Final Report Summary - NANOBE (Nano- and microtechnology -based analytical devices for online measurements of bioprocesses)
Executive summary:
There is a growing need for effective monitoring of the micro-organisms and bioprocesses used in the sustainable production of fuels, chemicals and pharmaceuticals. The aim of the NANOBE consortium was to develop a compact analysis tool for reaction monitoring applications in the industrial biotechnology industry.
NANOBE system was designed flexible in terms of what analytes to measure. However, in the NANOBE project the primary example of the application was all along the production of organic acids in yeast saccharomyces cerevisiae. Therefore, lactic acid production in genetically modified s. cerevisiae strain was focused on when testing the whole NANOBE system.
The complete NANOBE system was constructed on the basis of the commonly agreed structure of the system and the functions of the individual modules developed in form of several stand-alone units: sampling; sample treatment and delivery; in-situ monitoring of pH, dissolved oxygen (dO2) and carbon dioxide (pCO2); cell counting, sorting and lysis; enzyme linked immunosorbent assay (ELISA) based protein analysis; and finally capillary electrophoresis with mass spectrometry (CE-MS) to analyse intra- and extracellular metabolites. The planned full setup occupied an area of 2 x 2 m2 including the bioreactor and its peripheries. Most of the required space was occupied by auxiliary components such as pumps, valves or electrical power supplies. The virtual system control software (Labview) controlled the individual analysis and core system modules.
In total, only four analysis cycles were carried out with the integrated system at the very end of the project. This was mainly due to the prolonged analysis cycle of the system and repeating plumbing problems. The monitored dissolved oxygen concentrations were in good agreement with the values measured using conventional sensor. The pH and CO2 sensors worked otherwise well but there were some issues with the optical sensor attachment that can be easily solved in future. The sample was successfully transported from the bioreactor to the ELISA and MS modules in two cycles out of four. One out of the two successful system operation cycles resulted in successful data readout in both the ELISA and MS modules. The results from the ELISA analysis suggested that the dilution was not totally reliable. The CE-MS analysis results were qualitatively in good agreement with data from the manually-taken sample using the same Microsaic mass spectrometer.
The main achievements of the project were:
1. two new analysis tools (micro mass spectrometer already in the market and automated platform for unattended ELISA tests at industrialised prototyping level)
2. new innovative methods and devices developed for dead-volume free µL-scaled sampling and sample handling. The 20 µl sample volume is most probably the smallest sample volume of which so many different analytes can be measured
3. developed digital microfluidics for handling 1 µL droplets that push the boundaries of working with really small sample volume so that the initial sample can be separated for several analytical devices
4. demonstration of tools for counting and sorting of dead and alive yeast cells with added cell concentration estimation
5. new tools developed for lysis of the yeast cells
6. new optical sensor technology developed for autoclavable sensors for pH in the range of three to six
7. more robust optical sensors for measuring dissolved CO2
8. one patent application
9. nineteen scientific journal publications and 15 conference presentations.
Overall, the online monitoring tools developed in the NANOBE project can help to increase the production rate, yield and concentration of the final product of a fermentation process. These improvements in process monitoring may be crucial for the economic viability of a new bio-based product.
Project context and objectives
Project context
Biotechnological processes are increasingly used for the production of chemicals, fuels and materials. The world market for non-food biotechnology-based products was EUR 45 billion in 2006. Advances in genome analysis, systems biology and synthetic biology enable the engineering of cells as efficient hosts for production of a variety of compounds for various industrial sectors. These include bioactive compounds and drugs for pharma industry, fuels such as bioethanol, platform chemicals and biopolymer precursors for the chemical industry, as well as industrial enzymes for industrial applications. These developments are generating an even greater number of micro-organisms and other cells which can be used as 'production organisms'. The expanding number of tailor made production organisms and processes will necessitate better measurement and control of the processes in order to ensure the highest possible productivity.
In an attempt to reduce our dependency on oil, the European Union's (EU) Technology Platform for Sustainable Chemistry emphasizes fuel production using biotechnology and the development of efficient production processes for chemicals. Sustainability means that these bioprocesses need to be highly efficient and use minimal amount of resources. To achieve these goals one needs to understand the performance of the production organism during the process and be able to continuously monitor and control the process itself.
The development of efficient bioprocesses requires rapid, high-throughput analysers. For example, the selection of the most optimal production strains is limited by the absence of integrated, online tools for monitoring of strain performance during the process. Another bottleneck is the lack of rapid high-throughput systems for optimisation of production conditions. Consequently, closed loop feedback control of the production process is rarely implemented. Multi-parameter analyses, particularly for special compounds and intracellular biomarkers, require tedious manual sampling and various very slow off-line analyses.
The increasing number of production organisms and bioprocesses sets demands for fast and versatile product measurement tools for applications ranging from strain screening to large-scale production. Overall, the requirements are to reduce the efforts needed for process development, running time and the cost from the initial idea to the production scale. Even one month shorter development time can be worth of millions of Euros for a pharmaceutical product if it helps to bring the product earlier to the market. Faster process development confers competitiveness. Entering the market place too late is quite often a disaster for the market penetration of a new drug.
In addition, the United States of America (USA) Food and Drug Administration (FDA) process analytical technology (PAT) initiative is about to revolutionize biotechnology-based processes used for the production of pharmaceuticals. The PAT initiative provides for real-time monitoring and control of the state of the production process within a certain ?design space? as opposed to the conventional 'one recipe' strategy. The PAT initiative is driving demand for reliable real-time monitoring tools in bioprocesses.
1. organism development creates a great number of strain variants. Real-time analysis of the performance of the strains will aid in screening of strains and selection of the best production organism in miniaturised or small scale bioreactors.
2. production conditions are optimised using real-time measurement devices in bench-top bioreactors to accelerate scale-up to full production
3. real-time measurement devices can be applied in process control, monitoring and quality control of the production phase
4. overall, the use of the real-time measurement devices has potential to shorten the whole development process by approximately 40 %.
Improved bioprocess control and faster organism development drives a need for a device system which fulfils the following requirements:
1. the ability to take rapid real-time measurements of product and biomarker levels during cultivation, thereby eliminating time-consuming manual analysis. Frequent measurements can be taken without sample storage. In addition, automated data collection allows identification of those time intervals at which selected samples should be stored (e.g. for future quality assurance / quality validation requirements).
2. a versatile analyte range, adjustable to measure various parameters and molecules (including intra-cellular analytes) depending on the process. The analyte range to be measured by the system should be flexible with only changes to the method and without any changes to the system.
3. the ability to collect representative, low-volume samples from large and small scale cultures without endangering the integrity of the cultivation. In addition, all in situ sensing or sampling devices need to be cleanable and sterilisable before and after each sample. Sample volumes will depend on the sensitivity of the tool but should be as small as possible, in particular when sampling from small scale cultures.
4. reproducible performance throughout the duration of long cultivations, which is achieved to a great extent via automation and using ex-situ devices. In-situ devices in particular are vulnerable to fouling and drift as they often cannot be changed or cleaned during cultivations. In addition, ex situ devices are easier to maintain and calibrate during long cultivations.
5. a compact size and a reasonable price - both for the infrastructure and the consumables. Lower costs and smaller size enable higher throughput via multiplication, which is needed especially in the organism screening and selection stages. Smaller systems are cheaper to install and can be sited closer to the process reducing the infrastructure required. However, the size is not the primary criterion of the device development; the overall performance of the system is of greater importance.
Scientific and technical objectives of the project
The aim of the NANOBE project was to construct a versatile tool for the real-time analysis of several compounds and biomarkers in bioreactor processes. The tool was targeted to the control and optimisation of production processes and to acceleration of development of production organisms for applications in industrial biotechnology. Micro and nanofabrication techniques were used to exploit the scaling laws associated with microfluidic devices to reduce analysis time and sample volume. The added value of the NANOBE approach is in the ability to measure certain biomarkers, indicative of cell performance and state.
The NANOBE approach aimed to a flexible platform for real-time measurement of both intra and extra-cellular analytes. One motivation was to avoid the need to develop a custom probe for measurement of each individual analyte. The versatile measurement tool would require only a change in method to enable the measurement of another new analyte. The device was to be designed so that it could be coupled to range of bioreactors, from novel micro-bioreactors to conventional industrial production bioreactors.
Automated sample handling for measurements of intra and extra-cellular analytes was of importance because some key biomarkers exist only inside cells. A general requirement was sensitivity. Sensitivity minimises the amount of sample required. Small sample volumes have the advantage that many samples can be taken from a culture without disturbing the process. The sampling and analysis of individual cells was desirable because we need to record variation in single cell performance, rather than observing the performance of the whole population of cells. In addition, accurate cell counts from a culture were paramount in order to relate productivity to cell biomass.
The application
NANOBE system was designed so that it is flexible in terms of what analytes to measure. In practise this means that different modules of the NANOBE system can be selected and combined for various applications. In the NANOBE project the primary example of the application was all along the production of organic acids in yeast saccharomyces cerevisiae. This related to several publicly funded research projects at the technical research centre of Finland (VTT).
It was seen important to develop the NANOBE system so that it can cope with different analytes, different cell types and different cultivation conditions in the future. However, only one application example was to be focused on when testing the whole NANOBE system. The main application example was lactic acid production in genetically modified s. cerevisiae strain. Strain is based on CEN.PK113-16B. Strain has a lactate dehydrogenase gene ldhL from lactobacillus helveticus on a plasmid. Individual selected analytes are chosen so that they are related to lactic acid production.
The analytes were chosen so that during the cultivation of the strain it is possible to follow how the beginning and the end of the lactic acid production pathway operate at the gene expression, enzyme activity and intracellular metabolite level, in addition to consumption of glucose and production of lactic acid. So far it is unknown for instance how the levels of intracellular pathway components change during different phases of the production process. Measurement of the levels may provide additional information into the cell-level regulation of the lactic acid production and open new possibilities for controlling the process in order to improve the production rate. Lactic acid is the precursor for poly-lactic acid, which is used in the production of biodegradable plastics such as cups and plastic bags.
Project results
Main scientific and technological (S&T) results
Sampling
A significant effort in the NANOBE project was put on developing a sampling system, which allows a time-discrete sampling of dead-volume free µl-scaled samples and a subsequent separation of cells and fermentation broth. This compact sampling and filtration system makes an interface between bioreactors and the rest of the NANOBE measurement tool. The sampling probe needs to be sterilisable. It also needs to fit the industrial standard size. The probe should preferably also be applicable to the sampling of different cell types (e.g. yeast, bacterial, filamentous, plant). In addition, integration of filtration and optical in situ sensing (e.g. in situ sensing for dissolved gaseous analytes) were to be included to create an intelligent probe in a conventional package which will fit bioreactors. The sampling and filtration system supplies the downstream analysis modules with the raw sample for cell counting, the retentate for intracellular substance analysis after cell lysis and the filtrate for extracellular substance analysis.
The sampling probe and the filtration module were realised by means of conventional technologies, whereas the integrated fluid management was partly realised with conventional and with a new on-chip pinching valving/pumping technology. The beneficiaries involved in the sampling probe development were Institut für Bioprosess und Analysenmesstechnik e.V. (iba), in Germany; Department of Microsystems Engineering (IMTEK), University of Freiburg, in Germany; and PreSens Precision Sensing GmbH (PreSens), also in, Germany.
Sampling probe specifications
At the beginning of the project the sampling probe specifications were discussed with all beneficiaries and defined for saccharomyces cerevisiae and streptomyces fermentation.
Dead-volume free sampling
Iba developed a sampling probe that based on a two-capillary concept patented by them. It could be proved for the sampling specifications of saccharomyces cerevisiae and streptomyces fermentation by iba. The sampling probe was designed so that the sample could be embedded into two air plugs, which enabled a time-discrete sampling in µL scale (20 µL of yeast and 50 µL of streptomyces) without dead-volume. The sampling probe was successfully tested for fermentations with overpressure of 3 bar.
In order to withdraw a defined sample volume from a bioreactor operating with a gas flow rate up to 2 vvm, it was necessary to develop a special probe head adapter. This adapter excluded the air bubbles of the fermentation broth from the sampling tube but not from the sensor spots which are additionally located on the probe head. The sampling probe was successfully purged with buffer and compressed air so that repeated sampling of streptomyces over a fermentation period of up to 10 days was no problem.
Filtration in µL-scale
In order to separate the cells and the supernatant from each other a filtration module based on conventional technology was developed by iba. It is thus appropriate for usual sterilisation procedures like sterilisation in an autoclave. Using channel spacers having different heights it is possible to apply the filtration module for taking 20 µL samples as well as for 50 µL samples. It is possible to use filters made of ceramics, polymers or paper. As the filtration element an asymmetric polyethersulfone membrane was chosen for yeast and an asymmetric ceramic filter for streptomyces. A purging procedure was developed to clean the filter between the subsequent samplings. As the module is connected to the bioreactor by means of a sterile barrier downstream the bioreactor, it is also possible to exchange the filter without danger of contaminating the fermentation broth in case the filter is blocked.
Integrated Braille-actuated fluid management
A microfluidic chip was developed as an optional component to replace part of the tubing-based fluidic network of the sampling module. In order to achieve accurate fluid switching and low-volume pumping in nL- and µL-scale with minimal external components, an on-chip valving technique using actuation with a Braille display was developed. In this technique a thin membrane is deformed by the actuation of a Braille pin, resulting in complete blockage of the microchannel.
The microfabrication process of the PDMS fluidic structure with a thin deformable membrane was established by IMTEK. Test experiments for the valving function was successful with water up to flow rates exceeding the range of interest in the NANOBE project, proving the effectiveness of the developed technique for implementation in the sampling module. Additionally, a sequential actuation of multiple valve elements could be successfully used for fluid pumping. The quasi-peristaltic motion of the fluidic structure resulted in a flow rate of up to 6 µL/min.
In the final NANOBE demonstrator the role of the microfluidic chip developed by IMTEK was to perform the splitting of the raw sample in two parts, one for cell counting and the other for filtration and for continuously dilution of the one part of the raw sample to guarantee a fast and accurate mixing. A fluidic structure was designed and tested to intermittently extract small volumes (100 to 500 nL) from a continuously flowing stream. Also, a dilution structure was developed utilising Braille-actuated active mixing of two continuously flowing streams.
In situ analysis of pH, pCO2 and pO2
A further challenge for the sampling probe development was the in situ analysis of pH, pCO2 and pO2. As the existing pH sensor was not suitable for yeast culture due to its limited dynamic range of 5.5 to 8.5 PreSens had to develop a new pH sensor to cover the more acidic pH range from 3 to 5.5. For pCO2 a sensor was developed which covers typical dynamic ranges (1 to 25 %) for yeast fermentations. The pre-existing dO sensor was suited also for this type of fermentations, dynamic range 0.02 to 100 % oxygen.
Sensors for all three analytes were never applied in the same fermentation before. Therefore a common sterilisation method for all three sensor types had to be developed by PreSens. For the first time an autoclavable pH sensors was demonstrated. The autoclaving conditions were determined by the pCO2 sensor which was more critical than the other two sensors. Autoclaving in the absence of oxygen is required. Many cleaning procedures were investigated but none was compatible with all three sensors. Therefore a disposable concept had to be developed. The spots were integrated onto polycarbonate discs which could be mounted to the probe head and exchanged after the fermentation.
Integrated sensor and sampling probe
A functional conventional sampling probe for a standard port of 12 mm diameter with three integrated optical sensors for gas/pH monitoring were constructed and manufactured by iba. This is reducing the amount of ports needed in the bioreactor, saves time during assembly and reduces the risk of contamination by reducing amount of components. Optical fibres used to read the optical sensor spots have to be removed before autoclaving and can be easily integrated in the sampling probe after autoclaving.
The construction of the sampling probe was optimised for a usage with any bioreactor independently of the position of the agitation motor by separating the sterile and overpressure barrier from the probe. A tube squeezing valve protects the further downstream modules from high pressure inside the bioreactor during cultivation and from contamination. The sampling and filtration system can be easily coupled sterile with the sampling probe via this valve. The newly developed probe-head adapter guarantees bubble-free sampling from the fermentation broth without any influence on the accuracy of the sensor measurements. As the raw sample is embedded into air plugs, samples can actually be withdrawn without dead-volume. The volume of the sample can be easily adapted from 10 µL to a few hundreds of microlitres if necessary.
Testing of the sampling probe in a lab scale process
The sampling probe with its integrated optical sensors was tested in a lab scale process of yeast at iba regarding to its functionality, biofouling effects and reproducibility in comparison to a conventional tubular probe as sampling system.
The sampling procedure in the NANOBE setup lasts approximately six and a half minutes. The volume of the sample was determined by measuring the sample length in the tube directly after withdrawing and weighting the sample after pumping the sample through the tube squeesing valve. With both methods a good reproducibility of sample volume was determined.
The cell counting was performed with repeat determination by means of a Neubauer counting chamber. During four days of fermentation the cell count was determined. During the first 28 hours the cell counts of both sampling probes were almost the same. In the next days the differences of the cell counts of both probes first increased, then adjusted to each other before the differences increased again (to 15.5 % on average). Therefore it was not a cumulative effect. The observed sample volume increase of 3 % within the tube cannot result in such an increase or dilution of the sample, which would explain the differences in the cell counting between the two sampling probes. But it indicates an attachment at the inner sides of the tubings. Adhesion and biofouling is always a topic in microfluidic devices, but only some adhesion on the two-capillary probe was observed. Additionally, the different tubing materials and the tubing diameters together with the flow rates of the samples could influence the sample content. However, the cell counting method itself showed a high standard deviation, so that the differences are only speculative. Only further investigations could finally explain it.
Cell handling tools
In the NANOBE project the cell handling tools were developed to provide information about the concentration and physiological state of the production organism and to perform cell lysis as well as sample handling. Both Ecole Polytechnique Fédérale de Lausanne (EPFL), in Switzerland and Centre National de la Recherche Scientifique (CNRS), situated in Lille, France, participated in the development of the cell handling tools.
Cell sorting and counting
Cell sorting and counting devices were developed by EPFL. These devices were designated for processing of samples with cells extracted from the bioreactor by means of the sample probe. The main components are cell sorting by dielectrophoresis and cell counting by impedance measurements. The sorting and counting module allows the separation of living and dead yeast cells and the counting of the separated populations. The viability of the sample is obtained by dividing the number of living cells counted by the total number of counted cells. The cell density is obtained by dividing the number of counted cells by the sample volume. After they are sorted, the cells flow through microelectrodes arranged in coulter counter configuration.
The connections of the sorting/counting chip were adapted to be able to connect the device to the other NANOBE subsystems with tubing. The tubing of small inner diameter was used and the tubings were arranged so that the sedimentation at the inlet and outlet of the chip and the dead volumes were reduced.
The remaining dead volume of a few microliters still represented a long transit time at the speed of the sorting system. Therefore, EPFL developed a pressure regulated as well as a switchable bypass which allowed the sample to pass faster through the tubing and only a fraction of the sample was analysed with the sorting/counting chip. The switchable bypass was successfully tested with a sample of yeast cells and suitable values of the order of 1 µl/min for the flow entering the sorting/counting system.
During the last year of the project the cell analyser module was reworked to use impedance spectroscopy instead of cell sorting, in order to reduce complexity and improve reliability. Parameters for counting as well as density and viability measurements were adapted for use with NANOBE in order to enable the use of higher flow rates. The result was good, the new system allowed for considerably higher flow rates of around 1 µl/min without use of a bypass.
Cell lysis
Cell lysis is used to break the cell membrane in order to extract intracellular components for analysis. Both chemical and electrical lysis have been investigated in the NANOBE project, although electrical lysis is preferred because components present in the chemical lysing solution could lead to fouling of the downstream analysis. Cell lysis was investigated by both EPFL and CNRS. EPFL focused on electrical lysis whereas CNRS worked both on electrical and chemical lysis.
Chemical lysis
CNRS studied the chemical lysis of Saccharomyces sereviciae yeast cells. First, CNRS determined with cellytic from Sigma Aldrich if chemical lysis would be compatible with the technology used in the sample transceiver device and with the NANOBE specifications.
Electrical lysis
Electrical lysis makes use of high electrical fields to disrupt the cell membrane. Higher electrical fields yield better lysis efficiency, but can also lead to the formation of bubbles at the microelectrode surface. An alternating voltage at a frequency of 50 kHz was used by EPFL to reduce this effect while still achieving efficient cell lysis. Specific microchannel designs can be optimised to locally concentrate the electrical fields in the lysis regions.
These three designs have been evaluated by EPFL using red blood cells, as the lysis of those cells is easily observable optically by the diffusion of haemoglobin out of the cell. The third design with so-called liquid electrodes was chosen for the NANOBE project. In this design the current density and thus the risk of electrochemical reactions is reduced. In addition the electrical field is uniform on the side of the main channel and the length of the lysis region and the transit time of the cells is largest of these three options without reducing the electrical field. This leads to better lysis efficiency and higher throughput. Two different flow cytometry methods were used to evaluate the lysis efficiency with yeast cells: cell sorting by dielectrophoresis and multiple-frequency impedance measurements. The measurements indicated that the intracellular components were successfully extracted from the cells. However, an effective lysis of the cells at higher flow rates required the application of higher voltages (100 V) which resulted in bubble creation at the electrodes and clogging of the channel. The long term operation of the device was thus impossible and the device was not suitable for the NANOBE project purposes.
To overcome the bubbling and clogging problems, EPFL adopted a new technology based on three-dimensional (3D) carbon electrodes. EPFL demonstrated high lysis efficiency (90 %) with high flow rate (up to 50 ul/min) while the lower flow rates (below 20 µl/min) tended to show slow bubble creation and cell sedimentation inside the microchannel. Despite of the good lysis results with optimised operation the EPFL device was not well compatible with the NANOBE system specification and suffered reliability problems.
CNRS had developed parallel electrical lysis technologies in the NANOBE project. The lysis device was made of conductive silicon layer sandwiched between two isolating glass slides. These 3D electrodes allow constant electrical field along the full channel height. The chip was powered by a high alternating current (AC) voltage to enable disruption of the tough yeast cell envelope without bubble formation. The CNRS lysis chip was selected to be implemented to the integrated NANOBE analysis system.
Filtration
Three filtration systems were required in the NANOBE workflow:
1. a filter after cell lysis, removing the cell debris from the lysate
2. a filter before EPFL density/viability chip to remove dust but letting cells flow through
3. a micro-debubbler to avoid air entering into the density/viability chip.
EPFL tested commercial filters for the removal of cell lysis debris as well as for protection of the cell counter chip. The commercial inline filters from IDEX were found to be suited for the removing of the cell debris from the lysate (filter pore size 1 µm) and for the removal of dust (filter pore size 10 µm). Moreover, a micro-debubbler was developed and tested to be put just in front of the cell counter. The micro-debubbler was found to be suited for the removal of the air between samples, as well as smaller bubbles.
Sample transceiver
A special sample transceiver was developed by CNRS to perform the delivery of the supernatant or lysed cell samples delivered by the sample probe to the individual analysis modules.
The sample transceiver module uses electrowetting on dielectric (EWOD) technology to manipulate and deliver samples to the analysis system. The liquids are provided and delivered to/from the sample transceiver through micro-channels. In the sample transceiver the incoming liquid is digitised to small droplets (volume 1 µl) and all the operations in the sample transceiver are made by moving droplets in the central area of the device. Various bio-chemical operations can be done in digital fluidic mode. At the output end the droplets are combined to specified sample reservoir area from where they can be pumped onwards to the analysis tools downstream the NANOBE system via capillary connections. The functionalities available in the sample transceiver are droplet generation, sample mixing/diluting and sample exportation.
The sample transceiver was tested with cell lysate, supernatants corresponding to different cells cultures, the buffer used for CE analysis and PBS with different concentrations. All these fluids were to be used in the NANOBE application. Based on these tests CNRS defined the conditions on these fluids to be manipulated by electrowetting in the sample transceiver. The surface tension and bio contents of the droplets were the key parameters to enable the manipulation of these liquids. It was observed that too low surface tension or too high bio-concentration of the droplets made the droplet displacement difficult or even impossible. The results showed that the movement feasibility of the supernatant in yeast nitrogen base (YNB) is better than the one in yeast extract peptone-dextrose (YPD). For 'cell in buffer', the two solutions both need to be worked by 100 time dilution. The results confirmed that with a moderate dilution with water it was possible to handle all the fluids needed for NANOBE application.
Micro-ELISA analysis of proteins
The NANOBE system has two analysis paths for intra- and extracellular compounds. For the first one, DiagnoSwiss has adapted their micro-ELISA technology to automated immunoassays and to online process monitoring of extracellular product protein concentrations. Furthermore, DiagnoSwiss has evaluated the feasibility of mRNA level determination using magnetic beads and electrochemical detection in their ImmuSpeed platform.
Coupling with the sample transceiver
The specifications for the coupling of the micro-ELISA platform to the sample transceiver were set according to the fluidic layout of the integrated system, the unit operations of the sample transceiver and of the micro-Elisa platform and the related timings of the analysis cycles. For the immunoassays, the total analysis cycle was set to 40 minutes, with the sample being sent from the transceiver to the ELISA module every 20 minutes, but with two samples in duplicates being analysed simultaneously with a 40 minute cycle.
The coupling system consisted of a peristaltic pump connected to the sample transceiver in order to transfer the samples in wells of a sample collector and placed in the micro-ELISA platform. The robotic head of the micro-ELISA platform will then transfer the samples into the microtiter-plate placed in the instrument for incubation with the capture antibody and then further analysis by electrochemical detection in the microchip.
Adaptation for the messenger ribonucleic acid (mRNA) analysis
In the NANOBE system, investigations were made to evaluate whether expressed mRNA levels of the selected genes can be measured using DiagnoSwiss' ImmuSpeed device developed primarily for sandwich assay of proteins and see whether it could be adapted also for measuring the expression of mRNA molecules.
These studies demonstrated the feasibility of transposing the TRAC methodology to the electrochemical micro-chip format and that a reliable assay protocol could be developed for HXK1 mRNA detection. However, the obtained results were only semi-quantitative and, in its current state, the assay performances did not show sufficient sensitivity and reproducibility for direct application to online monitoring of yeast cultures. Further efforts should be placed in order to investigate whether the assay performances can be further improved to reach the very low detection limit required for real applications.
CE separation of analytes
The second analysis path of the NANOBE real-time monitoring tool was a coupled CE-MS system for measurement of small molecules, such as substrates, metabolic products and intermediate biomarker metabolites. Originally the CE was planned to be applied in multiple ways in the NANOBE project to analyse metabolic compounds, nucleic acids and fluorescent cells. It could also be used for the detection of intracellular enzyme activities. In yeast fermentation the most important analytes are extracellular sugars and organic acids. In streptomyces spp fermentation the most important analytes are polyketides rapamycine, nigericine and/or tacrolimus depending on the strain. The design of the online analysis system was such that the (above mentioned) analytes in the cultivation broth can be quantified in concentration range 100 mg/L to 100 g/L. The required sample volume from the sample transceiver to the CE is 1 µl.
During the project VTT designed, fabricated and analysed CE chips with and without MS connectivity option. Laser induced fluorescence (LIF) was used as a detection method in the case of CE chips without the MS option while the MS was naturally the choice for detection in the case of other chip design.
MS detection of analytes
Microsaic developed further its micro-fabricated mass spectrometer technology to be integrated with VTT's chip based CE technology. The objective was to develop the first fully micro-fabricated CE-MS system and to provide online detection and definitive identification of analytes separated by the CE module. In the NANOBE system MS detection is used to analyse extracellular substrates and metabolites. A stand-alone 'Spraychip nanospray' module and controller was designed, built and tested by Microsaic. The development of negative ionisation mode was completed before delivering the module to VTT for testing. VTT successfully completed initial trials of the nanospray unit interfaced to one of their commercial mass spectrometers. The Spraychip nanospray module was interfaced to a chip scale breadboard atmospheric interface mass spectrometer developed by Microsaic and the functionality of the integrated system was successfully demonstrated.
Sample capture for storage and off-line analysis
A sample capture for off-line surface assisted laser desorption/ionisation - desorption ionisation on silicon surface (SALDI-DIOS) analysis was developed at CNRS. In NANOBE, the DIOS chip was planned to be used to store and analyze off-line both extra- and intracellular metabolites and products in saccharomyces cerevisiae and streptomyces fermentations. In yeast fermentation the most important analytes are extracellular sugars. In streptomyces fermentation the most important analytes are rapamycine, nigericine and/or tacrolimus depending on the strain. Limit of detection of these compounds is in the range of nmol-pmol/µL. CNRS evaluated possible substrate candidates for the DIOS analysis. CNRS evaluated also the ion affinity chromatography (IMAC) technique for functionalisation of the nanowires. The superomniphobic surfaces developed by CNRS allow handling of droplets having a low surface tension due to either surfactant or some biomolecules. It was shown that depending on their morphology, the silicon nanowires offered liquid-repellent character with very good robustness over a wide range of surface energy.
Towards the end of the project CNRS studied the interaction of graphene oxide and reduced graphene oxide with bovine serum albumin (BSA). The results suggested that graphene oxide has a high capacity for protein adsorption. This property was finally exploited for the displacement of BSA using EWOD with minimum non-specific adsorption. The SALDI-DIOS was not part of the final system demonstrator.
System integration
Finally, the complete NANOBE system was developed in form of several stand-alone units that were physically integrated to a complete system. All partners were involved in defining the detailed specifications of all modules and their interfaces while IMTEK was responsible for the collection and management of the system specifications. The NANOBE system was constructed on the basis of the commonly agreed structure of the system and the function of the individual modules.
The following modifications to the complete NANOBE system had to be made based on the results of the submodule interface testing:
1. due to software communication problems with the impedance spectrometer and some issues with the sampling probe, no cell viability measurements could be performed simultaneously with the other components of the system
2. the high-pressure resistance of the silica capillary connections of the cell lysis chip and the limited pressure tolerance of the microvalves used for flow switching prevented the use of the cell lysis chip in the online monitoring system for the final testing. This is a result of a long capillary connection between the modules that was not planned at the beginning of the project.
3. the supernatant dilution ratio was necessary to be increased to 1:50 from the original 1:10 to avoid the damaging of the sample transceiver electrodes.
Testing of the NANOBE system
The genetically modified yeast cell strain of VTT was cultivated in the bioreactor and used for interface development and integrated system testing purposes. During the testing the pH and oxygen concentration was monitored with conventional sensors for comparison with the PreSens optical sensors. Cell samples were manually sampled with a syringe at the same time as the automated sampling at the IBA sampling probe. Off-line analysis of the manually-taken samples were performed using the CE-MS and ELISA modules and with off-line high performance liquid chromatography (HPLC) for comparison.
Integrated NANOBE system
Despite the original plan of completing the online analysis cycle in 25 minutes with a 20 minute cycle, the finally tested analysis cycle turned out to require 80 minutes. In the course of the timing and flow rate adjustments during the interface development, all process steps at which the sample or washing liquid is delivered into external microfluidic chips needed to be prolonged considerably. The resulting online analysis sequence was such that the sample was delivered to the transceiver approximately 45 minutes after automated sampling. The time dedicated to the sample preparation in the sample transceiver was around 6 minutes. The CE-MS analysis and ELISA measurement of the supernatant would finish at 60 minutes and 100 minutes after sampling, respectively. The diluted raw sample reached the cell counting chip approximately 15 minutes after sampling, although the impedance measurements and viability analysis were not performed online.
In total, only four analysis cycles were carried out with the integrated system. There were two cycles in which the sample was successfully transported from the bioreactor to the ELISA and MS modules. One out of the two successful system operation cycles resulted in successful data readout in both the ELISA and MS modules. The results from the ELISA analysis suggest that the diluted supernatant from iba sampling/filtration module may have been more concentrated than the theoretical 1:50. If this could be confirmed, it can explain the difficulty to transport the sample in the sample transceiver. The CE-MS analysis results were qualitatively in good agreement with data from the manually-taken sample using the same Microsaic mass spectrometer. The monitored dissolved oxygen concentrations were in good agreement with the values measured using conventional sensor. The oxygen monitoring continued for two days without major drift.
The failure modes in the three unsuccessful runs were all different. In cycle one, the sample reached the inlet of the mass spectrometer but did not reach the detector due to clogging of the vacuum interface. The ELISA module also failed to proceed beyond the sample/bead conjugate incubation due to software failure on the master PC side. In cycle two, several electrodes on the sample transceiver were damaged due to repeated use from interface development to sample testing. In cycle four, an unexpectedly viscous sample was transported from the iba sampling/filtration module to the sample transceiver which indicates that the supernatant dilution of 1:50 ratio was not achieved, most probably because of air bubble generation at one fluidic junction. This is also an explanation for the discrepancy between the IgG concentration in the online sample and manually-taken sample in the ELISA measurements.
In summary, the complexicity of the whole NANOBE system caused several irregularities in the sample transport, which could not be finally eliminated. The reason for this is that the development of all modules had to be finished before the interfaces could be developed. It was tried to adjust the interfaces with the state of development of the modules, but this was time consuming and expensive, so that it could not always be done.
Testing of an alternative integrated system configuration
An alternative online operation of the integrated system was also tested. In this configuration, the CE module was directly connected to the supernatant outlet of the sampling/filtration module. This is one variant of the system for applications that do not require ELISA measurements. Since there is no need of sample distribution to multiple measurement paths, the sample transceiver is also omitted.
The outcome of the alternative configuration test was similar to that in the full system test; one set of measurement data was obtained in three test cycles. One failure mode was the clogging at the vacuum interface of the mass spectrometer and the other was leakage at one fluidic connection between the sampling/filtration module and the CE module.
Results on IgG determination
For the two samples delivered by the sample transceiver, only one sample resulted in an electrochemical signal corresponding to an enzymatic reaction. The second signal gave a negative slope corresponding to a blank point. The difference between the spiked concentration and the measured concentrations in the undiluted and diluted samples may be explained by a partial decomposition of the IgG molecules during the fermentation, by a loss of IgG during the filtration of the medium, by dilution errors and by artefacts between the signals obtained for calibration points measured in buffer solution and the real samples. Unfortunately, the lack of experimental data does not allow one to understand the discrepancy between the expected and the measured IgG concentrations.
These limited number of data points prevents to identify clearly the source of errors, to qualify the real performance of the developed assay and to benchmark the results against measurements that could be obtained with the same samples using conventional microtiter-plates with optical detection.
CE-MS analysis of extracellular metabolites
During the testing of the integrated NANOBE system, altogether five samples out of eight were successfully analysed online with the CE-MS system while being part of the whole NANOBE system. In addition, all of the eight samples were analysed off-line as well. Absolute quantification was not possible for three reasons:
1. the optimisation of the CE method was not complete
2. dilution of the sample upstream in the Nanobe system was not accurate eventually
3. ionisation at the MS varied resulting in large uncertainties.
Furthermore, the delivery of the sample from the bioreactor was too slow. Quantification problems two and three could have been and can be solved by using internal standard in the cultivation. Normally internal standard is not used directly in the cultivation as it may affect the performance of the cells or it may be consumed by the cells.
Online ethanol extraction and CE-MS analysis of intracellular metabolites
The analysis of intracellular metabolites was not possible with the NANOBE setup, because the time in the NANOBE setup for the treatment and lysis of the cells would have been too long and the metabolism of the cells would have had time to adapt to the conditions in the sampling system and would not have reflected the status of the cells in the reactor. A separate setup was built to demonstrate that the analysis of intracellular metabolites was possible with the online CE-MS device.
Several difficulties occurred during these tests, vac chips in the MS got clogged and there was also some pressure build-up caused by the narrow tubing. It was also noticed that the m/z values varied between analyses in the mass spectrometer which made identification of the peaks difficult.
In the future research of the CE-MS, the CE method development will be continued. Higher CE voltages and temperature control would increase sensitivity and make the CE peaks sharper. Non-aqueous background electrolytes (BGEs) would help with the ionisation of analytes in electrospray and might also help to prevent the clogging of vac chips. The use of internal standard would be essential for the quantification since there is always variation in the MS signal due to day-to-day variation in the efficiency of ionisation.
Overall the project succeeded to connect the MS detection developed by Microsaic to the CE separation by VTT and this CE-MS system to the rest of the NANOBE system. Furthermore, VTT was able to develop a method for online ethanol extraction in order to enable online analysis of intracellular metabolites in connection of the CE-MS system. Measurement of extracellular and intracellular metabolites in connection to bioreactor cultivations online was demonstrated for the first time.
However, the methods are not yet quantitative and further work is required. Problems encountered in flowing the electrophoretic solvents used in the CE directly to MS are a major issue that needs to be solved by developing the CE system and the method as well as the ionisation.
Microsaic's MS as instrument was very easy to operate; one of the most advanced features of the device was the software. Compared to conventional mass spectrometers changing parts was much easier and faster. Also, the vacuum was reached very rapidly, in just 30 minutes. The electrospray of this device was very promising, because it was able to produce stabile electrospray despite very low liquid flow rates of CE system.
Cell analyser tests with VTT's own sampling system
As the tests of the cell counter/analyser chip were not successful in the integrated NANOBE system, VTT set up a simpler system for connecting the EPFL chip to a bioreactor. PBS flow was maintained steady using a pressurised and pressure-controlled vessel.
The developed impedance analysis system can efficiently discriminate living and dead yeast cells and determine the cell density from the average cell speed. The system can function with the sampling probe and gives good results with yeasts. Unfortunately the combination of different problems starting from existing commercial instruments led to the fact that there was no time for producing results to really demonstrate that the cell and viability measurement chip works in the NANOBE system. Results outside the NANOBE system have been presented and published and show that the technology works. When connected to a bioreactor with a simpler system some measurements for cell density and viability directly from the cultivation could be done. However, there is great promise with this technology as cell density and viability measurements are the top priority for measuring in bioprocesses. Tests of the optical probes were also carried out.
Conclusions
The NANOBE integrated system was constructed only a couple of months before the end of the project, occupying one laboratory in VTT. All hardware was connected with each other as planned, with the exception of the cell lysis chip. The resulting system created a long pathway (to some system parts such as the EWOD, which reduced the timing precision and made inclusion of the lysis chip impossible. The fully integrated system was operated for online handling and analysis of genetically modified yeast fermentation.
After a significant amount of software modifications, the fluidic interfacing between the individual modules functioned successfully with acceptable repeatability of the transported sample volume. However, due to the long tubings the timing of fluidic transport to and from the sample transceiver saw non-negligible deviations, which made fully automated operation of the entire system impractical. Therefore, triggering of the module operation was performed manually although automated triggering was implemented in the master program of the system.
Online sample handling and analysis using the integrated system functioned flawlessly in one run out of four attempts. Failure modes in the unsuccessful runs were all different, appearing in different sections of the system. This means that the robustness of the modules have to be increased.
Comparisons of the measured mass spectrum from the online and reference offline analyses indicated that the sample has been successfully transported through the entire system without major contamination from the various splitting, mixing and cleaning steps during the automated operation. The ELISA readouts of spiked IgG concentrations showed a significant deviation from the offline measurement, but the two were on the same order of magnitude. Considering that the dilution ratio of the supernatant sample in the sampling/filtration module is yet to be quantitatively evaluated, the outcome is encouraging for further development of the system.
Overall, the NANOBE concept of an automated online bioanalysis system based on microfluidic core components was proven to be promising in principle. It was shown that a minute amount of sample with a volume of 20 µL can be analysed online in two paths without any manual handling. Two different configurations of the system components were tested, both showing results that are comparable to conventional off-line methods, which was a good demonstration of the modular interchangeability of the system. Improvements in robustness and accuracy of sample handling are keys for achieving constant operation for online monitoring over the full duration of fermentations.
Potential impact
The NANOBE project delivered an integrated real time bioprocess analysis system for measuring of extra- and intracellular compounds. It provided a system for bioprocess monitoring that was not existent at the time when the project was started. The NANOBE system is well beyond the state of the art in its concept and individual components. Smart systems integration led to increased functionality, since the sampling is now solved in a way which enables the real online analytics during the bioprocess. Different lab-on-a-chip modules enable measurement of various analyte types such as cells, low molecular weight compounds, proteins and specific mRNAs. The system provides improvement in terms of automation, analysis time and sensitivity. Further development is needed to increase the robustness of the integrated NANOBE system. The analysis platform can significantly improve real-time feedback control of large-scale production processes as well as screening and optimisation of production organisms and conditions. The project combined in an innovative manner special expertise in microfluidics, nano- and microfabrication techniques, chip-scale mass spectrometry, photonics, electronics, sensor technologies and biotechnology.
There is an increasing need for more rigorous bioprocess monitoring due to the increasing number of tailor-made production organisms and bioprocesses that will be required for sustainable production of fuels, chemicals and pharmaceuticals. Development of novel production organisms is benefited from the recent advances in genomics, metabolic engineering, synthetic biology and systems biology, which in turn provide tools and understanding to design analysis methods for those cellular and process parameters and biomarkers that are critical for the most efficient performance of the process. The increasing number of non-traditional bioprocesses with high demands on energy efficiency and yields sets a demand for multiparameter analyses that also enable better feedback control. In addition, using the NANOBE platform, right time points for sample collection for more comprehensive off-line analyses can be chosen. With the NANOBE system also individual cells can be analysed. This provides means to count non-viable and nonproductive cells in the population. On the other hand cells with intracellular biomarkers that indicate cell stress etc. can be counted. The biomarkers can also be used for cell sorting and for instance selection of cells with best viability and productivity for further cultivations. It is also noteworthy that the technologies developed allow synchronisation of cells. This is of fundamental importance since until now the great majority of biological data has been generated from populations where the cells are in different points of the cell cycle and have different physiology. Thus, it is to be noted that the NANOBE analysis platform is not only useful for monitoring and control of production processes but is an excellent tool for strain screening and process optimisation, as well as for basic studies on cell physiology and cellular phenomena including systems biology analyses. Overall, the online monitoring tools developed in the NANOBE project can help to increase the production rate, yield and concentration of the final product of a fermentation process. These improvements in process monitoring may be crucial for the economic viability of a new bio-based product.
NANOBE platform offers reduced process development times for existing bioprocesses. This will enable the cost reduction e.g. in pharmaceutical industry. Firstly, time is saved for custom manufacturing companies in their process development services. Secondly, the time that it takes for a drug to reach the market will be reduced. Thirdly, time savings in product development also confers competitiveness. Also e.g. product purification and process characterisation at small scale prior to production of large scale GMP batches can be done much quicker by use of online real time assays, which will be implemented and tested in this NANOBE project. The FDA PAT Initiative is demanding a higher amount of analytical methods during the bioprocess similar to the analytics in the downstream processing. The NANOBE project is directly addressed this challenge by providing a real time and on line measurement platform for multiple analytes.
A further cost effective point is the possibility of miniaturisation, which results in a higher parallelisation of bioprocess approaches. Furthermore, some modules need less space as the known systems; e.g. MS and the combination of gas/pH sensors with the sampling probe. Last one results in saving ports for other analytics in the bioreactor, saving time during assembly and reduces the risk of contamination by reducing the amount of components.
Bioreactor is an essential tool for bioprocess research and development. It can be used for wide range of applications varying from general process optimisation to gene mapping tasks. This fact also leads up to a wide range of different type bioreactor designs suitable for different applications. The real time measurement platform developed in the NANOBE project can be realised in a way that it will suit different types of bioreactors and different types of cultivations. The experiences and conclusions obtained during the NANOBE project are of general value for bioprocess monitoring and very useful for other research areas like cell and tissue engineering, lab-on a chip development and standardisation.
Although the measurement platform developed in the NANOBE project was tested only with a certain yeast based bioprocess, the perspective still is to apply it to monitor also cultivations based on different organisms in the future. The extracellular analyses developed in NANOBE are very generic and can be used with most organisms. The complexity of the culture medium may however cause adjustment in the device cleaning and analysis methods.
The main dissemination activities and exploitation of results
Dissemination activities
In the NANOBE project the dissemination of knowledge was done in multiple ways. An essential part of the dissemination of the results of the project results was to publish them in major international scientific journals and conferences. Numerous scientific journal papers have been published, accepted for publication or submitted by the end of the project. In addition several conference presentations were held in international conferences and workshops.
An Industrial platform (IPF) was established in the beginning of the project. It comprised of representatives of application and equipment manufacturing companies. The members of the IPF had an access to monitor the progress of the project. Three workshops were arranged to disseminate the project objectives and results to the IPF members. The material presented to the IPF members have been distributed to all members by e-mail.
A project website was set up in the beginning of the project. The public extranet can be found in http://www.vtt.fi/NANOBE
Altogether, the NANOBE project produced 19 reviewed journal publications, 15 conference presentations and 17 other presentations to disseminate the project results.
Exploitation of results
Both general advancement of knowledge and commercially exploitable research and development results have been generated in the project. General advancement of knowledge relates to the developed technologies, devices and fabrication methods. Significant progress has been made in understanding biopollution and understanding of the interaction between small biomolecules in a liquid phase and nanotextured solid surfaces. Generated knowledge will be exploited in scientific research and development for various biological, chemical, environmental and life science applications by the beneficiaries. Generated new knowledge will be integrated also into the student education.
One patent application was filed in the project. With the new patent with the topic 'Sampling probe for withdrawing samples from aerated liquids' iba shows that the sampling probe consists of a wide application range.
The developed sampling port combining three analytes and sampling is believed to gain interest even without the attached analytical devices of NANOBE. The consortium hopes that positive feedback from the IPF and from other industries will arise in the near future and lead therefore to a commercialisation of this part of NANOBE. PreSens, DiagnoSwiss and Microsaic Systems will also exploit commercially those results of the research and development work that are the most mature. PreSens will profit from developments on the individual sensors for dO, dCO2 and pH. Diagnoswiss will use the prototype of automated microchip platform for further demonstration of immunoassay applications. Microsaic Systems has incorporated the technology developed during the NANOBE project into a new product, the 3 500 MiD. This miniature mass spectrometer was launched at the Lab Automation show in January 2011.
The end user companies have shown interest especially on the single component developments in the NANOBE project. These include the sampling probe, mass spectrometer, mRNA analysis and optical sensors.
Overall, the online monitoring tools developed in the NANOBE project could help to increase the production rate, yield and concentration of the final product of a fermentation process. These improvements in process monitoring may be crucial for the economic viability of a new bio-based product.
Project website: http://www.vtt.fi/NANOBE
Coordinator
Päivi Heimala, Senior Scientist, VTT Technical research center of Finland, PO Box 1000, FI-02044 VTT, Finland
Telephone: +35-820-7226637
Mobile: +35-840-5078004
Fax: +35-820-7227012
E-mail: paivi.heimala@vtt.fi
There is a growing need for effective monitoring of the micro-organisms and bioprocesses used in the sustainable production of fuels, chemicals and pharmaceuticals. The aim of the NANOBE consortium was to develop a compact analysis tool for reaction monitoring applications in the industrial biotechnology industry.
NANOBE system was designed flexible in terms of what analytes to measure. However, in the NANOBE project the primary example of the application was all along the production of organic acids in yeast saccharomyces cerevisiae. Therefore, lactic acid production in genetically modified s. cerevisiae strain was focused on when testing the whole NANOBE system.
The complete NANOBE system was constructed on the basis of the commonly agreed structure of the system and the functions of the individual modules developed in form of several stand-alone units: sampling; sample treatment and delivery; in-situ monitoring of pH, dissolved oxygen (dO2) and carbon dioxide (pCO2); cell counting, sorting and lysis; enzyme linked immunosorbent assay (ELISA) based protein analysis; and finally capillary electrophoresis with mass spectrometry (CE-MS) to analyse intra- and extracellular metabolites. The planned full setup occupied an area of 2 x 2 m2 including the bioreactor and its peripheries. Most of the required space was occupied by auxiliary components such as pumps, valves or electrical power supplies. The virtual system control software (Labview) controlled the individual analysis and core system modules.
In total, only four analysis cycles were carried out with the integrated system at the very end of the project. This was mainly due to the prolonged analysis cycle of the system and repeating plumbing problems. The monitored dissolved oxygen concentrations were in good agreement with the values measured using conventional sensor. The pH and CO2 sensors worked otherwise well but there were some issues with the optical sensor attachment that can be easily solved in future. The sample was successfully transported from the bioreactor to the ELISA and MS modules in two cycles out of four. One out of the two successful system operation cycles resulted in successful data readout in both the ELISA and MS modules. The results from the ELISA analysis suggested that the dilution was not totally reliable. The CE-MS analysis results were qualitatively in good agreement with data from the manually-taken sample using the same Microsaic mass spectrometer.
The main achievements of the project were:
1. two new analysis tools (micro mass spectrometer already in the market and automated platform for unattended ELISA tests at industrialised prototyping level)
2. new innovative methods and devices developed for dead-volume free µL-scaled sampling and sample handling. The 20 µl sample volume is most probably the smallest sample volume of which so many different analytes can be measured
3. developed digital microfluidics for handling 1 µL droplets that push the boundaries of working with really small sample volume so that the initial sample can be separated for several analytical devices
4. demonstration of tools for counting and sorting of dead and alive yeast cells with added cell concentration estimation
5. new tools developed for lysis of the yeast cells
6. new optical sensor technology developed for autoclavable sensors for pH in the range of three to six
7. more robust optical sensors for measuring dissolved CO2
8. one patent application
9. nineteen scientific journal publications and 15 conference presentations.
Overall, the online monitoring tools developed in the NANOBE project can help to increase the production rate, yield and concentration of the final product of a fermentation process. These improvements in process monitoring may be crucial for the economic viability of a new bio-based product.
Project context and objectives
Project context
Biotechnological processes are increasingly used for the production of chemicals, fuels and materials. The world market for non-food biotechnology-based products was EUR 45 billion in 2006. Advances in genome analysis, systems biology and synthetic biology enable the engineering of cells as efficient hosts for production of a variety of compounds for various industrial sectors. These include bioactive compounds and drugs for pharma industry, fuels such as bioethanol, platform chemicals and biopolymer precursors for the chemical industry, as well as industrial enzymes for industrial applications. These developments are generating an even greater number of micro-organisms and other cells which can be used as 'production organisms'. The expanding number of tailor made production organisms and processes will necessitate better measurement and control of the processes in order to ensure the highest possible productivity.
In an attempt to reduce our dependency on oil, the European Union's (EU) Technology Platform for Sustainable Chemistry emphasizes fuel production using biotechnology and the development of efficient production processes for chemicals. Sustainability means that these bioprocesses need to be highly efficient and use minimal amount of resources. To achieve these goals one needs to understand the performance of the production organism during the process and be able to continuously monitor and control the process itself.
The development of efficient bioprocesses requires rapid, high-throughput analysers. For example, the selection of the most optimal production strains is limited by the absence of integrated, online tools for monitoring of strain performance during the process. Another bottleneck is the lack of rapid high-throughput systems for optimisation of production conditions. Consequently, closed loop feedback control of the production process is rarely implemented. Multi-parameter analyses, particularly for special compounds and intracellular biomarkers, require tedious manual sampling and various very slow off-line analyses.
The increasing number of production organisms and bioprocesses sets demands for fast and versatile product measurement tools for applications ranging from strain screening to large-scale production. Overall, the requirements are to reduce the efforts needed for process development, running time and the cost from the initial idea to the production scale. Even one month shorter development time can be worth of millions of Euros for a pharmaceutical product if it helps to bring the product earlier to the market. Faster process development confers competitiveness. Entering the market place too late is quite often a disaster for the market penetration of a new drug.
In addition, the United States of America (USA) Food and Drug Administration (FDA) process analytical technology (PAT) initiative is about to revolutionize biotechnology-based processes used for the production of pharmaceuticals. The PAT initiative provides for real-time monitoring and control of the state of the production process within a certain ?design space? as opposed to the conventional 'one recipe' strategy. The PAT initiative is driving demand for reliable real-time monitoring tools in bioprocesses.
1. organism development creates a great number of strain variants. Real-time analysis of the performance of the strains will aid in screening of strains and selection of the best production organism in miniaturised or small scale bioreactors.
2. production conditions are optimised using real-time measurement devices in bench-top bioreactors to accelerate scale-up to full production
3. real-time measurement devices can be applied in process control, monitoring and quality control of the production phase
4. overall, the use of the real-time measurement devices has potential to shorten the whole development process by approximately 40 %.
Improved bioprocess control and faster organism development drives a need for a device system which fulfils the following requirements:
1. the ability to take rapid real-time measurements of product and biomarker levels during cultivation, thereby eliminating time-consuming manual analysis. Frequent measurements can be taken without sample storage. In addition, automated data collection allows identification of those time intervals at which selected samples should be stored (e.g. for future quality assurance / quality validation requirements).
2. a versatile analyte range, adjustable to measure various parameters and molecules (including intra-cellular analytes) depending on the process. The analyte range to be measured by the system should be flexible with only changes to the method and without any changes to the system.
3. the ability to collect representative, low-volume samples from large and small scale cultures without endangering the integrity of the cultivation. In addition, all in situ sensing or sampling devices need to be cleanable and sterilisable before and after each sample. Sample volumes will depend on the sensitivity of the tool but should be as small as possible, in particular when sampling from small scale cultures.
4. reproducible performance throughout the duration of long cultivations, which is achieved to a great extent via automation and using ex-situ devices. In-situ devices in particular are vulnerable to fouling and drift as they often cannot be changed or cleaned during cultivations. In addition, ex situ devices are easier to maintain and calibrate during long cultivations.
5. a compact size and a reasonable price - both for the infrastructure and the consumables. Lower costs and smaller size enable higher throughput via multiplication, which is needed especially in the organism screening and selection stages. Smaller systems are cheaper to install and can be sited closer to the process reducing the infrastructure required. However, the size is not the primary criterion of the device development; the overall performance of the system is of greater importance.
Scientific and technical objectives of the project
The aim of the NANOBE project was to construct a versatile tool for the real-time analysis of several compounds and biomarkers in bioreactor processes. The tool was targeted to the control and optimisation of production processes and to acceleration of development of production organisms for applications in industrial biotechnology. Micro and nanofabrication techniques were used to exploit the scaling laws associated with microfluidic devices to reduce analysis time and sample volume. The added value of the NANOBE approach is in the ability to measure certain biomarkers, indicative of cell performance and state.
The NANOBE approach aimed to a flexible platform for real-time measurement of both intra and extra-cellular analytes. One motivation was to avoid the need to develop a custom probe for measurement of each individual analyte. The versatile measurement tool would require only a change in method to enable the measurement of another new analyte. The device was to be designed so that it could be coupled to range of bioreactors, from novel micro-bioreactors to conventional industrial production bioreactors.
Automated sample handling for measurements of intra and extra-cellular analytes was of importance because some key biomarkers exist only inside cells. A general requirement was sensitivity. Sensitivity minimises the amount of sample required. Small sample volumes have the advantage that many samples can be taken from a culture without disturbing the process. The sampling and analysis of individual cells was desirable because we need to record variation in single cell performance, rather than observing the performance of the whole population of cells. In addition, accurate cell counts from a culture were paramount in order to relate productivity to cell biomass.
The application
NANOBE system was designed so that it is flexible in terms of what analytes to measure. In practise this means that different modules of the NANOBE system can be selected and combined for various applications. In the NANOBE project the primary example of the application was all along the production of organic acids in yeast saccharomyces cerevisiae. This related to several publicly funded research projects at the technical research centre of Finland (VTT).
It was seen important to develop the NANOBE system so that it can cope with different analytes, different cell types and different cultivation conditions in the future. However, only one application example was to be focused on when testing the whole NANOBE system. The main application example was lactic acid production in genetically modified s. cerevisiae strain. Strain is based on CEN.PK113-16B. Strain has a lactate dehydrogenase gene ldhL from lactobacillus helveticus on a plasmid. Individual selected analytes are chosen so that they are related to lactic acid production.
The analytes were chosen so that during the cultivation of the strain it is possible to follow how the beginning and the end of the lactic acid production pathway operate at the gene expression, enzyme activity and intracellular metabolite level, in addition to consumption of glucose and production of lactic acid. So far it is unknown for instance how the levels of intracellular pathway components change during different phases of the production process. Measurement of the levels may provide additional information into the cell-level regulation of the lactic acid production and open new possibilities for controlling the process in order to improve the production rate. Lactic acid is the precursor for poly-lactic acid, which is used in the production of biodegradable plastics such as cups and plastic bags.
Project results
Main scientific and technological (S&T) results
Sampling
A significant effort in the NANOBE project was put on developing a sampling system, which allows a time-discrete sampling of dead-volume free µl-scaled samples and a subsequent separation of cells and fermentation broth. This compact sampling and filtration system makes an interface between bioreactors and the rest of the NANOBE measurement tool. The sampling probe needs to be sterilisable. It also needs to fit the industrial standard size. The probe should preferably also be applicable to the sampling of different cell types (e.g. yeast, bacterial, filamentous, plant). In addition, integration of filtration and optical in situ sensing (e.g. in situ sensing for dissolved gaseous analytes) were to be included to create an intelligent probe in a conventional package which will fit bioreactors. The sampling and filtration system supplies the downstream analysis modules with the raw sample for cell counting, the retentate for intracellular substance analysis after cell lysis and the filtrate for extracellular substance analysis.
The sampling probe and the filtration module were realised by means of conventional technologies, whereas the integrated fluid management was partly realised with conventional and with a new on-chip pinching valving/pumping technology. The beneficiaries involved in the sampling probe development were Institut für Bioprosess und Analysenmesstechnik e.V. (iba), in Germany; Department of Microsystems Engineering (IMTEK), University of Freiburg, in Germany; and PreSens Precision Sensing GmbH (PreSens), also in, Germany.
Sampling probe specifications
At the beginning of the project the sampling probe specifications were discussed with all beneficiaries and defined for saccharomyces cerevisiae and streptomyces fermentation.
Dead-volume free sampling
Iba developed a sampling probe that based on a two-capillary concept patented by them. It could be proved for the sampling specifications of saccharomyces cerevisiae and streptomyces fermentation by iba. The sampling probe was designed so that the sample could be embedded into two air plugs, which enabled a time-discrete sampling in µL scale (20 µL of yeast and 50 µL of streptomyces) without dead-volume. The sampling probe was successfully tested for fermentations with overpressure of 3 bar.
In order to withdraw a defined sample volume from a bioreactor operating with a gas flow rate up to 2 vvm, it was necessary to develop a special probe head adapter. This adapter excluded the air bubbles of the fermentation broth from the sampling tube but not from the sensor spots which are additionally located on the probe head. The sampling probe was successfully purged with buffer and compressed air so that repeated sampling of streptomyces over a fermentation period of up to 10 days was no problem.
Filtration in µL-scale
In order to separate the cells and the supernatant from each other a filtration module based on conventional technology was developed by iba. It is thus appropriate for usual sterilisation procedures like sterilisation in an autoclave. Using channel spacers having different heights it is possible to apply the filtration module for taking 20 µL samples as well as for 50 µL samples. It is possible to use filters made of ceramics, polymers or paper. As the filtration element an asymmetric polyethersulfone membrane was chosen for yeast and an asymmetric ceramic filter for streptomyces. A purging procedure was developed to clean the filter between the subsequent samplings. As the module is connected to the bioreactor by means of a sterile barrier downstream the bioreactor, it is also possible to exchange the filter without danger of contaminating the fermentation broth in case the filter is blocked.
Integrated Braille-actuated fluid management
A microfluidic chip was developed as an optional component to replace part of the tubing-based fluidic network of the sampling module. In order to achieve accurate fluid switching and low-volume pumping in nL- and µL-scale with minimal external components, an on-chip valving technique using actuation with a Braille display was developed. In this technique a thin membrane is deformed by the actuation of a Braille pin, resulting in complete blockage of the microchannel.
The microfabrication process of the PDMS fluidic structure with a thin deformable membrane was established by IMTEK. Test experiments for the valving function was successful with water up to flow rates exceeding the range of interest in the NANOBE project, proving the effectiveness of the developed technique for implementation in the sampling module. Additionally, a sequential actuation of multiple valve elements could be successfully used for fluid pumping. The quasi-peristaltic motion of the fluidic structure resulted in a flow rate of up to 6 µL/min.
In the final NANOBE demonstrator the role of the microfluidic chip developed by IMTEK was to perform the splitting of the raw sample in two parts, one for cell counting and the other for filtration and for continuously dilution of the one part of the raw sample to guarantee a fast and accurate mixing. A fluidic structure was designed and tested to intermittently extract small volumes (100 to 500 nL) from a continuously flowing stream. Also, a dilution structure was developed utilising Braille-actuated active mixing of two continuously flowing streams.
In situ analysis of pH, pCO2 and pO2
A further challenge for the sampling probe development was the in situ analysis of pH, pCO2 and pO2. As the existing pH sensor was not suitable for yeast culture due to its limited dynamic range of 5.5 to 8.5 PreSens had to develop a new pH sensor to cover the more acidic pH range from 3 to 5.5. For pCO2 a sensor was developed which covers typical dynamic ranges (1 to 25 %) for yeast fermentations. The pre-existing dO sensor was suited also for this type of fermentations, dynamic range 0.02 to 100 % oxygen.
Sensors for all three analytes were never applied in the same fermentation before. Therefore a common sterilisation method for all three sensor types had to be developed by PreSens. For the first time an autoclavable pH sensors was demonstrated. The autoclaving conditions were determined by the pCO2 sensor which was more critical than the other two sensors. Autoclaving in the absence of oxygen is required. Many cleaning procedures were investigated but none was compatible with all three sensors. Therefore a disposable concept had to be developed. The spots were integrated onto polycarbonate discs which could be mounted to the probe head and exchanged after the fermentation.
Integrated sensor and sampling probe
A functional conventional sampling probe for a standard port of 12 mm diameter with three integrated optical sensors for gas/pH monitoring were constructed and manufactured by iba. This is reducing the amount of ports needed in the bioreactor, saves time during assembly and reduces the risk of contamination by reducing amount of components. Optical fibres used to read the optical sensor spots have to be removed before autoclaving and can be easily integrated in the sampling probe after autoclaving.
The construction of the sampling probe was optimised for a usage with any bioreactor independently of the position of the agitation motor by separating the sterile and overpressure barrier from the probe. A tube squeezing valve protects the further downstream modules from high pressure inside the bioreactor during cultivation and from contamination. The sampling and filtration system can be easily coupled sterile with the sampling probe via this valve. The newly developed probe-head adapter guarantees bubble-free sampling from the fermentation broth without any influence on the accuracy of the sensor measurements. As the raw sample is embedded into air plugs, samples can actually be withdrawn without dead-volume. The volume of the sample can be easily adapted from 10 µL to a few hundreds of microlitres if necessary.
Testing of the sampling probe in a lab scale process
The sampling probe with its integrated optical sensors was tested in a lab scale process of yeast at iba regarding to its functionality, biofouling effects and reproducibility in comparison to a conventional tubular probe as sampling system.
The sampling procedure in the NANOBE setup lasts approximately six and a half minutes. The volume of the sample was determined by measuring the sample length in the tube directly after withdrawing and weighting the sample after pumping the sample through the tube squeesing valve. With both methods a good reproducibility of sample volume was determined.
The cell counting was performed with repeat determination by means of a Neubauer counting chamber. During four days of fermentation the cell count was determined. During the first 28 hours the cell counts of both sampling probes were almost the same. In the next days the differences of the cell counts of both probes first increased, then adjusted to each other before the differences increased again (to 15.5 % on average). Therefore it was not a cumulative effect. The observed sample volume increase of 3 % within the tube cannot result in such an increase or dilution of the sample, which would explain the differences in the cell counting between the two sampling probes. But it indicates an attachment at the inner sides of the tubings. Adhesion and biofouling is always a topic in microfluidic devices, but only some adhesion on the two-capillary probe was observed. Additionally, the different tubing materials and the tubing diameters together with the flow rates of the samples could influence the sample content. However, the cell counting method itself showed a high standard deviation, so that the differences are only speculative. Only further investigations could finally explain it.
Cell handling tools
In the NANOBE project the cell handling tools were developed to provide information about the concentration and physiological state of the production organism and to perform cell lysis as well as sample handling. Both Ecole Polytechnique Fédérale de Lausanne (EPFL), in Switzerland and Centre National de la Recherche Scientifique (CNRS), situated in Lille, France, participated in the development of the cell handling tools.
Cell sorting and counting
Cell sorting and counting devices were developed by EPFL. These devices were designated for processing of samples with cells extracted from the bioreactor by means of the sample probe. The main components are cell sorting by dielectrophoresis and cell counting by impedance measurements. The sorting and counting module allows the separation of living and dead yeast cells and the counting of the separated populations. The viability of the sample is obtained by dividing the number of living cells counted by the total number of counted cells. The cell density is obtained by dividing the number of counted cells by the sample volume. After they are sorted, the cells flow through microelectrodes arranged in coulter counter configuration.
The connections of the sorting/counting chip were adapted to be able to connect the device to the other NANOBE subsystems with tubing. The tubing of small inner diameter was used and the tubings were arranged so that the sedimentation at the inlet and outlet of the chip and the dead volumes were reduced.
The remaining dead volume of a few microliters still represented a long transit time at the speed of the sorting system. Therefore, EPFL developed a pressure regulated as well as a switchable bypass which allowed the sample to pass faster through the tubing and only a fraction of the sample was analysed with the sorting/counting chip. The switchable bypass was successfully tested with a sample of yeast cells and suitable values of the order of 1 µl/min for the flow entering the sorting/counting system.
During the last year of the project the cell analyser module was reworked to use impedance spectroscopy instead of cell sorting, in order to reduce complexity and improve reliability. Parameters for counting as well as density and viability measurements were adapted for use with NANOBE in order to enable the use of higher flow rates. The result was good, the new system allowed for considerably higher flow rates of around 1 µl/min without use of a bypass.
Cell lysis
Cell lysis is used to break the cell membrane in order to extract intracellular components for analysis. Both chemical and electrical lysis have been investigated in the NANOBE project, although electrical lysis is preferred because components present in the chemical lysing solution could lead to fouling of the downstream analysis. Cell lysis was investigated by both EPFL and CNRS. EPFL focused on electrical lysis whereas CNRS worked both on electrical and chemical lysis.
Chemical lysis
CNRS studied the chemical lysis of Saccharomyces sereviciae yeast cells. First, CNRS determined with cellytic from Sigma Aldrich if chemical lysis would be compatible with the technology used in the sample transceiver device and with the NANOBE specifications.
Electrical lysis
Electrical lysis makes use of high electrical fields to disrupt the cell membrane. Higher electrical fields yield better lysis efficiency, but can also lead to the formation of bubbles at the microelectrode surface. An alternating voltage at a frequency of 50 kHz was used by EPFL to reduce this effect while still achieving efficient cell lysis. Specific microchannel designs can be optimised to locally concentrate the electrical fields in the lysis regions.
These three designs have been evaluated by EPFL using red blood cells, as the lysis of those cells is easily observable optically by the diffusion of haemoglobin out of the cell. The third design with so-called liquid electrodes was chosen for the NANOBE project. In this design the current density and thus the risk of electrochemical reactions is reduced. In addition the electrical field is uniform on the side of the main channel and the length of the lysis region and the transit time of the cells is largest of these three options without reducing the electrical field. This leads to better lysis efficiency and higher throughput. Two different flow cytometry methods were used to evaluate the lysis efficiency with yeast cells: cell sorting by dielectrophoresis and multiple-frequency impedance measurements. The measurements indicated that the intracellular components were successfully extracted from the cells. However, an effective lysis of the cells at higher flow rates required the application of higher voltages (100 V) which resulted in bubble creation at the electrodes and clogging of the channel. The long term operation of the device was thus impossible and the device was not suitable for the NANOBE project purposes.
To overcome the bubbling and clogging problems, EPFL adopted a new technology based on three-dimensional (3D) carbon electrodes. EPFL demonstrated high lysis efficiency (90 %) with high flow rate (up to 50 ul/min) while the lower flow rates (below 20 µl/min) tended to show slow bubble creation and cell sedimentation inside the microchannel. Despite of the good lysis results with optimised operation the EPFL device was not well compatible with the NANOBE system specification and suffered reliability problems.
CNRS had developed parallel electrical lysis technologies in the NANOBE project. The lysis device was made of conductive silicon layer sandwiched between two isolating glass slides. These 3D electrodes allow constant electrical field along the full channel height. The chip was powered by a high alternating current (AC) voltage to enable disruption of the tough yeast cell envelope without bubble formation. The CNRS lysis chip was selected to be implemented to the integrated NANOBE analysis system.
Filtration
Three filtration systems were required in the NANOBE workflow:
1. a filter after cell lysis, removing the cell debris from the lysate
2. a filter before EPFL density/viability chip to remove dust but letting cells flow through
3. a micro-debubbler to avoid air entering into the density/viability chip.
EPFL tested commercial filters for the removal of cell lysis debris as well as for protection of the cell counter chip. The commercial inline filters from IDEX were found to be suited for the removing of the cell debris from the lysate (filter pore size 1 µm) and for the removal of dust (filter pore size 10 µm). Moreover, a micro-debubbler was developed and tested to be put just in front of the cell counter. The micro-debubbler was found to be suited for the removal of the air between samples, as well as smaller bubbles.
Sample transceiver
A special sample transceiver was developed by CNRS to perform the delivery of the supernatant or lysed cell samples delivered by the sample probe to the individual analysis modules.
The sample transceiver module uses electrowetting on dielectric (EWOD) technology to manipulate and deliver samples to the analysis system. The liquids are provided and delivered to/from the sample transceiver through micro-channels. In the sample transceiver the incoming liquid is digitised to small droplets (volume 1 µl) and all the operations in the sample transceiver are made by moving droplets in the central area of the device. Various bio-chemical operations can be done in digital fluidic mode. At the output end the droplets are combined to specified sample reservoir area from where they can be pumped onwards to the analysis tools downstream the NANOBE system via capillary connections. The functionalities available in the sample transceiver are droplet generation, sample mixing/diluting and sample exportation.
The sample transceiver was tested with cell lysate, supernatants corresponding to different cells cultures, the buffer used for CE analysis and PBS with different concentrations. All these fluids were to be used in the NANOBE application. Based on these tests CNRS defined the conditions on these fluids to be manipulated by electrowetting in the sample transceiver. The surface tension and bio contents of the droplets were the key parameters to enable the manipulation of these liquids. It was observed that too low surface tension or too high bio-concentration of the droplets made the droplet displacement difficult or even impossible. The results showed that the movement feasibility of the supernatant in yeast nitrogen base (YNB) is better than the one in yeast extract peptone-dextrose (YPD). For 'cell in buffer', the two solutions both need to be worked by 100 time dilution. The results confirmed that with a moderate dilution with water it was possible to handle all the fluids needed for NANOBE application.
Micro-ELISA analysis of proteins
The NANOBE system has two analysis paths for intra- and extracellular compounds. For the first one, DiagnoSwiss has adapted their micro-ELISA technology to automated immunoassays and to online process monitoring of extracellular product protein concentrations. Furthermore, DiagnoSwiss has evaluated the feasibility of mRNA level determination using magnetic beads and electrochemical detection in their ImmuSpeed platform.
Coupling with the sample transceiver
The specifications for the coupling of the micro-ELISA platform to the sample transceiver were set according to the fluidic layout of the integrated system, the unit operations of the sample transceiver and of the micro-Elisa platform and the related timings of the analysis cycles. For the immunoassays, the total analysis cycle was set to 40 minutes, with the sample being sent from the transceiver to the ELISA module every 20 minutes, but with two samples in duplicates being analysed simultaneously with a 40 minute cycle.
The coupling system consisted of a peristaltic pump connected to the sample transceiver in order to transfer the samples in wells of a sample collector and placed in the micro-ELISA platform. The robotic head of the micro-ELISA platform will then transfer the samples into the microtiter-plate placed in the instrument for incubation with the capture antibody and then further analysis by electrochemical detection in the microchip.
Adaptation for the messenger ribonucleic acid (mRNA) analysis
In the NANOBE system, investigations were made to evaluate whether expressed mRNA levels of the selected genes can be measured using DiagnoSwiss' ImmuSpeed device developed primarily for sandwich assay of proteins and see whether it could be adapted also for measuring the expression of mRNA molecules.
These studies demonstrated the feasibility of transposing the TRAC methodology to the electrochemical micro-chip format and that a reliable assay protocol could be developed for HXK1 mRNA detection. However, the obtained results were only semi-quantitative and, in its current state, the assay performances did not show sufficient sensitivity and reproducibility for direct application to online monitoring of yeast cultures. Further efforts should be placed in order to investigate whether the assay performances can be further improved to reach the very low detection limit required for real applications.
CE separation of analytes
The second analysis path of the NANOBE real-time monitoring tool was a coupled CE-MS system for measurement of small molecules, such as substrates, metabolic products and intermediate biomarker metabolites. Originally the CE was planned to be applied in multiple ways in the NANOBE project to analyse metabolic compounds, nucleic acids and fluorescent cells. It could also be used for the detection of intracellular enzyme activities. In yeast fermentation the most important analytes are extracellular sugars and organic acids. In streptomyces spp fermentation the most important analytes are polyketides rapamycine, nigericine and/or tacrolimus depending on the strain. The design of the online analysis system was such that the (above mentioned) analytes in the cultivation broth can be quantified in concentration range 100 mg/L to 100 g/L. The required sample volume from the sample transceiver to the CE is 1 µl.
During the project VTT designed, fabricated and analysed CE chips with and without MS connectivity option. Laser induced fluorescence (LIF) was used as a detection method in the case of CE chips without the MS option while the MS was naturally the choice for detection in the case of other chip design.
MS detection of analytes
Microsaic developed further its micro-fabricated mass spectrometer technology to be integrated with VTT's chip based CE technology. The objective was to develop the first fully micro-fabricated CE-MS system and to provide online detection and definitive identification of analytes separated by the CE module. In the NANOBE system MS detection is used to analyse extracellular substrates and metabolites. A stand-alone 'Spraychip nanospray' module and controller was designed, built and tested by Microsaic. The development of negative ionisation mode was completed before delivering the module to VTT for testing. VTT successfully completed initial trials of the nanospray unit interfaced to one of their commercial mass spectrometers. The Spraychip nanospray module was interfaced to a chip scale breadboard atmospheric interface mass spectrometer developed by Microsaic and the functionality of the integrated system was successfully demonstrated.
Sample capture for storage and off-line analysis
A sample capture for off-line surface assisted laser desorption/ionisation - desorption ionisation on silicon surface (SALDI-DIOS) analysis was developed at CNRS. In NANOBE, the DIOS chip was planned to be used to store and analyze off-line both extra- and intracellular metabolites and products in saccharomyces cerevisiae and streptomyces fermentations. In yeast fermentation the most important analytes are extracellular sugars. In streptomyces fermentation the most important analytes are rapamycine, nigericine and/or tacrolimus depending on the strain. Limit of detection of these compounds is in the range of nmol-pmol/µL. CNRS evaluated possible substrate candidates for the DIOS analysis. CNRS evaluated also the ion affinity chromatography (IMAC) technique for functionalisation of the nanowires. The superomniphobic surfaces developed by CNRS allow handling of droplets having a low surface tension due to either surfactant or some biomolecules. It was shown that depending on their morphology, the silicon nanowires offered liquid-repellent character with very good robustness over a wide range of surface energy.
Towards the end of the project CNRS studied the interaction of graphene oxide and reduced graphene oxide with bovine serum albumin (BSA). The results suggested that graphene oxide has a high capacity for protein adsorption. This property was finally exploited for the displacement of BSA using EWOD with minimum non-specific adsorption. The SALDI-DIOS was not part of the final system demonstrator.
System integration
Finally, the complete NANOBE system was developed in form of several stand-alone units that were physically integrated to a complete system. All partners were involved in defining the detailed specifications of all modules and their interfaces while IMTEK was responsible for the collection and management of the system specifications. The NANOBE system was constructed on the basis of the commonly agreed structure of the system and the function of the individual modules.
The following modifications to the complete NANOBE system had to be made based on the results of the submodule interface testing:
1. due to software communication problems with the impedance spectrometer and some issues with the sampling probe, no cell viability measurements could be performed simultaneously with the other components of the system
2. the high-pressure resistance of the silica capillary connections of the cell lysis chip and the limited pressure tolerance of the microvalves used for flow switching prevented the use of the cell lysis chip in the online monitoring system for the final testing. This is a result of a long capillary connection between the modules that was not planned at the beginning of the project.
3. the supernatant dilution ratio was necessary to be increased to 1:50 from the original 1:10 to avoid the damaging of the sample transceiver electrodes.
Testing of the NANOBE system
The genetically modified yeast cell strain of VTT was cultivated in the bioreactor and used for interface development and integrated system testing purposes. During the testing the pH and oxygen concentration was monitored with conventional sensors for comparison with the PreSens optical sensors. Cell samples were manually sampled with a syringe at the same time as the automated sampling at the IBA sampling probe. Off-line analysis of the manually-taken samples were performed using the CE-MS and ELISA modules and with off-line high performance liquid chromatography (HPLC) for comparison.
Integrated NANOBE system
Despite the original plan of completing the online analysis cycle in 25 minutes with a 20 minute cycle, the finally tested analysis cycle turned out to require 80 minutes. In the course of the timing and flow rate adjustments during the interface development, all process steps at which the sample or washing liquid is delivered into external microfluidic chips needed to be prolonged considerably. The resulting online analysis sequence was such that the sample was delivered to the transceiver approximately 45 minutes after automated sampling. The time dedicated to the sample preparation in the sample transceiver was around 6 minutes. The CE-MS analysis and ELISA measurement of the supernatant would finish at 60 minutes and 100 minutes after sampling, respectively. The diluted raw sample reached the cell counting chip approximately 15 minutes after sampling, although the impedance measurements and viability analysis were not performed online.
In total, only four analysis cycles were carried out with the integrated system. There were two cycles in which the sample was successfully transported from the bioreactor to the ELISA and MS modules. One out of the two successful system operation cycles resulted in successful data readout in both the ELISA and MS modules. The results from the ELISA analysis suggest that the diluted supernatant from iba sampling/filtration module may have been more concentrated than the theoretical 1:50. If this could be confirmed, it can explain the difficulty to transport the sample in the sample transceiver. The CE-MS analysis results were qualitatively in good agreement with data from the manually-taken sample using the same Microsaic mass spectrometer. The monitored dissolved oxygen concentrations were in good agreement with the values measured using conventional sensor. The oxygen monitoring continued for two days without major drift.
The failure modes in the three unsuccessful runs were all different. In cycle one, the sample reached the inlet of the mass spectrometer but did not reach the detector due to clogging of the vacuum interface. The ELISA module also failed to proceed beyond the sample/bead conjugate incubation due to software failure on the master PC side. In cycle two, several electrodes on the sample transceiver were damaged due to repeated use from interface development to sample testing. In cycle four, an unexpectedly viscous sample was transported from the iba sampling/filtration module to the sample transceiver which indicates that the supernatant dilution of 1:50 ratio was not achieved, most probably because of air bubble generation at one fluidic junction. This is also an explanation for the discrepancy between the IgG concentration in the online sample and manually-taken sample in the ELISA measurements.
In summary, the complexicity of the whole NANOBE system caused several irregularities in the sample transport, which could not be finally eliminated. The reason for this is that the development of all modules had to be finished before the interfaces could be developed. It was tried to adjust the interfaces with the state of development of the modules, but this was time consuming and expensive, so that it could not always be done.
Testing of an alternative integrated system configuration
An alternative online operation of the integrated system was also tested. In this configuration, the CE module was directly connected to the supernatant outlet of the sampling/filtration module. This is one variant of the system for applications that do not require ELISA measurements. Since there is no need of sample distribution to multiple measurement paths, the sample transceiver is also omitted.
The outcome of the alternative configuration test was similar to that in the full system test; one set of measurement data was obtained in three test cycles. One failure mode was the clogging at the vacuum interface of the mass spectrometer and the other was leakage at one fluidic connection between the sampling/filtration module and the CE module.
Results on IgG determination
For the two samples delivered by the sample transceiver, only one sample resulted in an electrochemical signal corresponding to an enzymatic reaction. The second signal gave a negative slope corresponding to a blank point. The difference between the spiked concentration and the measured concentrations in the undiluted and diluted samples may be explained by a partial decomposition of the IgG molecules during the fermentation, by a loss of IgG during the filtration of the medium, by dilution errors and by artefacts between the signals obtained for calibration points measured in buffer solution and the real samples. Unfortunately, the lack of experimental data does not allow one to understand the discrepancy between the expected and the measured IgG concentrations.
These limited number of data points prevents to identify clearly the source of errors, to qualify the real performance of the developed assay and to benchmark the results against measurements that could be obtained with the same samples using conventional microtiter-plates with optical detection.
CE-MS analysis of extracellular metabolites
During the testing of the integrated NANOBE system, altogether five samples out of eight were successfully analysed online with the CE-MS system while being part of the whole NANOBE system. In addition, all of the eight samples were analysed off-line as well. Absolute quantification was not possible for three reasons:
1. the optimisation of the CE method was not complete
2. dilution of the sample upstream in the Nanobe system was not accurate eventually
3. ionisation at the MS varied resulting in large uncertainties.
Furthermore, the delivery of the sample from the bioreactor was too slow. Quantification problems two and three could have been and can be solved by using internal standard in the cultivation. Normally internal standard is not used directly in the cultivation as it may affect the performance of the cells or it may be consumed by the cells.
Online ethanol extraction and CE-MS analysis of intracellular metabolites
The analysis of intracellular metabolites was not possible with the NANOBE setup, because the time in the NANOBE setup for the treatment and lysis of the cells would have been too long and the metabolism of the cells would have had time to adapt to the conditions in the sampling system and would not have reflected the status of the cells in the reactor. A separate setup was built to demonstrate that the analysis of intracellular metabolites was possible with the online CE-MS device.
Several difficulties occurred during these tests, vac chips in the MS got clogged and there was also some pressure build-up caused by the narrow tubing. It was also noticed that the m/z values varied between analyses in the mass spectrometer which made identification of the peaks difficult.
In the future research of the CE-MS, the CE method development will be continued. Higher CE voltages and temperature control would increase sensitivity and make the CE peaks sharper. Non-aqueous background electrolytes (BGEs) would help with the ionisation of analytes in electrospray and might also help to prevent the clogging of vac chips. The use of internal standard would be essential for the quantification since there is always variation in the MS signal due to day-to-day variation in the efficiency of ionisation.
Overall the project succeeded to connect the MS detection developed by Microsaic to the CE separation by VTT and this CE-MS system to the rest of the NANOBE system. Furthermore, VTT was able to develop a method for online ethanol extraction in order to enable online analysis of intracellular metabolites in connection of the CE-MS system. Measurement of extracellular and intracellular metabolites in connection to bioreactor cultivations online was demonstrated for the first time.
However, the methods are not yet quantitative and further work is required. Problems encountered in flowing the electrophoretic solvents used in the CE directly to MS are a major issue that needs to be solved by developing the CE system and the method as well as the ionisation.
Microsaic's MS as instrument was very easy to operate; one of the most advanced features of the device was the software. Compared to conventional mass spectrometers changing parts was much easier and faster. Also, the vacuum was reached very rapidly, in just 30 minutes. The electrospray of this device was very promising, because it was able to produce stabile electrospray despite very low liquid flow rates of CE system.
Cell analyser tests with VTT's own sampling system
As the tests of the cell counter/analyser chip were not successful in the integrated NANOBE system, VTT set up a simpler system for connecting the EPFL chip to a bioreactor. PBS flow was maintained steady using a pressurised and pressure-controlled vessel.
The developed impedance analysis system can efficiently discriminate living and dead yeast cells and determine the cell density from the average cell speed. The system can function with the sampling probe and gives good results with yeasts. Unfortunately the combination of different problems starting from existing commercial instruments led to the fact that there was no time for producing results to really demonstrate that the cell and viability measurement chip works in the NANOBE system. Results outside the NANOBE system have been presented and published and show that the technology works. When connected to a bioreactor with a simpler system some measurements for cell density and viability directly from the cultivation could be done. However, there is great promise with this technology as cell density and viability measurements are the top priority for measuring in bioprocesses. Tests of the optical probes were also carried out.
Conclusions
The NANOBE integrated system was constructed only a couple of months before the end of the project, occupying one laboratory in VTT. All hardware was connected with each other as planned, with the exception of the cell lysis chip. The resulting system created a long pathway (to some system parts such as the EWOD, which reduced the timing precision and made inclusion of the lysis chip impossible. The fully integrated system was operated for online handling and analysis of genetically modified yeast fermentation.
After a significant amount of software modifications, the fluidic interfacing between the individual modules functioned successfully with acceptable repeatability of the transported sample volume. However, due to the long tubings the timing of fluidic transport to and from the sample transceiver saw non-negligible deviations, which made fully automated operation of the entire system impractical. Therefore, triggering of the module operation was performed manually although automated triggering was implemented in the master program of the system.
Online sample handling and analysis using the integrated system functioned flawlessly in one run out of four attempts. Failure modes in the unsuccessful runs were all different, appearing in different sections of the system. This means that the robustness of the modules have to be increased.
Comparisons of the measured mass spectrum from the online and reference offline analyses indicated that the sample has been successfully transported through the entire system without major contamination from the various splitting, mixing and cleaning steps during the automated operation. The ELISA readouts of spiked IgG concentrations showed a significant deviation from the offline measurement, but the two were on the same order of magnitude. Considering that the dilution ratio of the supernatant sample in the sampling/filtration module is yet to be quantitatively evaluated, the outcome is encouraging for further development of the system.
Overall, the NANOBE concept of an automated online bioanalysis system based on microfluidic core components was proven to be promising in principle. It was shown that a minute amount of sample with a volume of 20 µL can be analysed online in two paths without any manual handling. Two different configurations of the system components were tested, both showing results that are comparable to conventional off-line methods, which was a good demonstration of the modular interchangeability of the system. Improvements in robustness and accuracy of sample handling are keys for achieving constant operation for online monitoring over the full duration of fermentations.
Potential impact
The NANOBE project delivered an integrated real time bioprocess analysis system for measuring of extra- and intracellular compounds. It provided a system for bioprocess monitoring that was not existent at the time when the project was started. The NANOBE system is well beyond the state of the art in its concept and individual components. Smart systems integration led to increased functionality, since the sampling is now solved in a way which enables the real online analytics during the bioprocess. Different lab-on-a-chip modules enable measurement of various analyte types such as cells, low molecular weight compounds, proteins and specific mRNAs. The system provides improvement in terms of automation, analysis time and sensitivity. Further development is needed to increase the robustness of the integrated NANOBE system. The analysis platform can significantly improve real-time feedback control of large-scale production processes as well as screening and optimisation of production organisms and conditions. The project combined in an innovative manner special expertise in microfluidics, nano- and microfabrication techniques, chip-scale mass spectrometry, photonics, electronics, sensor technologies and biotechnology.
There is an increasing need for more rigorous bioprocess monitoring due to the increasing number of tailor-made production organisms and bioprocesses that will be required for sustainable production of fuels, chemicals and pharmaceuticals. Development of novel production organisms is benefited from the recent advances in genomics, metabolic engineering, synthetic biology and systems biology, which in turn provide tools and understanding to design analysis methods for those cellular and process parameters and biomarkers that are critical for the most efficient performance of the process. The increasing number of non-traditional bioprocesses with high demands on energy efficiency and yields sets a demand for multiparameter analyses that also enable better feedback control. In addition, using the NANOBE platform, right time points for sample collection for more comprehensive off-line analyses can be chosen. With the NANOBE system also individual cells can be analysed. This provides means to count non-viable and nonproductive cells in the population. On the other hand cells with intracellular biomarkers that indicate cell stress etc. can be counted. The biomarkers can also be used for cell sorting and for instance selection of cells with best viability and productivity for further cultivations. It is also noteworthy that the technologies developed allow synchronisation of cells. This is of fundamental importance since until now the great majority of biological data has been generated from populations where the cells are in different points of the cell cycle and have different physiology. Thus, it is to be noted that the NANOBE analysis platform is not only useful for monitoring and control of production processes but is an excellent tool for strain screening and process optimisation, as well as for basic studies on cell physiology and cellular phenomena including systems biology analyses. Overall, the online monitoring tools developed in the NANOBE project can help to increase the production rate, yield and concentration of the final product of a fermentation process. These improvements in process monitoring may be crucial for the economic viability of a new bio-based product.
NANOBE platform offers reduced process development times for existing bioprocesses. This will enable the cost reduction e.g. in pharmaceutical industry. Firstly, time is saved for custom manufacturing companies in their process development services. Secondly, the time that it takes for a drug to reach the market will be reduced. Thirdly, time savings in product development also confers competitiveness. Also e.g. product purification and process characterisation at small scale prior to production of large scale GMP batches can be done much quicker by use of online real time assays, which will be implemented and tested in this NANOBE project. The FDA PAT Initiative is demanding a higher amount of analytical methods during the bioprocess similar to the analytics in the downstream processing. The NANOBE project is directly addressed this challenge by providing a real time and on line measurement platform for multiple analytes.
A further cost effective point is the possibility of miniaturisation, which results in a higher parallelisation of bioprocess approaches. Furthermore, some modules need less space as the known systems; e.g. MS and the combination of gas/pH sensors with the sampling probe. Last one results in saving ports for other analytics in the bioreactor, saving time during assembly and reduces the risk of contamination by reducing the amount of components.
Bioreactor is an essential tool for bioprocess research and development. It can be used for wide range of applications varying from general process optimisation to gene mapping tasks. This fact also leads up to a wide range of different type bioreactor designs suitable for different applications. The real time measurement platform developed in the NANOBE project can be realised in a way that it will suit different types of bioreactors and different types of cultivations. The experiences and conclusions obtained during the NANOBE project are of general value for bioprocess monitoring and very useful for other research areas like cell and tissue engineering, lab-on a chip development and standardisation.
Although the measurement platform developed in the NANOBE project was tested only with a certain yeast based bioprocess, the perspective still is to apply it to monitor also cultivations based on different organisms in the future. The extracellular analyses developed in NANOBE are very generic and can be used with most organisms. The complexity of the culture medium may however cause adjustment in the device cleaning and analysis methods.
The main dissemination activities and exploitation of results
Dissemination activities
In the NANOBE project the dissemination of knowledge was done in multiple ways. An essential part of the dissemination of the results of the project results was to publish them in major international scientific journals and conferences. Numerous scientific journal papers have been published, accepted for publication or submitted by the end of the project. In addition several conference presentations were held in international conferences and workshops.
An Industrial platform (IPF) was established in the beginning of the project. It comprised of representatives of application and equipment manufacturing companies. The members of the IPF had an access to monitor the progress of the project. Three workshops were arranged to disseminate the project objectives and results to the IPF members. The material presented to the IPF members have been distributed to all members by e-mail.
A project website was set up in the beginning of the project. The public extranet can be found in http://www.vtt.fi/NANOBE
Altogether, the NANOBE project produced 19 reviewed journal publications, 15 conference presentations and 17 other presentations to disseminate the project results.
Exploitation of results
Both general advancement of knowledge and commercially exploitable research and development results have been generated in the project. General advancement of knowledge relates to the developed technologies, devices and fabrication methods. Significant progress has been made in understanding biopollution and understanding of the interaction between small biomolecules in a liquid phase and nanotextured solid surfaces. Generated knowledge will be exploited in scientific research and development for various biological, chemical, environmental and life science applications by the beneficiaries. Generated new knowledge will be integrated also into the student education.
One patent application was filed in the project. With the new patent with the topic 'Sampling probe for withdrawing samples from aerated liquids' iba shows that the sampling probe consists of a wide application range.
The developed sampling port combining three analytes and sampling is believed to gain interest even without the attached analytical devices of NANOBE. The consortium hopes that positive feedback from the IPF and from other industries will arise in the near future and lead therefore to a commercialisation of this part of NANOBE. PreSens, DiagnoSwiss and Microsaic Systems will also exploit commercially those results of the research and development work that are the most mature. PreSens will profit from developments on the individual sensors for dO, dCO2 and pH. Diagnoswiss will use the prototype of automated microchip platform for further demonstration of immunoassay applications. Microsaic Systems has incorporated the technology developed during the NANOBE project into a new product, the 3 500 MiD. This miniature mass spectrometer was launched at the Lab Automation show in January 2011.
The end user companies have shown interest especially on the single component developments in the NANOBE project. These include the sampling probe, mass spectrometer, mRNA analysis and optical sensors.
Overall, the online monitoring tools developed in the NANOBE project could help to increase the production rate, yield and concentration of the final product of a fermentation process. These improvements in process monitoring may be crucial for the economic viability of a new bio-based product.
Project website: http://www.vtt.fi/NANOBE
Coordinator
Päivi Heimala, Senior Scientist, VTT Technical research center of Finland, PO Box 1000, FI-02044 VTT, Finland
Telephone: +35-820-7226637
Mobile: +35-840-5078004
Fax: +35-820-7227012
E-mail: paivi.heimala@vtt.fi
