Efficacy of intervention strategies for bioremediationof crude oil in marine systems and effects onindigenous hydrocarbonoclastic bacteria

Boyd A. McKew,1* Frédéric Coulon,1

Michail M. Yakimov,2 Renata Denaro,2

Maria Genovese,2 Cindy J. Smith,1,3

A. Mark Osborn,1,3 Kenneth N. Timmis1,4 andTerry J. McGenity1

1Department of Biological Sciences, University of Essex,Wivenhoe Park, Colchester CO4 3SQ, UK.2Istituto per Ambiente Marino Costiere, IAMC-CNR,Sezione di Messina, Italy.3Department of Animal and Plant Sciences, TheUniversity of Sheffield, Sheffield S10 2TN, UK.4Division of Microbiology, Helmholtz Center for InfectionResearch, Inhoffenstrasse 7, D-38124 Braunschweig,Germany.

Summary

There is little information on how different strategiesfor the bioremediation of marine oil spills influencethe key indigenous hydrocarbon-degrading bacteria(hydrocarbonoclastic bacteria, HCB), and hence theirremediation efficacy. Therefore, we have used quan-titative polymerase chain reaction to analyse changesin concentrations of HCB in response to interventionstrategies applied to experimental microcosms.Biostimulation with nutrients (N and P) produced nomeasurable increase in either biodegradation or con-centration of HCB within the first 5 days, but after15 days there was a significant increase (29%; P <0.05) in degradation of n-alkanes, and an increase ofone order of magnitude in concentration of Thalas-solituus (to 107 cells ml-1). Rhamnolipid bioemulsifieradditions alone had little effect on biodegradation,but, in combination with nutrient additions, provokeda significant increase: 59% (P < 0.05) more n-alkanedegradation by 5 days than was achieved with nutri-ent additions alone. The very low Alcanivorax cellconcentrations in the microcosms were hardly influ-enced by addition of nutrients or bioemulsifier, butstrongly increased after their combined addition,

reflecting the synergistic action of the two types ofbiostimulatory agents. Bioaugmentation with Thalas-solituus positively influenced hydrocarbon degrada-tion only during the initial 5 days and only of then-alkane fraction. Bioaugmentation with Alcanivoraxwas clearly much more effective, resulting in 73%greater degradation of n-alkanes, 59% of branchedalkanes, and 28% of polynuclear aromatic hydrocar-bons, in the first 5 days than that obtained throughnutrient addition alone (P < 0.01). Enhanced degrada-tion due to augmentation with Alcanivorax continuedthroughout the 30-day period of the experiment. Inaddition to providing insight into the factors limitingoil biodegradation over time, and the competition andsynergism between HCB, these results add weight tothe use of bioaugmentation in oil pollution mitigationstrategies.

Introduction

Oil pollution represents a significant threat to the marineenvironment. Although large spills such as that from thePrestige oil tanker in 2002 (63 000 tonnes) are relativelyrare, in 2005 alone there were 21 tanker spills of between7 and 700 tonnes and three in excess of 700 tonnes(ITOPF, 2006). Fortunately, natural biodegradation pro-cesses remove considerable quantities of oil from themarine environment, with indigenous bacteria playing thedominant role (Leahy and Colwell, 1990). As a result,since the earliest oil spills, there has been increasinginterest in the application of bioremediation to enhance oildegradation and thereby mitigate ecological damagecaused by oil spills. Such bioremediation strategies areoften based on stimulating the indigenous microbialcommunity (biostimulation), or involve amendment of thenaturally occurring microbial community with an inoculumof hydrocarbon-degrading bacteria (bioaugmentation).

Typically the rate-limiting factor for oil degradation in themarine environment is inorganic nutrient concentration,and in particular, nitrogen (N) and phosphorous (P). Bio-stimulation by the addition of nutrients has been the mostwidely practised bioremediation strategy, and it is welldocumented that degradation of oil is often significantlyenhanced by this method (Bragg et al., 1994; Swannell

Received 7 November, 2006; accepted 12 February, 2007. *Forcorrespondence. E-mail bamcke@essex.ac.uk; Tel. (+44) 1206872547; Fax (+44) 1206 872592.

Environmental Microbiology (2007) 9(6), 1562–1571 doi:10.1111/j.1462-2920.2007.01277.x

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd

et al., 1995; 1999; Venosa et al., 1996; Wright et al., 1997;Röling et al., 2002). For example, hydrocarbon degrada-tion was enhanced by a factor of 8.6 during one of thebiggest biostimulation treatments, in which approximately50 000 kg of N and 5000 kg of P were added over suc-cessive summers following the Exxon Valdez spill (Bragget al., 1994).

Bioavailability of the hydrocarbons present in crude oilcan also be a limiting factor owing to their hydrophobicnature and low water solubility. These properties coupledwith the fact that bacteria must come into contact with thehydrocarbons so that membrane-bound oxygenases caninitiate degradation mean that bioavailability can begreatly enhanced by emulsification of oil (Ron and Rosen-berg, 2002). Consequently, most hydrocarbon-degradingbacteria have been found to produce biosurfactants (Ronand Rosenberg, 2002). The addition of surfactants thatsupplement those produced naturally is another widelyadopted strategy to enhance degradation of oil. Surfac-tants reduce surface tension and increase the surfacearea of hydrophobic compounds such as oil, thereforeincreasing its bioavailability. However, the effectiveness ofsurfactant additions in increasing the rate of oil degrada-tion has been mixed (Leahy and Colwell, 1990). Gener-ally, biosurfactants are preferred to synthetic surfactantsas they are selective, less toxic, biodegradable and gen-erally more effective at increasing hydrocarbon bioavail-ability (Rambeloarisoa et al., 1984; Harvey et al., 1990;Iwabuchi et al., 2002; Ron and Rosenberg, 2002).

While there is evidence that bioremediation, and inparticular biostimulation strategies, can sometimesenhance oil degradation, there is limited information onhow these strategies affect the in situ oil-degradingcommunity. In recent years it has been established thatbacteria from the genera Thalassolituus, Alcanivorax andCycloclasticus play important roles in marine oil degrada-tion (Dyksterhouse et al., 1995; Geiselbrecht et al., 1998;Yakimov et al., 1998; Kasai et al., 2002; Hara et al., 2003;Yakimov et al., 2004; 2005). Most recently, our studies onthe Thames Estuary have also shown that these bacteriaare dominant in hydrocarbon- and crude oil-enrichedmicrocosms (McKew et al., 2007), and that co-addition ofN and P enhanced crude oil degradation (Coulon et al.,2007).

Bioaugmentation, the addition of oil degraders, is con-troversial, and in many cases has proved to be ineffective(Tagger et al., 1983; Venosa et al., 1996; MacNaughtonet al., 1999). As some bioaugmentation studies may havebeen carried out with bacteria not optimally adapted forsurvival and growth within the marine environment, wehave focused attention on marine oil-degrading special-ists, the hydrocarbonoclastic bacteria (HCB), namelyAlcanivorax borkumensis (Yakimov et al., 1998) andThalassolituus oleivorans (Yakimov et al., 2004; McKew

et al., 2007). We have investigated the effects of bioaug-mentation with these organisms upon hydrocarbon deg-radation in marine waters, and additionally compared therole of bioaugmentation with biostimulation via nutrient (Nand P) and/or rhamnolipid biosurfactant amendment.Importantly, we have assessed the effects of thesevarious bioremediation strategies on cell concentrationsof Thalassolituus, Alcanivorax and Cycloclasticus withinthe oil-degrading community using real-time quantitativepolymerase chain reaction (Q-PCR). Our findings offerinsights into the interplay between different hydrocarbon-degrading bacterial species during bioremediation, dem-onstrate the importance of Alcanivorax for bioremediation,and affirm the potential role of bioaugmentation as a strat-egy for mitigation of oil-spill pollution.

Results

Seawater microcosms containing 0.1% v/v weatheredcrude oil were established with nine different treatmentscomprising permutations of amendments with nutrients (Nand P), bioemulsifiers and added HCB (Table 1). Micro-cosms were destructively sampled at three different timepoints: day 5 (Fig. 1), day 15 (Fig. 2) and day 30 (Fig. 3),at which times, the concentration of n-alkanes, branchedalkanes and polynuclear aromatic hydrocarbons (PAHs)remaining in the microcosms were determined. At eachtime point, the abundance of Thalassolituus, Alcanivoraxand Cycloclasticus were quantified by Q-PCR amplifica-tion of their functional genes (alkB, alkB2 and phnA,encoding alkane hydroxylases or aromatic ring-hydroxylating dioxygenase respectively).

After 5 days (Fig. 1), no significant hydrocarbon degra-dation had occurred in the oil-only microcosms (Oil). Addi-tion of nutrients on their own (NP) had no observable effecton degradation, while addition of bioemulsifier (JBR-0.1and JBR-1) led to a small but not significant reduction ofn-alkane concentrations (9–25%) (Fig. 1A). In contrast, inthose microcosms that received simultaneous addition ofnutrients and emulsifier, n-alkanes were degraded by 41%(NP + JBR-1) and 52% (NP + JBR-0.1), which was signifi-cantly greater (F = 9.25; P < 0.001) than in both the oil-only(Oil) and nutrient-enriched (NP) microcosms. Furthermore,in all of the bioaugmented microcosms, degradation ofn-alkanes was significantly greater than in the oil-onlymicrocosms, with the highest levels of degradation(66%) found in the Alcanivorax-enriched microcosm(NP + Alcanivorax). For the branched alkanes, pristaneand phytane (Fig. 1B), there was again no degradationapparent in the oil-only (Oil) and nutrient-enriched micro-cosms (NP). Significant degradation of branched alkaneswas seen only in those microcosms augmented with Alca-nivorax (F = 6.4; P < 0.001) and in particular when addedsingly (NP + Alcanivorax), in which 58% degradation was

Effectiveness of crude oil bioremediation strategies 1563

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

found. Similarly, total PAH degradation was also the great-est (50% degradation; F = 3.85; P < 0.001) in microcosmsin which Alcanivorax was added singly (NP + Alcanivorax)(Fig. 1C).

Quantitative polymerase chain reaction analysis(Fig. 1D) showed that at 5 days, Thalassolituus numbers

had increased in microcosms containing oil only (Oil) andin all microcosms subjected to biostimulation (NP,JBR-0.1, JBR-1, NP + JBR-0.1 and NP + JBR-1) withcell numbers of between 5 ¥ 105 and 4 ¥ 106 cells ml-1.In microcosms augmented with Alcanivorax(NP + Alcanivorax), Thalassolituus numbers increased to

Table 1. Microcosm design.

Microcosmdesignation

WeatheredForties crude oil0.1% v/v

Nutrients (20 mg l-1

NH4NO3 and10 mg l-1 KH2PO4)

Rhamnolipid bioemulsifierJBR-215 (% v/vactive ingredient)

Alcanivoraxborkumensis (typestrain SK2)(cells ml-1)

Thalassolituusoleivorans(strain SLHC162b)(cells ml-1)

Oil +NP + +JBR-0.1 + 0.1JBR-1 + 1NP + JBR-0.1 + + 0.1NP + JBR-1 + + 1NP + Alcanivorax + + 1 ¥ 105

NP + Thalassolituus + + 1 ¥ 105

NP + both A+T + + 5 ¥ 104 5 ¥ 104

Summary of nutrient and bioemulsifier amendments and bioaugmentation with Alcanivorax and Thalassolituus within the oil-enriched microcosms.No-oil controls contained 20 ml of seawater only. Killed controls contained 20 ml of seawater, plus 0.1% v/v weathered Forties crude oil andmercuric chloride (300 mg l-1).

A B

DC

Fig. 1. Hydrocarbon degradation and abundance of hydrocarbonoclastic bacteria after 5 days in seawater microcosms enriched with crude oiland subjected to bioremediation treatments (as indicated). Degradation of (A) n-alkanes, (B) branched alkanes (pristane + phytane) and (C)PAHs, determined by GC-MS; and (D) the abundance of hydrocarbonoclastic bacteria (Thalassolituus, Alcanivorax and Cycloclasticus)quantified by Q-PCR (means � SE; n = 3). Dashed lines indicate levels of hydrocarbons (means � SE) measured in the killed controls. Theasterisk (*) indicates microcosms in which degradation was significantly greater (P < 0.05) than the oil-only (Oil) microcosms. The hash symbol‘#’ indicates microcosms in which degradation was significantly greater (P < 0.05) than both the oil-only (Oil) and nutrient-enriched (NP)microcosms. The cross symbol ‘¥’ indicates microcosms in which degradation was not significant (P > 0.05) relative to the killed controls.

1564 B. A. McKew et al.

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

5 ¥ 105 cells ml-1. In the microcosms to which Thalassoli-tuus had been added initially (NP + Thalassolituus andNP + both A+T) Thalassolituus cell numbers were signifi-cantly higher (F = 59; P < 0.001), by up to three orders ofmagnitude (8 ¥ 107 to 1 ¥ 108 cells ml-1). Alcanivoraxincreased to 7 ¥ 107 cells ml-1 in the microcosms where itwas added singly (NP + Alcanivorax), approximatelytwofold higher than it was in the microcosms in which itwas added simultaneously with Thalassolituus (NP + bothA+T). Alcanivorax, however, was only detected in thosemicrocosms to which it had been added initially, with theexception of one of the replicate microcosms containing a1% bioemulsifier addition (JBR-1), where Alcanivorax waspresent at 2 ¥ 104 cells ml-1. There was little variation innumbers of Cycloclasticus at day 5, with populationsincreasing to ~106 cells ml-1 in all microcosms regardlessof the bioremediation treatment. Thalassolituus, Alcanivo-rax and Cycloclasticus were not detected in the originalseawater, nor in the no-oil controls at any stage of theexperiment (data not shown).

After 15 days (Fig. 2) n-alkane degradation hadincreased in all microcosms (Fig. 2A). In the oil-only

microcosms (Oil), 40% of total n-alkanes were degraded.Emulsifier additions alone (JBR-0.1 and JBR-1) still hadno significant effect on alkane degradation when com-pared with the oil-only microcosms (Oil). As at day 5,n-alkane degradation (64–78%) was significantly greater(F = 20.9; P < 0.001) in microcosms amended with bothnutrients and bioemulsifier (NP + JBR-0.1 andNP + JBR-1) and in all of the bioaugmented microcosmswhen compared with the oil only (Oil) microcosms.However, in contrast to day 5, degradation of n-alkanes inthe nutrient-enriched microcosms (NP) had nowincreased to 69%. Pristane and phytane remained unde-graded in the oil-only (Oil) and nutrient-enriched (NP)microcosms (Fig. 2B), and while there was some degra-dation in the nutrient-plus-bioemulsifier microcosms(NP + JBR-0.1 and NP + JBR-1), it was still only in theAlcanivorax-augmented microcosms (NP + Alcanivorax)where degradation was significantly (F = 4.7; P < 0.001)greater than in the nutrient-enriched microcosm (NP),although most of this degradation had occurred within thefirst 5 days. By day 15, PAH degradation had increased inall microcosms, including that containing only oil (Oil), with

A B

DC

Fig. 2. Hydrocarbon degradation and abundance of hydrocarbonoclastic bacteria after 15 days in seawater microcosms enriched with crudeoil and subjected to bioremediation treatments (as indicated). Degradation of (A) n-alkanes, (B) branched alkanes (pristane + phytane) and (C)PAHs, determined by GC-MS; and (D) the abundance of hydrocarbonoclastic bacteria (Thalassolituus, Alcanivorax and Cycloclasticus)quantified by Q-PCR (means � SE; n = 3). Dashed lines indicate levels of hydrocarbons (means � SE) measured in the killed controls. Theasterisk (*) indicates microcosms in which degradation was significantly greater (P < 0.05) than the oil-only (Oil) microcosms. The hash symbol‘#’ indicates microcosms in which degradation was significantly greater (P < 0.05) than both the oil-only (Oil) and nutrient-enriched (NP)microcosms. The cross symbol ‘¥’ indicates microcosms in which degradation was not significant (P > 0.05) relative to the killed controls.

Effectiveness of crude oil bioremediation strategies 1565

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

38% of total PAHs now degraded (Fig. 2C). As at day 5,degradation of branched alkanes and also PAHs in theAlcanivorax-augmented microcosms (NP + Alcanivorax)remained significantly greater (F = 15.2; P < 0.001) (72%)than in the oil-only microcosms (Oil).

At day 15, Thalassolituus numbers were in the region of107 cells ml-1 in those microcosms to which it was origi-nally added, but in addition, Thalassolituus cell numbershad increased to similar levels in microcosms withnutrients only (NP) and nutrients plus emulsifier(NP + JBR-0.1 and NP + JBR-1) (Fig. 2D). In the oil-only(Oil) and Alcanivorax-augmented (NP + Alcanivorax)microcosms Thalassolituus numbers were significantlylower (F = 6.9; P < 0.001) than in all other microcosms, byup to two orders of magnitude. Alcanivorax numbersremained highest (107 cells ml-1) in the Alcanivorax-augmented microcosms, but it was now also detected inthe nutrient-plus-bioemulsifier microcosms (NP + JBR-0.1and NP + JBR-1) at 106 and 107 cells ml-1, respectively, aswell as in the 1% bioemulsifier-only microcosms (JBR-1)at 104 cells ml-1. However, Alcanivorax remained undetec-ted in the oil-only (Oil) and nutrient-enriched (NP)

microcosms, in the microcosms with the lower bioemulsi-fier concentration (JBR-0.1) and in the Thalassolituus-augmented microcosms (NP + Thalassolituus). Cyclo-clasticus numbers were the lowest (6 ¥ 105 cellsml-1) in the Thalassolituus-augmented microcosms(NP + Thalassolituus), and ~106 cells ml-1 in most othermicrocosms. Highest numbers of Cycloclasticus (3 ¥ 107

cells ml-1) were found in microcosms containing 1%bioemulsifier (JBR-1 and NP + JBR-1).

After 30 days (Fig. 3) significantly greater n-alkane deg-radation (between 70% and 85%) was found in all micro-cosms than in the oil-only (Oil) microcosms (56%), withthe exception of those amended with bioemulsifier only(JBR-0.1 and JBR-1) (Fig. 3A). There was minimalchange in degradation of branched alkanes (Fig. 3B) fromday 15 to day 30. It is notable that degradation of pristaneand phytane positively correlated (r = 0.88; P < 0.01) withnumbers of Alcanivorax.

Similar levels of degradation of PAHs were evident in allmicrocosms (Fig. 3C), ranging from 65% in the oil-onlymicrocosms (Oil) to 78% in the Alcanivorax-augmentedmicrocosm (NP + Alcanivorax), which was the only micro-

A B

DC

Fig. 3. Hydrocarbon degradation and abundance of hydrocarbonoclastic bacteria after 30 days in seawater microcosms enriched with crudeoil and subjected to bioremediation treatments (as indicated). Degradation of (A) n-alkanes, (B) branched alkanes (pristane + phytane) and (C)PAHs, determined by GC-MS; and (D) the abundance of hydrocarbonoclastic bacteria (Thalassolituus, Alcanivorax and Cycloclasticus)quantified by Q-PCR (means � SE; n = 3). Dashed lines indicate levels of hydrocarbons (means � SE) measured in the killed controls. Theasterisk (*) indicates microcosms in which degradation was significantly greater (P < 0.05) than the oil-only (Oil) microcosms. The hash symbol‘#’ indicates microcosms in which degradation was significantly greater (P < 0.05) than the oil-only (Oil) and nutrient-enriched (NP)microcosms. The cross symbol ‘¥’ indicates microcosms in which degradation was not significant (P > 0.05) relative to the killed controls.

1566 B. A. McKew et al.

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

cosm with significantly (F = 70.5; P < 0.01) greater PAHdegradation than the oil-only microcosms (Oil). By day 30,97–99% of all naphthalene compounds were degraded inall microcosms (data not shown), leaving mainly the morerecalcitrant three- and four-ring PAHs. In the Alcanivorax-augmented microcosms, degradation of even thefour-ring compounds: pyrene, C1-pyrene, C2-pyrene,chrysene and C1-Chrysene (60% degradation), wassignificantly greater (F = 8.1; P < 0.001) than in theoil-only (Oil), nutrient-enriched (NP) and nutrient-plus-bioemulsifier (NP + JBR-0.1 and NP + JBR-1) micro-cosms (26%, 27%, 32–33% degradation, respectively;data not shown). Numbers of Alcanivorax were positivelycorrelated (r = 0.77; P < 0.01) with the total amount ofdegradation of these high-molecular-weight four-ringPAHs.

Alcanivorax remained undetected in the oil-only (Oil)and the Thalassolituus-augmented (NP + Thalassolituus)microcosms (Fig. 3D), but could now be detected at levelsof ~106 cells ml-1 in the nutrient-enriched microcosms(NP), and ~104-105 cells ml-1 in the bioemulsifier-treatedmicrocosms (JBR-0.1 and JBR-1). Alcanivorax cellnumbers were significantly greater (F = 60.5; P < 0.001) inmicrocosms amended with nutrient-plus-bioemulsifieradditions (NP + JBR) with ~107 cells ml-1, and also in theAlcanivorax-augmented microcosm (NP + Alcanivorax),where numbers were the highest (~108 cells ml-1). Cyclo-clasticus cell numbers ranged from ~105 to 106 cells ml-1

in most microcosms. However, Cycloclasticus numberswere significantly (F = 7.4; P < 0.001) greater (107

cells ml-1) in those microcosms augmented with Alcanivo-rax (NP + Alcanivorax) and in the NP + JBR-1 microcosmwhen compared with the oil-only (Oil) and nutrient-enriched (NP) microcosms. Thalassolituus numbersranged from ~106 to 107 in most microcosms, butsignificant differences were only found betweenthe Thalassolituus-augmented microcosms, in whichnumbers were the highest (2 ¥ 107 to 6 ¥ 107 cells ml-1),and the Alcanivorax-augmented (NP + Alcanivorax) andoil-only (Oil) microcosms, in which they were at theirlowest (9 ¥ 105 and 1 ¥ 106 cells ml-1 respectively).

Discussion

Biostimulation specifically encouraged growthof the main HCB

At 15 and 30 days, biostimulation by N and P addition (toalleviate nutrient limitation) proved an effective strategyfor enhanced oil degradation, a phenomenon that hasoften been observed (Lindstrom et al., 1991; Swannellet al., 1995; 1999; Venosa et al., 1996; Wright et al., 1997;Röling et al., 2002; Coulon et al., 2007). Increasednumbers of hydrocarbon degraders following nutrient bio-

stimulation have been shown by MPN (Most ProbableNumber) counts (Lindstrom et al., 1991; Venosa et al.,1996), and in the Thames Estuary an increase of oneorder of magnitude has been observed (Coulon et al.,2007). However, in this study, using Q-PCR, nutrient addi-tion has been shown to specifically enhance numbers ofThalassolituus, Alcanivorax and Cycloclasticus species,which we have previously shown to play the dominantroles in n-alkane, branched alkane and PAH degradationrespectively (McKew et al., 2007). Following nutrient addi-tion, we observed an increase in degradation ofn-alkanes, which are the largest component of crude oil,comprising around 65% of the total extractable hydrocar-bons within the weathered Forties crude oil. We proposethat this nutrient-enhanced n-alkane degradation is prima-rily mediated by Thalassolituus, which increased innumbers by more than one order of magnitude relative tothe oil-only microcosms (Figs 2D and 3D).

Alcanivorax, in particular, responded favourably to bio-stimulation treatments, as, although it remained undetec-ted in the oil-only microcosms, Alcanivorax numbers weresignificantly enhanced, first by combined nutrient andbioemulsifier treatments, but also when addedindependently. The requirement for nutrient addition tostimulate growth of Alcanivorax has been observed pre-viously (Kasai et al., 2002) and there is evidence thatdifferent genotypes of Alcanivorax are adapted to differentconcentrations of nutrients (see Head et al., 2006).

Biosurfactants enhance early degradation ofhydrocarbons when nutrients are not limiting

Emulsification of oil via addition of the rhamnolipid biosur-factant JBR-215 as the sole bioremediation strategy hadno significant effect on degradation at any stage of theexperiments, suggesting that if nutrients are limiting,increasing bioavailability of the oil by emulsification willhave a minimal effect on degradation. While biostimula-tion with nutrients was effective in the longer term, thisstrategy had no significant effect on degradation by day 5.However, the large increases in degradation seen in thosemicrocosms that were supplemented with both bioemul-sifier and nutrients, and also in all bioaugmented micro-cosms, suggest that, in the early stages, it is not onlynutrients that are a limiting factor, but that degradation isco-limited by hydrocarbon bioavailability and/or by limitednumbers of hydrocarbon-degrading bacteria. The greaterdegradation in those microcosms supplemented with bothnutrients and bioemulsifier (compared with nutrientsalone) suggests that by increasing hydrocarbon bioavail-ability the lag phase of microbial growth can be shortened.Exponential growth is not possible if the surface area of oilis limiting, and biomass will only increase arithmetically(Ron and Rosenberg, 2002). In the nutrient-only micro-

Effectiveness of crude oil bioremediation strategies 1567

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

cosms degradation of n-alkanes by day 15 increased tolevels similar to microcosms amended with both nutrientsand bioemulsifier, and so it is likely that the in situoil-degrading community produced their own surfactantsand consequently that hydrocarbon bioavailability was nolonger a co-limiting factor.

Bioaugmentation could be an effective remediationstrategy for the crucial early stages of an oil spill

Although combined bioaugmentation and nutrient addi-tions resulted in enhanced degradation of n-alkanes bydays 15 and 30, these treatments were no more effectiveat enhancing degradation than nutrient additions alone.However, bioaugmentation was effective in the earlystages of degradation of n-alkanes, as shown by theadditions of Thalassolituus and Alcanivorax indepen-dently, or together, that resulted in significantly increaseddegradation at day 5. In some cases, addition of alloch-thonous bacteria has been reported as unsuccessful, atleast via culture-based analysis, with such organismsdisappearing from oil-impacted waters within 30 days(Tagger et al., 1983). In contrast, in our study, cellnumbers of both Alcanivorax and Thalassolituusincreased and remained high throughout the experiment,probably because they are the naturally dominantbacteria in hydrocarbon-degrading communities from theThames Estuary (Coulon et al., 2007; McKew et al.,2007).

Addition of a non-PAH-degrading bacterium enhancesPAH degradation

Bioaugmentation with Alcanivorax led to enhanced deg-radation of the branched alkanes (pristane and phytane)over the course of the entire 30-day experiment. Thisconfirms our previous finding that in the Thames Estuary,it is Alcanivorax that is primarily responsible for branchedalkane degradation within crude oil (McKew et al., 2007).This is further supported by significant positive correla-tions between Alcanivorax cell numbers and degradationof these substrates. More surprising, however, was thepositive effect that bioaugmentation with Alcanivorax hadon PAH degradation. Alcanivorax borkumensis strain SK2does not degrade PAHs (Yakimov et al., 1998) yetdegradation of a further 28% of PAHs occurred by day 5 inAlcanivorax-augmented microcosms, relative to thenutrient-enriched microcosms, suggesting a potentialsynergistic relationship between Alcanivorax and PAH-degrading Cycloclasticus or other PAH degraders. Bacte-ria that cannot grow on particular hydrocarbons havepreviously been shown to impart an important synergisticeffect in the degradation of petroleum (Rambeloarisoaet al., 1984). While A. borkumensis cannot degrade

PAHs, it produces a powerful extracellular glucose lipidbiosurfactant (Yakimov et al., 1998) that may increasebioavailability of PAHs, enabling enhanced degradationby Cycloclasticus. This potential synergism is further sup-ported by the fact that at day 30, microcosms that werebioaugmented with Alcanivorax (NP + Alcanivorax), andalso those supplemented with nutrients and 1% bioemul-sifier (NP + JBR-1), were the only microcosms in whichCycloclasticus numbers were significantly greater (F =7.458; P < 0.01) than in the nutrient-enriched (NP) andoil-only (Oil) microcosms. Also at day 30, Alcanivorax andCycloclasticus numbers were positively correlated (r =0.629; P < 0.01).

Alternatively, Alcanivorax may enhance PAH degrada-tion by utilizing the alkyl side-chains on methylated PAHswithin the crude oil, as previously suggested by Iwabuchiand colleagues (2002), when Alcanivorax was dominantin PAH-enriched cultures. Alternatively, increased degra-dation of PAHs may not be directly attributable toincreased numbers of Alcanivorax, but instead to thereduced numbers of Thalassolituus. Competition betweenThalassolituus and Alcanivorax has been previously dem-onstrated (Yakimov et al., 2005; McKew et al., 2007) andit has been further suggested that Thalassolituus mayproduce metabolites that are inhibitory to other bacteria(Yakimov et al., 2005). This may explain why Alcanivoraxcould not be detected following bioaugmentation withThalassolituus. Similarly, although Thalassolituus is not indirect competition with Cycloclasticus for carbon andenergy sources, it is in competition for nutrients, andso Cycloclasticus may also be inhibited indirectly byThalassolituus. This potential antagonistic effect byThalassolituus is further supported by the observationthat following bioaugmentation with Thalassolituus(NP + Thalassolituus), Alcanivorax (NP + Alcanivorax) orboth (NP + both A+T) there was a significant negativecorrelation between the abundance of Thalassolituus andAlcanivorax (r = -0.461; P < 0.05), and Thalassolituusand Cycloclasticus (r = -0.451; P < 0.05).

Bioaugmentation with A. borkumensis at the earlystages of an oil spill could be an importantbioremediation strategy

The damage caused by an oil slick is most pronouncedwhen it reaches the shoreline and its toxic effects arefrequently dose dependent. Consequently, rapid (bio)re-mediation is essential to reduce hydrocarbon concentra-tions to subcritical levels, and thereby minimizeecosystem impairment. In this study we have highlightedthe factors that limit biodegradation of oil at different timesover a 30-day period, and shown how some of the mostimportant HCB are affected. The addition of nutrients ledto more rapid n-alkane degradation, and, although in the

1568 B. A. McKew et al.

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

long term biostimulation treatments other than nutrientsupplementation seem to be unnecessary, the additionalapplication of a bioemulsifier may have a dramatic impacton degradation in the crucial few days immediately follow-ing an oil spill. All of the bioremediation strategies testedhad an impact on the community composition, butthe growth of Alcanivorax populations in particular wasdependent on biostimulation with either nutrients and/orbioemulsifier. For bioaugmentation to be successful theaddition of appropriate bacteria, with consideration toboth the environment and pollutant, is of paramountimportance. Here we have demonstrated that bioaugmen-tation with A. borkumensis was particularly effective, asits addition significantly increased degradation ofn-alkanes, branched alkanes and PAHs during the first5 days. It is appreciated that in vitro experiments may notnecessarily reflect what happens in situ following a real oilpollution event (see Head et al., 2006). In particular, theproblems of the dilution of added nutrients, bioemulsifiersand microorganisms into marine ecosystems need to beovercome, while increased grazing pressure may addi-tionally influence the bacterial populations. Moreover,marine oil spills are unpredictable and occur on a vastscale and so the availability of a viable inoculum could beproblematic. Nevertheless, our results suggest that suchapproaches have an exciting potential as bioremediationstrategies, and warrant further research in the field.

Experimental procedures

Sample site and sample collection

Water samples were collected in February 2006 fromStanford le Hope saltmarsh (51°30′N, 0°27′E) neighbouringthe Shell-Haven and Coryton BP oil refineries, situated atthe mouth of the Thames Estuary on the south-east coast,Essex, UK. The site has been characterized previously(Coulon et al., 2007). Triplicate 2 l of seawater samples werecollected in sterile bottles at high tide for microcosmexperiments.

Microcosm design

Aerobic microcosms were established in triplicate in sterile40 ml vials with PTFE septa containing 20 ml of seawateramended with 0.1% v/v weathered Forties crude oil (weath-ered by distillation at 230°C to remove the volatile fraction,and to sterilize) and amended with different combinations ofnutrients, a rhamnolipid bioemulsifier JBR-215 (Jeneil Biosur-factant Company, WI, USA) and live exponential-phasehydrocarbon-degrading bacteria (Table 1) pre-grown in artifi-cial seawater minimal media ON7Ra (Dyksterhouse et al.,1995) with tetradecane (0.1% w/v) as the sole carbon source.Triplicate killed controls (seawater plus 0.1% v/v weatheredForties crude oil and mercuric chloride at 300 mg l-1) weresampled at each time point to determine whether any hydro-carbon losses were abiotic. All microcosms were incubated at

12°C on an orbital shaker (110 rpm). Triplicate microcosmswere analysed at days 5, 15 and 30 to determine degradationof petroleum hydrocarbons by gas chromatography-massspectrometry (GC-MS), and to measure numbers of Alca-nivorax, Cycloclasticus and Thalassolituus by Q-PCR usingseparate specific primers.

Gas chromatography-mass spectrometry analysis

Hydrocarbons were solvent extracted from microcosms with6 ml of hexane–dichloromethane (1:1) and samples furtherdiluted (three parts extracted sample into seven parts hexanecontaining deuterated alkanes C10d22, C19d40, C30d62 andPAHs naphthalened8, anthracened10, perylened12, chrysened12

as internal standards at 5 mg ml-1 and 2 mg ml-1 respectively).Alkanes and PAHs were identified and quantified using aThermo Trace GC gas chromatograph coupled to a ThermoTrace DSQ® mass spectrometer as described previously(McKew et al., 2007). The extraction efficiency was 89%, andthe variation of the reproducibility of extraction and quantifi-cation of samples was determined by successive extractionsand injections (n = 6) of the same sample, and estimated tobe �8%.

DNA extraction

Cells were pelleted from 2 ml of the seawater microcosms bycentrifugation (16 000 g for 5 min at 4°C), re-suspended in0.5 ml of 1 M sodium phosphate buffer (pH 8.0), and 0.5 ml ofphenol–chloroform–isoamyl alcohol (25:24:1), and then lysedby bead beating in Lyzing Matrix B tubes (Bio 101 System,Q-Biogene®) using a Mikro-dismembrator-U (B. Braun,Biotech International) at 2000 rpm for 30 s. After centrifuga-tion (16 000 g for 1 min at 4°C), 450 ml of the upper aqueouslayer was removed and further purified with chloroform–isoamyl alcohol (24:1), and the DNA precipitated, pelletedand washed as described previously (McKew et al., 2007),prior to re-suspension in 400 ml of sterile distilled water.

Quantitative polymerase chain reaction

Primer Express® Software v. 2.0 (Applied Biosystems, FosterCity, CA, USA) was used to design three primer pairs forQ-PCR analysis. Primers were based on the Alcanivoraxalkane hydroxylase (alkB2) gene (ALCalkB2F867 5′-CGCCGTGTGAATGACAAGGG-3′ and ALCalkB2R999 5′-CGACGCTTGGCGTAAGCATG-3′), Thalassolituus alkanehydroxylase (alkB) gene (THALalkBF125 5′-GACGTCGCCACACCTGCC-3′ and THALalkBR342 5′-GGGCCATACAGAGCAAGCAA-3′) and Cycloclasticus aromaticring-hydroxylating dioxygenase (phnA) gene (CYCphnAF2435′-CGTTGTGCGCATAAAGGTGCGG-3′ and CYCphn-AR388R 5′-CTTGCCCTTTCATACCCCGCC-3′). All primerswere checked for specificity using BLAST searches for short,nearly exact matches. All of the sequences in the GenBankdatabase, except the desired targets, had more than twomismatches with the primers. Standard curves for Q-PCRwere constructed using known amounts of target templategenerated by PCR amplification of the target gene fromgenomic DNA. The resulting amplicon was then purified using

Effectiveness of crude oil bioremediation strategies 1569

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

the QIAquick PCR purification kit (Qiagen), prior to quantifi-cation using a Nanodrop® ND-1000 spectrophotometer. Thetarget abundance for DNA standards was calculated assum-ing a molecular mass of 660 Da for double-stranded DNAusing the following formula:

Target abundance = 6.023 ¥ 1023 (copies mol-1) ¥ standardconc. (g ml-1)/MW (g mol-1)

DNA standard curves for each gene were then created usingdilution series ranging from 5 ¥ 101 to 5 ¥ 107 amplicons ml-1.For each of the three genes, DNA isolated from the triplicatemicrocosm samples, and from the initial seawater samples,together with no-template controls (NTC) were used inQ-PCR amplifications in triplicate with each replicate onseparate runs. Reactions were performed on an ABI 7000Sequence Detection System (Applied Biosystems) with initialdenaturation for 5 min at 95°C, followed by 40 cycles of 95°Cfor 15 s and 60°C for 1 min. Each 25 ml of reaction contained1 ml of template, 12.5 ml of 2¥ SYBR® Green PCR Master Mix(Applied Biosystems) and 100 nM of each primer. A dissocia-tion protocol was run at the end of each SYBR® Greenreal-time PCR reaction to verify that only the expected ampli-fication product was generated. In all experiments, appropri-ate negative controls were subjected to the same procedureto exclude or detect any possible contamination. Ampliconnumbers were quantified against the standard curve usingthe ABI Prism 7000 sequence detection software (AppliedBiosystems) using automatic analysis settings for the Ctvalues and baseline settings. The limit of detection for allthree genes was set at 3.3 cycles lower than the Ct value ofthe NTC (Smith et al., 2006), which corresponds with a geneabundance of 3 ¥ 104, 5 ¥ 103 and 1 ¥ 104 cells per millilitre ofseawater for Thalassolituus alkB, Alcanivorax alkB2 andCycloclasticus phnA respectively. Detected target geneswere converted to cell density (cells ml-1) assuming a singlecopy per genome, as demonstrated for Alcanivorax(Schneiker et al., 2006). This assumption was further indi-rectly confirmed for all three target genes using the Q-PCRprimers for amplification from known amounts of genomicDNA, where the chromosome copy number was calculatedfrom the known genome sizes of 3.12, 2.9 and 2.2 Mb forAlcanivorax (Schneiker et al., 2006), Cycloclasticus (Buttonet al., 1998) and Thalassolituus (Yakimov et al., 2004)respectively.

Acknowledgements

This work was funded by the UK Natural EnvironmentResearch Council (NER/S/A/2003/11237) and the EuropeanCommunity Project COMMODE (EVK3-CT2002-00077) witha contribution from the European Community Project FACEIT[STREP-01839(GOCE)]. We would like to thank all partnersof the COMMODE project for their useful discussions.

References

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.

Button, D.K., Robertson, B.R., Lepp, P.W., and Schmidt, T.M.(1998) A small, dilute-cytoplasm, high-affinity, novel bacte-rium isolated by extinction culture and having kinetic con-stants compatible with growth at ambient concentrations ofdissolved nutrients in seawater. Appl Environ Microbiol 64:4467–4476.

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.

Dyksterhouse, S.E., Gray, J.P., Herwig, R.P., Lara, J.C., andStaley, J.T. (1995) Cycloclasticus pugetii gen. nov, sp. nov,an aromatic hydrocarbon-degrading bacterium frommarine-sediments. Int J Syst Bacteriol 45: 116–123.

Geiselbrecht, A.D., Hedlund, B.P., Tichi, M.A., and Staley,J.T. (1998) Isolation of marine polycyclic aromatic hydro-carbon (PAH)-degrading Cycloclasticus strains from theGulf of Mexico and comparison of their PAH degradationability with that of Puget Sound Cycloclasticus strains. ApplEnviron Microbiol 64: 4703–4710.

Hara, A., Syutsubo, K., and Harayama, S. (2003) Alcanivoraxwhich prevails in oil-contaminated seawater exhibits broadsubstrate specificity for alkane degradation. Environ Micro-biol 5: 746–753.

Harvey, S., Elashvili, I., Valdes, J.J., Kamely, D., and Chakra-barty, A.M. (1990) Enhanced removal of Exxon Valdezspilled oil from Alaskan gravel by a microbial surfactant.Bio-Technology 8: 228–230.

Head, I.M., Jones, D.M., and Röling, W.F.M. (2006) Marinemicroorganisms make a meal of oil. Nat Rev Microbiol 4:173–182.

ITOPF (2006) Oil Tanker Spill Statistics: 2005. London, UK:The International Tanker Owners Pollution Federation(ITOPF).

Iwabuchi, N., Sunairi, M., Urai, M., Itoh, C., Anzai, H., Naka-jima, M., and Harayama, S. (2002) Extracellular polysac-charides of Rhodococcus rhodochrous S-2 stimulate thedegradation of aromatic components in crude oil by indig-enous marine bacteria. Appl Environ Microbiol 68: 2337–2343.

Kasai, Y., Kishira, H., Sasaki, T., Syutsubo, K., Watanabe, K.,and Harayama, S. (2002) Predominant growth of Alcanivo-rax strains in oil-contaminated and nutrient-supplementedsea water. Environ Microbiol 4: 141–147.

Leahy, J.G., and Colwell, R.R. (1990) Microbial degradationof hydrocarbons in the environment. Microbiol Rev 54:305–315.

Lindstrom, J.E., Prince, R.C., Clark, J.C., Grossman, M.J.,Yeager, T.R., Braddock, J.F., and Brown, E.J. (1991)Microbial populations and hydrocarbon biodegradationpotentials in fertilized shoreline sediments affected by theT/V Exxon Valdez oil spill. Appl Environ Microbiol 57:2514–1522.

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

MacNaughton, S.J., Stephen, J.R., Venosa, A.D., Davis,G.A., Chang, Y.J., and White, D.C. (1999) Microbial popu-lation changes during bioremediation of an experimental oilspill. Appl Environ Microbiol 65: 3566–3574.

1570 B. A. McKew et al.

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571

Rambeloarisoa, E., Rontani, J.F., Giusti. G., Duvnjak, Z., andBertrand, J.C. (1984) Degradation of crude oil by a mixedpopulation of bacteria isolated from sea-surface foams.Mar Biol 83: 69–81.

Röling, W.F.M., Milner, M.G., Jones, D.M., Lee, K., Daniel,F., Swannell, R.J.P., and Head, I.M. (2002) Robust hydro-carbon degradation and dynamics of bacterial communitiesduring nutrient-enhanced oil spill bioremediation. ApplEnviron Microbiol 68: 5537–5548.

Ron, E.Z., and Rosenberg, E. (2002) Biosurfactantsand oil bioremediation. Curr Opin Biotechnol 13: 249–252.

Schneiker, S., Martins dos Santos, V.A.P., Bartels, D., Bekel,T., Brecht, M., Buhrmester, J., et al. (2006) Insights intomarine oil-degradation: the genome sequence of Alcani-vorax borkumensis, a cosmopolitan and efficienthydrocarbon-degrading bacterium. Nat Biotechnol 24:997–1004.

Smith, C.J., Nedwell, D.B., Dong, L.F., and Osborn, A.M.(2006) Evaluation of quantitative polymerase chainreaction based approaches for determining gene copyand gene transcript numbers in environmental samples.Environ Microbiol 8: 804–815.

Swannell, R.P.J., Croft, B.C., Grant, A.L., and Lee, K. (1995)Evaluation of bioremediation agents in beach microcosms.Spill Sci Technol B 2: 151–159.

Swannell, R.P.J., Mitchell, D., Lethbridge, G., Jones, D.,Heath, D., Hagley, M., et al. (1999) A field demonstration ofthe efficacy of bioremediation to treat oiled shorelines fol-

lowing the Sea Empress incident. Environ Technol 20:863–873.

Tagger, S., Bianchi, A., Julliard, M., Le Petit, J., and Roux, B.(1983) Effect of microbial seeding of crude oil in seawaterin a model system. Mar Biol 78: 13–20.

Venosa, A.D., Suidan, M.T., Wrenn, B.A., Strohmeier, K.L.,Haines, J.R., Eberhart, B.L., et al. (1996) Bioremediation ofan experimental oil spill on the shoreline of Delaware Bay.Environ Sci Technol 30: 1764–1775.

Wright, A.L., Weaver, R.W., and Webb, J.W. (1997) Oil biore-mediation in salt marsh mesocosms as influenced by Nand P fertilization, flooding, and season. Water Air Soil Poll95: 179–191.

Yakimov, M.M., Golyshin, P.N., Lang, S., Moore, E.R.B.,Abraham, W.R., Lunsdorf, H., and Timmis, K.N. (1998)Alcanivorax borkumensis gen. nov., sp. nov., a new,hydrocarbon-degrading and surfactant-producing marinebacterium. Int J Syst Bacteriol 48: 339–348.

Yakimov, M.M., Giuliano, L., Denaro, R., Crisafi, E., Cherni-kova, T.N., Abraham, W.R., et al. (2004) Thalassolituusoleivorans gen. nov., sp. nov., a novel marine bacteriumthat obligately utilizes hydrocarbons. Int J Syst Evol Micro-biol 54: 141–148.

Yakimov, M.M., Denaro, R., Genovese, M., Cappello, S.,D’Auria, G., Chernikova, T.N., et al. (2005) Natural micro-bial diversity in superficial sediments of Milazzo Harbor(Sicily) and community successions during microcosmenrichment with various hydrocarbons. Environ Microbiol7: 1426–1441.

Effectiveness of crude oil bioremediation strategies 1571

© 2007 The AuthorsJournal compilation © 2007 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 9, 1562–1571