TY - JOUR
AU - Schmid, Andreas
AB - Redesigning biology towards specific purposes requires a functional understanding of genetic circuits. We present a quantitative in-depth study on the regulation of the methanol-specific MOX promoter system (PMOX) at the single-cell level. We investigated PMOX regulation in the methylotrophic yeast Hansenula (Ogataea) polymorpha with respect to glucose-mediated carbon catabolite repression. This promoter system is particularly delicate as the glucose as carbon and energy source in turn represses MOX promoter activity. Decoupling single cells from population activity revealed a hitherto underrated ultrasensitivity of the MOX promoter to glucose repression. Environmental control with single-cell technologies enabled quantitative insights into the balance between activation and repression of PMOX with respect to extracellular glucose concentrations. While population-based studies suggested full MOX promoter derepression at extracellular glucose concentrations of ∼1 g L−1, we showed that glucose-mediated catabolite repression already occurs at concentrations as low as 5 × 10–4 g L–1. These findings demonstrate the importance of uncoupling single cells from populations for understanding the mechanisms of promoter regulation in a quantitative manner. Envirostat, MOX promoter, carbon catabolite repression, single-cell analysis, Hansenula polymorpha INTRODUCTION Promoter regulation is traditionally investigated at high cell number populations in shake flasks or stirred tank bioreactors. However, at a population level, heterogeneous promoter activities in single cells remain hidden behind the macroscopic population average (Elowitz et al.2002; Schmid et al.2010; Delvigne and Goffin 2014). Physicochemical changes in the microenvironment of single cells are an important cause for this phenotypic heterogeneity. Extracellular inhomogeneities that create physicochemical gradients arise from mixing effects and metabolic activity of the population. Such effects obscure the quantitative relation between macroscopic extracellular conditions and promoter activity (Dunlop and Ye 1990; Lara et al.2006; Dusny and Schmid 2015). However, controlling the activity of a promoter system in a given technical setup by strain and process design depends on quantitative insight into the underlying regulatory mechanism. An important example for processes in which promoters determine the efficiency of the biological system represents the production of recombinant proteins with yeasts (Nielsen 2013). The application of yeasts as protein-producing cell factories was in the first place enabled by the discovery of performant promotor systems (Ahmad et al.2014). Most of these promoter systems were derived from methanol metabolism in methylotrophic yeasts (Cos et al.2006). On the basis of these promoter systems, the production of heterologous proteins in yeasts has matured into a multibillion-dollar industry (Mattanovich et al.2012) Until now, the highest reported extracellular protein titers (13.5 g L−1) have been achieved with Hansenula Polymorpha (current name Ogataea polymorpha) as expression host (Yamada, Maeda and Mikata 1994; Mayer et al.1999). The high protein titer was enabled by using the strong formate dehydrogenase (FMD) promoter from H. polymoprha's native methanol utilization pathway. In combination with very high copy numbers of the target gene (the strain contained ∼120 copies), the strain was able to efficiently synthesize and secrete phytase under strong glucose limitation conditions for achieving derepression of the FMD promoter. Besides the FMD promoter, the peroxisomal MOX (methanol oxidase) promoter (PMOX) is a comparably strong, but even more tightly regulated promoter for the large-scale production of recombinant proteins with H. polymorpha (Gellissen 2000). The regulatory mechanism controlling MOX promoter activity in glucose-fueled protein production processes is carbon catabolite repression. The presence of glucose in the extracellular medium strongly represses the activation of the MOX promoter, while its absence initiates high transcriptional activity (Egli et al.1980; Kensy, Engelbrecht and Buchs 2009; Suppi et al.2013). Glucose-6-phosphate, the first glycolytic product of glucose, is assumed to be the molecule involved directly in carbon catabolite repression of the MOX promoter (see Fig. 1) (Teusink et al.1998; Suppi et al.2013). Signaling elements controlling carbon catabolite repression by hexoses, such as the recently identified HXS1 hexose transporter-like sensor, are responsible for the regulation of the MOX promoter (Stasyk et al.2008). However, the bifunctional role of glucose as essential carbon source and simultaneous repressor of PMOX activity makes this system a particularly delicate subject of study. Figure 1. View largeDownload slide Proposed mechanism of carbon catabolite repression of the MOX promoter by glucose in the methylotrophic yeast H. polymorpha according to Suppi et al. (2013). Glucose is taken up from the extracellular medium and transported into the cytosol via the HXT (hexose transporter) system. Glucose is subsequently phosphorylated by hexokinase or glucokinase enzyme activity. Glucose-6-phopshate acts as a repressor on the MOX promoter. Figure 1. View largeDownload slide Proposed mechanism of carbon catabolite repression of the MOX promoter by glucose in the methylotrophic yeast H. polymorpha according to Suppi et al. (2013). Glucose is taken up from the extracellular medium and transported into the cytosol via the HXT (hexose transporter) system. Glucose is subsequently phosphorylated by hexokinase or glucokinase enzyme activity. Glucose-6-phopshate acts as a repressor on the MOX promoter. In this context, three central questions are important for understanding the response of the MOX gene promoter to extracellular glucose. First, is the MOX promoter activity homogenously distributed among single cells? Second, how dynamic is MOX promoter activity in single cells as a result of extracellular influences? Third, is the averaged MOX promoter activity measured at the bulk level biased by continuously changing cultivation conditions due to the glucose metabolization of the population? Innovative analytical concepts are required for answering these questions. Microfluidics exploits the physical principles prevailing at the microscale and enables controlling the extracellular environment of a single cell. Laminar flow dominates in microfluidic devices and replaces turbulent flow conditions. This facilitates predicting concentration profiles in microstructured single-cell cultivation devices (Weibel, Diluzio and Whitesides 2007; Colin, Squires and Bocquet 2012; Westerwalbesloh et al.2015). Extracellular substrate concentrations are kept constant with continuous medium flow around the single cell, and secreted metabolites are immediately removed (Dusny and Schmid 2015). The isolated cell is hence uncoupled from bias that arises due to population activity (Dusny et al.2012). The Envirostat system enables contactless isolation and cultivation of an isolated single cell with negative dielectrophoresis (nDEP) in a continuous laminar medium flow (Kortmann et al.2009). With this single-cell cultivation system, the cellular phenotype can be assigned to the physicochemical properties of the extracellular environment. Complementary to single-cell microfluidics, powerful analytical concepts for characterizing promoter activities in single cells with fluorescent reporter proteins are available (Locke and Elowitz 2009). Despite this fact, the regulation of gene expression in isolated single cells and its connection to extracellular conditions has scarcely been investigated until now (Gruenberger, Wiechert and Kohlheyer 2014). We demonstrate how engineering of cellular microenvironments with single-cell technologies provides quantitative insight into the functional mechanism of glucose-mediated repression by the MOX promoter system. We employed a H. polymorpha strain that synthesized GFP (green fluorescent protein) under the control of the MOX promoter. This enabled linking GFP fluorescence intensities and promoter activities in individual yeast cells on the basis of quantitative fluorescence microscopy. Dynamics and degrees of glucose repression of PMOX in single yeast cells were evaluated with four principally different population and single-cell cultivation systems. These systems differed in terms of scale, population size, mode of cultivation and environmental control. We observed distinct differences of these systems with respect to the extracellular glucose concentration at which catabolite repression of PMOX set in. Single-cell analysis with the Envirostat revealed repression of PMOX at extracellular glucose concentrations that were more than four orders of magnitude lower as reported previously MATERIALS AND METHODS Strains Hansenula polymorpha RB11 MOX-GFP was used as recombinant host strain throughout this study. The recombinant clone originates from the methylotrophic yeast H. polymorpha RB11 parental strain (which is Ura− due to a deficiency in the oritidine-5′-phosphate decarboxylase gene (ura3) (Zurek et al.1996). Hansenula polymorpha RB11 MOX-GFP expresses GFP gene under the control of the MOX promoter system as a fluorescent reporter for MOX activity. The strain harbors eight copies of the expression cassette. Therefore, titration with transcription factors is unlikely to occur. As the strain originates from a single clone that underwent a selection assay, it can be assumed that the cultures are homogeneous in terms of gene copy number. The employed vector system for transformation lacks an α-mating factor secretion signal, leading to the intracellular accumulation of GFP. Ura− was complemented with oritidine-5′ phosphate decarboxylase gene from Saccharomyces cerevisiae located on the vector. Hansenula polymorpha strains were kindly donated by Dr. Michael Piontek (ARTES Biotechnology GmbH, Langenfeld, Germany). Media For precultures, yeast peptone dextrose (YPD) medium containing 2% (w/v) glucose, 1% (w/v) yeast extract and 2% (w/v) peptone was used. For agar plate preparation, YPD medium was additionally supplemented with 2% (w/v) Agar. All experimental investigations on MOX promoter activity were carried out with minimal medium SYN8, containing different amounts of carbon source. SYN8 is a modified variant of the SYN6 high cell density medium (Gellissen et al.2005). SYN8 was rationally designed by exchange of the phosphate source and further ion reduction. This prevented the formation of precipitate and offered an electrical medium conductivity of 1 S m−1 for optimal trapping performance of single cells via nDEP. Furthermore, it permitted identical growth rates and morphology of H. polymorpha as compared to the original SYN6 cultivation medium. SYN8 medium was composed of 2 g L−1 sodium hexametaphosphate ((NaPO3)6), 3 g L−1 ammonium sulfate ((NH4)2SO4), 1.5 g L−1 potassium chloride (KCl), 0.15 g L−1 sodium chloride (NaCl) and 3 g L−1 magnesium sulfate heptahydrate (MgSO4 × 7 H2O). The basic salt solution was filter-sterilized after adjusting the pH to 6.4 with 10 M sodium hydroxide solution. One liter of sterile basic salt solution was subsequently complemented by the sequential addition of 6.67 mL L−1 microelements, 3.33 mL L−1 trace elements, 2 mL L−1 calcium chloride and 6.67 mL L−1 vitamin solution. The microelement solution consisted of 10 g L−1 EDTA (C10H14N2Na2O8 × 2 H2O), 10 g L−1 ammonium iron (II) sulfate hexahydrate ((NH4)2 Fe(SO4)2) × 6 H2O), 0.8 g L−1 copper (II) sulfate pentahydrate (CuSO4 × 5 H2O), 3 g L−1 zinc (II) sulfate heptahydrate (ZnSO4 × 7 H2O) and 4 g L−1 manganese (II) sulfate monohydrate (MnSO4 × H2O). The trace element solution consisted of 0.2 g L−1 nickel (II) sulfate hexahydrate (NiSO4 × 6 H2O), 0.2 g L−1 cobalt (II) chloride hexahydrate (CoCl2 × 6 H2O), 0.2 g L−1 boric acid (H3BO3), 0.2 g L−1 potassium iodide (KI) and 0.2 g L−1 sodium molybdate dihydrate (Na2MoO4 × 2 H2O). The calcium chloride stock solution contained 150 g L−1 CaCl2. The vitamin stock solution contained 0.06 g L−1 D(+)-biotin and 20 g L−1 thiamine chloride hydrochloride. For batch growth experiments with carbon source contents above 0.5% (w/v) and for chemostat cultivations, the medium was additionally buffered by the addition of 27.3 g L−1 MES (Pufferan, C6H13NO4S × H2O). Unless stated otherwise, all employed chemicals were purchased from Sigma-Aldrich (Steinheim, Germany) and were of highest purity available. Correlation of optical density to cell dry weight and cell titer Biomass concentrations were always determined by measuring the optical density of cell suspensions with a Genesys 20 Spectronic photometer (Thermo Fisher Scientific, Waltham, USA). For cell dry weight–optical density correlation, 2 mL Eppendorf cups were dried for 72 h at 105°C prior to usage. In order to avoid analytical errors due to moist condensation, the Eppendorf cups were cooled down in a sealed desiccator filled with dry silica gel. Fifty milliliters of H. polymorpha RB 11 cell suspension from mid-exponential growth phase was taken and diluted 1:50, 1:100, 1:200 and 1:400 with fresh medium. The optical density at 600 nm wavelength (OD600) of the diluted cell suspension was determined in quadruplicate, respectively, with a Genesys 20 Spectronic photometer (Thermo Fisher Scientific). After that 0.5, 1 and 2 mL of the respective diluted cell suspensions were transferred into the dried and weighted Eppendorf cups and centrifuged at 13 000 rpm and 4°C for 10 min in a Biofuge fresco centrifuge (Kendro Laboratory Products GmbH, Langenselbold, Germany). The supernatants were carefully discarded; the cell pellets were washed three times with ammonium acetate buffer (50 mM) and centrifuged at 13 000 rpm and 4°C for 10 min. The supernatants were carefully discarded; the cell pellets were dried for 72 h at 85°C. After drying, the Eppendorf cups containing the dry cell pellets were weighed. The differences in weight of empty and cell pellet-filled Eppendorf cups were used to determine the correlation of cell dry weight (CDW) to optical density (OD). For H. polymorpha RB 11 and its recombinant derivatives, the ratio of CDW [g L−1] and optical density OD600 [–] was 0.448 ± 0.03. With a microscope Zeiss Microscope Observer D1 (Carl Zeiss Microscopy GmbH, Göttingen, Germany) and disposable C-Chip hemocytometer DHC‐N01 improved by Neubauer (InCyto, Chungnam-do, Korea), the correlation of optical density and cell titer was determined by counting cells in cell suspensions with know OD600 from 0.1 to 1 in bright-field mode and at ×400 magnification. The ratio of cell titer [108 cells mL−1] to optical density OD600 [–] was 1.116 ± 0.098. Glucose analytics Glucose concentrations in cultivation supernatants were determined by a photometric assay and HPLC analytics. Photometric glucose determination was carried out with an EnzytecTM D-Glucose Kit (R-Biopharm AG, Darmstadt, Germany). The samples for glucose measurements were diluted to concentrations between 0.05 and 0.5 g L−1. Photometric measurements were performed with a Tecan infinite M200 multiwell plate reader (Tecan Group AG, Maennedorf, Switzerland). For HPLC glucose analytics of the cultivation supernatants, an Aminex HPX-87-H column (Bio-Rad Laboratories, Hercules, CA, USA) was used with a LaChrom Elite HPLC system (Hitachi High Technologies America, Inc., Pleasanton, CA, USA). The flow rate of the mobile phase (5 mM sulfuric acid) was set to 1.0 mL min−1. The column temperature was set to 40°C. Detection of the glucose was performed with an L-2420 UV–Vis and an L-2490 refractive index detector. Cultivation workflow All employed cultures were kept at –80°C for long-term storage. For cryopreservation, 1000 μL of an exponentially growing yeast culture in YPD complex medium were supplemented with 250 μL sterile 50% (w/v) glycerol solution, shortly vortexed and immediately frozen at –80°C. For colony plating, small amounts of frozen cell suspension from cryostocks were scratched with a sterile platinum loop, streaked onto YPD agar plates and incubated at 30°C for 48 h. A temperature of 30°C was applied as the standard cultivation temperature for all subsequent experiments as this temperature could be precisely set and controlled with all four employed cultivation technologies. After visible colonies emerged, a single colony was picked with a sterile platinum loop from the agar plate to inoculate a sterile 15 mL Falcon tube containing 5 mL of YPD medium. This preculture was incubated at 30°C and 280 rpm in a rotary shaker (Shaker KS-15, Edmund Bühler GmbH, Hechingen, Germany). After cultivation in complex medium, the overnight preculture was used to inoculate 25 mL of sterile mineral medium (SYN8 or SYN8-MES, supplemented with 0.5%–2% (w/v) carbon source) in a 250 mL shake flask, which was again incubated overnight at 30°C and 280 rpm shaking frequency. With the final overnight precultures, main cultures for batch and continuous cultivations were inoculated. The optical density was generally adjusted to OD600 = 0.1 and the main cultures were used for subsequent experiments. Batch cultivation Batch cultivations were carried out in baffled shake flasks, filled with 10% (v/v) of the nominal volume of the flask. The main culture was inoculated with the overnight preculture to an OD600 of 0.1 and cultivated at 30°C and 280 rpm in a rotary shaker (Shaker KS-15, Edmund Bühler GmbH). Setup and operation of parallelized continuous cultivations in miniaturized chemostats Continuous and parallelized cultivations of H. polymorpha RB11 MOX-GFP were conducted with a customized minichemostat system based on the original setup of Nanchen, Schicker and Sauer (2006). Seventeen milliliters sterile Hungate tubes (16 × 125 mm) (Bellco Glass Inc., Vineland, USA) were sealed with a butyl rubber septum and 13 mm Hungate septum stoppers. The tubes were used as bioreactor vessels with a working volume of 10 mL. The bioreactor tubes were immersed in in a Julabo 5 M water bath at 30°C (Julabo GmbH, Seelbach, Germany) by a Julabo MP thermostat (Julabo GmbH, Seelbach, Germany) to maintain a constant cultivation temperature. For efficient mixing of the cell suspension and dispersion of air bubbles, the bioreactor tubes were equipped with magnetic bars and the whole water bath setup was placed on a custom stage with magnetic stirrer plates below. Aerobic cultivation conditions were maintained by aeration of the tubes with filter-sterilized and pre-humidified air at a flow rate of 20 mL min−1, corresponding to a normalized aeration rate of 2 vvm. Air was delivered with a six-channel peristaltic pump (ECOLINE VC-MS/CA4-12, Ismatec, Wertheim-Mondfeld, Germany), ensuring pressurization of the bioreactor tubes. The filter-sterilized air flow was guided through a Hungate tube filled with 10 mL of sterile water for pre-humidification and minimization of evaporation losses of growth medium by downstream air flow. The off gas was drawn into a moisture trap including a cotton-filled syringe body to prevent contamination via the air outlet tubing. Pressurization prevented contamination of the system as encountered with the original setup of Nanchen et al. and enabled stable long-term cultivations for up to 2 months. Continuous supply of fresh growth medium was facilitated with fused-silica feed capillaries with an inside diameter of 250 μm (TSP250350, Polymicro Technologies, Phoenix, USA) that were pierced through the butyl rubber septum. The capillaries were connected via a luer-lock sterile filter with tubings attached to a high-precision multi-channel peristaltic pump P3 (IPC-N ISM 937, Ismatec, Wertheim-Mondfeld, Germany). A constant level of cell suspension was maintained by placing the tip of a disposable drain tube in an appropriate height inside the bioreactor tubes. Excess cell suspension was continuously withdrawn using another high-precision multi-channel peristaltic pump P3 (IPC-N ISM 937, Ismatec) at a constant flow rate of 0.1 mL min−1. Withdrawn cell suspension was collected inside sterile 500 mL Erlenmeyer flasks. For cell sampling, the rubber septum was sanitized with 70% EtOH and a Bunsen burner was placed beside the reactor system to expose it to the sterile cone of burner flame. After air flow equilibration, 1 mL of cell suspension was collected with a sterile 1 mL disposable syringe and immediately placed on ice. Glucose-limited continuous cultivations of H. polymorpha RB11 MOX-GFP were carried out using buffered SYN8-MES mineral medium containing 0.5% (w/v) of glucose by default. Preparation and application of low-melt agarose pads Low-melt (LM) agarose pads were prepared for the monofocal positioning of individual yeast cells, enabling quantitative epifluorescence time-lapse microscopy without analytical bias due to out-of-focus cells (Young et al.2012). For agarose melt preparation, 75 mg of LM agarose (Carl Roth GmbH & Co.KG, Karlsruhe, Germany) were added to 5 mL SYN8 basic salt solution and dissolved by repeated, gentle heating in a microwave oven operated at 750 W, yielding a 1.5% (w/v) aqueous agarose melt. The hot agarose melt was left to cool at room temperature (RT) for 15 min. Carbon source and heat sensitive medium compounds, such as vitamins and microelement stock solutions, were complemented and gently vortexed for 2 min. For pad casting, an arbitrary smooth surface was covered with hydrophobic Parafilm M® (Bemis NA, Wisconsin, USA) and round 18 × 18 mm cover glass slips with a thickness of 175 μm (VWR International GmbH, Darmstadt, Germany) were placed on the paraffin film-coated surface. The cover slips were cleaned with aqueous 70% (v/v) EtOH and sterile water. Six hundred microliters of agarose melt was pipetted onto the respective glass cover slips, avoiding air bubble formation. The liquid drops of agarose melt were immediately covered by another clean cover glass slip, forming a glass-agarose-glass sandwich. Surface tension of the liquid melt provided a homogeneous thickness of the pad. The agarose sandwiches were covered and left at RT for 45 min to complete polymerization. For cell seeding, the top cover glass slip was carefully removed and 0.5 μL of cell suspension from a mid-exponential H. polymorpha RB11 MOX-GFP culture, diluted to OD600 = 1 with fresh SYN8 medium, were pipetted on the agarose pads. For glucose repression experiments, the cell pellet was washed twice before seeding with ice-cold SYN8 medium, containing the respective amount of glucose that was used during the experiment. After ∼5 min of drying, the loaded agarose pad was flipped and put upside down into a μ-Dish 35 mm microscope dish equipped with a glass bottom (thickness 175 μm) (Ibidi GmbH, Martinsried, Germany). The μ-Dish was sealed with parafilm and immediately transferred to the microscope stage and analyzed. For time-lapse growth and gene expression measurements with agarose pads, the microscope system was set up in a 30°C chamber in order to maintain a constant temperature during cultivation. Single-cell isolation and cultivation employing the Envirostat system Prior to cultivation experiments, the Envirostat microfluidic chip was sanitized with 40% (v/v) EtOH solution. Subsequently, the chip was extensively flushed with sterile cultivation medium in order to remove residual ethanol and to equilibrate and prime the microfluidic network. Dissolved oxygen concentrations of the employed cultivation media were measured with a pO2 electrode (Bioengineering, Wald, Switzerland). For growth experiments with single cells and microcolonies, exponentially growing, non-glucose-limited cells from shake flasks were harvested and diluted to a cell titer of ∼1 × 106 cells mL−1. Five hundred nanoliters of the diluted cell suspension was immediately introduced into the pre-tempered microfluidic chip system. Single cells were isolated from bulk injection samples employing nDEP-cell isolation and microfluidic guidance according to Fritzsch et al. (2013). Singularized cells were contactless trapped in an electrode cage with nDEP. The physical principle of nDEP trapping is based on the induction of a dipole moment in the cell or particle, which physically interacts with an external electrical field (Schnelle, Muller and Fuhr 2000). Depending on the applied frequency of the polarity change, an attracting force by positive DEP or repelling force by nDEP is present (Pohl and Hawk 1966) The cells were positioned equidistant from the eight electrodes in the minimum of the force field. The electrodes were driven by a Cytocon radio frequency generator system (PerkinElmer, Waltham) and controlled with a customized version of the software Switch (PerkinElmer). The octupole cage for cell trapping was operated in ROTX (rotating) mode at 2.6 Vrms and a frequency of 6.25 MHz by default. The ROTX field mode generates a continuously rotating dielectric field that is optimal for observing and measuring the dimension of the trapped single cells. A continuous medium flow of 7.2 nL min−1 for recombinant H. polymorpha RB11 strains was applied to ensure immediate removal of metabolites and unlimited availability of nutrients and oxygen. The medium was conveyed with low pulsation employing syringe pumps SP210IWZ (World Precision Instruments Inc., Sarasota, USA) and Cetoni Nemesys 14:1 (Cetoni GmbH, Korbußen, Germany), equipped with 10 or 50 μL glass syringes (ILS Innovative Labor Systeme GmbH, Stuetzerbach, Germany). Both pump systems were remotely controlled from a computer. Cultivation temperature was kept constant at 30°C with a customized temperature control system basing on TED 200C digital temperature controller (Thorlabs, Newton, USA) equipped with two peltier elements (CP 1.0-63-08L, MELCOR, Wolfheze, the Netherlands), a thermo sensor (AD590, Analog Devices, USA) and a customized water cooling system (Kortmann et al.2009). The temperature shift induced by nDEP Joule heating was compensated with this temperature control system (Jaeger, Mueller and Schnelle 2007). Quantitative fluorescence microscopy All quantitative measurements of epifluorescence intensity in H. polymorpha RB11 MOX-GFP were carried out using an inverted Zeiss Microscope Observer D1 equipped with an Axio Cam MRm CCD camera and a HXP 120 C halogen light source for fluorescence excitation at maximum intensity (Carl Zeiss Microscopy GmbH, Jena, Germany). Microscopic images were taken at a total magnification of ×1000, employing an oil-immersion objective lens Plan-Apochromat 100x/1.40 Oil M27 (Carl Zeiss Microscopy GmbH, Jena). Bright-field pictures were taken with LED illumination at a brightness corresponding to an applied voltage of 4.1 V. As fluorescence filters sets, shift-free 43 HE CY3 and 38 HE eGFP were used. A neutral density filter was used to decrease fluorescence excitation light intensities to 20% or 2%, minimizing phototoxicity and fluorescence bleaching during time-lapse experiments. Image processing was carried out with AxioVision Release 4.8.2, (Carl Zeiss Microscopy GmbH, Jena). Exposure times were either 50 or 200 ms, according to sample fluorescence intensity. RESULTS The glucose-mediated carbon catabolite repression of the MOX promoter in Hansenula polymorpha was investigated with four different cultivation technologies. The employed systems differed with regard to mode of nutrient supply (continuous or batch), total cell number and control over the extracellular environment (Fig. 2). Population analyses were carried out in shake flasks and continuous glucose-limited chemostats. For single-cell analyses, agarose pads and the Envirostat system were used to determine PMOX activity repression/derepression characteristics in controlled environments. All experiments were carried out in SYN8 medium at a cultivation temperature of 30°C for comparing MOX promoter function order and cross-platform comparison. Figure 2. View largeDownload slide The four different cultivation and analysis platforms used in this study for investigating glucose-mediated carbon catabolite repression of the MOX promoter in H. polymorpha. Figure 2. View largeDownload slide The four different cultivation and analysis platforms used in this study for investigating glucose-mediated carbon catabolite repression of the MOX promoter in H. polymorpha. Populations in batch environment: MOX promoter analysis in shake flask cultivations Hansenula polymorpha RB 11 MOX-GFP populations were cultivated in shake flasks with 0.5% (w/v) glucose as sole carbon source. MOX promoter activity in single cells was microscopically assayed as intracellular GFP fluorescence intensity. Promoter derepression as a result of glucose consumption due to the metabolic activity of surrounding cells was monitored during the course of the cultivation. During the initial growth phase on 0.5% (w/v) glucose in batch mode, no GFP synthesis activity in single cells of the culture was detectable. This indicated that extracellular glucose tightly repressed PMOX activity. Besides GFP fluorescence, glucose uptake and suspended biomass were also monitored during the course of cultivation (see Fig. 3). Figure 3. View largeDownload slide Cultivation of H. polymorpha RB11 MOX-GFP in shake flasks. Cells were cultivated in buffered SYN8-MES mineral medium with 0.5% glucose (w/v) as whole carbon source at 280 rpm and 30°C. MOX promoter activity was followed by quantitative fluorescence microscopy. The gray area indicates the standard deviation of the measured GFP fluorescence intensities. Figure 3. View largeDownload slide Cultivation of H. polymorpha RB11 MOX-GFP in shake flasks. Cells were cultivated in buffered SYN8-MES mineral medium with 0.5% glucose (w/v) as whole carbon source at 280 rpm and 30°C. MOX promoter activity was followed by quantitative fluorescence microscopy. The gray area indicates the standard deviation of the measured GFP fluorescence intensities. Samples were taken periodically and confirmed exponential growth of both cultures at specific growth rates of 0.29 and 0.3 h−1, respectively. Glucose was consumed at specific rates of 5.83 and 4.94 mmol gcdw h−1 until the cultures reached the stationary phase after 9 h of growth. Micromolar concentrations of glycerol and ethanol could be detected in cultivation supernatants directly before the cells reached the stationary growth phase. Occasionally, MOX promoter activity in terms of GFP expression could be observed after another 30 min of cultivation. This was most likely due to glucose depletion and the resulting release of carbon catabolite repression. During the next five hours of incubation in the stationary growth phase, an increase in total GFP fluorescence was observed. Fluorescence intensity reached its maximum value at a total cultivation time of 11.5 h. At this time point, the trace amounts of secreted glycerol and ethanol had been completely metabolized. Subsequently, GFP fluorescence slowly faded. Interestingly, GFP fluorescence intensities were generally weak and broadly distributed among single cells of the population. The majority of cells showed no GFP fluorescence despite the fact that extracellular glucose was depleted and hence glucose-depression conditions prevailed. The analysis of the GFP fluorescence distribution revealed the extent of intrapopulation heterogeneity in MOX promoter activity upon glucose depletion in batch cultivations (see Fig. 4). Figure 4. View largeDownload slide Distribution of single-cell MOX promoter activity in a H. polymorpha RB11 MOX-GFP population cultivated in shake flasks. (A) Frequency plots of fluorescence distribution, corresponding to the amount of intracellular GFP during the course of the cultivation experiment. The red curves correspond to a log-normal fit of single cell fluorescence data. (B) Bright-field micrographs of H. polymorpha RB11 MOX-GFP. (C) Fluorescence images of H. polymorpha RB11 MOX-GFP cells corresponding to the shown bright-field images at an exposure time of 200 ms (80% neutral density filter). The number of analyzed individual cells per time point was at least n = 47. Figure 4. View largeDownload slide Distribution of single-cell MOX promoter activity in a H. polymorpha RB11 MOX-GFP population cultivated in shake flasks. (A) Frequency plots of fluorescence distribution, corresponding to the amount of intracellular GFP during the course of the cultivation experiment. The red curves correspond to a log-normal fit of single cell fluorescence data. (B) Bright-field micrographs of H. polymorpha RB11 MOX-GFP. (C) Fluorescence images of H. polymorpha RB11 MOX-GFP cells corresponding to the shown bright-field images at an exposure time of 200 ms (80% neutral density filter). The number of analyzed individual cells per time point was at least n = 47. Interestingly, a few individual cells sporadically showed GFP synthesis activity after depletion of glucose. In the inactive subpopulation, no GFP synthesis could be detected even at highest possible detector sensitivity. The total fraction of cells showing fluorescence above the detection threshold was below 1.3% of the total number of cells. This distribution of PMOX activity did not change at glucose-derepressing conditions upon glucose depletion in the stationary growth phase. Despite thorough monitoring of glucose concentrations in the cultivation supernatants, it was not possible to directly link an extracellular glucose concentration to PMOX derepression. Quantitative conclusions on the connection between glucose, glucose uptake and MOX promoter activity could hence not be drawn. This was due to constantly changing environmental conditions as a result of glucose metabolization during the batch cultivations. Populations in macroscopically controlled environments: MOX promoter analysis in glucose-limited chemostats Continuous cultivations generally enable to define the glucose uptake rate via the dilution rate D. This capability is useful for studying a catabolite repression-regulated promoter system such as PMOX. In comparison to discontinuous cultivations, the extracellular environment is in a macroscopic state of equilibrium, since the growth medium is continuously exchanged. Carbon source and nutrients are continuously resupplied, and the extent of extracellular metabolite accumulation is moderate. In order to determine PMOX repression dynamics with respect to the dilution rate, H. polymorpha cells were cultivated in continuous, glucose-limited chemostats at dilution rates ranging from D = 0.2 h−1 to D = 0.03 h−1. At dilution rates of D = 0.09 h−1 and above, individual cells of H. polymorpha synthesized GFP with a low frequency of occurrence. The fraction of fluorescent cells at these relatively high dilution rates resembled the distribution of GFP-synthesizing cells in the stationary growth phase of the batch cultivation experiments (see Fig. 5). At dilution rates of D ≤ 0.06 h−1, a large fraction of individual cells synthesized GFP. This observation indicated that the repression of the MOX promoter was relieved due to decreased glucose influx. The average MOX promoter activity further increased at a dilution rate of D = 0.03 h−1. Along with increasing GFP fluorescence intensities, intrapopulation heterogeneity in terms of MOX promoter activity increased as well. Figure 5. View largeDownload slide Average GFP fluorescence intensity of single H. polymorpha RB11 MOX-GFP cells cultivated in glucose-limited chemostats in dependence on the dilution rate. The gray areas correspond to the measured standard deviation of GFP fluorescence in single cells. The minimum number of analyzed cells per applied dilution rate was at least n = 52. Figure 5. View largeDownload slide Average GFP fluorescence intensity of single H. polymorpha RB11 MOX-GFP cells cultivated in glucose-limited chemostats in dependence on the dilution rate. The gray areas correspond to the measured standard deviation of GFP fluorescence in single cells. The minimum number of analyzed cells per applied dilution rate was at least n = 52. However, a quantitative connection between GFP synthesis activity and extracellular glucose concentration could not be made with the chemostat experiments. Residual glucose concentrations in the supernatant were generally low and only present in trace amounts. At a dilution rate of D = 0.2 h−1, a sharp decrease in biomass concentration occurred. Despite stable biomass concentrations during the following 10 residence times, residual glucose was detectable in the outflow of the reactor, pointing to an equilibrium state of the culture close to cell wash out. The intrapopulation heterogeneity in terms of MOX promoter activity was investigated for quantifying the activity range of chemostat-cultivated cells (see Fig. 6). Figure 6. View largeDownload slide Distributions of single-cell MOX promoter activity in continuous, glucose-limited cultivations of H. polymorpha RB11 MOX-GFP. (A) Frequency plots of single-cell GFP fluorescence distribution experiments. The red curves correspond to log-normal fittings of the fluorescence intensity data. (B) Bright-field images of H. polymorpha RB11 MOX-GFP. (C) Fluorescence images of H. polymorpha RB11 MOX-GFP cells corresponding to the shown bright-field images at an exposure time of 100 ms (80% neutral density filter). The number of analyzed individual cells per dilution rate was at least n = 52. Figure 6. View largeDownload slide Distributions of single-cell MOX promoter activity in continuous, glucose-limited cultivations of H. polymorpha RB11 MOX-GFP. (A) Frequency plots of single-cell GFP fluorescence distribution experiments. The red curves correspond to log-normal fittings of the fluorescence intensity data. (B) Bright-field images of H. polymorpha RB11 MOX-GFP. (C) Fluorescence images of H. polymorpha RB11 MOX-GFP cells corresponding to the shown bright-field images at an exposure time of 100 ms (80% neutral density filter). The number of analyzed individual cells per dilution rate was at least n = 52. As described before, dilution rates above 0.06 h−1 resulted in strong PMOX repression. Less than 2.1% of the investigated cells exhibited detectable GFP fluorescence. At dilution rates of 0.03 and 0.06 h−1, the MOX promoter was active in more than 99% of the cells. A significant increase in the average specific single-cell fluorescence intensity at D = 0.03 h−1 compared to D = 0.06 h−1 was observed as well. A closer inspection of the fluorescence intensity distribution revealed a pronounced heterogeneity in GFP fluorescence intensities of single cells. The measured intensities deviated more than 500% among individual cells. However, as experienced with batch cultivations, a quantitative connection between extracellular glucose concentration and MOX promoter activity could not be established with glucose-limited chemostat cultivations. Single cells in batch environments: MOX analysis with agarose pads A temporal development of MOX promoter activity may be a response to environmental glucose concentrations. Therefore, agarose pads were used as discontinuous single-cell cultivation systems. Cells were spatially confined in one focal plane when growing on pads. This enabled tracking PMOX activities in individual yeast cells over extended time periods. Agarose pads are cultivation systems resembling the environment in shake flasks. However, the ratio of biomass to available growth medium, and consequently glucose, was several orders of magnitude lower on agarose pads than in shake flasks. The metabolic activity of single cells growing on pads was hence not sufficient to macroscopically deplete glucose. Local depletion in the microenvironment of single cells could not be excluded as metabolized glucose was solely resupplied by diffusion. Agarose pads were prepared with different amount of glucose corresponding to extracellular concentrations from 1 to 0 g L−1. GFP synthesis in single H. polymorpha RB11 MOX-GFP cells was followed during the course of the cultivations with quantitative fluorescence microscopy. Artifacts due to phototoxicity were minimized by choosing extended analysis intervals of 30 min at lowest possible exposure times and fluorescence excitation intensities. Mean intensities of GFP fluorescence of single cells after a total cultivation time of 2 h are given in Fig. 7. Figure 7. View largeDownload slide Average glucose-dependent MOX promoter activities of H. polymorpha RB11 MOX-GFP cells cultivated on agarose pads with different glucose concentrations. The GFP fluorescence intensities were measured after 2 h of cultivation. The gray area corresponds to the measured standard deviation of GFP fluorescence in single cells. The number of analyzed individual cells per concentration was at least n = 44. Figure 7. View largeDownload slide Average glucose-dependent MOX promoter activities of H. polymorpha RB11 MOX-GFP cells cultivated on agarose pads with different glucose concentrations. The GFP fluorescence intensities were measured after 2 h of cultivation. The gray area corresponds to the measured standard deviation of GFP fluorescence in single cells. The number of analyzed individual cells per concentration was at least n = 44. MOX promoter activity was completely repressed at extracellular glucose concentrations from 0.2 to 1 g L−1. At 0.1 g L−1 glucose, approximately 2 out of 100 analyzed cells showed detectable GFP fluorescence. At a glucose concentration of 0.05 g L−1, GFP synthesis was detectable in a large fraction of single cells. This observation indicated that the derepression threshold of the MOX promoter was reached (see Fig. 8E and F). GFP fluorescence further increased with decreasing extracellular glucose concentrations. The maximum average PMOX activity was reached at an extracellular glucose concentration of 0.01 g L−1, but sharply decreased at lower concentrations. Similar levels of PMOX activity were determined for cultivations on agarose pads that contained no glucose (see Fig. 8A–C). Glucose concentrations of <0.01 g L−1 triggered GFP synthesis in more than 99% of the analyzed cells on the agarose pads. The amount of GFP was heterogeneously distributed among the individual cells. In contrast to shake flask and chemostat cultivations, a definite glucose-derepression threshold concentration of 0.05 g L−1 could be determined with agarose pads. An extracellular glucose concentration of 0.01 g L−1 supported maximal PMOX activity. Another important aspect was the fact that agarose pads allowed to analyze PMOX activity in single cells in a time-resolved manner. Derepression kinetics and activity profiles of single cells could be therefore investigated in detail (see Fig. 8). Figure 8. View largeDownload slide Dynamics of GFP fluorescence as a function of extracellular glucose concentration in single H. polymorpha cells cultivated on agarose pads. Concentrations ranging from (A) 0 g L–1 up to (F) 1 × 10–1 g L–1 glucose were used to study MOX activity dynamics in single H. polymorpha RB11 MOX-GFP. The number of investigated cells per condition was n = 20. Figure 8. View largeDownload slide Dynamics of GFP fluorescence as a function of extracellular glucose concentration in single H. polymorpha cells cultivated on agarose pads. Concentrations ranging from (A) 0 g L–1 up to (F) 1 × 10–1 g L–1 glucose were used to study MOX activity dynamics in single H. polymorpha RB11 MOX-GFP. The number of investigated cells per condition was n = 20. GFP fluorescence development as a result of PMOX derepression was tracked in 20 individual cells per applied glucose concentration. Apart from mean fluorescence intensities at the end of the derepression experiment, the time-resolved analysis revealed the kinetic aspects of derepression. In general, MOX promoter activity was detectable 30 min after seeding of the cells to the agarose pads. This coincides with the maturation time of GFP (Chalfie et al.1994; Cubitt et al.1995). Obviously, the transcriptional activity controlled by PMOX is immediately initiated upon glucose derepression. Maximal GFP intensities of single cells occurred after 1.5 h of incubation, with a minor tendency towards a subsequent decrease after 2 h. GFP intensities of individual cells occasionally exceeded mean intensities at low glucose concentrations more than 3-fold. Thus, PMOX regulation was altered in the highly fluorescent cells. A possible explanation may be hidden in intrinsic factors resulting in a reduced sensitivity to carbon catabolite repression of some cells. GFP fluorescence distributions of cells grown on agarose pads after 2 h of incubation time were investigated (see Fig. 9). Figure 9. View largeDownload slide Glucose-dependent heterogeneity of MOX promoter activity in a H. polymorpha RB11 MOX-GFP population cultivated on agarose pads. (A) Frequency plots of fluorescence distribution, corresponding to the amount of intracellular GFP during the course of the cultivation experiment. Red fitting curves represent log-normal distribution data fitting. (B) Bright-field micrographs H. polymorpha RB11 MOX-GFP. (C) Fluorescence images of H. polymorpha RB11 MOX-GFP cells corresponding to the shown bright-field images. The number of analyzed individual cells per glucose concentration was at least n = 44 up to n = 100. Figure 9. View largeDownload slide Glucose-dependent heterogeneity of MOX promoter activity in a H. polymorpha RB11 MOX-GFP population cultivated on agarose pads. (A) Frequency plots of fluorescence distribution, corresponding to the amount of intracellular GFP during the course of the cultivation experiment. Red fitting curves represent log-normal distribution data fitting. (B) Bright-field micrographs H. polymorpha RB11 MOX-GFP. (C) Fluorescence images of H. polymorpha RB11 MOX-GFP cells corresponding to the shown bright-field images. The number of analyzed individual cells per glucose concentration was at least n = 44 up to n = 100. The majority of cells were expressing the GFP gene when glucose was only present in trace amounts in the agarose pads, as observed before in shake flasks and chemostat cultivations. The pronounced intrapopulation heterogeneity indicated that besides carbon catabolite repression as the superordinate regulatory control mechanisms for PMOX transcriptional activity, other intrinsic or extrinsic factors affected promoter activity as well. Changes in the chemical composition of the microenvironment of single cells due to their metabolic activity are also a conceivable cause for the observed heterogeneity. While insignificant in shaken or stirred systems, glucose resupply by diffusion might have become limiting on agarose pads, which lead to glucose depletion and PMOX derepression (Doran 1995). The diffusivity of glucose in agarose gels (D = 0.6 × 10−9 m2 s−1) is close to its aqueous diffusivity. The average surface-specific glucose uptake rate of H. polymorpha is rglucose = 2.34 × 10–8 mol m−2 s−1. Thus, local glucose depletion did not bias the analysis of PMOX derepression (Jones and Bellion 1991; Lundberg and Kuchel 1997; Schmid et al.2010). An isolated cell in a fully controlled environment: MOX promoter analysis with the Envirostat To uncouple carbon catabolite repression from biasing effects due to metabolic activity of surrounding cells, PMOX activity dynamics were investigated in isolated yeast cells with the microfluidic Envirostat system. Continuous perfusion of an isolated cell with fresh growth medium during cultivation and analysis in the Envirostat enabled controlling the glucose concentration in the extracellular environment. The analyzed single cells were dynamically perturbed by changing from a high glucose concentration to a low concentration. In contrast to population-based cultivation approaches, such chemical perturbations could only be carried out at high precision with the Envirostat. In comparison to agarose pads, surface-induced influences on the cellular phenotype could be excluded due to contactless trapping. The methodological concept for the analysis of MOX promoter activity in isolated cells with the Envirostat is exemplarily illustrated in Fig. 10. Figure 10. View largeDownload slide The Envirostat as a continuous and single-cell-based cultivation system for the investigation of MOX promoter dynamics at a single-cell level. A single cell or a budding doublet cell was trapped in a 3D electrode arrangement via nDEP under continuous perfusion with cultivation medium. In the depicted experiment, the perfusion medium was changed from SYN8 with a glucose concentration of 0.5% (w/v) to medium without any glucose right after trapping of the cell. Fluorescence images were taken at an interval of 30 min in the GFP channel with an exposure time of 200 ms (80% neutral density filter). Figure 10. View largeDownload slide The Envirostat as a continuous and single-cell-based cultivation system for the investigation of MOX promoter dynamics at a single-cell level. A single cell or a budding doublet cell was trapped in a 3D electrode arrangement via nDEP under continuous perfusion with cultivation medium. In the depicted experiment, the perfusion medium was changed from SYN8 with a glucose concentration of 0.5% (w/v) to medium without any glucose right after trapping of the cell. Fluorescence images were taken at an interval of 30 min in the GFP channel with an exposure time of 200 ms (80% neutral density filter). PMOX derepression and GFP synthesis kinetics were comparable to those over served with agarose pads, as can be seen from the microscopy images. Detectable GFP fluorescence occurred 30 min after the change to glucose-reduced medium. PMOX activity was measured in 16 individual single-cell experiments at four different glucose concentrations (see Fig. 11) under dynamic perturbation of the cells. A strikingly different response of the MOX promoter activity to glucose was observed in comparison to investigations with the three other cultivation systems (see Fig. 11C and D). The MOX promoter was repressed by trace amounts of extracellular glucose. GFP synthesis was detectable only at 5 × 10−4 g L−1 of extracellular glucose. This was one order of magnitude lower than the determined glucose concentration that allowed PMOX activity during experiments with agarose pads (see Fig. 11B). Full derepression occurred when the isolated cells were perfused with glucose-free medium (See Fig. 11A). At an extracellular glucose concentration of 5 × 10−3 g L−1, one out of four cells showed high and steadily increasing intracellular GFP fluorescence during the course of the cultivation. Most likely, this particular single cell had a reduced susceptibility to carbon catabolite repression (see Fig. 11C). Figure 11. View largeDownload slide Time-resolved analysis of MOX promoter activity in single H. polymorpha cells cultivated with the Envirostat system at different glucose concentrations in the perfusion medium. Figure 11. View largeDownload slide Time-resolved analysis of MOX promoter activity in single H. polymorpha cells cultivated with the Envirostat system at different glucose concentrations in the perfusion medium. Threshold glucose concentrations for glucose-mediated carbon catabolite repression of the MOX promoter Significant differences of population-based and single-cell-based cultivations were striking when comparing extracellular glucose concentrations at which MOX promoter derepression occurred. Full PMOX derepression occurs at an approximated glucose concentration of 1 g L−1 (Suppi et al.2013) according to population-based experiments. Glucose-limited chemostats with baker's yeast reveal that residual glucose concentrations in continuous cultivations are between 0.05 and 0.03 g L−1 at low dilution rates (Diderich et al.1999). Transferred to this study, the extracellular glucose concentration at which PMOX derepression occurred can be estimated as 0.05 g L−1 in continuous cultures of H. polymorpha at a dilution rate of D = 0.06 h−1. This is more than 20 times lower as the corresponding glucose concentration estimated from discontinuous batch cultures in shake flasks. The derepression concentration of extracellular glucose for the MOX promoter was again decreased by an order of magnitude with 5 × 10−3 g L−1 in agarose pads. The MOX promoter was ultrasensitive to glucose-mediated carbon catabolite repression with the ultimate increment of environmental control in the Envirostat system. PMOX activity was only detectable at micromolar glucose concentrations of 5 × 10−4 g L−1 (see Fig. 12). Full PMOX derepression in isolated cells only occurred in the absence of glucose in the perfusion medium. Figure 12. View largeDownload slide Cross-platform comparison of the determined glucose concentrations at which PMOX derepression occurred. Glucose concentrations for PMOX derepression in shake flasks are based on our experiments and supported by data from Suppi et al. (2013). Residual glucose concentrations in glucose-limited chemostat cultivation were estimated based on data from Diderich et al. (1999). Figure 12. View largeDownload slide Cross-platform comparison of the determined glucose concentrations at which PMOX derepression occurred. Glucose concentrations for PMOX derepression in shake flasks are based on our experiments and supported by data from Suppi et al. (2013). Residual glucose concentrations in glucose-limited chemostat cultivation were estimated based on data from Diderich et al. (1999). In summary, the differences between the employed cultivation systems concerning influences of metabolic activity and environmental control were directly reflected in the glucose concentrations at which derepression of the MOX promoter system in H. polymorpha occurred. DISCUSSION Single-cell analyses disclosed the relation of extracellular glucose and MOX promoter activity in Hansenula polymorpha. Environmental control in the Envirostat single-cell analysis system demonstrated that the MOX promoter activity is ultrasensitive to glucose-mediated carbon catabolite repression. The repressing extracellular glucose concentration was determined to be as low as 5 × 10−4 g L−1. Suppi et al. (2013) reported full PMOX activation at glucose concentrations of 1 g L−1 in H. polymorpha. This exceeds the activity threshold concentration determined in this study by more than four orders of magnitude. However, in the cited study, PMOX activity was measured in colonies grown on agar plates over several days. Local glucose depletion in the agar medium as a result of metabolization apparently distorted the obtained results. Previously reported studies on the interplay of glucose and transcriptional initiation by the MOX promoter obviously underrated the biasing effects of population activity (Kramarenko et al.2000). Our results clearly substantiate the influence of different cultivation systems with principally different scales and degrees of environmental control on the experimental readout. It was not possible to link glucose concentration and PMOX activity due to the rapid consumption of glucose by the population in shake flasks. Furthermore, MOX promoter activity was hardly detectable in single cells of the population and decreased again after 2 h of glucose depletion. We speculate that two main reasons are responsible for the observed PMOX activity patterns. First, GFP synthesis was discontinued due to the lack of glucose for anabolic and catabolic reactions. The GFP itself was then hydrolyzed to serve as energy source or to refill depleted amino acid pools. Internally stored carbohydrates that are metabolized to glycolytic intermediates in the stationary growth phase, such as glucose-6-phosphate, could also have repressed PMOX activity in batch cultivations. Second, ethanol was detected in the cultivation supernatants, which is known to have a strong repressing effect on the MOX promoter (Eggeling and Sahm 1978). Ethanol accumulation was likely to pose a main reason for the generally low PMOX activity despite full consumption of extracellular glucose. However, the lack of an extracellular carbon and energy source not necessarily entails interruption of gene expression. Intracellular storage compounds of microorganisms provide sufficient resources for de novo synthesis of protein in the stationary growth phase (Gefen et al.2014). For wild-type H. polymorpha, it was shown that expression of genes from its methanol metabolism is reinitiated after the depletion of glucose in the stationary growth phase (Eggeling and Sahm 1978). Glucose-limited chemostats enabled an estimation of the extracellular glucose threshold concentration for PMOX derepression. However, the determined concentration was three orders of magnitude higher as observed for isolated cells in the Envirostat (Diderich et al.1999). Apparently, glucose concentrations that are sensed by the cells can be neither controlled nor measured with sufficient precision in population-based chemostats. Stirred chemostat cultivation systems feature local substrate and oxygen inputs by feed needles or spargers. Such systems inevitably exhibit substantial concentration gradients within the cultivation vessel due to delayed mixing and substrate distribution effects (Delvigne et al.2006; Lara et al.2006). For example, locally increased glucose concentrations in the vicinity of the medium feed needle can be assumed to trigger MOX promoter repression. This is in contrast to stagnant zones with low mixing. Here, glucose is fully consumed, which in turn entails derepression of PMOX. Regulatory systems such as promoters can rapidly respond to such short-term alterations of environmental conditions (Ozbudak et al.2004; Narula et al.2012). This supports the hypothesis that local glucose gradients could have influenced PMOX activity in chemostats. Despite these obvious technical limitations, glucose supply and hence intracellular glucose-6-phosphate pools could be macroscopically controlled in chemostat cultivations. The MOX promoter activity inversely scaled with the applied dilution rate, hence decreasing extracellular glucose and simultaneously reduced glucose flux into the cells entailed increasing PMOX activities. However, statements on promoter dynamics in individual cells could not be made since bulk cultivation approaches did not permit tracking of promoter dynamics in single cells. In order to further push the analytical limits and minimize the effects of glucose metabolization, agarose pads were used as a discontinuous, single-cell based cultivation platform. Agarose pads offer a high ratio of growth medium volume to total biomass volume. Due to the low cell densities on the pads, the extracellular environment was defined and maintained during the analysis. Glucose depletion in the microenvironment of the cells was considered to bias the response of the PMOX system to glucose on agarose pads. However, a kinetic analysis of surface-specific glucose uptake rates and resupply of glucose by diffusion suggested that no local glucose depletion occurs on agarose pads. The glucose depression concentration for PMOX was determined as 5 × 10−3 g L−1 with the agarose pads. This was again one order of magnitude lower as measured in chemostat cultivations, but higher than determined with the Envirostat system. A couple of possible mechanisms could explain these results. For example, PMOX-inhibiting metabolites accumulated in the vicinity of the cells on agarose pads. Since metabolites were rapidly removed from the extracellular environment of the isolated cells in the Envirostat, secreted metabolites could not influence PMOX activity. Another aspect of interest was how the actual uptake rate of glucose influences PMOX activity. Glucose-6-phosphate is assumed to directly act as a repressing signal molecule for PMOX. The uptake flux of glucose and the kinetics of its subsequent phosphorylation by hexokinases or glucokinases might hence have been a crucial parameter that governed promoter activity (Rolland et al.2000; Laht et al.2002; Genu et al.2003; Suppi et al.2013). The determined glucose threshold concentrations for PMOX carbon catabolite repression were far below the Km value of both low affinity (Km = 0.31 g L−1) and high affinity (Km = 0.011 g L−1) glucose transporter system. This indicates that the net glucose influx is indeed likely to be a determinant for PMOX activity (Karp and Alamae 1998). As previously proposed for Escherichia coli, a metabolic flux sensor for glucose could regulate the extent of carbon catabolite repression in yeast (Kochanowski et al.2013). However, kinetic aspects only come into full effect at low glucose concentrations with uptake velocities far below the Vmax of the hexose uptake system. This substantiates the ultrasensitivity of the MOX promoter system to glucose-mediated repression. Consequently, all cells within the population should have exhibited PMOX activity under strong glucose-derepression conditions. However, substantial heterogeneity in promoter activity in single H. polymorpha cells was observed with all four employed cultivation systems. This indicated principally different susceptibilities of PMOX to glucose repression in individual H. polymorpha cells (Carlson 1999). Differences in gene expression due to gene copy number variations could be ruled out as the origin for this behavior as the MOX-GFP expression cassettes were stably integrated into the host genome. PMOX is a native promoter system in H. polymorpha that tightly regulates expression of the methanol oxidase gene in the intrinsic methanol degradation pathway (Stockmann et al.2009). Thus, host-intrinsic regulatory circuits rather than genetic traits of individual cells are likely to be responsible for differential PMOX activity. The observed phenotypic variability under homogeneous cultivation conditions in the Envirostat could be assigned to stochastic events that act on the regulation of such native transcriptional control mechanisms (Avery 2006). It has recently been shown that host-intrinsic regulatory systems can be particularly prone to heterogeneous activation (Lindmeyer et al.2015). Such differential gene expression can be caused by fluctuations in location and abundance of molecules that are at the top of the regulatory regime (Elowitz et al.2002; Swain, Elowitz and Siggia 2002). Similar stochastic effects are also known to occur e.g. during the cell cycle in yeast (Okabe and Sasai 2007). Phenotypic heterogeneity can also be actively promoted in bet-hedging strategies, which increase the chance of survival upon drastic environmental changes. (Bachmann et al.2013; Nikel et al.2014). In case of H. polymorpha and PMOX, such an unforeseen event could be the sudden presence of actually toxic methanol. These aspects are important targets for further studies that aim at explaining the observed behavior of the MOX promoter. In a technical context, the efficiency of PMOX in H. polymorpha is heavily dependent on the presence of extracellular glucose. Maximizing productivity with this particular expression system requires careful control of glucose limitation for effective PMOX derepression. However, glucose still has to be present to supply sufficient carbon and energy for anabolic and catabolic reactions. Measurements of intracellular glucose-6-phopsphate concentrations in response to different environmental conditions could deliver key insight for improving the efficiency PMOX in process scenarios. With these insights granted by single-cell analysis, new process and strain engineering strategies can now be designed that account for the ultrasensitivity of PMOX to extracellular glucose. Do we understand the control mechanisms of genetic circuits involved in carbon catabolite repression phenomena in yeast? To our knowledge, this is the first study that investigated promoter regulation uncoupled from metabolic population activity in yeast. It could be demonstrated that results on PMOX regulation from apparently controlled bulk cultivations can be significantly misleading. The metabolic activity of surrounding cells was identified as a significant source of bias. This bias can be eliminated by uncoupling environment and cellular physiology with single-cell technologies. By constraining the influencing parameters on the biological system, the functioning of biological elements such as promoters can be mechanistically understood in a quantitative manner. The obtained insight might then be used to understand and even predict the behavior of the biological elements in more complex environments such as populations. The unique opportunities of microfluidic single-cell analysis might thus be exploited for the uncovering of hidden control mechanisms and for optimization of biological systems applied in biotechnological processes. The authors thank for his excellent technical assistance. Conflict of interest.None declared. REFERENCES Ahmad M Hirz M Pichler Het al.   Protein expression in Pichia pastoris: recent achievements and perspectives for heterologous protein production Appl Microbiol Biot  2014 98 5301 17 Google Scholar CrossRef Search ADS   Avery SV. Microbial cell individuality and the underlying sources of heterogeneity Nat Rev Microbiol  2006 4 577 87 Google Scholar CrossRef Search ADS PubMed  Bachmann H Fischlechner M Rabbers Iet al.   Availability of public goods shapes the evolution of competing metabolic strategies P Natl Acad Sci USA  2013 110 14302 7 Google Scholar CrossRef Search ADS   Carlson M. Glucose repression in yeast Curr Opin Microbiol  1999 2 202 7 Google Scholar CrossRef Search ADS PubMed  Chalfie M Tu Y Euskirchen Get al.   Green fluorescent protein as a marker for gene expression Science  1994 263 802 5 Google Scholar CrossRef Search ADS PubMed  Colin A Squires TM Bocquet L Soft matter principles of microfluidics Soft Matter  2012 8 10527 9 Google Scholar CrossRef Search ADS   Cos O Ramon R Montesinos JLet al.   Operational strategies, monitoring and control of heterologous protein production in the methylotrophic yeast Pichia pastoris under different promoters: a review Microb Cell Fact  2006 5 17 Google Scholar CrossRef Search ADS PubMed  Cubitt AB Heim R Adams SRet al.   Understanding, improving and using green fluorescent proteins Trends Biochem Sci  1995 20 448 55 Google Scholar CrossRef Search ADS PubMed  Delvigne F Goffin P Microbial heterogeneity affects bioprocess robustness: dynamic single-cell analysis contributes to understanding of microbial populations Biotechnol J  2014 9 61 72 Google Scholar CrossRef Search ADS PubMed  Delvigne F Lejeune A Destain Jet al.   Modelling of the substrate heterogeneities experienced by a limited microbial population in scale-down and in large-scale bioreactors Chem Eng J  2006 120 157 67 Google Scholar CrossRef Search ADS   Diderich JA Schepper M van Hoek Pet al.   Glucose uptake kinetics and transcription of HXT genes in chemostat cultures of Saccharomyces cerevisiae J Biol Chem  1999 274 15350 9 Google Scholar CrossRef Search ADS PubMed  Doran PM Bioprocess Engineering Principles/Pauline M. Doran  London SA Academic Press 1995 Dunlop EH Ye SJ Micromixing in fermentors: metabolic changes in Saccharomyces cerevisiae and their relationship to fluid turbulence Biotechnol Bioeng  1990 36 854 64 Google Scholar CrossRef Search ADS PubMed  Dusny C Fritzsch FS Frick Oet al.   Isolated microbial single cells and resulting micropopulations grow faster in controlled environments Appl Environ Microb  2012 78 7132 6 Google Scholar CrossRef Search ADS   Dusny C Schmid A Microfluidic single-cell analysis links boundary environments and individual microbial phenotypes Environ Microbiol  2015 17 1839 56 Google Scholar CrossRef Search ADS PubMed  Eggeling L Sahm H Derepression and partial insensitivity to carbon catabolite repression of the methanol dissimilating enzymes in Hansenula polymorpha Eur J Appl Microbiol Biotechnol  1978 5 197 202 Google Scholar CrossRef Search ADS   Egli T Vandijken JP Veenhuis Met al.   Methanol metabolism in yeasts - Regulation of the synthesis of catabolic enzymes Arch Microbiol  1980 124 115 21 Google Scholar CrossRef Search ADS   Elowitz MB Levine AJ Siggia EDet al.   Stochastic gene expression in a single cell Science  2002 297 1183 6 Google Scholar CrossRef Search ADS PubMed  Fritzsch FS Rosenthal K Kampert Aet al.   Picoliter nDEP traps enable time-resolved contactless single bacterial cell analysis in controlled microenvironments Lab Chip  2013 13 397 408 Google Scholar CrossRef Search ADS PubMed  Gefen O Fridman O Ronin Iet al.   Direct observation of single stationary-phase bacteria reveals a surprisingly long period of constant protein production activity P Natl Acad Sci USA  2014 111 556 61 Google Scholar CrossRef Search ADS   Gellissen G. Heterologous protein production in methylotrophic yeasts Appl Microbiol Biot  2000 54 741 50 Google Scholar CrossRef Search ADS   Gellissen G Kunze G Gaillardin Cet al.   New yeast expression platforms based on methylotrophic Hansenula polymorpha and Pichia pastoris and on dimorphic Arxula adeninivorans and Yarrowia lipolytica - a comparison FEMS Yeast Res  2005 5 1079 96 Google Scholar CrossRef Search ADS PubMed  Genu V Godecke S Hollenberg CPet al.   The Hansenula polymorpha MOX gene presents two alternative transcription start points differentially utilized and sensitive to respiratory activity Eur J Biochem  2003 270 2467 75 Google Scholar CrossRef Search ADS PubMed  Gruenberger A Wiechert W Kohlheyer D Single-cell microfluidics: opportunity for bioprocess development Curr Opin Biotechnol  2014 29C 15 23 Google Scholar CrossRef Search ADS   Jaeger MS Mueller T Schnelle T Thermometry in dielectrophoresis chips for contact-free cell handling J Phys D Appl Phys  2007 40 95 105 Google Scholar CrossRef Search ADS   Jones JG Bellion E Methylamine metabolism in Hansenula polymorpha: an in vivo 13C and 31P nuclear magnetic resonance study J Bacteriol  1991 173 4959 69 Google Scholar CrossRef Search ADS PubMed  Karp H Alamae T Glucose transport in a methylotrophic yeast Hansenula polymorpha FEMS Microbiol Lett  1998 166 267 73 Google Scholar CrossRef Search ADS PubMed  Kensy F Engelbrecht C Buchs J Scale-up from microtiter plate to laboratory fermenter: evaluation by online monitoring techniques of growth and protein expression in Escherichia coli and Hansenula polymorpha fermentations Microb Cell Fact  2009 8 68 Google Scholar CrossRef Search ADS PubMed  Kochanowski K Volkmer B Gerosa Let al.   Functioning of a metabolic flux sensor in Escherichia coli P Natl Acad Sci USA  2013 110 1130 5 Google Scholar CrossRef Search ADS   Kortmann H Chasanis P Blank LMet al.   The Envirostat - a new bioreactor concept Lab Chip  2009 9 576 85 Google Scholar CrossRef Search ADS PubMed  Kramarenko T Karp H Jarviste Aet al.   Sugar repression in the methylotrophic yeast Hansenula polymorpha studied by using hexokinase-negative, glucokinase-negative and double kinase-negative mutants Folia Microbiol (Praha)  2000 45 521 9 Google Scholar CrossRef Search ADS PubMed  Laht S Karp H Kotka Pet al.   Cloning and characterization of glucokinase from a methylotrophic yeast Hansenula polymorpha: different effects on glucose repression in H. polymorpha and Saccharomyces cerevisiae Gene  2002 296 195 203 Google Scholar CrossRef Search ADS PubMed  Lara AR Galindo E Ramirez OTet al.   Living with heterogeneities in bioreactors: understanding the effects of environmental gradients on cells Mol Biotechnol  2006 34 355 81 Google Scholar CrossRef Search ADS PubMed  Lindmeyer M Meyer D Kuhn Det al.   Making variability less variable: matching expression system and host for oxygenase-based biotransformations J Ind Microbiol Biot  2015 42 851 66 Google Scholar CrossRef Search ADS   Locke JC Elowitz MB Using movies to analyse gene circuit dynamics in single cells Nat Rev Microbiol  2009 7 383 92 Google Scholar CrossRef Search ADS PubMed  Lundberg P Kuchel PW Diffusion of solutes in agarose and alginate gels: 1H and 23Na PFGSE and 23Na TQF NMR studies Magn Reson Med  1997 37 44 52 Google Scholar CrossRef Search ADS PubMed  Mattanovich D Branduardi P Dato Let al.   Recombinant protein production in yeasts Methods Mol Biol  2012 824 329 58 Google Scholar PubMed  Mayer AF Hellmuth K Schlieker Het al.   An expression system matures: a highly efficient and cost-effective process for phytase production by recombinant strains of Hansenula polymorpha Biotechnol Bioeng  1999 63 373 81 Google Scholar CrossRef Search ADS PubMed  Nanchen A Schicker A Sauer U Nonlinear dependency of intracellular fluxes on growth rate in miniaturized continuous cultures of Escherichia coli Appl Environ Microb  2006 72 1164 72 Google Scholar CrossRef Search ADS   Narula J Devi SN Fujita Met al.   Ultrasensitivity of the Bacillus subtilis sporulation decision P Natl Acad Sci USA  2012 109 E3513 22 Google Scholar CrossRef Search ADS   Nielsen J. Production of biopharmaceutical proteins by yeast: advances through metabolic engineering Bioengineered  2013 4 207 11 Google Scholar CrossRef Search ADS PubMed  Nikel PI Silva-Rocha R Benedetti Iet al.   The private life of environmental bacteria: pollutant biodegradation at the single cell level Environ Microbiol  2014 16 628 42 Google Scholar CrossRef Search ADS PubMed  Okabe Y Sasai M Stable Stochastic dynamics in yeast cell cycle Biophys J  2007 93 3451 9 Google Scholar CrossRef Search ADS PubMed  Ozbudak EM Thattai M Lim HNet al.   Multistability in the lactose utilization network of Escherichia coli Nature  2004 427 737 40 Google Scholar CrossRef Search ADS PubMed  Pohl HA Hawk I Separation of living and dead cells by dielectrophoresis Science  1966 152 647 9 Google Scholar CrossRef Search ADS PubMed  Rolland F De Winde JH Lemaire Ket al.   Glucose-induced cAMP signalling in yeast requires both a G-protein coupled receptor system for extracellular glucose detection and a separable hexose kinase-dependent sensing process Mol Microbiol  2000 38 348 58 Google Scholar CrossRef Search ADS PubMed  Schmid A Kortmann H Dittrich PSet al.   Chemical and biological single cell analysis Curr Opin Biotechnol  2010 21 12 20 Google Scholar CrossRef Search ADS PubMed  Schnelle T Muller T Fuhr G Trapping in AC octode field cages J Electrostat  2000 50 17 29 Google Scholar CrossRef Search ADS   Stasyk OG Maidan MM Stasyk OVet al.   Identification of hexose transporter-like sensor HXS1 and functional hexose transporter HXT1 in the methylotrophic yeast Hansenula polymorpha Eukaryote Cell  2008 7 735 46 Google Scholar CrossRef Search ADS   Stockmann C Scheidle M Dittrich Bet al.   Process development in Hansenula polymorpha and Arxula adeninivorans, a re-assessment Microb Cell Fact  2009 8 22 Google Scholar CrossRef Search ADS PubMed  Suppi S Michelson T Viigand Ket al.   Repression vs activation of MOX, FMD, MPP1 and MAL1 promoters by sugars in Hansenula polymorpha: the outcome depends on cell's ability of sugar phosphorylation FEMS Yeast Res  2013 13 219 32 Google Scholar CrossRef Search ADS PubMed  Swain PS Elowitz MB Siggia ED Intrinsic and extrinsic contributions to stochasticity in gene expression P Natl Acad Sci USA  2002 99 12795 800 Google Scholar CrossRef Search ADS   Teusink B Diderich JA Westerhoff HVet al.   Intracellular glucose concentration in derepressed yeast cells consuming glucose is high enough to reduce the glucose transport rate by 50% J Bacteriol  1998 180 556 62 Google Scholar PubMed  Weibel DB Diluzio WR Whitesides GM Microfabrication meets microbiology Nat Rev Microbiol  2007 5 209 18 Google Scholar CrossRef Search ADS PubMed  Westerwalbesloh C Grunberger A Stute Bet al.   Modeling and CFD simulation of nutrient distribution in picoliter bioreactors for bacterial growth studies on single-cell level Lab Chip  2015 15 4177 86 Google Scholar CrossRef Search ADS PubMed  Yamada Y Maeda K Mikata K The phylogenetic relationships of the hat-shaped ascospore-forming, nitrate-assimilating Pichia species, formerly classified in the genus Hansenula Sydow et Sydow, based on the partial sequences of 18S and 26S ribosomal RNAs (Saccharomycetaceae): the proposals of three new genera, Ogataea, Kuraishia, and Nakazawaea Biosci Biotech Bioch  1994 58 1245 57 Google Scholar CrossRef Search ADS   Young JW Locke JC Altinok Aet al.   Measuring single-cell gene expression dynamics in bacteria using fluorescence time-lapse microscopy Nat Protoc  2012 7 80 8 Google Scholar CrossRef Search ADS   Zurek C Kubis E Keup Pet al.   Production of two aprotinin variants in Hansenula polymorpha Proc Biochem  1996 31 679 89 Google Scholar CrossRef Search ADS   © FEMS 2016. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
TI - The MOX promoter in Hansenula polymorpha is ultrasensitive to glucose-mediated carbon catabolite repression
JF - FEMS Yeast Research
DO - 10.1093/femsyr/fow067
DA - 2016-09-01
UR - https://www.deepdyve.com/lp/oxford-university-press/the-mox-promoter-in-hansenula-polymorpha-is-ultrasensitive-to-glucose-Z9YcWSW0Ll
SP - fow067
VL - 16
IS - 6
DP - DeepDyve
ER -