Engineered single-chain insulins with increased thermostability and solubility and their use in animal cell culture applications

Abstract

The present invention relates to engineered single-chain insulin variants of animal origin and their use in cell culture applications such as cultivated meat, biopharmaceutical manufacturing, and research use. The variants are characterized by its increased thermostability and solubility in comparison to native mature insulins.

Description

[0001] ENGINEERED SINGLE-CHAIN INSULINS WITH INCREASED THERMOSTABILITY AND SOLUBILITY AND THEIR USE IN ANIMAL CELL CULTURE APPLICATIONS SPECIFICATION

[0002] 1. FIELD OF THE INVENTION

[0003] The present invention relates to engineered single-chain insulin variants of animal origin and their use in cell culture applications such as cultivated meat, biopharmaceutical manufacturing, and research use. The variants are characterized by its increased thermostability and solubility in comparison to native mature insulins.

[0004] 2. BACKGROUND OF THE INVENTION

[0005] Insulin is a peptide hormone that regulates glucose, lipid, and protein metabolism. Insulin is secreted by the beta cells of the pancreas and is essential for the maintenance of blood glucose homeostasis (Norton et al., 2022). At cellular levels, insulin promotes glucose uptake by increasing GLUT4 translocation to the plasma membrane. Additionally, insulin triggers mitogenic signals that promote cell growth, cell division, and migration (Le et al., 2023).

[0006] Insulin and its engineered analogs have been used for diabetes therapy for several decades. Aside from its therapeutic role, insulin also has many applications for the cultivation of animal cells in serum-free media. Many cell types require a media supplementation with insulin to consume glucose and proliferate, making it essential in applications such as cultivated meat, biopharmaceutical manufacturing, and research use. However, the limited stability and solubility of this protein increases production and distribution costs, impacting directly on high market prices. For this reason, it is important to develop novel animal insulin variants with enhanced thermostability and solubility, specifically designed for cell culture applications.

[0007] Native insulin is produced by animal cells as a precursor called proinsulin, in which the A- and B-chains are joined by the C-peptide (31 residues). The C-peptide is further cleaved in the secretory pathway to produce functional mature insulin.

[0008] Single-chain insulins (SCIs) are analogs in which the A- and B-chains are joined by designed small linkers (6-11 residues). In contrast to proinsulin and other insulin precursors used by the industry, SCIs can bind insulin receptors (IRs) with high affinity and trigger the same functional responses as native insulin. Since SCI analogs are functional as single polypeptide chains, they can be producedin microbial hosts without cleaving the linker, reducing the number of downstream processing steps. Additionally, the linker can be designed to increase the thermostability and solubility of the protein. Altogether, the improved properties of SCIs can significantly reduce production costs and facilitate the supply chain in comparison to native insulin (Mao et al., 2019). These inexpensive analogs are particularly interesting for formulating serum-free media for large-scale applications such as cultivated meat and biomanufacturing, where native insulin prices are prohibitive.

[0009] Patent US8883449B2 describes single-chain insulins with linkers composed of 10 residues, with the sequence X1-X2-X3-X4, where XI is T, K, or R; X2 is P, K, or D; X3 is K, P, or E; X4 is a heptapeptide selected from several sequence combinations. The invention mentions that the analogs have better stability and solubility, but no experimental data is presented related to this.

[0010] Patent US10392429B2 discloses single-chain insulins with linkers composed of 6-11 residues, with the sequence X1-X2-X3-X4-X5, where XI is D or E; X2 is D or E; X3 is hepta-, hexa-, penta-, tera-, tri- or di-peptide of any sequence; X4 is R or K; X5 is R or K. Additionally the invention describe substitutions in the positions of A- and B-chains: A(8) is R, K or H; A(14) is E or D; B(29) is D or E. The invention also presents data on the increased thermostability of the analogs.

[0011] The present invention describes novel single-chain insulin variants with improved thermostability, and solubility than previously reported single-chain insulins, specifically designed for inexpensive production process and its use in cell culture media.

[0012] 3. SUMMARY OF THE INVENTION

[0013] The present invention focused on engineered single-chain insulins that can be used in animal cell culture applications, providing more thermostability and solubility optimized to reduce cost and facilitate the supply chain in comparison with the commercially available native insulin. These variants provide a solution for the need for an insulin that maximize its activity in a cost-effective way for the formulation of serum-free media for large-scale applications such as cultivated meat, biomanufacturing, and research markets.

[0014] In general, these variants are single-chain insulins derived from animal origin Bos Iannis. Sus scrofa, Ovis aries. Capra hircus. Bubalus bubalis, Oryctolagus cunicuhis. Gallus gallus, Meleagris gallopavo, Anas platyrhynchos, Anser anser. among other species commonly used in meat and dairy industries. The variants contain an A-chain and B-chain linked by a C-peptide of 6 amino acid residues designed as Cl, C2, C3, C4, C5, and C6. The connecting polypeptide has the composition as follows: Cl is arginine (R); C2 is glutamic acid (E) or aspartic acid (D); C3 is valine or alanine (V or A); C4 is glutamic acid (E); C5 is glycine (G) or glutamine (Q) or asparagine (N); and C6 is proline (P). The variants can have any combinations of the following substitutions in the A-chain comparedto native insulin: A8 position substituted by H, E, or S; A9 position substituted by T; A10 position substituted by V, I, or T; and A14 position substituted by E, K or Q. The variants can also have any combinations of the following substitutions in the B-Chain compared to native insulin: B29 position substituted by an acidic residue such as E or D; B30 position substituted by P or E.

[0015] In particular, the linker was designed to provide folding stability to the variants by protecting the hydrophobic exposed residues in the A- and B-chain, reducing the fibrillation propensity, and increasing thermostability and solubility. The linker also has enough flexibility to enable threading to the site 1 of the insulin receptor allowing a bioactive binding mode. The substitutions in the A14 position increase the stability of these variants enhancing the electrostatic charge on the protein surface while the substitutions in the A8 position increase the folding stability by protecting some residues of the helix of A-chain. Finally, the substitutions in the B29 position help to stabilize the folding by forming a salt bridge between the end of the B-chain and the start of the linker. The present invention features a set of insulin variants that exhibit higher thermostability, increased solubility, and less aggregation propensity.

[0016] 4. BRIEF DESCRIPTION OF THE FIGURES

[0017] Figure 1. Structural comparison of the native insulin and the single-chain insulins presented in this invention. (A) Native insulins are composed of two chains, -A and -B (colored in light and dark gray, respectively), which are connected by two interchain disulfide bridges. (B) Single-chain insulins presented in this invention (exemplified with the bM2 variant) has a linker of 6 residues (highlighted in black) that joins both chains into a single polypeptide. The positions modified to generate more stable and soluble variants are shown as spheres. This image was created with PyMOL.

[0018] Figure 2. Dimeric forms of bM2 and bM2-664 insulins. The bM2 (A) and bM2-664 (B) insulin variants can form dimers through backbone interactions between their antiparallel -strands present in the B-chain. Both variants are shown in cartoon and surface representation, where A- and B- chains are colored in dark and light gray, respectively. The linkers are highlighted in black. This image was created with PyMOL.

[0019] Figure 3. The bM2 and I1M2-664 insulins can be threaded by the insulin receptor aCT domain.

[0020] Protein-protein docking analyses predict that these variants are able to interact with the isoform A of the bovine insulin receptor. This figure, created with PyMOL, shows the possible mechanism of bM2 (A) and bM2-664 (B) interacting with the receptor. These variants can adopt an “open” conformation where the flexible linker can extend to thread with the aCT domain to bind with the insulin receptor. In both complexes, the receptor is colored in dark gray in “surface” representation, while the insulin variants are depicted in “cartoon” representation.Figure 4. Expression vectors for insulin variants. (A) Transcriptional units design for the expression of bM2, bM7, bMIO, bM25, and bM2-664 variants in Pichia pasloris. exemplified for bM2 variant. The genes are under the control of A0X1 promoter and A0X1 terminator. (B) Plasmid map of the expression vector PICZa used for the expression of the variants, exemplified for bM2. Figure 5. Screening of insulin variants in deep well plates. (A) Western blot analyses of supernatants obtained from deep well cultures of bM2, bM7, bMIO, and bM25 strains. Insulin variants were detected using a mouse monoclonal anti-insulin B-chain antibody and HRP-conjugated anti-mouse secondary antibody. (B) SDS-PAGE analysis of supernatants obtained from deep well cultures of bM2-664 strain. The first variants (bM2, bM7, bMIO, and bM25) were detected by western blot due to low expression levels while the optimized variant (bM2-664) detection was evident by SDS-PAGE due to higher expression.

[0021] Figure 6. Production of bM2 and bM2-664 variants in 10 L bioreactor. (A, C) Biomass, total protein, and insulin in the fermentation process of bM2 (A) and bM2-664 (C) insulins. “B” is the batch stage, “F” is the fed-batch stage, and “I” is the induction stage. In the “X” axis is shown the fermentation time (hours), in the left “Y” axis is shown the grams of wet cell weight per liter (gWCW / L), and in the right “Y” axis are shown total protein concentration (mg / L) and bM2 or bM2-664 protein concentration (mg / L). (B, D) SDS-PAGE (12-20% acrylamide) analysis of fermentation samples obtained at different culture times of bM2 (B) and bM2-664 (D) strains.

[0022] Figure 7. Downstream process of bM2 and bM2-664 variants. The supernatants obtained after the centrifugation and microfiltration operations were precipitated using zinc chloride. After centrifugation and supernatant removal, the crystals were resuspended, and the solutions were microfiltered. Finally, the concentrated insulin solutions were lyophilized to obtain samples as a dry powder.

[0023] Figure 8. Bioactivity assays of bM2 and bM2-664 variant in C2C12 cells. The bioactivity of the purified bM2 (A), bM2-664 (B), and human native insulin were tested by quantifying the proliferation of C2C12 cells in serum-free media. The cells were cultivated for 72 h, stained with Hoechst, and automated counted using an imaging reader. Control: media without insulin.

[0024] Figure 9. Thermostability assays of bM2 and bM2-664 variants. (A) bM2 and human native insulin samples were incubated at two different temperatures, 45°C and 60°C, for one week. After the incubations, the proteins were used for the bioactivity assays in C2C12 cells. Negative control: media without insulin, positive control: samples stored at -20°C. (B) The same assay described in (A) was performed to compare bM2-664 and human native insulin.

[0025] 5. DETAILED DESCRIPTION OF THE INVENTIONThe present invention is focused on one embodiment of recombinant single-chain insulin variants designed to eliminate the necessity for post-fermentation processing while exhibiting enhanced stability and solubility compared to mature native insulin. These engineered variants were designed for animal cell culture, including applications such as research use, cultivated meat production, and pharmaceutical biomanufacturing (e.g., antibodies and vaccine production). The present invention provides insulin variants that contain insulin B-chain and A-chain polypeptides derived from animal origin connected by a short 6-residue polypeptide linker. These single-chain variants were designed using A- and B-chain sequences sourced from animal insulins, including bovine, chicken, porcine, and ovine insulins, by way of non-limiting examples. The short linker sequence was obtained from the native C-peptide of proinsulin (a precursor of mature insulin), enabling the construction of singlechain insulins of 57 residues. In one embodiment, the amino acid sequences of the A-chain, B-chain, and C-peptide of the bovine insulin, for comparative purposes, are provided below as SEQ ID NOs: 1-3:

[0026] SEQ ID NO: 1 (bovine B chain):

[0027] “FVNQHLCGSHLVEALYLVCGERGFFYTPKA”

[0028] SEQ ID NO: 2: (bovine A chain):

[0029] “GIVEQCCASVCSLYQLENYCN”

[0030] SEQ ID NO: 3 (bovine C-peptide)

[0031] “RREVEGPQVGALELAGGPGAGGLEGPPQKR”

[0032] Based on the results of in silico analysis (example 1), four single-chain variants derived from bovine sequence were selected: bM2, bM7, bMIO, and bM25. These variants were subsequently expressed in Pichia pastoris. Among these, the bM2 variant composed by the linker “REVEGP”, exhibited the best expression and biological activity in vitro assays (example 3 and 7). This variant was called bM2 because its linker derived from the 2thposition of the sliding window approach of the bovine proinsulin (example 1). The amino acid sequence of the bM2, bM7, bMIO, and bM25 variants are provided in SEQ ID NOs: 4-7, respectively. The A- and B-chains derived from native bovine insulin are highlighted in underline, while the 6-residue linker originating from the C-peptide is presented in italics:

[0033] SEQ ID NO: 4 (bM2):

[0034] “FVNQHLCGSHLVEALYLVCGERGFFYTPKA^ / .J / .GVMIVEQCCASVCSLYQLENYCN”

[0035] SEQ ID NO: 5 (bM7):

[0036] “FVNQHLCGSHLVEALYLVCGERGFFYTPKAAA / V'ONAGIVEQCCASVCSLYQLENYCN”

[0037] SEQ ID NO: 6 (bMIO):

[0038]

[0039] SEQ ID NO: 7 (bM25):

[0040] “FVNOHLCGSHLVEALYLVCGERGFFYTPKAGPPQKRGIVEOCCASVCSLYOLENYCN”

[0041] Following the design approach of variant bM2, a single-chain insulin similar to variant bM2 was generated in this invention, incorporating A- and B-chain polypeptide sequences and the linker from chicken insulin. The amino acid sequence of this variant, named cM2, is shown in SEQ ID NO: 8. The A- and B-chains derived from native chicken insulin are indicated by underlining, while the 6-residue linker originating from the C peptide is depicted in italics:

[0042] SEQ ID NO: 8 (cM2):

[0043] “AANQHLCGSHLVEALYLVCGERGFFYSPKARDFEQPGIVEOCCHNTCSLYOLENYCN”

[0044] This is a feature of the present invention, where the bM2 and cM2 variants exhibit higher thermodynamic stability than mature bovine insulin at 37°C (Table 1), as well as possessing greater solubility due to the amino acid composition of the linker (presence of charged residues such as glutamine and arginine, which increase its solubility). It is also presumed that these variants, because of their design, are less prone to fibrillation, showing an isoelectric point close to pH 5.5, with a soluble formulation feasible under neutral conditions (pH 7.0) (Table 1). Additionally, these variants can form stable dimers according to AlphaFold predictions through hydrogen bonds generated between the backbone atoms of residues PheB24, PheB25, and TyrB26 of the B-chain, which can fold into antiparallel -strands (Figure 2). The cM2 variant simultaneously reports better stability than bM2, also exhibiting greater solubility according to the CamSolpH server (Table 1). It is presumed that the cM2 variant shows biological activity similar to the bM2 variant, which has been experimentally tested (Example 7), considering the protein-protein docking essays showed that this variant has a similar ZDOCK interaction score to the M2 variant (Table 1, Figure 3).

[0045] Table 1. Stabilities, solubilities and affinity scores to insulin receptor of insulin variants.

[0046] Insulins Mature bovine insulin bM2 cM2

[0047] AAG foldX (kcal / mol) 0 (14.2) -5.4 -6.8

[0048] AAG Rosetta (REU) 0 (-36.4) -43.8 -49.5 molecular weight (Da) 5721.6 6389.3 6423.2

[0049] pl 5.52 5.30 5.47

[0050] charge at 5.5 -0.6 -1.4 -0.9

[0051] charge at 7.0 -4.5 -5.5 -5.5

[0052] Camsol score -0.014 ± 0.04 0.296 ± 0.03 0.890 ± 0.07 ZDOCK score 2050.2 ± 60.4 2068.1 ± 68.5 2033.0 ± 48.5linker - REVEGP RDVEQP AAG foldX was calculated at 37°C with pH 7.0, relaxing the structure obtained from AlphaFold2 with the ‘RepairPDB’ program.

[0053] AAG Rosetta was calculated in Rosetta Energy Units (REU) from a relaxed structure with ‘minimize with cst’ and ‘scorejd2’ programs.

[0054] pl and charge at certain pH were estimated with the IPC2 server.

[0055] The mean and standard deviation of the Camsol scores were measured over a pH range of 3 to 9 from values obtained from the CamSolpH server.

[0056] The mean and standard deviation of ZDOCK scores were calculated from top predictions of the insulin / receptor complexes performed with ZDOCK.

[0057] The present invention is not limited to bovine and chicken single-chain insulins with two similar polypeptide linkers ("REVEGP" or "RDVEQP", respectively). It is also envisaged that linkers similar in sequence to these two examples may also be present in animal insulins such as sheep and porcine insulins (sM2 and pM2, respectively), considering that the linker fragment is relatively conserved at the C-peptide level in animal insulins. These insulins are shown as having SEQ ID NOs: 9-10, as follows:

[0058] SEQ ID NO: 9 (sM2):

[0059] “FVNOHLCGSHLVEALYLVCGERGFFYTPKAREEEGPGIVEOCCAGVCSLYQLENYCN”

[0060] SEQ ID NO: 10 (pM2):

[0061] "FVNQHLCGSHLVEALYLVCGERGFFYTPKAR 4ENPGIVEQCCTSICSLYQLENYCN"

[0062] The variants of animal insulins of the present invention contain a linker represented by "C1C2C3C4C5C6" with the following amino acid composition:

[0063] • Ci: At the first position of the linker, there must be an arginine (R).

[0064] • C2: At the second position of the linker, it must contain a negatively charged amino acid, which can be glutamic acid or aspartic acid (E or D).

[0065] • C3: At the third position of the linker, it must have a neutral hydrophobic amino acid, which can be valine or alanine (V or A).

[0066] • C4: At the fourth position of the linker, it must contain glutamic acid (E).

[0067] • C5 : At the fifth position of the linker, it must contain glycine (G) or a non-charged polar amino acid, which can be glutamine or asparagine (Q or N).

[0068] • Ce: At the sixth and last position of the linker, it must contain a proline (P).

[0069] These insulin variants may introduce other minor conservative modifications in the linker sequence in the positions C2, C3, and C5, especially those substitutions of amino acids with similar chemical properties, without affecting the scope of the present patent.Table 2. Variants designed with Al with lower aggregation and better thermodynamic stability than the bovine M2.

[0070] model sap score AAG(relax)

[0071] bM2-682 34.67 -3.76

[0072] bM2-664 38 -7.84

[0073] bM2-173 39.26 -0.54

[0074] bM2-668 40.54 -13.06

[0075] bM2-368 41.34 -2.3

[0076] bM2 42.07 0

[0077] The sap score was estimated on monomeric structures minimized with the “relax” Rosetta’s protocol. The AAG values were calculated from the total score reported from the “relax” protocol in comparison to the M2.

[0078] In the second design stage, the objective was to optimize M2 variants for better expression. For this purpose, additional substitutions in A- and B-chains were introduced using artificial intelligence (Al) and rational design to reduce aggregation propensity and improve the expression potential of the variants (Example 1). The results showed 5 variants of the 89 Al-designed insulins that exhibit a smaller hydrophobic surface than the bM2 insulin with better stability (Table 2). The amino-acid sequence of these optimized variants derived from the bovine M2 insulin using Al, as disclosed in Table 2, are shown in SEQ ID NOs: 11-15. Residues highlighted in bold indicate substitutions in the A- and B-chains of the wild-type insulin. The A- and B-chains derived from native bovine insulin are indicated by underline, while the 6-residue linker originating from the C peptide is depicted in italics.

[0079] SEQ ID NO: 11 (bM2-682) “FVNOHLCGSHLVEALYLVCGERGFFYTPDPREFEGPGIVEOCCNSTCSLKQLENYCN”

[0080] SEQ ID NO: 12 (bM2-664):

[0081] “FVNOHLCGSHLVEALYLVCGERGFFYTPDPREFEGPGIVEOCCESICSLKQLENYCN”

[0082] SEQ ID NO: 13 (bM2-173):

[0083] “FVNOHLCGSHLVEALYLVCGERGFFYTPKAREFEGPGIVEOCCSSVCSLEQLENYCN”

[0084] SEQ ID NO: 14 (bM2-668):

[0085] “FVNOHLCGSHLVEALYLVCGERGFFYTPEPREFEGPGIVEOCCNTICSLEQLENYCN”

[0086] SEQ ID NO: 15 (bM2-368) “FVNOHLCGSHLVEALYLVCGERGFFYTPEEREFEG GIVEOCCESVCSLOQLENYCN” Additionally, in the present invention, two variants were designed rationally based on the bM2 and cM2 insulins to improve the expression of these variants by including substitutions of charged amino acids at positions B29, A8, and A14. The substitution of residues B29 and A14 with glutamic acid (E) allowed reducing the isoelectric point of the bM2 and cM2 variants to a more acidic pH (below5.5), increasing their solubility at a neutral pH (between 7.0 and 8.0) (Table 3). Also, the substitution of lysine with glutamic acid at position B29 helps to stabilize the folding by forming a salt bridge between the carboxyl group of B29E and the amino group of residue C1R. Meanwhile, the glutamic acid at position A14 in the bM2 and cM2 variants increases the stability of these variants while enhancing the electrostatic charge of their surface. Otherwise, the substitution of tyrosine with histidine at position A8 of the M2 variant increases its folding stability by protecting residues A1-A4 of the first helix of chain A.

[0087] The amino-acid sequence of single-chain insulin variants of the present invention derived from bM2 and cM2 using rational design are given in SEQ ID NO: 16-17. Residues highlighted in bold indicate substitutions in the A- and B- chains. The A- and B-chains derived from native insulin are indicated by underline, while the 6-residue linker originating from the C-peptide is depicted in italics.

[0088] SEQ ID NO: 16 (bM2-7)

[0089] “FVNOHLCGSHLVEALYLVCGERGFFYTPEA / / J7-.G7)GIVEOCCHSVCSLEOLENYCN”

[0090]

[0091] Table 3. Solubility and pl predictions of the optimized M2 insulin variants.

[0092] model camsol score&pl+pH 5.5 charge* pH 7.4 charge* molecular weight* cM2-3 0.98 ± 0.14 4.75 -3.8 -8.4 6390.1 cM2 0.89 ± 0.07 5.47 -0.9 -5.5 6423.22 bM2-368 0.77 ± 0.21 4.44 -5.2 -9.4 6471.26 bM2-664 0.75 ± 0.11 4.75 -3.3 -7.5 6439.3 bM2-7 0.62 ± 0.09 4.82 -3.7 -8.4 6422.23 bM2-682 0.59 ± 0.04 4.95 -2.4 -6.5 6412.24 bM2-173 0.55 ± 0.04 5.02 -2.4 -6.5 6371.23 bM2-668 0.53 ± 0.12 4.57 -4.3 -8.4 6453.29 bM2 0.24 ± 0.04 5.3 -1.4 -5.5 6389.29 &: The mean and standard deviation of the Camsol scores were measured over a pH range of 3 to 9 using values obtained from the CamSolpH server. The IPC method was used for pKa correction.

[0093] +: Isoelectric points (pl), charges at certain pH levels, and the molecular weight of the insulin variants (Da) were estimated using the IPC2 server.

[0094] The features of the bovine and chicken optimized M2 variants designed with artificial intelligence or rational design, compared to native bovine insulin (blN), are shown in Table 4. These variants theoretically exhibit higher solubility than native insulin and the first M2 variant, with some, such as bM2-682, bM2-668, and bM2-664, showing better monomer stability than these controls. Notable among these variants are bM2-7, bM2-668, bM2-368 and bM2-664, where dimer folding analyseshighlight that these variants can form more stable dimers than the M2 insulin (AG dimer < -230.92 REU) with less exposure of hydrophobic residues on their surface (Agg < 66.29, value reported to M2), indicating a lower tendency to aggregate.

[0095] Table 4. Stabilities, solubilities and aggregation predictions of the bovine insulin variants.

[0096] AG

[0097] modelB29 B30 Cl C2 C3 C4 C5 C6 A8 A9 A10 A14 monomer Sol AG dimer Agg blN K A A S V Y -115.7 -0.06 -212.9 84.77 bM2-682 D P R E V E G P N S T K -124 0.59 -223.82 73.39 bM2-173 K A R E V E G P S S V E -112.6 0.55 -226.38 69.21 M2 K A R E V E G P A S V Y -111.8 0.24 -230.92 66.29 bM2-7 E A R E V E G P H S V E -109.9 0.62 -231.97 63.24 bM2-668 E P R E V E G P N T I E -124 0.53 -233.88 61.99 bM2-368 E E R E V E G P E S V Q -101.8 0.77 -240.08 62.02 bM2-664 D P R E V E G P E S I K -120.5 0.75 -270.31 58.91 “AG monomer” is the average value of the “total score” obtained from the replicates minimized with the “relax” Rosetta protocol from Alphafold2 models generated for each variant. These values were measured in Rosetta energy units (REU).

[0098] “Sol” is the average solubility score of each variant calculated from the CamSol scores estimated for a pH range of 3 to 9 according to the predictions of the CamSolpH server.

[0099] “AG dimer” is the folding energy score of the dimer in Rosetta energy units (REU) determined by the Cao protocol. The lower values suggest more favorable folding energy.

[0100] “Agg” is the sum of contributions from hydrophobic amino acids present on the surface of the dimer (sap score). The higher values indicate that these proteins have a major tendency to aggregate.

[0101] blN is the bovine native insulin that was used as a control.

[0102] Table 5 shows that the single-chain chicken insulin variants (cM2 and cM2-3) exhibit better solubility and folding stability of the monomers compared to native chicken insulin (cIN) with the insertion of the M2 linker derived from the chicken C-peptide (linker “RDVEQP”). Natively, chicken insulin shows better solubility and higher stability for dimer formation than its bovine homologous variant, and these characteristics have been enhanced by the insertion of the linker and the incorporation of single-site mutations at residues B29 and A14 (Tables 4 and 5).

[0103] These results suggest the cM2-3, bM2-668, and bM2-664 variants should exhibit a better expression in Pichia pasloris. because they reported favorable folding energy in the formation of dimers with less tendency to aggregate (Tables 4 and 5).

[0104] Table 5. Stabilities, solubilities and aggregation predictions of the chicken insulin variants.AG

[0105] ModelB29 B30 Cl C2 C3 C4 C5 C6 A8 A9 A10 A14 monomer Sol AG dimer Agg cIN K A - - - - - - H N T Y -106.3 0.71 -235.07 53.8 cM2 K A R D V E Q P H N T Y -114.1 0.89 -240.86 63.57 CM2-V3 E A R D V E Q P H N T E -109.8 0.98 -247.16 54.79

[0106] “AG monomer” is the average value of the “total score” obtained from the replicates minimized with the “relax” Rosetta protocol from Alphafold2 models generated for each variant. These values were measured in Rosetta energy units (REU). “Sol” is the average solubility score of each variant calculated from the CamSol scores estimated for a pH range of 3 to 9 according to the predictions of the CamSolpH server.

[0107] “AG dimer” is the folding energy score of the dimer in Rosetta energy units (REU) determined by the Cao protocol. The lower values suggest more favorable folding energy.

[0108] “Agg” is the sum of contributions from hydrophobic amino acids present on the surface of the dimer (sap score). The higher values indicate that these proteins have a major tendency to aggregate.

[0109] cIN is the chicken native insulin which was used as control.

[0110] Regarding their potential biological activity, the protein-protein docking results of these variants binding with the insulin receptor, along with Cao’s protocol (Cao et al., 2022), suggest that selected candidates can bind to site 1 of the isoform A of the bovine insulin receptor with interaction energy similar to or greater than the bM2 insulin. Although the bM2-7 and bM2-664 variants show a lower binding energy with the receptor, these values suggest they can still favorably bind to the insulin receptor as they report a binding free energy change of less than -30 REU, covering a molecular surface contact greater than 450 A2according to Cao’s protocol (Table 6).

[0111] Table 6. Evaluation of the binding interaction of the bM2 insulin variants to site 1 of isoform A of the bovine insulin receptor.

[0112] model AAG binding (REU) contact molecular surface (A2)

[0113] cM2-3 -90.07 680.21

[0114] cM2 -83.85 647.56

[0115] bM2-668 -86.08 623.74

[0116] bM2-368 -82.87 631.65

[0117] bM2-173 -80.28 625.26

[0118] bM2-682 -79.73 623.22

[0119] bM2 -81.81 601.69

[0120] bM2-V7 -77.81 591.56

[0121] bM2-664 -76.44 590.09

[0122] The change in binding free energy (AAG binding) and the contact area of insulin when threaded with the receptor were estimated using the Cao protocol from the docking predictions performed with ZDOCK of insulin in an open conformationbound to the insulin receptor. The lower the binding free energy (AAG binding < 0), and larger contact area suggest more favorable the insulin-receptor binding.

[0123] In Table 7, the substitutions present in the single-chain insulins derived from the present invention are summarized. The linker has the general sequence "R[E,D][V,A]E[G,Q,N]P". In positions B29 and B30 of the described single-chain insulins, these variants natively possess a lysine followed by an alanine (KA), which can be replaced by a negatively charged amino acid, such as glutamic and aspartic acid followed by a proline (for example, EP or DP), or by two subsequent negatively charged amino acids in sequence, such as EE. As for positions A8 to A10 of the single-chain insulins of the present invention, position A8 can be replaced by a charged or polar amino acid, such as Glu, Ser, or His, followed by a Thr at position A9 and maintaining Vai, Iso, or Thr residues at position A10. In positions A9 and A10 of the single-chain insulins, there must be a polar uncharged residue, such as Ser, Thr, Asn, or Gly, followed by a hydrophobic one, such as Vai or He, or another polar uncharged residue, such as Thr. Finally, at position A14, tyrosine must be replaced by a charged or uncharged polar amino acid, such as Glu, Lys, or Gin.

[0124] Figure 1 summarizes the general structure and modified positions of the insulin variants presented in this invention and their comparison with the native insulin. In concordance with the above, the singlechain insulin variants of the present inventions comprise the amino acid sequence selected from the group consisting of SEQ ID NOs: 4-17.

[0125] Table 7. List of substitutions in the bM2 insulins provided in the present invention.

[0126] positions

[0127]

[0128] On the other hand, in another embodiment, the present invention discloses a nucleic acid encoding any of the insulin variant described above.

[0129] In another embodiment, the present invention describes an expression vector that comprises the nucleic acid previously disclosed.

[0130] In one embodiment, the present invention shows a method for producing the insulin variants above disclosed, which comprises the general steps of: (a) introducing the expression vector of claim 11 into a cell capable of expressing the insulin variant; (b) culturing said recombinant cell; and(c) recuperating the insulin variant from the supernatant. In a preferred embodiment, the cell canbe a prokaryotic or eukaryotic cell. In another embodiment, the cells can be, for example, but not limited to, yeast, insects, and mammal cells.

[0131] Lastly, the present invention also discloses in another embodiment the use of the insulin variants previously described to be used in cell culture media for growing animal cells, including in industrial applications such as cultivated meat, biopharmaceutical manufacturing, and research use.

[0132]

[0133] 6. EXAMPLES

[0134] EXAMPLE 1: BIOINFORMATIC DESIGN OF SINGLE-CHAIN INSULIN VARIANTS WITH INCREASED THERMOSTABILITY AND SOLUBILITY.

[0135] In order to design these insulin variants, two strategies were employed to derive a functional peptide linker from the C-petide of bovine proinsulin. The first strategy utilized was the " sliding window" approach, which involved employing a fixed-size window ranging from 6 to 9 residues in length and moving it along the sequence of the C-peptide to obtain a set of linkers. The second strategy involved taking various combinations of the N- and C-terminal ends of the C-peptide and joining them to create linkers ranging from 6 to 9 residues in length. A total of 119 single-chain insulin variants were evaluated using an in silico pipeline that combined rational design tools and artificial intelligence (Al) to filter out the most stable and soluble single-chain insulin variants capable of binding to the bovine insulin receptor.

[0136] Initially, these designed variants were modeled as monomers using the AlphaFol d2 Colab notebook accessible through the ColabFold initiative (Mirdita et al., 2022). To build these models, the "alphafold2 _ptm" model was employed with 3 recycling iterations, utilizing custom templates and running a MMseqs2 search against the UniRefl 00 database to generate a diverse multiple sequence alignment (MSA). Amber force field was applied to relax the predicted structures. The best-predicted model for each variant was evaluated using FoldX 4.0 and Rosetta tools (BuB et al., 2018; Fleishman et al., 2011) to measure its folding stability (AAG) in kcal / mol or Rosetta energy units (REU) compared to mature bovine insulin. Moreover, the most stable variants were evaluated using the DeepSoluE web server (Wang & Zou, 2023) to predict their probability of solubility and determine their solubility status ("soluble" or "insoluble"). Predicted soluble variants were subsequently tested on the CamSolpH server (Oeller et al., 2023) to determine their solubility profiles compared to mature insulin at a range of pH (between 3 to 9). Finally, the best candidates were modeled with AlphaFold2 (Evans et al., 2022) in complex with site 1 of the bovine isoform A insulin receptor to obtain these single-chain variants in an "open" conformation (using the model "alphafold2_multimer_v3" with 3 recycle iterations). These variants were extracted from the best predicted models followed by evaluation against several replicates of the receptor site to measure their ability to bind to the insulin receptor through protein-protein docking assays using the ZDOCK tool (Chen et al., 2003). After that, the best docking predictions were tested using the CAO protocol with Rosetta (Cao et al., 2022) to determine the interaction energy of these designed insulins with the insulin receptor.

[0137] In order to reduce the aggregation propensity and improve the expression potential, additional substitutions in A- and B-chains were introduced using artificial intelligence (Al) and rational design.For this purpose, a conservation analysis was carried out to identify positions with higher variability and lower conservation, whose mutations could improve the formation of stable dimers and hexamers, and more soluble monomers. Based on A12CO results (Pei & Grishin, 2001), the positions with lower conservation were B29 and B30 in chain B, and A8, A9, A10, and A14 in chain A (Figure 1). For the IA approach, 683 unique sequences analogous to the bovine M2 variant were generated using ProteinMPNN (Dauparas et al., 2022) with a sampling temperature of 0.5, using the template from the bM2 / isoform A complex of the bovine insulin receptor. Of these variants, 89 were predicted to exhibit higher solubility and thermal stability than the M2 variant according to analyses conducted with the DeepSoluE and DeepSTABp servers. To evaluate their thermodynamic stabilities and aggregation propensity, the three-dimensional structures of these selected variants were modeled at the monomer level using AlphaFol d2, selecting the model with the best accuracy (using “alphafold_ptm” model with 3 recycles), minimizing the energy of the best-predicted models with Rosetta's “relax” protocol to measure their aggregation surface using the "sap score" (Voynov et al., 2009).

[0138] EXAMPLE 2: VECTORS AND CLONING OF bM2, bM7, bM10, bM25, and bM2-664 INSULIN VARIANTS.

[0139] The following section describes the vectors and associated cloning for generating bM2, bM7, bM10, bM25, and bM2-664 strains of Pichia pastoris.

[0140] Codon-optimized sequences for the thermostable insulin variants were synthesized by a third-party Gene Synthesis service that included cloning into PUC57 vectors with custom restriction sites. Custom sites selected were EcoRI and Notl in the 5’ and 3’ ends, respectively. These restriction sites were used for further cloning into a multi cloning site of PICZa-A expression vector (Figure 4B). The transcriptional units design consisted of, from 5’ to 3’ end: the methanol induced AOX promoter from X-33, the alpha-factor secretion signal from Saccharomyces cerevisiae. the codon-optimized sequence for the specific insulin variant, TGA stop codon, untranslated Myc and His-tags, and AOX terminator from X-33 (Figure 4A). Integration to genomic DNA is targeted to the AOX promoter region via digestion of the plasmid with Pmel restriction enzyme.

[0141] Cloning comprised the following steps, 1- double digestion with EcoRI / Notl of PUC57 vectors carrying the sequence for variant and of the PICZa-A acceptor vector, 2- ligation reaction with T4 ligase, 3- chemocompetent E. coli TOP10 transformation with selection of transformants in 25 ug / ml Zeocin LB agar, 4- colony PCR screening of correctly assembled vectors using primers PAOX and TAOX, 5- extraction and purification of the plasmids through commercially available Miniprep kits, and finally 6- sequence verification through Capillary Electrophoresis Sequencing.Table 8. Nucleic acid optimized sequences for the expression of bM2, bM7, bMIO, bM25, and bM2-664 Insulin variants in Pichia pastoris.

[0142]

[0143] Table 9. Nucleic acid sequence of the transcriptional unit used for expression of the insulin variants. Italics sequence corresponds to the promoter region. Underlined italics highlight the Pmel restriction site. Underlined sequence corresponds to the CDS region. [*] denotes the region of the optimized codogenic sequence for each of the insulin variants. Plain text denotes the 3’ UTR region, including the terminator.

[0144]

[0145]

[0146] EXAMPLE 3: CONSTRUCTION OF bM2, bM7, bMIO, bM25, AND bM2-664 PRODUCING STRAINS.

[0147] Transformation and selection of Pichia pastoris

[0148] bM2, bM7, bMIO, bM25, and bM2-664 vectors were transformed into the parental wild-type strain Pichia pastoris X-33 by electroporation method (Lin-Cereghino et al., 2022). This protocol comprises initial growing for 24 h in YPD (Yeast extract 10 g / L, Peptone 20 g / L, Dextrose 20 g / L) media and sequential resuspension of the cells in sorbitol as a preparation for the following electroporation. Cells were selected in YPDS-Zeocin media (Yeast extract 10 g / L, Peptone 20 g / L, Dextrose 20 g / L, Sorbitol 182 g / L, Agar 20 g / L, Zeocin™ 100 mg / L). Selection of the transformants that correctly integrated the construct was performed by colony PCR of 30 colonies. These PCRs were carried out using primers that attach to flanking regions of the Insulin gene variants (to promoter and terminator sequences, SEQ ID NO: 24, GACTGGTTCCAATTGACAAGC and SEQ ID NO: 25, GCAAATGGCATTCTGACATCC). Up to thirty positive colonies per strain were cultivated in deep well plates for 72 h to test the production of each insulin variant.

[0149] Screening of insulin variants-producing strains in deep well plates

[0150] The screening of insulin-producing strains was carried out on 96 deep well plates. Cultures consisted of 3 defined stages: Pre-culture, growth phase, and induction phase (24 h, 24 h, and 48 h, respectively). In pre-culture stage, colonies were stabbed from an YPD agar plate to inoculate in 4 ml YPG media culture tubes at 30°C and 1000 rpm for 24 h. After the pre-culture stage, cells were inoculated at an initial ODeoo of 2 into a new 96 deep well plates containing 600 pl of YPG media (Yeast extract 10 g / L, Peptone 20 g / L, Glycerol 20 g / L). Inoculated plate was cultured at 1000 rpm and 30°C. To initiate induction phase, YPG culture media was replaced for BMMY (Yeast extract 10g / L, Peptone 20 g / L, YNB 13.4 g / L, Biotin 4 mg / L, Monopotassium phosphate 5.7 g / L, Dipotassium phosphate 10 g / L, Methanol 2% v / v) induction media after centrifugation and resuspension of grown biomass. Culture conditions were kept as described above. To detect insulin production, supernatant was analyzed by SDS-PAGE and western blot assays.

[0151] For the SDS-PAGE (18% acrylamide), 30 pl of supernatant was loaded into the gels. In the case of bM2, bM7, bMIO, and bM25, SDS-PAGE gels showed no evident band at the expected size. Due to the low expression, these variants were detected only by western blot. Western blots were performed using mouse monoclonal anti-insulin B-chain (Santa Cruz Biotechnology) as the primary antibody. HRP-conjugated anti-mouse secondary antibody (Santa Cruz Biotechnology) was used for chemiluminescent detection of the bands. Among all variants, bM2 showed the best expression levels and colony-to-colony consistency. On the other hand, only a few colonies of bM7, bMIO, and bM25 showed detectable insulin secretion (Figure 5A).

[0152] In the case of the optimized version of bM2, called bM2-664, its expression was clearly noticeable by SDS-PAGE, indicating higher levels of protein secretion than the other variants (Figure 5B).

[0153] EXAMPLE 4: PRODUCTION OF bM2 AND bM2-664 VARIANTS IN 10 L BIOREACTOR

[0154] The pre-inoculum cultures were prepared by transferring 1 mL of glycerol stock of bM2 and bM2-664 strain to 20 mL of YPG medium (containing per liter: Glycerol 20 g, Peptone 20 g, Yeast extract 10 g) in a 100 mL flask and growing for 7 h at 30°C and 200 rpm. The inoculums were prepared in 1 L baffled flask at 120 rpm and 30°C with 200 mL of YPG medium.

[0155] Fed-batch fermentations were carried out in 10 L Applikon bioreactors. The fermentations begin with an initial ODeoo=2 and 7.5 L of basal salt medium (BSM) (containing per liter: Glycerol 40 g, Phosphoric acid 13.5 mL, Citric acid-EEO 4.37 g, Calcium chloride-2H2O 0.5 g, Sodium chloride 0.25 g, Magnesium sulfate-7H2O 7.5 g, Potassium hydroxide 2 g, Potassium sulfate 9 g, Antifoam 0.1 mL, PTM1 4.35 mL). PTM1 is the trace elements solution (containing per liter: Sulfuric acid 5 mL, Copper sulfate-SFFO 6 g, Potassium iodide 0.09 g, Manganese sulfate-FFO 3 g, Sodium molybdate-2H2O 0.2, Boric acid 0.02 g, Cobalt chloride-OEEO 0.92 g, Zinc sulfate-7H2O 42.2 g, Iron sulfate-7H2O 65 g, Biotin 0.2 g).

[0156] The fermentations consisted of the following phases: 15 h of batch phase, 5 h of fed-batch (constant flow of 2.7 mL / min of glycerol), 2 h of glycerol - methanol adaptation (constant flow of 0.5 mL / min of glycerol and exponential feeding of methanol), and 46-47 h with exponential methanol induction. The parameters set in the induction phases were Xo: 260 g / L, Yxs: 1.2 gWCW / g-methanol, Vo: 8.3L, p: 0.01 1 / h Sf: 792 g / L. My-Control system was used to maintain the cultures at 30°C, pH 5.0 (using 28% ammonium hydroxide), and >30% DO2 in the growth phase. The gas flow was fixed at 2 VVM and the stirrer speed varied between 600 - 1000 rpm. In the induction phase, a temperature of 25°C, pH 6.0 and >30% DO2 were set. Culture samples were collected for quantification of biomass, total protein, bM2 and bM2-664 insulin concentrations.

[0157] For the bM2 culture, the maximum specific growth rate pmax was 0.18 h'1for glycerol consumption and the specific growth rate set in the induction phase was 0.01 h’1. The final biomass concentration reached was 310 gWCW / L. The methanol yield on biomass in the induction phase was 1.1 gWCW / g-methanol. After 69 hours of fermentation, the total protein content in the supernatant was 820 mg / L, of which 70% (574 mg / L) was the recombinant protein bM2 (Figure 6 A, B).

[0158] In the case of bM2-664 culture the pmax was 0.2 h'1for glycerol consumption and the specific growth rate set in the induction phase was 0.01 h'1The final biomass concentration reached was 330 gWCW / L. The methanol yield on biomass in the induction phase was 0.9 gWCW / g-methanol. After 68 hours of fermentation, the total protein content in the supernatant was 2300 mg / L, of which 70% (1610 mg / L) was the recombinant protein bM2-664 (Figure C, D).

[0159] EXAMPLE 5: PURIFICATION OF bM2 AND bM2-664 VARIANTS FROM 10 L BIOREACTOR BY ZINC PRECIPITATION:

[0160] A scheme with the downstream process (DSP) is shown in Figure 7. The supernatant obtained after the centrifugation and microfiltration operations was precipitated using zinc chloride. The conditions of the precipitation were: pH 4.6 (below isoelectric point of 5.3), 1.25 g / L of phenol, and 0.46 g / L of zinc chloride (Hazra et al., 2021, 2023). The supernatant was added with phenol and zinc chloride and the pH was adjusted at 4.6 utilizing HC14 M. It was left in agitation for 30 min at 4°C, and then overnight without movement at 4°C. The precipitated solution was centrifuged at 15.000 g for 30 min. The bM2 and bM2-664 insulin crystals were resuspended with a buffer Tris-HCl 20 mM pH 7.5 + 2% Trehalose solution until it achieved a concentration of 20 g / L of each insulin. The pH was adjusted at 7.0 utilizing NaOH 2.5 M. The solution was filtered with a syringe filter with a PES membrane of 0.22 pm. Finally, the insulins were lyophilized to obtain dry powders. The results obtained after the precipitation process are shown in Table 10.

[0161] Table 10. Purity and recovery obtained in the resuspended crystal of bM2 and bM2-664 after the precipitation process.

[0162]

[0163]

[0164] EXAMPLE 6: SOLUBILITY ASSAYS OF bM2 and bM2-664 VARIANTS

[0165] For the solubility assays the lyophilized bM2 and bM2-664 insulins were solubilized in PBS Buffer pH 7.4 at four different expected concentrations: 10, 50, 100 y 150 g / L. After the solubilization, the solution was centrifuged at 13.000 g for 5 min to eliminate undissolved protein. The supernatants were analyzed by Bradford assay to quantify the soluble protein. The Soluble insulin / Total insulin (S / T) ratio should be 1 if the protein is fully soluble at the expected concentrations and should decrease if the maximum soluble concentration is reached and the solution is saturated. For reference, the solubility of human native insulin dissolved in buffer Tris-HCl pH 7.4 is approximately 20 g / L (Glidden et al., 2018).

[0166] The solubility results of bM2 and bM2-664 are detailed in Table 11. In the case of bM2, at 150 g / L of total insulin, the maximum soluble concentration reached was 106 g / L with an S / T ratio of 0.71, indicating saturation of the solution. The bM2-664 variant showed even better solubility, reaching a maximum soluble concentration of 147 g / L with an S / T ratio of 0.98, suggesting that the solution is not yet saturated. These results validate the bioinformatic predictions of the increased solubility of these variants and the observed differences between bM2 and bM2-664 (Table 4).

[0167] Table 11. Concentrations obtained in solubility assays of bM2 and bM2-664

[0168]

[0169]

[0170] EXAMPLE 7: BIOACTIVITY ASSAYS OF bM2 and bM2-664 VARIANTS

[0171] To test the bioactivity of the purified bM2 and bM2-664 variants, C2C12 cells (CRL-1772, ATCC), an immortalized mouse myoblast cell line was used as an experimental model. The cells were seeded at 3000 cells per well in a Matrigel-coated 96-well culture plate. Cells were cultivated for 24h at 37°C %, 5% CO2 in DMEM medium 2% fetal bovine serum (FBS). After 24 h, the medium was replaced for the following serum-free medium: DMEM / F12 basal medium, L-ascorbic acid-2-phosphate magnesium (64 mg / L), Sodium selenium (14 pg / L), albumin (1 g / L), insulin (20 mg / L), transferrin (10 mg / L), FGF-2 (100 pg / L), and TGF- 1 (2 pg / L). Medium formulated with human native insulin was used as positive control, while medium without insulin was used as negative control. The lyophilized bM2 and b2-664 proteins were resuspended in PBS Buffer pH 7.4 to a concentration of 2 g / L and used as stock solution for the formulation of experimental media at different final concentrations. The cells were cultivated for 72 h at 37°C %, 5% CO2. After treatment time, cells nuclei were stained with Hoechst and automatically counted using a cytation imaging reader.

[0172] Human native insulin showed a dose-dependent effect on the proliferation of the cells, reaching their maximum effect at 1 mg / L and promoting a 67% increase in cell number (Figure 8A). On the other hand, bM2 achieves its highest effect at 10 mg / L, promoting an 80% increase in cell number. At 1 mg / L bM2 showed a slightly lower effect (58%) compared to human insulin (67%). Both insulins, native human and bM2 showed a decreased effect at concentrations below 0.1 mg / L. These results confirmed that bM2 is a bioactive protein, able to stimulate cell proliferation in animal cells.

[0173] The bM2-664 variant exhibited a dose-dependent proliferative effect more similar to native insulin than the bM2 variant. At a concentration of 1 mg / L both native insulin and bM2-664 promoted a 40% increase in cell number (Figure 8B). The fold-change variation in the effect of native insulin between the two assays can be attributed to inherent variability in the cell lines between passages.

[0174] EXAMPLE 8: THERMOSTABILITY ASSAYS OF bM2 and bM2-664 VARIANTSFor the thermostability assays, resuspended human native insulin, bM2, and bM2-664 samples were incubated at two different temperatures, 45°C and 60°C, for one week. After the incubations, the proteins were used for the bioactivity assays following the same protocol described in example 7 but using a fixed concentration of 1 mg / L of each insulin. Human native insulin, bM2, and bM2-664 stored at -20°C, were used as positive controls, while media without insulin served as the negative control.

[0175] Results are illustrated in Figure 9. Both human native insulin and bM2 stored at -20°C showed a significant effect on cell number (48% and 34% increase, respectively). After incubation at 45°C, human native insulin decreases its activity to 35%, while incubation at 60°C results in complete inactivation. The basal effect of bM2 (sample stored at -20°C) is slightly lower than the native insulin at the assay concentration (1 mg / L), which is consistent with what was observed in the bioactivity assay (Example 7). However, unlikely native insulin, the bM2 variant was shown to be thermostable at 45°C and 60°C without any significant loss of activity (Figure 9A).

[0176] For the bM2-664 variant, an 11% decreased activity was observed after incubating the protein for one week at 60°C, compared to the sample stored at -20°C. However, bM2-664 incubated at 60°C does not perform significantly differently from native insulin stored at -20°C. As expected, native insulin was fully inactivated at 60°C (Figure 9B).

[0177] These results demonstrate that the bM2 and bM2-664 variants have enhanced thermostability compared to human native insulin.7. BIBLIOGRAPHY

[0178] BuB, O., Rudat, J., & Ochsenreither, K. (2018). FoldX as Protein Engineering Tool: Better Than Random Based Approaches? Computational and Structural Biotechnology Journal, 16, 25—33. https: / / doi.Org / 10.1016 / j.csbj.2018.01.002

[0179] Cao, L., Coventry, B., Goreshnik, I., Huang, B., Sheffler, W., Park, J. S., Jude, K. M., Markovic, I., Kadam, R. U., Verschueren, K. H. G., Verstraete, K., Walsh, S. T. R., Bennett, N., Phal, A., Yang, A., Kozodoy, L., DeWitt, M., Picton, L., Miller, L., ... Baker, D. (2022). Design of protein-binding proteins from the target structure alone. Nature, 605(7910), 551—560. https: / / doi.org / 10.1038 / s41586-022-04654-9

[0180] Chen, R., Li, L., & Weng, Z. (2003). ZDOCK: An initial-stage protein-docking algorithm. Proteins:

[0181] Structure, Function, and Bioinformatics, 52(1), 80-87. https: / / doi.org / 10.1002 / prot.10389 Dauparas, J., Anishchenko, I., Bennett, N., Bai, H., Ragotte, R. J., Milles, L. F., Wicky, B. I. M., Courbet, A., de Haas, R. J., Bethel, N., Leung, P. J. Y., Huddy, T. F., Pellock, S., Tischer, D., Chan, F., Koepnick, B., Nguyen, H., Kang, A., Sankaran, B., ... Baker, D. (2022). Robust deep learning-based protein sequence design using ProteinMPNN. Science, 375(6615), 49—56. http s : / / doi . org / 10.1126 / science.add2187

[0182] Evans, R., O’Neill, M., Pritzel, A., Antropova, N., Senior, A., Green, T., ZIdek, A., Bates, R., Blackwell, S., Yim, J., Ronneberger, O., Bodenstein, S., Zielinski, M., Bridgland, A., Potapenko, A., Cowie, A., Tunyasuvunakool, K., Jain, R., Clancy, E., ... Hassabis, D. (2022). Protein complex prediction with AlphaFold-Multimer (p. 2021.10.04.463034). bioRxiv. https: / / doi.Org / 10.1101 / 2021.10.04.463034

[0183] Fleishman, S. J., Leaver-Fay, A., Com, J. E., Strauch, E.-M., Khare, S. D., Koga, N., Ashworth, J., Murphy, P., Richter, F., Lemmon, G., Meiler, J., & Baker, D. (2011). RosettaScripts: A Scripting Language Interface to the Rosetta Macromolecular Modeling Suite. PLOS ONE, 6(6), e20161. https: / / doi.org / 10.1371 / journal.pone.0020161

[0184] Glidden, M. D., Aldabbagh, K., Phillips, N. B., Carr, K., Chen, Y.-S., Whittaker, J., Phillips, M., Wickramasinghe, N. P., Rege, N., Swain, M., Peng, Y., Yang, Y., Lawrence, M. C., Yee, V. C., Ismail-Beigi, F., & Weiss, M. A. (2018). An ultra-stable single-chain insulin analog resists thermal inactivation and exhibits biological signaling duration equivalent to the native protein. Journal of Biological Chemistry, 293(1), 47—68. https: / / doi.org / 10.1074 / jbc.M117.808626Hazra, P., Buddha, M., Reddy, C., & Gupta, I. (2023). Large-scale crystallization as an intermediate processing step in insulin downstream process: Explored advantages and identified tool for process intensification. Bioprocess and Biosystems Engineering, 6(12), 1765—1776. https: / / doi.org / 10.1007 / s00449-023-02931-z

[0185] Hazra, P., Sreenivas, S., Venkatesan, K., Patale, M. B., Chatterjee, A., Ramprabu, N., Shaikh, A. M., & Kusumanchi, M. (2021). A novel peptide design aids in the expression and its simplified process of manufacturing of Insulin Glargine in Pichia pastoris. Applied Microbiology and Biotechnology, 105(8), 3061-3074. https: / / doi.org / 10.1007 / s00253-021-11224-y

[0186] Lin-Cereghino, J., Naranjo, C. A., & Lin-Cereghino, G. P. (2022). Competent Cell Preparation and Transformation of Pichia pastoris. In V. Mapelli & M. Bettiga (Eds.), Yeast Metabolic Engineering: Methods and Protocols (pp. 113-120). Springer US. https: / / doi.org / 10.1007 / 978- l-0716-2399-2_7

[0187] Le, T. K. C., Dao, X. D., Nguyen, D. V., Luu, D. H, Bui, T. M. H., Le, T. H., Nguyen, H. T., Le, T.

[0188] N., Hosaka, T., & Nguyen, T. T. T. (2023). Insulin signaling and its application. Frontiers in Endocrinology, 14. https: / / doi.org / 10.3389 / fendo.2023.1226655

[0189] Mao, R., Chen, Y., Chi, Z., & Wang, Y. (2019). Insulin and its single-chain analogue. Applied Microbiology and Biotechnology, 103(21), 8737—8751. https: / / doi.org / 10.1007 / s00253-019- 10170-0

[0190] Mirdita, M., Schutze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., & Steinegger, M. (2022).

[0191] ColabFold: Making protein folding accessible to all. Nature Methods, 19(6), 679-682. https: / / doi.org / 10.1038 / s41592-022-01488-l

[0192] Norton, L., Shannon, C., Gastaldelli, A., & DeFronzo, R. A. (2022). Insulin: The master regulator of glucose metabolism. Metabolism - Clinical and Experimental, 129. https: / / doi.Org / 10.1016 / j.metabol.2022.155142

[0193] Oeller, M., Kang, R., Bell, R., Ausserwbger, H., Sormanni, P., & Vendruscolo, M. (2023). Sequencebased prediction of pH-dependent protein solubility using CamSol. Briefings in Bioinformatics, 24(2), bbad004. https: / / doi.org / 10.1093 / bib / bbad004

[0194] Pei, J., & Grishin, N. V. (2001). AL2CO: Calculation of positional conservation in a protein sequence alignment. Bioinformatics, 17(8), 700-712. https: / / doi.Org / 10.1093 / bioinformatics / 17.8.700 Voynov, V., Chennamsetty, N., Kayser, V., Helk, B., & Trout, B. L. (2009). Predictive tools for stabilization of therapeutic proteins. mAbs, 1(6), 580-582. https: / / doi.org / 10.4161 / mabs.L6.9773Wang, C., & Zou, Q. (2023). Prediction of protein solubility based on sequence physicochemical patterns and distributed representation information with DeepSoluE. BMC Biology, 21( \ ), 12. https: / / doi.org / 10.1186 / sl2915-023-01510-8

Claims

CLAIMS1. Single-chain insulin variants from animal origin, wherein A-chain and B-chain are linked by a C-peptide of 6 amino acid residues designed as C1C2C3C3C4C5C6, wherein:• Ci: Arginine (R).• C2: Glutamic acid (E) or aspartic acid (D).• C3: Valine (V) or alanine (A).• C4: Glutamic acid (E).• C5: Glycine (G) or glutamine (Q) or asparagine (N).• Ce: Proline (P).Wherein the animal origin includes the A-chain and B-chain sequences of insulin from the following species: Bos Iannis. Sus scrofct. Ovis aries, Capra hircus, Bubalus bubalis, Oryctolagus cuniculus, Gallus gallus, Meleagris gallopavo, Anas platyrhynchos, Anser anser, among other species commonly used in meat and dairy industries.

2. The insulin variant of claim 1, wherein the B29 position is substituted by an acidic residue such as E or D.

3. The insulin variants of claim 1 or 2, wherein the B30 position is substituted by P or E.

4. The insulin variants of any of claims 1 to 3, wherein the A8 position is substituted by H or E or S.

5. The insulin variants of any of claims 1 to 4, wherein the A14 position is substituted by E or K or Q6. The insulin variants of any of claims 1 to 5, wherein the A9 position is substituted by T.

7. The insulin variants of any of claims 1 to 6, wherein the A10 position is substituted by V or I or T.

8. The insulin variants of any of claims 1-7, wherein the insulin variants comprise the amino acid sequence selected from the group consisting of SEQ ID NOs: 4-17.

9. A nucleic acid wherein said nucleic acid encodes the insulin variants of any of claims 1 to 7.

10. An expression vector wherein said expression vector contains the nucleic acid of claim 9.

11. Method for producing the insulin variants of any of claims 1 to 7, wherein it comprises the steps of: (a) introducing the expression vector of claim 10 into a cell capable of expressing the insulin variant; (b) culturing said recombinant cell; and (c) recovering the insulin variant from the supernatant.

12. Use of insulin variants of claims 1 to 7 in cell culture media for growing animal cells, preferably in cultivated meat, biopharmaceutical manufacturing, and research use.

Images