Polycistronic baculovirus expression of SUGT1 enables high-yield production of recombinant leucine-rich repeat proteins and protein complexes

Abstract

The SHOC2-MRAS-PPP1CA (SMP) complex is a holoenzyme that plays a vital role in the MAP kinase signaling pathway. Previous attempts to produce this challenging three-protein complex have relied on co-infection with multiple viruses and the use of affinity tags to attempt to isolate functional recombinant protein complexes. Leucine-rich repeat containing proteins have been historically challenging to express, and we hypothesized that co-expression of appropriate chaperones may be necessary for optimal production. We describe here how the SUGT1 chaperone can, in conjunction with polycistronic protein expression in baculovirus-infected insect cells, dramatically enhance production yield and quality of recombinant SHOC2, the SMP complex, and other leucine-rich repeat proteins.

1. Introduction

The developmental diseases Noonan syndrome and Neurofibromatosis type 1 are often caused by germline mutations in the genes encoding RAS and RAS effector proteins, and mutations in these genes are also observed in more than 30% of all human cancers (1). Specific RAS mutations are more prevalent in different types of cancer and there are both upstream and downstream effector proteins that contribute to the development of cancer that are attractive targets for research (2). RAS activates a signaling cascade by binding and activating members of the RAF kinase family, including BRAF and RAF1 proteins. RAF kinases are initially produced in an autoinhibited state in which the adapter protein 14-3-3 is bound in a phosphorylation-dependent manner to two phosphorylated serine residues, which in the case of RAF are pS259 and pS621 (3). In order for signaling through RAS to occur, dephosphorylation of pS259 is required to allow activation of the kinase. A three-protein complex comprised of SHOC2, MRAS, and PPP1CA (known collectively as SMP) has been identified that specifically dephosphorylates RAF1 pS259 and the analogous BRAF pS365 (4). MRAS is a member of the RAS family of proteins and shares regulatory and effector proteins (5). SHOC2 is an evolutionarily well-conserved scaffolding protein comprised primarily of leucine rich repeats (LRRs) (6). Scaffold proteins are important regulators in many signaling pathways and act by binding key proteins into complexes, localizing pathway proteins together, and controlling positive and negative feedback. SHOC2 contributes to modulation of RAF signaling activity through interaction with multiple proteins including RAS, PPP1CA, SCRIB, ERBIN, and HUWE1 (6). SCRIB and ERBIN are LAP (LRR and PDZ domain) proteins which mediate SMP activity by interacting with PPP1CA and SHOC2 to disrupt SMP/RAF1 signaling while HUWE1 is involved in ubiquitination of SHOC2 and RAF1 (6,7). The final protein in the complex, PPP1CA, is one of a limited number of serine/threonine phosphatases which can form hundreds of distinct protein-protein complexes to specifically dephosphorylate proteins (8,9). It is unclear how MRAS and SHOC2 target the phosphatase to specifically activate RAF kinases; elucidating the mechanism of this process is important both for understanding a vital RAS signaling component and potentially enabling drug discovery to target this process. Producing large amounts of this complex has been a significant challenge due to poor solubility caused by misfolding of multiple proteins in the complex.

Highly pure, well folded, and functional protein is the goal of recombinant protein expression. Some proteins are highly expressed in soluble form and are easy to purify, while others can be either insoluble or purify as soluble aggregates. Often, expression conditions such as temperature, time of expression, or cell line can enable better protein expression without resorting to altering the target protein. Even after protein expression, the lysis methodology and purification process can make a difference in the yield and quality of the final protein product. However, some proteins cannot be rescued by changing either expression system or purification changes. These proteins need assistance in folding during expression which can only be achieved by co-expressing chaperones and/or co-chaperones. The Heat shock protein 90 (Hsp90) chaperone system is one of the most highly conserved systems among many species and has been extensively studied (10,11). In addition, this system has many co-chaperones that interact with both specific client proteins and the heat shock proteins to promote protein folding and stability. CDC37 is a co-chaperone well known to interact with Hsp90 client kinases, while other co-chaperones such as Hop, p23, and Aha1 act by either stimulating or inhibiting Hsp90 ATPase activity. A study using high throughput LUMIER assays and mass spectrometry identified interaction networks including chaperones, co-chaperones, and client proteins (12). In addition to describing the previously known specificity of the CDC37:kinase interaction, this study found two other domain specific co-chaperones families, Sgt1 and NUDC proteins. Sgt1 is an essential yeast protein (the human homolog is known as SUGT1) and has been found to strongly interact with proteins with LRRs and has a high LUMIER score for interaction with SHOC2 (12,13). Lacking direct interaction with Hsp90, SUGT1 has been implicated in multiple studies to be a co-chaperone for substrate transfer between Hsp70 and Hsp90 in the same way that HOP does for glucocorticoid receptors (13,14). Although the link between SUGT1 and LRR-containing proteins has been known for many years, specific client proteins for SUGT1 have not been extensively reported, especially in the context of recombinant protein expression. The work presented here shows the improvement of SMP complex production using the co-chaperone SUGT1. Co-expression of SUGT1 with SHOC2, PPP1CA, and MRAS resulted in the production of highly pure, functional complexes by both folding SHOC2 and stabilizing PPP1CA to facilitate binding to MRAS. Additionally, we demonstrate the significant benefit of polycistronic co-expression of the SUGT1 chaperone as compared with the more common method of co-infection. We also show that it is possible to isolate similarly pure and high-yield SHOC2:PPP1CA complexes with KRAS4b, HRAS, and NRAS using co-expression of SUGT1. Further testing of SUGT1 co-expression with other LRR proteins such as SCRIB and ERBIN shows significant differences in the way in which SUGT1 might impact protein folding during expression. Overall, these data add another weapon to the armory for high-yield production of challenging protein targets and protein complexes containing the ubiquitous LRR motifs.

2. Materials and Methods

2.1. DNA

DNA constructs for protein expression were generated using Gateway recombinational cloning or combinatorial Gateway Multisite cloning from a variety of sequence validated Gateway Entry clones (15). In all cases, Entry clones encoding human proteins were produced from synthetic templates constructed by ATUM (Newark, CA) and DNA was optimized for protein expression in insect cells. Templates were flanked by Gateway recombination sites as noted in Table 1 and fully sequence validated. Several constructs contain protease sites (ENLYFQ/G) for tobacco etch virus (TEV) protease where noted (tev). Final baculovirus expression clones shown in Table 2 were generated by LR recombination or multisite LR recombination as per the manufacturer’s instructions (Thermo Fisher Scientific, Waltham, MA). Destination vectors include pDest-8 (native baculovirus expression, Thermo Fisher Scientific), pDest-636 (His6-MBP fusion for baculovirus expression, Addgene #159574), and pDest-623 (multisite att4-att3 baculovirus expression, Addgene #161878). After recombination reactions were completed, final expression clones were validated by agarose gel electrophoresis and restriction digest and were used to generate bacmid DNA with the Bac-to-Bac system (Thermo Fisher Scientific, Waltham, MA) using the manufacturer’s instructions.

Table 1:

Entry clone constructs utilized in this manuscript

Entry ID	Entry Clone Protein	Gateway sites
R725-E99	MRAS(1-178) Q71L	att4-att1r
R745-E05	His6-MBP-tev-PPP1CA(2-330)	att1-att2
R735-E10	tev-SHOC2(58-564)	att1-att2
R725-E100	SHOC2	att2r-att3
R751-E20	SUGT1	att4-att5
R745-E30	MRAS(1-178) Q71L	att5r-att1r
R939-E30	HRAS(1-169) Q61L	att5r-att1r
R939-E34	NRAS(1-169) Q61L	att5r-att1r
R949-E92	KRAS4b(1-169) Q61L	att5r-att1r
R745-E50	tev-SCRIB(1-418)	att1-att2
R745-E51	tev-ERBIN(1-430)	att1-att2
R751-E14	SUGT1	att1-att2

Table 2:

Expression clone constructs utilized in this manuscript

Clone ID	Protein Expressed	Entry 1	Entry 2	Entry 3	Entry 4	DV
R735-M96–623	SMP complex	R725-E99	R745-E05	R725-E100	N/A	623
R735-M99–623	SMP complex + SUGT1	R751-E20	R745-E30	R745-E05	R725-E100	623
R745-M53–623	SHP complex + SUGT1	R751-E20	R939-E30	R745-E05	R725-E100	623
R745-M57–623	SNP complex + SUGT1	R751-E20	R939-E34	R745-E05	R725-E100	623
R745-M45–623	SKP complex + SUGT1	R751-E20	R949-E92	R745-E05	R725-E100	623
R745-X50–636	H6-MBP-tev-SCRIB	R745-E50	N/A	N/A	N/A	636
R745-X51–636	H6-MBP-tev-ERBIN	R745-E51	N/A	N/A	N/A	636
R735-X10–636	H6-MBP-tev-SHOC2(58–564)	R735-E10	N/A	N/A	N/A	636
R751-X14–008	SUGT1	R751-E14	N/A	N/A	N/A	008

2.2. Protein expression in insect cell culture

Baculovirus was produced by transfection of bacmid DNA into Sf9 cells adapted to serum-free suspension and grown in SF900-III SFM media (Thermo Fisher Scientific, Waltham MA). Briefly, 100 ml of Sf9 cells were set at 1.5 x 10⁶ cells/ml and transfected using DNA:Cellfectin II (Thermo Fisher Scientific, Waltham MA) lipid complex. The DNA:Cellfectin II complex was prepared using 70 ul of prepared bacmid DNA and 250 ul Cellfectin II, both diluted in 500 ul SF900 III media and mixed together. The transfected culture incubated for 120 hours at 27°C in 250 ml Optimum Growth Flasks (Thomson Instrument Company Oceanside, CA) shaking at 105 RPM on an Innova 44 shaker with a 2” orbit (New Brunswick Scientific, Edison, NJ). After harvest of the cell culture supernatant at 120 hours post transfection, the baculovirus titer was measured by qPCR using TaqMan Gene Expression Assay for baculovirus protein GP64 and TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific, Waltham MA).

Protein expression was carried out in Tni-FNL cells isolated and adapted at the Frederick National Laboratory for Cancer Research (16). All proteins were expressed in 1-2 liters of SF900 III SFM media (Thermo Fisher Scientific, Waltham MA) set at 7 x 10⁵ cells/ml in 2.8 L or 5 L Optimum Growth Flasks (Thomson Instrument Company Oceanside, CA) the day before infection so that cells would double to ~1.5 x 10⁶ cells/ml on the day of infection for large scale purification. For small scale purification 50 mls of Tni-FNL cells were set at 1.5 x 10⁶ cells/ml on the day of infection. Single baculovirus expressions were infected at a multiplicity of infection (MOI) of 3 and multi-baculovirus expressions were tested at a variety of MOI combinations before choosing the optimal conditions in which we infected at MOI of 6 (SHOC2 or SMP) and MOI of 3 (SUGT1). Infected cell culture was incubated for 72 hours at 21°C, shaking at 105 RPM on an Innova 44 shaker with a 2” orbit (New Brunswick Scientific, Edison, NJ). Cells were harvested and cell pellets stored at −80°C until purification.

2.3. Protein Purification

Small-scale purification was carried out using a modified version of the procedure first described previously (17). Cell pellets were thawed, resuspended and lysed in 5 ml lysis buffer (20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP) per 50 ml of culture. Lysis was performed using the LV-1 Microfluidizer (Microfluidics Corp., Westwood, MA) for 2 passes at 7,000 psi on ice. Lysates were then clarified by ultracentrifugation at 100,000 x g for 30 minutes at 4°C. Proteins were purified on the Phynexus MEA (Biotage, Uppsala, Sweden) at 4°C using 40 μL bed volume (BV) nickel-charged IMAC tips (Biotage, Uppsala, Sweden). The columns were washed in a 96-well deepwell plate with 20BV H2O and then equilibrated in 20BV of 20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP, 25 mM imidazole. The column tip was loaded with 100 μL of sample and then washed with 40BV equilibration buffer. Next, the protein was eluted using 3 x 2BV bump elutions using lysis buffer with 125 mM, 250 mM, and 500 mM imidazole. Total protein, soluble protein, column flowthrough, and column elutions were analyzed by SDS-PAGE and Coomassie-staining.

For large-scale purification, cell pellets were thawed, resuspended and lysed in 100 ml lysis buffer, 20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP, per liter of culture. Lysis was performed using the M-110EH Microfluidizer (Microfluidics Corp., Westwood, MA) for 2 passes at 7,000 psi on ice. Lysates were then clarified by ultracentrifugation at 100,000 x g for 30 minutes at 4°C and filtered through a 250 ml Autofil 0.45 μm High Flow PES Bottle Top Filter (Thomas Scientific, Swedesboro NJ) and used immediately. All proteins were purified on NGC chromatography systems (Bio-Rad Laboratories, Hercules CA) using immobilized metal affinity chromatography (IMAC) and size exclusion chromatography (SEC). The flowrate for all IMAC steps was 5 ml/min and for SEC was 1ml/min. Clarified lysate was adjusted to 35 mM imidazole and loaded onto a 5 ml Ni Sepharose High Performance nickel-charged column (GE Healthcare, Chicago, IL) equilibrated in 20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP, 35 mM imidazole. The column was washed for 5 column volumes (CV) with the equilibration buffer and protein was eluted using a gradient of 7-100% Buffer B (20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP, 500 mM imidazole) for 20 CV, and column elution collected in 5 ml fractions. Total protein, soluble protein, and elution fractions were analyzed by SDS-PAGE and Coomassie-staining. Appropriate fractions were pooled and dialyzed into 2 liters 20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP with His6-TEV protease at a 1:20 (v/v) protease:pool volume ratio overnight at 4°C. The digested material was processed by a second IMAC to separate the cleaved His6-MBP tag on the PPP1CA from the SMP complex. The second IMAC column was equilibrated with 20 mM HEPES pH 7.4, 300 mM NaCl, 1 mM TCEP and proteins were eluted with a shallow gradient from 0-10% Buffer B for 10CV followed by a 10-100% Buffer B gradient for 10CV. Elution fractions containing the desired complex by Coomassie stained SDS-PAGE were pooled from the shallow gradient and concentrated using a 30K Amicon Ultra 15 Centrifugal Filter Units (Millipore Sigma, Burlington MA) down to 5 ml for size exclusion chromatography. The concentrated sample was applied to an SEC column (HiLoad 16/600 Superdex 200, Cytiva, Chicago IL) equilibrated in final buffer (20 mM HEPES pH 7.4, 150 mM NaCl, 1 mM TCEP). Final buffer was flowed over the column at 1.2 ml/min and fractions were pooled based on Coomassie-stained SDS-PAGE and a final yield concentration was determined by measuring A280 (Nanodrop 2000C spectrophotometer (Thermo Fisher Scientific, Waltham MA).

2.4. RAF kinase dephosphorylation assay

PPP1CA and SMP complex phosphatase activity was tested on purified RAF1 and BRAF proteins. RAF1 and BRAF were diluted in 20 mM HEPES pH 7.4, 150 mM NaCl, 1 mM TCEP, 2 mM MgCl₂, 2 mM MnCl₂ to a final concentration of 686 nM. 20 ul of diluted RAF1/BRAF sample was mixed with 20 ul of 204 nM SMP or PPP1CA and incubated at 30°C for 30 minutes. After 30 minutes, 40 ul of 2x NuPAGE LDS Sample Buffer (Thermo Scientific, Waltham, MA) was added to the tube and samples were boiled for 5 minutes to stop the reaction. Western blots were prepared by electrophoresing 10 ul of each sample on an SDS-PAGE gel, transferring samples to a PVDF membrane via iBlot (Thermo Scientific, Waltham MA) using standard manufacturer’s conditions, and probing for RAF1 pS259/BRAF pS365 (Abcam #ab173539), RAF1 pS621/BRAF pS729 (Abcam #ab4767), and anti-His6 to detect tagged BRAF and RAF1 proteins (Abcam #ab18184). Final images were taken using an Odyssey CLx (LI-COR Biosciences, Lincoln NE).

3. Results/Discussion

3.1. Optimization of expression constructs for SMP complex production

Initial SHOC2, MRAS, and PPP1CA constructs (Fig. 1A) tested were designed with three different purification tags to be used with multiple affinity purifications for high-purity final protein. The glutathione-S-transferase (GST) tagged MRAS had low levels of expression but high solubility. However, both the TwinStrep-tagged SHOC2 and His6-tagged PPP1CA had high levels of expression but were almost completely insoluble in our initial tests (Supplemental Fig. S1, panel A) leading to production of a complex at very low yield (0.1 mg/L) and poor purity (Fig. 1C, left panel)

Fig. 1. Optimization of constructs for SMP protein complex production.

A. Initial constructs tested included a GST tag (yellow) on MRAS, a TwinStrep tag (blue) on SHOC2, and a His6x2 tag (green) on PPP1CA, all containing a tobacco etch virus protease site between the tag and protein of interest. B. Optimized constructs have untagged MRAS and SHOC2 and a single His6 tag (green) with a maltose binding protein (MBP, red) on PPP1CA, separated by a tobacco etch virus protease site. C. SDS-PAGE/Coomassie staining analysis of final purified proteins for initial constructs (left) and optimized constructs (right). Protein bands representing SHOC2 (S), MRAS (M), and PPP1CA (P) are noted with arrows. Molecular weight markers are shown in kilodaltons.

Maltose binding protein (MBP) is a common solubility enhancing tag used in both prokaryotic and eukaryotic expression systems (18). Based on success with other proteins in our laboratory, we redesigned all of the SMP constructs to resemble our standard MBP-based protein expression strategy more closely. Placing MBP tags on all three components improved the apparent solubility of each protein but the overall yield and quality of purified complex was not significantly improved (Supplemental Fig. S1, panel B). After testing multiple tags and expression conditions using these constructs, we settled on an optimal combination of native full-length SHOC2, native MRAS G-domain (amino acids 1-178, lacking the C-terminal hypervariable tail), and full-length His6-MBP-tev-PPP1CA containing a tobacco etch virus (TEV) protease site (tev) between the fusion tag and PPP1CA. (Fig. 1B). We decided to tag only the PPP1CA component since it was the lowest expressing of the three components and tagging only that subunit ensured that any unbound SHOC2 and MRAS would flow through the initial IMAC purification. Using this combination of constructs, we were able to increase both the yield and purity of the complex (Table 3, Fig. 1C) from 0.1 mg/L to 2 mg/L. However, purity levels were still insufficient for downstream structural biology needs, and the triple infection required for production of the complex was suboptimal for final complex yield and reproducibility.

Table 3:

Purification yields of SHOC2-MRAS-PPP1CA complexes

Proteins	Viruses used	Protein Yield (mg/L)	Avg. Yield (mg/L)
SHOC2 + MRAS + PPP1CA	3	3.0, 1.9, 1.1, 2.1	2.0 ± 0.7
SHOC2 + MRAS + PPP1CA + SUGT1	4	5.8, 3.7, 3.7	4.4 ± 1.0
SHOC2/MRAS/PPP1CA	1	5.5, 5.7, 3.5, 4.6	4.8 ± 0.9
SHOC2/MRAS/PPP1CA + SUGT1	2	26.5, 27.0, 27.0	26.8 ± 0.2
SHOC2/MRAS/PPP1CA/SUGT1	1	33.1, 37.5, 27.2	32.6 ± 4.2

3.2. SUGT1 co-expression with SHOC2 and SMP complex

Several literature reports suggest that chaperone/co-chaperone systems specific to a variety of protein domains, such as leucine-rich repeats like those found in SHOC2, were vital to the proper intracellular production of correctly folded protein complexes (12,13). We hypothesized that co-expression of co-chaperones might enable higher yield and solubility of our highly overexpressed recombinant proteins. Initially, we tested co-expression of human SUGT1 (a homolog of yeast SGT1) on the production of SHOC2. Fig. 2A shows total and soluble protein expression and microscale purification of His6-MBP-SHOC2 alone (left) and in combination with SUGT1 (right). This test clearly showed a dramatic improvement in both SHOC2 solubility and purification yield. Expression of SHOC2 in the absence of SUGT1 produced less than ~20% soluble protein, while addition of the co-chaperone led to 100% soluble SHOC2. Based on this result we performed a quadruple infection with SUGT1 and all three components of the SMP complex (Fig. 1B), which more than doubled the final yield of the complex from 2 mg/L to 4.4 mg/L (Fig. 2B, Table 3). Pooled fractions from this purification were assessed by preparative size exclusion chromatography (SEC), and based on known molecular weight standards, the complex migrated as expected for a 122 kDa protein complex (Fig. 2C).

Fig. 2. SUGT1 co-expression improves SHOC2 and SMP complex production.

A. SDS-PAGE/Coomassie staining analysis of small scale IMAC chromatography of full length His-6-MBP-SHOC2 alone (left) and with SUGT1 co-expression (right), M – molecular weight standards noted in kilodaltons, T – total protein expression, S – soluble protein/column load, FT – column flowthrough. E1 – 125 mM imidazole elution fraction, E2 – 250 mM imidazole elution fraction, E3 – 500 mM imidazole elution fraction. B. SDS-PAGE/Coomassie staining analysis of large scale IMAC chromatography of SMP and SUGT1 quadruple infection, W – column wash. C. SEC chromatogram of SMP and SUGT1 quadruple infection with molecular weight standards noted at appropriate elution volumes in kilodaltons.

3.3. Polycistronic expression of the SMP protein complex

The multiplicity of infection (MOI) for a baculovirus infection is a vital parameter in recombinant protein expression in insect cells, particularly when infecting cells with multiple baculoviruses. The Trichoplusia ni Tni-FNL cell line supports high protein expression, but relatively low virus amplification compared to Sf9 cells (16,19). The reduced virus amplification capacity of Tni-FNL cells means that an MOI of at least 3 is required to obtain at least one viral particle per cell for greater than 95% of the cells in culture (20). It is especially crucial when expressing multiple proteins that form a complex, that the correct MOI ratio between each virus is determined to permit expression of each protein in equimolar amounts. However, during a multi-baculovirus infection it is nearly impossible to ensure that each cell is infected with the same amount of each virus or that each cell is infected with at least one copy of each virus. The result is a heterogenous sample of cells which may express 1, 2, or 3 proteins at varying levels, which can lead to significant accumulation of failed complexes or insoluble proteins. To solve this problem, many laboratories have developed methods to generate polycistronic baculoviruses which express multiple genes from one virus (21,22). In our laboratory, we have addressed this problem by building a series of Gateway Entry clones, each containing a baculovirus polyhedrin promoter transcribing a particular gene of interest. We then can use multisite Gateway recombinational cloning to generate polycistronic constructs by linking together these individual transcription cassettes in a single viral genome (15).

Using a single baculovirus with the optimal SHOC2, MRAS, and PPP1CA constructs increased the yield of the SMP complex from 2 mg/L to 4.8 mg/L, more than doubling the yield compared with the triple infection (Table 3, Fig. 3A). This yield improvement likely resulted from every cell now expressing all three proteins in equal amounts. Further co-infection of the SMP polycistronic baculovirus with a virus containing SUGT1 increased the yield an additional 6-fold to 26.8 mg/L (Fig. 3C). Even though both the SMP polycistronic virus alone and with SUGT1 produced very similar amounts of protein in the initial cell extracts (“T” lanes, Fig. 3A and Fig. 3C), SMP complex formation and yield without SUGT1 is highly limited by the reduced levels of properly folded SHOC2 and stable PPP1CA. With SUGT1 co-infection, the majority of the PPP1CA is folded into the complex and the SHOC2 is removed from the flowthrough of the IMAC purification (“FT” lanes, Fig. 3A and Fig. 3C).

Fig. 3. Polycistronic production of SUGT1 and SMP complexes further improves yield and quality of the SMP complex.

A. SDS-PAGE/Coomassie staining analysis of large scale IMAC chromatography of polycistronic SMP without SUGT1. B. IMAC chromatogram of polycistronic SMP without SUGT1. C. SDS-PAGE/Coomassie staining analysis of large scale IMAC chromatography of polycistronic SMP with co-infection of SUGT1. D. IMAC chromatogram of polycistronic SMP with co-infection of SUGT1. E. SDS-PAGE/Coomassie staining analysis of large scale IMAC chromatography of polycistronic SMP including SUGT1. F. IMAC chromatogram of polycistronic SMP including SUGT1. M – molecular weight standards noted in kilodaltons, T – total protein expression, S – soluble protein/column load, FT – column flowthrough, W – column wash.

To further improve the system, we combined the three SMP proteins and SUGT1 all on a single polycistronic baculovirus. With this new virus, we were able to generate SMP complex at a yield of 32.6mg/L (Table 3). Fig. 3E shows the purification of this complex and notably demonstrates that nearly all SHOC2 and PPP1CA protein is now present in the proper SMP complex. The IMAC chromatograms (Fig. 3D as compared with Fig. 3B) also highlight how the presence of SUGT1 increases the purifiability of the complex while still leaving some excess PPP1CA uncomplexed in the righthand peak. This is likely due to the unequal ratio of viruses often observed in co-infection, which can lead to incomplete complex formation in some cells. In the polycistronic virus which expresses all 4 proteins from the same virus (Fig. 3F), this peak is nearly eliminated, suggesting that most of the purified protein represents the full SMP complex. In total, the combination of vector, expression, and purification optimization increased the overall yield of SMP complex by over 300-fold from 0.1mg/L to 32.6mg/L, and dramatically improved purity of the final complex.

Purified SMP complexes were validated in a dephosphorylation assay which showed that they could remove the specific phosphoserine residues on RAF1 (pS259) and BRAF (pS365) without dephosphorylation of other phosphoserines (Fig. 4), indicating that the complexes produced were functional. The PPP1CA phosphatase alone was unable to dephosphorylate these residues, demonstrating the necessity of SHOC2 and MRAS for targeting the phosphatase to the appropriate substrate on the RAF kinases. In addition, the high yield of extremely pure proteins produced using this method led to the generation of crystals of both SHOC2 and SMP complex which are currently being used to elucidate the high-resolution crystal structure of these proteins for the first time (Bonsor D., et al, manuscript in preparation).

Fig. 4. Assay for phosphatase activity of purified SMP complexes on BRAF and RAF1 proteins.

Western blot analysis of RAF proteins incubated with SMP complexes using an overlay of antibodies against total RAF1/BRAF (green) and phosphospecific sites pS259 and pS621 on RAF1 or pS365 and pS729 on BRAF (red). A. Activity on RAF1 pS259 and pS621. B. Activity on BRAF pS365 and pS729. Complexes containing MRAS (SMP), KRAS (SKP), NRAS (SNP), and HRAS (SHP) are compared with untreated control protein, PPP1CA phosphatase alone, or phage lambda phosphatase.

3.4. SUGT1 co-expression with other SHOC2-containing protein complexes

In addition to the SMP complex, we also produced SHOC2/PPP1CA complexes with other RAS isoforms, including KRAS (SKP), HRAS (SHP), and NRAS (SNP). Initially, in the absence of SUGT1 co-expression, we were unable to produc significant quantities of these complexes, which are believed to have lower affinity than the SMP complex. However, when combined in the 4-piece polycistronic baculovirus format used for SMP, we were able to isolate at high yield the same 1:1:1 functional complex as observed with SMP (Fig. 5). As expected from the lower affinity of the complexes, the yields were significantly lower than we saw with the SMP complex (Table 4), but still high enough to produce sufficient high-quality protein for biochemical and structural experiments.

Fig. 5. SHOC2/PPP1CA complexes with other RAS isoforms.

SDS-PAGE/Coomassie staining analysis of final purified proteins for SNP (left), SKP (middle), and SHP (right) complexes. Each complex was made using the same SHOC2, PPP1CA and SUGT1 polycistronic construct with various RAS G-domain isoforms as noted.

Table 4:

Purification yields of SHOC2-RAS-PPP1CA complexes with different RAS isoforms

Construct	RAS isoform	Yield (mg/L)	Average (mg/L)
SMP	MRAS	33.1, 37.5, 27.2	32.6 ± 4.2
SNP	NRAS	18.5, 23.5, 22.2	21.4 ± 2.1
SKP	KRAS	11.1, 10.6, 14.5	12.1 ± 1.7
SHP	HRAS	12.3, 23.2, 24.7	20.0 ± 5.5

3.5. SUGT1 co-expression with other LRR-containing proteins

SCRIB is another LRR-containing protein that is poorly folded when expressed recombinantly in insect cells. We co-expressed SCRIB with SUGT1 to determine experimentally whether the co-chaperone might have a similar impact on SCRIB as it did on SHOC2. As shown in Fig. 6A, expression of SCRIB in the absence of SUGT1 produces high levels of soluble protein, but the majority of the protein fails to bind the IMAC column. This likely represents soluble aggregate and/or misfolded protein due to the lack of co-chaperone for SCRIB folding. When SUGT1 is co-expressed using a second virus (Fig. 6B), binding of SCRIB to the IMAC column is enhanced. Although the protein is highly expressed both with and without SUGT1, the final yield increased from 9.3 mg/L to 26.4 mg/L despite the suboptimal use of co-infection instead of co-expression of SUGT1.

Fig. 6. Effect of SUGT1 co-expression on production of SCRIB.

The SCRIB construct used for testing co-expression with SUGT1 contains the LRR domains including amino acids 1-418 and excludes other domains. A. SDS-PAGE/Coomassie staining analysis of IMAC of SCRIB (1-418) alone. B. SDS-PAGE/Coomassie staining analysis of IMAC of SCRIB (1-418) with SUGT1 co-expression by co-infection. M – molecular weight standards noted in kilodaltons, T – total protein expression, S – soluble protein/column load, FT – column flowthrough, W – column wash.

ERBIN is an LRR-containing protein involved in the regulation of the activity of the SMP complex. Expressed alone (Fig. 7A and Fig. 7B), ERBIN is mostly aggregated, and solubility tag cleavage is poor (“Pool 1”), with only a small percentage (“Pool 2”) of monomeric protein that purifies with many higher molecular weight contaminants. With co-expression of SUGT1 (Fig. 7C and Fig. 7D), there is an additional aggregate pool that co-elutes with SUGT1 and cleaved protein (“Pool 2”), a pool that is cleaved ERBIN with SUGT1 (“Pool 3”), and the monomer pool (“Pool 4”). This collection of different fractions on the SEC suggests that the SUGT1 binds ERBIN but gets stuck and cannot cycle completely through the chaperone system to fold the remainder of the aggregated protein. It is possible that additional co-chaperones could be needed to fully allow ERBIN to fold completely. In comparison to SHOC2, both SCRIB and ERBIN have many more predicted disordered regions and additional functional domains which could account for the difference in the ability of SUGT1 to completely rescue the production of these proteins.

Fig. 7. Effect of SUGT1 co-expression on production of ERBIN.

The ERBIN construct used for testing co-expression with SUGT1 contains the LRR domains including amino acids 1-430 and excludes other domains. A. SDS-PAGE/Coomassie staining analysis of the SEC aggregate peak (pool 1) and included peak (pool 2) for ERBIN alone. B. SEC chromatogram of ERBIN alone. C. SDS-PAGE/Coomassie staining analysis of the SEC aggregate peaks (pool 1 and pool 2) and included peaks (pool 3 and pool 4) for ERBIN with SUGT1 co-expression. D. SEC chromatogram of ERBIN with SUGT1 co-expression. Molecular weight standards (M) on SDS-PAGE gels are noted in kilodaltons. Locations of molecular weight standard elution on SEC columns are denoted by vertical arrows and relative sizes are indicated in kilodaltons.

4. Conclusions

We have demonstrated the ability to significantly improve yields of recombinant protein complexes containing LRR proteins using a combination of construct design, co-chaperone co-expression using SUGT1, and polycistronic baculoviruses. It is likely that for these proteins, any enhancement in production utilizing solubility tags only increases the level of soluble aggregate formation, and in the absence of the proper co-chaperone folding system, production of high-yield and high-purity proteins would be challenging. While the largest improvement seen was in the production of the SMP complex, we also have shown that other proteins in the LRR family such as ERBIN and SCRIB are also improved to different levels using this SUGT1 system. It remains to be seen why some proteins may perform better than others or whether some proteins may require additional chaperones to complete their proper folding. However, it is clear that SUGT1 plays a major role in the folding process, and it provides a significant weapon in the arsenal of chaperone-mediated recombinant protein production which should be considered for the production of many important proteins that contain leucine-rich repeats. In addition, we have provided another example that the optimal production of protein complexes in baculovirus-infected insect cells benefits significantly from polycistronic viruses that ensure proper expression of all proteins in each cell. As structural biology and drug discovery efforts move consistently towards research into larger multiprotein complexes, the need to ensure consistent high-quality production of these proteins will continue to grow.

Supplementary Material

1

Supplemental Fig. S1. IMAC purification of initial constructs prior to optimization. A. SDS-PAGE/Coomassie staining analysis of large scale IMAC chromatography of co-infection of GST-MRAS, TwinStrep-SHOC2, and His6-PPP1CA proteins. B. SDS-PAGE/Coomassie staining analysis of large scale IMAC chromatography of co-infection of His6-MBP-SHOC2, His6-MBP-PPP1CA, and His6-MBP-MRAS proteins. M – molecular weight standards noted in kilodaltons, T – total protein expression, S – soluble protein/column load, FT – column flowthrough, W – column wash.

Highlights.

• Improved yields and purity of LRR proteins and multiprotein complexes through chaperone co-expression, DNA construct design, and use of polycistronic baculovirus.

• Over 300-fold yield increase of SMP protein complex from 0.1 mg/L to 32.6 mg/L.

• SUGT1 co-expression is a highly effective technique for recombinant LRR protein expression and purification.

• Polycistronic baculovirus infection is ideal for production of multiprotein complexes

Abbreviations

CV

column volume

IMAC

immobilized metal ion affinity chromatography

LRR

leucine-rich repeat

MWCO

molecular weight cut-off

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

SEC

size exclusion chromatography

SMP

SHOC2/MRAS/PPP1CA protein complex

TEV

tobacco etch virus proteins