Yu Mi,Yuan Gao ,Daidi Fan *,Zhiguang Duan Rongzhan Fu Lihua Liang Wenjiao Xue ,Shanshan Wang

1 Shaanxi R&D Center of Biomaterials and Fermentation Engineering,School of Chemical Engineering,Northwest University,Xi'an 710069,China

2 Shaanxi Institute of Microbiology,Xi'an 710043,China

3 School of Biological Science and Engineering,Shaanxi University of Technology,Hanzhong 723001,China

Keywords:Fusion expression Human collagen α1(I)chain Insulin Protein stability Pichia pastoris

A B S T R A C T To enhance the stability of recombinant human collagen α1(I)chains(rhCOL1A1)in production and purification stages,a gene fragment fusing COL1A1 and insulin protein coding domains was synthesized and inserted into the pPIC9K expression vector.The fusion peptide-expressing Pichia pastoris strain was created by transformation.After optimization of shake flask cultures,the ultimate intracellular expression level of the insulin-collagen α1(I)chain fusion protein(INS-COL1A1)reached about 300 mg·L-1,and no obvious protein degradation was found in the fermentation and purification processes.The His-tagged recombinant fusion protein was detected by western blotting and was effectively purified using Ni2+-chelating chromatography.A prominent improvement in the stability of INS-COL1A1 was observed compared to rhCOL1A1 in vitro,and the rhCOL1A1 released from the fusion protein was studied by LC–MS/MS and in bioassays.The results showed that the purified rhCOL1A1 was consistent with the native protein in amino acid composition and had a similar biological compatibility.To our knowledge,this is the first study to demonstrate the use of insulin as a fusion protein to improve the stability of easily degradable proteins.

1.Introduction

Collagen is a major structural element of virtually all connective and interstitial tissues in parenchymal organs.It simultaneously contributes to the stability of those tissues and organs and maintains their structural integrity.Due to its favorable binding capacity,collagen could serve as a network-forming ingredient and as a component of a delivery system for drugs,growth factors,or cells[1,2].However,most collagen products applied at present are derived from animal tissues.These commodities,isolated from tissues,exhibit significant lot-to-lot variability and may be contaminated with potential immunogenic and infective agents considered pernicious to human health.For many years,substantial research and work have been committed to creating an efficient means of producing collagens.By the end of the last century,researchers have produced collagen using recombinant systems such as Escherichia coli and Pichia pastoris[3–5].In addition,some collagen-like proteins were designed and expressed to connect with other polymers for engineering biomaterials[6].

Because of the vulnerable structure and low productivity of full length collagen chains[5],almost all of these previous studies focused on the expression of triple-helical collagen segments or collagen single chain fragments.These segmental or artificial proteins do not possess all the active sites nor the potential bioactivity of full-length natural collagen.The expression of full-length collagen single chains in previous studies was accompanied by an obvious degradation phenomenon[7,8].The authors declared that the degradation of rhCOL3A1 begins from the N-terminal portion of the protein.Considering the large molecular weight and linear structure of collagen,preventing degradation is the main priority in expression of full length collagen chains[5,9,10].Although protein stability has been studied[11,12],there is still no effective solution regarding this problem.

The natural insulin molecule is composed of two polypeptides,an A chain and a B chain,which are joined by two pairs of disulfide bonds with an additional intramolecular disulfide bond in the A chain[13–19].Recently,several reports suggest the use of protein fusion expression to improve production or facilitate purification of desired proteins[20–22].In this study,a fusion protein composed of the human collagen α1(I)chain and insulin was designed and expressed in P.pastoris GS115 to allow for the straightforward manipulation,post-translational modification,and high expression level of recombinant protein.In the recombinant plasmid used in this study,two chains of insulin were located at either end of collagen.We reasoned that insulin could form a closed loop structure with the collagen chain to protect the terminal ends of collagen.Moreover,a His-tag was present at the end of the expression cassette to facilitate subsequent purification.The stabilities of the fusion protein and rhCOL1A1 were subsequently studied.

2.Materials and Methods

2.1.Strains,plasmids and reagents

2.2.Construction of the expression vector

The human insulin cDNA(GenBank:JF909299.1)was synthesized and inserted into the pPICZαA vector by Sangon Biotech(Shanghai,China)to create vector pPICZ-INS.Sequences encoding the TEV enzyme(tobacco etch virus protease)cleavage site were added to the 3′end of the insulin B chain and the 5′end of the A chain respectively.Moreover,a His-tag sequence was added at the 3′end of the A chain.The vector pPIC9K-COL1A1 containing the collagen α1(I)gene was constructed previously in our laboratory.Sequence verification of the collagen gene in vector pPIC9K-COL1A1 was consistent with standard data(NCBI reference sequence:NM_000088.3).

To construct the expression vector,the insulin B chain was amplified from pPICZ-INS with the 3′end TEV cleavage site coding sequence,and an Eco RI restriction site was introduced at the 5′end.The insulin Achain with the TEV cleavage site at the 5′end was amplified from the same vector and included a His-tag at the 3′end,followed by a Not Irestriction site.Afterwards,the gene coding the collagenα1(I)chain was amplified from pPIC9K-COL1A1 using PCR,introducing consensus sequences at both ends of the two amplified fragments above.As shown in Fig.1,the three DNA fragments were fused by overlapping PCR and ligated into the pPIC9K vector to form the recombinant expression vector pPIC9K-F.All primers used in this study are shown in Table 1.The completed vector was transformed into competent E.coli DH5αcells for verification and amplification.Recombinant colonies were selected on LB plates by scoring for kanamycin resistance.Subsequently,colony PCR was performed to select for correct recombinants.Then,plasmids collected from the selected colonies were verified by restriction endonuclease digestion and sequencing.

2.3.P.pastoris transformation

After sequence identification,verified recombinant plasmids were linearized with Sal I and transformed into competent P.pastoris GS115 cells,which were prepared according to standard procedures.Each 10 μg linearized recombinant vector was mixed with 100 μl of competent cells,and the mixture was immediately transferred into a pre-chilled 2 mm electroporation cuvette and incubated on ice for 5 min before transformation.An Electro Cell Manipulator 630(Harvard Apparatus,USA)was used for electroporation.Charging voltage,capacitance,and resistance during electroporation were 1.5 kV,25 μF,and 200 Ω,respectively.Transformants were selected on YPD plates containing 0.5 and 4.0 mg·ml-1G418 at 28 °C for 3 days.Colonies grown on G418-resistance medium were selected for genomic PCR verification.

Fig.1.The construction diagram of expression plasmid.

Table 1List of primers used in polymerase chain reaction

2.4.Shake flask culture

Screened transformants of different G418 resistances were cultivated in shake flasks to produce the fusion protein.Individual colonies were selected and grown in 5 ml of YPD(1%yeast extract,2%peptone,2%glucose)in an incubator at 28 °C and shaken at 220 r·min-1for approximately 24 h.The cells were then harvested by centrifugation at 1500×g for 10 min at 4°C and inoculated into 25 ml of BMGY(1%yeast extract,2%peptone,100 mmol·L-1potassium phosphate,pH 6.0,1.34%yeast nitrogen base,4×10-7biotin,1%glycerol)and cultured at 28 °C and shaken at 220 r·min-1until an OD600(Optical Density)of 2–8.Subsequently,the cells were collected by centrifugation at 1500×g for 10 min at 4°C.Cell pellets were resuspended in 200 ml of BMMY(1%yeast extract,2%peptone,100 mmol·L-1potassium phosphate,pH 6.0,1.34%yeast nitrogen base,4×10-7biotin,1%methanol)at an initial OD600of 1.0 for inducible expression.Methanol was supplemented every 24 h to maintain a final concentration of 1 vol%for induction.

2.5.Protein purification and enzyme cleavage

After induction,the entire medium was collected and centrifuged at 8000×g for 20 min.Subsequently,intracellular proteins were obtained using the One Step Yeast Active Protein Extraction kit(Sangon Biotech,China)and concentrated by ultrafiltration.Then the His-tagged INSCOL1A1 was purified using Ni2+-chelating chromatography with an AKTA purifier(GE Healthcare,Sweden)that had been equilibrated with binding buffer containing 20 mmol·L-1imidazole.After washing to baseline absorbance with the same buffer,the bound fraction was eluted with 200 mmol·L-1imidazole at a flow rate of 2 ml·min-1.The purified fusion protein was pooled,dialyzed,and then lyophilized.

To separate the collagen chain from insulin,the lyophilized fusion protein was resuspended in cleavage buffer to a final concentration of 5 mg·ml-1.TEV protease(Solarbio,Beijing,China)was subsequently added to a final concentration of 1 U of enzyme per 50 μg of fusion protein.Optimized pH,reaction time,and other conditions were verified by repeated experiments.After incubation for 2 h at 30°C and pH 8.0,the reaction mixture was treated by ultrafiltration through a 50 kDa membrane(Millipore,USA)to remove insulin and the TEV enzyme.Finally,the purity,identity,and concentration of rhCOL1A1 were checked by SDS-PAGE and measured using Quantity One software with the protein marker as a standard.

2.6.SDS-PAGE and western blotting

SDS-PAGE analysis was performed after protein harvest.The concentration of different proteins in the gel was measured by Quantity One software by comparing to the standard protein marker with known concentration combined with the Biuret method.Then,proteins in the separation gel were transferred to PVDF membranes(Millipore,USA).After blocking overnight at 4 °C with TBS(10 mmol·L-1Tris–HCl,150 mmol·L-1NaCl,pH 7.5)containing 1%bovine serum albumin,the membrane was incubated with a mouse monoclonal antibody against the 6x-His-tag(Tiangen,Beijing,China)over night at 4°C.A secondary anti-mouse IgGconjugated to sheep radish peroxidase(Tiangen,Beijing,China)was incubated with the membrane for 1 h at 25°C at a 1:500 dilution in PBS.Immunoreactivity was detected using the HRPDAB development kit(Tiangen,Beijing,China).

2.7.Stability research of the fusion protein and collagen

The stability of the recombinant fusion protein INS-COL1A1 was compared to the rhCOL1A1 single chain by examining several aspects below.The stability of the two proteins under different pH was carried out at 4°C for 24 h.Thermal stability was measured by incubating the salt-less protein solution at 4 °C,25 °C and 30 °C for 12,24,48,and 72 h at the optimal pH.For resistance to enzymatic hydrolysis,the two proteins were incubated in appropriate buffers with 1 U of trypsin(Thermo)per 50 μg of purified protein for 0.5,1,2,or 4 h at pH 6.0 and 25°C.Data concerning protein degradation rates were detected by SDS-PAGE and densitometry analysis.

2.8.LC–MS/MS analysis of rhCOL1A1

Liquid chromatography coupled with tandem mass spectrometry(LC–MS/MS)analysis was performed for identification of the expression product.The rhCOL1A1 was separated by SDS-PAGE as shown previously.Target protein bands were cut from SDS-PAGE gels for LC–MS/MS analysis performed by the Beijing HuaDa Protein Research and Development Center(China).The result was compared with standard data acquired from UniProt(|P02452|CO1A1_HUMAN).

So during my parenting years, I joined in the fun, on occasion yelled Boo! , dressed my children in fun costumes and took them around to neighbors for safe trick-or-treating

2.9.Bioassay of rhCOL1A1

Abioassay was performed on L929 cells(mouse fibroblast cells)to investigate whether rhCOL1A1 obtained from the engineered strain has a stimulatory effect on cell proliferation.L929 cells were grown in RPMI-1640 medium,at a density of 104cells/well in 96-well plates,at 37°C and 5%CO2in a humidified incubator.After 24 h of incubation,the medium of each well was replaced with prepared samples(medium with rhCOL1A1),positive control(medium with commercial COL1A1)and fresh medium as a negative control.After incubation for 72 h,the media was replaced with 50 μl of MTT(5 mg·ml-1)and 100 μl of fresh culture medium and incubated for another 4 h.Finally,100 μl of DMSO was added as solvent.The OD of each culture was measured at 490 nm using a microplate reader(Bio-Rad,USA).The relative cell growth rate was calculated with the formula below:

All experiments were performed in triplicate and data were analyzed using SAS(release 8.1,SAS Institute,Cart,NC).The value of p＜0.05 was considered to be statistically significant.

3.Results

3.1.Construction and transformation of expression vector

In this study,the recombinant vector pPIC9K-F was constructed to enhance the stability of the full-length collagen single chain.The human insulin coding sequence was obtained from the NCBI database(GenBank:JF909299.1)and synthesized by Sangon Biotech.The gene encoding the collagen α1(I)chain was amplified from vector pPIC9KCOL1A1,which was previously generated by our laboratory.Afterwards,these fragments were fused via overlapping PCR and ligated into pPIC9K to construct the recombinant plasmid pPIC9K-F.Two sections of the insulin gene were linked,respectively,to either side of the collagen gene,with TEV enzyme cleavage sites in between.It has previously been reported that degradation of rhCOL3A1 begins at the N-terminal portion of the protein[8].In this study,it was reasoned that the presence of insulin would allow collagen to form a circular structure so that both ends of the collagen single chain would be protected.In addition,a His-tag at the end of the fusion protein was added to simplify post-purification.Colony PCR and sequencing results showed that the sequence of the constructed plasmid was completely consistent with our design.Afterwards,the recombinant vector was linearized and successfully transformed into P.pastoris GS115 cells.Strains with high resistance in YPD plates containing 0.5 or 4.0 mg·ml-1G418 were selected for further examination.

3.2.Protein expression

We selected pPIC9K as the expression vector on account of its alpha signal peptide from Saccharomyces cerevisiae,which could secrete exogenous proteins into the extracellular culture medium and facilitate subsequent purification.However,as shown in Fig.2A,the electrophoresis result showed that no protein was found in the fermentation broth.We tentatively speculated that the fusion protein was expressed as expected but could not be secreted into the culture medium because of the formation of a cyclic structure induced by insulin.Therefore,we next examined the intracellular level of protein expression in these strains.

After cell lysis and centrifugation,SDS-PAGE analysis of intracellular soluble proteins was performed.As shown in Fig.2B,the protein expression levels in strains with different G418 resistance were distinctly different.Protein production in strains with a 4.0 mg·ml-1G418 resistance was apparently higher than strains with low G418 resistance.The molecular weight of the recombinant protein expressed by the screened strain was about 140 kDa,approximately,1.4 times that of the theoretical molecular weight(100 kDa).This phenomenon is consistent with the previous reports[7,8].It appeared because the migration of collagen in gels differs from that of globular proteins[23].Subsequently,the high resistant strain was selected to proceed with optimization of fusion protein expression.Shake flask cultures were carried out under various conditions;namely,temperatures between 28 and 32°C,varying pH,and methanol concentrations ranging from 0.1 to 2 vol%.Optimal production of fusion protein was obtained at 28°C,pH 5.0 and with 1 vol%methanol.As shown in Fig.3A,expression levels reached their highest value after 36 h of induction.The approximate yields were measured using Quantity One software,comparing to a standard marker.Consequently,the yield of fusion protein was determined to be about300 mg·L-1.In addition,no degradation of the fusion protein was detected in this process.

Fig.3.Fusion protein expressed by P.pastoris.(A)Intracellular protein of engineering strain after 12,24,36,and 48 h induction;(B)curves of protein expression level.

3.3.Purification and cleavage of fusion protein

Fig.2.SDS-PAGE analysis of recombinant strains.(A)M,protein standard marker;lanes 1–4,extracellular protein expressed by positive control(engineered GS115 possesses pPIC9KCOL1A1),negative control(blank GS115),recombinant strain with low or high G418 resistance after 48 h induction;(B)M,protein standard marker;lanes 1–3,intracellular protein expressed by negative control,strain with low or high G418 resistance after 48 h induction.

The His-tag at the end of INS-COL1A1 offered a valid one-step purification protocol using Ni2+-chelating chromatography.After cell lysis,concentration,and ultrafiltration,the protein solution was loaded onto an affinity chromatography column.Undesired proteins were effectively removed by washing with binding buffer,and most of the fusion protein was collected following elution.The ultimate purity of INSCOL1A1 achieved was at least 90%.To obtain rhCOL1A1,purified INSCOL1A1 was treated with TEV protease.Digestion was completed after a 2 h incubation at pH 8.0 and a temperature of 30°C.Afterwards,rhCOL1A1 was purified to homogeneity by removing insulin and TEV protease using a 50 kDa ultrafiltration membrane.This protein gave a single band of about 125 kDa as visualized by SDS-PAGE in Fig.4A.This purification protocol was highly effective in obtaining abundant rhCOL1A1,and 231 mg of pure product was acquired from 1 L of fermentation broth.The yield and purity of INS-COL1A1 and rhCOL1A1 at several critical steps are summarized in Table 2.Total protein concentration was measured by the Biuret method using commercial type I collagen from bovine tendon(Solarbio,China)as a standard[7].The quantity of the protein of interest was calculated by quantifying each lane of the gel using Quantity One software.Finally,the purity of rhCOL1A1 was measured by densitometry test after SDS-PAGE analysis and was discovered to be almost 91%.

3.4.Stability of the fusion protein INS-COL1A1 in vitro

The proteins INS-COL1A1 and rhCOL1A1 were purified as described before.Thermal stability,resistance to different pH and enzymatic hydrolysis of these two proteins were determined.Previously,studies have shown that different salt concentrations in solution influence the stability of collagen[24].In order to detect their stability,lyophilized proteins were dissolved in deionized water and the following experiments were performed.After treatment,the ratio of the residual protein content to the initial protein content was used to reflect the anti-degradation properties of rhCOL1A1 and INSCOL1A1.Fig.5A shows that INS-COL1A1 has a better stability at different pH values,and the optimal pH is nearly 6.0.After incubation at 4 °C,25 °C,and 30 °C,there were degradation phenomena observed in both protein solutions.We found that the thermal stabilities revealed by INS-COL1A1 and rhCOL1A1 were distinctly different.As shown in Fig.5B,the INS-COL1A1 fusion protein degraded slowly at 25°C and 30 °C and was stable at 4 °C.More than 90%of the fusion protein remained after incubation at 4°C for 72 h.On the other hand,rhCOL1A1 was more unstable in either condition and nearly 70%of rhCOL1A1 was degraded after 72 h of incubation at 30°C.In testing for resistance to enzyme hydrolysis,both proteins showed more than 90%degradation after 4 h of treatment.The degradation rate of INS-COL1A1 was slightly slower than that of rhCOL1A1.These results suggest that the fusion protein is much more stable than rhCOL1A1,and no obvious degradation was found in the fermentation and purification processes.

Fig.4.SDS-PAGE and western blotting.(A)SDS-PAGE analysis.M represents protein marker;lanes 1 and 2 mean intracellular protein of blank GS115 and engineered strain;3 and 4,harvest and residual after affinity chromatography;5,fusion protein after cleavage;6,rhCOL1A1;7,TEV enzyme and insulin removed from fusion protein.(B)Western blotting result.

Table 2Purification of INS-COL1A1 and rhCOL1A1 from 1 L culture

Fig.5.Degradation curves of INS-COL1A1 and rhCOL1A1.Here INS represents fusion protein INS-COL1A1 and COL represents recombinant human COL1A1 for convenient.(A)Proteins were incubated at diverse pH;(B)proteins were incubated at 4 °C,25 °C,and 30 °C for different times;(C)proteins were treated with trypsin(1 U of trypsin per 50 μg of protein).

3.5.LC–MS/MS analysis

To identify the recombinant collagen chain,LC–MS/MS was applied for protein qualitative analysis.SDS-PAGE gel slices containing the target protein were cut and analyzed.As a control,the theoretical molecular weight of rhCOL1A1 is about 100 kDa,and the isoelectric point calculated by DNAMAN software is about 7.6.Compared with standard data acquired from UniProt(UniProtKB-P02452),sequence coverage of the recombinant product was approximate 85%.It was confirmed that the protein we obtained was human collagen α1(I).Moreover,Fig.6 shows that the molecular weight of rhCOL1A1 is about 105 kDa and has a pI of 7.7,similar to standard data.This result demonstrated that this research study successfully obtained the desired product.

Fig.6.LC–MS/MS analysis of rhCOL1A1.

3.6.Biological activity determination

To confirm that rhCOL1A1 cleaved from the fusion protein was active,its mitogenic capacity was investigated using L929 cells.As shown in Fig.7B,cells cultured with rhCOL1A1 behaved similarly to cell treated with a positive control(commercial COL1A1 acquired from cattle),resulting in analogous cell densities after 1 and 3 days of culture.After 5 days,the mitogenic activity of rhCOL1A1 was consistent with the positive control,leading to a parallel relative cell growth rate.Both the two samples promoted growth of cells throughout the whole process compared to the negative control(fresh medium with no additional protein).These results suggest that rhCOL1A1 produced in P.pastoris has good biocompatibility approaching that of commercial products.

4.Discussion

Recent research on protein stability has made great progress[11,12].However,although there are many studies addressing the recombinant expression of collagen,none have focused on full length collagen as it is unstable and difficult to maintain,research and utilize.Full length single chain collagen,which retains all the potential active sites of collagen,was explored in the present research.The initial intent of this experiment was to establish an effective strategy to improve the stability of single chain collagen.Yet,the innovation and significance of the current study exceeded our expectations.We found that this method could improve the stability of single chain collagen and prevent its degradation by fusion expression,simplifying post-purification processes.Meanwhile,this method allows for the possibility to purify other unstable proteins.The results showed that the fusion protein was stable during the expression and purification process.In addition,insulin was obtained during the purification process additionally.This approach may allow for production of two products at the same time,which require further investigation of insulin activity.In other words,this strategy has a great economic significance and is suitable for further development after detailed exploration.

Fig.7.Bioassay of rhCOL1A1.(A)Electronic microscope photographs of cell cultured with negative control,rhCOL1A1 and positive control for5 days were marked as 1,2 and 3.(B)Relative cell growth rate of L929 cells cultured with negative control,rhCOL1A1 and positive control after 1,3,and 5 days.

Reports indicate that the C-peptide is necessary for the formation of insulin,and there are many modifications[25].The natural form consists of 31 amino acids while the shortest one is three amino acids(AAK),but there is no provision for the longest[26].However,there are some requirements so that it does not react directly with the two chains of insulin and does not affect the expression and secretion of insulin.Here we placed the insulin Aand B chain on both ends of the collagen single chain so that collagen was presentas the C-peptide of insulin.Collagen does not react with insulin.The formation of insulin led collagen to form a closed loop and alleviated the degradation phenomenon.At the same time,TEV protease cleavage sites were added between insulin and the collagen chain to facilitate subsequent separation.Compared with previous methods used in laboratory to purify rhCOL3A1[7,8],this protocol is much simpler.Of note,the production of the fusion protein was relatively lower than production of rhCOL3A1 obtained previously[7,8].Therefore,it is necessary to further improve the yield for large-scale applications.Collagen is easy to express in P.pastoris and the yield is relatively high[3,7].There is a lot of room for improvement in the fermentation process.

According to our hypothesis,there would be an interaction between the fusion protein molecules.For example,the insulin A and B chains in different fusion protein molecules could bind resulting in the formation of fusion protein dimers or even multimers.However,there were no higher molecular weight proteins other than the target product found in this experiment.Only the fusion protein INS-COL1A1 was formed without the formation of protein polymer.The reason for this is not yet clear and we may find the polymerization phenomenon in future studies.In addition,the spatial structure and physiological function of the fusion protein should be further studied.

To our knowledge,this experiment is the first to use insulin as a fusion protein to improve the stability of easily degradable proteins,which is of a great significance and is valuable to further research.