The present invention relates to a method to select variants of an enzyme wherein the variants show increased activity under a certain environmental condition. Cells with that enzyme variant are encapsulated by a hydrogel, which serves as a selection marker.

Backqround

Compartmentalization is a defining feature of living systems. Cells, tissues and organs exhibit an extraordinary degree of spatial segregation in form and function, which is necessary for carrying out the daily activities of life. In an effort to advance synthetic biological systems, researchers have focused on developing compartmentalization strategies that function at the molecular-to-cellular length scales. Along these lines, encapsulation of individual cells inside synthetic conformal hydrogels has attracted significant attention. The ability of conformal synthetic hydrogels to serve as optical labels, act as steric protective layers, and separate distinct cell types make these systems potentially useful for rare cell isolation and in vivo cellular therapies.

Several early demonstrations of cellular encapsulation technology utilized formation of multicellular aggregates on the scale of 0.1 -1 mm relying on, for example, self-assembly of oppositely charged polyelectrolyte chains or incorporation of cells in agarose emulsions. Polymerization-based approaches have also been developed, which rely on cross-linking hydrogel materials from monomeric precursor solutions. Photoinitiators have also been conjugated to antibodies that specifically recognize target cells and generate free radicals upon light exposure. In such systems, target cells within a heterogeneous mixture can be selectively encapsulated by photo-initiated polymerization in the presence of monomers. Enzymatic initiation systems for cellular encapsulation have also been demonstrated, for example peroxidase-mediated dityrosine cross-coupling for forming synthetic hydrogels capsules around cells. The general limitation of the previously reported methods is that the encapsulation is performed in a top-down fashion, meaning the cells are extrinsically labelled with compounds that induce the polymerization/encapsulation reactions.

Based on the above-mentioned state of the art, the objective of the present invention is to provide means and methods to enable a bottom-up approach where cellular encapsulation relies on inducible gene expression from within the cells to trigger the encapsulation. This objective is attained by the subject-matter of the independent claims of the present specification. Terms and definitions

The term GOx in the context of the present specification relates to glucose oxidase.

The term HRP in the context of the present specification relates to horse radish peroxidase.

The term fluorescent dye in the context of the present specification relates to a small molecule capable of fluorescence in the visible or near infrared spectrum. Examples for fluorescent labels or labels presenting a visible color include, without being restricted to, fluorescein isothiocyanate (FITC), rhodamine, allophycocyanine (APC), peridinin chlorophyll (PerCP), phycoerithrin (PE), alexa Fluors (Life Technologies, Carlsbad, CA, USA), dylight fluors (Thermo Fisher Scientific, Waltham, MA, USA) ATTO Dyes (ATTO-TEC GmbH,

Siegen, Germany), BODIPY Dyes (4,4-difluoro-4-bora-3a,4a-diaza-s-indacene based dyes) and the like.

A polymer of a given group of monomers is a homopolymer (made up of a multiple of the same monomer); a copolymer of a given selection of monomers is a heteropolymer constituted by monomers of at least two of the group.

The term polypeptide in the context of the present specification relates to a molecule consisting of a linear chain of 50 or more amino acids connected by peptide bonds. The amino acid sequence of a polypeptide may represent the amino acid sequence of a whole (as found physiologically) protein, fragments, or circular permutations thereof.

The term peptide in the context of the present specification relates to a molecule consisting of up to 50 amino acids, in particular 8 to 30 amino acids, more particularly 8 to 15 amino acids, that form a linear chain wherein the amino acids are connected by peptide bonds.

Detailed description of the invention

A first aspect of the invention relates to a method for selecting an improved variant of a polypeptide. This polypeptide in its native sequence is capable of reacting a substrate to a product [such being capable of reacting a substrate to a product defined as“activity”] under a first set of conditions, but is not capable of

retaining activity after having been exposed to a second set of conditions, or showing activity when exposed to said second set of conditions;

and wherein the variant [which differs from the native sequence by at least one deletion, insertion or substitution of an amino acid in the polypeptide sequence] is an improved variant if it is capable of retaining activity

after having been exposed to said second set of conditions, and/or

when being exposed to said second set of conditions. The method comprises the steps of:

a. generating a plurality of cells, wherein each cell of said plurality expresses [i.e. produces on protein level] one variant of said polypeptide and particularly presents said variant on the cell surface;

b. exposing said plurality of cells to said second condition;

c. [optionally, exposing said plurality of cells to said to the first condition]; d. contacting said plurality of cells with said substrate [thereby producing the

reaction product if the variant is capable of retaining activity under the second set of conditions] and a hydrogel precursor mix capable of forming a hydrogel in the presence of the product,

e. separating the cells that are encapsulated in hydrogel.

The“first condition” can be seen as the set of parameters under which the native, non-variant polypeptide is active, and the“second condition” is a set of parameters, different from the fist condition in at least one parameter, in which the native, non-variant polypeptide is not active, or is significantly less active, and for which an active variant is to be selected.

In certain embodiments, the precursor mix comprises soluble polymers capable of forming a hydrogel via crosslinking. In certain embodiments, the concentration of said soluble polymers is adjusted so that cells that produce an improved variant of the polypeptide, are

encapsulated in the hydrogel.

The set of conditions that can be varied to render a first and a second condition includes, without being limited to, pH, temperature, solvent conditions (particularly concentration of ions, nature of ions, concentration of non-ionic solvates of the aqueous reaction medium).

The polypeptide is either presented on the cell’s surface or said product of the polypeptide is membrane permeable and diffuses out of the cell.

In certain embodiments, the polypeptide variant is expressed in each of the plurality of cells as a single copy transgene. In the experiments reported in the example section of the present specification, there is only one copy of a plasmid encoding a variant of the polypeptide per cell. This is based on the origin of replication of the plasmid used in the examples, pYD1 , which is a‘single-copy’ plasmid.

In certain embodiments, the second condition is reverted or changed to the first condition by changing the temperature or washing the cells in order to change the pH or the solute concentration. In certain embodiments, a detectable label is incorporated into the hydrogel during or after crosslinking. Incorporating a detectable label facilitates automatic selection of cells that express an active variant under the second set of conditions.

In certain embodiments, said detectable label is covalently bound to said soluble polymers.

In certain embodiments, the product of the polypeptide, in other words the product of the active variant, is a crosslinking starter itself. In certain other embodiments, the product is a compound that is converted into a crosslinking starter by one or more enzymes that are present in the hydrogel precursor mix, and which are contacted with the plurality of cells concomitantly in order to facilitate encapsulation.

In certain embodiments, said polypeptide is a laccase, a cellulase, an oxidoreductase, a peroxidase, a peptide ligase, a protease or a FAD-dependent oxidoreductase.

In certain embodiments, said polypeptide is selected from glucose oxidase, D-amino acid oxidase, peroxidase, urate oxidase, lipase, laccase, L-amino acid oxidases, asparaginase, arginase, ornithine decarboxylase, superoxide dismutase, xanthine oxidase, uric acid oxidase, tyrosinase (another name: polyphenol oxidase), hemocyanin, haemoglobin, myoglobin, catechol oxidase, and prophenoloxidase. In certain embodiments, said polypeptide is selected from glucose oxidase, D-amino acid oxidase, peroxidase, and urate oxidase.

In certain embodiments, said enzyme is a peptide ligase, a protease or a peroxidase, particularly horse-radish peroxidase.

In certain embodiments, said polypeptide is a FhC^-producing polypeptide.

In certain embodiments, said product is H ₂ O ₂ .

In certain embodiments, said substrate is a monosaccharide, particularly glucose, or an amino acid, particularly tyrosine or phenylalanine, or a phenol.

In certain embodiments, said polypeptide in its native sequence is encoded by a native nucleic acid, wherein the method comprises the steps:

i. mutating said native nucleic acid, e.g. by deletion, insertion or substitution of at least one nucleic acid base pair in the nucleic acid sequence, yielding a plurality of different mutated nucleic acids;

ii. introducing each nucleic acid of said plurality of different mutated nucleic acids into a cell and expressing the mutated nucleic acid, thereby yielding said plurality of cells, each producing one variant of the polypeptide. In certain embodiments, said plurality of different mutated nucleic acids is generated via gene shuffling, error prone PCR or degenerate primers.

There are several methods to obtain mutated nucleic acids. Any method of nucleic acid diversification known to the skilled molecular biologist can be applied to obtain mutated nucleic acids.

In certain embodiments, the first set of conditions differs from the second set of conditions in a. temperature (particularly said second set of conditions is characterized by higher temperature than the first set), and/or

b. pH and/or

c. molality (ion concentration) and/or presence of certain polymers such as PEG, other polymeric molecules used to increase viscosity, as well as other anions or cations.

In certain embodiments, said second set of conditions is characterized by 55°C to 65°C. In certain embodiments, said second set of conditions is characterized by a pH of 3 to 5 or a pH of 8 to 10. In certain embodiments, said second set of conditions is characterized by an ion concentration of 0.001 to 7.0 mol/L. In certain embodiments, said second set of conditions is characterized by an ion concentration of 0.01 to 5.0 mol/L. In certain embodiments, said second set of conditions is characterized by an ion concentration of 0.05 to 2 mol/L. In certain embodiments, said second set of conditions is characterized by an ion concentration of 0.1 to 1 mol/L.

In certain embodiments, exposure to said second set of conditions is 1 to 60 minutes, particularly 10 minutes.

In certain embodiments, the detectable label is a fluorescent dye, or biotin.

In certain embodiments, said soluble polymers comprise a polypeptide or modified alginate or modified chitosan, particularly wherein the modification comprises

alginate or chitosan covalently modified by a hydroxyaryl moiety, particularly a phenol moiety, or

a polypeptide comprising tyrosine residues,

and optionally, a detectable label selected from a fluorescent dye and biotin.

In certain embodiments, said hydrogel precursor mix is modified chitosan and the hydrogel is pH-sensitive. pH-sensitive hydrogels shrink as a function of changing pH.

In certain embodiments, step e is performed using FACS, size filtration, enzymatic cell lysis, osmotic cell lysis. In certain embodiments, the plurality of cells are yeast cells.

Wherever alternatives for single separable features such as, for example, an isotype protein or detectable label are laid out herein as“embodiments”, it is to be understood that such alternatives may be combined freely to form discrete embodiments of the invention disclosed herein. Thus, any of the alternative embodiments for a detectable label may be combined with any of the alternative embodiments of condition and these combinations may be combined with any polypeptide mentioned herein.

The invention is further illustrated by the following examples and figures, from which further embodiments and advantages can be drawn. These examples are meant to illustrate the invention but not to limit its scope.

Brief description of the figures

Fig. 1 shows modifications on alginate and chitosan.

Fig. 2 Overview of genetically encoded hydrogel encapsulation system for single cells. Scheme depicting gene induction and yeast display of GOx homodimer, followed by enzyme-mediated cross-linking of phenolated fluorescent monomers into a conformal hydrogel surrounding the cell. Modified monomers carry fluorescein and phenol moieties depicted as a green circle and yellow star, respectively.

Fig. 3 Flow cytometry analysis of cells displaying GOx and enzymatically

encapsulated in fluorescent hydrogel. (A) Flow cytometry analysis of (i) a yeast population displaying GOx and stained with anti-Xpress antibody; (ii) yeast enzymatically encapsulated in fluorescent alginate hydrogel; and (iii) doubly stained yeast population confirming surface display of GOx (Alexa fluor 555-labelled anti-Xpress antibody) and encapsulation in fluorescent hydrogel (fluorescein-modified alginate). (B) Cell reference mixtures used to test the fidelity of the encapsulation process. Decreasing amounts of GOx positive cells in the starting mixtures were detected through analytical flow cytometry following the encapsulation reaction. An example of a flow cytometry plot with the gate used for the analysis is shown on the right.

Fig. 4 Morphology of alginate capsules. (A) Yeast cells individually encapsulated in a conformal alginate shell following a 10 minute reaction. The shell was not observable in bright field (A, top). Fluorescein incorporated in the alginate macromonomers allowed direct visualization of the hydrogel shells in X, Y, and Z projections (A, bottom) using confocal fluorescence microscopy. (B) Hydrogel particle diameters (including cell diameter) were measured at different reaction times (n=10 cells for each time point).

Fig. 5 Overview of GOx directed evolution using enzymatic hydrogel

polymerization. (A) Structural model of GOx homodimer from A. niger (PDB code 1 cf3). The N-terminal residues subjected to genetic diversification are highlighted in orange. The constant region residues are shown in blue. The FAD cofactor is shown as a surface plot in green. (B) Chemical structure of modified alginate used for cell encapsulation. (C) Scheme depicting gene induction and yeast display of GOx homodimer, followed by enzyme-mediated cross-linking of phenolated fluorescent alginate into a conformal hydrogel surrounding the cell. (D) Scanning confocal fluorescence microscopy images of single yeast cells encapsulated in fluorescent alginate shells.

Fig. 6 Kinetic parameters and thermal stability of WT and evolved mutant GOx.

(A) Kinetic plots showing initial rates of product formation vs. substrate concentration for WT and GOx mutants. (B) Mutation sites and kinetic parameters for WT and mutant GOx. (C) Thermostability of WT and mutant GOx was compared using normalized initial activity following heat shock at a given temperature (n=3 for all samples).

Fig. 7 Thermal denaturation curves of GOx wild type (WT) and mutant enzymes (M1 ,

M2, M3) analyzed detecting changes in the fluorescence emission ratio using the Prometheus NT.48. Each experiment was run in triplicate and at a protein concentration of 0.1 mg/ml.

Fig. 8 SDS-PAGE of purified Aga2-GOx wt, M1 , M2, and M3 fusion proteins. Each sample (1 pg) was loaded respectively in non denaturing form, after denaturation with 2-Mercaptoethanol and incubation at 95°C, or after denaturation and incubation for 1 hour with 1000 units of EndoH enzyme. The glycosylated samples appear as smears due to variations in glycosylation levels. Treatment with EndoH caused cleavage of chitobiose from high mannose and some hybrid N-linked oligosaccharides to the revealing a more defined pattern as compared with the wild type.

Fig. 9 Strategies to isolate encapsulated cells. (A) Immunity to enzymatic lysis.

Genetically heterogeneous yeast cell populations containing 10% pGAL-GOx positive cells were encapsulated in fluorescent alginate and treated with lytic enzymes. The fraction of genotype-positive pGAL-GOx colonies was quantified before and after enzymatic lysis. The recultivated population contained only pGAL-GOx positive cells which were protected from enzymatic lysis by the hydrogel layer. (B) Size exclusion by filtration. Encapsulated yeast cells in fluorescent alginate (green) were retained by the polycarbonate filter with 10 pm pores (dark spots). Scale bar indicates 100 pm. (C) A yeast cell population containing 10% pGAL-GOx-positive cells was encapsulated in alginate and processed by filtration through a 10 pm pore filter membrane. A majority of non-encapsulated cells were removed, resulting in 95% pGAL-GOx positive cells retained on the filter.

Fig. 10 pH-responsive chitosan capsules. (A, top) Single budding yeast cell

encapsulated in a fluorescent chitosan capsule at pH 6. The chitosan hydrogel is swollen and invisible in bright field. (A, bottom) Single budding yeast cell encapsulated in fluorescent chitosan at pH 8. The polymer shell is collapsed around the cell. At pH 8, the chitosan layer around the cells is also visible in bright field mode as indicated by the red arrows. Scale bars indicate 10 pm.

(B) Size distribution of chitosan-encapsulated single cells following resuspension in aqueous buffers with increasing pH from 6 to 8. Average capsules diameters and standard deviations: 67.6 ± 20.5 pm at pH 6, 27.8 ± 12.8 pm at pH 7, and 13.2 ± 4 pm at pH 8 (n>25 cells for each pH).

Fig. 11 Comparison of wild type and mutant GOx chemical and thermal stability.

(A) Residual activity of mutant enzymes and wild type GOx after incubation with increasing concentrations of GdnHCI. GOx residual activity was tested through cell encapsulation. The ratio of cell encapsulated has been plotted. The values are normalized to the number of cell encapsulated at 0 M GdnHCI.

(B) Thermal stability of the mutant enzymes and wild type GOx assayed through cell encapsulation after incubation of the yeast cells at increasing temperatures. The ratio of encapsulated cells are normalized to the number of cells encapsulated after incubation of the same population at 30°C.

Fig. 12 Expression and display of DAOx enzyme. EBY100 yeast cells carrying the plasmid pYD1_Aga2_RgDAOx were grown for 24h at 30°C in -Trp medium supplemented with 2% glucose. In order to induce protein expression and surface display, the culture was transferred to -Trp medium with 0.2% Glucose and 1.8% galactose (OD ₆ oo=0.4). Protein expression was induced for 48 h at 20°C. The display of the fusion protein Aga2_DAOx on the yeast cell was verified through antibody staining of the C-terminal HisTag as follows. Cells were incubated with the primary mouse Anti-HisTag antibody. Afterwards the samples were washed and then incubated with the secondary anti-mouse antibody conjugated with Alexa Fluor 594 fluorophore. The samples were then analysed through flow cytometry. As negative control, the same cell population was treated with only the secondary antibody.

Fig. 13 Amplex Red enzymatic assay to detect DAOx activity. 1 million yeast cells from the Rg-DAOx positive population (Fig. DAOx_1 ) were mixed with the reactants which included 35 mM D-Alanine, 4 mM HRP and 100 mM Amplex Red. The activity of the enzyme was detected by reading the fluorescence produced by Amplex Red when converted to Resorufin by HRP in the presence of H2O2. As negative control, the same reaction reaction mixture was added to 1 million yeast cells displaying Aga2 but lacking the target DAOx enzyme.

Fig. 14 Time course cell encapsulation experiment. One half million cells from the population staining positive for the expression and display of RgDAOx (Fig. DAOx_1 ) were mixed with the cell encapsulation reaction mixture (0.125% modified alginate, 4.5 uM HRP, 35 mM D-ALA, 20%PEG) and incubated at RT for increasing amount of time. The reaction was stopped at each time point and the samples were analyzed through flow cytometry. Cells positive for the display of the enzyme acquired red fluorescence (y-axis), while encapsulated in fluorescent hydrogel acquired green fluorescence (x-axis). By t=10 min., most of the cells that tested positive for enzyme expression also shifted towards the right, indicating successful encapsulation. The gate R1 in the image indicates all the events positive for RgDAOx display and for

encapsulation in fluorescent alginate.

Fig. 15 Drop in experiment. Increasing concentrations of RgDAOx positive cells stained for the enzyme display (Fig. DAOx_1 ) were mixed with yeast cells displaying just the anchor protein Aga2 at final Oϋboo = 0.5, and then used for the cell encapsulation procedure. Cell encapsulation was achieved with 0.125% modified alginate, 4.5 uM HRP, 35 mM D-Alanine, 20% PEG incubating the reaction for 10 min at RT. The gate R2 includes all the events positive for both RgDAOx display and encapsulation in fluorescent alginate.

Fig. 16 Effect of PEG on cell encapsulation. 500Ό00 cells from RgDAOx population were mixed with varying amount of polyethylene glycol as well as the cell encapsulation mixture (0.125% modified alginate, 4.5 uM HRP, 35 mM D- ALA). PEG was added in order to increase the viscosity and does not participate in the polymerizaiton reaction. The samples were incubated for 10 min. Afterwards the samples were analysed through flow cytometry. Fig. 17 Expression and display of HRP enzyme. EBY100 yeast cells carrying the plasmid pCT2_Aga2_HRP were grown for 24h at 30°C in -Trp medium supplemented with 2% glucose. In order to induce protein expression and display, the culture was transferred to a -Trp medium with 0.2% Glucose and 1.8% galactose, 3.6 mM Aminolevulinic acid and 0.2 mM Ferric citrate

(OD6OO=0.4). The protein expression was induced for 24h at 20°C. The display of the fusion protein Aga2_HRP on the yeast cell was verified through antibody staining of the C-terminal protein tag cMyc. The cells were incubated with the primary mouse Anti-cMyc antibody. Afterwards the samples were incubated with the secondary anti-mouse antibody conjugated with fluorescein. The samples were then analysed through flow cytometry. As negative control, the same cell population was treated just with the secondary antibody.

Fig. 18 Amplex Red enzymatic assay to detect HRP activity. 1 million yeast cell from the HRP positive population (Fig. HRP_1 ) were mixed with the reaction mix containing 2 mM H2O2 and 100 uM Amplex Red. The activity of the enzyme was detected by reading the fluorescence produced by Amplex Red substrate when processed by HRP in presence of H2O2. As negative control the same reaction mix was added to 1 million yeast cells displaying Aga2 but not the target HRP enzyme.

Fig. 19 HRP driven cell encapsulation experiment. 400Ό00 yeast displaying HRP were mixed with the encapsulation reaction mixture (0.125% modified alginate, 100 mM Glucose, 4.5 uM GOx). Differing concentrations of soluble HRP were added to the bulk reaction mixture. The reactions were incubated at RT for 10 and the samples then analyzed by flow cytometry. Cell positive for encapsulation shifted towards the right (green fluorescence) as indicated by the gate R1.

Fig. 20 Expression and display of UOx enzyme. EBY100 yeast cells carrying the plasmid pYD1_Aga2_UOx_HisTag were grown for 24h at 30°C in -Trp medium supplemented with 2% glucose. In order to induce protein expression and display the culture was transferred to -Trp medium with 0.2% Glucose and 1.8% galactose (OD ₆ oo=0.4). The protein expression was induced for 24h at 20°C. The display of the fusion protein Aga2_UOx on the yeast cell was verified through antibody staining of the C-terminal His Tag. Cells were incubated with the primary mouse Anti-His antibody. Afterwards the samples were incubated with the secondary anti-mouse antibody conjugated with fluorescenin. The samples were then analysed through flow cytometry. As negative control, the same cell population was treated just with the secondary antibody.

Fig. 21 Amplex Red enzymatic assay to detect UOx activity. 4 millions yeast cell from the UOx positive population were mixed with the reaction mix containing 4 mM Uric acid, 4.5 uM HRP, 100 uM Amplex Red. The activity of the enzyme was detected by reading the fluorescence produced by Amplex Red substrate when processed by HRP in presence of H2O2. As negative control the same reaction mix was added to 4 million yeast cells displaying Aga2 but not the target UOx enzyme.

Materials and methods

Materials

All the chemicals used in this work were purchased from Sigma Aldrich if not otherwise specified. The primary and secondary antibody were purchased from Thermo Fisher Scientific. Restriction enzymes were bought from New England Biolabs. Sodium Alginate (Viscosity 1 %: 100 - 200 mPa.s) was purchased from Duchefa Biochemie. Zymolyase 100T was purchased from Roth.

Preparation of fluorescent alginate and chitosan with phenol groups

Alginate and chitosan with phenol moieties and fluorescein were prepared through carbodiimide activation chemistry (Sakai, S.; Kawakami, K. Acta Biomater. 2007, 3 (4), 495- 501 .). Sodium alginate (avg. Mw 70,000 Da) was dissolved in 50 mM MES buffer pH 6 at a final concentration of 10 mg/ml_. After the alginate was completely dissolved, tyramine hydrochloride, NHS and EDC were added to the sodium alginate solution at 7, 1 .2, 3.9 mg/ml_ respectively. Finally, 6-aminofluorescein was added to a final concentration of 0.25 mg/ml_. Chitosan (avg. Mw 50,000-190,000 Da) was dissolved in a pH 2 solution of HCI at a final concentration of 10 mg/ml_. After the chitosan was completely dissolved, the pH was raised to 4.5 by adding NaOH. Then phloretic acid, NHS and EDC were added to the chitosan solution at 7, 1.2,3.9 mg/ml_ respectively. Finally, fluorescein was added to a final concentration of 0.25 mg/ml_. Both the reactions were incubated at room temperature for 16 to 20 hours with vigorous stirring and then precipitated dropwise into 80% ethanol aqueous solutions. The alginate was then washed with 80% ethanol and the final product dissolved in water before being lyophilized. Chitosan was then washed with 80% ethanol and then acetone. The final product was dissolved in an acidic solution at pH 2. After dissolution of the modified chitosan, the pH was raised to 5 by adding NaOH before being lyophilized. The success of the conjugations was confirmed through ¹ H-NMR.

Amplification and cloning of the Wild-type GOx gene

The GOx wild type gene was amplified from the genome of Aspergillus niger strain 4247 (LGC Standards) using the primers F1 (SEQ ID NO 1 : 5’-

G CAT ACG G AT C CAT G C AG ACT CTC CTTGT GAG CTC G C-3’ ) and R1 (SEQ ID NO 2: 5’- G CAT AC CT C G AGT C ACT G CAT G G AAG CAT AAT CTTC C-3’ ) , and cloned using BamHI and Xho\ restriction sites into the yeast plasmid pYD1 for protein display (gift from Dane Wittrup, Addgene plasmid #73447, Kieke, M. C.; Cho, B. K.; Boder, E. T.; Kranz, D. M.; Wittrup, K. D. Protein Eng. 1997, 10 (1 1 ), 1303-1310.)· After sequence confirmation, the plasmid pYD1 -GOx was transformed into Saccharomyces cerevisiae EBY100 following a typical lithium acetate transformation (Gietz, R. D.; Woods, R. A. Methods Enzymol. 2002, 350, 87-96.) procedure and selecting the positive colonies on SD agar 2% (w/v) glucose plates lacking tryptophan (- Trp). Resulting colonies were cultivated in liquid SD -TRP liquid medium with 2% glucose for 24 hours at 30°C with continuous shaking at 200 rpm. Protein expression and display was then induced by transferring the culture to a fresh liquid medium lacking tryptophan containing 0.2% (w/v) glucose and 1.8% (w/v) galactose, and shaking for 24 hours at 30°C.

Amplex Red enzymatic assay

Yeast cells displaying Glucose Oxidase (GOx) were used to perform an Amplex Red enzymatic assay in order to test for the functionality of the expressed enzyme. The reaction mixture consisted of ~10 ⁶ yeast cells, 100 mM Glucose, 4.5 mM HRP and 10 pM Amplex Red in 50mM of Phosphate Buffer (pH 7.4). The fluorescence was read at 590 nm every 1 min for 10 min.

GOx antibody labelling

The expression and display of GOx was confirmed by antibody labelling. -2 x 10 ⁶ induced yeast cells were washed with 1 ml. PBS containing 0.1 % BSA and then resuspended and incubated at room temperature for 30 min with 1/200 dilution of the primary anti-Xpress antibody (stock concentration of 1.2 mg/ml_). After incubation, the cells were washed with 1 ml. of ice-cold PBS + 0.1 % BSA and then resuspended in ice cold buffer containing 1/200 dilution of the goat anti mouse IgG secondary antibody (stock concentration 2 mg/ml_) conjugated with Alexa Fluor 555. After 20 min incubation on ice, the cells were pelleted, washed with cold buffer and resuspended just before flow cytometry. As a negative control, the same yeast cells carrying the plasmid pYD1-G0x wild type were treated only with the secondary antibody conjugated with Alexa Fluor 555.

Cell encapsulation in fluorescent alginate and chitosan

After induction of protein expression, yeast cells displaying GOx were washed with 50 mM sodium phosphate buffer pH 7.4 and resuspended in the same buffer at a final Oϋboo of 0.2. Glucose, HRP and the modified alginate were added to the cell sample at final concentrations of 100 mM, 4.5 mM, and 0.125% (w/v), respectively. The reaction was gently mixed and incubated at room temperature. After 10 min, four volumes of 50 mM sodium phosphate buffer pH 7.4 were added and the samples were analyzed through flow cytometry or used for single cell sorting. Cell encapsulation in chitosan was performed at the exactly same conditions as for the alginate but replacing the sodium phosphate buffer (pH 7.4) with 50 mM MES buffer at pH 6. When used, mCherry was added to the encapsulation reaction mixture at a final concentration of 4.5 mM.

Flow cytometry, FACS and downstream analysis

All the single and double stained yeast cells were analyzed with the Attune NxT (Thermo Fisher Scientific) flow cytometer equipped with a 488 nm and a 561 nm laser. Yeast cells were sorted using a MoFLo XDP cell sorter from Beckman Coulter equipped with 488 nm and 561 nm lasers, and with a 100 pm nozzle. Prior to encapsulation and sorting, the yeast cells were washed and stained with Propidium Iodide (final concentration of 4 pg/mL). The cells positive for the encapsulation reaction and negative for the staining with Propidium Iodide were sorted in single cell mode.

Cytoprotection against enzymatic lysis

A mixture of yeast cells containing less than 10% pGAL-GOx genotype positive cells mixed together with 90% cells carrying an empty pYD1 plasmid was used to perform a standard cell encapsulation reaction in fluorescent alginate. After cell encapsulation, Zymolyase was added to the reaction mix to a final concentration of 300 U/ml together with 1 mM Dithiothreitol (DTT). The reaction was incubated at 35°C with shaking at 1000 rpm for 60 minutes. The cells were then transferred to a -Trp glucose liquid medium and grown for 40 hours before being used for single cell sorting onto -Trp agar plates containing 2% Galactose. After 48 hours, colonies derived from single cell sorted events were assayed for the expression of GOx through an ABTS top agar assay. Glucose (333 mM), ABTS (7 mM) and HRP (2U/ml) were mixed with 2% agar solution and poured on the plate containing the sorted colonies. After 2 to 5 minutes, green halos started to appear above the colonies expressing and displaying the GOx on the cell wall. As a negative control the same yeast population was processed in parallel but without treatment with the lytic enzymes mixture. Yeast cell enrichment by filtration

A yeast population containing less than 10% pGAL-Gox genotypic positive cells was used to perform a standard cell encapsulation reaction in fluorescent alginate. After encapsulation the sample was filtered through polycarbonate membranes with 10 pm pores sizes (Whatman) by applying vacuum. Afterwards the membrane was washed three time with 50 mM sodium phosphate buffer at pH 7.4. Finally, the encapsulated cells retained by the filter were recovered by rinsing the the membrane with sodium phosphate buffer (pH 7.4) and analyzed with flow cytometry. pH-responsive chitosan capsules

Multiple samples of yeast cells displaying GOx were washed with 50 mM MES buffer pH 6 and resuspended in the same buffer at a final Oϋboo of 0.2. Glucose, HRP and the

fluorescent modified chitosan were added to the cells at final concentrations of 100 mM, 4.5 pM, and 0.125% (w/v), respectively. The reactions were gently mixed and incubated at room temperature. After 20 min, different samples were diluted respectively with four volumes of 50 mM MES buffer pH 6, 50 mM phosphate buffer pH 7.4 or 50 mM phosphate buffer pH 8.2 to stop the encapsulation reaction. The pH of each sample was measured and if needed corrected to final pH values of 6, 7 or 8. The size of the capsules at each pH was analysed through microscopy by measuring the diameter of multiple randomly picked cell-hydrogel units (n>25 for each pH value).

GOx mutant library construction by error prone PCR

A GOx mutant library was prepared by amplifying the region between nucleotides 4 and 513 of the GOx wild type gene in the pYD1 through error prone PCR (GeneMorph II, Agilent) using primers F2 and R2 (Table 1 ). The backbone of the recipient vector was amplified through PCR using primers F3 and R3 (Table 1 ). The error prone PCR product and the pYD1 linearized backbone were then co-transformed in EBY100 following the gap repair method -, leading to a library of -4 · 10 ⁵ clones. The average frequency of 8 mutations per kb was determined by sequencing the plasmids of five randomly selected clones. After selection of the positive clones on selective medium agar plates, the colonies were collected and grown for 24 hours before being induced for protein expression and display.

Table 1. List of DNA primers used in this work

FACS-based screening

Yeast cells were sorted using a MoFLo XDP cell sorter from Beckman Coulter equipped with 488 nm and 561 nm lasers, and with a 100 pm nozzle. Prior to sorting, the yeast mutant libraries were treated with 5M guanidinium chloride for 10 minutes, washed 3 times, and stained with Propidium Iodide (final concentration of 4 pg/mL) before being used for the cell encapsulation assay. All the cells positive for the encapsulation reaction and negative for the staining with Propidium Iodide were sorted in single cell mode.

After sorting, single cells were cultured in liquid glucose medium lacking tryptophan for two days at 30°C and then spotted on an SD agar 2% (w/v) galactose plate lacking tryptophan in order to induce the GOx protein expression and display. After 2 days incubation at 30°C the colonies were used for an ABTS topagar assay in order to test for the expression of GOx. The assay was performed by mixing 2% (w/v) agar with an equal volume of ABTS reaction solution containing 666 mM Glucose, 14 mM ABTS and 150 pg HRP and by pouring the mix on the colonies. After a few minutes green halos were observed around the colonies expressing GOx . Furthermore, after culturing the cells in liquid SD glucose medium lacking tryptophan the plasmids were isolated using the Zymoprep Yeast Plasmid Miniprep II kit. The sequences of the GOx mutant genes were obtained through Sanger sequencing.

Protein expression and purification Yeast colonies carrying the pYD1 vector with GOx wild type and mutant genes were grown in -TRP liquid medium with 2% (w/v) glucose for 24 hours at 30°C with continuous shaking at 200 rpm. The protein expression and display was induced by transferring the cells at initial Oϋboo qί 0.4 to a -TRP liquid medium containing the 1.8% (w/v) of galactose and 0.2 % (w/v) of glucose and by growing the cultures at 30°C and 200 rpm shaking. After 24h the cells were pelleted down and washed with 20 mM Hepes buffer pH 8 before to be resuspended at an Oϋboo equal to 50 in the same buffer containing 0.1 mM of DTT. The reactions were incubated for 2.5 hours at room temperature with gentle shaking in order to allow the reduction of the disulfide bonds between Aga2 and Aga1 proteins responsible for the anchoring of the proteins of interest on the yeast surface. Afterwards the cells were pelleted down and the supernatant was loaded on an HiTrap ion exchange column. Aga2- GOx WT and Aga2-GOx mutants (M1 , M2, and M3) were eluted by applying a NaCI gradient from 0 to 1 M. The eluted proteins were collected in 1 ml fractions and all samples were tested for GOx activity through ABTS assay. The fractions containing functional GOx were isolated and concentrated through a 100 kDa cutoff Vivaspin ultrafiltration column. The different GOx variants were then compared in there native, denatured and deglycosylated form on SDS Page electrophoresis gels. In order to cleave all the N-linked mannose structures and compare the protein in their deglycosylated form, each protein (1 pg) was denatured and then treated with 1000 units of EndoH (NEB) enzyme at 37°C for 1 h.

Enzyme kinetic studies

After protein purification the kinetic analyses of each variant were performed by using the ABTS assay with different concentrations of glucose from 0.4 to 50 mM. The other components of the reactions were 50 ng of Aga2-GOx WT or mutant enzymes, 4.5 mM HRP, 2 mM ABTS in 100 mM Sodium Phosphate Buffer at pH 7.4. The concentration of HRP used in the kinetic reactions is ~ 1000 times higher than the GOx concentration and was experimentally confirmed not to limit the rate of the reaction. All kinetic studies were performed in triplicates by following the absorbance of the ABTS substrate at 405 nm following processing by the GOx/HRP cascade. The absorbance values were converted to rate of product formation (mM s ¹ H2O2 ) by using a standard curve. The enzyme kinetic data were fitted to the Michaelis and Menten equation using the GraphPad Prism 5 software in order to determine the K _m and k _cai values for each enzyme variant. The molar concentration of Aga2-GOx WT and M1 , M2, and M3 mutants was calculated to be 4.74 nM for each reaction using the Lambert-Beer law and considering the extinction coefficient at 280 nm based on the amino acid sequence of the respective proteins.

Thermostability studies

The thermostability of the soluble GOx wt and mutant enzymes was determined by incubation 50 ng of each enzyme in 50 mM sodium phosphate buffer (pH 7.4) at increasing temperatures from 50 to 62.5 for 10 minutes. The sample was then cooled and residual activity of GOx was calculated through ABTS assay as described. For each reaction the initial rate was calculated in the linear range of the curve. For each enzyme variant, the rate of the reaction catalyzed by that sample incubated at 50°C was considered 100% of activity; in order to calculate the residual activity, the initial reaction rates at other temperature for the same protein were normalized onto this value. Temperature-dependent differential scanning fluorescence was performed using the Prometheus platform from NanoTemper (Munich).

Example 1:

The presented reaction system relies on an enzymatic cascade involving GOx and HRP to cross-link phenol groups grafted onto diverse macromonomers species that are presented in the medium to the cells. GOx is a highly glycosylated homodimeric flavoprotein, 160 kDa in size that catalyzes the oxidation of b-D-glucose to D-gluconolactone and hydrogen peroxide using molecular oxygen as an electron acceptor. Recently the inventors demonstrated the use of GOx in combination with Fenton’s reagent for quantifying enzyme activity based on fluorescent hydrogel formation. In the current system, GOx and HRP work in tandem as a bi-enzymatic initiation system for polymerization of phenolated macromonomers.

To synthesize enzyme-crosslinkable species, macromonomers were modified with phenol and fluorescein groups using carbodiimide activation chemistry. Sodium alginate was conjugated with tyramine and aminofluorescein (Fig. 1 ), and chitosan was coupled with fluorescein and phloretic acid (Fig. 1 ). Following extensive washing and lyophilization, the presence of the aromatic groups of phenol and fluorescein moieties was confirmed using proton nuclear magnetic resonance ( ¹ H-NMR) spectroscopy. The inventors tested the ability of the GOx/HRP cascade to cross-link phenolated monomers in free solution (i.e., without cells) by combining 10 mg/ml_ of modified alginate or chitosan macromonomers with 0.5 mM GOx and 4.5 mM HRP in sodium phosphate buffer (pH 7.4). Upon addition of 100 mM D- glucose, the solution rapidly gelled within a few seconds. Negative controls lacking the HRP or GOx remained liquid. These observations confirmed that the modified macromonomers could be rapidly cross-linked through the GOx/HRP cascade.

The inventors sought to adapt the enzyme-mediated polymerization system to modify, visualize, and isolate specific cell lines amidst genetically heterogeneous cell populations based purely on physical properties of the hydrogel capsules. In order to connect the hydrogel polymerization with the genotype of a specific cell line, the inventors incorporated GOx into a eukaryotic cellular display system, and tested the ability of cells displaying GOx to auto- encapsulate in hydrogel shells. HRP, phenolated fluorescent macromonomers, and glucose were then presented in the medium to GOx-displaying cells (Figure 2) to initiate the encapsulation reaction. To express GOx on the cell wall, a gene encoding Aspergillus niger GOx was cloned in frame with the yeast a-agglutinin subunit Aga2p to create an artificial protein display construct. Translocation and anchoring of Aga2p-GOx to Agal p on the cell wall were verified by immunostaining with Alexa Fluor 555-conjugated antibodies against a protein tag located between Aga2p and GOx in the fusion construct. Using analytical flow cytometry, the inventors observed an increase in fluorescence for a major fraction of the cell population (~ 60%), indicating successful display of the enzyme (Figure 3A, i). The activity of the displayed enzyme was tested using a coupled HRP/Amplex red assay, which confirmed GOx activity and therefore successful homodimerization of GOx on the cell wall.

Cells displaying functional GOx homodimers were next used to develop the cell encapsulation system. 2 x 10 ⁶ yeast cells/mL displaying GOx were suspended in a solution of 100 mM glucose, 4.5 mM HRP, and 0.125% (w/v) of phenolated alginate. After 10 min. reaction at room temperature, the samples were analyzed and a significant percentage (-60%) of cells corresponding to the same percentage successfully stained for GOx exhibited strong green fluorescence derived from fluorescein, consistent with the formation of fluorescent hydrogel shells around the cells (Figure 3A, ii). The inventors further tested the encapsulation procedure with a yeast population displaying GOx that had previously been stained with Alexa Fluor 555 for the display of GOx on the cell wall. As shown in Figure 3A, iii, after incubation in the reaction mixture a majority of the cell population prestained with red fluorescent antibodies targeting GOx also gained green fluorescence signal due to the formation of the fluorescent hydrogel. This result confirmed that cells expressing GOx were the same cells in the mixture capable of auto-encapsulating in fluorescent hydrogels.

To further prove the specificity of the cellular encapsulation system, the inventors tested the encapsulation procedure using different cell reference mixtures containing diluted amounts of GOx-positive cells at a fixed total number of cells. After encapsulation, the flow cytometry plots showed a precise correspondence between the percentage of GOx positive cells present in the cell mixtures and the number of fluorescein stained cells detected after encapsulation (Figure 4B). Moreover, 25 gated cells sorted in single cell mode from a mixture containing 10% GOx positive events were re-cultured and singularly tested for GOx activity. All of them tested positive for functional GOx, indicating the fidelity of the

encapsulation process despite the presence of a limited number of positive events in the starting sample.

Example 2

Directed evolution of accelerated or stabilized GOx variants is of considerable interest for the food and textile industries, as well as for blood glucose detection, deoxygenation of buffers for single-molecule fluorescence applications, and oxygen inhibitor removal for radical polymerization reactions. GOx is highly glycosylated and active as a homodimeric complex that binds 2 FAD cofactor molecules within a 160 kDa globular folded structure. The inventors developed the current system relying on GOx and HRP acting in tandem as a bi-enzymatic initiation system for polymerization of phenolated fluorescent macromonomers. Alginate macromonomers were synthesized from sodium alginate (Mw = 70,000 g/mol) by modification with tyramine and aminofluorescein (Figure 5B) . ¹ H-NMR confirmed the presence of aromatic phenol and fluorescein groups in the resulting polymer, as previously reported.

The approach for directed evolution of GOx is presented in Figure 5C. The inventors amplified the full length GOx gene sequence from the genome of Aspergillus niger and cloned it into the pYD1 yeast display vector in frame with the a-agglutinin protein Aga2p. The sequence was confirmed by Sanger sequencing and the plasmid was transformed into S. cerevisiae EBY100. EBY100 contains a genomic copy of Agal p under the control of a strong galactose promoter, which partners with the episomal pYD1 plasmid containing the gene of interest under control of the galactose promoter and located downstream of the Agal p domain. The inventors note that EBY100 is advantageous for GOx directed evolution because it supports glycosylation and can display large proteins. Indeed this strain was previously used for GOx directed evolution using a multi-phase emulsion system for reaction compartmentalization. Display of the enzyme was induced by galactose, inducing a chromosomal copy of Agal p and a plasmid copy of Aga2p-GOx. Immunostaining using anti- Xpress protein tag antibodies labeled with Alexa Fluor 555 was used to verify translocation and anchoring of Aga2p-GOx to Aga1 p on the outer cell wall. Analytical flow cytometry results indicated that -60% of cells in the population were fluorescently labeled, consistent with successful display of GOx. A positive result using an HRP/amplex red assay confirmed the activity of displayed GOx. Yeast cells displaying functional GOx homodimers were encapsulated in alginate hydrogels when 2 x 10 ⁶ yeast cells/mL pGAL-GOx positively induced cells were suspended in 100 mM glucose, 4.5 mM HRP, and 0.125% (w/v) of phenolated alginate. Following a 10 minute incubation, analysis with flow cytometry indicated that - 60% of cells were fluorescent in the green channel due to fluorescein. Confocal fluorescence microscopy (Figure 5D) indicated the presence of uniform and continuous fluorescent hydrogel shells encapsulating the individual cells. The inventors note that the system relies on a balance between reaction and diffusion of glucose, H2O2, HRP, and macromonomers, therefore, the cell concentration in the reaction medium is a crucial parameter. At high GOx positive cell concentrations of > 1 x 10 ⁷ cells/mL, the entire solution was found to cross-link and form a gel, while at an intermediate concentration range (2.5 x 10 ⁶ - 1 x 10 ⁷ cells/mL), small hydrogel aggregates encapsulating multiple cells were observed. Maintaining cell concentrations at or below 2 x 10 ⁶ cells/mL resulted in singly encapsulated cells with no cross reactivity even after extended incubation. The inventors validated the hydrogel encapsulation reaction as a cell screening procedure by screening cell mixtures of diluted pGAL-GOx positive cells in a background of pYD1 empty plasmid-containing cells, and observing a direct correspondence of selected positive cell concentrations in FACS plots. Green fluorescent cells sorted in single-cell mode from a reference mixture containing 10% pGAL-GOx positive cells were re-cultured without any noticeable effect of the gel capsules on cell growth rate, and all were found to be positive for GOx when tested using ABTS top agar assay. This resulted indicated there was a very low to zero background labeling of cells during the screening procedure, eliminating false positive events from the sorted cell mixtures.

The inventors next implemented the hydrogel encapsulation system for the isolation of GOx variants with enhanced properties. The inventors targeted mutagenesis to the N-terminus of the protein, which accounts for the dimer interface. The inventors applied a random mutagenesis approach, generating a pool of GOx mutants through error prone PCR targeting the first 500 nucleotides of the gene. This mutagenized PCR product was then transformed directly into yeast together with the linearized pYD1 plasmid and assembled in vivo via gap repair, resulting in a library of ~4 x 10 ⁵ GOx variants. By sequencing the plasmids of 5 randomly selected clones from the yeast library, we calculated an average mutation frequency of 8 mutations per kb.

For library screening, the inventors used high concentrations of guanidinium hydrochloride (GdnHCI) to disrupt the structure of GOx homodimers. The inventors found that short term exposure to 5M GdnHCI followed by washing in PBS buffer completely inhibited the cellular encapsulation reaction triggered by the parent enzyme. In addition to irreversibly denaturing the parent enzyme, incubation of the yeast cells with 5M GdnHCI also resulted in partial cell death, however -25% of cells survived the treatment and could be sorted and propagated without requiring plasmid extraction or re-transformation. This screening strategy allowed us to exclude cells that performed as well as or worse than wild type GOx, while maintaining a significant number of living cells capable of being isolated and expanded in culture. The inventors observed no adverse effects of the gel capsules on cell growth following

encapsulation and sorting. Encapsulated cells were able to break out of the gel capsules and propagate uninhibited in growth media.

The selection protocol involved incubation of the mutant library with 5M GdnHCI for 10 minutes, followed by washing and addition of hydrogel encapsulation reagents to the cell mixture in a one-pot reaction. This approach assayed for residual GOx activity following GdnHCI treatment, whereby the presence of the hydrogel shell was correlated with more active or more stable GOx variants displayed on the cell wall. All cells trapped in fluorescent alginate and not stained by propidium iodide (PI), a fluorescent dye used to label the dead cells, were isolated through single cell sorting and recultured for subsequent protein expression and characterization.

The ultrahigh throughput nature of the 10-min one-pot encapsulation method allowed the analysis of more than 5x10 ⁶ library members in a single sorting session of less than 1 hour, providing -10-fold coverage of the mutant library. Selected cells were recultured and tested individually to compare the stability and activity of the displayed GOx variants to the parent enzyme. Following selection, the residual activity of 27 mutants along with wild type GOx was tested in the cell display format. Each isolated clone was induced for GOx expression and incubated at elevated temperature (from 50-70 °C) for varying times, followed by exposure to the hydrogel encapsulation reagents. The fraction of encapsulated cells was then analyzed using flow cytometry (data not shown). The three GOx mutants that performed best under thermal stress in the cell surface-based assay were then purified as soluble Aga2-GOx fusion proteins and further analyzed. The obtained wild type parental Aga2-GOx and the three mutant Aga2-GOx enzymes are denoted WT, and M1 , M2, and M3, respectively.

The three evolved mutants (M1 , M2, M3) and WT enzyme were cleaved from the cell surface using dithiothreitol (DTT) to reduce the disulfide bonds between Agal p and Aga2p- GOx. Soluble Aga2-GOx was then purified from the reduced cell supernatants using ion exchange chromatography. The resulting proteins were tested for kinetic parameters using a cascade HRP/ABTS assay (Figure 6A). All three mutants exhibited improved activity (Figure 6B), with M1 having 4.9-fold greater catalytic efficiency. M3 exhibited the highest k _cai value, which was 3.96-fold higher than WT. The K _m values for M1 , M2, and M3 were furthermore slightly lower than WT but not hugely improved. This was consistent with the selection step which was performed at a saturating substrate concentration of 100 mM glucose. The observation that the mutants exhibit improvement in /c _cat and catalytic efficiency, but not significantly improved K _m values was consistent with a genetic library specifically responding to the applied selection pressure.

Significantly, sequence analysis (Figure 6B) showed that the improvements in enzyme activity were achieved as a result of only a single point mutation in the case of M1 and M3, or as the result of 2 mutations close together in primary sequence space in the case of M2.

Each mutant presented a mutation which replaced a Leucine with Proline in position 9 or 13. In addition, M2 contained an additional mutation that replaced the Alanine at position 16 with Threonine. The identified mutations were all located within the leader peptide (first 22 amino acids) of the GOx protein sequence linking Aga2 and GOx. Since prior work on directed evolution of GOx in heterologous systems only considered the protein structural core and excluded the leader sequence, the mutations identified here are newly reported mutants. The ultrahigh throughput nature of the encapsulation method allowed the analysis of more than 5 million GOx mutants in a single sorting session of less than 1 hour, covering the size of the mutant library about 10 times. The cells selected and isolated were recultured and tested individually to compare the stability of the displayed GOx variants to the wild type enzyme. The residual activity of three mutants (Mut1 , Mut2, and Mut9) along with wild type GOx was tested after incubation with increasing concentrations of the GdnHCI. Following incubation at a given concentration of GdnHCI, cells were washed and exposed to hydrogel reaction components. The fraction of encapsulated cells was then analyzed using flow cytometry (Figure 11 A).

The three GOx mutants exhibited improved stability to GdnHCI denaturing conditions, with a significant fraction of cells retaining residual activity after treatment with 5 M GndCI. Under this condition, cells displaying wild type GOx meanwhile were completely non-encapsulated due to the denatured/inactive enzyme.

Next the inventors investigated the thermal stability of M1 , M2, and M3. The thermal stability of the displayed mutants and wild type GOx was tested by incubating the cells at increasing temperatures for 10 minutes before assaying for residual enzymatic activity through cell encapsulation and analytical flow cytometry. As shown in Figure 11B, all three mutants remained active after incubation at high temperatures, showing some residual activity even after incubation at 70°C, ~10 degrees higher than the highest temperature sustained by the wild type enzyme. Based on these results, it is clear that the cell encapsulation system is valuable for screening large libraries using a one-pot method, and for correctly identifying enhanced mutant enzymes. The potential throughput for screening enzymes in this manner is limited only by the speed of flow cytometry (-10 ⁷ cells / hour).

Thermostability of the evolved mutants was evaluated by measuring normalized initial substrate turnover rate following 10 minute heat shock (Figure 6C). All three enzymes showed improvement in thermostability, with the largest improvement occurring at 55 °C, where the WT enzyme exhibited 59% residual activity while M1 , M2, and M3 remained 74, 73, and 72% active, respectively. Thermostability of WT, M1 , M2 and M3 was further analyzed using nano differential scanning fluorescence (Figure 7), which confirmed two primary denaturation peaks occurring at 53°C and 61 °C. For all three mutants, the peak at 61 °C was the more prominent of the two peaks, while for wild type the peak at 53°C was dominant, consistent with the mutants exhibiting a thermally stabilized dimer conformation that significantly outperformed WT. The lower temperature denaturation peak was attributed to Aga2-GOx monomers.

Since GOx is known to be highly glycosylated when expressed in yeast, the inventors hypothesized that changes in GOx glycosylation resulting from the selected mutations could be responsible for the observed activity enhancement. With the introduction of a threonine residue, one of the isolated mutations, A16T in M2 is a candidate residue for O-linked glycosylation. The other three isolated mutations introduced a proline in place of leucine at position 9 or 13 (Figure 6B) directly adjacent to a serine residue, which is also a potential O- linked glycosylation site. The inventors found that digestion of WT, M1 , M2, and M3 glycoproteins using endoglycosidase-H to cleave N-linked high mannose chains resulted in differently sized bands for M1-M3 as compared with WT when analyzed using SDS-PAGE (Figure 8), suggesting that O-linked glycosylation of the mutants conferred activity and stability enhancement. Since availability for glycosylation depends on structure, the inventors can speculate that the L9P and L13P mutations resulted in disordering of the alpha helix linker region and improved accessibility of the adjacent serine residues for O- mannosyltransferases. This additional O-linked glycosylation site apparently resulted in improved thermodynamic stability of dimerized Aga2-GOx not only when displayed on the cell wall, but also as a soluble fusion protein.

All of the mutations identified in our screen are located within the first 22 amino acids of the GOx protein sequence. This region at the N-terminus of the enzyme is not part of the dimer interface, but in fact is part of a signal peptide that is typically cleaved from the mature protein when expressed in its soluble form in A. niger. Nonetheless, the inventors found this region to be of critical importance for the stability of GOx in their display fusion construct, having probably a relevant role in the spatial orientation of the displayed monomers fused with Aga2p. Since no structural data is available on the N-terminal region of GOx, the inventors used in silico homology modeling of the Mut1 , Mut2, and Mut9 sequences to gain insight into the potential mechanism of stabilization of our discovered mutants. The modeling suggests that the Leucines in position 9 and 13 are part of an alpha helix corresponding and that insertion of a Proline in either one of the two positions breaks this helix. These broken helices could conceivably allow the GOx catalytic core to explore more diverse orientations, thereby facilitating the formation of more stable homodimers. Typically, thermal-stability improvements of only between 0.5 and 1 °C are obtained via single amino acid substitutions in catalytic enzymes. This surprising increase in thermostability of ~10 °C of our mutant enzymes attributable to only a single amino acid substitution is probably attributable to the complexity of the environment surrounding the anchored fusion enzyme on the yeast cell wall, where a slight modification in orientation results in major changes in the stability of the GOx dimer.

Example 3

The inventors characterized the morphology of the alginate hydrogel capsules using confocal fluorescence microscopy. In bright field mode, the alginate layer was invisible (Figure 4A, top). In fluorescence mode, a well-defined conformal alginate hydrogel is seen surrounding the cells (Figure 4A, bottom). The orthogonal views show a homogenous distribution of hydrogel surrounding the dark cell mass in the center.

The inventors investigated the kinetics of gel formation by measuring the thickness of the hydrogel shells using fluorescence microscopy following incubation of the reaction for various lengths of time. Figure 4B shows that the alginate shells grew rapidly during the first few minutes, reaching a particle diameter of ~30 pm including the entrapped cell. This rapid growth was followed by a plateau phase wherein extended incubation times (>10 min) did not increase the size of the hydrogel shells. This is an advantageous feature of the system that the inventors attribute to auto-inhibition of the reaction. Auto-inhibition is expected upon formation of the alginate shell which provides a diffusion barrier that prevents new macromonomers from reaching the GOx/HRP enzymes at the cell surface. In this case, the reaction will only promote the formation of radical species in the proximity of the cell surface, therefore the extremely short half-life of the radical species does not allow the free radicals to encounter and react with unpolymerized monomers after the hydrogel reaches a critical thickness.

The inventors tested whether the hydrogel shells could entrap proteins present in the medium and serve as a diffusion barrier for macromonomers and proteins. The cell encapsulation experiment was performed in the presence of mCherry, a ~29 kDa fluorescent protein. The results showed that mCherry was trapped and held within the gel network over several hours. The inventors expect that HRP and macromonomers, with their higher molecular weight of ~44 kDa and ~70 kDa respectively, are similarly limited in their diffusion by the forming gel. The inventors note that in the kinetic growth experiments, it is not possible that the growth plateau is due to consumption of the substrate since the reaction is performed at saturating concentrations of glucose, HRP and macromonomer species. The inventors tested the cytocompatibility of the alginate shells by measuring growth curves of yeast alone, in the presence of the reaction mixture but lacking HRP, and encapsulated in the hydrogel capsules. All three of these growth curves were found to be the same, demonstrating that the hydrogel capsules did not inhibit cell growth. The cells are able to grow and divide inside the capsules until they eventually break free from the alginate hydrogel capsules.

The artificial alginate layer formed around cells displaying functional GOx also served as an enhanced safeguard against cytotoxic agents. The protective function of the alginate shell was tested through incubation of a heterogeneous population of yeasts containing less than 10% of GOx positive encapsulated cells with Zymolyase, which is a mixture of lytic enzymes that digests the yeast cell wall leading to osmotic lysis and cell death. After incubation with 300 U/ml of Zymolyase for 60 minutes, the cell mixture containing 10% GOX-positive cells was recultivated and sorted in single cell mode onto agar plates. The clonal colonies were then assayed for GOx activity and the fraction of GOx-positive colonies was found to be 100%. This result indicated that the hydrogel capsules conferred resistance to enzymatic lysis to GOx- positive cells (Figure 9A) with zero background level survival. Although the hydrogel capsule is porous and allows the diffusion of small nutrients and molecules, it acts as diffusion barrier for large macromolecules such as proteins and enzymes and therefore inhibited Zymolyase enzymes from attaching to and digesting the cell wall. The porosity of the alginate gel layer and as a consequence the efficacy of the diffusion barrier can be tuned by varying the concentration of crosslinkable moieties in the conjugation reaction.

Example 4

The alginate layer formed around the GOx positive cell population provides new physical properties to the encapsulated cells. The inventors exploited the significant increase in cell size that occurs upon enzymatic encapsulation in a filtration process to separate encapsulated cells from background populations. The inventors applied a vacuum reverse- filtration strategy using polycarbonate filters. In this process, populations of yeast cells containing— 10% pGAL-GOx genotypic positive cells were first encapsulated in fluorescent alginate and then filtered through 10 pm pores size filters. As expected the bare yeast cells (2 to 5 pm in diameter) could easily pass through the filter and were discarded in the flow through. Encapsulated yeast cells (20 to 30 pm in diameter), however, were retained on the filter and could be recovered for further analysis. Figure 4B shows a micrograph of green fluorescent hydrogel capsules adhered onto the polycarbonate filter membrane. The dark spots in the image are the membrane pores. Following this filtration procedure, the yeast cell population encapsulated in fluorescent hydrogel and positive for the pGAL-GOx gene was enriched from ~ 15% to—95% within two filtration steps, each requiring only -2 minutes of processing time (Figure 4C). Both the cell survival and filtration procedures have relevant applications for the enrichment of genetic libraries in directed evolution experiments, or in automated genomics applications where a specific genetically tagged cell line can be separated from genetically diverse mixtures, thus providing a rapid, high-throughput, scalable and affordable alternative to cell sorting.

Example 5

As an alternative to alginate, the inventors also investigated the use of the polysaccharide chitosan for encapsulation of cells. Chitosan has been used in biomedical applications and drug delivery systems due to its biocompatibility and its mechanical and chemical properties. Dissolution of chitosan in aqueous buffers is achieved only at acidic pH < 6, when the amino groups of the polymer are protonated. This solubility switch in aqueous environments causes the swelling and deswelling of the polymer based on the pH. The pH sensitivity of chitosan hydration presented an interesting feature for cell encapsulation applications by providing a tool to tune the permeability and elasticity of the polymer and regulate the accessibility to the encapsulated cells. Chitosan with phenol and fluorescein moieties (Scheme 2) was used for cell encapsulation using a slightly modified protocol at pH 6. Phenolated chitosan showed similar behaviour to the modified alginate, however, the chitosan shells showed higher stability and less tendency to form non-specific interactions with filter membranes as compared to alginate. Chitosan capsules were found to be less prone to aggregation and the gel capsules did not break under mechanical stress such as during centrifugation and filtration.

The pH sensitivity of the chitosan capsules was investigated by varying the pH of the solution containing encapsulated cells from 6 to 8 and observing the behaviour of the hydrogel capsules using fluorescence microscopy. The inventors note that yeast cells are stable and remain viable in a pH range of 6 to 8, therefore tuning the capsules diameter with pH is compatible with in vivo processing and propagation of the cells. Yeast cells encapsulated with fluorescent chitosan at pH 6 presented a well-defined and water swollen shell that was not visible in bright field mode but was clearly visible in fluorescence mode (Fig 5 upper panels). By resuspending the same cell sample in basic buffer (pH 8), the inventors observed a radical reduction in the size of the gel layer around the cells. At pH > 7.5, the chitosan capsules were clearly visible in both bright field and fluorescence modes (Figure 5 lower panels). The pH- tunable solubility of the chitosan provided a means to visualize the hydrogel layer surrounding the cells in bright field mode, eliminating the requirement of fluorescence microscopy to visualize the capsules. This is beneficial since high energy light sources can cause cell damage under prolonged exposure. To correlate the size of the capsules with the solution pH, the inventors incubated a population of chitosan encapsulated cells at increasing pH from 6 to 8 and imaged them using fluorescence microscopy. The inventors found a clear decrease and narrowing of the size distribution from -80 pm to -15 pm upon increased the pH from 6 to 8. Collapse of the gel at pH 8 is expected to alter the elasticity and porosity of the gel. The ability to finely control physical features of the artificial capsules can be relevant for different biochemical separation processes and controlled release of proteins from the cells in many biotechnological and medical applications.

Example 6

The first step toward high-throughput screening of new candidate enzymes is successful expression of the wild type enzyme. Figure 12 demonstrates that D-amino acid oxidase from Rhodotorula gracilis can be successfully expressed and displayed on the yeast cell wall using Aga1/Aga2 as anchors. Successful staining for the C-terminal polyhistidine tag on the protein indicates successful expression. A critical step for our system is the folding/export of enzymes to the cell wall in a functional form. Figure 13 demonstrates that under our expression conditions (described previously), the inventors can display FAD-dependent enzymes in a functional form (correctly folded and containing the FAD cofactor).

The purpose of figure 14 is to demonstrate 2 points: (1 ) specificity of the encapsulation reaction and (2) time course. By staining with the anti-histidine antibody (y-axis) and then encapsulating the cells (x-axis), the inventors observe that only cells displaying the enzyme become encapsulated, whereas those that do not display the enzyme do not become encapsulated (no green fluorescence). The reaction is therefore specific only to cells displaying the enzyme, and the neighboring cells are not encapsulated. The second point is to note that the reaction is rapid and completed in 10 minutes. In some cases, shorter or longer times can be used (2-30 minute reaction times).

Figure 15 demonstrates a‘mock’ screening experiment to validate the approach. The inventors are interested in the ability of the encapsulation system to allow isolation of rare cells in a large background of cells. For this purpose, the inventors first‘drop-in’ a given percentage of cells displaying the WT enzyme into a background of cells lacking the enzyme. The inventors are interested to understand if the fraction of cells isolated by encapsulation and cell-sorting corresponds to the fraction dropped in to the starting population. Indeed, they find a direct correlation between the percentage of positive DAOx cells present in each starting sample and the fraction of cells found in the R2 gate which are positive for both display of the enzyme and encapsulation in modified Alginate. This demonstrates the ability to screen and enrich cells from a large background population.

For enzymes that show low expression levels or low catalytic turnover rates, the diffusion of H2O2 and alginate macromonomers away from the cell surface over the time course of the reaction can result in the critical threshold of radicals not being produced near the cell surface, resulting in no encapsulation. To counteract this effect, the inventors found that addition of PEG as a viscosity enhancer greatly increases the sensitivity and allows us to encapsulate cells that would otherwise not be able to carry out the reaction (Fig. 16).

Figure 17 demonstrates that with the inventors’ plasmids and expression protocol, they can successfully display the porphyrin/heme-dependent enzyme horseradish peroxidase on the outer yeast cell wall. Almost 50% of the yeast population tested positive for HRP expression by antibody staining, indicating a display of the full-length protein.

Display of enzymes that are correctly folded and contain the correct co-factor is non-trivial. Here the inventors demonstrate the ability to successful express, fold, and display porphyrin/heme-dependent HRP on the outer cell wall of yeasts (Fig. 18). The enzyme shows strong activity relative to a negative control lacking the enzyme.

For enzymes with low expression levels or low catalytic turnover rates, there may not be sufficient production of H2O2 to trigger the reaction. In these cases, the inventors found that addition of soluble HRP to the medium can increase the sensitivity of the gel polymerization reaction (Fig. 19).

Figure 20 confirms successful expression of the FAD-dependent enzyme uric acid oxidase, which is an approved biomedical therapeutic enzyme. More than 50% of the yeast population tested positive for protein display by the antibody staining procedure, indicating successful display of the full-length protein.

In figure 21 , the inventors are showing the functionality of the enzyme uric acid oxidase using the yeast Aga1/Aga2 anchor system. Uric acid oxidase is a FAD-dependent enzyme and forms a homotetramer in the active state.

Discussion The inventors report a genetically encoded system for single eukaryotic cell encapsulation in well-defined and robust synthetic hydrogel capsules. Physical properties of the gel including thickness and pH-responsiveness allowed for simple cell separation and cell lysis procedures to be carried out to isolate pGAL-GOx-positive genotypic cells with high efficiency. Such a system is adaptable to a large class of bio-initiators, proteins and enzymes that can trigger radical polymerization providing a valuable tool for visualization, protection and isolation of genetically high value cells from heterogeneous cell populations. The inventors explored the use of two different hydrogel materials, alginate and chitosan, exploiting the robust mechanical properties for size-based filtration and the cytoprotective properties for resistance to lytic enzymes. This new artificial eukaryotic cellular phenotype can provide advantages in many synthetic cellular selection or evolution campaigns. This research points towards the development of synthetic extracellular matrices or artificial biofilms that can be programmed at the genetic level. This exciting area of bottom-up cell-directed materials synthesis is one of the areas where the inventors expect the presented reaction system to find significant applications.