WO2009132820A1 - Suspension arrays for species-species interaction studies - Google Patents

Abstract

The present invention relates to a porous, gel-like particle for incubation of candidates and sensors within close vicinity to each other comprising one or more sensor and one or more candidate cells within the particle's interior and methods for incubation of candidates and sensors as well as means of applying the invention for discovery and optimization of cells and microorganisms.

Description

Suspension arrays for species-species interaction studies

FIELD OF THE INVENTION

The present invention relates to a device and method for the screening of biomolecules and biological systems that produce such compounds. The invention furthermore relates to the selection, production, and application of biomolecules and cells of microorganisms identified by this method.

BACKGROUND OF THE INVENTION Alexander Fleming's discovery of the first antibiotic for human application hallmarks one the biggest successes of life-science and biotechnological research. An observation as trivial as the suppression of bacterial growth in the vicinity of rampantly-growing fungi led to the discovery of penicillin. Up to now, an uncountable number of human lives have been saved for sake of this observation and for sake of the scientific verve and diligence of Dr. Fleming.

Dr. Fleming recognized that one of the two species dominated the other. In other words, a "sensor" while referring to the bacterial, antibiotic sensitive species responded to a "candidate" namely the fungal, antibiotic producing species.

The concept of employing living microbes or cells as sensors is still of considerable importance for research and development in bio- and life-sciences and application examples comprise the detection of vitamins, amino acids, antibiotics, sugars, or metal-ions. On the technical side, the concept of suppression or induction of a certain sensor-strain located in the vicinity of a candidate-strain is applied in a large number of different formats.

One of the most abundant formats probably still is the cultivation-dish (e.g. Petri-dish). Filled with gel-like solid support and nutrients, said surfaces enable growth of cultivatable microbial species in a petrified fashion. Upon growth, biomolecules such as proteins, metabolites, or chemicals formed upon specific reactions catalyzed by the candidate strain and hereinafter consistently referred to as "biomolecules" can be formed. Said biomolecules may then interact with the co-incubated sensor-strain. In this way, positive as well as negative interactions between two or more species or - more precisely - between the compounds released by the candidates and taken-up or "sensed" by any other means by the sensors can be studied.

If for instance a new antibiotic substance is sought, the sensor might be a pathogenic microorganism or a model for a pathogen and the candidates might be a collection of potential antibiotica producers, e.g. a library of recombinant Streptomyces variants. Positive candidates could then be straightforwardly identified as no sensor strain would grow in their direct neighborhood. If, in another example, a strain for production of a natural amino-acid is sought, the sensor could be an E. coli strain with an auxotrophy for the said amino-acid, and the candidate could be a library of genetically modified potential producers of the said amino-acid. In such a setup a positive feedback regulation is studied since sensor strains would only grow in the vicinity of candidates that do secrete the amino acids that is sought but not in the vicinity of negative ones. In yet another embodiment, the sensor could be a so-called "two- hybrid" that allows the identification of two interaction partners. For instance a genetically modified E. coli or yeast strain expressing a protein serving as a drug target would serve as a sensor or "bait". "Preys" or candidates could be a selection of yeast cells carrying artificial chromosomes and expressing heterologous pathways and therefore possessing the potential to secret compounds that may interact with the protein target presented by the sensor. In principle, an almost infinite number of different experimental setups could be designed for the purpose of indentifying appropriate candidates upon response of vicinal sensors.

However, one of the shortcomings of the approach stems from the difficulty of downscaling and miniaturization of the required experimental setup. As said, currently employed screening and discovery schemes mainly rely on the application of Petri- dishes which are rather space-intense arrays. As a rule of thumb, studying of 10'0OO species-species interactions requires array volumes of at least 100 ml_. Success- rates, i.e. the identification of a positive candidate, however, may require that very large numbers of candidates readily exceeding millions are screened. Array-volumes may therefore become prohibitively large and costs for maintenance, logistics, and analyses of large array volumes may not justify the potential outcome of the experimental efforts anymore.

Alternative approaches for high throughput experimenting and cultivation feature the highly parallelized cultivation of a large number of independent reaction vessels, typically arrayed in a secondary structure and referred to as microtiter plates. Within an individual well of a microtiter plate, strains can be grown in liquid media or on solid support and in a monoseptic fashion. Employed for species-species interaction studies, microtiter plate based assays thus have the potential advantage that undesired cross-talking between candidate strains does not take place. In addition, strains grown in liquid media may express different genes and secrete different products if compared to those grown on solid support thereby broadening the scope of the technology.

Even more importantly, microtiter plates also feature growth of suspended mammalian or insect cells. This is of a considerable value as mammalian cells often serve as in vitro models for certain diseases such as cancer or diabetes Il and are therefore nothing else but disease-specific sensors. Moreover, mammalian cells and in particular recombinant cell-lines may secrete proteins that can up- or down regulate certain pathways in malign cells and also provide a rich source of candidates. The study of species-species interactions of mammalian cells can thus give rise to the discovery of novel proteins or hormones with drug-like properties.

If it comes to the cultivation of cells of microorganisms, microtiter plates do unfortunately not feature an array density higher than 1 '000 to 2'00O experiments per 200 mL array volume (1536 microtiter plate). Technical and physical limitations such as evaporation losses, mixing, wall-wetting, mixing, and the lack of appropriate equipment for accurate dosing cells or microorganisms to microtiter-plate wells restrict the versatility and scalability of microtiter-plate based screenings in practice.

As presented here, there is a need for more effective techniques that overcome the limits posed by the inflexible and bulky reaction arrays currently employed as support for species-species interaction studies. New, space-effective methods for the highly parallel incubation and analysis of a larger number of candidates and sensors, located within close proximity to each other and able to interact by means of diffusion are required. DESCRIPTION OF THE INVENTION

The abovementioned objectives are solved by the device and method for the screening of biomolecules and candidates that potentially produce such compounds. Accordign to the invention, both candidate and sensor are located in a gel-like particle and generate an analyzable signal upon interaction.

More specifically, the invention comprises a device for co-incubation of candidates and sensors within a spatial array, in particular:

a) at least one gel-like particle, comprising one or more sensor and one or more candidate cells within the particle's interior

b) a porous, gel-like structure of the said particle enabling that nutrients can be supplied to the sensor and candidate cells by means of diffusion

c) sufficient stability of the said particle allowing to incubate it under conditions such as required to activate the metabolism of the sensor or candidate strains

d) optical properties of said particles enabling that fluorescence or color or signals s or variations of the particle opaqueness or extinction originating from the sensors and / or the candidates can be recorded

e) a material composition allowing the isolation of candidates from the particle by means and under conditions that guarantee that at least one candidate cell survives that re-isolation procedure or that part or all of the genetic information of the candidates can be reisolated.

The invention disclosed herein relates further to a method for co-incubation of sensor and candidate strains comprising:

a) providing one or more particles, each comprising one or more sensor cells and one or more candidate cells within the particle's interior,

b) supply of nutrient and / or reagents, to said particles c) incubation of said particles under conditions suited to activate the metabolism of the candidates and / or sensors,

d) identification of responding sensors by a spectrophotometrical method or by eye, and

e) isolation of candidates or genetic information derived from the candidates from the particles.

The invention further relates to the manufacture of biomolecules and/or by biological systems identified by the devise or method described hereinbefore, and further to the use of these biomolecules, in particular as therapeutically active agents or food additives.

In a preferred embodiment the present inventions allows the study of several 100'0OO isolated species-species interactions at array volumes of less than 100 milliliter, more preferably of less than 10 milliliter and even more preferably of less than 1 milliliter. Particles are preferably composed of a hydrogel that allows nutrients and reagents to come into contact with the embedded candidates and sensors by means of diffusion. For example, the embedded candidates and sensors can be a bacillus-library screened for formation of new bacteriostatic peptides, and the sensor can be a pathogen E. coli. In this case, the medium could be a complex medium based on e.g. yeast extract, salts, and tryptically digested peptides. In another embodiment, the sensors could be cells of a specific tumor cell line and candidates could be a retroviral Chinese hamster ovary cell-library. In this case the mediums could be a CHO PF medium.

In principles any species can be harbored within an individual particle and even combinations of three or more different species can be applied, as long as the particle provides sufficient space for the embedded cells. Similarly, any medium can be supplied to a particle and to the harbored species as long as particle's integrity is maintained.

In a preferred embodiment some or all components in the medium diffuse into the particles. Once diffused into a particle, media components come into contact with the encapsulated species whereupon some or all of the species within the particle start to grow or become metabolically active. In yet another preferred embodiment, particles are soaked with medium, e.g. by incubation of the particles for a view minutes, and than coated by one or more layers of a certain material. An example of an appropriate layer material is polylysine, which can be readily used for coating of alginate beads. Upon coating, particles may or may not be penetrable for some or all of the components in a medium but species should still respond either with growth or by metabolic activity on the medium present in the capsule prior to coating. In principle any mode of media supply is appropriate as long as some or all of encapsulated species respond with growth or metabolic activity.

Particles are composed of materials and are produced by technologies that have no or only a minor effect on cell viability. An example for an appropriate particle-material is calcium-alginate and an example of a technology employed for particle synthesis is laminar jet-breakup. Respective devices are, for instance, available from company Nisco Engineering in Zurich. The particle may however as well be produced from agarose and by aid of flow-focusing (Nisco Engineering, Zurich) or emulsion polymerization technologies (One-cell Systems, Cambridge, MA). In principle any material and technology can be employed as long as at least one sensor cell and one candidate cell survives the embedment procedure and as long as nutrients can be supplied to the embedded cells.

According to the invention, particles preferably have a spherical shape. Further preferred are spherical particles with a narrow diameter distribution. Even more preferred are particles with a distribution of diameters of less than 30%, and yet even more preferred are particles with a distribution of diameters of less than 5%. The distribution of diameters depends on the particle synthesis methods but it is obvious to somebody skilled in the art, that particle populations with a wide distribution of diameters can be fractionated into populations of narrow distribution of diameters by one several well established techniques such as sieving, flotation, or sedimentation.

The optimal absolute diameter of the spherical particles depends on the analysis method used for the discrimination between positive and negative events. In a preferred embodiment, the analysis is done by flow cytometry and the preferred diameter of the particles is lower than the diameter of the nozzle of the flow cytometric instrument. Further preferred are particle diameters lower than the half of the diameter of the nozzle of the flow cytometric instrument. A preferred diameter is 400 micro m, even more preferred 150 micro m, and yet even more preferred less that 40 micro m.

In a preferred embodiment, candidate and sensors respond to the supply of nutrients by nutrient-consumption. Upon nutrient supply, both sensor and candidate therefore become metabolically active. Metabolically active candidates may then synthesize biomolecules, such as proteins and metabolites, and excrete them into the particle interior. In one embodiment said biomolecules accumulate within the particle. Upon accumulation, concentrations within the particle increase up to a level at which a sensor response is stimulated. In another preferred embodiment, biomolecules leave the particle by diffusion and are subsequently readily diluted in the medium. Yet, as this process is a dynamic equilibrium, a sensor-response of a colony harbored in the same particle as a positive candidate can still be stimulated by biomolecules produced by a candidate. In principle any mode of biomolecule formation as well as any condition within respect to the timely resolved localization of the biomolecules is appropriate as long as biomolecules can mediate a sensor-response.

Even measures for influencing the localization of biomolecules e.g. by promoting their accumulation within the particle can be undertaken. Particles may for instance be incubated in a hydrophobic solvent such as dodecane. Under such conditions, the hydrophobic phase acts as a barrier promoting the accumulation of charged or hydrophilic metabolites or proteins within the particle. In another embodiment, particles may be incubated in an aqueous solute of a slightly acidic or basic pH. As the pH within the particle will thus change as well, net-charges of proteins, chemicals or metabolites are influenced and water-solubility and, thus, mean residence-times of said compounds in the particle interior can be triggered.

In another preferred embodiment, particles are incubated under conditions guaranteeing that both candidate and sensor strain, grow at a pre-adjusted, desired growth rate. In one embodiment, particles are incubated for a certain time under conditions guaranteeing that only the candidate strain grows whereas growth of the sensor strain is halted, thereby allowing the candidate strain to accumulate larger amounts of biomolecules prior to activation, i.e. growth of the sensor. Sensor-growth may than be started by for instance the addition of a supplement, a vitamin, an inducer, or by a medium change. In another embodiment, growth rates may be adjusted such that sensor and candidate strains grow at a similar rate. In this way, both biological events, i.e. the formation and excretion of biomolecules and their uptake or binding by or onto the sensor strain, are harmonized. This is especially important if biomolecule formation rates are of first order, i.e. proportional to the amount of growing candidate cells harbored by a particle. Means for growth harmonization are therefore a convenient technique for guaranteeing that sensors are suseptiveat time at which biomolecules are formed. In one embodiment growth rates are adjusted by selecting an appropriate medium or medium ingredient featuring the desired growth rates for the candidate as well as for the sensor. Measures for influencing the growth by medium design comprise for instance pH adjustments, the choice of a specific carbon source, or the addition of compounds slowing down or accelerating growth of either of the strains. In another preferred embodiment, growth rates are adjusted via the incubation temperature. All cells or organisms have a maximal growth rate at a certain optimal temperature or temperature range. Both, higher or lower incubation temperature as the optimum will therefore lead to exceedingly reduced growth rates. For a larger number of different cells or microorganisms, thus generally temperatures can be found at which both the sensors and the candidate grow at rather similar rates. This is especially important if larger libraries of recombinant variants of cells or microorganisms are screened.

According to the invention, sensor strains are labeled by a fluorescence marker, which however has not to be present at any time. In one embodiment, said markers are fluorescent proteins that are endogenously synthesized by the sensors. Examples for such fluorescent proteins are blue fluorescent proteins (CFP), red-fluorescent proteins (RFP), green fluorescent proteins (GFP), or yellow fluorescent proteins (YFP). In another embodiment, said fluorescent proteins are only synthesized if an appropriate inducer is added to the cells. Examples of such inducers are isopropyl β-D-1- thiogalactopyranoside (IPTG), dicyclopropylketone (DCPK) tetracycline, or galactose. In yet another preferred embodiment, sensors are auto-fluorescent, i.e. fluoresce due to an unknown or not further specified mechanism. An example of an auto-fluorescent sensor is Pseudomonas fluorescence. In another preferred embodiment, sensors are labeled by labeling techniques known in the art. Examples are tagging of sensors by fluorescently antibodies, cell specific dyes or fluorophores. In principle any kind of labeling-technology can be applied as long as the labeling step does not compromise particle integrity.

Sensors can respond on the presence or absence of a biomolecule formed by candidates in many different ways. In one preferred embodiment, sensors respond on the presence of a metabolite or protein with growth. In another preferred embodiment, the sensors respond by synthesizing a certain protein, e.g. a fluorescent protein. In yet another preferred embodiment, metabolites or proteins formed by a positive candidate do inhibit the sensor. Upon inhibition, sensors may stop synthesis of proteins, not start to grow, or even lyse. In yet another preferred embodiment, inhibited sensors may not fluoresce as upon inhibition protein synthesis is halted or cells do not become metabolically active while non-inhibited sensors will readily synthesize fluorescent proteins. Alternatively, the lysed sensors can be fluorescently labeled, for example with propidium iodide. According to the invention, any kind of sensors response is sufficient as long as the response of the sensor can be detected in a fast and reliable fashion.

Preferred methods for the identification of responding sensors and therefore for particles containing potentially positive candidates are high throughput analyses and sorting techniques such as flow cytometry or particle sorting. Examples of appropriated devices are FACS devices (BD Biosciences), COPAS devices (Union

Biometrica), or CyFlow devices (Partec). If a cell or particle sorting technology is employed, particles are preferably kept in suspension prior to the analysis. Positive candidates may then be identified due to the absence or presence of fluorescent sensors within the same particle. In a preferred embodiment, the fluorescent signals of the subpopulations of positive and negative candidates are completely discriminated and the subpopulations can be sorted into pure fractions. Yet in another preferred embodiment of the invention the fluorescent signals of the subpopulations of positive and negative candidates are overlapping and the subpopulations can be sorted into enriched fractions. In another preferred embodiment, responding sensors are identified by aid of an imaging technology or by light- or fluorescence-microscopy. If imaging or microscopy technologies are employed, particles are preferably arrayed on a surface prior to analysis. Examples for appropriate surfaces are microscopic slides, Petri-dishes or illuminated lab-tables. The presence or absence of fluorescent sensors in a particle is then indicative for the presence or absence of a positive candidate strain. After analysis, particles containing positive candidates are rapidly recovered from the surface, for instance by a picking robot or by manual manipulation.

Though not a requirement, it is to be understood that for practical reasons the candidate strains might also be fluorescently labeled. Preferably, candidates have another florescence label as sensor strains thereby enabling discrimination between sensors and candidates upon a fluorescence analysis. The application of fluorescently labeled candidate strains can for instance be advantageous under conditions under which a larger number of sensor strains show an unspecific response, i.e. a respond even in the absence of a candidate. An example for such an unspecific response is the spontaneous reversion of a genetically engineered auxotrophic sensor strain to the prototroph wild-type isolate.

At any stage after synthesis and prior to analysis, particles might be standardized by a presorting step. Such a standardizations-step might for instance be desired in order to adjust an average number of sensors and candidates after synthesis or incubation of the particles. Preferred pre-sorting technologies are cell or particle technologies similar or identical to those employed for identification of positive candidates. A preferred number of candidates per particles is 10 or less, more preferably 3 or less and yet even more preferably exactly 1. A preferred number of sensors per particle is 100 or less, more preferably 30 or less and even more preferably 10 or less.

In the following the invention will be further explained by specific examples which are provided for the purpose of illustration only and are not to be construed as limiting the scope of the claims.

FIGURES Figure 1 : Fluorescent microscopic photographs of 40 micrometer particles with encapsulated microorganisms:

Upper panel: Paenibacillus polymyxa (a bacillus strain producing the antibiotic polymycin active against Gram negative bacteria), in average 3 colonies per bead and E. coli JM101 pGFP (a GFP expressing Gram negative strain), in average 15 colonies per bead.

Lower panel: E. coli JM101 pGFP, in average 15 colonies per bead.

Both types of particles have been incubated 14 hours in 20 % OB media containing 0.3 phenylethanol followed by incubation in TSB media for 6 hours. As indicated by the lower colony diameters and the lower fluorescence, growth of JM 101 pGFP is clearly inhibted in the presence of coencapsulated Paenibacillus polymyxa.

Figure 2: Histogram plot of the green fluorescence peak high measured by a FACS Aria device. 50'0OO events of a 1 to 2 mixture of positive to negative particles (see also Figure 1) were analysed.

EXAMPLES

Example 1: Antibiotic activity of Paenibacillus polymyxa on coencapsulated E. coli JM101 pGFP

Paenibacillus polymyxa is a bacillus strain that produces polymyxin, an antibiotic compound active against Gram negative bacteria.

Two cell suspensions were prepared. One contained only the sensors strain E. coli JM101 pGFP whereas the other one contained both, the sensor strain as well as Paenibacillus polymyxa being representative for a candidate. P. polymyxa was taken from the culture which was incubated in 5 mL ISP medium (40 g yeast extract, 10 g malt extract, 4 g dextrose in 1 L MQ water) for 30 hours on a rotary shaer (30⁰C; 220 rpm) until the OD 600 nm reached 3.58. E. coli was taken from an LB-overnight culture (10 gr bacto tryptone, 5 g yeast extract, 5 g NaCI in 1 L MQ water; 37°C; 200 rpm; OD 600 nm of 8.48). 3 ml of each cell suspension was centrifuged at 13.200 rpm for 2 minutes and resuspended in 1 volume of TRIS pH 7 containing 0.9% NaCI.

Both cell suspensions were thoroughly mixed with two volumes of 3% sterile filtered sodium alginate solution (Sigma, low viscosity) containing 0.9 % sodium chloride. Subsequently, the suspensions were encapsulated using a commercial J30 flow focusing encapsulator unit from NISCO Engineering (nozzle diameter 250 micron; flow rate 0.2 ml/min; pressure drop 115 mbar; hardening solution 100 mM sterile calcium chloride; hardening time 30 min). Afterwards the resulting particles were isolated by aid of a 20 micron sieve.

The particles were washed with sterile water and incubated for 12 hours at 30⁰C in two separate batches in a petri dish as 10% (v/v) suspensions in OB medium (8 g yeast extract, 2 g bacto tryptone, 2 g glycerol in 1 L MQ water) containing 0.3 % phenylethanol. The overnight incubated particles were isolated using a 20 μm sieve and washed thoroughly with water so as to eliminate phenylethanol. The sieved particles were further diluted to 10% (v/v) in YPD medium (25 g peptone, 5 g yeast extract, and 20 g dextrose in 1 L MQ water) and incubated at 30° on a 12 multi well plate. Each well contained 1 ml. of medium and 1 mL of 10% (v/v) diluted particles.

After 6 hours incubation, particles of either batch were withdrawn and photographed by help of an Axiostar plus fluorescence microscope from Zeiss (ex. 480/30 nm, beam splitter 505 nm, em. 535/40 nm; illumination time 300 msec). As indicated (see figure 1), the sensor E. coli JM101 pGFP formed much smaller colonies in particles also containing Paenibacillus polymyxa as the reference strain thereby suggesting that growth was inhibited by antibiotic compounds formed by the candidate.

Fractions of either batch were diluted with 0.9 % sterile filtered sodium chloride to give 5% (v/v) particle suspensions. The particle suspensions were mixed in a ratio of 1 to 2 analyzed by a FACS Aria flow cytometer (BD Biosciences) at a rate of 4'00O - 6'0OO events per second. The beads were gated using the forward and side scattering and the green fluorescence peak high (ex. 488 nm, em. 530/30 nm; PM = 400 V) is plotted as a histogram (see figure 2). From the histogram plot two clearly distinct populations with an event ratio of 1 to 2 can be distinguished.

Example 2: Screening for a strain with improved vitamin production and secretion Vitamin B1 (thiamine pyrophosphate), which can be synthesized by microorganisms, plants, and fungi but not by mammals, is a cofactor of a number of important enzymes in carbohydrate and amino acid metabolism. This example illustrates the application of the invention for the selection of a thiamine-overproducing Bacillus.

Bacillus subtilis is subjected to in vivo mutagenesis by chloramine to a randomly diversified library. The intracellular level of thiamine products in logarithmic- or stationary-phase wild-type Bacillus subtilis is in the order of 100-200 μg/liter, while almost no products are excreted into the medium. For mutagenesis an early stationary culture is diluted ten times by 50 mM Tris-HCI buffer (pH 7) containing 56 μM chloramine and incubated for 30 minutes at 37°C. The mutagenesis procedure is stopped by the addition of 20 mM sodium thiosulfate, the cells are recovered by centrifugation, washed twice with 50 mM Tris-HCI buffer (pH 7), and resuspended in 10 mM Tris-HCI buffer (pH7) containing 0.9% sodium chloride. E. coli JM101 cells constitutively expressing a green fluorescent protein gene are added to the suspension at a ratio of 10:1 (E. coli cells to Bacillus subtilis cells).

The cell suspension is thoroughly mixed with two volumes of 3% sterile filtered sodium alginate solution (Sigma, low viscosity) containing 0.9 % sodium chloride. Subsequently, the cells are encapsulated using a commercial J30 flow focusing encapsulator unit from NISCO Engineering (nozzle diameter 250 micron; flow rate 0.2 ml/min; pressure drop 80 mbar; hardening solution 100 mM sterile calcium chloride; hardening time 30 min) and the resulting gelleous particles are isolated with the help of a 20 micron sieve.

The particles are washed with sterile water and incubated for 24 hours at 37°C in low phosphate minimal medium containing 100 mM Tris-HCI buffer pH7, 0.5% glucose, 0.04% sodium glutamate, 50 mM ammonium chloride, 5 mM potassium dihydrogen phosphate, 1 mM magnesium sulfate, 1 mM calcium chloride, 15O mM sodium chloride, and 1x MT trace elements.

Afterwards, the fraction of the particles with diameter smaller than 40 micron is isolated by sieving and diluted with 0.9 % sterile filtered sodium chloride to give a 5% (v/v) bead suspension. The suspension is analyzed and sorted by help of FACSAria flow cytometer from BD Biosciences at a rate of 6'00O events per second. The Bacillus subtilis strains are released by treating the beads with 50 mM citrate and separated from E. coli JM101 by methods well known in the art. The strains obtained in such a way are characterized with respect to production and excretion of thiamine.

Example 3: Screening for novel antimicrobial compounds

The appearance of multidrug resistant pathogenic microorganisms is a growing problem. This example shows the use of the invention for the discovery and development of novel antimicrobial compounds.

Gallidermin is a lanthionine containing polypeptide antibiotic encoded by a gene cluster in Staphylococcus gallinarum. For the development of novel antimicrobial compounds the cluster is subjected to in vitro mutagenesis by error-prone PCR. The reaction (100 μl) contains 50 mM KCI, 10 mM Tris-HCI (pH 9), 6.5 mM MgCI₂, 0.1%

Triton X-100, 10 μl DMSO, 0.5 mM MnCI₂, 1 mM dNTPs, 15 pM of each forward and reverse primer, 20 ng of genomic template DNA, and 2.5 U of Taq DNA polymerase, and is placed in a Perking Elmer thermal cycler well. After 5 min at 95°C, the thermal cycler performs 25 cycles of the following steps: 1 min at 95°C, 1 min at 55°C, 7 min at 72⁰C. After a final step of 10 min at 72°C the PCR product is purified and cloned into a Staphylococcus vector. The resulting DNA library is then transformed into a

Staphylococcus gallinarum strain that has the gallidermin cluster deleted. Subsequently, the cells are co-encapsulated with a multidrug resistant Staphylococcus aureus harboring a constitutively expressed gene encoding a green fluorescent protein.

For doing so, the cell suspension is thoroughly mixed with an equal volume of 3% sterile filtered sodium alginate solution (Pronova, high G, low viscosity) containing 0.9 % sodium chloride. Subsequently the cell are encapsulated using a commercial Var D laminar jet break-up encapsulator unit from NISCO Engineering (nozzle diameter 150 micron; flow rate 3.8 ml/min; frequency 1050 Hz; hardening solution 100 mM sterile calcium chloride; hardening time 30 min) and the resulting gel-like particles with diameters in the range of 345 to 360 microns are isolated with the help of a 250 micron sieve. The beads are washed with sterile water and incubated for 24 hours at 37°C in medium 21 containing 5% yeast extract, 2% sodium chloride, and 0.5% maltose. After washing with sterile water, the beads are analyzed with respect to their green fluorescence. For doing so, the particles are suspended in water (1000 particles per ml) and analyzed by help of a COPAS Biosorter, from Union Biometrica at a rate of 35 events per second.

The compartments that yield the lowest signals are selected and the synthesized variant polypeptides from the corresponding Staphylococcus gallinarum strains are characterized by sequencing of the corresponding genes. The mutations identified in this way are optionally combined by methods well known in the art, such as DNA shuffling or rational design by site directed mutagenesis. Subsequently, a second cycle of particle production and selection is performed and potential candidates are isolated and characterized.

Example 4: Screening of a metagenomic DNA library

In recent years, metagenomic approaches to identify proteins with specific properties have gained much interest. This example illustrates the application of the invention described herein for the identification of proteins involved in the synthesis of vitamin B12.

Total DNA from marine samples is isolated and fragments of 300-800 kbp are generated by employing a competition of EcoRI and EcoRI methylase. Partially digested DNA is electrophoresed in a 1% Seaplaque GTG agarose CHEF gel to compress DNA fragments >300 kbp into a zone of limited mobility and subsequently isolated. The DNA isolated in such a way is cloned into the pJS97/98 yeast artificial chromosome (YAC) system to generate a YAC library. Subsequently, spheroblasts of Saccharomyces cerevisiae strain YPH252 are generated and transformed by standard techniques and transformants are transferred to DOB-UT agar plates (2% dextrose,

0.17% yeast nitrogen base without amino acids, 0.5% ammonium sulfate, 0.72 g L^"1

UraTrp dropout powder, 2.2% Noble agar). S. cerevisiae harboring the YAC library are mixed with E. coli 113-3 constitutively expressing a gfp gene in a ration of 1 to 10 in 10 mM Tris pH 7 containing 0.9% sodium chloride. Two volumes of 5% agarose in 0.9% sodium chloride previously molten and cooled to 40⁰C are thoroughly mixed with one volume of the cell suspension previously heated to 40°C. This mixture is subsequently dispersed into sterile mineral oil containing 0.2% Span 80 and vortexed at full speed for 30 s and is placed in an ice bath. The particles are isolated with the help of a 20 micron sieve and washed thoroughly with 0.9 % sodium chloride. The particles are then incubated in cobalamine free medium for 24 hours at 30⁰C. After washing the particles are placed on an UV-screen and compartments harboring bright fluorescent colonies of E. coli 113-3 are isolated.

From these particles, the co-encapsulated clones of the YAC library are recovered and characterized further for their capability to produce and excrete vitamin B12.

Examples for bacterial sensor strains

Actithiazic acid by Mycobacterium smegmatis DSMZ 43465, Alanine by Pediococcus pentosaceus DSMZ 20206, long chain Aldehydes by Vibrio harveyi DSMZ 2332, Amino acids Lactobacillus plantarum DSMZ 20205 or Pediococcus acidilactici DSMZ 20238, p-Aminobenzoic acid by Clostridium acetobutylicum DSMZ 792, Gluconobacter oxydans subsp. suboxydans DSMZ 50049, Lactobacillus plantarum DSMZ 20205, Lactobacillus plantarum DSMZ 20246, or Neurospora crassa DSMZ 894, or Arginine by Alcaligenes faecalis subsp. faecalis DSMZ 30030, Escherichia coli DSMZ 301 , Lactobacillus rhamnosus DSMZ 20021 , or Pediococcus acidilactici DSMZ 20238, Aspartic acid by Lactobacillus rhamnosus DSMZ 20245, Azothioprine by Lactobacillus rhamnosus DSMZ 20021 , Benzalkonium chloride Micrococcus luteus DSMZ 1790, Biotin by Lactobacillus pentosus DSMZ 20314, Lactobacillus plantarum DSMZ 20205. Neurospora sitophila DSMZ 1130, or Saccharomyces cerevisiae DSMZ 2155, Chitin synthase by Phycomyces blakesleeanus DSMZ 1359, Choline by Neurospora crassa DSMZ 2968, Coenzyme M by Methanobrevibacter ruminantium DSMZ 1093, Cystine by Lactobacillus plantarum DSMZ 20205, Cytosine by Lactobacillus brevis DSMZ 20556 or Lactobacillus paracasei subsp. paracasei DSMZ 46331 , D(+)-Malate by Pseudomonas fluorescens DSMZ 2764, Folic acid by Bacillus coagulans DSMZ 2308, Enterococcus hirae DSMZ 20160, Lactobacillus rhamnosus DSMZ 20021 , Lactobacillus rhamnosus DSMZ 20022, or Pediococcus pentosaceus DSMZ 20206, Fructose by Lactobacillus fructosus DSMZ 20349, D-Galactose Pseudomonas saccharophila DSMZ 654, Galacturonic acid Paenibacillus polymyxa DSMZ 36, Glucose, isotopic carbon pattern Leuconostoc pseudomesenteroides DSMZ 20193, Glutamic acid by Escherichia coli DSMZ 301 , Lactobacillus rhamnosus DSMZ 20021 , or Lactobacillus rhamnosus DSMZ 20245, Glutamine by Escherichia coli DSMZ 2304, Haloprogin by Candida albicans DSMZ 1386, Histidine by Escherichia coli DSMZ 301 , Inositol by Hanseniaspora uvarum DSMZ 70788, Neurospora crassa DSMZ 1129, or Schizosaccharomyces pombe var. pombe DSMZ 2791, lsoleucine by Lactobacillus plantarum DSMZ 20205, Leucine by Lactobacillus plantarum DSMZ 20205 or Proteus hauseri DSMZ 30118, Lysine by Escherichia coli DSMZ 1099, Lysozyme by Micrococcus luteus DSMZ 20030, Mannose by Pseudomonas saccharophila DSMZ 654, 6-Mercaptopurine Lactobacillus rhamnosus DSMZ 20021 , Methionine by Lactobacillus plantarum DSMZ 20205, Mevalonic acid by Lactobacillus fructivorans DSMZ 20350 or Lactobacillus homohiochii DSMZ 20571 , Myristic acid by Vibrio harveyi DSMZ 2332, Nicotinamide by Lactobacillus fructosus DSMZ 20349, Nicotinic acid Gluconobacter oxydans subsp. suboxydans DSMZ 50049, Lactobacillus plantarum DSMZ 20205, or Lactobacillus rhamnosus DSMZ 20021 , Ornithin decarboxylase by Phycomyces blakesleeanus DSMZ 1359, Panthenol by Gluconobacter oxydans DSMZ 2343 or Pediococcus acidilactici DSMZ 20238, Pantothenic acid by Gluconobacter oxydans subsp. suboxydans DSMZ 50049, Lactobacillus pentosus DSMZ 20314, Lactobacillus plantarum DSMZ 20205, Lactobacillus plantarum DSMZ 20246, Lactobacillus rhamnosus DSMZ 20021 , Morganella morganii subsp. morganii DSMZ 30117, or Saccharomyces cerevisiae DSMZ 70424, Phenylalanine by Lactobacillus plantarum DSMZ 20205, Phenylmercuric acetate byMicrococcus luteus DSMZ 1790, Phenylmercuric nitrate by Micrococcus luteus DSMZ 1790, Pyridoxal by Enterococcus hirae DSMZ 20160, Lactobacillus delbrueckii subsp. delbrueckii DSMZ 20074 or Lactobacillus rhamnosus DSMZ 20021 , Pyridoxamine byEnterococcus hirae DSMZ 20160, Pyridoxine by Hanseniaspora uvarum DSMZ 2768 or Neurospora sitophila DSMZ 1130, Pyrithiamine by Lactobacillus fermentum DSMZ 20391 , Pyruvate oxidation factor by Enterococcus faecalis DSMZ 20409, Riboflavin by Lactobacillus rhamnosus DSMZ 20021 , Serine by Lactobacillus rhamnosus DSMZ 20245, Succinic acid by Claviceps purpurea DSMZ 715, Sulfur by Aspergillus niger DSMZ 1957, Tetrahydrofolic acid by Pediococcus pentosaceus DSMZ 20206, Thiamine by Lactobacillus fermentum DSMZ 20391 or Weissella viridescens DSMZ 20410, Thimerosal by Micrococcus luteus DSMZ 1790, Threonine by Clostridium perfringens DSMZ 798, Thymidine by Lactobacillus paracasei subsp. paracasei DSMZ 46331 , Trace elements by Aspergillus niger DSMZ 2182, Tryptophan by Clostridium perfringens DSMZ 798 or Lactobacillus plantarum DSMZ 20205, Tyrosine by Clostridium perfringens DSMZ 798, Uracil by Lactobacillus brevis DSMZ 20556 or Lactobacillus paracasei subsp. paracasei DSMZ 46331 , Valine by Lactobacillus plantarum DSMZ 20205 or Proteus hauseri DSMZ 30118, Vitamin B1 by Hanseniaspora uvarum DSMZ 70788, Vitamin B12 by Escherichia coli DSMZ 4261 , Vitamin B13 Lactobacillus delbrueckii subsp. lactis DSMZ 20076, Vitamin B6 by Escherichia coli DSMZ 2769, and Vitamin B7 Hanseniaspora uvarum DSMZ 2768.

Examples for mammalian sensors cell-lines:

Mammalian or insect cells used as models and sensors for diabetes II, preferably pancreatic islets, or isolated beta cells, from rats, mice, or humans used for assessing insulin-releasing activity or glucose-dependent expression of the glucagon-like peptide GLP-1 , or isolated human beta-cells, preferably beta TC3 insulinoma for assessing intracellular calcium concentration and HIT-T15 cells for ligand-binding experiments, or primary hepatocytes from rats or humans, preferably used for the identification of inhibitors of the human liver glycogen phosphorylase a (HLPGa), or rat liver hepatocyctes from fastened rats used for identification of inhibitors of the Carnitine Palmitoyltransferases (CPT-1), or muscle cells of cell-lines L6 and C2C12 used as sensors for detecting modulation of glucose-uptake rates, glycogen-syntheses, or aldose reductase inhibition, or primary adipose cells and the 3T3-L1 adipocyte cell line as sensors for modulation of antilipolytic effects, expression of the glucose transporters GLUT1 and GLUT4, transport of amino acids by System A. insulin- stimulated glucose transport, or the transfected Drosophila SL-3 cells used for sensing of compounds modulating the retinoic acid recepot related orphan receptor alpha (RORA), or cell line C3H10T1/2 for measuring transcription of the peroxisom proliferator-activated receptor PPAR gamma. Or mammalian cells used as models as sensors for cancer, preferably K-Ras negative and positive DLD-1 cells used for identification of K-Ras specific, or Fanconi pathway-deficient and proficient cancer cells serving as a model and sensor for Fanconi anemia.

Claims

1. A porous, gel-like particle for co-incubation of candidates and sensors comprising one or more sensor and one or more candidate cells within the particle's interior.

2. The particle according to claim 1 , whereas the sensor cells are labeled by a fluorescent marker.

3. The particle according to claim 2, whereas the fluorescent marker is a protein.

4. The particle according to one of the preceding claims, whereas the fluorescent marker is a substrate or a product from a reaction catalyzed by an enzyme synthesized by the sensor.

5. The particle according to one of the preceding claims, whereas the number of sensors per particle is 1-100, preferably 1-30, more preferably 1-10.

6. The particle according to one of the preceding claims, whereas the number of candidates per particle is 1-10, preferably 1-3, more preferably 1.

7. A suspension array containing particles according to one of the preceding claims whereas more than 100'0OO particles are suspended in a volume of less than 100 ml, preferably less than 10 ml, more preferably less than 1 ml.

8. A method for co-incubation of sensor and candidates comprising: a) providing one or more particles, each comprising one or more sensor cells and one or more candidate cells within the particle's interior; b) supply of nutrient and / or reagents to said particles; c) incubation of said particles under conditions suited to activate the metabolism of the candidates and / or sensors; d) identification of responding sensors by a spectrophotometrical method or by eye, and e) isolation of candidates or genetic information derived from the candidates from those particles that display a sensor response.

9. The method according to claim 8, whereas the sensor yields an increasedfluorescent signal when a positive candidate is present in the same particle.

10. The method according to claim 8, whereas the sensor yields a decreased fluorescent signal when a positive candidate is present in the same particle.

11. The method according to claim 9, whereas the candidates are potential producers or overproducers of antibiotic compounds, amino acids, vitamins, chemicals, peptides or proteins.

12. The method according to claim 10, whereas the candidates are potential producers or overproducers of vitamin B7 (biotin), vitamin B12 (cobalamin), vitamin D3 (cholcalciferol) or metabolites or derivatives thereof.

13. The method according to claim 8, whereas the activity of the metabolism or the growth rate of candidates and/or sensors is adjusted by temperature and/or medium composition.