Present interest in hydrocarbon biodegradation is motivated for a large part by their presence as environmental pollutants (Atlas 1984; Leahy & Colwell 1990). Crude oil is a major sea pollutant and petroleum products, such as gasoline or diesel and fuel oils, are the most frequent organic pollutants of soils and ground-waters. The biodegradation of hydrocarbons has a high ecological significance as it constitutes the major process for remediation of the contaminated areas. Research aims at a better understanding of the fate of polluting hydrocarbons (natural attenuation) and of the prospects for biotechnological improvement of biodegradation processes (engineered bioremediation). A key point, which is only partially understood, is the mechanism of uptake by micro-organisms of these strongly hydrophobic compounds. For long-chain alkanes, which are practically non-water soluble, two uptake modes are generally considered (Boulton & Ratledge 1984; Singer & Finnerty 1984; Haferburg et al. 1986; Hommel 1994): interfacial accession (direct contact of cells with hydrocarbon droplets) and biosurfactant-mediated hydrocarbon uptake (cell contact with so-called accommodated or with solubilized hydrocarbons). Such mechanisms are also involved in the biodegradation of poorly soluble, polycyclic aromatic hydrocarbons (Déziel et al. 1996; Bouchez et al. 1997). A question often raised concerning hydrocarbon biodegradation in the environment is the importance of the contribution of biosurfactants to the process. However, this complex question remains a matter of debate, as conflicting results have been reported regarding the efficiency of biosurfactants in promoting hydrocarbon degradation in various situations (Haferburg et al. 1986). Another approach was followed in the present work. A series of bacterial strains representative of the microflora degrading long-chain alkanes was isolated and identified. Physiological properties useful as criteria for diagnosing mechanisms of alkane uptake were selected and their determination was applied to the strain collection. The approach aimed at widening the basis of observations concerning alkane uptake with respect to bacterial diversity, and at discerning the relative distribution of either mode of uptake in alkane-degrading bacterial strains.
The relative distribution of the modes of hydrocarbon uptake, used by bacteria of the environment for the degradation of long chain alkanes, has been evaluated. The first mode of uptake, direct interfacial accession, involves contact of cells with hydrocarbon droplets. In the second mode, biosurfactant mediated transfer, cell contact takes place with hydrocarbons emulsified or solubilized by biosurfactants. Sixty one strains growing on hexadecane were isolated from polluted and non polluted soils and identified. The majority (61%) belonged to the Corynebacterium Mycobacterium Nocardia group. Criteria selected for characterizing hexadecane uptake were cell hydrophobicity, interfacial and surface tensions and production of glycolipidic extracellular biosurfactants. These properties were determined in flask cultures on an insoluble (hexadecane) and on a soluble (glycerol or succinate) carbon source for a subset of 23 representative strains. Exclusive direct interfacial uptake was utilized by 47% of studied strains. A large proportion of strains (53%) produced biosurfactants. The data on cellular hydrophobicity suggested the existence of two distinct alkane transfer mechanisms in this group. Accordingly, tentative assignments of biosurfactant mediated micellar transfer were made for 11% of the isolated strains, and of biosurfactant enhanced interfacial uptake for 42%.
Materials and methods
Trypticase soy agar (TSA) was obtained from Biomérieux. The vitamin-supplemented mineral salt medium (MSM4) used contained (g l−1): 2·8 KH2PO4, 16·7 Na2HPO4.12H2O, 3·6 KNO3, 0·1 MgSO4.7H2O, 0·001 FeSO4.7H2O, trace elements and vitamins (Bouchez et al. 1995).
Isolation of strains
A series of bacterial strains was isolated, by selective enrichment, from soils polluted by hydrocarbons (gasoline, diesel oil, coal tar) or from unpolluted (garden and forest) soils, for their capacity to use hexadecane as sole source of carbon and energy. Enrichment was conducted with 1 g soil in 250 ml flasks containing MSM4 (100 ml) with hexadecane (1·8 ml) as carbon source. After 1 week of enrichment (30 °C, orbital shaking at 160 rev min−1), cultures were plated on MSM4 agar plates containing hexadecane on the inside wall of the lid. Morphologically different colonies were then purified on TSA. The degradative capacity of purified strains was verified by growth on hexadecane-MSM4 agar plates. Strains from other origins, using hexadecane as sole source of carbon and energy, were also used; strain GL1 was isolated from a soil of a manufactured gas plant (Arino et al. 1996), and strains PyrGe1, Mu1.4 and NapRu1 were isolated for their capacity to use pyrene, anthracene and naphthalene, respectively, as growth substrate. Strains were stored at −80 °C in 50% v/v glycerol/physiological salt solution.
Identification of strains
Identification was performed by numerical taxonomy with API kits (Biomérieux). Gram-positive rods and Gram-negative, oxidase-negative rods were identified on API Coryne and API 20 NE kits, respectively. Gram-positive, catalase-positive cocci were identified on API Staph kits.
Cultures were incubated in duplicate in 250 ml flasks containing 100 ml MSM4. Glycerol, succinate and hexadecane (12 g carbon l−1) were used individually as sole sources of carbon and energy. Flasks were inoculated (1% v/v culture grown on the same medium) and incubated on a rotary shaker (160 rev min−1) at 30 °C. Growth was followed by optical density (O.D.) measurement at 600 nm and by nitrate consumption for cultures grown on soluble substrate and on hexadecane, respectively. Residual nitrate concentrations of culture supernatant fluid were determined with nitrate reductase (Beutler & Wurst 1986) using commercial kits (Boehringer Mannheim).
Glycoside and glycolipid determination.
Glycosides were evaluated on supernatant fluids of cultures by the colorimetric method of Dubois et al. (1956). Glucose was used as a standard and results were expressed in glucose equivalents. Each result was the average of three determinations, and standard deviation was within 5%. Glycolipids were recovered as follows. Supernatant fluid of cultures was heated at 100 °C for 10 min, acidified to pH 2 after cooling and centrifuged in order to eliminate extracellular proteins. Subsequently, deproteinized supernatant fluid was extracted three times with ethyl acetate (1/1 v/v); finally, after solvent evaporation, extracted glycolipids were taken up in methanol. The different glycolipidic types were separated by analytical thin layer chromatography (TLC) carried out on silica gel plates 60 F-254 (Merck) using the solvent system CHCl3/CH3OH/ H2O (65/25/4 in volume). Detection was performed with the Molish reagent (Arino et al. 1996).
Interfacial tension and surface tension measurement.
Interfacial tension against hexadecane (γi) and surface tension (γs) were determined, at 30 °C, on filtered (0·45 μm Analypore filters, OSI, Elancourt, France) supernatant fluid of cultures with, respectively, the De Nouy ring method and the Wilhelmy plate method, using a K-12 tensiometer (Krüss, Hamburg, Germany). Stabilization was allowed to occur until standard deviation of 10 successive measurements was less than 0·4 mN m−1. Each result was the average of 10 determinations after stabilization.
Cell hydrophobicity was measured by bacterial adherence to hydrocarbons (BATH) according to a method similar to that described by Rosenberg et al. (1980). The cells were washed twice and resuspended in a buffer salt solution (16·9 g l−1 K2HPO4, 7·3 g l−1 KH2PO4) to give an O.D. at 600 nm of 0·5. The cell suspension (2 ml) with 100 μl hexadecane added was vortex-shaken for 3 min in haemolysis tubes (10 × 100 mm). After shaking, hexadecane and aqueous phases were allowed to separate for 1 h. The O.D. of the aqueous phase was then measured at 600 nm. Hydrophobicity is expressed as the percentage of adherence to hexadecane calculated as follows: 100 × (1–O.D. of the aqueous phase/OD of the initial cell suspension). For a given sample, three independent determinations were made and the standard deviation was within 5%.
Strain isolation and identification
Sixty-one different bacterial strains were selected for their capacities to use hexadecane as sole source of carbon and energy, 57 of them having being isolated on this compound. Phenotypic identification showed that strains belonged to various genera (Fig. 1), with the Corynebacterium/Mycobacterium/Nocardia (CMN) group (Barksdale 1970) predominating.
Thirty-four categories of strains were defined, each one regrouping strains with the same phenotypic characteristics on API kits and identical colonial morphology when grown on TSA and on MSM4-hexadecane plates (data not shown). For further studies, 23 strains (presented in Tables 1, 2 and 3) were chosen. They belonged to 19 different categories of strains representing 73·5% of the total. For the two largest categories, several strains were tested in order to study the homogeneity of substrate uptake characteristics.
|Strain name||Strain identification||Relative frequency |
|Cell hydro- |
|GL1||Pseudomonas aeruginosa||6·6||hexadecane |
|1·3 (S) |
|Au1||Pseudomonas aeruginosa||1·6||hexadecane |
|2·8 (S) |
|Es1||Pseudomonas fluorescens||1·6||hexadecane |
|0·5 (S) |
|0·4 (Ps) |
|Vi1||Alcaligenes faecalis||1·6||hexadecane |
|0·5 (Ps) |
|Strain name||Strain identification||Relative frequency |
|Cell hydrophobicity |
|Fl2||Micrococcus sp.||3·3||77||<2||28·0||0·5 (S)|
|Cb1||Micrococcus sp.||1·6||21||16·1||50·0||0·5 (S)|
|Ou3||Micrococcus sp.||1·6||92||15·6||49·5||2·1 (Ps)|
|Ac7||Staphylococcus caprae||1·6||83||26·4||37·9||0·2 (Ps)|
|Strain name||Strain identification||Relative frequency |
|Cell hydro- |
|Ou2||Rhodococcus equi||19·7 |
|Fo2||Rhodococcus equi||8·2||hexadecane |
|0·2 (Ps) |
|HeA1||Rhodococcus equi||4·9||hexadecane |
|0·2 (Ps) |
|HdGe1||Rhodococcus equi||3·3||hexadecane||37||21·0||51·4||1·0 (Ps)|
|PyrGe1||Rhodococcus equi||3·3||hexadecane |
|0·8 (Ps) |
|NapRu1||Rhodococcus equi||1·6||hexadecane |
|0·1 (Ps) |
|Ju1·2||Corynebacterium jeikeium||6·6||hexadecane||89||22·8||47·3||0·4 (S)|
|Mu1·4||Corynebacterium sp.§||1·6||hexadecane |
|0·3 (Ps) |
Characterization of the cultures in relation to the mode of substrate uptake
Several types of determinations were selected for diagnosing the modes of substrate uptake used by the strains in alkane-grown cultures. They involved cell hydrophobicity, which constitutes an important parameter in interfacial uptake, and surface and interfacial tensions of culture supernatant fluids, which provide basic information concerning biosurfactant-mediated substrate transfer. For further characterization of biosurfactant production, the production of glycolipids, one of the main classes of biosurfactants, was also studied. First, glycoside concentrations in culture supernatant fluid were determined. Then, the glycosides were tentatively classified as glycolipidic biosurfactants (S) when they were amphiphilic (as shown by TLC migration after ethyl acetate extraction) or as polysaccharides (Ps) when they were not. These properties were determined at the end of the growth phase but kinetic studies performed with strains GL1 and NapRu1 showed that they did not undergo large variations after the initial growth phase (data not shown).
In order to promote biosurfactant production, cultures were conducted, with nitrate as nitrogen source, until nitrate depletion (Guerra-Santos et al. 1984; Syldatk et al. 1985; Ramsay et al. 1988; Arino et al. 1996; Desai & Banat 1997). Moreover, in most cases, two carbon sources were tested, one insoluble (hexadecane) and one soluble (glycerol or succinate), to compare their influence on biosurfactant production.
The results obtained with Gram-negative rods are presented in Table 1. Pseudomonas aeruginosa GL1 and Au1 presented identical characteristics in cultures on hexadecane or on glycerol, i.e. low cell hydrophobicity and high biosurfactant production resulting in low surface and interfacial tensions. Cultures of Ps. fluorescens Es1 also presented very low surface and interfacial tensions, but higher cell hydrophobicity, when grown on hexadecane. During growth of strain HeB2, the glycosides produced were polysaccharides which, consequently, reduced surface and interfacial tensions only moderately. It can be noted that, although no glycolipidic biosurfactants were produced by Alcaligenes Vi1, interfacial tension was notably decreased, suggesting the production of another class of biosurfactants.
Each of the three strains of Micrococcus tested presented different culture characteristics when grown with hexadecane as sole source of carbon and energy (Table 2). For Micrococcus Ou3, high cell hydrophobicity, relatively high surface tension and intermediate interfacial tension of the culture supernatant fluid were observed. The high amount of polysaccharides produced may explain the interfacial tension decrease. The complex role of extracellular polysaccharides in hydrocarbon uptake has been well documented in the case of emulsan produced by Acinetobacter calcoaceticus strains (Rosenberg & Rosenberg 1981; Sar & Rosenberg 1983; Rosenberg 1993; Marin et al. 1996). Cultures of Micrococcus Fl2 were different, as the glycosides produced were glycolipids and the surface and interfacial tensions of the culture supernatant fluid were low. Micrococcus Cb1 also produced amphiphilic glycosides which, surprisingly, only moderately decreased interfacial and surface tensions of culture supernatant fluids. It can be supposed that, in this case, concentrations of amphiphilic glycosides were under the critical micellar concentration (CMC). During growth of Staphylococcus caprae Ac7, compounds active at the air/aqueous phase interface (medium surface tension), but inactive at the interface of the aqueous and hydrophobic phases, were produced (Table 2).
The results obtained with the strains of the CMN group are presented in Table 3. Most of the strains of this group (Rhodoccocus Fo2 and La2.1, two strains of the same category, Rhodoccocus PyrGe1 and NapRu1, and Corynebacterium Mu1.4, Co1 and Ju1.2) presented identical culture characteristics on soluble and on insoluble substrates (when tested on both), i.e. high values for cell hydrophobicity and for surface and interfacial tensions. The glycosides produced were polysaccharides except for strains PyrGe1 grown on glycerol and Ju1.2 grown on hexadecane where glycolipids were found. During culture of Coryne. pseudodiphteriticum Ac2 and of R. equi Ou2 on hexadecane, glycolipids were produced that decreased surface and interfacial tensions. It is notable that, for strain Ou2, culture characteristics were very different depending on whether growth took place on hexadecane or on glycerol, a point already observed in the case of Ps. fluorescens Es1 (Table 1). Properties of cultures on hexadecane of strains of the same category, BB1, SGB1 and Ou2, were quite similar although in the same category, cultures of DuB1 presented a lower cell hydrophobicity.
Strain classification by mode of substrate uptake
Strains were classified according to their predominant mode of substrate uptake. It was considered that the main characteristics of interfacial uptake were a high cellular hydrophobicity to allow cell adherence to the hexadecane phase, and a high interfacial tension of culture supernatant fluid proving that there was no biosurfactant production. The requisites retained for strictly biosurfactant-mediated transfer were a low interfacial tension as well as a low cell hydrophobicity, preventing cell adherence to hydrophobic interfaces. It is of interest to note that this classification, as presented in Table 4, points to the existence of strains that could employ interfacial uptake, but for which biosurfactant production took place and was thus susceptible to enhanced hexadecane uptake.
|Predominant mode of |
|Interfacial uptake||high||high||Rhodococcus equi Fo2|
|high||high||R. equi La2·1|
|high||high||R. equi NapRu1|
|high||high||R. equi HeA1|
|high||high||R. equi PyrGe1|
|high||high||Corynebacterium sp. Co1|
|high||high||Corynebacterium sp. Mu1·4|
|high||high||Coryne. jeikeium Ju1·2|
|high||high||Staphylococcus caprae Ac7|
|medium||high||Gram-negative rod HeB2|
|medium||high||R. equi HdGe1|
|Biosurfactant-mediated||low||low||Pseudomonas aeruginosa Au1|
|uptake||low||low||Ps. aeruginosa GL1|
|Biosurfactant-enhanced||high||low||Micrococcus sp. Fl2|
|interfacial uptake||high||low||R. equi Ou2|
|high||low||R. equi BB1|
|high||low||R. equi SGB1|
|high||low||Coryne. pseudodiphteriticum Ac2|
|high||medium||Micrococcus sp. Ou3|
|medium||low||Ps. fluorescens Es1|
|medium||medium||R. equi DuB1|
|medium||medium||Alcaligenes faecalis Vi1|
|medium||medium||Micrococcus sp. Cb1|
Most of the strains isolated belonged to genera known for their capacity to degrade alkanes (Britton 1984; Goodfellow 1992; Rosenberg 1992). All soil samples tested (polluted as well as non-polluted) yielded hexadecane-degrading strains, confirming the widespread distribution in the environment of such strains. This fact probably reflected the ubiquitous presence of long-chain alkanes from plants in the environment (Kolattukudy 1976; Berdiéet al. 1995).
Broadly speaking, the results concerning hexadecane uptake were in line with the modes of substrate transfer previously described in the literature (Singer & Finnerty 1984; Haferburg et al. 1986; Goswami & Singh 1991; Hommel 1994), direct interfacial uptake and biosurfactant-mediated uptake, but they gave new physiological and environmental insights into the diversity of alkane uptake mechanisms, as detailed below.
Direct interfacial uptake appeared to be the most frequent mechanism. Assuming, as suggested by the data, that strains belonging to the same category presented the same characteristics of substrate transfer as the selected strains, an interfacial mechanism was employed by 46·7% of strains from the set. A natural habitat for such strains may be one that maximizes hydrocarbon interfaces such as the phylloplane of waxy plants. Neu & Poralla (1990) reported the isolation of bacteria with hydrophobic cell surfaces from such sources.
The effective production of biosurfactants is usually viewed as the obvious criterion for the existence of biosurfactant-mediated hydrocarbon transfer. In the present study, both hydrophilic and hydrophobic bacteria were able to produce biosurfactants, as reported by Neu & Poralla (1990). Several hydrophobic strains, however, did not produce biosurfactants. These observations do not corroborate a recently held view associating high cell hydrophobicity with biosurfactant production capacity (Pruthi & Cameotra 1997). The mode of action of biosurfactants is usually considered to promote both substrate emulsification and solubilization. In fact, the case where biosurfactants act mainly by solubilization of the alkane, thus resulting in the formation of micelles with a hydrophilic outer layer, appears to be a mode of transfer well suited to hydrophilic cells, allowing efficient contact with the alkane-degrading bacteria. Among the strains studied, such a mechanism (micellar transfer) appears to be appropriate for the two Ps. aeruginosa strains, although this is not thought to be a general pattern for this species (Zhang & Miller 1994). It can be noted that the biosurfactants produced by Ps. aeruginosa GL1 have been previously characterized as rhamnolipids (Arino et al. 1996) and that biosurfactants produced by Ps. aeruginosa Au1 were likely to be rhamnolipids as well, as suggested by their migration pattern on TLC (migration identical to that of GL1 surfactants). These observations were in agreement with the literature (Guerra-Santos et al. 1984; Syldatk et al. 1985; Parra et al. 1989; Robert et al. 1989). Using the dilution method (Bosch et al. 1988), we observed (data not shown) that the low interfacial and surface tensions values corresponded to biosurfactant concentrations above the CMC determined (30-fold the CMC for biosurfactants produced by GL1 at the end of growth on hexadecane). Such conditions appear favourable to alkane solubilization and indeed, we observed solubilization of hexadecane (around 15 mg l−1) in culture supernatant fluid of strain GL1.
A different situation was observed for a large group of bacteria (42·2% of the set) exhibiting both a high or medium hydrophobicity and production of biosurfactants. Assuming that such strains utilized both direct interfacial uptake and micellar transfer poses a problem as the solubilization of alkanes implies the existence of hydrophilic micelles. The data on bacterial cell hydrophobicity suggest that for hydrophobic biosurfactant-producing bacteria, the alkane transfer mechanism involves, primarily, substrate emulsification rather than solubilization, emulsified alkane droplets containing hydrophobic regions on their surface. This is in line with the observation that supernatant fluids of hexadecane-grown cultures of several strains of this group (such as Micrococcus Fl2, Corynebacterium Ac2 and Pseudomonas Es1) were found to produce emulsions that remained stable after several months. In fact, the question of the mode of action of biosurfactants in promoting emulsification (Roy et al. 1979; Rosenberg & Rosenberg 1981; Neu & Poralla 1990) or solubilization (Roy et al. 1979; Rosenberg & Rosenberg 1981; Zhang & Miller 1995) of substrate, or in modifying cell hydrophobicity (Zhang & Miller 1994), is long standing. In a marine bacterium, the coexistence of solubilization and emulsification mechanisms has been proposed (Husain et al. 1997). The hydrophobicity data presented here indicate the occurrence of both mechanisms but with one or the other usually being privileged, depending on the strain concerned.
Biosurfactants were observed in significant amounts for a large proportion (53·3%) of the set of strains studied. This is somewhat high compared with other authors’ data (Bosch et al. 1988; Parra et al. 1989; Neu & Poralla 1990; Mercadéet al. 1996). Production of biosurfactants was also observed during growth on a soluble carbon source, which is in agreement with Haferburg et al. (1986). These observations suggest a broader role for biosurfactants than just hydrocarbon uptake. As discussed by Neu (1996), a likely possibility is the more general participation in adhesion and de-adhesion interactions between micro-organisms and interfaces. Biosurfactant-enhanced interfacial uptake of alkanes could be one aspect of this general role.
The authors thank V. Bardin for strain isolation, H. Bouzrara for participation in strain characterization and C. Dalmazzone for helpful discussions.
* Present address: Laboratoire de Maîtrise des Technologies Agro-Industrielles, Université de la Rochelle, Avenue Marillac, 17042 La Rochelle cedex 01, France.