Vol. 58, No. 10APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Oct. 1992, p. 3276-32820099-2240/92/103276-07$02.00/0Copyright X 1992, American Society for Microbiology

Enhanced Octadecane Dispersion and Biodegradation by aPseudomonas Rhamnolipid Surfactant (Biosurfactant)

YIMIN ZHANG AND RAINA M. MILLER*

Department of Soil and Water Science, University ofArizona, Tucson, Anizona 85721

Received 20 March 1992/Accepted 20 July 1992

A microbial surfactant (biosurfactant) was investigated for its potential to enhance bioavailability and,hence, the biodegradation of octadecane. The rhamnolipid biosurfactant used in this study was extracted fromculture supernatants after growth of Pseudomonas aeruginosa ATCC 9027 in phosphate-limited proteosepeptone-glucose-ammonium salts medium. Dispersion of octadecane in aqueous solutions was dramaticallyenhanced by 300 mg of the rhamnolipid biosurfactant per liter, increasing by a factor of more than 4 ordersof magnitude, from 0.009 to >250 mg/liter. The relative enhancement of octadecane dispersion was muchgreater at low rhamnolipid concentrations than at high concentrations. Rhamnolipid-enhanced octadecanedispersion was found to be dependent on pH and shaking speed. Biodegradation experiments done with an

initial octadecane concentration of 1,500 mg/liter showed that 20%o of the octadecane was mineralized in 84 hin the presence of 300 mg of rhamnolipid per liter, compared with only 5% octadecane mineralization when no

surfactant was present. These results indicate that rhamnolipids may have potential for facilitating thebioremediation of sites contaminated with hydrocarbons having limited water solubility.

The bioavailability of an organic compound is a function ofits solubility, mobility, and sorption or desorption in theenvironment. Bioavailability may be the critical limitingfactor controlling biodegradation rates for many organiccompounds having low water solubility. These compoundsresist biodegradation at least in part because of limitedtransport into cells, as demonstrated by Miller and Bartha(16) in a study of alkane biodegradation. This work showedthat transport limitations could be overcome by encapsula-tion of alkanes (C18 and C36 compounds) into small unila-mellar vesicles (liposomes). Encapsulation increased theaqueous dispersion of the alkanes by intercalation of thealkanes into the vesicle phospholipid bilayers, which wereapproximately 20 to 50 nm in diameter. Furthermore, encap-sulation facilitated the delivery of alkanes to bacterial mem-brane-bound enzymes, most likely via fusion of the vesicleswith cell membranes. While the liposome-encapsulated al-kanes were efficiently degraded by a hydrocarbon-utilizingPseudomonas sp., the high cost of encapsulation precludesits use in bioremediation. However, this work suggested thatfurther investigation into methods to increase the aqueousdispersion of recalcitrant compounds is necessary.

Relatively low concentrations of surfactants can alsoenhance the aqueous dispersion of slightly water-solublecompounds (11). Surfactants can be synthetic (detergents)but are also naturally produced by plants, animals, and manydifferent microorganisms (24). A property that all surfactantshave in common is that they reduce the surface tension of aliquid medium. For example, distilled water has a surfacetension of 73 dyn/cm. An effective microbially producedsurfactant (biosurfactant) can lower this value to <30dyn/cm (12). The amount of surfactant needed to achieve thelowest possible surface tension is defined as the criticalmicelle concentration (CMC). Alternatively, the CMC canbe defined as the surfactant concentration at which theconcentration of a free monomer ceases to increase and anyfurther monomer added will form micellar structures (1). In

* Corresponding author.

the process of micelle formation, the surfactant moleculesthat aggregate to form micelles have the ability to surroundslightly soluble molecules, a process that effectively dis-perses or emulsifies them into the aqueous phase (21). TheCMCs of biosurfactants typically range from 1 to 200 mg/liter(12).We postulated that biosurfactants would enhance the

aqueous dispersion of organic compounds having limitedwater solubility in two ways: (i) biosurfactants would de-crease the surface and interfacial tensions in the culturemedium, thereby increasing the aqueous dispersion of or-ganic compounds at the molecular level, and (ii) physicalbiosurfactant interactions with slightly water-soluble organiccompounds would increase their aqueous dispersion, theincrease below the CMC being due to hydrophobic interac-tions between biosurfactant monomers and slightly solubleorganic compounds (22) and the increase above the CMCbeing due to biosurfactant encapsulation of slightly solubleorganic compounds into micellar or bilayer aggregates. Inthis report, we quantitate the ability of a rhamnolipid toincrease the aqueous dispersion of a slightly soluble modelorganic compound and to enhance the rate of biodegradationof that model organic compound. The use of biosurfactantsis attractive because they are natural products, they arebiodegradable, and they have potential for use in in situremediation.The biosurfactant used in this study was a rhamnolipid

produced by Pseudomonas aeruginosa ATCC 9027. Wechose this biosurfactant because (i) it is a glycolipid, which isthe most commonly isolated type of biosurfactant (4), and (ii)members of the genus Pseudomonas are common soil mi-croorganisms. A previous characterization of rhamnolipidsproduced by various Pseudomonas spp. (5-7, 9, 10, 19)indicated that a mixture of rhamnolipids containing eitherone or two rhamnose residues and two lipid chains isproduced (Fig. 1) (13, 18). A slightly water-soluble alkane,n-octadecane (C18), was chosen as a model compound forseveral reasons. The water solubility of octadecane is verylow, 0.006 mg/liter (21), and a study of octadecane wouldallow a comparison with results obtained from previous

3276

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from

BIOSURFACTANT-ENHANCED OCTADECANE BIODEGRADATION 3277

HO 0

{CH3\\3COCH-CH2COO-CH-CH2COOH

OH OR Ri K2FIG. 1. Structure of a rhamnolipid. In a monorhamnolipid, R =

H or R = CO-CH=CH-(CH2)6-CH3; in a dirhamnolipid, R =

rhamnosyl or R = rhamnosyl-O-CO-CH=CH(CH2)6-CH3. R, andR2 = C7HI5 (13,14,18).

octadecane liposome encapsulation studies (16). In thisstudy, the effects of rhamnolipid concentration, pH, andshaking speed on the aqueous dispersion of octadecane weredetermined. Octadecane mineralization rates with P. aerug-

inosa ATCC 9027 in the presence of various concentrationsof rhamnolipid were determined by measuring 14CO2 evolu-tion from the radioactive substrate. Experiments were per-

formed at 23°C, below the melting point of octadecane, tomaintain the compound in a solid physical state.

MATERIALS AND METHODSBiosurfactant production. P. aeruginosa ATCC 9027 was

obtained from the American Type Culture Collection (Rock-ville, Md.). The culture was maintained on Pseudomonasagar P (Difco, Detroit, Mich.) slants and transferredmonthly. The organism was grown as described by Mulliganet al. (17) to induce biosurfactant production. In brief, P.aeruginosa was inoculated into 25 ml of Kay's minimalmedium, composed of NH4H2PO4 (0.3%), K2HPO4 (0.2%),glucose (0.2%), FeSO4 (0.5 mg of Fe per liter), andMgSO4. SO4 (0.1%), in a 125-ml flask (23). The preculturewas incubated with gyratory shaking at 250 rpm for 24 h at37°C, and then 2 ml was used to inoculate 200 ml ofphosphate-limited proteose peptone-glucose-ammoniumsalts (PPGAS) medium in a 1,000-ml flask. PPGAS mediumis composed of NH4C1 (0.02 M), KCl (0.02 M), Tris-HCl(0.12 M), glucose (0.5%), proteose peptone (1%), andMgSO4 (0.0016 M), adjusted to pH 7.2 (3). This flask wasincubated at 37°C with gyratory shaking at 250 rpm, withperiodic aseptic removal of samples to monitor surfacetension. The initial surface tension of PPGAS medium was

66 dyn/cm. The surface tension of the medium started todecrease at 6 h, coinciding with the production of a bluepigment. By 48 h, the surface tension of the medium reached29 dyn/cm and did not decline further. Rhamnolipid was

harvested at 60 h.Rhamnolipid extraction and measurement. Rhamnolipid

was recovered from the culture supernatant after the re-

moval of cells by centrifugation at 6,800 x g for 20 min.Rhamnolipid was then precipitated by acidification of thesupematant to pH 2.0 and centrifugation at 12,100 x g for 20min. The precipitate was dissolved in 0.05 M bicarbonate(8.6), reacidified, and recentrifuged at 12,100 x g for 20 min.Following centrifugation, the precipitate was extracted withchloroform-ethanol (2:1) three times. The organic solventwas evaporated on a rotary evaporator, and the yellowishoily residue was dissolved in 0.05 M bicarbonate (pH 8.6).The biosurfactant concentration was then estimated inde-pendently by surface tension measurement and by L-rham-nose measurement.

Surface tension was measured with a SensaDyne (Milwau-kee, Wis.) 6000 surface tensiometer, which uses the maxi-mum bubble pressure technique. This technique avoidserrors that can be caused by foam or debris on the solution

surface by measuring surface tension within the body of thetest fluid (20). The sensitivity of the surface tension mea-surements was <1 mg/liter for rhamnolipid concentrationsbetween 0 and 50 mg/liter, and the measurements were foundto be reproducible to +0.1 dyn/cm. Rhamnolipid was alsoquantified by measurement of L-rhamnose by the 6-deoxy-hexose method (2). In brief, biosurfactant solutions weretreated with 70% sulfuric acid and boiled for 10 min. Afterthe samples had cooled, thioglycolic acid was added to afinal concentration of 0.059%. Samples were incubated in thedark for 3 h and then measured spectrophotometrically at420 nm. Standard curves were prepared with L-rhamnoseobtained from Sigma (St. Louis, Mo.).

Effect of pH on surface tension. Rhamnolipid was diluted to50 mg/liter in 0.1 M pH 7.0 phosphate buffer. The solutionpH was adjusted by the addition of HCI or NaOH, and thesurface tension was measured at 25°C.Aqueous dispersion tests. Octadecane (99% pure) was

purchased from Aldrich Chemical Co. (Milwaukee, Wis.).[14C]Octadecane, labeled at the C-1 position (specific activ-ity, 3.6 mCi/mmol; 98% pure), was obtained from Sigma.A1propriate amounts of a mixture of octadecane and['C]octadecane (0.78 g/liter, with a specific activity of 15,uCi/mmol for rhamnolipid dispersion tests; 1 mg/liter, with aspecific activity of 1.1 mCi/mmol for determination of thesolubility of octadecane in water) in chloroform were addedto test tubes (16 by 100 mm). The chloroform was evapo-rated to allow the octadecane to coat the bottom of the testtubes. After evaporation of the solvent, 1 ml of rhamnolipidsolution was added. The test tubes were incubated at 37°C ina water bath for 30 s, to melt the coated octadecane and thencooled at room temperature until the octadecane solidifiedon the surface of the solution. The test tubes were thenincubated at 23°C at an appropriate shaking speed. After 24h, 0.1 ml was carefully removed from the bottom of the testtubes to avoid floating of any particulate octadecane on thetop of the liquid surface and was added to 5 ml of ScintiverseBD (Fisher, Pittsburgh, Pa.). Radioactivity was determinedwith a Packard (Meriden, Conn.) Tri-Carb liquid scintillationcounter (model 1600 TR).To confirm that this method did not exaggerate dispersion

results and also to characterize the size of the dispersedoctadecane particles, we filtered some octadecane-rhamno-lipid solutions through 30,000- or 300,000-molecular-weight(MW)-cutoff membranes (Millipore Corp., Bedford, Mass.)and determined the radioactivity in the filtrates.Rhamnolipid solutions were prepared in 0.1 M pH 7.0

phosphate buffer. Tests were done in triplicate, with theexception of the test for octadecane solubility in water, forwhich five replicates were done. Each experiment wasrepeated twice.

Biodegradation tests. The biodegradation of octadecaneand rhamnolipid was determined by measurement of theincrease in the protein concentration as an indication of cellgrowth and by rhamnolipid determination (based on rham-nose). Octadecane mineralization was measured by thequantification of 14CO2 and 14C-volatile compounds. Forprotein and rhamnose determinations, octadecane was dis-solved in chloroform, and the mixture was used to coat thebottom of 250-ml flasks. The chloroform was evaporated,and 50 ml of mineral salts medium (16) containing rhamno-lipid was added to each flask. The flasks were warmed in a370C water bath for 30 s to melt the coated octadecane andthen cooled at room temperature until the octadecane solid-ified on the surface of the medium. The flasks were theninoculated with a 5% inoculum of P. aeruginosa ATCC 9027

VOL. 58, 1992

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from

3278 ZHANG AND MILLER

80 -

2 7O0

1-6O0

*m 50

E*8 40-I-

z 30-

20-(

80

70

60

50

01 34

0 1 2 3 4 5Rhamnolipid Conc. (mg/l)

9

Is CMC"'sI

------- --

.................... ...... ... . .. . .

25 50 75 100 125 150 175 200

Rhamnolipid Concentration (mg/i)FIG. 2. Effect of increasing rhamnolipid concentrations (Conc.) on surface tension (Surf. Tens.). Rhamnolipid solutions were prepared in

0.1 M pH 7.0 phosphate buffer. Surface tension was measured at 25°C. Symbols: 0, experimental data; -- -, best-fit exponential function;, best-fit linear functions at c5 and >50 mg of rhamnolipid per liter. (Inset) Expanded view of the graph for 0 to 5 mg of rhamnolipid per

liter. Symbols: *, experimental data; , best-fit linear function.

grown in Kay's minimal medium at 37°C for 24 h. The flaskswere incubated with gyratory shaking (200 rpm) at 23°C.Periodically, samples were taken from each flask for rham-nolipid and protein determinations. For rhamnolipid analy-ses, 1-ml samples were filtered through a 0.22-p,m-pore-sizefilter, and L-rhamnose was quantified. For protein analyses,1-ml samples were heated for 10 min with 0.1 ml of 1 NNaOH (to lyse cells), and the protein content was deter-mined by the method of Lowry et al. (14).For mineralization experiments, a mixture of octadecane

and [14C]octadecane (1.55 g/liter; specific activity, 0.37 ,uCi/mmol) was used to coat the bottom of modified 125-mlmicro-Fembach flasks (Wheaton, Milville, N.J.), designedfor the collection of 14C02 and 14C-volatile compounds (15).The solvent was evaporated, and 20 ml of mineral saltsmedium (16) containing rhamnolipid was added to each flask.The octadecane was melted and the flasks were inoculatedand incubated as described above. The micro-Fernbachflasks were periodicallyv flushed through a series of six trapsto collect '4C02 and 1 C-volatile organic compounds (15).

RESULTS

Effect of rhamnolipid on surface tension. Figure 2 showsthe dependence of surface tension on the biosurfactantconcentration. Surface tension decreased rapidly from 72 to30 dyn/cm with small increases in the rhamnolipid concen-tration up to 50 mg of rhamnolipid per liter. Further in-creases in the rhamnolipid concentration only slowly re-duced the surface tension from 30 to 29 dyn/cm. Once thesurface tension reached 29 dyn/cm, the further addition ofrhamnolipid had no effect. The CMC was determined from asemilog plot of surface tension versus rhamnolipid concen-tration to be 40 mg/liter (data not shown). The data in Fig. 2were fitted (r2 = 0.968) to an exponential function (seebroken curve) which is described by the equation y = a +

be-', where a = 29.5 dyn/cm, b = 0.117 dyn/cm, and c =42.5 mg liter-1. However, two separate linear functionsdescribed a better fit (r2 = 0.999) for the data at rhamnolipidconcentrations of <5 mg/liter and >50 mg/liter (see solidlines and inset). These functions are described by the equa-tion y = a + bx, where a = 72.1 dyn/cm and b = -4.26dyn/cm/mg/liter for 0 to 5 mg of rhamnolipid per liter and a= 29.6 dyn/cm and b = -0.00385 dyn/cm/mg/liter for 50 to200 mg of rhamnolipid per liter. The close fit obtained for thedata with these two different linear functions implies thatthere are different behaviors for the rhamnolipid in solution.Low rhamnolipid concentrations (<5 mg/liter) have a strongeffect on surface tension, while high concentrations (>50mg/liter) have a negligible effect.

Dispersion of octadecane in the biosurfactant solution. Asshown in Figure 3, rhamnolipid dramatically increased theaqueous dispersion of octadecane. For instance, in thepresence of 50 mg of rhamnolipid per liter, the aqueousdispersion of octadecane was 125 mg/liter, an increase ofmore than 4 orders of magnitude over the solubility ofoctadecane in pure water (0.009 + 0.005 mg/liter). The datain Fig. 3 suggest that there are two mechanisms of disper-sion, one at low rhamnolipid concentrations (region 1) andanother at higher concentrations (region 2). At low rhamno-lipid concentrations (1 to 50 mg/liter), there was a sharpincrease in octadecane dispersion. After this sharp increase,there was a linear dependence of octadecane dispersion onbiosurfactant concentration from 50 to 500 mg/liter. Rham-nolipid also caused gross visible changes in the physical stateof the excess insoluble octadecane. Instead of several largeflakes of octadecane floating on the liquid surface, in thepresence of rhamnolipid, the insoluble octadecane was dis-persed as fine particles on the surface of the solution.

Estimation of sizes of octadecane-rhamnolipid aggregates.For estimation of the sizes of the octadecane-rhamnolipidaggregates, a solution containing 200 mg of rhamnolipid per

APPL. ENvIRON. MICROBIOL.

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from

BIOSURFACTANT-ENHANCED OCTADECANE BIODEGRADATION 3279

350

0zz-

cw._

cco

300

250

200

150

100

50

00 50 100 150 200 250 300 350 400 450 500

Rhamnolipid Concentration (mg/i)FIG. 3. Effect of rhamnolipid concentration on tlte aqueous

dispersion of octadecane. Solutions containing rhamnolipid and["4C]octadecane (0.78 g/liter) were incubated with gyratory shakingat 100 rpm and 23°C for 24 h. The aqueous dispersion of octadecanewas measured as described in Materials and Methods. Symbols: *,experimental data; -, best-fit linear functions at c5 and .50 mgof rhamnolipid per liter.

liter, 0.78 g of octadecane per liter, and 15 ,uCi of [14C]oc-tadecane per mmol was shaken for 10 h at 200 rpm. Aliquotsof this solution were filtered through 300,000- or 30,000-MWfilters. The concentrations of octadecane in the filtrate were231 + 0 mg/liter in the 300,000-MW filter and 126 + 6 mg/literin the 30,000-MW filter. These results indicated that all of theoctadecane that was measured as dispersed (see Fig. 3)passed through the 300,000-MW filter. The size distributionof dispersed octadecane was approximately 55% in aggre-gates of less than 30,000 MW and 45% in aggregates ofbetween 30,000 and 300,000 MW.

Effect of shaking speed and time on octadecane dispersion inan aqueous solution. The aqueous dispersion of octadecanewas dependent on both the time that octadecane-rhamno-lipid solutions were incubated together and the shakingspeed during incubation. The dispersion of octadecane in-creased with time of incubation until approximately 10 to 12h and then remained constant (data not shown). The disper-sion of octadecane increased to 75 mg/liter in the presence ofrhamnolipid (125 mg/liter) with no shaking in 10 h. Thedispersion of octadecane under the same conditions but withshaking (100 rpm) increased to 150 mg/liter. Further in-creases in shaking speed increased octadecane dispersiononly slightly (data not shown).

Effect of pH on the surface tension and aqueous dispersionof octadecane. The surface tension of rhamnolipid solutionswas quite sensitive to pH (Fig. 4). The surface activity of therhamnolipid was highest between pH 7.0 and 7.5. As the pHwas increased above 7.5, there was a slight decrease insurface activity that resulted in an increase in surface tensionfrom 30 to 32 dyn/cm. After increasing to 32 dyn/cm at pH 8,the surface tension of the solutions remained relativelystable, even at pH 11. On the other hand, as the pH wasdecreased from 7.0 to 5.0, surface activity decreased signif-icantly, resulting in an increase in surface tension from 30 to

42

0

C;

-P

IU,

Cs

40

38

36

34

32

30

5 6 7 8 9 10 11

pHFIG. 4. Effect of pH on the surface tension of rhamnolipid

solutions. The pH of rhamnolipid solutions (50 mg/liter) in 0.1 Mphosphate buffer was varied, and the surface tension was measuredat 25'C.

>40 dyn/cm. Visually, the rhamnolipid solutions were clearabove pH 7.0 but became turbid below this pH. Precipitationof rhamnolipid was not apparent until the pH was decreasedbelow 5.0.As might be expected given the effect of pH on the surface

activity of rhamnolipids, octadecane dispersion in the pres-ence of rhamnolipids was also influenced by pH (Fig. 5).Maximum enhancement of aqueous octadecane dispersionoccurred at approximately pH 7.0, with dispersion decreas-ing with an increase in pH. Octadecane dispersion alsodecreased from pH 7.0 to 6.0 but unexpectedly increasedagain from pH 6.0 to 5.5. Below pH 5.5, experimental errorin the measurement of octadecane dispersion became veryhigh because of the precipitation of rhamnolipid.Rhamnolipid-enhanced mineralization of octadecane. As

shown in Fig. 6, P. aeruginosa ATCC 9027 did not degraderhamnolipid either alone or in the presence of octadecane, asmeasured on the basis of protein (Fig. 6A) or rhamnolipid(Fig. 6B) concentrations. The culture medium used in these

200-

~i175

as 150

125

75

50

250

5 6 7 8 9 10pH

FIG. 5. Effect of pH on octadecane dispersion in rhamnolipidsolutions. Solutions containing rhamnolipid (200 mg/liter) and["4C]octadecane (0.78 g/liter) were adjusted to the appropriate pHand incubated with gyratory shaking at 100 rpm and 23°C for 24 h.The aqueous dispersion of octadecane was measured as described inMaterials and Methods.

VOL. 58, 1992

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from

3280 ZHANG AND MILLER

200

- 150

o 100,l;

.050lou05

0

125

Cu

-

0

o

I*_

0

eq

as

csle

0 25 50 75 100 125 150 175Incubation Time (hours)

75 _

50 _

25

0

0 25 50 75 100 125 150 175

Incubation Time (hours)FIG. 6. Biodegradation of octadecane and rhamnolipid by P.

aeruginosa ATCC 9027. Rhamnolipid (100 mg/liter) or a mixture ofoctadecane (1.55 g/liter) and rhamnolipid (100 mg/liter) was incu-bated in mineral salts medium with P. aeruginosa with gyratoryshaking (200 rpm) at 23°C. Biomass protein (A) and rhamnolipid (B)concentrations were measured as described in Materials and Meth-ods. Symbols: A, rhamnolipid alone; 0, rhamnolipid and octade-cane.

experiments was a mineral salts medium amended with 1.55g of octadecane per liter. Figure 7 shows the mineralizationof octadecane by P. aeruginosa ATCC 9027 in the presenceof various concentrations of rhamnolipid. P. aeruginosa

mineralized 5.1% of the substrate in 84 h with no rhamno-lipid present. Rhamnolipid increased octadecane mineraliza-tion to 10.2% (100 mg/liter), 12.8% (125 mg/liter), and 19.9%(300 mg/liter). Surface tension measurements of the culturesupernatants showed that P. aeruginosa ATCC 9027 did notproduce rhamnolipid during growth on octadecane in themineral salts medium.

25

20 -

15 -

10 -

5-

0 10 20 30 40 50 60 70 80 90

Incubation Time (hours)FIG. 7. Effect of rhamnolipid concentration on the mineraliza-

tion of octadecane by P. aeruginosa ATCC 9027. A mixture of["4C]octadecane (1.55 g/liter) and rhamnolipid was incubated inmineral salts medium with P. aeruginosa ATCC 9027 with gyratoryshaking (200 rpm) at 23°C. Mineralization was measured as de-scribed in Materials and Methods. Symbols: 0, no rhamnolipid; V,rhamnolipid at 100 mg/liter; V, rhamnolipid at 125 mg/liter; El,rhamnolipid at 300 mg/liter.

DISCUSSION

Rhamnolipid-enhanced dispersion of octadecane. The re-sults showed that a rhamnolipid biosurfactant enhanced theaqueous dispersion of octadecane by more than 4 orders ofmagnitude, from 0.009 to 320 mg/liter. As shown in Fig. 3,the dramatic increase in octadecane dispersion seemed to bedue to two different mechanisms, one at low (region 1) andone at high (region 2) rhamnolipid concentrations. In region1, the octadecane/rhamnose ratio was 3.7:1. In this region,octadecane solubility may be due to the rapid decrease insurface tension and/or may be due to hydrophobic interac-tions between octadecane and the hydrophobic tails ofrhamnolipid monomers (22). Taken together, the results inFig. 2 and 3 suggest that a lowering of surface tension may bethe dominant mechanism by which octadecane dispersion isenhanced at low rhamnolipid concentrations. Examinationof these figures shows that the effect of low rhamnolipidconcentrations on octadecane dispersion (Fig. 3) mirrors theeffect of low rhamnolipid concentrations on surface tension(Fig. 2). In both cases, low rhamnolipid concentrations havea strong effect on surface tension and octadecane dispersionand increasing rhamnolipid concentrations have a diminish-ing effect on both surface tension (Fig. 2) and the relativeenhancement of apparent octadecane dispersion (Fig. 3).At rhamnolipid concentrations higher than the CMC (re-

gion 2, Fig. 3), the octadecane/rhamnose ratio decreased to0.3:1, resulting in a slower rate of octadecane dispersionwith increasing rhamnolipid concentrations. At these higherrhamnolipid concentrations, increasing octadecane disper-sion is most likely due to physical association with rhamno-lipid aggregates. The physical interaction between octade-cane and rhamnolipids seems to be analogous to thepreviously described intercalation of octadecane into phos-pholipid bilayers during liposome formation (16). For rham-nolipids, the case is more complex, because a number of

: B.

APPL. ENvIRON. MICROBIOL.

r

100la 0 a

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from

BIOSURFACTANT-ENHANCED OCTADECANE BIODEGRADATION 3281

different rhamnolipid structures that are dependent on pHcan form (8). The type of rhamnolipid structure formed isespecially sensitive to pH in the range of 6.0 to 7.0, mostlikely because of the rhamnosyl moiety, which has a pKa of5.6 (8). At a low pH (<6.0), the rhamnosyl moiety is at least50% uncharged and rhamnolipids form liposome-like vesi-cles. Between pH 6.0 and 6.6, rhamnolipids form eitherlamella-like structures or lipid aggregates, and above pH 6.8,when the rhamnosyl moiety is negatively charged, micellesform.The effect of pH on octadecane dispersion seems to be

correlated with the type of rhamnolipid aggregate that ispresent in solution. The dispersion of octadecane was high-est at pH 7.0 (Fig. 5), corresponding to rhamnolipids being ina micellar formation. We speculate that in this case, theoctadecane was intercalated into the rhamnolipid micelles,which may then have fused with the outer membrane of thecell in a manner similar to that described for the liposomedelivery of octadecane (16). Octadecane dispersion de-creased dramatically from pH 7.0 to 6.0 (Fig. 5), correspond-ing to particulate and lamellar rhamnolipid structures. FrompH 6.0 to 5.5, octadecane dispersion increased again, corre-sponding to rhamnolipids forming liposome-like vesiclestructures.

In comparison with liposome encapsulation of octadecaneby phospholipids, which increased the aqueous dispersion ofoctadecane by more than 6 orders of magnitude, to 7.5g/liter, rhamnolipid enhancement of octadecane dispersionto 0.32 g/liter is a modest 4 orders of magnitude (16). Thereare several possible reasons for the large difference in thesolubilization of octadecane by rhamnolipids and phospho-lipids. First, the octadecane-rhamnolipid association wasdriven by a much smaller energy input (shaking) than theoctadecane-phospholipid association (sonication). In fact,sonication of rhamnolipid-octadecane solutions does in-crease octadecane dispersion to levels comparable to thoseachieved with liposome encapsulation (data not shown).Second, the octadecane/rhamnolipid ratio in biosurfactantexperiments (3:1) was higher than the octadecane/phospho-lipid ratio in liposome experiments (1:1).

Effect of rhamnolipid-enhanced octadecane dispersion onmineralization. The results in Fig. 7 indicate that mineraliza-tion rates can be increased significantly by rhamnolipid-enhanced octadecane dispersion. The enhancement of min-eralization rates was linearly dependent on the rhamnolipidconcentration at a statistically significant level (P = 0.05) forthe tested range of 0 to 300 mg of rhamnolipid per liter.However, the fourfold increase in mineralization was notnearly as high as the >104-fold increase in the aqueousdispersion of octadecane. Therefore, it is clear that althoughrhamnolipids increase the dispersion of octadecane, theoctadecane is still not freely bioavailable.A large proportion (45%) of the dispersed octadecane was

shown to be associated with aggregates that passed througha 300,000-MW filter but not through a 30,000-MW filter. Theremainder of the dispersed octadecane was able to passthrough a 30,000-MW filter. Further work will be necessaryto determine the optimal aggregate size for biodegradation.

It should be emphasized that these experiments representa realistic environmental scenario in that the octadecane wasnot mixed with the rhamnolipid prior to inoculation with P.aeruginosa ATCC 9027. Therefore, initial octadecane dis-persion was very low and maximum dispersion was reachedonly after approximately 10 h. Figure 7 shows that very littlebiodegradation occurred at any rhamnolipid concentrationuntil approximately 10 h.

The results of this study have exciting implications for thebioremediation of sites containing recalcitrant compoundshaving limited water solubility. The rhamnolipid used in thisstudy served as an effective dispersion agent, increasing themineralization of octadecane fourfold in a pure-culture batchexperiment. Although shaking was used in batch experi-ments, rhamnolipids can significantly increase the aqueousdispersion of octadecane to 75 mg/liter even without energyinput (shaking). Thus, rhamnolipids may have the potentialto facilitate the biodegradation of hydrocarbons in contam-inated soil as well as aqueous environments. Future studiesare needed to examine (i) rhamnolipid-hydrocarbon interac-tions with bacterial cells, (ii) the effects of rhamnolipids onthe rates of biodegradation of slightly water-soluble organiccompounds in soil and aqueous environments, and (iii) thefate of rhamnolipids in the environment.

ACKNOWLEDGMENTS

This research was supported by the Subsurface Science Program,Office of Health and Environmental Research, U.S. Department ofEnergy. Support was also provided by the University of ArizonaAgricultural Experiment Station (Hatch Project ARZT 136452 H21-032).

REFERENCES1. Cantor, C. R., and P. R. Schimmel. 1980. Biophysical chemis-

try, part III: the behavior of biological macromolecules, p.1327-1371. W. H. Freeman & Co., New York.

2. Chandrasekaran, E. V., and J. N. Bemiller. 1980. Constituentanalysis of glycosaminoglycans. Methods Carbohydr. Chem.8:89-96.

3. Cheng, K. J., J. D. Ingram, and J. W. Costerton. 1970. Releaseof alkaline phosphatase from cells of Pseudomonas aeruginosaby manipulation of cation concentration and of pH. J. Bacteriol.104:748-753.

4. Cooper, D. G., and J. E. Zajic. 1980. Surface-active compoundsfrom microorganisms. Adv. Appl. Microbiol. 26:229-253.

5. Edwards, J. R., and J. A. Hayashi. 1965. Structure of arhamnolipid from Pseudomonas aeruginosa. Arch. Biochem.Biophys. 111:415-421.

6. Guerra-Santos, L., 0. Kippeli, and A. Fiechter. 1984. Pseudo-monas aeruginosa biosurfactant production in continuous cul-ture with glucose as carbon source. Appl. Environ. Microbiol.48:301-305.

7. Hitsatsuka, K., T. Nakahara, and K. Yamada. 1971. Formationof rhamnolipid by Pseudomonas aeruginosa and its function inhydrocarbon fermentation. Agric. Biol. Chem. 35:686-692.

8. Ishigami, Y., Y. Gama, H. Nagahora, M. Yamaguchi, H. Naka-hara, and T. Kamata. 1987. The pH-sensitive conversion ofmolecular aggregates of rhamnolipid biosurfactant. Chem. Lett.5:763-766.

9. Itoh, S., H. Honda, F. Tomita, and T. Suzuld. 1971. Rhamnolipidproduced by Pseudomonas aeruginosa on n-paraffin (mixture ofC12, C13 and C14 fractions). J. Antibiot. 24:855-859.

10. Itoh, S., and T. Suzuki. 1972. Effect of rhamnolipids on growthof Pseudomonas aeruginosa mutant deficient in n-paraffin-utilizing ability. Agric. Biol. Chem. 36:2233-2235.

11. Klevens, H. B. 1954. Solubilization of polycyclic hydrocarbons.J. Phys. Colloid Chem. 54:283-298.

12. Lang, S., and F. Wagner. 1987. Structure and properties ofbiosurfactants. Surfactant Sci. Ser. 25:21-45.

13. Linhardt, T. J., R. Bakhit, L. Daniels, R. Mayerl, and W.Pickenhagen. 1989. Microbially produced rhamnolipid as asource of rhamnose. Biotechnol. Bioeng. 33:365-368.

14. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall.1951. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

15. Marinucci, A. C., and R. Bartha. 1979. Apparatus for monitor-ing the mineralization of volatile 14C-labeled compounds. Appl.Environ. Microbiol. 38:1020-1022.

VOL. 58, 1992

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from

3282 ZHANG AND MILLER

16. Miller, R. M., and R. Bartha. 1989. Evidence for liposomeencapsulation for transport-limited microbial metabolism ofsolid alkanes. Appl. Environ. Microbiol. 55:269-274.

17. Mulligan, C. N., G. Mahmourides, and B. F. Gibbs. 1989. Theinfluence of phosphate metabolism on biosurfactant productionby Pseudomonas aeruginosa. J. Biotechnol. 12:199-210.

18. Rendell, N. B., G. W. Taylor, M. Somerville, H. Todd, R.Wilson, and P. J. Cole. 1990. Characterization of Pseudomonasrhamnolipids. Biochim. Biophys. Acta 1045:189-193.

19. Robert, M., M. E. Mercade, M. P. Bosch, J. L. Parra, M. J.Espuny, M. A. Manresa, and J. Guinea. 1989. Effect of thecarbon source on biosurfactant production by Pseudomonasaeruginosa 44T1. Biotechnol. Lett. 11:871-874.

20. Schramm, L. L., R. G. Smith, and J. A. Stone. 1984. Asurface-tension method for the determination of anionic surfac-

APPL. ENVIRON. MICROBIOL.

tants in hot water processing of athabasca oil sands. ColloidsSurf. 11:247-263.

21. Singer, M. E., and W. R. Finnerty. 1984. Microbial metabolismof straight-chain and branched alkanes, p. 1-59. In R. Atlas(ed.), Petroleum microbiology. Macmillan Publishing Co., NewYork.

22. Somasundaran, P., and B. M. Moudgii. 1973. The effect ofdissolved hydrocarbon gases in surfactant solutions on frothflotation of minerals. J. Colloid Interface Sci. 476:290-299.

23. Warren, A. J., A. F. Ellis, and J. J. R. Campbell. 1960.Endogenous respiration of Pseudomonas aeruginosa. J. Bacte-riol. 79:875-880.

24. Zajic, J. E., and C. J. Panchel. 1976. Bio-emulsifiers. Crit. Rev.Microbiol. 5:39-66.

on February 11, 2018 by guest

http://aem.asm

.org/D

ownloaded from