Introduction

Silkworm has been proven to be one of the most efficient and popularly used eukaryotic expression tools [1, 2]. Recently the silkworm has become an ideal multicellular eukaryotic model system for basic research. Its larvae have a lot of advantages as a biofactory because: (1) they can be easily reared using mulberry leaves on large scale at much lower cost or artificial diet throughout the year; (2) their bodies are large and easy to manipulate; (3) they have a relatively short life cycle (approximately 7 weeks); (4) their genetics and biology have been well documented; and (5) the domesticated moths cannot fly and hence it is easy and safe for management. Therefore, it is very promising to use the silkworm as a vector for large-scale industrial mass production [3].

Glycoproteins are a major subclass of proteins distinguished by the presences of oligosaccharide side chains, or glycans, covalently linked to the polypeptide background. Glycoproteins have been implicated in a wide variety of important biochemical and biological functions, including protein stability, immune function, enzymatic function, cellular adhesion and others. Unfortunately, there is no therapeutic protein produced in insect system to date, due to the expressed glycoproteins are paucimannose-typed, named paucimannosidic N-glycans, rather than the complex, terminally sialylated N-glycans in mammalian cells [4–7].

In N-glycosylation pathway, the early steps are probably identical as both the insect and mammalian. Subsequently, however, these two pathways diverge. Early studies indicated that insect protein N-glycosylation pathway included the enzymes involved in N-glycan trimming, but not those involved in elongation in mammalian cells [8–14]. In mammalian cells, paucimannosidic N-glycans were further changed to terminally sialylated N-glycans. The sequential series of reactions catalyzed by various enzymes such as N-acetylglucosaminyltransferase II (Gn-TII), galactosyltransferases (Gal-Ts), and sialyltransferases (Sial-Ts), which can convert the hybrid structures to “complex” type N-glycans. N-acetylglucosaminyltransferase III (Gn-TIII) catalyzes a competing reaction that adds an N-acetylglucosamine to the core mannose residue in the hybrid precursor and initiates production of “bisected” complex N-glycans [15–17]. Comparing the insect expression system with mammalian one, we should emphasize that the major processed N-glycans produced by insects lack terminal sialic acid residues. This is important because sialic acids are negatively charged, peripheral sugars that govern many different properties of glycoproteins importantly including their clearance rates in the mammalian circulatory system, which would impact the clinical efficacy of insect-derived recombinant glycoprotein product.

In this paper, our aim is to transfer the human genes of Gn-TII, Gal TI, 2,6-Sial-T (ST6 GalII) and 2,3-Sial-T (ST3GalIII) to silkworm genome in cell lines through transgenesis to establish a transgenic Bm cell line of piggyBac transposon-derived targeting expression of humanized glycoproteins.

Materials and methods

Materials

HFL-I cell line was purchased from cell pool of CAS in Shanghai, China. Escherichia coli TG1, vectors pSL1180-A3 pSL1180 pXL-BacII and pXL-BacII-GFP were preserved in our Lab. pSilencer™ 4.1-CMV-neo vector was purchased from Applied Biosystems. The vector pigA3GFP-hIGF-ie-neo was gift from Suzhou University, China. The Premix Ex Taq ® Version 2.0 (loading dye mix), primeSTARTM HS DNA polymerase and pMD-18T were purchased from Takara Biotechnology (Dalian) Co., Ltd., china.

Cloning of the human N-glycosylation genes

The HFL-I cells were collected by centrifugating at 300×g, 5 min. Total RNA was isolated from the collected cells using RNAprep Pure kit (Tiangen, China) according to the user manual. 1 μg total RNA was used as template to synthesize cDNA using PrimeScript II 1st Strand cDNA Synthesis Kit (Takara). For the cloning of the human N-glycosylation genes and construction of the transformation vectors, following primers were set (Table 1).

Table 1 Primers used in this study
Full size table

PCR was performed on the resulting cDNAs. The reaction was carried out with Takara Ex Taq polymerase for 35 amplification cycles (95 °C/10 s, 68 °C/90 s), finally elongation at 72 °C, 10 min. PCR product was examined by electrophoresis in 1 % agarose gel with the ethidium bromide staining. The PCR amplified product was recovered and cloned into pMD18-T vector, then being verified by sequencing.

For the A3 promoter and SV40 PolyA sequence, the vectors pSL1180-A3 or pSilencer™ 4.1-CMV-neo were used as templates for PCR according to above primers and conditions.

Construction of recombinant transformation plasmids with Gn-TII, ST3, GalT and ST6

The Gn-TII and ST6 genes along with their relevant A3 promoters and SV40 PolyA were firstly subcloned to vector of pSL-1180, and then the expression cassettes were further subcloned to pXL-BacII (PiggyBac transposition vector). The resulting vectors were designated as pXL-BacII-Gn-TII and pXL-BacII-ST6.

The expression cassette for ST3 was firstly spliced using the method described by Li et al. [18] with modification. Briefly, two round PCR were performed using primeSTARTM HS DNA Polymerase(Takara),the first PCR with primers pairs A3-F(ST3) & A3-R(ST3), ST3-F(ST3) & ST3-R(ST3) and PolyA-F(ST3) & PolyA-R(ST3) was under the following conditions: 94 °C, 2 min; 35 cycles of 98 °C, 10 s; 68 °C, 1 min 30 s, followed by store at 4 °C forever. PCR product was examined by electrophoresis in 1 % agarose gel with the ethidium bromide staining. The PCR amplified product was purified using a DNA Gel Extraction Kit (Axygen) and used as templates in the next PCR with primer pair A3-F(ST3) & PolyA-R(ST3) under 94 °C, 2 min; 35 cycles of 98 °C, 10 s; 68 °C, 3 min, followed by store at 4 °C forever. PCR product was examined and purified as mention above. Then the purified PCR product was digested by BamHI and XbaI and directly subcloned to pXL-BacII, named as pXL-BacII-ST3. The resulting vector was further confirming by sequencing.

pXL-BacII-GalT was constructed similarly as pXL-BacII-ST3 except for primer pairs of A3-F(GalT) & A3-R(GalT), ST3-F(GalT) & ST3-R(GalT) and PolyA-F(GalT) & PolyA-R(GalT), and digestion by BamHI and EcoRI.

The BamHI and XbaI fragment of pXL-BacII-ST3 was inserted to pXL-BacII-Gn-TII digested by the same enzymes, and then expression cassette of Neo digested by EcoRI from pigA3GFP-hIGF-ie-neo (gift from Prof. Gong in Suzhou University) was further cloned to the vector resulting in a new vector designated as pXL-BacII-Gn-TII-ST3-Neo. The XbaI and XhoI fragment of pXL-BacII-ST6 was inserted to pXL-BacII-GalT digested by the same enzymes, and then expression cassette of GFP digested by BamHI from pXL-BacII-GFP was further cloned to the vector resulting in a new vector designated as pXL-BACII-GalT-GFP-ST6.

Cloning of the transgenic Bm cell line

The Bm N cells were seeded in a 6-well tissue culture plate with 5 × 105 cells per well, and cultured in TC-100 medium with 10 % FBS. The cells were transfected with the vector DNA of pXL-BACII-Gn-TII-ST3-Neor, pXL-BACII-GalT-GFP-ST6 and helper vector DNA (3:3:1, total 4 μg) and transfectant Lipofectin2000 according to the instruction (Invitrogen Corporation, USA).The green fluorescence was observed 24 h post-transfection.

The transfected cells were passaged at a 1:5 into fresh growth medium 24 h post transfection. After the cells adhered to the plates, G418 (Sangon, Shanghai) was added to the medium at the concentration of 1,000 μg/ml. The media were renewed every 3–5 days. G418-resistant clones were isolated by limiting dilution in 96-well plates. Followed stepwise amplification, the cell clones were cultured continually with the selection of G418 to establish the transgenic Bm cell line.

PCR and RT-PCR analysis of transgenic Bm cell line

Total DNA or RNA was extracted and purified from the cloned transgenic Bm cells and stored at −80 °C. The primers of GnTII, GalT, ST3 and ST6 (Table 1) were used to amplify the relative sequences. Amplification was carried out with Takara Ex Taq polymerase for 35 amplification cycles (95 °C/10 s, 68 °C/90 s), finally elongation at 72 °C, 10 min.

A sample (1–2 μg) of total RNA was used as template to synthesize cDNA using superscript II reverse transcriptase (Invitrogen). Continually, PCR was conducted using cDNA as template according to above conditions.

Splinkerette-PCR

The detailed protocol was described elsewhere [28]. Except for EcoRV omitting in this study, all the other steps were the same. Briefly, the genomic DNA was digested using Sau3A1, followed by ligation the digestion product with the synthesized adaptors, and then purified ligation product was used as the template for nested PCR. The primary PCR using Splink1 & SP-L or Splink1 & SP-R1 as primers and primeSTARTM HS DNA as DNA polymerase was performed as follows: 3 min at 95 °C and then 35 cycles of 10 s at 98 °C, 3 min at 68 °C, with a final extension of 10 min at 72 °C. The primary PCR product was used as template of secondary PCR. The secondary PCR using Splink2 & SP-L or Splink2 & SP-R2 as primers and ExTaq as DNA polymerase was perform as follows 15 min at 95 °C and then 35 cycles of 10 s at 98 °C, 30 s at 60 °C, 3 min at 72 °C, with a final extension of 5 min at 72 °C. Then the secondary PCR was cloned to pMD-18T Vector for sequencing. The sequencing results were further analyzed in the SilkDB database (http://silkworm.genomics.org.cn/silksoft/silkmap.html).

Characterization of N-linked glycan by MALDI-TOF MS

The experimental procedures used, including the chromatographic and mass spectrometric conditions, have been described previously [19, 20], with slight modifications in the preparation of the N-glycans.

15 ml of normal and transgenic Bm cells were collected by centrifuged twice at 500×g for 10 min at 4 °C, and add the lysis solution (30 mM Tris, 7 M Urea, 2 M thiourea, 4 %CHAPS, pH 8.5) to 500 μl. The mixture was fully lysed for 1 h in an ice bath and sonicated for 3 min, then centrifuged twice at 20,000×g for 30 min at 4 °C. The supernatant was taken to dryness in a Speed Vac concentrator. 200 μg of dried sample was reconstituted with 500 μl water containing 0.1 % trifluoroacetic acid, and proteolyzed with trypsin and chymotrypsin, and further digested with PNGase F (New England Biolabs, MA, USA) to release N-glycans. After removal of the peptides by Sep-Pack reversed-phase cartridges (Waters, MA, USA), the reducing ends of the N-glycans were subjected to matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS). The instrument was set as 20-kV accelerating voltage, laser shots at 1,500 per spectrum, the spectra with mass range from 900 to 4,000 Da. External calibration was performed with horse heart myoglobin trypsin peptide (Sigma, USA). The identification of N-glycan structures was based on the database http://www.expasy.ch/tools/glycomod/.

Results

Cloning of human Gn-TII, ST3, GalT, ST6, A3 promoter and SV40 PolyA

For constructing the recombinant transformation plasmids, the A3 promoter and SV40 PolyA sequence were PCR-amplified from the vectors pSL1180-A3 or pSilencer™ 4.1-CMV-neo, being verified by sequencing. Meanwhile, the main human N-glycosylation genes Gn-TII, ST3, GalT and ST6 were synthesized from HFL-I cDNA by RT-PCR (Fig. 1).

Fig. 1
figure 1

Schematic representations of recombinant transformation plasmids PXL-BACII-Gn-TII/ST3/Neor (a) and PXL-BACII-GalT/GFP/ST6 (b)

Full size image

The cloned Gn-TII, ST3, GalT, ST6, A3 promoter and SV40 PolyA were 1356, 1002, 1209, 1233, 674 and 270 bp respectively which are consistent with expected sizes.

Construction of recombinant transformation plasmids

The Gn-TII and ST3 genes along with the A3 promoter and SV40 PolyA were sub-cloned into the vector PXL-BACII (PiggyBac transposition vector) named pXL-BACII-Gn-TII-ST3. GalT and ST6 genes were sub-cloned into the vector pXL-BACII, named pXL-BACII-GalT-ST6.

Then the Neor and GFP genes were cut and recovered from the pigA3GFP-hIGF-ie-neo or pXL-BACII-GFP vectors with EcoRI or BamHI respectively, and sub-cloned into above vectors, designated as pXL-BACII-Gn-TII-ST3-Neor and pXL-BACII-GalT-GFP-ST6 (Fig. 1).

Establishment of the cloned transgenic Bm cell line

The Bm cells were transfected with the vector DNAs of pXL-BACII-Gn-TII-ST3-Neor and PXL-BACII-GalT-GFP-ST6 and helper vector DNA (3:3:1, total 4 μg) and transfectant Lipofectin2000. The green fluorescence was observed 24 h post-transfection (Fig. 2).

Fig. 2
figure 2

The fluorescence of Bm cells transfected with the vector DNA of PXL-BACII-Gn-TII/ST3/Neor and PXL-BACII-GalT/GFP/ST6 and helper vector DNA (3:3:1, total 4 μg)

Full size image

The transfected cells were passaged at a 1:5 into fresh growth medium 24 h post transfection. After the cells adhered to the plates, G418 (Sangon, Shanghai) was added to the medium at the concentration of 1,000 μg/ml. The media were renewed every 3–5 days. G418-resistant clones were isolated by limiting dilution in 96-well plates (Fig. 3).

Fig. 3
figure 3

Fluorescence of Bm cells transfected with transgenesis plasmids after 1,200 μg/ml G418 selecting for after 1,000 μg/ml G418 selecting for 14 days

Full size image

The cloned cells were adjusted to the density of 20 cells per ml media, and plated in a plate with 2 cells per hole. The positive mono-clone was confirmed under reverse microscope after 4–5 days, and selected 10 clones, which showed fluorescence under fluorescence microscope. The selected cell clones were cultured continually to establish the transgenic Bm cell line (Fig. 4).

Fig. 4
figure 4

Fluorescence of the establish transgenic Bm cell line. a Normal cells, light; b normal cells, UV; c transgenic Bm cells, light; d transgenic Bm cells, UV

Full size image

Confirmation of insertion in transformed cells by PCR or RT-PCR analysis

To verify that the insertion of the transgenesis was due to transposition events, the PCR and RT-PCR targeting the human N-glycosylation genes were conducted. Figure 5a showed four bands were amplified when using the primers of GnTII, GalT, ST3, ST6 and total DNA of cloned transgenic Bm cells as template, which is consisted with the size of genes respectively. Similarly the RT-PCR presented the same results (Fig. 5b).

Fig. 5
figure 5

Confirmation of insertion in transformed cells by PCR or RT-PCR analysis. a PCR using the primers of GnTII, ST3, GalT, ST6 respectively and total DNA of cloned transgenic Bm cells as template. a Lanes 1–4 DNA of normal Bm cells (non-transformed) as template, Lanes 5–8, DNA of transformed Bm cells as template. b RT-PCR

Full size image

Insert site confirmation by splinkerette-PCR

The sequencing results were shown in Table 2. The sequences between TTAA and GATC (SauIIIAI recognition site) were used for chromosome location in the SilkDB. The results were demonstrated in Fig. 6. These results supported that the recombinant transgenic vectors had been transposited into the genome of BmN cells successfully. However, the clone L2 showed that there may be random integration while transposition took place.

Table 2 Genomic DNA sequences flanking piggyBac insertions
Full size table
Fig. 6
figure 6

Chromosome location of the piggyBac insertions by splinkerette-PCR

Full size image

Characterization of N-linked glycan in normal and transgenic Bm cells by MALDI-TOF MS

The characterization of N-linked glycans in normal and transgenic Bm cells was analyzed by MALDI-TOF MS. Figure 7 and Table 3 showed both of normal and transgenic Bm cells presented eight of N-linked glycans. However five of glycans were specially identified in transgenic Bm cells, which mainly were hybrid or complex N-glycans.

Fig. 7
figure 7

MALDI-TOF mass spectrometry of N-glycans both in normal and transgenic Bm cells. Each sample was dissolved in 0.1 % TFA and proteolyzed with trypsin and chymotrypsin, and further digested with PNGase F to release N-glycans. The MALDI-TOF mass spectrum was acquired on laser shots at 1,500 per spectrum and spectra with mass range from 900 to 4,000 Da. Left N-glycans of normal Bm cells, Right N-glycans of transgenic Bm cells. Blue square N-acetyl glucosamine, green circle mannose, yellow circle galactose, red triangle fucose. (Color figure online)

Full size image
Table 3 Glycan types of normal and transgenic Bm cells
Full size table

Discussion

Many biomedically significant proteins, including antibodies, cytokines, anticoagulants, blood clotting factors, etc. are glycoproteins. Thus there is a high demand for systems that can be used to produce recombinant glycoproteins for basic research and clinic applications. Unfortunately, few currently available recombinant protein expression systems can produce higher eukaryotic glycoproteins with authentic carbohydrate side chains. Furthermore no currently available system can produce large amounts of recombinant glycoproteins in properly glycosylated form at relatively low cost [21–23].

The silkworm, Bombyx mori (Bm) is an ideal bio-factory to produce the exogenous proteins through two manners, baculovirus expression system (BES) and transgenesis. In fact, hundreds of proteins were expressed in the silkworm cell line and larvae. Although the Bm system presented limited trimming function such as glycosylation, phosphorylation, fatty acetylation etc. [24–26], but its processing is much simple compared to the mammalian expression system. The Bm system lacks terminal sialic acid residues in its expressed protein [27], which implied the glycoproteins expressed in silkworm are less complicated than that in mammalians. This limitation would impact the clinical efficacy of insect-derived recombinant glycoprotein product. So the objective of this work is to genetically engineer the silkworm to develop a new system that can express complex, terminally sialylated N-glycans like in mammalian cells.

We cloned the necessary genes in glycosylation of mammalian cells, such as N-acetylglucosaminyltransferase II (Gn-TII), galactosyltransferases (Gal-Ts), 2,6-Sial-T (ST6 GalII) and 2,3-Sial-T (ST3GalIII), and transformed them to silkworm genome in cell lines through transgenesis to establish a transgenic Bm cell line of piggyBac transposon-derived targeting expression of humanized glycoproteins.

Recently a variety of approaches have been developed for insertion-site isolation, including inverse-PCR, vectorette or splinkerette-PCR, panhandle-PCR, Alu-PCR, capture-PCR and boomerang DNA amplification. Although inverse-PCR has been a widely used approach for isolating insertion sites, it is relatively inefficient when compared with other PCR-based approaches. Splinkerette-PCR, which is a variant of ligation mediated PCR. Like any PCR-based method, it is easy to use, reliable, and efficient, and has become the most widely accepted technique for the amplification of transposon insertion sites [28].

Both the PCR or RT-PCR and splinkerette-PCR confirmed that the recombinant transgenic vectors had been transposited into the genome of BmN cells successfully. The GnTII, GalT, ST3, ST6 were inserted into the transgenic Bm cells.

We characterized the N-linked glycans both in normal and transgenic Bm cells by MALDI-TOF MS. The results showed both of normal and transgenic Bm cells presented eight of N-linked glycans. However five of glycans were specially identified in transgenic Bm cells, which mainly were hybrid or complex N-glycans. This implied that the hybrid or complex N-glycans in transgenic Bm cells were more similar with that in mammalian cells.

The study supplied a new insect cell line which is practical to express the production of “bisected” complex N-glycans.