Final Report Summary - BIOSURFING (New-to-nature biosurfactants by metabolic engineering: production and application)
Executive Summary:
The European Union has a lead position in terms of volume and revenue of biosurfactants and is therefore investing in research on these bio-based molecules. Indeed, biosurfactants offer great opportunities, but the current lack of diversity limits their penetration in a broad range of applications; biosurfactants are applied, but only in specific niche products. Therefore a consortium of European academic and industrial partners were working together in the BIOSURFING project to develop and test new types of biosurfactants. By developing a wider range of biosurfactants and evaluating their use in various applications (cleaning, cosmetics, medical and nanoscience), the ultimate goal of the project is to get biosurfactants to the supermarket shelf in day-to-day products available for every EU-citizen.
The glycolipid synthetic pathway of the biosurfactant producing yeast Starmerella bombicola was engineered to introduce structural variation in the biosurfactant molecules. Hence, tailor-made glycolipids with altered physical and chemical properties were created. During the strain engineering process, several useful tools were developed to allow fast genetic modification and fast evaluation of the fitness of the obtained strains. These tools can accelerate the time-to-market process for novel bio-based molecules. Much effort was put in the scale-up and design of an industrial relevant downstream process for each novel biosurfactants. Indeed, due to the new-no-nature structures, processes completely different form the existing ones had to be designed. Thanks to feed-back from the application partners, knowledge about the required purity and specifications was generated and methods were developed to monitor and remove contaminants, resulting in products of high purity. Due to the availability of these high-quality products and detailed specifications, we expect companies to be less reluctant to evaluate these compounds in their specific applications in the future.
Intense application testing was obviously also performed by the BIOSURFING project partners. Interesting observations were made regarding the self-assembling behaviour of the different congeners and molecules were evaluated for cleaning applications. Yet, it turned out to be challenging to simply apply a substitution based approach. There is some potential in the use of certain compounds in specific medical and cosmetic applications requiring anti-bacterial activity, and a limited number of novel glycolipids is being evaluated as an anticancer agent.
Several actions were taken to promote the BiIOSURFING project and disseminate the objectives and the results. Academic partners published their accomplishments in peer-reviewed journals or are preparing manuscripts accordingly. Furthermore, several contributions to conferences were made, both national and international, and as a poster or oral presentation. Some academic partners also used the work conducted during the Biosurfing project to illustrate their lectures and master and PhD students were engaged on the practical work, in this way also integrating the Biosurfing project in the educational programmes.
The Biosurfing website and both brochures were very useful to these above mentioned dissemination activities, also for the industrial partners. Furthermore, a dedicated workshop on Biosurfactants was organised in June 2015 and several of the Biosurfing partners were given the opportunity to give a talk there, nicely making a statement regarding the importance of our work towards the mixed academic and industrial public.
In general, existing staff was engaged to conduct the project work, and in total 10 additional staff members were hired (from which 6 are woman). Some Gender Equality Actions were taken such as the implement an equal opportunity policy and the use of targets to achieve a gender balance in the workforce.
Project Context and Objectives:
This project aims to create new-to-nature and tailor-made biosurfactants through metabolic engineering of the unconventional yeast Starmerella bombicola (= new taxonomic correct name of Starmerella bombicola). Biosurfactants produced by fermentation offer a worthy alternative to traditional surfactants, which are typically derived from non-renewable petrochemical resources and may cause environmental problems due to their ecotoxicity and poor biodegradability. Despite the clear advantages of biosurfactants, their overall use is hampered by the lack of structural variation. This is in sharp contrast to chemically produced surfactants where one can introduce variation by simply changing the building blocks. Structural variation is essential as (bio)surfactants find application in a very broad range of sectors. This project aims to alleviate this fundamental limitation by developing a generic biotechnological production technology for glycolipid biosurfactants. This will in turn significantly broaden the range of commercial biosurfactants, satisfying the need for structural diversity in the market. It is expected that this technology will result in a breakthrough penetration of glycolipid biosurfactants in the overall surfactant market, in this way helping to build the bio-based economy.
The mean objectives of the BIOSURFING project can be divided in three large interacting parts: (1) strain development, (2) production and (3) application evaluation. More detailed, following objectives will be purchased:
- Development of a sophisticated molecular toolbox for the genetic engineering of S. bombicola
- Metabolic and genetic engineering of S. bombicola for the production of new-to-nature tailor-made biosurfactants
- Development of analytical methods for structure elucidation of the novel compounds
- Evaluation of the strains by proteome, metabolome and transcriptome analysis;
- Providing feed-back in order to overcome limitations or further improve the strains
- Assessment of wild-type catabolic processes
- Optimisation of the fermentation process and down-stream processing for each target molecule to obtain a robust and efficient production system
- Production of sufficient quantities of test material for the application tests
- Physicochemical evaluation as emulsifier/surfactant and comparison with market leaders in the field
- Evaluation of (modified) glycolipids in nanomaterial applications
- Evaluation of the biological properties of natural and tailor-made biosurfactants
- The efficient dissemination and exploitation of the project results
- Establishment of an efficient management scheme to ensure the fulfilment of the scientific and valorisation objectives of BIOSURFING project
To summarise the above objectives, the very efficient biosurfactant producing yeast S. bombicola will be metabolically engineered by applying the recently developed meganuclease technology, allowing to control all structural parts of the glycolipid biosurfactant molecule: fatty acid tail, sugar moiety, acetylation and lactonization. For each target molecule, a fermentation and downstream process will be developed and the molecules will be evaluated for various applications (cleaning and cosmetics, medics and nanoscience). The project thus covers the whole innovation chain from basic research to production and application development. To achieve this goal, a complementary consortium of academic and industrial partners has been formed that covers the whole range of required expertises.
Project Results:
CONFIDENTIAL: this part contains confidential information. Do not make public.
PART I: STRAIN DEVELOPMENT
Development of a sophisticated molecular tool-box
For the metabolic engineering of S. bombicola, molecular tools needed to be established. Targeted expression and easy marker re-use could render the genetic engineering of S. bombicola more efficient and turn S. bombicola into an easy to manipulate platform organism for the production of tailor-made biosurfactants. Within the framework of this project, CELLECTIS has developed three main genome engineering technologies for S. bombicola based on their meganuclease and TALEN® technology.
Site specific endonucleases recognizing large DNA sequences from 22 base pairs (meganucleases) and up to 34 bp (TALEN®) can be used for gene delivery and integration at the specific site they recognize, based on a general principle of DNA cutting by the enzyme and repair by the cell. Indeed, due to the long recognition sequence, a site specific endonuclease is highly specific and often only cuts once in a whole genome allowing targeted integration. Three site specific endonuclease based strategies has been developed during the project. Both existing and custom site specific endonucleases were produced and delivered.
(i) Landing pads for genomic integration of transgenes
The genome of S. bombicola has been scanned for meganuclease sites and regions beneficial for expression (non-coding region, high expression area, etc.) were further examined. The 14 best ones have been selected as so-called “landing pads” for meganuclease meditated modification (Table 1). For seven loci several targeted meganucleases have been developed (only one failed), and these were validated in yeast using SSA annealing assay (Epinat JC, 2014). The meganuclease showing the best activity on its target was delivered to UGent. Seven more targets were identified and the in silico meganucleases were developed offering a whole range of potential landing paths. UGent executed tests with one selected meganuclease not interfering with any coding region.
Table1: List of identified meganucleases that were produced and validated for transgenes landing pads integration.
Name Hit sequence Validation
BioS01 GCAAGCCATCGTACCACATCTCGA 0.7
BioS03 TTTTGCCACTGTACGCCATTTTGA 0.9
BioS04 TCTTGCTGAATTACGTGGGATCTA 0.5
BioS05 ACCCCCCGTCTTACCGTGGCACAT 0.8
BioS06 GTTTTATCCTTTACAAAGTTTTGA 0.9
BioS07 TCTCCCCCCTGTAACCTATTTCAT 0.5
NB: Validation level varies from 0 (no activity) to 1 (max activity), cut off for validation based on CELLECTIS know how on meganuclease was determined at 0.4.
(ii) Re-usable markers
In this subtask, a second set of site specific endonuclease were chosen (Table2). This set does not recognize any target in the S. bombicola genome and were used to develop scar-less excisable selection marker constructs that thus will be re-usable. These re-usable marker constructs rely on flanking a functional selection marker with unique site specific endonuclease recognition sites, the whole being inserted between tandem DNA repeats. Once inserted in the genome, excision of the marker gene will occur through high efficiency intra-molecular homologous recombination triggered by cutting (Perez et al. 2005).
Table2: List of Meganucleases and TALEN® and their corresponding target site for Re-usable markers project
Name Description Sequence
C1221 natural mega TCAAAACGTCGTACGACGTTTTGA
ADCY9.1 custom mega CCCAGATGTCGTACAGCAGCTTGG
CAPSN1 custom mega CAGGGCCGCGGTGCAGTGTCCGAC
CAPT1.1 TALEN TCCGGGAACCCAGAGCTcacagccacgatcttAGACCCGAGCCCACAGA
An antibody allowing the detection of I-Cre I meganuclease and corresponding protocol for western blot meganuclease detection were also delivered.
In addition to these meganucleases and TALEN®, new generation of site specific endonucleases developed at CELLECTIS, CompactTALEN® (Beurdeley et al. 2013) were proposed, designed, produced and delivered. These CompactTALEN® should allow an easier delivery in S. bombicola which was an issue in the project. The CompactTALEN® were designed to target two markers available in S. bombicola: URA3 and GFP. For each marker target two different CompactTALEN® were produced and delivered (Table3).
Table3 : List of CompactTALEN® and their corresponding target site for marker removal project:
Name targeted sequence strand
URA3cT1 CGCAGACATCGGTTCCACTGTTAAGGCCCAATATGCA 1
URA3cT2 GTAACTCAGAGGTAGTTCGCACATCCAAGCTTGCGCAAAG -1
GFPcT1 GGCGTTCAGTGCTTTGCCCGTTACCCCGACCACATGAA 1
GFPcT2 GCGATACTTTGGTAAACCGTATCGAGCTCAAGGGCATT 1
(iii) Metabolic gene platforms
Based on the target genes for metabolic engineering, a third set of targets was chosen within genes relevant to the biosynthetic or biochemical pathway. Custom-meganucleases and/or TALEN® were designed, engineered, produced and validated to target these specific genes (Table4). They were designed for their capacity to induce knock-outs, allelic substitutions (including libraries of candidate gene substitutes) or gene or promoter swaps in S. bombicola within the gene targeted.
Table4 : List of Meganucleases and TALEN® with their corresponding target site chosen for the Metabolic gene platform.
Name Gene of Interest Type Validation
BioS15 invertase cabom01g09430 Meganuclease /
BioS15 invertase cabom01g09430 TALEN® 0.91+/- 0.01
BioS16 Leu2 cabom02g11170 Meganuclease /
BioS16 Leu2 cabom02g11170 TALEN® 0.74 +/- 0.02
BioS17 Trp1 cabom02g00890 TALEN® 0.98 +/- 0.01
BioS18 Ade1 cabom03g11310 TALEN® 0.97 +/- 0.02
BioS23 Aro7 cabom03g15270 TALEN® 0.84 +/- 0.02
BioS19 CYP52M1 cabom02g13890 TALEN® 0.98 +/- 0.01
BioS20 GT1 cabom02g13880 TALEN® 0.99 +/- 0.02
BioS24 AT cabom02g13870 TALEN® 0.96 +/- 0.02
BioS21 GT2 cabom02g13850 TALEN® 0.93 +/- 0.03
BioS25 lactonase cabom02g01720 TALEN® 0.95 +/- 0.02
BioS22 MDR cabom02g13860 TALEN® 0.95 +/- 0.04
In addition to the site specific endonucleases delivered, CELLECTIS provided a construct expressing the single chain Trex2 that has been shown to stimulate targeted mutagenesis induced by meganucleases by one log factor (Delacôte et al. 2013) and TALEN®.
Strain development
Several strains either producing sophorolipids with an altered composition of the individual components or producing entirely new glycolipids were developed. This are listed and commented below:
1. A strain producing 100 % acidic sophorolipids: this one behaves very similar compared to the wild type regarding growth, nutrient requirement and production of sophorolipids: sophorolipids can be produced at rates and in quantities equal to the wild-type.
2. A strain producing almost 100 % lactonic sophorolipids. Again, this strain behaves very similar compared to the wild type regarding growth, nutrient requirement and production of sophorolipids: sophorolipids can be produced at rates and in quantities equal to the wild-type.
3. A strain producing cellobioselipids. Although the novel compounds are produced, this is in low amounts and combined with the synthesis of intermediate glucolipids. The strain evaluation procedure described below gives some input on possible reason for this less performing behaviour. These will be implemented in the follow-up of the project and should result in a strain with improved production characteristics.
4. A strain producing C16 sophorolipids: the novel compounds are produced at high purity (few other congeners and low amount of C18 sophorolipids), yet the yield is only 5 g/L maximum, which is quite low when compared to the wild type. Some samples can be provided, but only for applications requiring a low amount of material input for testing.
5. An engineered S. bombicola strain producing branched C22 sophorolipids is still under construction. Yet, in order to allow the applications partners to perform test with these types of molecules prior to the strain finalization, a non-Starmerella host was successfully used for the synthesis of these compounds.
6. A strain producing glucolipids was used to obtain glucolipids in a mixture of none- and acetylated congeners. Again, lower production titters compared to the wild type were noticed; yet, the strain evaluation procedure described below gives some input on possible reason for this less performing behaviour. These will be implemented in the follow-up of the project and should result in a strain with improved production characteristics.
Strain evaluation
As described above, genetically engineered strains of S. bombicola were created which produce new-to-nature sophorolipids (SL). However, some of those newly generated strains are not able to produced equal amounts of product as compare to the wilt type (WT) S. bombicola strain. The reasons for this lower production rate are not well understood. By analysing the proteome and metabolome of those strain and compare it with the WT, we aimed to understand this phenomenon. Proteomic approaches were used to analyse the amount of the sophorolipid biosynthetic cluster gene products: cytochrome P450, glycosyltransferase A1, glycosyltransferase B1, acetyltransferase, SL transporter. The first studied mutant was created by knock-out of the cytochrome P450 monooxygenase, responsible for the first step of SL biosynthesis (fatty acid hydroxylation), this newly generated strain (CYPKO) lost the ability to produce sophorolipids. We must stress that this strain is therefore not used for scale-up and new-to-nature sophorolipid production, but is a good negative control to be included in the strain evaluation. When the acetyltransferase gene participating in sophorolipid acetylation was knocked out, this mutant (ATKO) produced only non-acetylated sophorolipids. Additionally, when the glucosyltransferase (ugtA1), responsible for the attachment of the first glucose to the hydroxylated fatty acid, was replaced with ugt1 from Ustilago maydis, this S. bombicola mutant (CL) produced a new kind of the cellobiose lipids.
To characterize S. bombicola mutants producing new-to-nature molecules we developed multiple protocols to analyse their protein and metabolite level. Firstly we optimized a protocol to analyse WT and mutants using an isobaric labelling proteomic approach - iTRAQ. This experiment confirmed that in all mutants the amount of SL cluster proteins is different form the WT. As SL cluster proteins are underrepresented in the mutants this can have a direct connection with biosurfactant yield. It is possible that during genetic manipulation a promotor or regulatory region of the SL cluster was modified which causes reduced expression of this cluster. On the other hand, presence of sophorolipids could generate a positive feedback, which activates the whole SL pathway. Therefore, the suppression of the product of the pathway could lead to its down-regulation.
As the iTRAQ approach is time consuming and expensive, we also developed more targeted approach, which allows faster and easier analysis of cluster gene products in new S. bombicola mutants. We established a multi-reaction monitoring (MRM) assay. This method is based on the selection of unique peptides of each targeted protein that are then monitored in a triple quadrupole type analyser. For each SL cluster protein we select two unique peptides and we compared their intensity between the WT and mutants. With this approach we validated the results obtained with iTRAQ. We confirmed that in CYPKO and CL mutant SL cluster protein expression dropt around 50% in comparison to WT, while in the ATKO of about 20%. Recent analysis showed equal SL cluster protein presence for the strain producing 100% acidic sophorolipids. This is in line with the observed production volumes. Interestingly, the first glucosyltransferase was clearly more present in the mutant then in the wild type.
This MRM assay is a new toolbox, which can be used to study SL cluster protein abundance in new S. bombicola mutants.
We also performed a metabolite analysis on the WT and the three mutants described above. We focused on fatty acids (FA) and soluble polar metabolites.
Extracted FA were converted to methyl ester of fatty acids (FAME) via transesterification. FAMEs were analysed (GC-MS) and they were identified using the NIST MS 2.0. database. Additionally, identification was based on comparing FAMEs retention times with those of the standards used to generate the standard curve. The FAMEs quantification was based on the extracted ion chromatogram (XIC) of the m/Z 74 fragment ion. This ion is characteristic for all FAMEs, representing the methylated carboxyl group.
Regarding FA, 40 peaks were detected by gas chromatography analysis (GC-MS) for the WT and the three mutants (CYP, AT, CL) extracts, but only 30 of them could be identified. When the S. bombicola fatty acids profile is compared with other described yeast like Sacharomyces cerevisiae not many differences are observed. However we can clearly see that there is a difference between S. bombicola WT, a good producer of sophorolipids, and the CYPKO, a strain with no sophorolipid production. The variation is most clear in case of the most abundant fatty acids: C16:0, C16:1, C18:0, C18:1, which are building blocks in sophorolipids biosynthesis. It is the first time that we can prove that indeed an additional pool of fatty acids is necessary for sophorolipids generation. Similar conclusion can be drown from observation of other detected fatty acids like for example branched fatty acids and dicarboxylic acids. Likely their presence in the WT is important for SL production as those fatty acids can be substrates for hydroxylated fatty acids. Contradictory, the odd numbered fatty acids are more abundant in the CYPKO mutant than in WT. Possibly this indicates a difference in composition of the cells membranes which can also be crucial for sophorolipids secretion.
For the polar metabolite analysis we used a targeted approach where we concentrated on LC-MS analysis of 84 standard metabolites with known elution profile and MW. We identified and quantified 44 metabolites in the WT and the three analysed mutants. We could confirm that the CL strain, producing small amount of cellobiose lipids, does not perform well and we suggest that it is under some kind of stress. It can be related to the problems during genetic manipulation of this strain or due to the intracellular accumulation of the cellobiose lipids. We indeed found out that this strain was based on a spontaneous ura3-negative mutant, which now turned out to be less vital as compared to the general used PT36 controlled ura3-negative strain. These findings will be taken into consideration for further strain improvement. Surprisingly the difference in the level of metabolites in CYPKO and WT are not as extreme as with the CL strain. It would mean that S. bombicola does not suffer much from losing this secondary metabolite pathway. We found indeed a few differences which can be connected with the higher need of the WT for energy, cofactors, nitrogen and glucose. Metabolites profiles of the WT and ATKO strains are very similar. This fits to the SL production pattern, as the ATKO strain is the best producing mutant in this experimental setup, but still lower than the WT.
We believe that the results collected from proteomic and metabolomics analysis can be useful to improve molecular engineering strategies in order to create better performing S. bombicola mutants producing increased amount of new-to-nature biosurfactants. Developed protocols and the MRM assay can be used for monitoring and selection the best new S. bombicola mutants for industrial application.
Related to the stain evaluation, product catabolism by the strain was evaluated as well. There are no indications that catabolism is an issue for cultivation and production conditions applied. Hence, no further strain modification was required regarding this matter.
PART II: PRODUCTION
Strains that successfully passed the first stage of the project were taken to scale-up and industrial production, whit special attention for product purity.
Three Starmerella bombicola strains created by the University of Ghent could proceed to scale-up and were delivered to the Bio Base Europe Pilot Plant (BBEPP) for process development and scale-up. Also a Rhodotorula bogoriensis strain was delivered for the production of a special type of branched sophorolipids as the engineering of its biosynthesis in S. bombicola was not yet successful. For these four strains the production processes (fermentation and purification) were first developed at the lab scale (7L) and subsequently scaled up to the small pilot scale (150L) and the purified products were delivered to the partners. For two strains, these processes were further scaled up to the 4500 L scale for which a lot of useful input was already gathered during the process development at the lab and small pilot scale. Robust, scalable and efficient fermentation processes with defined process parameters were developed giving rise to several batches of end product with limited variation. The latter enabled setting up the specifications for each product while at the same time large ‘sample’ amounts of the products were generated, which could be used for dedicated application testing.
Four structurally different glycolipid products were produced at the large (>100 kg) or small (> 100 g) pilot scale. Because of their structural variation, the application potential of these molecules was expected to be completely different.
During the project a lot of attention was given to the productivity of the fermentation processes, yield maximisation of the purification processes and maximisation of the purity for the final products. These objectives were obtained by an iterative process, which for some of the products required multiple rounds. Feedback of the project partners on the sample purity was also taken into account to decrease the presence of (specific) contaminants such as salt, remaining sugars, proteins, triglycerides and fatty acids in the final product(s).
Five products were finally generated in sample sizes, which were large enough for several of the project partners to perform dedicated application research. These molecules were:
- Lactonic sophorolipids
- Acetylated acidic sophorolipids
- Non-acetylated acidic sophorolipids
- Glucolipids
- Acetylated branched acidic sophorolipids
An LCA analysis was performed and it was quite surprising to see that the impact of the production of these molecules was larger as expected, as they were not significantly better (nor worse) as the impact of most (fossil based) reference products. However, this was mainly due to the agricultural production processes for rapeseed oil and glucose.
Now that these processes have been developed up to a mature level and large sample sizes are available for application testing, the possibility of using waste or side streams for their production can be evaluated. The latter would have a tremendous impact on the impact of the production of these molecules.
PART III: APPLICATION TESTING
Self-assembling behaviour and nanomaterials
The novel glycolipids were studied in detail for their self-assembling behaviour. In their most-common configurations, surfactants are made of one polar headgroup, which is hydrophilic, and one hydrophobic, non-polar tail. This ambivalent chemical structure allows them to have affinities for both polar (e.g. water) and non-polar (e.g. oil) phases. Under these conditions, micelles, spherical aggregates, form thus reducing the interfacial tension between the aqueous solution and the non-polar layer. It is probably the most simple and ubiquitous arrangement formed by surfactant molecules, and it is usually formed above a critical concentration known as the critical micelle concentration (cmc).
Besides spherical micelles, surfactants can self-organize into other aggregates depending on the medium in which they are dispersed, and on the respective size of the hydrophilic and hydrophobic parts of the molecule: common topologies are cylindrical micelles, planar bilayers or inverse micelles, for instance. If the most common surfactants are composed of the association of hydrophilic and hydrophobic blocks in simple head-to-tail configurations, more complex associations are also found. This is the case in most bio-based surfactants, constituting the topic of the BIOSURFING Project. In this case, the properties in solution of these new molecules composed of bio-derived building blocks are mostly unknown, despite their high potential in cosmetics, food and pharmaceutical industries.
The goal of UPMC in WP5 was to study the complex self-assembly behaviour of sophorolipids and the new-to-nature derivatives.
The state of the art on the sophorolipid self-assembly before BIOSURFING was the following:
- 2004: the first study on self-assembly properties of acidic deacetylated C18:1 sophorolipids was published by Zhou et al. (Langmuir, 2004, 20, 7926). The authors have shown the formation of giant ribbons at room temperature in water at acidic pH for concentrations in the range few mg/mL.
- 2010: the same compound was studied by Baccile et al. (Green Chem., 2010, 12, 1564) under similar conditions and they found that at acidic pH only micelles could form.
- 2011: Penfold et al. (Langmuir, 2011, 27, 8867) did show also the formation of micellar aggregates in water under acidic-neutral conditions but in their case the compound was a mixture of acetylated and non-acetylated sophorolipids.
No other work on this topic was published before the beginning of the BIOSURFING Project in October 2011. Since then, our contribution was very important and it can be summarized hereafter by type of compound:
1. Acidic deacetylated sophorolipids (SL C18:1)
Acidic conditions
Basic conditions
Silica templating
Nanoparticle functionalization
2. Acidic acetylated sophorolipids (SLa C18:1)
3. Acidic deacetylayed glucolipids (GL)
4. In-situ mechanisms of formation
1. Acidic deacetylated sophorolipids (SL C18:1)
This compound can be considered as a kind of reference as it is the one with the highest production yield and the most studied in the literature. In the initial part of the BIOSURFING Project we spent a lot of energy to better study and understand the behavior of this compound, in order to settle the ground reference for the new-to-nature ones, which we only received later on. The achievements for this compound are summarized in the following points, to which the appropriate reference is provided for more detailed information.
Acidic conditions: At pH below neutrality we have found that SL C18:1 form micelles in water, thus confirming our early data of 2010. Furthermore, using Small Angle Neutron Scattering, we have also found that micelles are uncharged objects in solution below pH 5 while they acquire a negative charge in the 5 < pH < 7 range, due to the deprotonation of the free COOH group. These results have been published in 2012 and 2013: N. Baccile et al. Soft Matter, 2013, 9, 4911; N. Baccile et al., ACS Nano, 2012, 6, 4763–4776.
In a next study, (S. Manet et al., J. Phys. Chem. B, 2015) we provide a detailed description of the structure of the micelle formed by SL C18:1 by combining Small Angle X-ray Scattering using synchrotron radiation and Molecular Dynamics modelling. In particular, the micelle is not constituted by the homogeneous distribution of the sophorolipid, as one would expect for classical surfactants, but by a more complex distribution of the sophorose and COOH groups, both hydrophilic. This is a unique result obtained for a bolaform compound. Furthermore, the micelle is no strict sphere, but has a more ellipsoid form.
Basic conditions: At pH above neutrality we have found the coexistence of micellar aggregates and larger objects of ill-defined structure, as published in N. Baccile et al., ACS Nano, 2012, 6, 4763–4776. In a more recent work (A.-S. Cuvier et al., Chem. – Europ. J., 2015), in which we combined Small Angle X-ray Scattering using synchrotron radiation and cryogenic Transmission Electron Microscopy, we were finally able to better describe the nature of the aggregates: they are composed of nanoscale platelets. The self-assembly in this medium was never reported before for sophorolipids.
Silica templating: Micelles can be used as templates to cast inorganic materials and in particular, the fact of removing micelles after the casting process allows the formation of nanoporous solids, very much studied in various fields. We have been pioneers in the use of sophorolipids as micellar templates for silica already in 2010, in which we did show the feasibility of the process. However, during BIOSURFING, we have strengthened our position by publishing several articles in which we give more synthesis details about templating thin silica layers (N. Baccile et al., J. Phys. Chem. C 2013, 117, 23899) and, for the first time, porous silica particles (B. Thomas, ACS Sus. Chem. Eng. 2014, 2, 512). Both the sugar part and the carboxylic group are believed to interact with the amine functions linked to the silica.
Nanoparticle functionalization: The free-standing COOH group of sophorolipids can also be used as capping agent for nanoparticles. The group of B. L. V. Prasad (Pune, India) did report before 2011 several strategies to stabilize metal nanoparticles in water using sophorolipids. During BIOSURFING, we have used sophorolipids C18:1 to stabilize metal oxides for the first time, and in particular the superparamagnetic iron oxide nanoparticles (N. Baccile, Phys. Chem. Chem. Phys., 2013, 15, 1606). We could show different stabilization strategies, the control over the size distribution but also the stability towards high ionic strength aqueous media, which is the key feature to employ these materials in biomedical applications. By controlling the parameters, different size distributions could be obtained.
2. Acidic acetylated sophorolipids (SLa C18:1)
This compound was provided to us by the Bio Based Europe Pilot Plant with the T21 code. Its self-assembly properties have been studied by applying the typical protocols developed for the SL C18:1 compound. We used Small Angle X-ray Scattering using synchrotron radiation and cryo-transmission electron microscopy to analyse it. We did find that in the mild acidic pH region, ellipsoidally-shaped micelles are formed in water for 5 w% solutions. However, upon decreasing the pH to 2, micellar length strongly increases by at least a factor 3 and large aggregates form in solution, which becomes turbid. This behaviour is in contrast with the typical SL C18:1 sample. These data are published in the PhD manuscript of Anne-Sophie Cuvier and will be the topic of a future communication in a peer-reviewed journal. This compound cannot be studied at high pH values due to the chemical fragility of the acetyl ester function.
SLa have been used in an attempt to obtain porous silica powders, in which the microbial glycolipids are the porogenic agents. The use of TEOS as inorganic precursor drove the system towards the formation of gels; however, by tuning the synthesis conditions, we could better control the interactions between SLa and silica and obtain a powdery material. Under TEM, one can observe silica samples with a network of non-organized pores. This is confirmed by the N2 adsorption/desorption isotherms and the high surface area obtained (> 400 m2/g), values which are comparable with the mesoporous silica samples typically obtained with classical surfactants like block copolymers. Interestingly, we did not have to remove the glycolipids using a heavy and costly calcination step. Although this result is very interesting, it is still hard to explain and more work would be needed.
3. Acidic deacetylayed glucolipids (GL)
This compound was provided to us by the Bio Based Europe Pilot Plant and its self-assembly behaviour was studied using the same protocols employed in the above-mentioned studies. We have in particular looked at its pH-dependent behaviour in the 0.5 – 5 w% region combining cryo-TEM and SAXS using synchrotron radiation. Very differently than all other compounds, GL forms entangled elongated micelles at basis pH and vesicles at pH below 6. These results are the first ones produced on this compound and will constitute an upcoming publication.
We have also used glucolipids to stabilize magnetic iron oxide nanoparticles in a two-step procedure. In particular, we could show for the first time that the vesicle-forming properties of glucolipids at pH 6 can fully stabilize the iron oxide particle colloidal dispersion after going through pH 2. The strong pH-dependent supramolecular behaviour of these glucolipids (vesicles at pH 6 and lamellar phase at pH 2) seem to explain our results.
4. In-situ mechanism of formation
One of the most complex aspect of the self-assembly of bio-based glycolipids is to understand, and put in evidence, the mechanism of formation depending on pH. SL C18:1 and SLa C18:1 behave similarly, but not exactly in the same way. GL and SL C18:1 have a very different behaviour. In all cases, the differences between these molecules are very small. In order to better understand such differences, we have carried out a series of pH-dependent SAXS studies, unique in the literature for these compounds, at the ID02 line of ESRF synchrotron in Grenoble (France). These data are still under analysis but there are indications for the evolution between nanoscale platelets and micellar aggregates in water at 0.5 w%. Advanced modelling of these data, in comparison with the other bio-based glycolipids will hopefully give a better insight in terms of pH/structure self-assembly behaviour in water.
Cleaning applications
Ecover received 8 samples in total. These samples were tested on their basic properties relevant for the cleaning industry like physic-chemical parameters (surface tension and foam behaviour) and aquatic toxicity. The samples were compared with reference sophorolipids.
A first experiment was set up to check the water solubility of the molecules provided. A limited water solubility has implications for incorporating the compound into a formula. Secondly impurities in the samples, like residual oil, can have an influence on the subsequent tests. The lactonic sample displayed low solubility, hampering some of the applications.
The minimal surface tension of each surfactant was measured using a Krüss DSA100 tensiometer. From the results obtained it could be concluded that none of the Biosurfing sophorolipid surfactants was able to reach a lower minimal surface tension than the reference wild-type sophorolipid as used in commercial available cleaning formulations.
Foam behaviour of the surfactants was reviewed on two aspects: foam ability and foam stability. It is known for wild-type sophorolipids that they are limited to non-foaming applications due to their limited properties. It was found that the acidic sophorolipid (T21) sample had high foam ability and low foam stability in comparison with wild type sophorolipid mixtures. The branched acidic C22 sample had low foam stability, probably due to the residual oil in the sample. Also the foam stability of the glucolipid sample was less in comparison with a mixture of wild type sophorolipids.
The aquatic toxicity was measured using a daphtoxkit F. The surfactants T21 and wild type sophorolipids were found to be practically non-toxic. Branched acidic C22 and the glucolipid mix were tested in a cleaning formulation. The branched acidic C22 sophorolipid cleaning formulation was found to be moderately toxic and the glucolipid mix cleaning formulation was found to be toxic.
From the samples obtained during this project, none of them seemed to have an added value when applied in the conditions applicable for the traditional sophorolipids, in comparison with the sophorolipids compounds used in the company’s commercial available cleaning products.
Werner & Mertz tried to find an optimal application area for the newly developed glycolipids in comparison with commercially available surfactants. To this end, glycolipids were tested in various cleaning formulations, and performance and behaviour tests were conducted and compared. The novel glycolipids were implemented in four different products; each representing a specific product category:
Acidic Cleaner 1 (sanitary cleaner) pH = 1.7 - 1.9
Neutral Cleaner 1 (manual dishwashing liquid) pH = 5.5 - 5.7
Neutral Cleaner 2 (floor cleaner) pH = 7.4 - 8.0
Alkaline Cleaner 1 (degreaser) pH = 9.6 - 9.9
For an optimal comparison between the new glycolipids and other surfactants, the nonionic surfactant in every cleaner was varied. The new glycolipids were checked against the following alkyl polyglucoside and sophorolipid:
Lauryl glucoside [1]
(Dodecyl glucoside, CAS: 59122-55-3, CH3(CH2)nO[C6H10O5]mH, n=11-14, m=1.4-1.6)
Standard Sophorolipid “Sophoclean” by company “Soliance” (for the first reporting period)
The following samples of new glycolipids were provided by “Bio Base Europe Pilot Plant VZW” and examined:
1. Sophorolipid T21 Acidic (KO)
2. Sophorolipid T23 C1 Lactonic (OE)
3. Glucolipid (GL01)
Cleaning Performance of the acidic cleaner 1 was examined by monitoring of lime soap removal and lime scale removal. Dynamic cleaning performance of the neutral cleaner 2 and alkaline cleaner 1 was tested with a burnt fat-dust soiling assay. Their self-acting cleaning performance was evaluated in a similar way, but without applying mechanical power. Furthermore, general properties such as emulsification, foaming and wetting ability, dynamic surface tension, stability and viscosity were determined as well. The findings for each product category are given below:
Acidic Cleaner 1
Glucolipid was not soluble and sophorolipid Acidic T 21 was not stable in the Acidic Cleaner formulation, so they cannot be recommended for the acidic system.
Although the remaining stable sophorolipid Lactonic T23 C1 sample showed a slightly better lime scale removal power, the lime soap removal ability was considerably weaker than the conventionally produced lauryl glucoside. Lauryl glucoside decreased the surface tension significantly more than Lactonic T23 C1 cleaner solution. The variation of the glycosides had no impact on the foaming ability.
Neutral Cleaner 1
The loss of emulsifying power of the three biosurfactants from the project in comparison to lauryl glucoside was too large to be an appropriate alterative in manual dishwashing agents.
Sophoclean that is produced by the untreated wild-type had the best foaming performance; the other samples had the same lower foamability. Lauryl glucoside and glucolipid had nearly the same increasing influence on the viscosity in the used salt thickening system. The sophorolipids had no effect on the viscosity. The use of the newly developed biosurfactants in manual dishwashing detergent formulas like Neutral Cleaner 1 can only be recommended if a loss of performance can be accepted.
Neutral Cleaner 2
Sophorolipid Acidic T21 and the Sophoclean standard were not stable in the Neutral Cleaner 2 formulation.
The dynamic cleaning performance concerning removal of a fat-dust soiling of the glucolipid is much and the performance of sophorolipid Lactonic T23 C1 is slightly weaker than the conventionally produced lauryl glucoside.
Self-acting cleaning performances of Neutral Cleaner 2 with lauryl glucoside and Lactonic T23 C1 were at the same level, but glucolipid had nearly no cleaning impact.
The wetting ability of the two remaining Biosurfing samples was not sufficient.
The variation of the non-ionic surfactant had no effect on the foaming ability of the cleaners.
The achieved cleaning abilities were not sufficient for the application area of floor cleaners. We cannot recommend using one of the tested samples in a similarly constructed floor cleaner.
Alkaline Cleaner 1
Besides the standard lauryl glucoside, only the glucolipid sample was stable in the pH-area of the Alkaline Cleaner 1.
The cleaning ability concerning dynamic removal of a fat-dust soiling of glucolipid was on a high level. Lauryl glucoside could not reach the performance of the biosurfactant.
Glucolipid is able to increase the quality of foam built by a foam spray trigger, but unfortunately glucolipid showed nearly no self-acting cleaning performance and a poor wetting ability.
To conclude, one can say that in none of the tested cleaning product categories the results were satisfying. This without exceptions, so a clear recommendation for one of the application areas cannot be stated. Although results in single performance tests fulfil the requirements, other results in the same type of product were not convincing enough.
Sophorolipids produced by the untreated wild-type of the used yeast have nearly the same properties as the new samples. We recommend continuing metabolic engineering on Starmerella bombicola to generate further biosurfactant compositions. Particular products based on a shorter fatty acid chain are expected to show the desired properties.
Applications in cosmetics
In the cosmetics industry there is a current trend to replace conventional cosmetic ingredients with naturally derived alternatives. In this view, sophorolipids T18C3, T21, T22, T23C1, T23C3, T37, T43 have been provided by partner UGent to partner COSMETIC to be tested. Based on the sophorolipids structure, various potential roles have been envisaged in cosmetic applications, such as emulsifiers, detergents, wetting agents, solubilisers or anti-microbial actives.
Within the frame of the investigation of the sophorolipid performance in such potential applications, a number of relevant physicochemical parameters have been initially calculated. We have calculated Cc and HLB values for all sophorolipids provided. Both these parameters are typically being considered in the state-of-the-art cosmetic science to guide the selection of appropriate emulsifiers to formulate a stable emulsion. HLB values calculated also gave us indications of different potential applications of the sophorolipids in the cosmetic industry.
We have studied the use of sophorolipids as solubilisers. Solubilisers are commonly used in cosmetic products to emulsify the fragrances/oils at molecular level and yield transparent solutions. T21 (the less purified acidic sophorolipid) performed well at ratios higher than 3:1 (solubiliser:oil) whereas the T37 (the acidic sophorolipid of high purity) was average and at ratios higher than 5:1. A possible explanation might be that the T21 contains some other impurity such as mixture of acidic sophorolipids acetylated/non-acetylated or lipid compounds that exhibit synergy with the acidic sophorolipid. T21 has the potential to be used as natural solubiliser.
Sophorolipids did not performed well as emulsifiers either at room temperature or at 80 °C. Even when using a high shear homogeniser the emulsions produced where not stable. Given that T21 performed well as solubiliser maybe it can create stable emulsions but at a much higher concentration than traditionally used emulsifiers.
Finally, our results demonstrate that the provided rhamnolipid can have applications in cosmetics where medium to low foaming is desirable similar to T21 as reported in deliverable 5.5.
Guided by the calculated HLB values for the provided sophorolipids (as reported in Deliverable 5.5) we have evaluated the potential use of the most promising candidate (T23C3- a diacetylated lactonic sophorolipid), as preservative booster in cosmetic formulations. Our results indicated that the provided preparations of lactonic diacetylated sophorolipids cannot have applications in cosmetics as preservative booster, at least at the concentration range investigated.
Another aspect studied was the antimicrobial activity. Our results confirmed literature reports on the targeted bacteriocidal activity of sophorolactone only on some Gram positive bacteria (Corynebacterium xerosis, Propionibacterium acnes).
When correcting for lactone content, purified lactone T43 was found as potent as commercial Sopholiance S against C. xerosis (the microorganism responsible for malodour under the armpit), when growth was assessed at 20h after inoculation. The MIC value we have determined for Sopholiance S (0.025 % w/w) is comparable to literature reports (0.05 % w/w without correction for lactone content, i.e. 0.03 % when corrected for a 60 % lactone content). Growth inhibition was retained longer in the case of T43, as opposed to Sophliance S. MBC values were also found comparable.
Regarding P. acnes (acne causing bacteria), our preliminary results do indicate a superior bacteriostatic activity of Biosurfing lactone T43 at 20h after inoculation, which is however not apparent, already at 27h of incubation. The lactone lack of solubility did not allow us test higher sophorolipid concentrations. Our results demonstrate in any case, a roughly comparable bacteriostatic potency.
A competitive advantage of Sophoroloactone preparation as dry powder (as in the case of Biosurfing T43) is that it is much more stable to hydrolysis, i.e. has a longer shelf-life when compared to the liquid preparation of Sopholiance S. Indeed, while performing our experiments we did notice a significant degradation of the lactone (measured by HPLC) to a value which is half the reported lactone concentration by the manufacturer! Also, when not taking into account the lactone content, Biosurfing T43 is a superior antimicrobial compared to Sopholiance S, which could also add a competitive, commercial advantage to T43.
The major comparative weakness however of T43 is that it is a GMO product, as opposed to Sopholiance S. Sopholiance S is approved by ECOCERT and COSMOS. Such designations could not be attributed to T43, due to its GMO nature, compromising its commercial value.
An acidic sophorolipid fraction T21 exhibited good solubilizing properties of perfumes and oils. Sophorolactones had good antimicrobial properties against selected gram positive bacteria such as Corynebacterium xerosis one microorganism responsible for the malodour of the armpit and Propionibacterium acnes the acne-causing bacteria.
Biological activity
There have been numerous preliminary reports stating that biosurfactants are suitable alternatives to some synthetic medicines and antimicrobial agents and may be used as safe and effective adjuvant or therapeutic agents. Some of these biosurfactants were described for their potential as biologically active compounds and applications in the medical field. Some of the lipopeptides biosurfactants (iturin and surfactin) have been implicated with high surface activities and antibiotic potential. Such compounds enhance or decrease the bacterial surface hydrophobicity an aspect which appears to be essential in relation to their effect on microbial pathogenicity and or susceptibility to antibiotics. In addition their low cytotoxicity to mammalian cells makes them suitable for use in many medical and microbial attack applications. Other glycolipids such as sophorolipids of Starmerella bombicola and rhamnolipids of Pseudomonas sp. have also been shown to have some antimicrobial activity particularly against gram-positive bacteria and varying levels of effects on the pathogenicity of some fungal and bacterial strains.
Biosurfactants are ideal compounds to increase O2 saturation to the surface of healing wounds, while providing a natural barrier to infection and aiding the migration of cells important in the healing response. To date there are few studies investigating the potential of these compounds to accelerate wound healing in vivo. Furthermore, there are some reports on the anticancer activity of natural and modified glycolipids.
To test such biological activates for the different tailor-made sophorolipids it is paramount to ensure the products are free of endotoxin contamination, thus allowing for reproducible and accurate testing of its biological activity. These different tailor-made biosurfactants can be tested for their ability singly or as a mixture thereof to increase cell turnover and differentiation (i.e. endothelium, fibroblasts and keratinocytes) in a range of biological assays which closely reflect cell turnover and differentiation in full-thickness cutaneous skin wounds. Once optimal dose-responses of the agent for increasing relevant cell turnover are established, the candidate agent can be further characterised in vivo. This involves creating a full thickness skin defect in biologically relevant animal models, where the rate of healing, histological, immunological, immunohistochemical and ultrastructural pathology of the wound are examined during the process of healing.
Finally the roles of biosurfactants in motility and interaction with cells and/or surfaces have been linked with the capability of bacteria to form biofilms. We observed that biosurfactants produced by yeasts such as sophorolipids interact differently with bacterial enzymes than with yeast ones. Sophorolipids negatively affect selected bacterial enzymes, while not affecting yeast enzymes while Rhamnolipids interacts negatively with selected yeast enzymes and not bacterial counterparts. The role of such interactions with microbial biofilm formation can be of great importance since such biofilms are implicated in many biological activities.
The Ulster University (UU) work package focussed on two major health technology challenges, relevant worldwide. Firstly, improving and expanding options for antimicrobial and antibacterial treatments and prevention of infection – within this area UU particularly targeted the scenario of wound healing. Secondly, the continuing requirement for novel options for cancer treatment.
Purified sophorolipids as test materials
Firstly, it is crucial to note that previously, much of the work (published literature and publicly available knowledge) that has investigated potential therapeutic and medical applications of has used non-purified samples or mixtures as test materials. The focus of the work described here is novel and innovative thanks to the commitment to using purified samples of predominantly single chemical structures. This approach generated reliable data which could be evaluated and discussed in relation to specific chemical structures, rather than complex mixtures. The advantage of this approach and the impact on the expanding knowledge-base cannot be understated.
The UU team found that endotoxin levels, measured by commercially available means, varied greatly between samples. However, investigations demonstrated that the test method applied could be a factor in this, in that the presence of β-glucan, a breakdown product of yeast cell walls, could interfere with endotoxin-detection assays and result in detection of false-positives. Despite this, with improved production and extraction methods developed by BIOSURFING collaborators, a selection of high-content or purified sophorolipid samples, with endotoxin levels within acceptable limits, became available for further analysis.
Wound healing and antimicrobial effects
The work investigating the effect of sophorolipids on wound healing has generated information with interesting implications. The UU approach was to screen non-purified samples which, despite being non-purified, contained high concentrations (>80% content) of single structured sophorolipids.
Simple, relatively high-throughput in vitro assays demonstrated that sophorolipids displayed antimicrobial activity against a range of clinically relevant bacteria: Pseudomonas aeruginosa, Enterococcus faecalis, Staphylococcus epidermidis and Salmonella typhimurium. It was observed that the Gram positive strains tested were more susceptible than the Gram negative strains. As well as reducing or inhibiting bacterial growth and biofilm formation at concentrations of between 1-5%, one of the most interesting and promising observations was the demonstration of antibiotic adjuvant activity of sophorolipids at sub-inhibitory concentrations. Generally, the lactonic sophorolipid was more effective in these assays compared to acidic sophorolipid. However, delivery of acidic sophorolipids in various culture media (e.g. nutrient broth for microbiology or variations on modified Eagle's medium for cell culture) proved to be easier than for lactonic types, which required DMSO or ethanol to aid solubility. This generated extra factors for investigations as some vehicle choices showed toxicity against the cells tested, and tolerable vehicle concentration became a limiting factor for the concentration range of lactonic sophorolipid that could be tested. Lactonic sophorolipids also displayed less favourable characteristics in some of the mammalian cell culture assays – for these reasons, assessment of the effect of sophorolipids on wound healing in vivo progressed using the purified acidic sophorolipid sample.
To our knowledge, this was the first study to apply sophorolipid topically to full thickness skin excision wounds. So far, the research has not detected any inherent benefit arising from applying acidic sophorolipid to wounds (within the remits of the models and structure tested). However, nor was any detrimental effect on wound healing been observed with sophorolipid application, and it is this observation, along with the preceding supporting data demonstrating anti-microbial, anti-biofilm and antibiotic adjuvant activity of sophorolipids which has exciting, and potentially life-saving implications.
Sophorolipids as anti-cancer agents
Due to the complexity and sensitivity of in vitro cell culture and in vivo assays, only purified, lyophilised sophorolipid samples were assessed in these models for anti-cancer activity. To our knowledge, the UU work package is the first to assess purified, single-structure sophorolipid samples as anti-cancer agents to relatively “normal” cells in culture, as well as a range of cancer cell-lines, specifically colorectal cancer cells. Furthermore, to our knowledge, this project was the first to evaluate potential anti-cancer activity of purified sophorolipid in vivo.
This research demonstrated a differential cytotoxic effect of purified acidic sophorolipids against cancer cell-lines only proving non- detrimental to “normal” control cell lines. This effect was not evident for the lactonic sophorolipid sample tested, proving detrimental to the latter, contrary to the general consensus currently in published literature. To further assess the anti-cancer potential of sophorolipids, both purified analogues were orally administrated in vivo to a well-established pre-cancerous mouse model; the Apcmin+/- mouse. This model develops spontaneous pre-cancerous intestinal polyps, precursors to intestinal adenomas. Lactonic sophorolipid exacerbated the intestinal pathology, evidenced by increased number and size of polyps, along with reduced haematocrit and increased spleen weight. Administration of acidic sophorolipid in this model resulted in no change to number and size of polyps however an increase in haematocrit and decrease in spleen size was observed. This was a positive outcome demonstrating a systematic effect of acidic sophorolipid. The effect of sophorolipid in a metastatic colon cancer model is still under evaluation.
Potential Impact:
CONFIDENTIAL: this part contains confidential information. Do not make public.
IMPACT
The European Union has a lead position in terms of volume and revenue of biosurfactants and is therefore investing in research on these bio-based molecules. Indeed, biosurfactants offer great opportunities, but the current lack of diversity limits their penetration in a broad range of applications; biosurfactants are applied, but only in specific niche products. Therefore a consortium of European academic and industrial partners are working together in the BIOSRUFING project to develop and test new types of biosurfactants. By developing a wider range of biosurfactants and evaluating their use in various applications (cleaning, cosmetics, medical and nanoscience), the ultimate goal of the project is to get biosurfactants to the supermarket shelf in day-to-day products available for every EU-citizen.
The glycolipid synthetic pathway in Starmerella bombicola was engineered to introduce structural variation in the produced biosurfactant molecules. We intend to expand the range of useful glycolipids beyond the natural variety and create tailor-made glycolipids with new and better physical and chemical properties. During the strain engineering process, several useful tools were developed to allow fast genetic modification and fast evaluation of the fitness of the obtained strains. These tools can in the future be applied to create and evaluate other S. bombicola strains producing not only novel glycolipids, but also other bio-compounds. In addition, this is a nice illustration of the value of the MRM method for strain development; this assay can be implemented for other industrial relevant strains as well and can help to bring bio-based product faster to the market.
The availability of dedicated and scaled processes for the production of very different biosurfactant products in combination with the availability of large sample sizes of these molecules is a very necessary step towards the commercialization of these new and interesting molecules. Only when sufficient and well-documented samples can be send out to interested companies, they will be willing to test them in their specific applications. During the BIOSURFING project we really invested in product purity and characterisation, as this turned out to be one of the bottlenecks in current application of biological compounds. Our work should help to remove some of the obstacles in the commercial implementation of biosurfactants.
The goal of UPMC was to study the complex self-assembly behaviour of sophorolipids and the new-to-nature derivatives. Indeed, most bio-based surfactants have more complex structures compared to common surfactants, resulting in more complex associations such as fibres and even transitions between different forms such as micelles, cylindrical micelles, planar bilayers, inverse micelles, large aggregates, liposomes etc. Nevertheless, the fundamental information on the self-assembling properties of the novel glycolipids in different conditions (such as pH and salinity) generates useful input for their application in cosmetics, food and pharmaceutical industries.
The results of WM and Ecover serve to assess newly developed biosurfactants for the use in various detergent categories. The possible consequence of suitability for this usage could be an exchange of conventional surfactants against biosurfactants. The production of biosurfactants is probably resource-saving in comparison to conventional surfactants manufacturing. As well, a lower impact on the environment after the use phase is expected. Yet, the novel glycolipids all display different behaviour and they cannot be implemented in an existing formulation by simply replacing a petroleum-based compound. Careful re-evaluation of formulations is required for proper and most optimal use.
One concern that came forward during the project is the GMO background of the strains. Indeed, the novel glycolipids are produced by carefully genetically modified organisms, but these yeasts themselves are no part of the final product. Nevertheless, there might be some issues regarding their use in cosmetics as the traditional wild type sophorolipids Sopholiance S produced by Soleance is approved by ECOCERT and COSMOS. Such designations could not be attributed to the Biosurfing materials, due to their GMO nature, compromising its commercial value, especially in cosmetics. On the other hand, this project could be used to open the discussion on the use of GMO’s and the hereto related regulations, not only on the political, but also on the public level.
Ulster University examined potential therapeutic applications of sophorolipids, in the contexts of two major challenges for healthcare in the 21st century: wound healing and associated microbial infection, and anti-cancer therapy. The data generated expands the current knowledge-base, and continuing research and development in this field has potential societal and economic impact.
Anti-microbial effects and wound healing
The problem of resistance to antimicrobial treatments is increasing and has been acknowledged internationally. National governments and international bodies are placing a large amount of focus and resources into developing and implementing strategies to combat this serious, growing, multi-faceted threat to aspects of “modern life”, in particular the resistance of bacteria to antibiotics. This decreased sensitivity impacts public health, veterinary and agricultural health, and economical health.
In addition to the specific diseases caused directly by pathogens, many invasive and surgical procedures carry a significant risk of microbial infection, as well as skin damage such as chronic wounds or burns. Multiple strains of multidrug resistant clinically relevant pathogenic species, such as Pseudomonas, Enterococcus or Salmonella, have been identified in cases of burns or surgical site infections, as well as posing risks to immunocompromised patients. It is also important to note that antibiotic administration is a standard preventative measure during cancer chemotherapy and following invasive surgery. The societal and financial burden arising from emerging antibiotic resistance, whilst complex and difficult to quantify, is widely considered to be significant (Table 1).
Table 1: Case study of estimated financial impact of cases of methicillin-resistant Staphylococcus aureus, data compiled from an example case study by Smith & Coast, 2012.
USA
- Additional cost of hospital treatment: $20,000 per patient per episode
Europe
- Patient deaths per year: 25,000
- Extra hospital days: 2.5 million
- Extra cost of hospital involvement: €900 million
The demonstrations of antibacterial effects of sophorolipids suggest a feasible role of sophorolipids in antimicrobial treatments, for example for topical applications to intact or wounded skin to prevent bacterial colonisation and resultant infections. Furthermore, the importance of adjuvant activity cannot be understated: these observations are clearly related to strategies to combat emerging drug resistance as outlined in the WHO antimicrobial resistance policy package (WHO, 2011). Pursuing research into defining a role for biosurfactants in reducing the effective dosages or minimal inhibitory concentrations of antibiotics necessary for treating or preventing infections in clinical settings should be encouraged.
Anti-cancer effects
The discovery, identification and development to trials of anti-cancer treatments prove a challenge in the cancer research field today. Most cancers prove resistant to certain treatments with unfavourable secondary effects as a consequence, such as death to normal organ resistant cells, causing detrimental effects and impacting on patient’s quality of life. This emphases the need for new natural chemotherapeutics which target specifically the tumour micro-environment (Department of Health (UK) 2011, WHO 2005).
The translational approach taken by Ulster University has improved the current knowledge-base for using a relatively cheap, renewable group of compounds as anti-cancer agents, not only by expanding the range of in vitro cancer models, but also by performing assays in vivo. In particular the observation of selective toxicity of one specific type of sophorolipid against cancer cells as opposed to normal cells is especially encouraging. Furthermore, the cost-effectiveness of using biosurfactants as anti-cancer agents is certainly worthy of further investigation: the role of biosurfactants in this field could be particularly useful in low- and middle-income countries, where cancer treatment options are less abundant, and cancer deaths are higher, relative to high-income countries (Cancer Research UK, WHO 2005).
Finally, this project can be an example to other projects dealing with Industrial Biotechnology for the production of (new-to-nature) biomolecules: lessons can be learnt and possible hurdles are clearer now.
DISSEMINATION
Several actions were taken to promote the Biosurfing project and disseminate the objectives and the results. Academic partners published their accomplishments in peer-reviewed journals or are preparing manuscripts accordingly. Furthermore, several contributions to conferences were made, both national and international, and as a poster or oral presentation. Specially the oral presentations always give rise to interesting discussions and contacts, very often also from interested companies eager to obtain some test material. Some academic partners also used the work conducted during the Biosurfing project to illustrate their lectures and master and PhD students were engaged on the practical work, in this way also integrating the Biosurfing project in the educational programmes.
Industrial partners disseminated the project on their website and company events. Some of these partners also referred to the Biosurfing project in company talks on public meetings.
The Biosurfing website and both brochures were very useful to these above mentioned dissemination activities and they can still be used in the near future. Furthermore, a dedicated workshop on Biosurfactants was organised in June 2015 and several of the Biosurfing partners were given the opportunity to give a talk there, nicely making a statement regarding the importance of our work towards the mixed academic and industrial public.
Details on the dissemination activities can be found in the later part of this final report.
EXPLOITATION
The majority of the described novel strains are protected by patent applications, either submitted during the project, or already anticipated on prior to the project initiation (e.g. strains producing acidic and lactonic sophorolipids, rhamnolipids, cellobioselipids). For some strains further confirmation of the results or further engineering is required before priority filling will be initiated, but this is monitored by the UGent Technology Transfer department. Together with this institute, InBio will safeguard the UGent IP position. After this project, this IP can be valorised by either licencing or by starting a spin-off company. Furthermore, certain research lines will be continued either in PhD-project or in novel collaborative projects.
The newly developed MRM method will stay of use for all genetic engineering work to be conducted in S. bombicola. Also after finishing the project, we will try to find a way to keep collaborating on this topic to apply and improve the method.
In the framework of the Biosurfing project, several innovative purification procedures for the novel glycolipids were developed. These are of particular interest when moving to industrial production and bringing molecules of a specific and repeatable purity on the market. The current knowledge is hard to protect by patenting, but will initially be kept in-house.
Several potential applications were tested and some yielded quite promising results:
1. Use of sophorolipids and glycolipids in deodorant or other related formulation to treat bad body odours and prevent microbial growth. There are potential applications in pharma, cosmetics and personal care. Yet, further testing is ongoing, and commercial use will depend on these outcomes. Patenting might be difficult due to existing overlapping patents.
2. Specific modified glycolipids display potential activity towards colon cancer. Other types are under evaluation. This currently requires further testing on different model systems, the option to take patents should be studied as there is quite some activity in this area.
3. Use of certain tailor glycolipids in nanomaterials. Requires further study regarding the stability of the system.
List of Websites:
http://www.biosurfing.ugent.be/
contact: Wim Soetaert (Wim.Soetaert@UGent.be) and Inge Van Bogaert (Inge.VanBogaert@UGent.be)
Prof. Dr. ir. Inge Van Bogaert
Laboratory of Industrial Biotechnology and Biocatalysis (InBio.be) -
Faculty of Bioscience Engineering -
Ghent University -
Coupure Links 653 -
9000 Ghent - Belgium
www.InBio.be
The European Union has a lead position in terms of volume and revenue of biosurfactants and is therefore investing in research on these bio-based molecules. Indeed, biosurfactants offer great opportunities, but the current lack of diversity limits their penetration in a broad range of applications; biosurfactants are applied, but only in specific niche products. Therefore a consortium of European academic and industrial partners were working together in the BIOSURFING project to develop and test new types of biosurfactants. By developing a wider range of biosurfactants and evaluating their use in various applications (cleaning, cosmetics, medical and nanoscience), the ultimate goal of the project is to get biosurfactants to the supermarket shelf in day-to-day products available for every EU-citizen.
The glycolipid synthetic pathway of the biosurfactant producing yeast Starmerella bombicola was engineered to introduce structural variation in the biosurfactant molecules. Hence, tailor-made glycolipids with altered physical and chemical properties were created. During the strain engineering process, several useful tools were developed to allow fast genetic modification and fast evaluation of the fitness of the obtained strains. These tools can accelerate the time-to-market process for novel bio-based molecules. Much effort was put in the scale-up and design of an industrial relevant downstream process for each novel biosurfactants. Indeed, due to the new-no-nature structures, processes completely different form the existing ones had to be designed. Thanks to feed-back from the application partners, knowledge about the required purity and specifications was generated and methods were developed to monitor and remove contaminants, resulting in products of high purity. Due to the availability of these high-quality products and detailed specifications, we expect companies to be less reluctant to evaluate these compounds in their specific applications in the future.
Intense application testing was obviously also performed by the BIOSURFING project partners. Interesting observations were made regarding the self-assembling behaviour of the different congeners and molecules were evaluated for cleaning applications. Yet, it turned out to be challenging to simply apply a substitution based approach. There is some potential in the use of certain compounds in specific medical and cosmetic applications requiring anti-bacterial activity, and a limited number of novel glycolipids is being evaluated as an anticancer agent.
Several actions were taken to promote the BiIOSURFING project and disseminate the objectives and the results. Academic partners published their accomplishments in peer-reviewed journals or are preparing manuscripts accordingly. Furthermore, several contributions to conferences were made, both national and international, and as a poster or oral presentation. Some academic partners also used the work conducted during the Biosurfing project to illustrate their lectures and master and PhD students were engaged on the practical work, in this way also integrating the Biosurfing project in the educational programmes.
The Biosurfing website and both brochures were very useful to these above mentioned dissemination activities, also for the industrial partners. Furthermore, a dedicated workshop on Biosurfactants was organised in June 2015 and several of the Biosurfing partners were given the opportunity to give a talk there, nicely making a statement regarding the importance of our work towards the mixed academic and industrial public.
In general, existing staff was engaged to conduct the project work, and in total 10 additional staff members were hired (from which 6 are woman). Some Gender Equality Actions were taken such as the implement an equal opportunity policy and the use of targets to achieve a gender balance in the workforce.
Project Context and Objectives:
This project aims to create new-to-nature and tailor-made biosurfactants through metabolic engineering of the unconventional yeast Starmerella bombicola (= new taxonomic correct name of Starmerella bombicola). Biosurfactants produced by fermentation offer a worthy alternative to traditional surfactants, which are typically derived from non-renewable petrochemical resources and may cause environmental problems due to their ecotoxicity and poor biodegradability. Despite the clear advantages of biosurfactants, their overall use is hampered by the lack of structural variation. This is in sharp contrast to chemically produced surfactants where one can introduce variation by simply changing the building blocks. Structural variation is essential as (bio)surfactants find application in a very broad range of sectors. This project aims to alleviate this fundamental limitation by developing a generic biotechnological production technology for glycolipid biosurfactants. This will in turn significantly broaden the range of commercial biosurfactants, satisfying the need for structural diversity in the market. It is expected that this technology will result in a breakthrough penetration of glycolipid biosurfactants in the overall surfactant market, in this way helping to build the bio-based economy.
The mean objectives of the BIOSURFING project can be divided in three large interacting parts: (1) strain development, (2) production and (3) application evaluation. More detailed, following objectives will be purchased:
- Development of a sophisticated molecular toolbox for the genetic engineering of S. bombicola
- Metabolic and genetic engineering of S. bombicola for the production of new-to-nature tailor-made biosurfactants
- Development of analytical methods for structure elucidation of the novel compounds
- Evaluation of the strains by proteome, metabolome and transcriptome analysis;
- Providing feed-back in order to overcome limitations or further improve the strains
- Assessment of wild-type catabolic processes
- Optimisation of the fermentation process and down-stream processing for each target molecule to obtain a robust and efficient production system
- Production of sufficient quantities of test material for the application tests
- Physicochemical evaluation as emulsifier/surfactant and comparison with market leaders in the field
- Evaluation of (modified) glycolipids in nanomaterial applications
- Evaluation of the biological properties of natural and tailor-made biosurfactants
- The efficient dissemination and exploitation of the project results
- Establishment of an efficient management scheme to ensure the fulfilment of the scientific and valorisation objectives of BIOSURFING project
To summarise the above objectives, the very efficient biosurfactant producing yeast S. bombicola will be metabolically engineered by applying the recently developed meganuclease technology, allowing to control all structural parts of the glycolipid biosurfactant molecule: fatty acid tail, sugar moiety, acetylation and lactonization. For each target molecule, a fermentation and downstream process will be developed and the molecules will be evaluated for various applications (cleaning and cosmetics, medics and nanoscience). The project thus covers the whole innovation chain from basic research to production and application development. To achieve this goal, a complementary consortium of academic and industrial partners has been formed that covers the whole range of required expertises.
Project Results:
CONFIDENTIAL: this part contains confidential information. Do not make public.
PART I: STRAIN DEVELOPMENT
Development of a sophisticated molecular tool-box
For the metabolic engineering of S. bombicola, molecular tools needed to be established. Targeted expression and easy marker re-use could render the genetic engineering of S. bombicola more efficient and turn S. bombicola into an easy to manipulate platform organism for the production of tailor-made biosurfactants. Within the framework of this project, CELLECTIS has developed three main genome engineering technologies for S. bombicola based on their meganuclease and TALEN® technology.
Site specific endonucleases recognizing large DNA sequences from 22 base pairs (meganucleases) and up to 34 bp (TALEN®) can be used for gene delivery and integration at the specific site they recognize, based on a general principle of DNA cutting by the enzyme and repair by the cell. Indeed, due to the long recognition sequence, a site specific endonuclease is highly specific and often only cuts once in a whole genome allowing targeted integration. Three site specific endonuclease based strategies has been developed during the project. Both existing and custom site specific endonucleases were produced and delivered.
(i) Landing pads for genomic integration of transgenes
The genome of S. bombicola has been scanned for meganuclease sites and regions beneficial for expression (non-coding region, high expression area, etc.) were further examined. The 14 best ones have been selected as so-called “landing pads” for meganuclease meditated modification (Table 1). For seven loci several targeted meganucleases have been developed (only one failed), and these were validated in yeast using SSA annealing assay (Epinat JC, 2014). The meganuclease showing the best activity on its target was delivered to UGent. Seven more targets were identified and the in silico meganucleases were developed offering a whole range of potential landing paths. UGent executed tests with one selected meganuclease not interfering with any coding region.
Table1: List of identified meganucleases that were produced and validated for transgenes landing pads integration.
Name Hit sequence Validation
BioS01 GCAAGCCATCGTACCACATCTCGA 0.7
BioS03 TTTTGCCACTGTACGCCATTTTGA 0.9
BioS04 TCTTGCTGAATTACGTGGGATCTA 0.5
BioS05 ACCCCCCGTCTTACCGTGGCACAT 0.8
BioS06 GTTTTATCCTTTACAAAGTTTTGA 0.9
BioS07 TCTCCCCCCTGTAACCTATTTCAT 0.5
NB: Validation level varies from 0 (no activity) to 1 (max activity), cut off for validation based on CELLECTIS know how on meganuclease was determined at 0.4.
(ii) Re-usable markers
In this subtask, a second set of site specific endonuclease were chosen (Table2). This set does not recognize any target in the S. bombicola genome and were used to develop scar-less excisable selection marker constructs that thus will be re-usable. These re-usable marker constructs rely on flanking a functional selection marker with unique site specific endonuclease recognition sites, the whole being inserted between tandem DNA repeats. Once inserted in the genome, excision of the marker gene will occur through high efficiency intra-molecular homologous recombination triggered by cutting (Perez et al. 2005).
Table2: List of Meganucleases and TALEN® and their corresponding target site for Re-usable markers project
Name Description Sequence
C1221 natural mega TCAAAACGTCGTACGACGTTTTGA
ADCY9.1 custom mega CCCAGATGTCGTACAGCAGCTTGG
CAPSN1 custom mega CAGGGCCGCGGTGCAGTGTCCGAC
CAPT1.1 TALEN TCCGGGAACCCAGAGCTcacagccacgatcttAGACCCGAGCCCACAGA
An antibody allowing the detection of I-Cre I meganuclease and corresponding protocol for western blot meganuclease detection were also delivered.
In addition to these meganucleases and TALEN®, new generation of site specific endonucleases developed at CELLECTIS, CompactTALEN® (Beurdeley et al. 2013) were proposed, designed, produced and delivered. These CompactTALEN® should allow an easier delivery in S. bombicola which was an issue in the project. The CompactTALEN® were designed to target two markers available in S. bombicola: URA3 and GFP. For each marker target two different CompactTALEN® were produced and delivered (Table3).
Table3 : List of CompactTALEN® and their corresponding target site for marker removal project:
Name targeted sequence strand
URA3cT1 CGCAGACATCGGTTCCACTGTTAAGGCCCAATATGCA 1
URA3cT2 GTAACTCAGAGGTAGTTCGCACATCCAAGCTTGCGCAAAG -1
GFPcT1 GGCGTTCAGTGCTTTGCCCGTTACCCCGACCACATGAA 1
GFPcT2 GCGATACTTTGGTAAACCGTATCGAGCTCAAGGGCATT 1
(iii) Metabolic gene platforms
Based on the target genes for metabolic engineering, a third set of targets was chosen within genes relevant to the biosynthetic or biochemical pathway. Custom-meganucleases and/or TALEN® were designed, engineered, produced and validated to target these specific genes (Table4). They were designed for their capacity to induce knock-outs, allelic substitutions (including libraries of candidate gene substitutes) or gene or promoter swaps in S. bombicola within the gene targeted.
Table4 : List of Meganucleases and TALEN® with their corresponding target site chosen for the Metabolic gene platform.
Name Gene of Interest Type Validation
BioS15 invertase cabom01g09430 Meganuclease /
BioS15 invertase cabom01g09430 TALEN® 0.91+/- 0.01
BioS16 Leu2 cabom02g11170 Meganuclease /
BioS16 Leu2 cabom02g11170 TALEN® 0.74 +/- 0.02
BioS17 Trp1 cabom02g00890 TALEN® 0.98 +/- 0.01
BioS18 Ade1 cabom03g11310 TALEN® 0.97 +/- 0.02
BioS23 Aro7 cabom03g15270 TALEN® 0.84 +/- 0.02
BioS19 CYP52M1 cabom02g13890 TALEN® 0.98 +/- 0.01
BioS20 GT1 cabom02g13880 TALEN® 0.99 +/- 0.02
BioS24 AT cabom02g13870 TALEN® 0.96 +/- 0.02
BioS21 GT2 cabom02g13850 TALEN® 0.93 +/- 0.03
BioS25 lactonase cabom02g01720 TALEN® 0.95 +/- 0.02
BioS22 MDR cabom02g13860 TALEN® 0.95 +/- 0.04
In addition to the site specific endonucleases delivered, CELLECTIS provided a construct expressing the single chain Trex2 that has been shown to stimulate targeted mutagenesis induced by meganucleases by one log factor (Delacôte et al. 2013) and TALEN®.
Strain development
Several strains either producing sophorolipids with an altered composition of the individual components or producing entirely new glycolipids were developed. This are listed and commented below:
1. A strain producing 100 % acidic sophorolipids: this one behaves very similar compared to the wild type regarding growth, nutrient requirement and production of sophorolipids: sophorolipids can be produced at rates and in quantities equal to the wild-type.
2. A strain producing almost 100 % lactonic sophorolipids. Again, this strain behaves very similar compared to the wild type regarding growth, nutrient requirement and production of sophorolipids: sophorolipids can be produced at rates and in quantities equal to the wild-type.
3. A strain producing cellobioselipids. Although the novel compounds are produced, this is in low amounts and combined with the synthesis of intermediate glucolipids. The strain evaluation procedure described below gives some input on possible reason for this less performing behaviour. These will be implemented in the follow-up of the project and should result in a strain with improved production characteristics.
4. A strain producing C16 sophorolipids: the novel compounds are produced at high purity (few other congeners and low amount of C18 sophorolipids), yet the yield is only 5 g/L maximum, which is quite low when compared to the wild type. Some samples can be provided, but only for applications requiring a low amount of material input for testing.
5. An engineered S. bombicola strain producing branched C22 sophorolipids is still under construction. Yet, in order to allow the applications partners to perform test with these types of molecules prior to the strain finalization, a non-Starmerella host was successfully used for the synthesis of these compounds.
6. A strain producing glucolipids was used to obtain glucolipids in a mixture of none- and acetylated congeners. Again, lower production titters compared to the wild type were noticed; yet, the strain evaluation procedure described below gives some input on possible reason for this less performing behaviour. These will be implemented in the follow-up of the project and should result in a strain with improved production characteristics.
Strain evaluation
As described above, genetically engineered strains of S. bombicola were created which produce new-to-nature sophorolipids (SL). However, some of those newly generated strains are not able to produced equal amounts of product as compare to the wilt type (WT) S. bombicola strain. The reasons for this lower production rate are not well understood. By analysing the proteome and metabolome of those strain and compare it with the WT, we aimed to understand this phenomenon. Proteomic approaches were used to analyse the amount of the sophorolipid biosynthetic cluster gene products: cytochrome P450, glycosyltransferase A1, glycosyltransferase B1, acetyltransferase, SL transporter. The first studied mutant was created by knock-out of the cytochrome P450 monooxygenase, responsible for the first step of SL biosynthesis (fatty acid hydroxylation), this newly generated strain (CYPKO) lost the ability to produce sophorolipids. We must stress that this strain is therefore not used for scale-up and new-to-nature sophorolipid production, but is a good negative control to be included in the strain evaluation. When the acetyltransferase gene participating in sophorolipid acetylation was knocked out, this mutant (ATKO) produced only non-acetylated sophorolipids. Additionally, when the glucosyltransferase (ugtA1), responsible for the attachment of the first glucose to the hydroxylated fatty acid, was replaced with ugt1 from Ustilago maydis, this S. bombicola mutant (CL) produced a new kind of the cellobiose lipids.
To characterize S. bombicola mutants producing new-to-nature molecules we developed multiple protocols to analyse their protein and metabolite level. Firstly we optimized a protocol to analyse WT and mutants using an isobaric labelling proteomic approach - iTRAQ. This experiment confirmed that in all mutants the amount of SL cluster proteins is different form the WT. As SL cluster proteins are underrepresented in the mutants this can have a direct connection with biosurfactant yield. It is possible that during genetic manipulation a promotor or regulatory region of the SL cluster was modified which causes reduced expression of this cluster. On the other hand, presence of sophorolipids could generate a positive feedback, which activates the whole SL pathway. Therefore, the suppression of the product of the pathway could lead to its down-regulation.
As the iTRAQ approach is time consuming and expensive, we also developed more targeted approach, which allows faster and easier analysis of cluster gene products in new S. bombicola mutants. We established a multi-reaction monitoring (MRM) assay. This method is based on the selection of unique peptides of each targeted protein that are then monitored in a triple quadrupole type analyser. For each SL cluster protein we select two unique peptides and we compared their intensity between the WT and mutants. With this approach we validated the results obtained with iTRAQ. We confirmed that in CYPKO and CL mutant SL cluster protein expression dropt around 50% in comparison to WT, while in the ATKO of about 20%. Recent analysis showed equal SL cluster protein presence for the strain producing 100% acidic sophorolipids. This is in line with the observed production volumes. Interestingly, the first glucosyltransferase was clearly more present in the mutant then in the wild type.
This MRM assay is a new toolbox, which can be used to study SL cluster protein abundance in new S. bombicola mutants.
We also performed a metabolite analysis on the WT and the three mutants described above. We focused on fatty acids (FA) and soluble polar metabolites.
Extracted FA were converted to methyl ester of fatty acids (FAME) via transesterification. FAMEs were analysed (GC-MS) and they were identified using the NIST MS 2.0. database. Additionally, identification was based on comparing FAMEs retention times with those of the standards used to generate the standard curve. The FAMEs quantification was based on the extracted ion chromatogram (XIC) of the m/Z 74 fragment ion. This ion is characteristic for all FAMEs, representing the methylated carboxyl group.
Regarding FA, 40 peaks were detected by gas chromatography analysis (GC-MS) for the WT and the three mutants (CYP, AT, CL) extracts, but only 30 of them could be identified. When the S. bombicola fatty acids profile is compared with other described yeast like Sacharomyces cerevisiae not many differences are observed. However we can clearly see that there is a difference between S. bombicola WT, a good producer of sophorolipids, and the CYPKO, a strain with no sophorolipid production. The variation is most clear in case of the most abundant fatty acids: C16:0, C16:1, C18:0, C18:1, which are building blocks in sophorolipids biosynthesis. It is the first time that we can prove that indeed an additional pool of fatty acids is necessary for sophorolipids generation. Similar conclusion can be drown from observation of other detected fatty acids like for example branched fatty acids and dicarboxylic acids. Likely their presence in the WT is important for SL production as those fatty acids can be substrates for hydroxylated fatty acids. Contradictory, the odd numbered fatty acids are more abundant in the CYPKO mutant than in WT. Possibly this indicates a difference in composition of the cells membranes which can also be crucial for sophorolipids secretion.
For the polar metabolite analysis we used a targeted approach where we concentrated on LC-MS analysis of 84 standard metabolites with known elution profile and MW. We identified and quantified 44 metabolites in the WT and the three analysed mutants. We could confirm that the CL strain, producing small amount of cellobiose lipids, does not perform well and we suggest that it is under some kind of stress. It can be related to the problems during genetic manipulation of this strain or due to the intracellular accumulation of the cellobiose lipids. We indeed found out that this strain was based on a spontaneous ura3-negative mutant, which now turned out to be less vital as compared to the general used PT36 controlled ura3-negative strain. These findings will be taken into consideration for further strain improvement. Surprisingly the difference in the level of metabolites in CYPKO and WT are not as extreme as with the CL strain. It would mean that S. bombicola does not suffer much from losing this secondary metabolite pathway. We found indeed a few differences which can be connected with the higher need of the WT for energy, cofactors, nitrogen and glucose. Metabolites profiles of the WT and ATKO strains are very similar. This fits to the SL production pattern, as the ATKO strain is the best producing mutant in this experimental setup, but still lower than the WT.
We believe that the results collected from proteomic and metabolomics analysis can be useful to improve molecular engineering strategies in order to create better performing S. bombicola mutants producing increased amount of new-to-nature biosurfactants. Developed protocols and the MRM assay can be used for monitoring and selection the best new S. bombicola mutants for industrial application.
Related to the stain evaluation, product catabolism by the strain was evaluated as well. There are no indications that catabolism is an issue for cultivation and production conditions applied. Hence, no further strain modification was required regarding this matter.
PART II: PRODUCTION
Strains that successfully passed the first stage of the project were taken to scale-up and industrial production, whit special attention for product purity.
Three Starmerella bombicola strains created by the University of Ghent could proceed to scale-up and were delivered to the Bio Base Europe Pilot Plant (BBEPP) for process development and scale-up. Also a Rhodotorula bogoriensis strain was delivered for the production of a special type of branched sophorolipids as the engineering of its biosynthesis in S. bombicola was not yet successful. For these four strains the production processes (fermentation and purification) were first developed at the lab scale (7L) and subsequently scaled up to the small pilot scale (150L) and the purified products were delivered to the partners. For two strains, these processes were further scaled up to the 4500 L scale for which a lot of useful input was already gathered during the process development at the lab and small pilot scale. Robust, scalable and efficient fermentation processes with defined process parameters were developed giving rise to several batches of end product with limited variation. The latter enabled setting up the specifications for each product while at the same time large ‘sample’ amounts of the products were generated, which could be used for dedicated application testing.
Four structurally different glycolipid products were produced at the large (>100 kg) or small (> 100 g) pilot scale. Because of their structural variation, the application potential of these molecules was expected to be completely different.
During the project a lot of attention was given to the productivity of the fermentation processes, yield maximisation of the purification processes and maximisation of the purity for the final products. These objectives were obtained by an iterative process, which for some of the products required multiple rounds. Feedback of the project partners on the sample purity was also taken into account to decrease the presence of (specific) contaminants such as salt, remaining sugars, proteins, triglycerides and fatty acids in the final product(s).
Five products were finally generated in sample sizes, which were large enough for several of the project partners to perform dedicated application research. These molecules were:
- Lactonic sophorolipids
- Acetylated acidic sophorolipids
- Non-acetylated acidic sophorolipids
- Glucolipids
- Acetylated branched acidic sophorolipids
An LCA analysis was performed and it was quite surprising to see that the impact of the production of these molecules was larger as expected, as they were not significantly better (nor worse) as the impact of most (fossil based) reference products. However, this was mainly due to the agricultural production processes for rapeseed oil and glucose.
Now that these processes have been developed up to a mature level and large sample sizes are available for application testing, the possibility of using waste or side streams for their production can be evaluated. The latter would have a tremendous impact on the impact of the production of these molecules.
PART III: APPLICATION TESTING
Self-assembling behaviour and nanomaterials
The novel glycolipids were studied in detail for their self-assembling behaviour. In their most-common configurations, surfactants are made of one polar headgroup, which is hydrophilic, and one hydrophobic, non-polar tail. This ambivalent chemical structure allows them to have affinities for both polar (e.g. water) and non-polar (e.g. oil) phases. Under these conditions, micelles, spherical aggregates, form thus reducing the interfacial tension between the aqueous solution and the non-polar layer. It is probably the most simple and ubiquitous arrangement formed by surfactant molecules, and it is usually formed above a critical concentration known as the critical micelle concentration (cmc).
Besides spherical micelles, surfactants can self-organize into other aggregates depending on the medium in which they are dispersed, and on the respective size of the hydrophilic and hydrophobic parts of the molecule: common topologies are cylindrical micelles, planar bilayers or inverse micelles, for instance. If the most common surfactants are composed of the association of hydrophilic and hydrophobic blocks in simple head-to-tail configurations, more complex associations are also found. This is the case in most bio-based surfactants, constituting the topic of the BIOSURFING Project. In this case, the properties in solution of these new molecules composed of bio-derived building blocks are mostly unknown, despite their high potential in cosmetics, food and pharmaceutical industries.
The goal of UPMC in WP5 was to study the complex self-assembly behaviour of sophorolipids and the new-to-nature derivatives.
The state of the art on the sophorolipid self-assembly before BIOSURFING was the following:
- 2004: the first study on self-assembly properties of acidic deacetylated C18:1 sophorolipids was published by Zhou et al. (Langmuir, 2004, 20, 7926). The authors have shown the formation of giant ribbons at room temperature in water at acidic pH for concentrations in the range few mg/mL.
- 2010: the same compound was studied by Baccile et al. (Green Chem., 2010, 12, 1564) under similar conditions and they found that at acidic pH only micelles could form.
- 2011: Penfold et al. (Langmuir, 2011, 27, 8867) did show also the formation of micellar aggregates in water under acidic-neutral conditions but in their case the compound was a mixture of acetylated and non-acetylated sophorolipids.
No other work on this topic was published before the beginning of the BIOSURFING Project in October 2011. Since then, our contribution was very important and it can be summarized hereafter by type of compound:
1. Acidic deacetylated sophorolipids (SL C18:1)
Acidic conditions
Basic conditions
Silica templating
Nanoparticle functionalization
2. Acidic acetylated sophorolipids (SLa C18:1)
3. Acidic deacetylayed glucolipids (GL)
4. In-situ mechanisms of formation
1. Acidic deacetylated sophorolipids (SL C18:1)
This compound can be considered as a kind of reference as it is the one with the highest production yield and the most studied in the literature. In the initial part of the BIOSURFING Project we spent a lot of energy to better study and understand the behavior of this compound, in order to settle the ground reference for the new-to-nature ones, which we only received later on. The achievements for this compound are summarized in the following points, to which the appropriate reference is provided for more detailed information.
Acidic conditions: At pH below neutrality we have found that SL C18:1 form micelles in water, thus confirming our early data of 2010. Furthermore, using Small Angle Neutron Scattering, we have also found that micelles are uncharged objects in solution below pH 5 while they acquire a negative charge in the 5 < pH < 7 range, due to the deprotonation of the free COOH group. These results have been published in 2012 and 2013: N. Baccile et al. Soft Matter, 2013, 9, 4911; N. Baccile et al., ACS Nano, 2012, 6, 4763–4776.
In a next study, (S. Manet et al., J. Phys. Chem. B, 2015) we provide a detailed description of the structure of the micelle formed by SL C18:1 by combining Small Angle X-ray Scattering using synchrotron radiation and Molecular Dynamics modelling. In particular, the micelle is not constituted by the homogeneous distribution of the sophorolipid, as one would expect for classical surfactants, but by a more complex distribution of the sophorose and COOH groups, both hydrophilic. This is a unique result obtained for a bolaform compound. Furthermore, the micelle is no strict sphere, but has a more ellipsoid form.
Basic conditions: At pH above neutrality we have found the coexistence of micellar aggregates and larger objects of ill-defined structure, as published in N. Baccile et al., ACS Nano, 2012, 6, 4763–4776. In a more recent work (A.-S. Cuvier et al., Chem. – Europ. J., 2015), in which we combined Small Angle X-ray Scattering using synchrotron radiation and cryogenic Transmission Electron Microscopy, we were finally able to better describe the nature of the aggregates: they are composed of nanoscale platelets. The self-assembly in this medium was never reported before for sophorolipids.
Silica templating: Micelles can be used as templates to cast inorganic materials and in particular, the fact of removing micelles after the casting process allows the formation of nanoporous solids, very much studied in various fields. We have been pioneers in the use of sophorolipids as micellar templates for silica already in 2010, in which we did show the feasibility of the process. However, during BIOSURFING, we have strengthened our position by publishing several articles in which we give more synthesis details about templating thin silica layers (N. Baccile et al., J. Phys. Chem. C 2013, 117, 23899) and, for the first time, porous silica particles (B. Thomas, ACS Sus. Chem. Eng. 2014, 2, 512). Both the sugar part and the carboxylic group are believed to interact with the amine functions linked to the silica.
Nanoparticle functionalization: The free-standing COOH group of sophorolipids can also be used as capping agent for nanoparticles. The group of B. L. V. Prasad (Pune, India) did report before 2011 several strategies to stabilize metal nanoparticles in water using sophorolipids. During BIOSURFING, we have used sophorolipids C18:1 to stabilize metal oxides for the first time, and in particular the superparamagnetic iron oxide nanoparticles (N. Baccile, Phys. Chem. Chem. Phys., 2013, 15, 1606). We could show different stabilization strategies, the control over the size distribution but also the stability towards high ionic strength aqueous media, which is the key feature to employ these materials in biomedical applications. By controlling the parameters, different size distributions could be obtained.
2. Acidic acetylated sophorolipids (SLa C18:1)
This compound was provided to us by the Bio Based Europe Pilot Plant with the T21 code. Its self-assembly properties have been studied by applying the typical protocols developed for the SL C18:1 compound. We used Small Angle X-ray Scattering using synchrotron radiation and cryo-transmission electron microscopy to analyse it. We did find that in the mild acidic pH region, ellipsoidally-shaped micelles are formed in water for 5 w% solutions. However, upon decreasing the pH to 2, micellar length strongly increases by at least a factor 3 and large aggregates form in solution, which becomes turbid. This behaviour is in contrast with the typical SL C18:1 sample. These data are published in the PhD manuscript of Anne-Sophie Cuvier and will be the topic of a future communication in a peer-reviewed journal. This compound cannot be studied at high pH values due to the chemical fragility of the acetyl ester function.
SLa have been used in an attempt to obtain porous silica powders, in which the microbial glycolipids are the porogenic agents. The use of TEOS as inorganic precursor drove the system towards the formation of gels; however, by tuning the synthesis conditions, we could better control the interactions between SLa and silica and obtain a powdery material. Under TEM, one can observe silica samples with a network of non-organized pores. This is confirmed by the N2 adsorption/desorption isotherms and the high surface area obtained (> 400 m2/g), values which are comparable with the mesoporous silica samples typically obtained with classical surfactants like block copolymers. Interestingly, we did not have to remove the glycolipids using a heavy and costly calcination step. Although this result is very interesting, it is still hard to explain and more work would be needed.
3. Acidic deacetylayed glucolipids (GL)
This compound was provided to us by the Bio Based Europe Pilot Plant and its self-assembly behaviour was studied using the same protocols employed in the above-mentioned studies. We have in particular looked at its pH-dependent behaviour in the 0.5 – 5 w% region combining cryo-TEM and SAXS using synchrotron radiation. Very differently than all other compounds, GL forms entangled elongated micelles at basis pH and vesicles at pH below 6. These results are the first ones produced on this compound and will constitute an upcoming publication.
We have also used glucolipids to stabilize magnetic iron oxide nanoparticles in a two-step procedure. In particular, we could show for the first time that the vesicle-forming properties of glucolipids at pH 6 can fully stabilize the iron oxide particle colloidal dispersion after going through pH 2. The strong pH-dependent supramolecular behaviour of these glucolipids (vesicles at pH 6 and lamellar phase at pH 2) seem to explain our results.
4. In-situ mechanism of formation
One of the most complex aspect of the self-assembly of bio-based glycolipids is to understand, and put in evidence, the mechanism of formation depending on pH. SL C18:1 and SLa C18:1 behave similarly, but not exactly in the same way. GL and SL C18:1 have a very different behaviour. In all cases, the differences between these molecules are very small. In order to better understand such differences, we have carried out a series of pH-dependent SAXS studies, unique in the literature for these compounds, at the ID02 line of ESRF synchrotron in Grenoble (France). These data are still under analysis but there are indications for the evolution between nanoscale platelets and micellar aggregates in water at 0.5 w%. Advanced modelling of these data, in comparison with the other bio-based glycolipids will hopefully give a better insight in terms of pH/structure self-assembly behaviour in water.
Cleaning applications
Ecover received 8 samples in total. These samples were tested on their basic properties relevant for the cleaning industry like physic-chemical parameters (surface tension and foam behaviour) and aquatic toxicity. The samples were compared with reference sophorolipids.
A first experiment was set up to check the water solubility of the molecules provided. A limited water solubility has implications for incorporating the compound into a formula. Secondly impurities in the samples, like residual oil, can have an influence on the subsequent tests. The lactonic sample displayed low solubility, hampering some of the applications.
The minimal surface tension of each surfactant was measured using a Krüss DSA100 tensiometer. From the results obtained it could be concluded that none of the Biosurfing sophorolipid surfactants was able to reach a lower minimal surface tension than the reference wild-type sophorolipid as used in commercial available cleaning formulations.
Foam behaviour of the surfactants was reviewed on two aspects: foam ability and foam stability. It is known for wild-type sophorolipids that they are limited to non-foaming applications due to their limited properties. It was found that the acidic sophorolipid (T21) sample had high foam ability and low foam stability in comparison with wild type sophorolipid mixtures. The branched acidic C22 sample had low foam stability, probably due to the residual oil in the sample. Also the foam stability of the glucolipid sample was less in comparison with a mixture of wild type sophorolipids.
The aquatic toxicity was measured using a daphtoxkit F. The surfactants T21 and wild type sophorolipids were found to be practically non-toxic. Branched acidic C22 and the glucolipid mix were tested in a cleaning formulation. The branched acidic C22 sophorolipid cleaning formulation was found to be moderately toxic and the glucolipid mix cleaning formulation was found to be toxic.
From the samples obtained during this project, none of them seemed to have an added value when applied in the conditions applicable for the traditional sophorolipids, in comparison with the sophorolipids compounds used in the company’s commercial available cleaning products.
Werner & Mertz tried to find an optimal application area for the newly developed glycolipids in comparison with commercially available surfactants. To this end, glycolipids were tested in various cleaning formulations, and performance and behaviour tests were conducted and compared. The novel glycolipids were implemented in four different products; each representing a specific product category:
Acidic Cleaner 1 (sanitary cleaner) pH = 1.7 - 1.9
Neutral Cleaner 1 (manual dishwashing liquid) pH = 5.5 - 5.7
Neutral Cleaner 2 (floor cleaner) pH = 7.4 - 8.0
Alkaline Cleaner 1 (degreaser) pH = 9.6 - 9.9
For an optimal comparison between the new glycolipids and other surfactants, the nonionic surfactant in every cleaner was varied. The new glycolipids were checked against the following alkyl polyglucoside and sophorolipid:
Lauryl glucoside [1]
(Dodecyl glucoside, CAS: 59122-55-3, CH3(CH2)nO[C6H10O5]mH, n=11-14, m=1.4-1.6)
Standard Sophorolipid “Sophoclean” by company “Soliance” (for the first reporting period)
The following samples of new glycolipids were provided by “Bio Base Europe Pilot Plant VZW” and examined:
1. Sophorolipid T21 Acidic (KO)
2. Sophorolipid T23 C1 Lactonic (OE)
3. Glucolipid (GL01)
Cleaning Performance of the acidic cleaner 1 was examined by monitoring of lime soap removal and lime scale removal. Dynamic cleaning performance of the neutral cleaner 2 and alkaline cleaner 1 was tested with a burnt fat-dust soiling assay. Their self-acting cleaning performance was evaluated in a similar way, but without applying mechanical power. Furthermore, general properties such as emulsification, foaming and wetting ability, dynamic surface tension, stability and viscosity were determined as well. The findings for each product category are given below:
Acidic Cleaner 1
Glucolipid was not soluble and sophorolipid Acidic T 21 was not stable in the Acidic Cleaner formulation, so they cannot be recommended for the acidic system.
Although the remaining stable sophorolipid Lactonic T23 C1 sample showed a slightly better lime scale removal power, the lime soap removal ability was considerably weaker than the conventionally produced lauryl glucoside. Lauryl glucoside decreased the surface tension significantly more than Lactonic T23 C1 cleaner solution. The variation of the glycosides had no impact on the foaming ability.
Neutral Cleaner 1
The loss of emulsifying power of the three biosurfactants from the project in comparison to lauryl glucoside was too large to be an appropriate alterative in manual dishwashing agents.
Sophoclean that is produced by the untreated wild-type had the best foaming performance; the other samples had the same lower foamability. Lauryl glucoside and glucolipid had nearly the same increasing influence on the viscosity in the used salt thickening system. The sophorolipids had no effect on the viscosity. The use of the newly developed biosurfactants in manual dishwashing detergent formulas like Neutral Cleaner 1 can only be recommended if a loss of performance can be accepted.
Neutral Cleaner 2
Sophorolipid Acidic T21 and the Sophoclean standard were not stable in the Neutral Cleaner 2 formulation.
The dynamic cleaning performance concerning removal of a fat-dust soiling of the glucolipid is much and the performance of sophorolipid Lactonic T23 C1 is slightly weaker than the conventionally produced lauryl glucoside.
Self-acting cleaning performances of Neutral Cleaner 2 with lauryl glucoside and Lactonic T23 C1 were at the same level, but glucolipid had nearly no cleaning impact.
The wetting ability of the two remaining Biosurfing samples was not sufficient.
The variation of the non-ionic surfactant had no effect on the foaming ability of the cleaners.
The achieved cleaning abilities were not sufficient for the application area of floor cleaners. We cannot recommend using one of the tested samples in a similarly constructed floor cleaner.
Alkaline Cleaner 1
Besides the standard lauryl glucoside, only the glucolipid sample was stable in the pH-area of the Alkaline Cleaner 1.
The cleaning ability concerning dynamic removal of a fat-dust soiling of glucolipid was on a high level. Lauryl glucoside could not reach the performance of the biosurfactant.
Glucolipid is able to increase the quality of foam built by a foam spray trigger, but unfortunately glucolipid showed nearly no self-acting cleaning performance and a poor wetting ability.
To conclude, one can say that in none of the tested cleaning product categories the results were satisfying. This without exceptions, so a clear recommendation for one of the application areas cannot be stated. Although results in single performance tests fulfil the requirements, other results in the same type of product were not convincing enough.
Sophorolipids produced by the untreated wild-type of the used yeast have nearly the same properties as the new samples. We recommend continuing metabolic engineering on Starmerella bombicola to generate further biosurfactant compositions. Particular products based on a shorter fatty acid chain are expected to show the desired properties.
Applications in cosmetics
In the cosmetics industry there is a current trend to replace conventional cosmetic ingredients with naturally derived alternatives. In this view, sophorolipids T18C3, T21, T22, T23C1, T23C3, T37, T43 have been provided by partner UGent to partner COSMETIC to be tested. Based on the sophorolipids structure, various potential roles have been envisaged in cosmetic applications, such as emulsifiers, detergents, wetting agents, solubilisers or anti-microbial actives.
Within the frame of the investigation of the sophorolipid performance in such potential applications, a number of relevant physicochemical parameters have been initially calculated. We have calculated Cc and HLB values for all sophorolipids provided. Both these parameters are typically being considered in the state-of-the-art cosmetic science to guide the selection of appropriate emulsifiers to formulate a stable emulsion. HLB values calculated also gave us indications of different potential applications of the sophorolipids in the cosmetic industry.
We have studied the use of sophorolipids as solubilisers. Solubilisers are commonly used in cosmetic products to emulsify the fragrances/oils at molecular level and yield transparent solutions. T21 (the less purified acidic sophorolipid) performed well at ratios higher than 3:1 (solubiliser:oil) whereas the T37 (the acidic sophorolipid of high purity) was average and at ratios higher than 5:1. A possible explanation might be that the T21 contains some other impurity such as mixture of acidic sophorolipids acetylated/non-acetylated or lipid compounds that exhibit synergy with the acidic sophorolipid. T21 has the potential to be used as natural solubiliser.
Sophorolipids did not performed well as emulsifiers either at room temperature or at 80 °C. Even when using a high shear homogeniser the emulsions produced where not stable. Given that T21 performed well as solubiliser maybe it can create stable emulsions but at a much higher concentration than traditionally used emulsifiers.
Finally, our results demonstrate that the provided rhamnolipid can have applications in cosmetics where medium to low foaming is desirable similar to T21 as reported in deliverable 5.5.
Guided by the calculated HLB values for the provided sophorolipids (as reported in Deliverable 5.5) we have evaluated the potential use of the most promising candidate (T23C3- a diacetylated lactonic sophorolipid), as preservative booster in cosmetic formulations. Our results indicated that the provided preparations of lactonic diacetylated sophorolipids cannot have applications in cosmetics as preservative booster, at least at the concentration range investigated.
Another aspect studied was the antimicrobial activity. Our results confirmed literature reports on the targeted bacteriocidal activity of sophorolactone only on some Gram positive bacteria (Corynebacterium xerosis, Propionibacterium acnes).
When correcting for lactone content, purified lactone T43 was found as potent as commercial Sopholiance S against C. xerosis (the microorganism responsible for malodour under the armpit), when growth was assessed at 20h after inoculation. The MIC value we have determined for Sopholiance S (0.025 % w/w) is comparable to literature reports (0.05 % w/w without correction for lactone content, i.e. 0.03 % when corrected for a 60 % lactone content). Growth inhibition was retained longer in the case of T43, as opposed to Sophliance S. MBC values were also found comparable.
Regarding P. acnes (acne causing bacteria), our preliminary results do indicate a superior bacteriostatic activity of Biosurfing lactone T43 at 20h after inoculation, which is however not apparent, already at 27h of incubation. The lactone lack of solubility did not allow us test higher sophorolipid concentrations. Our results demonstrate in any case, a roughly comparable bacteriostatic potency.
A competitive advantage of Sophoroloactone preparation as dry powder (as in the case of Biosurfing T43) is that it is much more stable to hydrolysis, i.e. has a longer shelf-life when compared to the liquid preparation of Sopholiance S. Indeed, while performing our experiments we did notice a significant degradation of the lactone (measured by HPLC) to a value which is half the reported lactone concentration by the manufacturer! Also, when not taking into account the lactone content, Biosurfing T43 is a superior antimicrobial compared to Sopholiance S, which could also add a competitive, commercial advantage to T43.
The major comparative weakness however of T43 is that it is a GMO product, as opposed to Sopholiance S. Sopholiance S is approved by ECOCERT and COSMOS. Such designations could not be attributed to T43, due to its GMO nature, compromising its commercial value.
An acidic sophorolipid fraction T21 exhibited good solubilizing properties of perfumes and oils. Sophorolactones had good antimicrobial properties against selected gram positive bacteria such as Corynebacterium xerosis one microorganism responsible for the malodour of the armpit and Propionibacterium acnes the acne-causing bacteria.
Biological activity
There have been numerous preliminary reports stating that biosurfactants are suitable alternatives to some synthetic medicines and antimicrobial agents and may be used as safe and effective adjuvant or therapeutic agents. Some of these biosurfactants were described for their potential as biologically active compounds and applications in the medical field. Some of the lipopeptides biosurfactants (iturin and surfactin) have been implicated with high surface activities and antibiotic potential. Such compounds enhance or decrease the bacterial surface hydrophobicity an aspect which appears to be essential in relation to their effect on microbial pathogenicity and or susceptibility to antibiotics. In addition their low cytotoxicity to mammalian cells makes them suitable for use in many medical and microbial attack applications. Other glycolipids such as sophorolipids of Starmerella bombicola and rhamnolipids of Pseudomonas sp. have also been shown to have some antimicrobial activity particularly against gram-positive bacteria and varying levels of effects on the pathogenicity of some fungal and bacterial strains.
Biosurfactants are ideal compounds to increase O2 saturation to the surface of healing wounds, while providing a natural barrier to infection and aiding the migration of cells important in the healing response. To date there are few studies investigating the potential of these compounds to accelerate wound healing in vivo. Furthermore, there are some reports on the anticancer activity of natural and modified glycolipids.
To test such biological activates for the different tailor-made sophorolipids it is paramount to ensure the products are free of endotoxin contamination, thus allowing for reproducible and accurate testing of its biological activity. These different tailor-made biosurfactants can be tested for their ability singly or as a mixture thereof to increase cell turnover and differentiation (i.e. endothelium, fibroblasts and keratinocytes) in a range of biological assays which closely reflect cell turnover and differentiation in full-thickness cutaneous skin wounds. Once optimal dose-responses of the agent for increasing relevant cell turnover are established, the candidate agent can be further characterised in vivo. This involves creating a full thickness skin defect in biologically relevant animal models, where the rate of healing, histological, immunological, immunohistochemical and ultrastructural pathology of the wound are examined during the process of healing.
Finally the roles of biosurfactants in motility and interaction with cells and/or surfaces have been linked with the capability of bacteria to form biofilms. We observed that biosurfactants produced by yeasts such as sophorolipids interact differently with bacterial enzymes than with yeast ones. Sophorolipids negatively affect selected bacterial enzymes, while not affecting yeast enzymes while Rhamnolipids interacts negatively with selected yeast enzymes and not bacterial counterparts. The role of such interactions with microbial biofilm formation can be of great importance since such biofilms are implicated in many biological activities.
The Ulster University (UU) work package focussed on two major health technology challenges, relevant worldwide. Firstly, improving and expanding options for antimicrobial and antibacterial treatments and prevention of infection – within this area UU particularly targeted the scenario of wound healing. Secondly, the continuing requirement for novel options for cancer treatment.
Purified sophorolipids as test materials
Firstly, it is crucial to note that previously, much of the work (published literature and publicly available knowledge) that has investigated potential therapeutic and medical applications of has used non-purified samples or mixtures as test materials. The focus of the work described here is novel and innovative thanks to the commitment to using purified samples of predominantly single chemical structures. This approach generated reliable data which could be evaluated and discussed in relation to specific chemical structures, rather than complex mixtures. The advantage of this approach and the impact on the expanding knowledge-base cannot be understated.
The UU team found that endotoxin levels, measured by commercially available means, varied greatly between samples. However, investigations demonstrated that the test method applied could be a factor in this, in that the presence of β-glucan, a breakdown product of yeast cell walls, could interfere with endotoxin-detection assays and result in detection of false-positives. Despite this, with improved production and extraction methods developed by BIOSURFING collaborators, a selection of high-content or purified sophorolipid samples, with endotoxin levels within acceptable limits, became available for further analysis.
Wound healing and antimicrobial effects
The work investigating the effect of sophorolipids on wound healing has generated information with interesting implications. The UU approach was to screen non-purified samples which, despite being non-purified, contained high concentrations (>80% content) of single structured sophorolipids.
Simple, relatively high-throughput in vitro assays demonstrated that sophorolipids displayed antimicrobial activity against a range of clinically relevant bacteria: Pseudomonas aeruginosa, Enterococcus faecalis, Staphylococcus epidermidis and Salmonella typhimurium. It was observed that the Gram positive strains tested were more susceptible than the Gram negative strains. As well as reducing or inhibiting bacterial growth and biofilm formation at concentrations of between 1-5%, one of the most interesting and promising observations was the demonstration of antibiotic adjuvant activity of sophorolipids at sub-inhibitory concentrations. Generally, the lactonic sophorolipid was more effective in these assays compared to acidic sophorolipid. However, delivery of acidic sophorolipids in various culture media (e.g. nutrient broth for microbiology or variations on modified Eagle's medium for cell culture) proved to be easier than for lactonic types, which required DMSO or ethanol to aid solubility. This generated extra factors for investigations as some vehicle choices showed toxicity against the cells tested, and tolerable vehicle concentration became a limiting factor for the concentration range of lactonic sophorolipid that could be tested. Lactonic sophorolipids also displayed less favourable characteristics in some of the mammalian cell culture assays – for these reasons, assessment of the effect of sophorolipids on wound healing in vivo progressed using the purified acidic sophorolipid sample.
To our knowledge, this was the first study to apply sophorolipid topically to full thickness skin excision wounds. So far, the research has not detected any inherent benefit arising from applying acidic sophorolipid to wounds (within the remits of the models and structure tested). However, nor was any detrimental effect on wound healing been observed with sophorolipid application, and it is this observation, along with the preceding supporting data demonstrating anti-microbial, anti-biofilm and antibiotic adjuvant activity of sophorolipids which has exciting, and potentially life-saving implications.
Sophorolipids as anti-cancer agents
Due to the complexity and sensitivity of in vitro cell culture and in vivo assays, only purified, lyophilised sophorolipid samples were assessed in these models for anti-cancer activity. To our knowledge, the UU work package is the first to assess purified, single-structure sophorolipid samples as anti-cancer agents to relatively “normal” cells in culture, as well as a range of cancer cell-lines, specifically colorectal cancer cells. Furthermore, to our knowledge, this project was the first to evaluate potential anti-cancer activity of purified sophorolipid in vivo.
This research demonstrated a differential cytotoxic effect of purified acidic sophorolipids against cancer cell-lines only proving non- detrimental to “normal” control cell lines. This effect was not evident for the lactonic sophorolipid sample tested, proving detrimental to the latter, contrary to the general consensus currently in published literature. To further assess the anti-cancer potential of sophorolipids, both purified analogues were orally administrated in vivo to a well-established pre-cancerous mouse model; the Apcmin+/- mouse. This model develops spontaneous pre-cancerous intestinal polyps, precursors to intestinal adenomas. Lactonic sophorolipid exacerbated the intestinal pathology, evidenced by increased number and size of polyps, along with reduced haematocrit and increased spleen weight. Administration of acidic sophorolipid in this model resulted in no change to number and size of polyps however an increase in haematocrit and decrease in spleen size was observed. This was a positive outcome demonstrating a systematic effect of acidic sophorolipid. The effect of sophorolipid in a metastatic colon cancer model is still under evaluation.
Potential Impact:
CONFIDENTIAL: this part contains confidential information. Do not make public.
IMPACT
The European Union has a lead position in terms of volume and revenue of biosurfactants and is therefore investing in research on these bio-based molecules. Indeed, biosurfactants offer great opportunities, but the current lack of diversity limits their penetration in a broad range of applications; biosurfactants are applied, but only in specific niche products. Therefore a consortium of European academic and industrial partners are working together in the BIOSRUFING project to develop and test new types of biosurfactants. By developing a wider range of biosurfactants and evaluating their use in various applications (cleaning, cosmetics, medical and nanoscience), the ultimate goal of the project is to get biosurfactants to the supermarket shelf in day-to-day products available for every EU-citizen.
The glycolipid synthetic pathway in Starmerella bombicola was engineered to introduce structural variation in the produced biosurfactant molecules. We intend to expand the range of useful glycolipids beyond the natural variety and create tailor-made glycolipids with new and better physical and chemical properties. During the strain engineering process, several useful tools were developed to allow fast genetic modification and fast evaluation of the fitness of the obtained strains. These tools can in the future be applied to create and evaluate other S. bombicola strains producing not only novel glycolipids, but also other bio-compounds. In addition, this is a nice illustration of the value of the MRM method for strain development; this assay can be implemented for other industrial relevant strains as well and can help to bring bio-based product faster to the market.
The availability of dedicated and scaled processes for the production of very different biosurfactant products in combination with the availability of large sample sizes of these molecules is a very necessary step towards the commercialization of these new and interesting molecules. Only when sufficient and well-documented samples can be send out to interested companies, they will be willing to test them in their specific applications. During the BIOSURFING project we really invested in product purity and characterisation, as this turned out to be one of the bottlenecks in current application of biological compounds. Our work should help to remove some of the obstacles in the commercial implementation of biosurfactants.
The goal of UPMC was to study the complex self-assembly behaviour of sophorolipids and the new-to-nature derivatives. Indeed, most bio-based surfactants have more complex structures compared to common surfactants, resulting in more complex associations such as fibres and even transitions between different forms such as micelles, cylindrical micelles, planar bilayers, inverse micelles, large aggregates, liposomes etc. Nevertheless, the fundamental information on the self-assembling properties of the novel glycolipids in different conditions (such as pH and salinity) generates useful input for their application in cosmetics, food and pharmaceutical industries.
The results of WM and Ecover serve to assess newly developed biosurfactants for the use in various detergent categories. The possible consequence of suitability for this usage could be an exchange of conventional surfactants against biosurfactants. The production of biosurfactants is probably resource-saving in comparison to conventional surfactants manufacturing. As well, a lower impact on the environment after the use phase is expected. Yet, the novel glycolipids all display different behaviour and they cannot be implemented in an existing formulation by simply replacing a petroleum-based compound. Careful re-evaluation of formulations is required for proper and most optimal use.
One concern that came forward during the project is the GMO background of the strains. Indeed, the novel glycolipids are produced by carefully genetically modified organisms, but these yeasts themselves are no part of the final product. Nevertheless, there might be some issues regarding their use in cosmetics as the traditional wild type sophorolipids Sopholiance S produced by Soleance is approved by ECOCERT and COSMOS. Such designations could not be attributed to the Biosurfing materials, due to their GMO nature, compromising its commercial value, especially in cosmetics. On the other hand, this project could be used to open the discussion on the use of GMO’s and the hereto related regulations, not only on the political, but also on the public level.
Ulster University examined potential therapeutic applications of sophorolipids, in the contexts of two major challenges for healthcare in the 21st century: wound healing and associated microbial infection, and anti-cancer therapy. The data generated expands the current knowledge-base, and continuing research and development in this field has potential societal and economic impact.
Anti-microbial effects and wound healing
The problem of resistance to antimicrobial treatments is increasing and has been acknowledged internationally. National governments and international bodies are placing a large amount of focus and resources into developing and implementing strategies to combat this serious, growing, multi-faceted threat to aspects of “modern life”, in particular the resistance of bacteria to antibiotics. This decreased sensitivity impacts public health, veterinary and agricultural health, and economical health.
In addition to the specific diseases caused directly by pathogens, many invasive and surgical procedures carry a significant risk of microbial infection, as well as skin damage such as chronic wounds or burns. Multiple strains of multidrug resistant clinically relevant pathogenic species, such as Pseudomonas, Enterococcus or Salmonella, have been identified in cases of burns or surgical site infections, as well as posing risks to immunocompromised patients. It is also important to note that antibiotic administration is a standard preventative measure during cancer chemotherapy and following invasive surgery. The societal and financial burden arising from emerging antibiotic resistance, whilst complex and difficult to quantify, is widely considered to be significant (Table 1).
Table 1: Case study of estimated financial impact of cases of methicillin-resistant Staphylococcus aureus, data compiled from an example case study by Smith & Coast, 2012.
USA
- Additional cost of hospital treatment: $20,000 per patient per episode
Europe
- Patient deaths per year: 25,000
- Extra hospital days: 2.5 million
- Extra cost of hospital involvement: €900 million
The demonstrations of antibacterial effects of sophorolipids suggest a feasible role of sophorolipids in antimicrobial treatments, for example for topical applications to intact or wounded skin to prevent bacterial colonisation and resultant infections. Furthermore, the importance of adjuvant activity cannot be understated: these observations are clearly related to strategies to combat emerging drug resistance as outlined in the WHO antimicrobial resistance policy package (WHO, 2011). Pursuing research into defining a role for biosurfactants in reducing the effective dosages or minimal inhibitory concentrations of antibiotics necessary for treating or preventing infections in clinical settings should be encouraged.
Anti-cancer effects
The discovery, identification and development to trials of anti-cancer treatments prove a challenge in the cancer research field today. Most cancers prove resistant to certain treatments with unfavourable secondary effects as a consequence, such as death to normal organ resistant cells, causing detrimental effects and impacting on patient’s quality of life. This emphases the need for new natural chemotherapeutics which target specifically the tumour micro-environment (Department of Health (UK) 2011, WHO 2005).
The translational approach taken by Ulster University has improved the current knowledge-base for using a relatively cheap, renewable group of compounds as anti-cancer agents, not only by expanding the range of in vitro cancer models, but also by performing assays in vivo. In particular the observation of selective toxicity of one specific type of sophorolipid against cancer cells as opposed to normal cells is especially encouraging. Furthermore, the cost-effectiveness of using biosurfactants as anti-cancer agents is certainly worthy of further investigation: the role of biosurfactants in this field could be particularly useful in low- and middle-income countries, where cancer treatment options are less abundant, and cancer deaths are higher, relative to high-income countries (Cancer Research UK, WHO 2005).
Finally, this project can be an example to other projects dealing with Industrial Biotechnology for the production of (new-to-nature) biomolecules: lessons can be learnt and possible hurdles are clearer now.
DISSEMINATION
Several actions were taken to promote the Biosurfing project and disseminate the objectives and the results. Academic partners published their accomplishments in peer-reviewed journals or are preparing manuscripts accordingly. Furthermore, several contributions to conferences were made, both national and international, and as a poster or oral presentation. Specially the oral presentations always give rise to interesting discussions and contacts, very often also from interested companies eager to obtain some test material. Some academic partners also used the work conducted during the Biosurfing project to illustrate their lectures and master and PhD students were engaged on the practical work, in this way also integrating the Biosurfing project in the educational programmes.
Industrial partners disseminated the project on their website and company events. Some of these partners also referred to the Biosurfing project in company talks on public meetings.
The Biosurfing website and both brochures were very useful to these above mentioned dissemination activities and they can still be used in the near future. Furthermore, a dedicated workshop on Biosurfactants was organised in June 2015 and several of the Biosurfing partners were given the opportunity to give a talk there, nicely making a statement regarding the importance of our work towards the mixed academic and industrial public.
Details on the dissemination activities can be found in the later part of this final report.
EXPLOITATION
The majority of the described novel strains are protected by patent applications, either submitted during the project, or already anticipated on prior to the project initiation (e.g. strains producing acidic and lactonic sophorolipids, rhamnolipids, cellobioselipids). For some strains further confirmation of the results or further engineering is required before priority filling will be initiated, but this is monitored by the UGent Technology Transfer department. Together with this institute, InBio will safeguard the UGent IP position. After this project, this IP can be valorised by either licencing or by starting a spin-off company. Furthermore, certain research lines will be continued either in PhD-project or in novel collaborative projects.
The newly developed MRM method will stay of use for all genetic engineering work to be conducted in S. bombicola. Also after finishing the project, we will try to find a way to keep collaborating on this topic to apply and improve the method.
In the framework of the Biosurfing project, several innovative purification procedures for the novel glycolipids were developed. These are of particular interest when moving to industrial production and bringing molecules of a specific and repeatable purity on the market. The current knowledge is hard to protect by patenting, but will initially be kept in-house.
Several potential applications were tested and some yielded quite promising results:
1. Use of sophorolipids and glycolipids in deodorant or other related formulation to treat bad body odours and prevent microbial growth. There are potential applications in pharma, cosmetics and personal care. Yet, further testing is ongoing, and commercial use will depend on these outcomes. Patenting might be difficult due to existing overlapping patents.
2. Specific modified glycolipids display potential activity towards colon cancer. Other types are under evaluation. This currently requires further testing on different model systems, the option to take patents should be studied as there is quite some activity in this area.
3. Use of certain tailor glycolipids in nanomaterials. Requires further study regarding the stability of the system.
List of Websites:
http://www.biosurfing.ugent.be/
contact: Wim Soetaert (Wim.Soetaert@UGent.be) and Inge Van Bogaert (Inge.VanBogaert@UGent.be)
Prof. Dr. ir. Inge Van Bogaert
Laboratory of Industrial Biotechnology and Biocatalysis (InBio.be) -
Faculty of Bioscience Engineering -
Ghent University -
Coupure Links 653 -
9000 Ghent - Belgium
www.InBio.be