Statistical Distribution of Halotolerant/Halophilic Biosurfactant- Producing Yeasts in Qua Iboe Estuarine Sediments

We isolated 100 morphologically-distinct halotolerant/halophilic yeasts from five estuarine sediments in Qua Iboe estuary, South-South Nigeria. A three-way analysis of variance (ANOVA) of the distribution was significant at p < 0.01; adjusted r = 0.851 suggesting that 85.1% of the variations in the number of halotolerant/halophilic yeast could be explained by the model. Sample, media and salt level main effects were all significant at p < 0.01, r = 0.732; 0.403 and 0.463 respectively. Only 17% of the yeasts demonstrated biosurfactant production potential by the oil displacement assay. Positive isolates were identified by macro/micro-morphological and physiological characterizations as species of Torulaspora, Pichia, Saccharomyces, Candida, Debaryomyces, Kluyveromyces, Schizosaccharomyces, Rhodotorula and Hortaea but the dominant halotolerant genus was Candida. Douglas creek sample harbored the highest number of halotolerant biosurfactant-producing yeasts probably by reason of its better proximity to Qua Iboe terminal where petroleum activity is high. Biosurfactants produced by all 17 yeasts could reduce surface tension to < 40 mN/m suggesting that halotolerant/halophilic yeasts in stressed environments secrete only effective surface-active compounds or not at all. Significant but moderate correlations existed between all correlate pairs involving oil displacement activity, surface tension, critical micelle concentration and emulsification activity of biosurfactants. The Pillai’s Trace test of a one-way multivariate analysis of variance (MANOVA) involving surface tension reduction, oil displacement activity and emulsification activity was significant at F (48,102) = 3119.004, p < 0.005, partial ƞ = 0.999.


Introduction
In spite of the huge and direct impact of pathogenic and spoilage organisms on mankind and their materials, interest in industrial organisms continues on a heightened scale owing to their huge economic, health, agricultural, environmental and academic concerns. It is very difficult in today's world to pinpoint an aspect of global life that is not affected by the invisible life forms. By industrial microorganism is meant all microorganisms that could be exploited by man for the purpose of production of value-added metabolites or the provision of valuable services through the concenpt of biotechnology.
At the forefront of industrial microorganisms are bacteria, yeasts and molds. Most industrial bioprocesses reported in literature make use of bacteria but only a few bacteria have been able to pass off as 'generally recognized as safe (GRAS)' because of the pathogenic status of most bacteria [14] The propensity towards infection arising from consumption of these metabolites explains the huge amounts of money spent in purification steps to rid industrial bacterial metabolites of toxins, pyrogen and the like [15]. The yeasts have their fair share of that concern but not nearly as much perhaps owing to the inherent difficulty to indulge in indiscriminate horizontal gene transfer that characterizes bacteria, especially conjugation. Therefore, a very good number of yeast genera, species and strains, and the industrial metabolites derived therefrom, meet the GRAS status [16]. A lot of attention has therefore been focused on yeast metabolites as alternatives to bacterial metabolites to reduce cost of production and that of final product, reduce health and environmental risks that usually attend the use of pathogenic bacteria in microbial fermentation bioprocesses.
One of such yeast exploitations is in the production of the most versatile and multifuctional microbial biomolecule of the 20 th and 21 st centurybiosurfactants [17]. Biosurfactants are biological, essentially microbially-derived, amphiphilic compounds that lower the surface and interfacial tension of substances by means of their hydrophilic and hydrophobic domains. Their surface and interfacial tension reduction potentials confer on them properties as diverse as emulsification, wetting, detergency, foaming and solubilization of hydrophobic organic compounds [18]. These properties have variously been exploited in industries like food, pharmaceutical, cosmetic, petroleum and detergent industries [19]. Some of these industrial applications especially food processing and enhanced oil recovery operate at high salt (NaCl) [20].
In this study, we investigated the distribution of halotolerant as well as halophilic yeast species in the estuarine sediments of the mangrove ecosystem of Qua Iboe River, Ibeno, Akwa Ibom State, Niger Delta region of Nigeria with potential for elaboration of surface-active compounds which would find applications in food processing and microbiallymediated enhanced recovery of oil (MEOR) in the petroleum industry.

Methodology 2.1
Sample location and collection Qua Iboe estuary is a mesotidal ecotone in the Qua Iboe River located along coordinates Lat. 4º30' -4º45'N and Long. 7º31' -8º45'E of Akwa Ibom State in the Niger Delta region of Nigeria. It experiences frequent intertidal fluctuations in salinity and is dominated by extensive intertidal mangrove and Nypa forested wetlands [21,22]. The estuary is frigned with several creeks most of which have direct opening into the coastal marine system of the Atlantic ocean. Such creeks help conduct petroleum products that spill chronically from the Qua Iboe Terminal (QIT) where petroleum production and distribution activities of Exxon Mobile dominate [23]. Our study locations included five creeks namely Iwo-okpom, Mkpanak, Upenekang, Okposo and Douglas creeks. Sample collection sites and their distances were located by means of the Global Positioning System (GPS). The samples were collected into properly labeled 7 cm diameter bottles and transported in ice-packed coolers to the laboratory for microbiological analyses.

2.2
Isolation od halotolerant/halophilic yeasts Halophily was defined as the ability of yeasts to grow in media containing 17% (~3 M) NaCl [24] and halotolerance was defined as the ability of yeast species to grow in media containing 8.75% (~1.5 M) NaCl. Three media were employed for the isolations following 100-fold dilution of sediment samples, to prevent loss of desired species due to over-dilution. Screening media employed included malt-extract yeast-extract agar (MYA), glucose yeast-extract agar (GYA) and glucose yeast-extract-peptone agar (GYPA). Media were made selective for yeast by incorporation of 50 µg/mL of chloramphenicol and medium pH adjusted to 4.0 using 0.1 M HCl/NaOH [25]. Each sample received treatment with all three media and quintuple replications of the arrangement were prepared for each treatment. A total of 150 plates were exposed to the isolation experiment; 75 for halophilic yeasts and 75 for halotolerant yeasts. All plates were incubated at 30ºC for 30 d. Cultural morphologies including pigmentation, elevation, margin, consistency and form were considered during isolation of morphologically-distinct halotolerant or halophilic yeast(s). Morphologically similar yeasts that occurred on more than one medium were only isolated from the medium where they made the most occurrence.

2.3
Screening of halotolerant/halophilic yeast isolates for biosurfactant production 2.3.1 Medium preparation for biosurfatant production by halotolerant/halophilic yeasts Minimal medium containing (g/L) KH 2 [27] was used for potential biosurfactant elaboration. The medium was supplemented with 17% NaCl (for halophilic yeasts) and 8.75 % NaCl for halotolerant yeasts. Glucose, olive oil or glycerol separately served as sole sources of carbon at 1% (v/v or w/v) and pH adjusted to 4.0 with 1 M HCl/IM NaOH. Twenty milliliter volume of medium was dispensed into 100 mL Erlenmeyer flasks and sterilized by autoclaving at 121°C for 15 min.

2.3.2
Biosurfactant production and evaluation Upon cooling, flasks were inoculated with 2% (v/v) 36 h-old broth culture of each yeast isolate. Broth cultures were prepared from the respective agar media on which the particular yeast was isolated minus agar-agar. Flasks were incubated at room temperature (28 ± 2°C) on a rotary shaker agitating at 150 rpm for 21 d.
Cell-free broth was obtained by centrifugation of flask content at 8,000 rpm for 10 min and subjecting the supernatant to membrane filtration using, first, 0.45 µM pore size filter (Millipore). The intact fermentation broth, sterile fermentation broth and yeast pellets all served as materials for biosurfactant screening.
The oil displacement assay of [28] was employed for biosurfactant screening. A linear relationship occurs between biosurfactant concentration and area of displaced oil [26] and the assay is sensitive enough to detect even nano-molar concentrations of the surfaceactive compound [28]. Briefly, 40 mL of distilled water was dispensed into a clean Petri dish of diameter 15 cm. Thereafter, 15 µL of Bonny medium crude oil was introduced at the center of the plate and allowed to equilibrate for 1 h producing a thin layer of oil film. The assay was performed by introducing 0.1 µL of intact fermentation broth (IFB) or cell-free supernatant (CFB) or needle point sample of yeast pellet (YP) at the center of the oil film. Displacement of oil film from surface was scored as positive, otherwise negative. Diameters of displaced oil film were measured after 30 s by means of a meter rule and areas of displaced oil calculated using the equation , where A is the area of biosurfactant-displaced oil film, r the radius and π a constant of value 3.14. Triplicate measurements were made at all times.

Assessment of biosurfactant properties 2.3.3.1 Assessment of oil displacement activity
The oil displacement activity was assessed by the method of Morikawa et al. (1993) as described in section 2.3.2.

Measurement of surface tension reduction
The surface tension reduction property was evaluated by the ring method of Du Nouy described in Rodrigues et al. [29] using CSC Du Nouy Tensiometer.

Determination of the critical micelle concentration
The efficiency of biosurfactant is given by its critical micelle concentration, defined as the minimum amount of biosurfactant required to form micelles and at which point further reduction in surface tension no longer occurs. This property was measured as described in Ekpenyong et al. [26]

Measurement of the emulsification activity of biosurfactant
The emulsification activities of all yeast biosurfactants were measured, using kerosene as hydrophobic compound. Equal volumes of cell-free biosurfactant broth and kerosene were mixed together in a test tube using vortex mixer (Vortex Mixer XH-C, PEC MEDICAL, USA) for 10 min. Thereafter, the mixture was allowed to stand for 24 h. Emulsion (the white layer) height and total height of mixture were measured. The ratio of the height of emulsion to total height of mixture multiplied by 100 was determined and presented as %E24.

2.4
Identification of yeasts Positive yeasts were identified by cultural morphology, sexual structures involving number, shape and size of ascospores, asexual structures like hyphae and/or pseudohyphae. Physiological studies included sugar assimilation and cycloheximide resistance and nitrogen utilization [30].

2.5
Statistical analyses Data generated were subjected to statistical analysis using descriptive statistics, pearson's correlation, two-way ANOVA and three-way ANOVA and oneway multivariate analysis of variance (MANOVA) using IBM-SPSS version 19. Bar graphs were plotted with Excel Software version 2007.

3
Results and discussion 3.1

Distribution of halotolerant/halophilic yeasts in sediment samples
The results of distribution of halotolerant and halophilic yeasts in Qua Iboe estuarine sediments are presented in Figure 1. A total of 100 morphologically-distinct types of yeasts were isolated from the five samples considered in the study. Fortyfive (45%) of the yeasts were halotolerant while 55 (55%) were able to grow in the presence of 17% NaCl as halophiles. Interestingly, all the isolates could grow optimally in the absence of salt suggesting that the definition of halophily in terms of amount of salt tolerated needs to be reviewed. Plemenitas et al. [31] reported that Hortaea werneckii is adapted to fluctuating concentrations of NaCl with growth optimum between 0.8 and 1.7 M. A truly halophilic yeast like Wallemia ichthiaphaga is reported to grow only in the presence of NaCl [32], however, we did not isolate it in this study.
When data was subjected to a three-way analysis of variance (ANOVA), the Levene's test of equality of error variances was not significant (F:1.165, p = .279 > 0.05) suggesting that the data did not violate the homogeneity of variance assumption and that we could proceed to interpret the results of the model. The results of test of between-subjects effect is presented as Table 1. The table reveals that the 3-way ANOVA model was significant at p < 0.01; r 2 = 0.880; adjusted r 2 = 0.851, and could explain 85.1% of the variations in the number of halotolerant/halophilic yeast isolated from the five different samples. The result of sample main effect was significant; p < 0.01, r 2 = 0.732 suggesting that 73.2% of the variations in the number of yeasts could be due to sample effect. Both media and salinity effects were respectively significant at r 2 = 0.403 and 0.463. The contributions of the four interaction effects were similarly significant at p < 0.01. The interaction between sample and media showed a coefficient of determination, r 2 , given in the table as Partial Eta squared, of 0.517 indicating that the interaction was significant enough to explain 51.7% of the variables about the data. Sample/salt requirement interaction could explain 59.0% of variations about the data. Interestingly, the little contribution of media/salt requirement interaction of only 9.0% was significant enough at p < 0.01 to mediate variations in the number of yeasts recovered from test samples. Finally, there was a significant three-way interaction among sample, media and salt requirement with an r 2 of 0.322 and capacity to influence 32.2% of the variations about the number of yeasts recovered from samples.
The interactive plots of the three-way interaction effects of the independent variables are presented in Figures 2 a and b. Because the error probabilities of the seven predictor terms were less than 0.05 and even 0.01, we rejected all seven null hypotheses and accepted the alternative hypotheses as true; that significant differences existed in the mean number of yeasts recovered from the five different samples using three different media exposed to two different salt levels.
The Bonferroni post-hoc multiple comparison test results showed that mean differences in number of yeast between Utan Iyatah sample and all other four samples were significant. On the contrary, results also showed that the mean difference in number of yeast obtained from Douglas Creek did not differ significantly (p = 0.111 > 0.05) from that of Mkpanak Creek sample. Non-significant mean differences also existed between Iwo-okpom and Mkpanak creek samples (p = 0.734 > 0.05), Iwo-okpom and Okposo samples (p =1.000 > 0.05) as well as between Mkpanak and Okposo samples (p = 0.176 > 0.05). Mean differences between all media pairs were significant at p < 0.01; 0.05 suggesting the strong influence of composition in mediating yeast colony development in isolation media.

3.2
Screening of halotolerant/halophilic yeasts for biosurfactant production Results of screen test for biosurfactant production by halotolerant/halophilic yeasts showed that out of the 100 yeasts screened for biosurfactant production using three different substrates, 17 (17%) could produce biosurfactant. Ten of the isolates were halotolerant yeasts while 7 were halophilic. Only DCHTM1 could produce biosurfactant in minimal medium when glucose served as carbon source and only UPHTM2 could produce biosurfactant when glycerol served as carbon source. All biosurfactantpositive isolates could produce biosurfactant in olive oil-minimal medium. This suggests that the major inducer substrates for biosurfactant production in those yeasts would be hydrophobic substrates.
In both DCHTM1 and UPHTM1, the bulk of surfaceactive compound was found associated with the yeast pellets. This suggests that the two yeast isolates will not be suitable for commercial production of the purified product by reason of the huge cost and tedion involved in product recovery. However, for enhanced oil recovery and hydrophobic compound remediation activities where purification is not an issue, isolate DCHTM1 would well pass with an oil displacement activity of 149.41 cm 2 which, from Table 2, is higher than those produced by 13 of the isolates that were positive in olive oil-minimal medium. Biosurfactant production, measured as area of displaced oil, was found in the cell-free broth (CFB) of 9 isolates. Four isolates had their surfaceactive compounds as cell-associated but location of the active compound in 4 isolates was indeterminate as both cell-free broth (CFB) and yeast pellets (YP) could produce significant area of displaced oil.  A two-way ANOVA of data, however, revealed that location of biosurfactant was not significantly different among the yeasts but that nature of substrate significantly (p = 0.038 < 0.05) influenced biosurfactant production in the isolates. A similar observation was earlier made by Ekpenyong et al. [26] in bacteria using five different substrates. The authors posited that the substrate chosen for biosurfactant production was a reflection of the use to which the organism would put the surface-active compound after synthesis.

Evaluation of biosurfactant properties
Since all 17 yeast isolates could produce biosurfactant in olive oil-minimal medium and the time to demonstration of biosurfactant release in fermentation medium was shorter in that medium than in any other, all evaluations were conducted using olive oil-minimal medium. Results of oil displacement activities, surface tension reduction potentials, emulsification activities and critical micelle concentration of biosurfactants are also presented in Table 2. The table shows that the best five biosurfactant-producing yeasts in terms of oil displacement were UPHTM2 (165.34 cm 2 ), OSHTG3 (160.59 cm 2 ), UPHTP1 (153.71 cm 2 ), DCHTM1 (149.41 cm 2 ) and DCHPG2 (145.69 cm 2 ). These isolates were obtained from Upenekang, Okposo and Douglas Creek samples. The codes of the isolates also indicate that the first four isolates were halotolerant yeasts while the last one was halophilic. The five poorest biosurfactant producers in terms of oil displacement activity were OSHPG1 (55.61 cm 2 ), UPHTM1 (73.72 cm 2 ), MKHPG1 (75.61 cm 2 ), DCHTM2 (91.63 cm 2 ) and OSHTP1 (95.22 cm 2 ).
The effectiveness of a biosurfactant is measured by its surface tension reduction potential. The results presented in Table 2 show that surface-active compounds from all 17 isolates could reduce the surface-tension of minimal broth to ≤ 40 mN/m. A biosurfactant is considered effective if it reduces surface tension of water from 72.00 mN/m to 40.00 mN/m at room temperature [33]. Five yeasts isolates had biosurfactants with surface tension reduction potential below 30.00 mN/m and these included UPHTM2 (28.48 mN/m), OSHTG3 (28.83 mN/m), OSHPG1 (29.57 mN/m), MKHPP2 (29.59 mN/m) and UPHPM2 (29.61 mN/m). Apart from UPHTM2 and OSHTG3 isolates, the remaining three yeasts with very effective biosurfactants were different from the yeasts with excellent oil displacement activity. This suggests some form of relationship between oil displacement activity and surface tension reduction potential.
To establish the efficiency of a biosurfactant, we measured the critical micelle concentration (cmc). The cmc answers the question: just how much biosurfactant is needed to bring about the surface tension reduction potential discussed in the earlier paragraph: the lower the cmc, the more efficient the biosurfactant. The results presented in Table 2 reveal that UYHPM2 was the most efficient biosurfactant with a cmc of 18.71 mg/L. Other efficient biosurfactants were from isolates OSHPG1 (19.38 mg/L), MKHPG1 (19.77 mg/L), DCHPG1 (21.54 mg/L) and DCHTM2 (22.67 mg/L). Once again, only isolate OSHPG1 has had an earlier mention in terms of excellence of activity.
Finally, results of emulsification activity of biosurfactants show that isolate DCHPG2 had the best emulsifying activity given by emulsification index of 77. 71%. Other biosurfactants with high emulsifying activities came from OSHTG3 (77.53%), DCHTM1 (76.56%), IWHTP1 (74.71%) and OSHTG2 (72.56%). Isolates OSHTG2 and IWHTP1 are now new entrants into the excellence activity group of biosurfactants. These results, as said earlier, only suggest some form of relationship among these activities. Table 3 and reveals that the four parameters of oil displacement activity (ODA), surface tension reduction potential (ST), critical micelle concentration (cmc) and emulsifying activity (E24) very significantly (p < 0.01) correlated among one another. Oil displacement activity (ODA) had a positive but moderate correlation (r = 0.674) with E24 but its relationship with ST was negative, weak but significant (r = -0.376, p = 0.006 < 0.01) and that with cmc was positive but moderate (r = 0.539, p = 0.000 < 0.01). Surface tension (ST) had moderately negative correlations of r = -0.433, p = 0.002 < 0.01 and r = -0.511, p = 0.000 < 0.01 with cmc and E24 respectively. Finally, cmc had a positive but weak correlation with E24 with a significant correlation coefficient of r = 0.370, p = 0.008 < 0.01.

Yeast identities
The identities of the 17 halotolerant/halophilic biosurfactant-producing yeasts isolated in this study are presented in Table 5. The identity of the isolates revealed a diverse genera of halotolerant yeasts in Qua Iboe Estuary. At least a single biosurfactant-producing halotolerant yeast was identified from each of the locations studied.

Candida:
This genus had the highest number of biosurfactant-producing halotolerant yeasts isolated from study samples. Candida versatilis strain DCHPG1 and Candida tolerans strain DCHPG2 were halophilic yeasts isolated from Douglas creek sample while Candida tropicalis strain OSHTP1 and Candida magnoliae strain OSHPG1were respectively halotolerant and halophilic yeasts isolated from Okposo creek sample. None of the Candida strains could grow in the presence of diazonium blue-B but whereas C. versatilis and C. tropicalis could grow in the presence of 0.01% of cycloheximide, C. tolerans and C. magnoliae could not. Candida tolerans and C. tropicalis could grow over a wide range of temperature (25 -40ºC) but those for the other species were narrow. .000 .000 .008 **. Correlation is significant at the 0.01 level (2-tailed). ODA-Oil displacement activity (cm 2 ); ST-Surface tension (dynes/cm); CMC-Critical micelle concentration (mg/L); E24-Emulsification activity (%). All strains of Candida in this study could use lysine as nitrogen source but only Candida versatilis and Candida magnoliae could assimilate nitrogen from the environmentally-abundant form of nitrogen, viz nitrate. The details of their carbon assimilation profiles are presented in Table 5. Their biosurfcatant production potentials also varied. Table 2 reveals that although the amount of biosurfactant produced by Candida mangoliae strain OSHPG1 was the least (55.61 cm 2 ) and therefore produced marginal emulsification activity of 58.23%, the biosurfactant was the most effective and efficient. However, C. tolerans biosurfactant is recommended for large-scale production for use in tertiary oil recovery, pharmaceutical formulations as well as in environmental bioremediation where its properties are required. Pichia: Three species of this genus were identified in this study and included the halotolerant strains of Pichia anomala strain DCHTM2, Pichia membranifaciens strain OSHTG3 isolated from Douglas and Okposo creeks respectively, and the halophilic Pichia farinosa strain MKHPG1 from Mkpanak Creek. Significantly, P. farinosa grew at 45ºC while P. anomala and P. membranifaciens had their temperature maximum at 37ºC and 35ºC respectively. Pichia membranifaciens displayed very fastidious requirement for assimilable nitrogen and could only assimilate nitrogen from D-glucosamine but P. anomala was the only one of the three that could assimilate nitrogen from nitrate. In terms of carbon assimilation, P. membranifaciens was also quite stringent, using only glucose, glycerol and ethanol as sole sources of carbon and energy out of the 19 carbon sources tested. While P. farinosa could assimilate carbon from L-sorbose and D-xylose, P. anomala could not but rapidly developed growth when sucrose served as carbon source suggesting that the yeast is a good producer of invertase which splits sucrose into the component monosaccharides for effective carbon assimilation from glucose. The three species also significantly differed in terms of colonial morphology as P. anomala was cream in colour, P. fabrinosa white and P. membranifaciens was yellow.
Only Pichia membranifaciens strain OSHTG3 is recommended for biosurfactant production given the excellent properties of its biosurfactant presented in Table 2: Mean oil displacement activity of 160.59 ± 0.16 cm 2 , mean surface tension reduction potential of 28.83 ± 0.14 mN/m, mean critical micelle concentration of 28.73 ± 0.18 mg/L and mean emulsification activity of 77.53 ± 0.27%. However, Thaniyavarn et al. [35] produced sophorolipid from Pichia anomala strain PY1 when grown on soybean oil as carbon source. Surface tension decreased to 28 mN/m with an oil displacement activity of 69.43 cm 2 . Our Pichia anomala strain could reduce surface tension to 30 mN/m with an oil displacement activity of 91.63 cm 2 at a cmc of 22.67 mg/L. The major sources of difference between the two reports could be attributaed to carbon source and NaCl stress.
Torulaspora: Two species of this genus were isolated from Douglas and Mkpanak creeks. The species from Douglas Creek was identified as Torulaspora globosa strain DCHTM1 while that from Mkpanak creek was Torulaspora delbrueckii strain MKHPP2. From the strain code of the yeast, T. globosa was halotolerant, growing in the presence of NaCl concentration of 8.75% (w/v) and T. delbrueckii was halophilic capable of growth in 17% (w/v) NaCl within 21 d. Very significant differences between the two strains were that T. globosa could grow in the presence of 0.01% cycloheximide while T. belbrueckii could not. While T. delbrueckii could grow at 25ºC, T. globosa could not. Interestingly, while T. globosa had a convex growth in agar media, T. delbrueckii appeared concave on glucose yeastextract peptone agar. Torulaspora globosa produced larger amount of biosurfactant than T. delbrueckii, however, the surface-active compound was less effective than that of the later. Table 5 Characterization and Identification of yeast isolates Table 5 Characterization and identification of yeast isolates cont.  castelli failed to assimilate carbon from methanol, Dgluconate and D-arabinose. Their potentials for biosurfactant production were very similar especially in amount of surface-active compound produced, its effectiveness and efficiency. This is the first report on biosurfactant production potential of these two species of Debaryomyces under high salt conditions.

Rhodotorula:
The characteristically red mucoid strain UPHTM2 was identified as Rhodotorula mucilaginosa while the pink butyrous colony was identified as Rhodotorula minuta. The redpigmenting mucoid Rhodotorula mucilaginosa strain UPHTM2 isolated from Upenekang sediments could assimilate nitrogen from nitrate and tryptophan while its pink counterpart could not. Contrariwise, R. minuta assimilated nitrogen from ethylamine while R. mucilaginosa could not. Neither of them could grow in the presence of 17% NaCl , so were described as halotolerant. Details of their differences and similarities in carbon assimilation is presented in Table 5. Not much difference was found in their biosurfactant production potentials with R. mucilaginosa releasing 165.34 cm 2 ; the highest and most effective (28.48 mN/m) of all 17 yeasts biosurfactants studied. No detailed description of biosurfactant production by Rhodotorula minuta exists in literature, therefore this is the first report of such a potential and under high salt levels. Kawahara et al. [36] reported the production of a glycoprotein biosurfactant from Rhodotorula mucilaginosa isolated from the Antarctica and demonstrated its ability to stabilize astaxanthin. Sousa et al. [37] also reported biosurfactant production by a strain of Rhodotorula mucilaginosa using pineapple residues, however, the biosurfactant was not nearly as effective (42.88 mN/m) as that from the present study (28.48 mN/m), neither was it produced under high NaCl levels.
The remaining four yeast genera had a species each and included Saccharomyces exiguus strain DCHTM3, Kluyveromyces marxianus strain OSHTG2, Schizosaccharomyces pombe strain UPHTM1 and the black mucoid yeast Hortaea werneckii strain UPHPM2. Apart from Hortaea werneckii, the other three were halotolerant yeasts. Hortaea werneckii strain UPHPM2 was the most fastidious of all yeasts studied being only able to grow on glucose and lactose out of the 19 carbon sources screened for carbon assimilation. Its preferred nitrogen sources included nitrate and nitrite. The yeast could only grow at temperatures between 25 and 35ºC. The yeast accumulated biosurfactant that could cause an oil displaced area of 109.60 cm 2 . Interestingly, this biosurfactant was the most efficient with a critical micelle concentration (cmc) of 18.71 mg/L. To the best of our knowledge, this is the first report on potential for biosurfactant production by a halotolerant strain of Hotaea werneckii, Saccharomyces exiguous, Schizosaccharomyces pombe. However, in this study, Kluyveromyces marxianus strain OSHTG2 produced a surface-active compound in the presence of 8.75% NaCl (w/v) which emulsified kerosene by 72.56% and was stable for more than 3 months at 4ºC. This result corroborates those by Lukondeh et al. [38] who reported the production of a natural mannan-protein emulsifying biosurfactant by Kluyveromyces marxianus strain FII 510700 when grown on lactosebased medium. Their bioemulsifier had an emulsification index of 70% when NaCl concentration was 5% as against the result of the strain in this study.

Conclusion
This study has identified a couple of novel yeast isolates with potentials to produce surface-active compounds under salt-stressed conditions. Excellent biosurfactant producer in the study included Rhodotorula mucilaginosa strain UPHTM2 that produced large amount of surface-active compound which was very effective and very efficient. Since all the biosurfactants produced by the 17 yeasts in this study could reduce surface tension to < 40 mN/m, we conclude that halotolerant/halophilic yeasts in stressed environments secrete only effective surfaceactive compounds or not at all. Most of these biosurfactants when, exploited could find applications in microbial enhanced oil recovery and