Journal of Chromatography A, 825 (1998) 149–159
`
`General approach for the development of high-performance liquid
`chromatography methods for biosurfactant analysis and purification
`*
`Sung-Chyr Lin , Yi-Chuan Chen, Yu-Ming Lin
`Department of Chemical Engineering, National Chung Hisng University, Taichung, Taiwan
`
`Received 22 June 1998; received in revised form 28 August 1998; accepted 28 August 1998
`
`Abstract
`
`A general approach for the development of HPLC methods for biosurfactant analysis and purification was proposed. By
`comparing the chromatograms of the cell-free fermentation broth, the ultrafiltration filtrate, and the ultrafiltration filtrate of a
`methanol–surfactant mixture, the peaks corresponding to biosurfactants can be identified without any prior structural
`information of the biosurfactants. It can be assumed that the peaks observed only on the chromatogram of the filtrate of
`methanol–surfactant mixture but not on the chromatogram of the filtrate are biosurfactant peaks. This approach can be
`applied for the development of a HPLC assay for any biosurfactants as long as the concentration of biosurfactants in the
`fermentation broth is higher than the critical micelle concentration. The HPLC methods thus developed can also be adapted
`for the preparation of homogeneous biosurfactant samples useful for chemical analysis for the elucidation of chemical
`structure of biosurfactants and for the determination of the physical properties of biosurfactants.
`1998 Elsevier Science
`B.V. All rights reserved.
`
`Keywords: Ultrafiltration; Surfactants; Lipopeptides; Surfactin
`
`1. Introduction
`
`Surfactants are amphiphilic molecules, consisting
`of hydrophilic and hydrophobic domains, which tend
`to partition preferentially at the interface between
`fluids of different degrees of polarity and hydrogen
`bonding. The formation of an ordered molecular
`layer at the interface lowers the interfacial tension
`and attributes to the unique surface properties of
`surfactants. Due to the unique interfacial behavior,
`surfactants find applications in various industrial
`
`*
`
`author. Fax: 1886
`Corresponding
`sclin@dragon.nchu.edu.tw
`
`4
`
`2852587, E-mail:
`
`foaming, de-
`involving emulsification,
`processes
`tergency, wetting and phase dispersion or solubiliza-
`tion.
`Many biological molecules exhibiting particularly
`high surface activity are classified as biosurfactants.
`Microbial biosurfactants included a wide variety of
`chemical structures, such as glycolipids,
`lipopep-
`tides,
`polysaccharide–protein
`complexes,
`phos-
`pholipids, fatty acids and neutral
`lipids [1–7]. In
`terms of physicochemical properties such as surface
`activity as well as pH and heat stability, many
`biosurfactants are comparable to synthetic surfactants
`[6]. Biosurfactants possess some advantages, such as
`low critical micelle concentration (CMC) and high
`biodegradability, over
`synthetic surfactants and,
`therefore, are particularly well suited for environ-
`
`0021-9673/98/$ – see front matter
`PII: S0021-9673( 98 )00709-2
`
`1998 Elsevier Science B.V. All rights reserved.
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 1 of 11
`
`(cid:211)
`(cid:211)
`

`
`150
`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`mental applications such as bioremediation and the
`dispersion of oil spills [8–12].
`Due to some technical and/or economic reasons,
`biosurfactants have not been employed extensively in
`industry. Like most microbial metabolites, biosurfac-
`tants exist in fermentation broth of complex com-
`position at relatively low concentrations, which often
`makes the costs associated with the isolation and
`purification of biosurfactants prohibitively high. To
`make the large-scale production of biosurfactants
`possible,
`it
`is generally necessary to undergo the
`time-consuming and labor-intensive strain improve-
`ment programs and the systematic medium optimi-
`zation studies. The success of strain improvement
`programs and medium optimization studies generally
`relies on the availability of efficient and specific
`analysis techniques for biosurfactants. Unfortunately,
`for most biosurfactants reported so far the techniques
`frequently employed for the detection of biosurfac-
`tants have been surface /interfacial tension measure-
`ments [13–17], which do not meet
`the desired
`criteria. The results of tension measurements for the
`quantification of biosurfactants are impractical and
`can be misleading in some instances. The correlation
`between surface/interfacial
`tension reduction and
`surfactant concentration holds for surfactant con-
`centrations below the CMC, at which the surface /
`interfacial tensions reach a minimum. However, at
`concentrations above the CMC,
`the reduction in
`surface/interfacial tensions becomes negligible due
`to the association of excess surfactant molecules into
`supramolecular structures such as micelles, making
`the estimation of surfactant concentration impossible
`without serial dilutions. Although the results of
`tension measurements of the serially diluted solu-
`tions can provide a rough estimation about how
`much higher the surfactant concentration is than the
`CMC, exact biosurfactant concentrations can not be
`quantified without information about the value of the
`CMC. For example,
`it has been reported that a
`Rhodococcus aurantiacus strain produced a glyco-
`lipid biosurfactant at a concentration as high as
`4003CMC [16]. However, the exact concentration
`of the glycolipid biosurfactant in the fermentation
`broth was still unknown because the CMC of the
`glycolipid biosurfactant has not been determined.
`The employment of surface/interfacial
`tension
`measurement for medium optimization studies is also
`
`tensions of surfactant
`that
`hindered by the fact
`solutions are also strongly affected by many parame-
`ters such as pH and ionic strength frequently investi-
`gated in medium optimization studies. Therefore,
`even at concentrations below its CMC, the extent of
`surface tension reduction does not always correspond
`to the level of biosurfactant
`in the fermentation
`broth.
`Another technique frequently employed for the
`characterization and quantification of biosurfactants
`has been thin-layer chromatography (TLC) [18,19].
`Although TLC analysis can provide qualitative and
`quantitative information about biosurfactants, time-
`consuming pre-purification procedures, such as pre-
`cipitation and organic extraction, are generally re-
`quired.
`Compared to tension measurements and TLC
`analysis, high-performance liquid chromatography
`(HPLC)
`represents an effective alternative for
`biosurfactant analysis with the desired sensitivity and
`selectivity. HPLC methods for quantitative analysis
`and/or
`for
`the purification of some lipopeptide
`biosurfactants have been reported [20–25]. However,
`the development of these HPLC methods generally
`required relatively pure biosurfactant samples, which
`cannot be obtained without
`tedious isolation and
`purification operations including HPLC. It is, there-
`fore, necessary to explore a general approach for the
`development of efficient HPLC methods for biosur-
`factant analysis and purification.
`In this study, a general approach incorporating
`ultrafiltration analysis was proposed for the develop-
`ment of HPLC analysis for biosurfactants without
`any prior structural or physicochemical information
`about the biosurfactants. The development of HPLC
`analysis for surfactin, a lipopeptide biosurfactant
`produced by Bacillus subtilis, was reported to dem-
`onstrate the feasibility of this approach. Neverthe-
`less,
`the proposed approach can be used for the
`development of HPLC analysis for practically any
`microbial surfactants. The techniques used are also
`useful for the preparation of homogeneous biosurfac-
`tant samples required for determining the CMC and
`for performing chemical analysis, such as Fourier
`transformation infrared (FT-IR) analysis and nu-
`clear magnetic resonance (NMR) spectroscopy, for
`the elucidation of chemical structures of biosurfact-
`ants.
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 2 of 11
`
`

`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`151
`
`2. Experimental
`
`2.1. Microorganism and growth conditions
`
`Bacillus subtilis ATCC 21332 (American Type
`Culture Collection, Rockville, MD, USA) was grown
`in a mineral salt medium supplemented with 4%
`glucose (Sigma, St. Louis, MO, USA) [26] at 308C
`for 48 h. Cells were removed from the fermentation
`broth by centrifugation at 12 000 g for 10 min.
`
`2.2. Ultrafiltration
`
`Solutions of sodium dodecyl sulfate (SDS, Sigma),
`2% (w/v), hexadecyltrimethylammonium bromide
`(CTAB, Sigma), 0.5% (w/v), lysozyme (Sigma), 0.1
`mg/ml, in 10 mM potassium phosphate, pH 6.0 and
`cell-free broth of B. subtilis were concentrated from
`10 ml to 2 ml by ultrafiltration with an Amicon
`magnetically stirred ultrafiltration cell (Beverly, MA,
`USA) assembled with ultrafiltration membranes of
`molecular mass cut offs (MWCOs) ranging from 500
`to 100 000 at operation pressure in the ranges of
`4
`5
`7?10
`to 2?10 Pa. Feeds and filtrates from all
`ultrafiltration runs were collected for HPLC analysis.
`In some experiments, methanol was added into the
`concentrate obtained by ultrafiltration to a final
`concentration ranging from 10 to 60% (v/v) before
`further concentrations were conducted.
`
`2.3. HPLC analysis
`
`All analytical experiments were performed by
`reversed-phase HPLC with a Jasco HPLC system
`(Tokyo,
`Japan)
`equipped with a C
`column
`18
`(Bondclone, 5 mm, 30033.9 mm, Phenomenex,
`Torrance, CA, USA). The mobile phase used for the
`analysis of SDS was a mixture of methanol–10 mM
`potassium phosphate buffer at pH 6.0 (75:25, v/v) at
`a flow-rate of 1 ml/min; the mobile phase used for
`the analysis of CTAB was a mixture of acetonitrile–
`water both containing 0.1% trifluoroacetic acid
`(TFA) (70:30, v/ v) at a flow-rate of 1 ml/min. The
`elution for SDS and CTAB analysis was monitored
`with a refractive index (RI) detector. The solvent
`system used for
`the analysis of
`lysozyme were
`mobile phase A (0.1% TFA in acetonitrile) and
`mobile phase B (0.1% TFA in water). The elution
`
`was conducted with a linear gradient from 40 to 45%
`A within 15 min at a flow-rate of 1 ml/min and
`monitored with an UV detector at 280 nm. Cell-free
`fermentation broth of B. subtilis and surfactin solu-
`tion were analyzed with a mobile phase consisting of
`a mixture of methanol, mobile phase A, and 10 mM
`potassium phosphate buffer at pH 6.0, mobile phase
`B. The elution was conducted with a linear gradient
`from 60 to 75% A within 35 min at a flow-rate of 1
`ml/min at 508C and monitored with UV at 210 nm.
`All HPLC experiments were repeated at least twice.
`For each assay, a sample of 10 ml was injected.
`
`3. Results and discussion
`
`Surface active molecules at concentrations above
`its CMC tend to aggregate spontaneously into sup-
`ramolecular micelles. The nominal molecular diame-
`ters of these surfactant micelles can be up to two-
`orders of magnitudes higher than that of the un-
`associated molecules.
`It
`is therefore possible to
`concentrate biosurfactants from fermentation broth
`by ultrafiltration with high MWCO membranes
`[27,28]. To demonstrate the feasibility of ultrafiltra-
`tion for the recovery of biosurfactant from fermen-
`tation broth, the recovery of SDS, an anionic surfac-
`tant, and CTAB, a cationic surfactant, from aqueous
`solutions by ultrafiltration with various MWCO
`membranes were conducted. The HPLC chromato-
`grams of the SDS solution and the filtrate collected
`from an ultrafiltration run with a MWCO 3000
`membrane are shown in Fig. 1a and b, respectively.
`The peak corresponding to SDS, eluted at 6.3 min,
`was only observed on chromatogram of the SDS
`solution, Fig. 1a, but not on the chromatogram of the
`filtrate, indicating that most of the SDS molecules
`associated into micelles with high nominal molecular
`diameters and thus were effectively concentrated in
`the retentate. Only trace amount of unassociated SDS
`molecules were detected in the filtrate. Similar
`results were also observed for CTAB. The HPLC
`chromatograms of the CTAB solution and the filtrate
`collected from ultrafiltration run with the MWCO
`3000 membrane are shown in Fig. 2a and b, respec-
`tively. The peak corresponding to CTAB, eluted at
`10.1 min, was only observed on chromatogram of the
`CTAB solution, Fig. 2a, but not on the chromato-
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 3 of 11
`
`

`
`152
`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`Fig. 1. Chromatograms of a 2% SDS solution (a) and of the filtrate collected from the concentration of SDS solution by ultrafiltration with a
`MWCO 3000 membrane (b). Ten ml of SDS solution was injected for each assay. The peak eluted at 6.3 min was identified as SDS. The
`peak eluted between 3 and 5 min was buffer front.
`
`gram of the filtrate. The losses of unassociated SDS
`and CTAB into the filtrate, defined as [surfactant
`concentration in the filtrate]/[surfactant concentra-
`tion in the feed], during ultrafiltration with various
`MWCO membranes are shown in Fig. 3. The loss of
`SDS during concentration by ultrafiltration with a
`membrane of MWCO as high as 10 000 was only
`8.19%, although the molecular mass of SDS is only
`288.4. The losses of SDS into the filtrates increased
`significantly to 14.13% with a MWCO 30 000
`membrane and to 73.92% with a MWCO 50 000
`membrane. These results indicates that the nominal
`molecular masses of most SDS micelles are between
`30 000 and 50 000, about two-orders of magnitude
`higher than that of the unassociated molecules. The
`
`losses of CTAB, a cationic surfactant with a molecu-
`lar mass of 364.5, were significantly lower than that
`of SDS, presumably due to the electrostatic repulsion
`between the charges on surfactant molecules and the
`charges on membrane surface. No significantly loss-
`es of CTAB were observed with membranes of
`MWCO below 30 000. The losses of CTAB into the
`filtrate with MWCO 50 000 and 100 000 membranes
`were 11.4 and 54.35%, respectively.
`is possible to
`Based on these observations,
`it
`identify peaks corresponding to biosurfactants with-
`out any prior structural or physicochemical infor-
`mation about
`the biosurfactant by comparing the
`chromatograms of the fermentation broth and of the
`filtrate from ultrafiltration experiments. However, it
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 4 of 11
`
`

`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`153
`
`Fig. 2. Chromatograms of a 0.5% CTAB solution (a) and of the filtrate collected from the concentration of CTAB solution by ultrafiltration
`with a MWCO 3000 membrane (b). Ten ml of CTAB solution was injected for each assay. The peak eluted at 9.1 min was identified as
`CTAB. The peak eluted between 3 and 5 min was buffer front.
`
`the peaks
`that some of
`should be pointed out
`observed only on the chromatogram of the fermen-
`tation broth but not on that of the filtrate may
`correspond to biological macromolecules, such as
`extracellular proteins or polysaccharides, other than
`biosurfactants. It is therefore necessary to develop a
`technique capable of differentiating peaks of biosur-
`
`factants and other macromolecular contaminants.
`Alcohols and acetone are capable of dissociating
`surfactant micelles into free molecules. The disso-
`ciated surfactant molecules with molecular diameters
`well below the MWCO of the ultrafiltration mem-
`brane employed for surfactant concentration are free
`to permeate the membrane and can therefore be
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 5 of 11
`
`

`
`154
`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`collected in the filtrate. The chromatograms of SDS
`solution containing 50% methanol and of the ultrafil-
`tration filtrate of the solution are shown in Fig. 4a
`and b. The peak corresponding to SDS was observed
`on both chromatograms, indicating that most SDS
`micelles were dissociated into free molecules by
`50% methanol and thus were collected in the filtrate.
`Similar phenomena were also observed for CTAB,
`Fig. 5. An extra peak eluted around 4.5 min corre-
`sponding to methanol was observed in Fig. 5. The
`presence of methanol in the samples also led to the
`small shift in CTAB retention time. The effective-
`ness of methanol in dissociating SDS and CTAB
`micelles was shown in Fig. 6. The degree of micelle
`dissociation was defined as [surfactant concentration
`in the filtrate]/[surfactant concentration in the surfac-
`
`Fig. 3. Losses of unassociated SDS (d) and CTAB (j) molecules
`into the filtrates during ultrafiltration with membranes of MWCO
`ranging from 500 to 100 000. The degree of surfactant loss during
`ultrafiltration was defined as [surfactant concentration in the
`filtrate]/[surfactant concentration in the feed]3100%.
`
`Fig. 4. Chromatograms of a 5% SDS solution containing 50% (v/v) methanol (a) and of the filtrate of the SDS–methanol solution (b). Ten
`ml of solution was injected for each assay. SDS peak, eluted at 6.3 min, was observed on both chromatograms.
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 6 of 11
`
`

`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`155
`
`Fig. 5. Chromatograms of a 1% CTAB solution containing 50% (v/v) methanol (a) and of the filtrate of the CTAB–methanol solution (b).
`Ten ml of solution was injected for each assay. CTAB peak, eluted at 10.1 min, was observed on both chromatograms. The peak eluted
`around 4.5 min was methanol which also led to a slight shift in retention time.
`
`tant–methanol solution]3100%. 75.6% of SDS mi-
`celles and 46.2% of CTAB micelles were dissociated
`with 50% of methanol. In the presence of 60%
`methanol, more than 90% of SDS micelles and
`CTAB micelles were dissociated.
`To confirm that the nominal molecular diameter of
`other extracellular macromolecules will not be sig-
`
`nificantly altered by the presence of methanol,
`similar ultrafiltration experiments with lysozyme, a
`protein with a molecular mass of about 14 000, were
`conducted. The chromatograms of lysozyme solu-
`tion, filtrate from ultrafiltration with a MWCO
`10 000 membrane with and without methanol (50%)
`were shown in Fig. 7. The peak eluted at 8.6 min
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 7 of 11
`
`

`
`156
`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`restricted. Similar
`through the membrane is not
`behavior was also observed for poly(ethylene glycol)
`(PEG) 6000 with a MWCO 3000 membrane (data
`not shown).
`These results indicate that surfactant micelles
`concentrated in the retentate can be separated from
`other macromolecules such as proteins by the addi-
`tion of appropriate amount of methanol. Although
`the permeation behavior of polysaccharides was not
`studied in this report, the ultrafiltration and HPLC
`experiments with PEG suggested that it may also be
`possible to separate dissociated surfactant molecules
`from extracellular polysaccharides by the addition of
`methanol. Therefore, it is proposed that biosurfactant
`peaks for any cell-free fermentation broth can be
`identified by comparing the chromatograms of the
`fermentation broth, of the ultrafiltration filtrate, and
`of the ultrafiltration filtrate obtained from biosurfac-
`tant
`solution containing appropriate amount of
`methanol. The peaks disappear from the chromato-
`gram of the ultrafiltration filtrate and reappear on the
`chromatogram of the filtrate with methanol can be
`identified as biosurfactants; the peaks observed only
`on the chromatogram of the fermentation broth but
`not on the chromatograms of the filtrates with or
`without methanol can be identified as macromole-
`cules. The feasibility of this approach was tested for
`the identification of biosurfactants produced by B.
`subtilis.
`B. subtilis has been shown to be effective for the
`production of surfactin, a highly active anionic
`lipopeptide biosurfactant [29]. The chromatograms
`of B. subtilis cell-free fermentation broth and of the
`ultrafiltration filtrate of the fermentation broth with a
`MWCO 10 000 membrane were shown in Fig. 8.
`The peaks eluted between 18 and 31 min on the
`chromatogram of the broth, Fig. 8a, were not ob-
`served on the chromatogram of the ultrafiltration
`filtrate, Fig. 8b, indicating that these peaks corres-
`ponded to surfactin micelles and/ or macromolecules
`retained in the concentrate of ultrafiltration. To
`further confirm the identities of these peaks, metha-
`nol was added to the ultrafiltration concentrate to a
`final concentration of 50%, and the resultant solution
`was further concentrated by ultrafiltration. The chro-
`matogram of the filtrate was shown in Fig. 9a. The
`peaks eluted between 18 and 31 min were observed
`again in the chromatogram,
`indicating that
`these
`
`Fig. 6. Effects of methanol concentration on the degrees of SDS
`(d) and CTAB (j) micelle dissociation, defined as [surfactant
`concentration in the filtrate] /[surfactant concentration in the
`surfactant–methanol solution]3100%. A MWCO 3000 membrane
`was employed.
`
`was identified as lysozyme. Unlike SDS or CTAB, a
`lysozyme peak was not observed on the chromato-
`grams for the filtrate with or without methanol,
`indicating that
`lysozyme can be concentrated by
`ultrafiltration and that the presence of 50% methanol
`does not change the nominal molecular diameter of
`lysozyme to such an extent
`that
`its permeation
`
`Fig. 7. Chromatograms of a lysozyme solution (0.1 mg/ml), of the
`filtrate of
`the lysozyme solution, and of
`the filtrate of
`the
`lysozyme–methanol solution. A MWCO 10 000 membrane was
`employed. Ten ml of solution was injected.
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 8 of 11
`
`

`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`157
`
`Fig. 8. Chromatograms of the cell-free B.subtilis fermentation broth (a) and of the filtrate of the fermentation broth (b). A MWCO 10 000
`membrane was employed. Ten ml of solution was injected for each assay.
`
`peaks corresponded to micelle-forming molecules,
`surfactin, instead of other extracellular macromole-
`cules, because as shown in Fig. 7 the peak corre-
`sponding to macromolecules such as proteins should
`not be observed on the chromatogram of the filtrate
`with methanol. The chromatogram of surfactin stan-
`dard from Sigma was shown in Fig. 9b. It can be
`observed that the peaks eluted between 18 and 31
`
`min in Fig. 9a and b were essentially identical,
`further confirming that these peaks corresponded to
`surfactin.
`The presence of more than one surfactant peaks on
`the chromatograms for surfactin standard was re-
`sulted from the existence of several surfactin struc-
`tures produced by B. subtilis. Like most secondary
`metabolites, surfactin consists of a family of lipopep-
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 9 of 11
`
`

`
`158
`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`Fig. 9. Chromatograms of the filtrate of the concentrated fermentation broth containing 50% methanol (a) and of the surfactin standard, 500
`mg/l, (b). A MWCO 10 000 membrane was employed. Ten ml of solution was injected for each assay.
`
`tides with similar chemical structures. So far at least
`nine different surfactin structures has been identified
`[24,27,30]. The relative areas of the surfactin struc-
`tures eluted between 18 and 23 min of the fermen-
`tation broth, Fig. 9a, are slightly different from those
`of the standard from Sigma, Fig. 9b. The difference
`
`in relative areas were resulting from the presence of
`different surfactin compositions. It has been reported
`that
`surfactin molecules of distinctive chemical
`structures are produced at different concentrations
`and ratios under different fermentation conditions
`such as medium formulation [30].
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 10 of 11
`
`

`
`S.-C. Lin et al. / J. Chromatogr. A 825(1998)149–159
`
`159
`
`4. Conclusions
`
`References
`
`Surface-active compounds at concentration above
`the CMC tend to associate into supramolecular
`structures such as micelles, which can be concen-
`trated effectively by ultrafiltration with high MWCO
`membranes. Surfactant micelles, formed under the
`influence of hydrophobic interaction, can be easily
`dissociated into unassociated molecules by the addi-
`tion of appropriate amount of alcohol. Unlike surfac-
`tant micelles with significantly higher nominal mo-
`lecular diameters, the unassociated surfactant mole-
`cules can permeate high MWCO membranes freely.
`Therefore, by comparing the chromatogram of the
`cell-free fermentation broth, such as Fig. 8a,
`the
`chromatogram of the ultrafiltration filtrate, such as
`Fig. 8b, and the chromatogram of the ultrafiltration
`filtrate of methanol–surfactant mixture, such as Fig.
`9a, the peaks corresponding to that of the biosurfac-
`tants can be identified without any prior structural
`information of the biosurfactants. In general, it can
`be assumed that
`the peaks observed only on the
`chromatogram of the filtrate of methanol–surfactant
`mixture but not on the chromatogram of the filtrate
`are biosurfactant peaks.
`This approach can be applied for the development
`of HPLC methods for any biosurfactants as long as
`the concentration of biosurfactant in the fermentation
`broth is higher than its CMC. If the concentration of
`the biosurfactant
`is below the CMC, pre-concen-
`tration by evaporation or freeze–drying is necessary
`to increase the biosurfactant concentration to above
`the CMC. This approach can also be used for the
`preparation of homogeneous biosurfactant samples
`necessary for performing chemical analyses such as
`NMR, FT-IR and fast atom bombardment (FAB) MS
`useful for the elucidation of chemical structure and
`for determining the physical properties of biosurfac-
`tants such as CMC [31].
`
`Acknowledgements
`
`This work was supported by a grant NSC 87-
`2214-E-005-007 from the National Science Council,
`Taiwan.
`
`[1] D.G. Cooper, J.E. Zajic, Adv. Appl. Microbiol. 26 (1980)
`229.
`[2] W.R. Finnerty, M.E. Singer, Dev. Ind. Microbiol. 25 (1984)
`31.
`[3] M. Parkinson, Biotech. Adv. 3 (1985) 65.
`[4] D.G. Cooper, Microbiol. Sci. 3 (1986) 145.
`[5] D. Haferburg, R. Hommel, R. Claus, H.P. Kleber, Adv.
`Biochem. Eng. Biotechnol. 33 (1986) 54.
`[6] G. Georgiou, S.C. Lin, M.M. Sharma, Bio/Technology 10
`(1992) 60.
`[7] S.C. Lin, J. Chem. Tech. Biotechnol. 66 (1996) 109.
`[8] S. Harvey, J. Elashvili, J.J. Valdes, D. Kamely, A.M.
`Chakrabarty, Bio/Technology 8 (1990) 228.
`[9] J.G. Leahy, R.R. Colwell, Microbiol. Rev. 54 (1990) 305.
`[10] A. Oberbremer, R. Muller-Hurtig, F. Wagner, Appl. Mi-
`crobiol. Biotechnol. 32 (1990) 485.
`[11] M.I. Van Dyke, H. Lee, J.T. Trevors, Biotech. Adv. 9 (1991)
`241.
`[12] A. Fiechter, TIBTECH 10 (1992) 208.
`[13] J. Akit, D.g. Cooper, K.I. Manninen, J.E. Zajic, Curr.
`Microbiol. 6 (1981) 145.
`[14] C.R. MacDonald, D.G. Coopper, J.E. Zajic, Appl. Environ.
`Microbiol. 41 (1981) 117.
`[15] R. Hommel, O. Stuwer, W. Stuwer, D. Haferburg, H.P.
`Kleber, Appl. Microbiol. Biotechnol. 26 (1987) 199.
`[16] B. Ramsay, J. McCarthy, L. Guerra-Santos, O. Kaepelli, A.
`Feitcher, A. Margaritis, Can. J. Microbiol. 34 (1988) 1209.
`[17] A. Persson, E. Oseterberg, M. Dostalek, Appl. Microbiol.
`Biotechnol. 29 (1988) 1.
`[18] M. Javaheri, G.E. Jenneman, M.J. McInerney, R.M. Knapp,
`Appl. Environ. Microbiol. 50 (1985) 698.
`[19] M.J. McInerney, M. Javaheri, D.P. Nagle, J. Ind. Microbiol.
`5 (1990) 95.
`[20] K.
`Jenny, O. Kaeppeli, A. Fiechter, Appl. Microbiol.
`Biotechnol. 36 (1991) 5.
`[21] C. Ullrich, B. Kluge, Z. Palacz, J. Vater, Biochemistry 30
`(1991) 6503.
`[22] S. Horowitz, W.M. Griffin, J. Ind. Microbiol. 7 (1991) 45.
`[23] S.C. Lin, M.M. Sharma, G. Georgiou, Biotechnol. Prog. 9
`(1993) 138.
`[24] S.C. Lin, M.A. Minton, M.M. Sharma, G. Georhiou, Appl.
`Environ. Microbiol. 60 (1994) 31.
`[25] M.M. Yakimov, K.N. Timmis, V. Wray, H.L. Fredrickson,
`Appl. Environ. Microbiol. 61 (1995) 1706.
`[26] D.G. Cooper, C.R. MacDonald, S.J.B. Duff, N. Kosaric,
`Appl. Environ. Micorbiol. 42 (1981) 408.
`[27] C.N. Mulligan, B.F. Gibbs, J. Chem. Tech. Biotechnol. 47
`(1990) 23.
`[28] S.C. Lin, H.J. Jiang, Biotechnol. Tech. 11 (1997) 413.
`[29] K. Arima, A. Kakinuma, A. Tamura, Biochem. Biophys.
`Res. Commun. 31 (1968) 488.
`[30] P. Pepoux, G. Michel, Appl. Microbiol. Biotechnol. 36
`(1996) 515.
`[31] S.C. Lin, K.G. Lin, C.C. Lo, Y.M. Lin, Enzyme Microb.
`Technol. 23 (1988) 267.
`
`PETITIONERS
`
`EXHIBIT NO. 1015 Page 11 of 11