrocarbon concentrations derived from the pres-ence of oil in the samples, 500 ml of chloroform was added toa 5 ml of water sample and mixed by vortexing. The chloro-form fraction was removed and the water re-extracted twoadditional times and the extracts combined. The extract wasdried over anhydrous sodium sulfate and analysed withoutfurther concentration. Hydrocarbon concentrations weremeasured using a GC-FID (Agilent, Santa Clara, USA)enabling individual quantification of the majority of the crudeoil constituents as previously described (Hazen et al., 2010).Selected samples were run on a GC/MSD (Agilent, SantaClara, USA) for compound identification. Results are reportedas TPH and quantification was accomplished by comparisonto known concentrations of diluted MC252 oil standards. Inthe samples containing the COREXIT dispersant, 26 com-pounds were grouped into hydrocarbon, glycol and DOSSfractions and assessed.

To test the hydrocarbon degradation potential of Colwelliasp. strain RC25, a 10% inoculum of washed cells was



incubated at 5°C in serum bottles containing 50 ml of modi-fied minimal marine medium with 100 ppm MC252 oil and 60ppm COREXIT. To ensure aerobic conditions, 5 ml of air wasinjected into the serum bottles at regular intervals during theincubation period. At 3, 6, 10, 14, 21, 28 and 37 days ofincubation, triplicate samples were harvested for hydrocar-bon analysis as detailed above, with the following modifica-tions: (i) the entire 50 ml volume was sacrificed for eachanalysis and (ii) the bottle was washed three times withchloroform and these washes were added to the analyses.

DNA extraction and sequencing of enrichmentsand flocs

Subsamples were taken from the enrichments before thestart of the experiment (day 0), and after 5 and 20 days ofincubation. Eighty millilitres of water (including visible flocswhen present) were split into two 50 ml tubes and centrifugedat 18 000 g for 15 min. The supernatant was filtered througha 0.22 mm Sterivex filter (MilliPore, Billerica, USA). Filters andpellets were stored at -80°C until extracted. When samplingfloc material alone, intact flocs were carefully removed fromthe water with a pipette. The flocs were carefully rinsed threetimes in sterile filtered milliQ water before transferring to amicrocentrifuge tube for DNA extraction.

DNA was extracted as described in Hazen and colleagues(2010) with the following modifications: cell pellets were dis-solved in extraction buffer and transferred along with theentire filter to a 1.5 ml PULSE tube (Pressure BioSciences,South Easton, USA). The PULSE tube was placed in a Baro-cycler NEP 3229 (Pressure BioSciences, South Easton,USA) and the cells were lysed using pressure cycling with 20cycles of 20 s at 35 000 psi and 10 s at atmospheric pres-sure. After pressure lysis the entire solution and filter weretransferred to the Lysing Matrix E tube (MP Biomedicals,Solon, USA).

16S rRNA gene sequences were amplified from the DNAextracts using the primer pair 926f/1392r as described inEngelbrektson and colleagues (2010), with an additionalwobble added to the 926F primer to improve coverage of thearchaea (5′-cct atc ccc tgt gtg cct tgg cag tct cag aaa ctY aaaKga att gRc gg-3’, including titanium adapter sequence). Thereverse primer included a 5 bp barcode for multiplexing ofsamples during sequencing. Sequencing of the PCR ampli-cons was performed at DOE’s Joint Genome Institute (JGI) orat the GeneChip™ Microarray Core (San Diego, USA) usingRoche 454 GS FLX Titanium technology, with the exceptionthat the final dilution was 1e-8 (Allgaier et al., 2010). The rawsequence reads and quality files were deposited into theNCBI sequencing read archive under project numberSRA049463. The 16S rRNA gene sequence of Colwelliastrain RC25 has been deposited to NCBI GenBank withAccession No. JQ627834.

Community profiling

Pyrotag sequences were analysed using the QIIME pipeline(Caporaso et al., 2010b). Briefly, 16S rRNA gene sequenceswere clustered with uclust (Edgar, 2010) and assigned toOTUs with 97% similarity. Representative sequences from

each OTU were aligned with Pynast (Caporaso et al., 2010a)using the Greengenes core set. Taxonomy was assignedusing the Ribosomal Database Project’s Naïve Bayesianclassifier (Wang et al., 2007) with a confidence of 80%.

SR-FTIR spectromicroscopy of flocs

SR-FTIR spectromicroscopy in the mid infrared region(~4000–650 cm-1 wavenumber) was used to detect petro-leum products, petroleum degradation products, as well asmacromolecules of biological origin as previously described(Hazen et al., 2010; Holman et al., 2010). The advantage ofthis approach is that it has a signal-to-noise ratio 100–1000times better than conventional FTIR approaches (Holmanet al., 2010). For the present study, SR-FTIR analyses wereconducted on fresh flocs, which were gently rinsed withsterile filtered, deionized water in order to minimize interfer-ence from molecules in the seawater. The floc was placedonto an infrared transparent ZnSe disc prior to imaging. Foreach SR-FTIR imaging measurement, the entire view-field ofthe floc was divided into equal-sized 5 mm ¥ 5 mm squaresbefore scanning. Photons were focused through the flocusing a Nicolet Nic-Plan IR microscope (with a numericalaperture objective of 0.65), which was coupled to a NicoletMagna 760 FTIR bench (Thermo Scientific, MA, USA). TheSR-FTIR transmission spectra at each position were col-lected using a single-element MCT (mercury-cadmium-telluride) detector at a spectral resolution of 4 cm-1 with eightco-added scans and a peak position accuracy of 1/100 cm-1.Background spectra were acquired from locations withoutany floc material and were used as reference spectra for bothsamples and standards to remove background H2O and CO2

absorptions. A data cube of position-associated infraredspectra was obtained following each SR-FTIR data acquisi-tion experiment. This data cube was then subjected to anarray of data processing calculations using both Matlab andThermo Electron’s Omnic version 7.3, which included thecomputational conversion of each transmission to absorptionspectrum and baseline removal. Spectral absorption peaks,which could be linked to target molecules in the flocs, wereintegrated. The relative concentration of a particular chemicalcomponent is presented as absorbance units (a.u.).

Acknowledgements

This work was supported by a subcontract from the Universityof California at Berkeley, Energy Biosciences Institute toLawrence Berkeley National Laboratory, the Berkeley Syn-chrotron Infrared Structural Biology Program, the JointGenome Institute (JGI) under its US Department of Energycontract DE-AC02-05CH11231, and by the Danish ResearchCouncil FTP (09-069890).

References

Adcroft, A., Hallberg, R, Dunne, J.P., Samuels, B.L., Galt, JA,Barker, C.H., and Payton, D. (2010) Simulations ofunderwater plumes of dissolved oil in the Gulf ofMexico. Geophys Res Lett 37: L18605, doi:10.1029/2010GL044689.

10 J. Bælum et al.


Allgaier, M., Reddy, A., Park, J.I., Ivanova, N., D’Haeseleer,P., et al. (2010) Targeted discovery of glycoside hydrolasesfrom a switchgrass-adapted compost community. PLoSONE 5: 1–9.

Atlas, R.M., and Hazen, T.C. (2011) Oil biodegradation andbioremediation: a tale of the two worst spills in U.S. history.Environ Sci Technol 45: 6709–6715.

Azam, F., and Long, R.A. (2001) Oceanography – sea snowmicrocosms. Nature 414: 495–498.

Azam, F., and Malfatti, F. (2007) Microbial structuring ofmarine ecosystems. Nat Rev Microbiol 5: 782–791.

Berto, D., Giani, M., Taddei, P., and Bottura, G. (2005) Spec-troscopic evidence of the marine origin of mucilages inthe Northern Adriatic Sea. Sci Total Environ 353: 247–257.

Bragg, J.R., Prince, R.C., Harner, E.J., and Atlas, R.M.(1994) Effectiveness of bioremediation for the Exxon-Valdez oil-spill. Nature 368: 413–418.

Brakstad, O.G., and Lodeng, A.G. (2005) Microbial diversityduring biodegradation of crude oil in seawater from theNorth Sea. Microb Ecol 49: 94–103.

Brakstad, O.G., Nonstad, I., Faksness, L.G., and Brandvik,P.J. (2008) Responses of microbial communities in Arcticsea ice after contamination by crude petroleum oil. MicrobEcol 55: 540–552.

Camilli, R., Reddy, C.M., Yoerger, D.R., Van Mooy, B.A.S.,Jakuba, M.V., et al. (2010) Tracking hydrocarbon plumetransport and biodegradation at Deepwater Horizon.Science 330: 201–204.

Caporaso, J.G., Bittinger, K., Bushman, F.D., DeSantis, T.Z.,Andersen, G.L., et al. (2010a) PyNAST: a flexible tool foraligning sequences to a template alignment. Bioinformatics26: 266–267.

Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K.,Bushman, F.D., et al. (2010b) QIIME allows analysis ofhigh-throughput community sequencing data. Nat Methods7: 335–336.

Coates, J.D., Lonergan, D.J., Philips, E.J.P., Jenter, H., andLovley, D.R. (1995) Desulfuromonas palmitatis sp. nov., amarine dissimilatory Fe III reducer that can oxidize long-chain fatty acids. Arch Microbiol 164: 406–413.

Coulon, F., McKew, B.A., Osborn, A.M., McGenity, T.J., andTimmis, K.N. (2007) Effects of temperature and biostimu-lation on oil-degrading microbial communities in temperateestuarine waters. Environ Microbiol 9: 177–186.

Dasanayaka, L.K., and Yapa, P.D. (2009) Role of plumedynamics phase in a deepwater oil and gas release model.J Hydro-Environ Res 2: 243–253.

Edgar, R.C. (2010) Search and clustering orders of magni-tude faster than BLAST. Bioinformatics 26: 2460–2461.

Engelbrektson, A., Kunin, V., Wrighton, K.C., Zvenigorodsky,N., Chen, F., et al. (2010) Experimental factors affectingPCR-based estimates of microbial species richness andevenness. ISME J 4: 642–647.

Francisc, D.E., Mah, R.A., and Rabin, A.C. (1973) Acridineorange-epifluorescence technique for counting bacteria innatural waters. Trans Am Microsc Soc 92: 416–421.

Garcia, M.T., Campos, E., Marsal, A., and Ribosa, I. (2009)Biodegradability and toxicity of sulphonate-based surfac-tants in aerobic and anaerobic aquatic environments.Water Res 43: 295–302.

Gertler, C., Gerdts, G., Timmis, K.N., and Golyshin, P.N.(2009) Microbial consortia in mesocosm bioremediationtrial using oil sorbents, slow-release fertilizer and bioaug-mentation. FEMS Microbiol Ecol 69: 288–300.

Hazen, T.C., Dubinsky, E.A., DeSantis, T.Z., Andersen, G.L.,Piceno, Y.M., et al. (2010) Deep-sea oil plume enrichesindigenous oil-degrading bacteria. Science 330: 204–208.

Holman, H.Y.N., Bechtel, H.A., Hao, Z., and Martin, M.C.(2010) Synchrotron IR spectromicroscopy: chemistry ofliving cells. Anal Chem 82: 8757–8765.

Huston, A.L., Krieger-Brockett, B.B., and Deming, J.W.(2000) Remarkably low temperature optima for extracellu-lar enzyme activity from Arctic bacteria and sea ice.Environ Microbiol 2: 383–388.

Judson, R.S., Martin, M.T., Reif, D.M., Houck, K.A., Knudsen,T.B., et al. (2010) Analysis of eight oil spill dispersantsusing rapid, in vitro tests for endocrine and other biologicalactivity. Environ Sci Technol 44: 5979–5985.

Kasai, Y., Kishira, H., Sasaki, T., Syutsubo, K., Watanabe, K.,et al. (2002) Predominant growth of Alcanivorax strains inoil-contaminated and nutrient-supplemented sea water.Environ Microbiol 4: 141–147.

Kessler, J.D., Valentine, D.L., Redmond, M.C., Du, M., Chan,E.W., et al. (2011) A persistent oxygen anomaly reveals thefate of spilled methane in the deep Gulf of Mexico. Science331: 312–315.

Kujawinski, E.B., Soule, M.C.K., Valentine, D.L., Boysen,A.K., Longnecker, K., et al. (2011) Fate of dispersantsassociated with the Deepwater Horizon oil spill. Environ SciTechnol 45: 1298–1306.

Lu, Z.M., Deng, Y., Van Nostrand, J.D., He, Z.L., Voordeck-ers, J., Zhou, A.F., et al. (2012) Microbial gene functionsenriched in the Deepwater Horizon deep-sea oil plume.ISME J 6: 451–460.

McKew, B.A., Coulon, F., Osborn, A.M., Timmis, K.N., andMcGenity, T.J. (2007) Determining the identity and roles ofoil-metabolizing marine bacteria from the Thames estuary,UK. Environ Microbiol 9: 165–176.

Marx, J.G., Carpenter, S.D., and Deming, J.W. (2009) Pro-duction of cryoprotectant extracellular polysaccharidesubstances (EPS) by the marine psychrophilic bacteriumColwellia psychrerythraea strain 34H under extremeconditions. Can J Microbiol 55: 63–72.

Mecozzi, M., and Pietrantonio, E. (2006) Carbohydrates pro-teins and lipids in fulvic and humic acids of sediments andits relationships with mucilaginous aggregates in the Italianseas. Mar Chem 101: 27–39.

Methé, B.A., Nelson, K.E., Deming, J.W., Momen, B.,Melamud, E., Zhang, X., et al. (2005) The psychrophiliclifestyle as revealed by the genome sequence of Colwelliapsychtreryhraea 34H through genomic and proteomicanalysis. Proc Natl Acad Sci USA 102: 10913–10918.

Naumann, D. (2000) Infrared Spectroscopy in Microbiology.Chichester, UK: John Wiley & Sons.

Operational Science Advisory Team (OSAT) (2010) Summaryreport for sub-sea and sub-surface oil and dispersantdetection: sampling and monitoring. 17 December 2010.[WWW document]. URL http://www.restorethegulf.gov/release/2010/12/16/data-analysis-and-findings.

Redmond, M.C., and Valentine, D.L. (2011) Natural gas andtemperature structured a microbial community response to



http://www.restorethegulf.gov/release/2010/12/16/data-analysis-and-findings

http://www.restorethegulf.gov/release/2010/12/16/data-analysis-and-findings

the Deepwater Horizon oil spill. Proc Natl Acad Sci USA.doi:10.1073/pnas.

Redmond, M.C., Valentine, D.L., and Sessions, A.L. (2010)Identification of novel methane-, ethane-, and propane-oxidizing bacteria at marine hydrocarbon seeps by stableisotope probing. Appl Environ Microbiol 76: 6412–6422.

Teralmoto, M., Suzuki, M., Okazaki, F., Hatmanti, A., andHarayama, S. (2009) Oceanobacter-related bacteria areimportant for the degradation of petroleum aliphatic hydro-carbons in the tropical marine environment. Microbiology155: 3362–3370.

Valentine, D.L., Kessler, J.D., Redmond, M.C., Mendes, S.D.,Heintz, M.B., et al. (2010) Propane respiration jump-startsmicrobial response to a deep oil spill. Science 330: 208–211.

Wang, L.P., Wang, W.P., Lai, Q.L., and Shao, Z.Z. (2010)Gene diversity of CYP153A and AlkB alkane hydroxylasesin oil-degrading bacteria isolated from the Atlantic Ocean.Environ Microbiol 12: 1230–1242.

Wang, Q., Garrity, G.M., Tiedje, J.M., and Cole, J.R. (2007)Naive Bayesian classifier for rapid assignment of rRNAsequences into the new bacterial taxonomy. Appl EnvironMicrobiol 73: 5261–5267.

Yakimov, M.M., Giuliano, L., Gentile, G., Crisafi, E., Cherni-kova, T.N., et al. (2003) Oleispira antarctica gen. nov., spnov., a novel hydrocarbonoclastic marine bacterium iso-lated from Antarctic coastal sea water. Int J Syst EvolMicrobiol 53: 779–785.

Yapa, P.D., Dasanayaka, L.K., Bandara, U.C., and Nakata, K.(2008) Modelling the impact of an accidental release ofmethane gas in deepwater. Oceans 1–4: 109–119.

Zahed, M.A., Aziz, H.A., Isa, M.H., and Mohajeri, L. (2010)Effect of initial oil concentration and dispersant on crude oilbiodegradation in contaminated seawater. Bull EnvironContam Toxicol 84: 438–442.

Supporting information

Additional Supporting Information may be found in the onlineversion of this article:

1. Equilibrium between CO2 and inorganic carbon.Fig. S1. Seawater:headspace partitioning of artificial CO2

(CO2 added to the experiment).2. Mass balances.Fig. S2. Mass balances for the microcosms. Mass balancecalculations were based on the amount of carbon (C) in thedifferent fractions – bottle surface, dissolved in water, CO2

and biomass. MC252 oil contains ~ 85% C, COREXIT 9500contains ~ 70% C, CO2 27.3% C, and we assume that theratio between CO2 produced and C built into biomass is70:30.Fig. S3. The ratio between straight chain alkanes heptade-cane (n-C17) and octadecane (n-C18) and the isoprenoidspristine (pris) and phytane (phy) decreased over time,showing possible preferential degradation of alkanes.

Please note: Wiley-Blackwell are not responsible for thecontent or functionality of any supporting materials suppliedby the authors. Any queries (other than missing material)should be directed to the corresponding author for thearticle.

12 J. Bælum et al.


Deepsea bacteria enriched by oil and dispersant from the Deepwater

Documents