Enhanced protein expression in the baculovirus/insect cell system using engineered SUMO fusions

Abstract

Recombinant protein expression in insect cells varies greatly from protein to protein. A fusion tag that is not only a tool for detection and purification, but also enhances expression and/or solubility would greatly facilitate both structure/function studies and therapeutic protein production. We have shown that fusion of SUMO (small ubiquitin-related modifier) to several test proteins leads to enhanced expression levels in E. coli. In eukaryotic expression systems, however, the SUMO tag could be cleaved by endogenous de-SUMOylases. In order to adapt SUMO-fusion technology to these systems, we have developed an alternative SUMO-derived tag, designated SUMOstar, which is not processed by native SUMO proteases. In the present study, we tested the SUMOstar tag in a baculovirus/insect cell system with several proteins, i.e. mouse UBP43, human tryptase beta II, USP4, USP15 and GFP. Our results demonstrate that fusion to SUMOstar enhanced protein expression levels at least 4-fold compared to either the native or His₆ tagged proteins. We isolated active SUMOstar tagged UBP43, USP4, USP15, and GFP. Tryptase was active following cleavage with a SUMOstar specific protease. The SUMOstar system will make significant impact in difficult-to-express proteins and especially to those proteins that require the native N-terminal residue for function.

The conjugation and de-conjugation of SUMO (small ubiquitin-related modifier) to target proteins is similar to the pathways described for ubiquitin [1,2,3]. Essentially, three ligating enzymes, E1, E2, and E3 work sequentially, in an ATP dependent cascade to couple SUMO through an isopeptide bond to the ε-NH₂-group of lysine residues of the acceptor protein. Unlike ubiquitin, however, SUMOylation does not result in targeting the protein for degradation by the 26S proteosome. Rather, SUMOylation appears to be involved in regulating transport to different intracellular compartments, e.g. the nucleus. In addition, some plasma proteins have been found to be SUMOylated [4] As an integral part of regulating intracellular trafficking, SUMO specific proteases (isopeptidases) are involved in maturation of, removal of and recycling of SUMO. In yeast, the SUMO (Smt3) specific protease Ulp1, removes the C-terminal tripeptide AlaThrTyr of preSUMO to generate mature SUMO with GlyGly at the C-terminus. The carboxylate of the terminal glycine residues is attached to lysyl-ε-NH2-groups in the target protein. Ulp1 and Ulp2 release SUMO from the target protein[5,6,7].

Recombinant fusion of yeast SUMO (Smt3) to the N-terminus of heterologous proteins for expression in bacteria has been shown to enhance both protein expression and solubility [8,9]. This SUMO “tag” can be removed by digestion with Ulp1 to generate the native N-terminus of the fusion partner [10]. Because Ulp1 recognizes exosites in the tertiary structure of SUMO as well as the Gly-Gly dipeptide sequence, it does not cleave other peptide bonds within the fusion partner. While this system works well in prokaryotes, it is likely that endogenous deSUMOylases would limit its utility in eukaryotes. We have created an R64T,R71E double mutant of yeast SUMO (smt3), termed SUMOstar, which is not recognized by native deSUMOylases as well as a mutant Ulp1, SUMOstar protease, that is capable of processing SUMOstar (Panavas, et al. manuscript in preparation). We have previously used this mutant and analogous mutants of human SUMOs 1 and 3 to obtain enhanced expression of sPLA2 in transiently transfected HEK293T and CHO-K1 cells [11]. This study was designed to test the utility of SUMOstar as an expression tag in a baculovirus/insect cell system and determine if it would yield the same enhancement in expression levels and solubility that we have shown with SUMO in bacteria and mammalian cells.

To this end, several proteins were chosen as test objects, green fluorescent protein (GFP), UBP43, USP4, USP15, and tryptase β. UBP43, USP4, USP15, and tryptase were selected because they are the subject of active drug discovery programs. UBP43 is the specific de-conjugating enzyme for ISG15,another small ubiquitin like modifier. Both ISG15 and UBP43 are induced by viral or bacterial infection and by interferon[12,13,14,15]. UBP43 has not been expressed in E. coli in an active form [16]. USP4 (UNP/Unph) is a deubiquitinase that shows both nuclear and cytoplasmic localization. It has been shown to be constitutively associated with the retinoblastoma protein (pRb), the autoantigen Ro52, itself an E3 ubiquitin ligase, and the A2a adenosine receptor. Deubiquitination of each of these proteins prolongs their intracellular lifetime. USP4 is also considered an oncogenic protein although the mechanism of oncogenesis is not understood. Elevated USP4 levels have been found in adenocarcinomas of the lung and injection of cells overexpressing USP4 leads to tumors in nude mice. [17,18,19,20]. USP15 is a Zincfinger containing deubiquitinase associated with the ubiquitin proteasome regulatory COP9 signalosome (CSN). It shows 61% identity with USP4 [21,22,23,24,25,26,20]. Tryptase is a tetrameric trypsin-like protease with a unique tetrameric structure. It is stored as an active enzyme in mast cell secretory granules and has been implicated in the etiology of asthma and other allergic and inflammatory disorders. Both human tryptase α and β form a ring-like tetramer with active sites facing an oval central pore. The tetrameric structure of tryptase α is stabilized by sulfated polysaccharides, e.g. heparin. In the absence of such polymeric anions, tryptase α reversibly converts to an inactive conformation and dissociates into monomers. Tryptase β is more stable although it also requires a polymeric anion for maximal activity. Tryptase β has previously been expressed and isolated in active form from insect cells [27,28,29,30,31,32,33]. GFP was used as an easily detected control protein [34].

Our results show that the native SUMO was indeed cleaved from the fusion proteins in vivo while the SUMOstar tag was not. Both SUMO and SUMOstar lead to enhanced expression of all of the test proteins. USP4 and USP15 were all active with the SUMOstar tag intact; however, tryptase β was only active after the SUMOstar tag was removed.

Materials and Methods

Gene Cloning

A series of pFastBac vectors with His₆, SUMO, His₆Sumo, gp67His₆, gp67SUMO, SUMOstar, His₆SUMOstar, gp67SUMOstar, and gp67His₆SUMOstar was prepared by primer annealing, PCR and subcloning. The SUMOstar template was derived from a bacterial vector generated previously at LifeSensors (Cat. 1101). The gene for Tryptaseβ II was the kind gift of Dr. Norman Schechter. The UBP43 gene was obtained from OpenBiosystem. USP4 and USP15 genes were kindly provided by Dr. Rohan Baker from The Australian National University. These genes were amplified by PCR using the forward and reverse primers shown below. Following purification, the genes were subcloned into pFastBac vectors.

UBP43 - BveI	5’-CGCGACCTGCATCGAGGTATGGGCAAGGGGTTTGGGCTCCTGAGG
	5’-CGCGACCTGCATGTCTAGATTAGGATCCAGTCTTCGTGTAAACCAAG
Tryptase - BpiI	5’-TTTGAAGACGAAGGTATCGTCGGGGGTCAGGAGG
	5’-TTTGAAGACGAAGCTTATTACGGCTTTTTGGGGACATAGTG
GFP - Eco31I/HindIII	5’-ATGATGGGTCTCTAGGTATGGTGAGCAAGGGCGAGGAGCT
	5’-CGCAAAGCTTGAGCTCTTACTTGTACAGCTCGTCCATGCCGA
USP4 - Esp3I/Esp3I-XbaI	5’-CGATCGTCTCTAGGTATGGCGGAAGGTGGAGGCTG
	5’- GCGCCGTCTCTCTAGATTAGTTGGTGTCCATGCTGCAAGCC
USP15 - Esp3I/Esp3I-XbaI	5’-CGTTCGTCTCTAGGTATGGCGGAAGGCGGAGC
	CGCCGTCTCTCTAGATTAGTTAGTG TGCATACAGTTTTCATTTTC

PCR was carried out with 0.5 units Platinum^® Tag DNA Polymerase High Fidelity according to the manufacturer’s instructions (Invitrogen, St. Louis, Mo). The amplified DNA was kept at 4°C prior to final purification which was accomplished on a Epoch miniprep column.

The vectors for UBP43, USP4 and USP15 were digested with Esp3I and XbaI. The vectors for tryptase and GFP were digested with Esp3I and HindIII. The purified PCR fragments were cut with restriction enzymes and isolated by agarose electrophoresis. The vector and insert fragments were ligated directionally and the resulting plasmid was transformed into XL10Gold competent cell (Stratagene, La Jolla, CA). Multiple colonies were screened for the correct ligation. The genes were confirmed by DNA sequencing (Agencourt, Beverly, MA).

Insect Cell Expression

pFastBac-genes were transformed into DH10Bac cells and plated onto agar plate containing kanamycin, gentamycin, tetracycline, and IPTG according to the manufacturers’ instructions (Invitrogen). White colonies were used for isolating Bacmid DNA by a modified procedure. Briefly the MX1, MX2 and MX3 buffers from GenCatch plasmid DNA miniprep kit (Epoch BioLabs) were used for extract DNA from 3 ml cultures of DH10Bac cells. Bacmid were precipitated by adding 70% isopropanol followed by a 30 min incubation at -20°C. The Bacmid DNA was collected by centrifugation at 16.1 Krcf for 15 min at room temperature. The DNA pellet was washed with 70% ethanol, air dried and dissolved in 40 μl of 1×TE buffer (10 mM Tris-HCl, pH 8.5, 10 mM EDTA) by gently tapping the 1.5 ml tube. The Bacmid DNA was stored at 4°C. Two micrograms of bacmid DNA were transfected into 2 × 10⁶ Sf9 cells in a T25 cell culture flask combined with PEI (polyethylenimine) at N/P ratio of about 40. After 5 days, 1 ml of this P1 virus was transferred to a new 4 ml sf9 cell culture in a T25 flask for P2 virus production. After 3 days, the P2 virus was harvested by centrifugation at 600 rcf for 5 min spin. An aliquot (50 -100 μl) of the clarified P2 virus preparation was used for P3 virus production which was harvested after 3 days of culture. The titer of the baculovirus was obtained by plague assay. The multiplicity of infection (MOI) studies were performed and cells and condition media were harvested 3 to 6 days post infection in Sf9 cells as described in the figure legends. Sf9 cells were maintained in Sf900II medium (Invitrogen) with antibiotics and anti-mytotics (Invitrogen) at 0.75 to 10 × 10⁶ cells per ml in 125 ml shake flask in 40 ml cultures.

Protein Isolation

Sf9 cells (1 × 10⁸) were infected with baculovirus harboring the gp67His₆SUMOstar-tryptase gene at an MOI of 4 × 10^-4 and cultured for 90 hours. One hundred ml of conditioned medium were harvested for the isolation of secreted tryptase. Briefly, the cells were removed by centrifugation at 3700 × g_av for 10 min.

The supernatant was isolated, made 1 mM in PMSF, and 1/10 volume of 10X PBS was added. Four mL of Ni-NTA resin was added and the resulting slurry was mixed at room temperature for 90 min. The beads were precipitated by centrifugation at 600 X g for five min, then washed by suspension and centrifugation as above with 50 mL PBS containing 25 mM imidazole and 0.35 M NaCl (WB). The resin was then suspended in WB and poured into a column. The column was washed and packed by passing an additional 20 mL of WB through. Tryptase was eluted with 5 mL of NaKPO₄ buffer, pH 6 containing 0.5 M imidazole. The eluted tryptase was pooled and concentrated in an Amicon Ultra centrifugal filter device using a YM-30 membrane. The buffer was exchanged to NaKPO4 buffer pH 6.0 containing 1 M NaCl by dilution and reconcentration. USPs 4 and 15 were isolated from SF9 cell lysates by Ni-NTA as described above. The lysate was prepared by incubating the cells with WB (50 mM Tris-HCl, pH 8.0, 150 mM NaCl) containing 1% NP-40 and 1 mM PMSF for 30m on ice followed by sonication. The lysate was clarified by centrifugation at 16,000 X g_av for 30 min. The supernatant was isolated and brought to 40 mL with WB before proceeding with the batch binding protocol described above. The bound protein was washed with WB containing 0.5M NaCl, 25 mM imidazole and eluted with WB containing 0.5M NaCl, 0.5 M imidazole. The pooled, concentrated enzymes were stored at −80°C in in 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 20% glycerol and 0.5 mM DTT.

Protein analyses

SDS-polyacrylamide electrophoresis was carried out on homogeneous 12% or 15% polyacrylamide gels. The separated proteins were stained with Commassie brilliant blue or transferred to nitrocellulose membranes in a BioRad semi-dry transfer apparatus for immunoblot analysis. Blotted proteins were localized with an monoclonal anti-His₆ antibody (Amersham Biosciences), followed by anti-mouse conjugated to HRP (Rockland). Chemiluminesence was detected using a Fujifilm LAS3000 (Edison, NJ). Coomassie gels were photographed using the Fujifilm LAS3000. Densitometric analysis of both Western blots and Coomassie blue stained gels was carried out using the Multigauge software supplied with the LAS3000. Protein concentrations were estimated using the Coomassie blue dye binding assay (BioRad, Hercules, Ca.)

Deubiquitinase Assays

As described in the text, the assay is based on the release of inhibition caused by fusion of a ubiquitin-like (UBL)-tag to PLA2 [35]. For UBP43, the UBL is ISG15, whereas, for USPs 4 and 15, the UBL is ubiquitin (Ub) itself. ISG15-PLA2 and Ub-PLA2 were kindly provided by Dr. Ben Nicholson of Progenra. Assays were carried out at room temperature in Greiner black, medium binding 96 well assay plates (E and K Scientific, Santa Clara, CA). Briefly, the deubiquitinases were diluted into freshly prepared assay buffer (50 mM Tris-HCl, pH 8.0, 2 mM CaCl₂, 2 mM β-mercaptomethanol) containing either 20 nM ISG15-PLA2 or 20 nM Ub-PLA2 and 20 μM NBD C₆-HPC (Invitrogen N3786). Negative controls contained either Sf9 cell crude extract (UBP43) or Ni-elution buffer (USPs 4 and 15). The release of active PLA2 was measured by the increase in fluorescence intensity of free NBD (exc. 460 nM; emm. 538 nm) using a Fluoroskan Ascent FL (Thermo Fisher Scientific, Waltham, MA) at either 20 or 60 s intervals over 1 hour. Data for each enzyme were fit to the integrated rate equation 1 using non-linear least squares.

$${\mathrm{A}}_{\mathrm{405}}={\mathrm{V}}_{\mathrm{s}}\left(\mathrm{t}\right)+\left({\mathrm{V}}_0-{\mathrm{V}}_{\mathrm{s}}\right)(\mathrm{1}-{\mathrm{e}}^{(-{\mathrm{k}}_{\mathrm{obs}}\left(\mathrm{t}\right)})/{\mathrm{k}}_{\mathrm{obs}}+\mathrm{c}$$ [1]

In this equation, V₀ is the initial velocity of substrate turn over by PLA2 prior to removal of the SUMOstar tag (≅ 0), V_s is the steady state velocity following removal, c is the initial Abs at 405nm, and k_obs is the psuedo-first order rate constant of activation. Under appropriate conditions, Vs is dependent on the amount of active PLA2 generated, while k_obs is dependent on the amount of deubiquitinase used.

Tryptase Assay

Purified tryptase was diluted into assay buffer (NaKPO₄ buffer, pH 6.0; 1M NaCl; 2 mg/ml heparin). Tryptase substrate Chromozym TH (Tosyl-Gly-Pro-Arg-p-nitroaniline, Cat. No. 10206849001, Roche Applied Science) dissolved in 5% DMSO was added to give a final concentration of 1 mM. SUMOstar specific protease, 0.5 μg with or without DTT, was added to remove the SUMOstar tag and initiate the assay. The absorbance at 405 nm was measured every 60s at room temperature in a Multiskan Ascent plate reader (Thermo Fisher Scientific, Waltham, MA). The data were fit to equation 1 as described above.

Results

Enhanced Intracellular Protein Expression Level by Attachment to the SUMOstar Tag

We have previously shown that the use of the Small Ubiquitin-like MOdifier, SUMO, as a fusion partner for recombinant expression of heterologous proteins in E. coli, resulted in increased expression levels and solubility. Because we anticipated that this benefit would be, at least partially, lost due to endogenous deSUMOylase activity in eukaryotic cells, we developed the R64T,R71E double mutant SUMO-tag that is not recognized by eukaryotic enzymes. We have termed this modified SUMO, SUMOstar (Panavas, et al., manuscript in preparation). In the present study, we examined whether fusion of heterologous proteins to SUMOstar would lead to enhancement of expression levels in a baculovirus/insect cell expression system. Conditions were optimized by varying the multiplicity of infection (MOI) and the time post infection for harvest. Protein expression levels were determined by SDS-PAGE. In the first set of experiments, GFP was expressed as a fusion with a panel of tags – SUMOstar (S*), His₆-SUMOstar (HS*), the gp67 signal sequence (gp67)-His₆-SUMOstar (GHS*), wild type SUMO (S), gp67 (G), gp67- His₆ (GH), and gp67-SUMO (GS). Expression levels were compared to results obtained with untagged GFP. Surprisingly, despite the presence of the gp67 leader sequence, most of the GFP expressed was retained in the cell (see discussion). Based on densitometry of Coomassie-stained SDS-gel, constructs containing SUMOstar exhibited over 3-fold higher levels of expression when compared to the non-SUMOstar constructs (Fig. 1A).; however, this analysis is complicated by the presence of an Sf9 protein which co-migrated with GFP. By Western blot analysis using an anti-His₆ antibody, which allows better quantitative comparison than Coomassie staining, GHS*-GFP levels were found to be > 6 fold higher than the level of GH-GFP (Fig. 1B.) Similar results were obtained for HS*-UBP43; a 3-fold enhancement (corrected for the increased molecular weight of the fusion protein) in expression compared to H-UBP43 (Fig. 2.)

Figure 1.

Analysis of the expression of GFP in Sf9 cells. Sf 9 cells (2 × 10⁶) were infected at an MOI of 0.05 with baculovirus harboring various GFP constructs. The cells were harvested at 120 hours post infection and lysed. (A) Equal amounts total, soluble protein were analyzed by SDS-PAGE and the separated proteins stained with Coomassie Blue. (B) Cell lysates expressing gp67His₆SUMOstar- and gp67His6-GFP constructs were also analyzed by western blot using an anti-His₆ antibody as the primary antibody. Images of the stained gels or western blots were made using a Fuji imager. Fuji image analysis software was used to estimate the relative amounts of each protein. Molecular weights, in kDa, of the marker proteins are shown next to each image. N.b. although the calculated MW of SUMO is 11.5 kDa, it migrates with an apparent MW of ~20 kDa on SDS-PAGE. The positions of native GFP and His₆SUMOstar-GFP are indicated. Gp67 = the gp67 signal sequence, His6 = the hexahis tag, SUMO = the SUMO fusion partner; and SUMO* = the SUMOstar fusion partner.

Figure 2.

Analysis of the expression of UBP43 in Sf9 cells. Cells (10⁴) in 24-well plates were infected with baculovirus harboring either His₆-UBP43 (H), SUMOstar-UBP43 (S*), or His₆SUMOstar-UBP43 (HS*) at different MOI as indicated. The cells were harvested at 140 hours post infection and SDS-PAGE analysis as described in the Materials and Methods. Equivalent amounts of cells were loaded in each lane. Following electrophoresis, the separated proteins were stained with Coomassie Blue. Image collection and analysis was carried out as described in the legend to Fig. 1.

We next wanted to determine whether fusion to SUMO or SUMOstar would lead to enhanced expression of secreted proteins. For this purpose, we examined conditioned culture medium from Sf9 cells infected with baculoviruses containing either GFP or tryptase fused to the gp67 signal sequence with either no tag or the native SUMO or SUMOstar tags by SDS-PAGE (Fig. 3A.). The results of this analysis are shown in Figure 3. Densitometric analysis of the Coomassie Blue staining intensity of the GFP or tryptase bands, indicated that at an MOI of 1, GS*-GFP was expressed at a higher level than either S-GFP or G-GFP. At MOIs of 0.1, 1 and 5, GS*-tryptase was expressed at a level 3-fold higher than that of G-tryptase. From these data, it is clear that fusion to SUMOstar led to increased levels of these recombinant proteins accumulating in the growth medium. On the other hand, comparison of the cell lysate results (Fig. 1B) for GHS*-GFP to the medium results (Fig. 3A, GS*-GFP) suggested that some of the expressed protein was being retained intracellulary despite the presence of the gp67 leader. To determine if tryptase was also being retained intracellularly, we analyzed cell lysates from cells expressing G-tryptase and GS*-tryptase (Fig. 3B). Based on this analysis, it is clear that a portion of the expressed tryptase was not secreted from the cells.

Figure 3.

Analysis of the secretion of Tryptase βII and GFP into the medium by Sf9 cells. (A) Cells (10⁴) were infected with baculovirus harboring either gp67-Tryptase (G), gp67SUMO-Tryptase (GS) or gp67SUMOstar-Tryptase (GS*) at MOIs of 0.1, 1 or 5 in 6 well plates. The corresponding fusions to GFP were only analysed at an MOI of 1. Conditioned medium from each well was collected at 72 hr post infection and clarified by centrifugation at 300 × g for 10 min. The clarified samples were concentrated using a 0.5 ml Microcon concentrator (MWCO 10 kDa, Millipore, Billerica, MA) at 14,000 × g for 40 min. Equal amounts of the concentrated media were loaded in each lane. (B) Aliquots of the cells expressing either G-Tryptase or GS*-tryptase were lysed and analyzed by SDS-PAGE in order to determine if the fusion proteins were being retained within the cells. Following electrophoresis, the separated proteins were stained with Coomassie Blue. Image collection and analysis was carried out as described in the legend to Fig. 1.

Isolation of active enzymes from insect cells

In order to determine if the SUMO-modified baculovirus/insect cell expression system would be useful for the isolation of purified proteins for biological or structural studies, we isolated GHS*-tryptase from conditioned medium and USPs 4 and 15 from cell lysates and tested their activities (Figs. 4 and 5). In order to increase the accumulation of secreted tryptase in the medium, we infected at an extremely low MOI (4 × 10^-4) which, in some instances, has been shown to lead to increased expression and secretion of proteins [36,37,38,39]. The secreted tryptase was readily isolated from conditioned medium by chromatography on a Ni-NTA column as described in the Materials and Methods section (Fig. 4A). Although we did not test it, we assume that the gp67 leader sequence was correctly processed during secretion. The resulting secreted HS*-tagged protein was not active as isolated but required removal of the HS*-tag. Hence, we incubated the purified tryptase in a phosphate buffer containing 1M NaCl, 10% glycerol, 2 mg/mL heparin, and 25 μg/mL SUMOstar protease at pH 6. Activation of tryptase was measured simultaneously by measuring the release of p-nitroanaline from Tosyl-Gly-Pro-Arg-pNA (Fig. 4B). The profile was consistent with a slow activation phase followed by increased turnover of the tryptase substrate as active enzyme was released from inhibition by SUMO.

Figure 4.

Isolation secreted SUMOstar-Tryptase from Sf9 conditioned medium. Secreted gp67His₆SUMOstar-Tryptase from 100 mL of conditioned medium was captured in batch mode by mixing with Ni-NTA as described in the Materials and Methods section. (A) SDS-PAGE analysis of aliquots of the starting condition medium (L), the unbound material (FT), the Wash (W) and pooled and concentrated eluate fractions (E). (B) The pooled and concentrated eluate fraction was assayed for tryptase activity as with and without removal of the SUMOstar fusion partner by incubation with SUMOstar protease using Chromozym TH as a substrate for tryptase as described in the Materials and Methods section. (●) Complete reaction mixture with SUMOstar protease, heparin, and SUMOstar-tryptase. ( ) Reaction mixture with heparin and SUMOstar-tryptase but without SUMOstar protease. ( ) Reaction mixture minus SUMOstar-tryptase and SUMOstar protease. In a separate experiment SUMOstar protease was found to have no activity against the tryptase substrate (data not shown.)

Figure 5.

Isolation of intracellular USP4 and USP15 and activity determination. Isolation of His₆SUMOstar-enzymes from Sf9 cell lysates was carried out as described in the Materials and Methods. (A) SDS-PAGE analysis of aliquots of samples from the isolation of USP4, L = the clarified cell lysate, FT = the unbound fraction from the Ni-NTA column, W = the low imidazole wash from the column and E = the pooled and concentrated elution fractions. The expected position of His₆SUMOstar-USP4 is indicated. (B) SDS-PAGE analysis of aliquots of samples from the isolation of USP15. Samples labed as per (A). The expected position of His₆SUMOstar-USP15 is indicated. The ability of each purified enzyme to process Ub-PLA2 to release active PLA2 was carried out as described in the Materials and Methods. (C) Ub-PLA2 processing activity of either 5 ( ) or 10 ( ) μL of the concentrated pool of USP4. (D) Ub-PLA2 processing activity of either 5 ( ) or 10 ( ) μL of the concentrated pool of USP15. For both, the line represents the fit of the data to equation 1. Boxed points were excluded from the calculation due to depletion of the PLA2 substrate.

Recombinant USP4 and USP15 were also readily isolated from insect cells as His₆-SUMOstar fusions (Figs. 5A and 5C). Progress curves demonstrating the isopeptidase activity of both enzymes are shown in Figure 5B and 5D. This assay takes advantage of the requirement of PLA2 to have a free NH₂-terminus for activity [35]. For USP4 and USP15, a ubiquitin-PLA2 fusion protein was used as the substrate and DUB activity was determined by measuring the activation of PLA2 following removal of the ubiquitin-tag (Fig. 5B and 5D). As isolated, both enzymes were active, without the need of removing the SUMOstar fusion partner.

Discussion

Expression of recombinantly derived, eukaryotic proteins in the baculovirus/insect cell system has seen increased usage in recent years in an attempt to circumvent problems associated with the incorrect folding, lack of disulfide bond formation, and lack of post-translational modifications found in prokaryotic systems. Based on our observation that fusion of the ubiquitin-like molecules in the SUMO family to a variety of proteins led to increased expression and solubility of many of these proteins when expressed in E. coli, we wanted to determine if fusion to SUMO would yield similar results in insect cells [22,40,41,18,42,43,44,45,46]. However, we anticipated that fusions to native SUMO would be subject to intracellular processing by endogenous deSUMOylases; hence, in addition to native SUMO, we tested a new SUMO mutant, termed SUMOstar, that is resistant to cleavage by yeast and human deSUMOylases. In addition to enhancing expression levels of recombinant proteins, use of the SUMO-fusion technology has the advantage that removal of SUMO by deSUMOylases results regenerates the native N-terminus of the target protein. The SUMO proteases recognize not only the Gly-Gly structure found at the C-terminus of SUMO, but also tertiary structure elements only present on correctly folded SUMO. As a result, digestion of SUMO-fusion proteins is limited to the junction between SUMO and the target protein. Since the SUMOstar tag is no longer recognized by native deSUMOylases, we were able to develop a mutated Ulp1 protease which binds to and processes SUMOstar as well as native SUMO although not as efficiently (Panavas et al. manuscript in preparation). As expected, our first series of experiments confirmed that fusion proteins containing native SUMO are processed by endogenous isopeptidases in insect cells (cf. lanes 1,2,4 of Fig. 1 with lanes 3 and 5). In addition, while fusion to SUMOstar did indeed lead to enhanced intracellular expression of soluble GFP and UBP43, fusion to native SUMO only gave expression levels equivalent to either untagged or His₆-tagged proteins. The enhancement in expression seen with SUMOstar is equivalent to that previously seen with native SUMO in E. coli [9].

Fusion to SUMOstar also led to increased levels of soluble recombinant proteins in the growth medium of the cells. Furthermore, significant amounts of SUMOstar-tryptase and SUMOstar-GFP were found to be retained in the cells despite the presence of the gp67 leader sequence. Defects in the secretion of overexpressed recombinant proteins in baculovirus infected cells have been described in the literature [47,48] and it has been proposed that reduced secretion may be a direct result of the overexpression and retention in the endoplasmic reticulum of the viral chitinase gene [49]. Tryptase was also partly retained within the cell, was soluble, and we were able to isolate and activate it by incubation with SUMOstar protease (data not shown). Purification of the secreted His₆-SUMOstar-tryptase from insect cell conditioned medium was facile and the enzyme could be reactivated by incubation with heparin and SUMOstar protease (Fig. 4 ●). Removal of SUMOstar from the enzyme was absolutely required for activity recovery since incubation of the His₆SUMOstar-tryptase with heparin alone did not generate active enzyme (Fig. 4, ). Human mast cell tryptase has been expressed in Pichia pastoris and in the baculovirus/insect cell system by others [50,51,52,31,33]. The enzyme was expressed as fusion with either the natural propeptide [51], a KEX2 propeptide [52], a truncated form of the propeptide [50] or as an N-terminal ubiquitin fusion [31,33]. In each case, an enterokinase (EK) site was engineered between the C-terminus of the fusion partner and the N-terminus of tryptase. Activation of the enzyme required removal of the fusion partner by digestion with EK; however, at least in the case of the ubiquitin fusion protein, an additional EK cleavage was found within tryptase itself. For efficient recovery of active tryptase, Selwood and colleagues were required to introduce a mutation in tryptase in order to eliminate this site. This concern is obviated by use of the SUMOstar/SUMOstar protease system.

Finally, we expressed members of the deubiquitinase family in active form in insect cells, USP4 and USP15. These enzymes were easily isolated in active form from crude lysates and did not require removal of SUMOstar to be active. Each of these enzymes has proven to be difficult to express as soluble active proteins in prokaryotes either due to their size or complex structures. As can be seen from Fig. 5, both USP4 and USP15 can be obtained in good yields from this system and, with minor optimization, would be suitable for biochemical and structural characterization.

In conclusion, we demonstrate that the SUMOstar-system enhances the expression and solubility of difficult-to-expressed recombinant proteins in insect cells. This tag was specially designed for eukaryotic expression systems in order to avoid premature processing of SUMO-fusion proteins by endogenous SUMO hydrolases. Since processing by SUMOstar protease is highly specific for the C-terminus of SUMO and is largely insensitive to the nature of the amino acid residue in the P1’ position, this system has the potential to allow the facile production of proteins which require well defined and specific amino acids at the NH₂-terminus to achieve proper activity, e.g. cytokines [53,54] or chemokines [55,56]. Also, precise processing by SUMOstar protease would led to a more homogeneous product. Thus the SUMOstar system in insect cells will be a powerful tool for structural and functional studies of variety of proteins.

Abbreviations

SUMO

small ubiquitin-like modifier

GFP

green fluorescent protein

NBD C₆-HPC

2-(6-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)hexanoyl-1-hexadecanoyl-sn-glycero-3-phosphocholine

USP

ubiquitin specific protease

UBP

ubiquitin protease