Method for ultra-high performance screening of biological objects

Abstract

The invention relates to the field of biotechnology and pharmaceutics. Proposed is a method for ultra-high performance screening of biological objects which is based on microfluidic generation of droplets of a biocompatible water-in-oil-in-water double emulsion, and also a method for producing a monodisperse biocompatible water-in-oil-in-water double emulsion. The invention can be used in diagnosing conditions and diseases in mammals, as well as for investigating biological objects.

Claims

1. A method for ultrahigh-throughput screening of a mixture of cells to isolate cells having an antimicrobial activity, the method comprising: generating a double microfluidic water-in-oil-in-water (W/O/W) emulsion comprising monodisperse droplets encapsulating individual cells or mixture of cells in individual droplet microcompartments, which monodisperse droplets have controlled diameters ranging from 20 to 90 μm, wherein said double water-in-oil-in-water (W/O/W) emulsion is generated by sequential microfluidic emulsification of a water phase comprising mixed streams of cells in the mixture of cells and an indicator that changes in fluorescence when activated by cells in the mixture of cells that have antimicrobial activity, wherein the indicator comprises target cells that change in fluorescence when inhibited or killed by the antimicrobial activity of cells in the mixture, and an oil phase to produce a water-in-oil emulsion, followed by microfluidic emulsification of water-in-oil emulsion in outer aqueous phase to produce monodisperse double water-in-oil-in-water (W/O/W) emulsion; isolating a population of monodisperse droplets containing cells having the antimicrobial activity by an ultrahigh-throughput method comprising fluorescence-activated cell sorting (FACS); optionally regenerating or culturing cells from isolated droplets; analyzing the genome or metabolism of the isolated cells having the specific antimicrobial activity; with or without regenerating or culturing—the isolated cells.

2. The method of claim 1, wherein the mixture of cells comprises microbiota.

3. The method of claim 1, wherein the mixture of cells comprises cells exhibiting different types or different levels of enzymatic or biocatalytic activities.

4. The method of claim 1, wherein the mixture of cells comprises a library of cellular phenotypes.

5. The method of claim 1, wherein generating a double water-in-oil-in-water (W/O/W) emulsion comprises monodisperse droplets encapsulating individual living cells, dividing living cells, or a mixture of cells in individual droplet microcompartments, wherein said living or living dividing cells comprise analyzed, isolated or regenerated cells.

6. The method of claim 1, wherein the indicator comprises target cells that are Streptococcus aureus cells.

7. The method of claim 1, wherein the mixture of cells comprises a library of cellular phenotypes having new artificial activities.

8. The method of claim 7, wherein the mixture of cells comprise oral microbiota; wherein the indicator is combination of a fluorogenic/fluorescent compound and a reporter target cell that changes in fluorescence when growth of the target cell is inhibited; and wherein the ultrahigh-throughput method comprises isolating droplets whose combination of fluorescence signals has changed indicating that target cell growth has been inhibited.

9. The method of claim 8, wherein the target cell produces green fluorescent protein (GFP) and its green fluorescence decreases after its growth is inhibited; wherein said ultrahigh-throughput method separates a subpopulation of cells having decreased or no green fluorescence; and wherein said subpopulation of cells have their 16S rRNA or whole genomes sequenced and/or are cultivated, fractionated, and have their components analyzed.

10. The method of claim 9, wherein the subpopulation of cells is cultured, metabolites are fractionated by reverse phase HPLC, and the anti-target cell activity of the fractions determined, and, optionally, have active fractions analyzed by mass spectrometry and have active metabolites identified.

11. The method of claim 1, wherein the indicator comprises a fluorogenic/fluorescent compound changing its fluorescence (level or spectrum) by interacting with the cells having a specific enzymatic or biocatalytic activity.

12. The method of claim 1, wherein the indicator comprises a reporter target cell, wherein fluorescence of the reporter target cell changes when it interacts with the cells having the antimicrobial activity.

13. The method of claim 1, wherein a combination of different fluorescent indicators probes is used for isolation of cells with the antimicrobial activity.

14. The method of claim 1, wherein the antimicrobial activity inhibits the growth of the target s cells.

15. The method of claim 1, wherein the antimicrobial activity kills the target cells.

16. The method of claim 1, wherein the water-in-oil emulsion is generated in a first hydrophobic chip and the resulting single oil-in-water emulsion is reemulsified in external aqueous phase in a second hydrophilic chip to generate the double water-in-oil-in-water emulsion.

17. The method of claim 16, wherein the chips have channels that range in diameter from 20 to 60 μm that generate a monodisperse double microfluidic emulsion having droplets with a controlled diameter ranging from 20 to 90 μm.

18. The method of claim 16, wherein the first hydrophobic chip has been made hydrophobic by treatment with a hydrophobizer that is trichloroctadecylsilane or AQUAPEL®.

19. The method of claim 1, wherein the surface of the second chip has been made hydrophilic by treatment with polyvinylalcohol.

20. The method of claim 1, wherein the ultrahigh-throughput method has a productivity of 1000-20000 events per second.

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) FIG. 1. Microfluidic platform for ultrahigh-throughput screening in droplets of microfluidic double emulsion.

(2) FIG. 2. Generation of double microfluidic emulsion by sequential emulsification in microfluidic chips.

(3) FIG. 3. The design of microfluidic chips used for screening of biocatalytic and antimicrobial activity.

(4) FIG. 4. Microfluidic chip for generation of 20 μm droplets, applied for encapsulation of cells in double emulsion droplets: scheme (top left) and photo (top right) of the general design of the microfluidic chip. Scheme (bottom left) and microscopy (bottom right) of the orifice of a microfluidic chip. Bar size (lower left corner) 0.1 mm.

(5) FIG. 5. Microfluidic chip for generation of 60 μm droplets, applied for encapsulation of cells in double emulsion droplets: scheme (top left) and photo (top right) of the general design of the microfluidic chip. Scheme (bottom left) and microscopy (bottom right) of the orifice of a microfluidic chip. Bar size (lower left corner) 0.1 mm.

(6) FIG. 6. Distribution of cells encapsulated in droplets of microfluidic double emulsion.

(7) FIG. 7. Yeast display of biocatalysts.

(8) FIG. 8. Schemes and kinetic constants of enzymatic reactions, applied for screening of biocatalysts in droplets.

(9) FIG. 9. The principal scheme of platform for ultrahigh-throughput screening of biocatalysts in double emulsion droplets.

(10) FIG. 10. Microscopy of a mixture of 1:10 active and inactive cells. The image was obtained by overlay of visible light and fluorescence microscopy using green fluorescence of product of the enzymatic reaction and the red fluorescence of mCherry fluorescent protein. The scale bar is 100 μm.

(11) FIG. 11. The flow cytofluorimetry of double emulsion droplets with encapsulated cells carrying the antibody Fab fragment, biocatalyst and artificial libraries containing 10% and 1% of active cells with anchored biocatalyst using the EK as an example.

(12) FIG. 12. Plates with clones regenerated from droplets before and after selection using FACS.

(13) FIG. 13. The dependence of the proportion of active cells and enrichment on the dilution of active cells with inactive ones.

(14) FIG. 14. The specific selection of biocatalysts from a mixture of enzymes.

(15) FIG. 15. (A) Crystal structure of BChE. The catalytic triad S198-H438-E325 and the mutated sequence 284-TPLSV-288 is shown in gray and yellow respectively. (B) The activity of individual clones before and after selection using gates G1-G3 with different levels of droplet fluorescence. The inset shows the selection of droplets carrying control cells (gray), the library (red), and wild type (blue) by fluorescence, as well as the fluorescence level of gates G1-G3 used for selection.

(16) FIG. 16. BChE mutants with different levels of activity result in droplets with different levels of fluorescence.

(17) FIG. 17. The selection of BChE mutants with different levels of activity. The proportion of mutants before and after selection (A) for a mixture in 1:1:1:1 ratio and in (B) 1:1:1:1000 ratio, as well as the corresponding enrichment (C) and (D). An asterisk indicates that the clone (cl 3) was not identified in the selected population.

(18) FIG. 18. The scheme of the reaction between and OP, as well as the structures of OPs used. Ki—inhibition constant; k1—phosphylation rate constant; k2—self-reactivation rate constant.

(19) FIG. 19. (A) The kinetics of BChE inhibition by paraoxon using wild type BChE (WT), cl 19 mutant enzyme selected for GDC, and mutants cl 14 and cl 15 selected for resistance against POX. (B) Mutant cl 14, unlike WT, demonstrate catalytic hydrolysis of POX-R. The concentration of both enzymes is 0.45 μM.

(20) FIG. 20. (A) A screening scheme based on cell-cell pairwise interactions for the selection of bacteria that inhibit Staphylococcus aureus (S. aureus) growth in droplets of a biocompatible microfluidic double emulsion. (B) Microscopy of droplets with encapsulated S. aureus+S. venezuelae and S. aureus+E. coli cells. Visible light microscopy, fluorescence microscopy of green fluorescence S. aureus (target), red fluorescence S. venezuelae (left effector) and far-red E. coli fluorescence (right effector), and overlay. The scale bar is 100 μm.

(21) FIG. 21. Plates with bacterial colonies regenerated from droplets before and after selection using FACS-based screening. S. aureus, colored with green, were identified based on their green fluorescence, E. coli—far-red fluorescence, and S. venezuelae—by specific colony morphology.

(22) FIG. 22. (A) The enrichment of S. venezuelae killers depends on its population. The data presented for a 10 and 1% population of S. venezuelae, as well as 1% for E. coli cells. (B) The selection of droplets with the lowest level of green fluorescence results in the selection of droplets with encapsulated S. venezuelae and empty droplets, which reduce the efficiency of selection.

(23) FIG. 23. The enrichment of S. venezuelae killers representing the population of 10 and 1% with S. aureus and E. coli as the target cells.

(24) FIG. 24. Identification of bacteria that inhibit the growth of S. aureus in the oral microbiota. (A) Formation of clearance zones on the plates with individual colonies of the oral microbiota species, overlayed with S. aureus agar. (B) Selection of the most effective inhibitors using plates.

(25) FIG. 25. An improved scheme for the selection of bacteria that inhibit the growth of S. aureus using the oral microbiota as a source of biodiversity.

(26) FIG. 26. Prediction of bacterial genera that inhibit the growth of S. aureus, according to the results of 16S rRNA sequencing.

(27) FIG. 27. Streptococcal clones selected using a microfluidic platform (droplets) showed significantly more efficient inhibition of S. aureus growth than clones obtained using conventional Waxman platform (plates).

(28) FIG. 28. P. aeruginosa inhibited the growth of S. aureus in coculture even at single cell level.

(29) FIG. 29. Fractionation of chloroform extract of P. aeruginosa culture medium and the analysis of the inhibitory activity of the obtained fractions.

(30) FIG. 30. (A) Synergistic inhibition of S. aureus growth by fractions 1 and 6, obtained by fractionation using HPLC. (B) Basic active compounds of fractions 1 and 6, identified using mass spectrometry.

DETAILED DESCRIPTION OF THE INVENTION

(31) In order to provide ultrahigh-throughput screening in droplets of microfluidic double emulsion (FIG. 1), it is necessary to use some highly specific mechanism activated by the desired phenotype. The activated mechanism should lead to a change in the fluorescence used as an analytical signal. The microfluidic compartmentalization of the library of phenotypes together with the mechanism in droplets of a biocompatible double water-in-oil-in-water (W/O/W) emulsion results in many individual microcompartments in which the mechanism can be activated, leading to a change in fluorescence. Thereafter, activator phenotypes are selected using fluorescence activated cell sorting (FACS), after which they can be analyzed by high-throughput sequencing directly without cultivation (which is crucial in the case of unculturable and slow-growing microorganisms) or subjected to cultivation. Cultivable activator phenotypes are subsequently characterized by genotypic, phenotypic, structural and functional analysis.

(32) The generation of a microfluidic double emulsion was carried out by the method of sequential emulsification in microfluidic chips (FIG. 2). The pressure controller supplied compressed air to four reservoirs with liquids: an aqueous phase carrying a library of phenotypes; water phase carrying mechanism; oil phase and external aqueous phase. The pressure of compressed air forced the fluids to flow through the teflon tubes leading to the chip. The flow rate of the fluids was determined by the flow controllers connected by a feedback loop to the pressure controller. Thus, the controller automatically maintained the set flow rate by increasing or decreasing the pressure in the reservoirs. Immediately before entering the chip, streams of water phases carrying a library of phenotypes and a mechanism were combined and entered into a hydrophobised chip.

(33) In the hydrophobic chip, the combined internal aqueous phase stream was split into separate droplets in the oil phase stream, and the resulting single water-in-oil emulsion entered to the hydrophilic chip. In the hydrophilic chip, the water-in-oil emulsion flow was split into separate droplets of a double water-in-oil-in-water (W/O/W) emulsion in a flow of external aqueous phase.

(34) The solution of phosphate-buffered saline, Tris-HCl, ions of divalent metals, growth media could be used as the internal aqueous phase. Mineral oil supplemented with 3% of the emulsifier Abil EM 180 or 2% Pico-Surf2 in Novec7500 fluorocarbon oil can be used as the oil phase. The external aqueous phase may contain 50 mM sodium or potassium phosphate buffer, 2% Pluronic and 0.1% Mowiol 23-88.

(35) The channels of the chips must have a higher wettability with respect to the dispersion medium, otherwise the emulsion generation is unstable. Hydrophobizers (trichloroctadecylsilane or Aquapel) were used for chemical modification while for stabilization of the hydrophilic surface polyvinyl alcohol was used. Mineral oil or fluorocarbon oil was used as the oil phase. Mineral oil has a high viscosity, reduces the transport of gases and leads to a floating emulsion, which complicates its use. In turn, the fluorocarbon oils HFE-7500 and FC-40 had a much lower viscosity, that enable to use them for generation of a double emulsion with a diameter of 20 μm. High gas permeability of fluorocarbon oils was used for screening combined with cultivation, and high density results in the formation of a sedimentation emulsion, more convenient for screening procedure.

(36) In order to ensure the biocompatibility of the emulsion, high molecular weight surfactants with extremely low interfacial transport capacity were used: Abil EM 180 for mineral oil, Pico-Surf 2 for fluorocarbon oil and Pluronic F-127 for the external aqueous phase. The destruction of the emulsion selected after FACS was mediated by its drying on a plate with a semi-solid culture medium, which resulted in 90±20% survival of the yeast cells.

(37) Microfluidic polydimethylsiloxane chips, obtained using soft lithography technology (FIG. 3) were used for microfluidic compartmentalization. The utilization of two chips for a single emulsification was much more simple than application of a single chip for generating a double emulsion, due to their interchangeability and straightforward chemical modification technology. The splitting of the dispersed phase occurred at the chip nozzle, where flows of immiscible liquids were joined.

(38) The chips had a 20 μm (FIG. 4) and 60 μm (FIG. 5) channels, which made it possible to generate a double microfluidic emulsion with a diameter of 20-90 μm. Reduction in the size of the channels results in an increase in throughput. The generation of microfluidic double emulsion occurred with a throughput up to 25000 events per second in the chip with 20 μm channels and 3000 events per second in the 60 μm chip.

(39) As it was mentioned previously, discrete microscopic objects are encapsulated in microfluidic droplets in accordance with the Poisson distribution (FIG. 6) due to the monodispersity of the microfluidic double emulsion. Thus, the lower the degree of filling. (the average number of cells in droplets), the higher the purity of selection resulted from the reduction of probability of coencapsulation of two or more cells into one droplet.

(40) Simultaneously, at λ<0.3, the proportion of droplets filled with cells rapidly decreases, which reduces the efficacy of screening. For λ=0.1, droplets carrying one cell represent 95.1% of the total number of filled droplets, however, 90.5% of the drops remain empty. Thus, to achieve a compromise between purity and screening efficacy, the most optimal is the range 0.3<λ<0.7. For screening of biocatalysts, the filling degree λ=0.5 was used, which corresponds to the maximum purity of selection of 78.7%. On the other hand, if it is required that all the drops are filled (which was used for S. aureus cells in the case of selection of antibiotic activity), it is necessary to use λ>>1. For λ=10, used to fill the droplets with S. aureus cells, the proportion of empty droplets is <0.005%.

(41) Screening Biocatalytic Activity in Droplets of Microfluidic Double Emulsion.

(42) Yeast Display of Biocatalysts and Highly Sensitive Detection of Biocatalytic Activity.

(43) Previously, screening of biocatalysts in droplets based on the principle of in vitro compartmentalization was carried out with approaches based on in vitro transcription/translation, lysis of individual bacterial cells in droplets, as well as bacterial and yeast display. See Agresti et al. Of all the above, the yeast display is the most versatile approach, as it enable folding of complex proteins, formation of correct disulfide bonds, as well as their glycosylation, which was of fundamental importance in the case of rhBChE, which production in the active form is possible only in eukaryotic cells. In order to simplify the identification of the cells that have activity in model experiments, we used the red fluorescent reporter protein mCherry. Simultaneously, it is necessary that the amount of mCherry and the enzyme correlate. To do this, their expression was carried out using a single promoter, and the coding sequences of mCherry and the enzyme were separated by a “self-processing” F2A peptide, providing simultaneous mCherry and enzyme production using a single mRNA transcript mediated by the ribosome skipping. Thus, the genetic construct used for the yeast display of biocatalysts (FIG. 7A) contained the methanol inducible AOX1 alcohol oxidase promoter, the sequence of the red fluorescent reporter protein mCherry, the “self-processed” F2A peptide, the leader peptide of human serum albumin (HSA) for extracellular transport, the enzyme, the hemagglutinin epitope (HA) for immunofluorescent detection of the anchored enzyme and the sequence SAG1, connected by a serine-glycine linker, providing the anchoring of enzyme on a yeast cell.

(44) Three different enzymes demonstrating phosphodiesterase (deoxyribonuclease I—DNase), protease (enteropeptidase—EK) and esterase (butyrylcholinesterase—BChE) activity (FIG. 7B) were used to demonstrate the versatility of the platform. Immunofluorescent staining indicated that all three enzymes were produced in an anchored form (a fluorescence halo was observed) on the surface of yeast cells. At the same time, mCherry localized intracellularly (uniform fluorescence of cells was observed). The yeast producing anchored enzymes and the control yeast producing the Fab antibody fragment were analyzed for biocatalytic activity using appropriate substrates (Ellmann reagent in the case of BChE, Gly-(Asp)4-Lys-naphthylamide in the case of EC and 5′FAM-AAAAAAACCCCCCCATATAGCGCGTTTTTTT-3′RTQ1 (SEQ ID NO: 1) in the case of DNase) and it was shown that all biocatalysts were produced in an active form.

(45) For screening in droplets, it is necessary that the enzymatic reaction product has a high hydrophilicity (which limits its transport through the oil layer to the external aqueous phase), as well as a fluorescence excitation/emission spectrum corresponding to the excitation lasers and emission filters of the cell sorter. Yeasts carrying anchored enzymes had a red fluorescent reporter protein. Hence, green (530/30 nm) and blue (450/50 nm) fluorescence was used to monitor the product. For the three enzymes, different fluorogenic substrates (FIG. 8) were used, based on FRET-pair, non-fluorescent amide, and a conjugate reaction, in which the product of the classical reaction catalyzed by BChE reacted with the non-fluorescent substrate to form a fluorescent adduct. The enzymes had different catalytic efficiencies ranging approximately 100 times toward these fluorogenic substrates. For the suspension of cells with anchored biocatalyst, the ratio of the rates of enzymatic and spontaneous hydrolysis ranged from 550 to 85, resulted in a signal-to-noise ratio of 33 to 6.5 when the reaction was performed in 384-well plates. Immunofluorescent staining and cytometry with fluorescent beads carrying a standard numbers of fluorophore showed that there are 7000±3000 molecules of biocatalysts on the surface of one yeast cell, which made it possible to evaluate the enzymatic constants of anchored biocatalysts.

(46) Selection Efficiency of Biocatalysts from a Mixture of Active and Inactive Cells.

(47) To assess the efficiency of screening of biocatalysts, we used artificial mixtures of active (producing anchored enzyme and red fluorescent reporter protein mCherry) and inactive (producing Fab fragment of the antibody without of the fluorescent reporter) yeast cells (FIG. 9). The compartmentalization of a mixture of active and inactive cells together with a fluorogenic substrate in droplets of a microfluidic double emulsion result in the formation of two populations with different levels of product fluorescence. Selection of droplets with the highest level of fluorescence using FACS led to enrichment with active cells. The microfluidic emulsion was highly monodisperse (FIG. 10), and the accumulation of the reaction product occurred exclusively in droplets carrying active yeast cells.

(48) The model screening of artificial libraries of biocatalysts was conducted in order to evaluate the selection efficiency (the degree of enrichment) for each biocatalyst after ultrahigh-throughput screening of droplets of a water-in-oil-in-water emulsion using FACS. Four samples were encapsulated together with the corresponding fluorogenic substrate under the same conditions: 1) yeast with anchored antibody Fab fragment (drops in which only the background level of the reaction occurs), 2) yeast with anchored biocatalyst (drops in which only the catalytic hydrolysis of the substrate takes place) 3) Biocatalyst/Fab 1:10 mixture, 4) Biocatalyst/Fab 1:100 mixture. After encapsulation, the samples were was analyzed and sorted using a BD FACSAria III FACS (FIG. 11).

(49) The population of droplets with a high level of fluorescence was selected using FACS in the mode of maximum purity of the screening with a productivity in the range of 10 000-20 000 events per second after the reaction took place in drops. The selected droplets were plated on agar plates and after 2-3 days the formation of yeast colonies was observed. Colonies regenerated from drops before and after selection were transferred on plates with methanol, where induction of expression of fluorescent reporter was observed (FIG. 12). As a result, efficient enrichment of each of the biocatalysts was observed.

(50) Ultrahigh-throughput screening of droplets was accomplished with different ratios of active and inactive cells in the range of 1:10-1:10.sup.5 (FIG. 13) in order to determine how the efficiency of screening depends on the ratio of active and inactive cells. The degree of enrichment of active cells from a mixture with inactive was close to the theoretical limit in the case of a 1:10 and 1:100 dilution. Even at extremely high 1:10.sup.5 dilution, more than 35% of active clones were observed after one round of selection, which is only two times less than the theoretical maximum.

(51) Thus, for cells with biocatalytic activity, an extremely high efficiency of selection from a mixture of active and inactive cells was shown using only one round of screening.

(52) Selectivity of the Selection of Biocatalysts from a Mixture of Biocatalysts with Different Specificity or Different Levels of Activity.

(53) The “artificial activity library” was used for the screening of biocatalysts from a mixture of cells displaying different enzyme activity and substrate specificity. “Artificial activity library” was prepared by mixing the yeast, carrying separately each of the three biocatalysts previously used and diluted 100 times with inactive cells. Thus, the proportion of cells of each type of activity was less than 1% in the mixture. The activity library was selected for each type of enzymatic activity separately using the corresponding fluorogenic substrate (FIG. 14). Cells with each type of biocatalytic activity were specifically selected for a given enzymatic activity with an efficiency close to the theoretical maximum, without enrichment with cells carrying a different type of activity.

(54) The BChE library was used to investigate how the efficiency and selectivity of selection depends on the activity level of the biocatalyst. The representativity of the library was about 3×10.sup.5 variants. The BChE library was made by randomization of 5 consecutive amino acids in the sequence of the wild type BChE (WT) 284-TPLSV-288 in acyl-binding loop located close to the active center of enzyme (FIG. 15A). The amino acid substitutions significantly reduced the activity of clones. The average library activity was less than 0.5% of activity of the wild type BChE. The yeast with the anchored BChE mutants were encapsulated together with the fluorogenic substrate in droplets of a microfluidic double emulsion. The resulting drops had different levels of fluorescence. They were selected using three G1-G3 gates with different levels of fluorescence G1>G2>G3 (FIG. 15B).

(55) The analysis of the activity of individual clones obtained before and after selection indicates the efficiency of the selection in the case of using each of the three gates. At the same time, the clones selected from the droplets with the highest level of fluorescence had the highest level of activity. Thus, it was qualitatively shown that the developed platform can be used to select biocatalysts with different levels of the same activity.

(56) In order to quantify the efficiency of selection of with different levels of the same activity, three clones with different levels of BChE activity, selected by biocatalysts using ultrahigh-throughput screening with gates G1-G3 were used. Mutant cl 13, selected from G3, had an insignificant level of activity (1.1% of WT BChE). This activity was approximately 2 times above the background activity of control cells measured in the 384 well plate. Mutant cl 8 and cl 3 were selected from G2 and G1, respectively, and had 5.4 and 55% of the WT BChE activity.

(57) The compartmentalization of cl 3, cl 8, cl 13 and control cells together with the substrate in the droplets led to the formation of four populations with different levels of fluorescence, corresponding to different levels of accumulated product: cl 3—high, cl 8—medium, cl 13—low (FIG. 16). The BChE mutants cl 3, cl 8 and cl 13, as well as control cells (Fab) were mixed with each other in an equal ratio and in the ratio of 1:1:1:1000 and screened using the gates “High”, “Medium”, “Low” (FIG. 16), corresponding to the high, medium and low fluorescence of the product. For the 1:1:1:1 ratio, we observed a specific enrichment of each of the BChE mutants with an efficiency close to the theoretical maximum (FIG. 17A, C). Simultaneously, in the case of a 1:1:1:1000 ratio, the enrichment of cl 3 and cl 8 also occurred with an efficiency close to the maximum, while the cl 13 enrichment efficiency was lower for an order of magnitude (FIG. 17B, D).

(58) Thus, if the concentration of cells with activity 2 times higher than background represent a population of more than 0.1%, they can be selected in a single round of screening. At the same time, in the case of cells displaying 5 times higher activity than the control, the efficient selection in single round of screening is possible even for the population which is less than 0.1%. However, the proportion of active cells may be less than 0.001% in the case of highly active wild type DNase, EK and BChE enzymes. The selection of droplets with medium and low fluorescence leads to the efficient elimination of highly active cl 3. This strategy can by used to ultrahigh-throughput identification of mutations leading to the loss of activity of enzymes.

(59) The Creation of New Biocatalysts with a Predesigned Activity Using Ultrahigh-Throughput Screening. Paraoxon (POX)

(60) The BChE library previously described was used for the ultrahigh-throughput selection of BChE mutants, resistant to organophosphates (OP). Yeast cells with anchored BChE mutants were incubated with OP paraoxon (POX) or soman coumarin analogue (GD.sub.C), washed out from the excess of OP and screened for the residual BChE activity.

(61) After the single round of screening, two mutants that showed resistance to POX inhibition (cl 14 and cl 15), and six mutants (5 copies of cl 19 mutant and one cl 14) with increased GD.sub.C resistance were selected. These mutants were sequenced and produced in mammalian cell line FreeStyle™ 293-F for kinetic studies to determine the reason of the resistance observed. The interaction between wild type BChE and OP (FIG. 18) leads to the fast (with a high bimolecular rate constant) and irreversible (k2˜0) inactivation of BChE resulted from phosphorylation of S198 in the active center.

(62) Unlike wild-type BChE (WT), cl 14 and cl 15 exhibited residual BChE activity even after prolonged inhibition by POX (FIG. 19A). Moreover, in contrast to WT, cl 14 and cl 15 showed catalytic hydrolysis of OP, which was identified using the fluorescent POX-R analog (FIG. 19B). After a rapid covalent modification of cl 14 and WT, leading to the release of a stoichiometric amount of resorufin, WT remains completely modified while cl 14 continues to react slowly with new POX-R molecules, which ultimately results in complete hydrolysis of POX-R.

(63) Thus, the mutants cl 14 and cl 15 display a new paraoxonase activity mediated by self-reactivation (k.sub.2>0). Mutant cl 14 had a high reactivity toward POX comparable with WT. Mutant cl 15 also demonstrated the catalytic hydrolysis of POX-R. While, in contrast to cl 14, cl 15 had a significantly lower reactivity toward POX (Table 1).

(64) TABLE-US-00001 TABLE 1 Kinetic constants of wild type BChE (WT) and mutants selected using ultrahigh- throughput screening according to their resistance against POX and GD.sub.C. POX Negative GD.sub.C Negative OP Selected control Selected control BChE WT cl 14 cl 15 cl 19 WT cl 14 cl 19 cl 15 k.sub.1/K.sub.i × 10.sup.2, M.sup.−1s.sup.−1 290 ± 30 260 ± 40  25 ± 4  140 ± 20 3 ± 1 0.5 ± 0.1 0.05 ± 0.01 2.0 ± 0.5 k.sub.2 × 10.sup.4, s.sup.−1 ND 1.0 ± 0.1 1.1 ± 0.1 ND ND ND ND ND

(65) Mutants selected for resistance to GD.sub.C did not display the catalytic hydrolysis of GDC, which was associated with the extremely rapid aging of soman adducts (t.sub.1/2<1 minute) compared to POX (˜8.6 hours) observed for WT BChE. In turn, the increased resistance of the mutant cl 19 selected for the resistance against GD.sub.C, was associated with more than 60 fold decrease in reactivity toward GD.sub.C.

(66) Thus, in one round of ultrahigh-throughput screening of a relatively small library with representativity of ˜3×10.sup.5 enzyme variants, new mutants resistant to OP inhibition were selected. Moreover, in the case of POX, BChE mutants with new artificial catalytic activity were created de novo. As a result, new catalytic antidotes were obtained based on the stoichiometric antidote BChE.

(67) Screening of Antibiotic Activity in Droplets of Microfluidic Double Emulsion

(68) The Model System for Discovery of Pairwise Interactions of Microorganisms.

(69) Biocompatibility is one of the most attractive advantages of the developed platform, which allows the cultivation of microorganisms in droplets of microfluidic double emulsion. To study the pairwise intercellular interactions of microorganisms in droplets, a simple model system was developed that simulates the interactions between three different types of cells that played the role of “target”, “killer” and “mate” (FIG. 20A). The developed scheme was used for the selection of bacteria producing antibiotics that inhibited the growth of pathogenic bacteria Staphylococcus aureus. The S. aureus strain, constitutively producing a green fluorescent reporter protein (GFP), was used as a “target”. Bacteria Streptomyces venezuelae, producing the antibiotic chloramphenicol and inhibiting the growth of S. aureus, were used as a “killer”. S. venezuelae produced a wide range of secondary metabolites, including prodignins with red fluorescence, which were used to identify their presence in the drops. E. coli cells were used as a neutral “mate” who did not inhibit the growth of S. aureus. E. coli cells produced the far-red fluorescent reporter protein Katushka2S.

(70) Thus, if S. aureus and S. venezuelae cells were coencapsulated in one droplet, S. venezuelae inhibited the growth of S. aureus and filled the entire volume of the droplet, leading to the formation of a droplet with a high level of red and a low level of green fluorescence. In the case that only S. aureus cells were encapsulated in the droplet, they were divided uncontrollably leading to the formation of a droplet with intense green fluorescence. If S. aureus and E. coli cells were coencapsulated in one droplet, both bacteria demonstrated division, which led to the droplet with a high level of green and far-red fluorescence. After cocultivation in droplets, the selection of droplets with the lowest level of green fluorescence led to the enrichment of S. venezuelae, but not E. coli. The expected results were observed in droplets (FIG. 20B). Every droplet with S. venezuelae, obtained after coencapsulation of S. aureus and S. venezuelae cells in droplets was characterized by intense red fluorescence but not green fluorescence. Most of the droplets that did not have red fluorescence had an intense green fluorescence. On the contrary, in course of coencapsulation of S. aureus and E. coli cells, the droplets with E. coli cells had intense far-red and green fluorescence simultaneously. Subsequent ultrahigh-throughput screening of droplets with the lowest level of green fluorescence using FACS led to the enrichment of S. venezuelae, but not E. coli, which was identified by cultivation of droplets before and after selection on plates (FIG. 21).

(71) The selection of biocatalysts was carried out on the principle of positive selection, i.e. selection of droplets with the highest level of fluorescence. However, unlike the positive selection, the negative selection of the least fluorescent droplets resulted in a significantly lower degree of enrichment (FIG. 22A). In the case of positive selection, an effective selection of active cells was observed even with their population of about 0.001%. In the case of negative selection, enrichment of killers, but not mates was observed. At the same time, the selection of killers with an efficiency close to the theoretical maximum occurred only if their population was more than 10%. If the proportion of S. venezuelae was less than 1%, the enrichment efficiency decreased.

(72) The observed decrease in the efficiency of selection was associated with the presence of a population of empty droplets, as well as drops in which S. aureus division did not occur due to their transition to persisters or death (FIG. 22B). In order to demonstrate that the utility of this platform is not limited to S. aureus cells as a target, E. coli cells were used. It was also shown that in the case of E. coli cells, inhibited by S. venezuelae, the negative selection leads to low enrichment efficiency of S. venezuelae, if their population is less than 1% (FIG. 23). The results suggest that to provide highly efficient screening based on negative selection of the effector cells representing a population of less than 1%, additional fluorescent reporters based on positive selection are necessary.

(73) Screening of the Oral Microbiota for Selection of Bacteria Inhibiting the Growth of S. aureus

(74) Despite its extremely high pathogenicity, S. aureus is rarely associated with dentoalveolar infections. Hence, there are some unknown natural effectors of the oral microbiota that inhibit the growth of S. aureus. The use of classical bacteriological screening approaches, based on plates (the so-called Waxman platform) showed the presence of bacteria that inhibit the growth of S. aureus in the oral cavity (FIG. 24A). The selected inhibitors showed different sizes of clearance zones (FIG. 24B). Clones with the clearest areas of the largest diameter were used for further work.

(75) In contrast to the classical Waxman platform, the developed microfluidic platform allows screening of considerably higher biological diversity, which was used to select the bacteria that inhibit the growth of S. aureus among the representatives of the oral microbiota species. For this, the screening scheme developed previously for pairwise interactions in droplets was modified by two additional fluorescent reporter signals to avoid the problem associated with the selection of empty drops and droplets with a low number of encapsulated S. aureus cells (FIG. 25).

(76) The cells of oral microbiota were coencapsulated together with an excess of S. aureus cells labeled with a red fluorescent dye sulfocyanin 5 (sCy5) in droplets of microfluidic double emulsion. The cocultivation of S. aureus and microbiota effectors resulted in four different variants I-IV. I—effector inhibited the growth of S. aureus and remained alive in the process of cocultivation. II—effector and S. aureus died. III—effector and S. aureus were cocultivated in a droplet and did not inhibit the growth of each other. IV—S. aureus inhibited the growth of the effector. In order to distinguish between these variants, Calcein Violet AM, a non-fluorescent, hydrophobic acetoxymethyl ester capable of penetrating through a layer of oil phase, was used.

(77) If the living cells are present in a droplet, their esterases hydrolyze acetoxymethyl ether, which in turn leads to the formation of a highly hydrophilic product with intense blue fluorescence, incapable of transport through the hydrophobic oil layer. Thus, the selection of a population of droplets with a high level of red, low green and high blue fluorescence leads to the selection of droplets with a high initial load of S. aureus cells that did not proliferate in droplets, while living cells other than S. aureus were present in the droplets.

(78) Since different bacteria have different growth rates and can inhibit the growth of each other on the plates, the selected drops were subjected to direct 16S rRNA and whole genome (WGS) high-throughput sequencing in order to identify slow-growing and unculturable bacteria that inhibit S. aureus growth in the droplets. The comparison of the number of readings before and after selection showed that, according both to the results of 16S rRNA sequencing (FIG. 26) and WGS, Propionibacterium acnes was the most enriched bacteria.

(79) 16S rRNA sequencing revealed two subpopulations of bacterial inhibitors selected with different efficacy. Bacteria that belong to the genera Propionibacterium, Stenotrophomonas, Sphingomonas, Pseudomonas, and Escherichia had the highest level of enrichment, while Corynebacterium, Janthinobacterium, Serratia, Enterobacter and Staphylococcus were also significantly enriched. Staphylococcus was the most common inhibitor of S. aureus among the oral microbiota. WGS confirmed the extremely efficient enrichment of slow-growing Propionibacterium acnes, also allowing the identification of inhibitor bacteria species. The Staphylococcus mitis group (S. pneumoniae, S. mitis, S. oralis and S. pseudopneumoniae), Prevotella dentalis (slow-growing), Staphylococcus epidermidis (well-known S. aureus effector) and Pseudomonas aeruginosa were significantly enriched after the selection with ultrahigh-throughput screening in droplets.

(80) The selected droplets were also cultivated to identify culturable S. aureus inhibitors. More than 90% of bacterial colonies, other than S. aureus, regenerated from droplets on the plates, belonged to the genus Staphylococcus. More than 64% of them were classified by mass spectrometry as Staphylococcus oralis. Selected clones produced metabolites that inhibit the growth of S. aureus. Staphylococcus oralis showed the greatest inhibitory dilution of the growth medium (up to 16 fold dilution). The clones obtained by selection with ultrahigh-throughput screening in droplets had a significantly higher inhibitory dilution of the growth medium than clones obtained using classical plate screening (FIG. 27).

(81) Cocultivation of diluted (<10.sup.6 CFU/ml) culture of Staphylococcus oralis and S. aureus, however, did not lead to inhibition of S. aureus growth, which, apparently, is associated with a much faster kinetics of S. aureus growth. At the same time, the physiological significance of Staphylococcus oralis can be very high. The real conditions differ significantly from what was observed during cocultivation, since saliva is a much poorer growth medium, and local concentrations of streptococci on the gum surface are much higher, than those that were used in liquid culture. In addition, effective inhibition of S. aureus was observed on plates, which also suggests that streptococci, in particular Staphylococcus oralis, can play an important physiological role as a protective barrier preventing the colonization of S. aureus.

(82) The Pseudomonas aeruginosa strain was selected exclusively using the microfluidic platform, completely inhibiting the growth of S. aureus in culture even on a single cell level (FIG. 28). Opportunistic pathogen P. aeruginosa is not a normal component of the oral microflora. It was present in oral microbiota in an extremely low concentration of <0.005%. This made its selection impossible using the classical Waxman platform. S. oralis, unlike P. aeruginosa, produced secondary metabolites that inhibit the growth of S. aureus inducibly in the process of cocultivation with S. aureus. The maximum level of induction corresponded to the minimum concentration of P. aeruginosa in the growth medium during cocultivation (FIG. 28) and correlated with blue fluorescence of pyoverdine. However, the growth inhibition of S. aureus was not mediated by pyoverdine.

(83) The chloroform extract of P. aeruginosa growth medium obtained after cocultivation of an excess of S. aureus over P. aeruginosa, was fractionated using reverse phase HPLC (FIG. 29). P. aeruginosa produced a wide range of secondary metabolites that inhibited the growth of S. aureus.

(84) Fractions 1 and 6 demonstrated highly efficient synergistic inhibition, leading to the death of S. aureus (FIG. 30A). The main active substances identified by mass spectrometry (FIG. 30B) were pyocyanin and phenazin-1-carboxylic acid (fraction 1) and 2-heptyl-4-hydroxyquinoline N-oxide (fraction 6). Pyocyanin and phenazin-1-carboxylic acid are catalase inhibitors, also acting as electron transport chain carriers, mediating the production of reactive oxygen species. While 2-heptyl-4-hydroxyquinoline N-oxide inhibits the oxidation pathway of hydroquinones. Thus, P. aeruginosa controls the growth of S. aureus using synergistic mechanisms that target the induction of oxidative stress and leading to the S. aureus death.