SPI “sandwich”: Combined SUMO‐Peptide‐Intein expression system and isolation procedure for improved stability and yield of peptides

Abstract

Recombinant peptide production in Escherichia coli is often accomplished through cloning and expression of a fusion protein. The fusion protein partner generally has two requirements: (a) it contains an affinity tag to assist with purification and (b) it can be cleaved off to leave only the desired peptide sequence behind. Common soluble fusion partners include small ubiquitin‐like modifier protein (SUMO), maltose‐binding protein (MBP), glutathione S‐transferase (GST), or intein proteins. However, heterologously expressed peptides can suffer from proteolytic degradation or instability. This degradation can pose a major issue for applications requiring a large amount of purified peptide, such as NMR structural assignments or biochemical assays. Improving peptide yield by testing various expression and isolation conditions requires a significant amount of effort and may not lead to improved results. Here, we cloned and expressed four different peptides as SUMO fusion proteins. These peptides (lactococcin A, leucocin A, faerocin MK, neopetrosiamide A) were truncated during expression and isolation as SUMO fusions, resulting in low yields of purified peptide. To prevent this degradation and improve yield, we designed a new expression system to create a “sandwiched” fusion protein of the form: His₆‐SUMO‐peptide‐intein (SPI). These sandwiched peptides were more stable and protected against degradation, resulting in improved yields (up to 17‐fold) under a set of standard expression and isolation procedures. This SPI expression system uses only two commercially available vectors and standard protein purification techniques, and therefore may offer an economical and facile route to improve yields for peptides that undergo degradation.

Abbreviations

DTT

dithiothreitol

FaeMK

faerocin MK

HPLC

high‐performance liquid chromatography

LcnA

lactococcin A

LeuA

leucocin A

MALDI

matrix‐assisted laser desorption ionization

MS

mass spectrometry

NeoA

neopetrosiamide A

SDS‐PAGE

sodium dodecyl sulfate‐polyacrylamide gel electrophoresis

SPI

SUMO‐peptide‐intein

SUMO

small ubiquitin‐like modifier

TOF

time of flight

1. INTRODUCTION

Ribosomally synthesized peptides are an important class of biomolecules that are ubiquitous in nature and act as signaling molecules, ¹ , ² antimicrobials, ³ , ⁴ , ⁵ mammalian hormones, ⁶ , ⁷ , ⁸ or toxins, ⁹ , ¹⁰ among many other roles. Many peptides have properties relevant to human health or industry, which has led to a large body of research dedicated to the study of these compounds. Some producer organisms are difficult to culture or produce tiny amounts of the peptide of interest, and so recombinant expression of the desired peptide in Escherichia coli is a common method to isolate these molecules for further study. ¹¹

Heterologous expression of peptides is often achieved by cloning the gene for the peptide of interest into a vector that encodes a soluble fusion protein. As a result of the relatively small size of peptides, it is desirable to isolate the peptide without the fusion protein attached in order to prevent any changes to structure or activity. ¹¹ This necessitates the use of a cleavable fusion protein that does not leave behind any residual amino acid residues on the peptide of interest. Fusion proteins should also contain an affinity tag to allow for separation from the cell lysate. ¹²

While there are many fusion proteins, affinity tags, and cleavable linkers reported in the literature, those most useful to researchers have a commercially available vector and a routine isolation procedure. ¹³ Most cleavable fusion protein systems are designed with the affinity tag and soluble fusion protein at the N‐terminus, followed by a protease recognition site, and then the peptide of interest. Common soluble fusion proteins include small ubiquitin‐like modifier (SUMO), maltose‐binding protein (MBP), thioredoxin, inteins, or glutathione S‐transferase (GST). ¹⁴ These proteins usually include a supplementary affinity tag to aid in isolation of the protein. SUMO fusion proteins do not require an additional cleavage sequence, as the SUMO protease recognizes the three‐dimensional structure of the SUMO protein and cleaves after a diglycine sequence at the C‐terminal end of the SUMO tag. ¹⁵ Inteins are self‐cleaving proteins that initiate cleavage with the addition of excess thiol, such as dithiothreitol (DTT), and may have an inserted chitin‐binding domain for affinity purification. ¹⁶ Other fusion proteins usually include a protease cleavage sequence, such as those for enterokinase, thrombin, tobacco etch virus protease, or Factor Xa protease, and may leave behind additional amino acid residues on the target protein after cleavage. ¹³ Creating a fusion protein with the peptide of interest at the N‐terminus and fusion partner at the C‐terminus can be accomplished using a C‐terminal intein. ¹⁷ This requires the first residue of the peptide of interest to be methionine. Other cleavage sequences listed above cannot readily be used at a fusion protein C‐terminus without leaving behind extra residues on the peptide of interest.

While many vectors and fusion protein systems exist, protein stability issues can still hamper attempts to isolate a peptide of interest. A common issue is degradation or truncation of the overexpressed protein during cellular expression. This may be due to proteolytic degradation, instability of the protein of interest, toxicity to the cell, or insolubility. ¹⁸ While expression optimization can lead to improved yields, this can be extremely time consuming as there are a large number of factors to potentially modify, including induction temperature, inducer concentration, optical density, induction time, buffer conditions, isolation temperature, protease inhibitors, detergents, solubility additives, or isolation procedures. ¹³ , ¹⁸

Our group and others have routinely cloned, expressed, and isolated both proteins ¹⁹ , ²⁰ , ²¹ and small peptides ²² , ²³ , ²⁴ as SUMO fusions in E. coli. While overproduction of proteins using this system has been successful in our group, we have observed, as described in this work, that SUMO fusions of small peptides including lactococcin A, leucocin A, faerocin MK, and neopetrosiamide A can be truncated during expression, resulting in low yield of the purified peptide.

Rather than trying to optimize expression and isolation conditions for each of these four peptides, we designed a “sandwiched” fusion protein system of the form: His₆‐SUMO‐peptide‐intein, which we called SPI. This fusion system was derived from a pET‐SUMO vector, containing an N‐terminal hexa‐histidine tag followed by the SUMO protein and then the peptide of interest. We then amplified the His₆‐SUMO‐peptide DNA sequence with polymerase chain reaction (PCR) and cloned it into a pTXB1 vector, which introduced a C‐terminal intein to the fusion protein. The four peptides were cloned and expressed under the same set of conditions, and then isolated according to a new procedure. All four peptides were purified with improved yields compared to the original SUMO fusion system, suggesting broad applicability of this method to different classes of peptides.

2. RESULTS AND DISCUSSION

2.1. Four SUMO‐peptide fusions with truncations and poor yield

We were interested in producing four peptides with varying structures and properties. Leucocin A (LeuA) and faerocin MK (FaeMK) are type IIa bacteriocins, a class of ribosomally synthesized antimicrobial peptides, produced by Leuconostoc gelidum UAL187 and Enterococcus faecium M3K31, respectively. ²⁵ , ²⁶ This class of bacteriocins have a characteristic YGNGV/L motif near their N‐termini, at least one disulfide bond, and a common fold containing an alpha helix and two beta strands. ³ Lactococcin A (LcnA) is a class II bacteriocin produced by Lactococcus lactis that lacks cysteine residues and disulfide bonds. ³ , ²⁷ Neopetrosiamide A (NeoA) is a peptide isolated from the marine sponge Neopetrosia sp. ²⁸ , ²⁹ that contains three disulfide bonds and has potential anti‐metastatic properties. ³⁰ These peptides range in sizes from ~3 to 6 kDa (Table S1).

We initially embarked on peptide production as SUMO fusion proteins using a standard isolation method (Figure 1a). Each peptide was fused to the His‐tagged SUMO protein and overexpressed in E. coli BL21(DE3). However, during isolation of each of these SUMO‐peptide fusion proteins, we observed His‐tagged SUMO fusion proteins of smaller molecular weight than expected (Figures 1b,c and S1–S3). This indicated that truncations were occurring in the C‐terminal region of the fusion proteins. Cleavage sites in each of the four peptide sequences were proposed based on the molecular weights of observed fragments (Figure 1d). This degradation led to low yields of pure peptide in all cases (Table 1), even with the addition of protease inhibitor cocktail tablets to the lysis buffer. Lowering the temperature during induction of the fusion protein expression in the cell cultures did not prevent degradation of target peptides either (data not shown). We initially explored other cell lines for peptide production and transformed the pET‐SUMO‐LeuA plasmid into T7 Express competent E. coli, another protease deficient cell line, but observed the same truncations (Figure S4).

FIGURE 1.

Isolation of SUMO‐peptide fusion proteins reveals C‐terminal truncations. (a) General isolation procedure for SUMO fusion proteins. (b) SDS‐PAGE of fractions from the first Ni‐NTA column during SUMO‐LeuA isolation. MW = molecular weight marker. FT = column flow through containing the cellular lysate. 20–500 = elution fractions containing 20–500 mM imidazole. (c) MALDI‐TOF MS of SUMO‐LeuA after dialysis. (d) Sequences of peptides used in this study. Putative cleavage sites observed during SUMO fusion expression are indicated by the red, vertical lines. Cleavage sites were proposed based on the m/z values observed in MALDI‐TOF MS that matched protein or peptide sequences with truncations from the C‐terminus

TABLE 1.

Yield of purified peptides obtained from fusion systems in this study

Peptide	Yield as SUMO fusion (mg/L)	Yield as SUMO‐intein fusion (mg/L)
LeuA	0.1	1.7
FaeMK	0.2	2.0
NeoA	0.5	1.4
LcnA	<0.1	0.7

2.2. SUMO‐peptide‐intein fusion system design and cloning

Since the addition of protease inhibitor tablets did not prevent truncation of the fusion proteins, which was observed even in the early stages of the isolation process, we hypothesized that degradation might have begun in the cells during expression. Rather than attempting to find the best expression conditions for each of the four peptides, we instead designed a universal expression system that would prevent this degradation from occurring in the first place. The proposed expression system “sandwiches” the peptide between two cleavable fusion proteins: an N‐terminal His₆‐SUMO protein and a C‐terminal intein (Figure 2a). We hypothesized that the increase in molecular weight and bulk of the C‐terminal intein may prevent degradation of the internal peptide sequence in this SUMO‐peptide‐intein (SPI) fusion protein.

FIGURE 2.

Design of “sandwiched” SUMO‐peptide‐intein (SPI) fusion protein. (a) Cartoon depiction of SPI fusion protein from N‐ to C‐ terminus. (b) Abbreviated intein cleavage mechanism. The first residue of the intein, Cys, catalyzes an N to S acyl shift of the peptide backbone after the last residue of the peptide sequence. Addition of an external thiol (usually DTT) results in a transthioesterification reaction that cleaves off the C‐terminal intein. The remaining thioester bond is then hydrolyzed to leave a C‐terminal carboxylic acid on the peptide of interest

We used PCR to amplify the His₆‐SUMO‐peptide DNA sequences and insert appropriate restriction sites for cloning into the expression vector pTXB1. This vector encodes a C‐terminal intein with an inserted chitin‐binding domain affinity tag that can be self‐cleaved upon the addition of DTT to leave behind an N‐terminal protein of interest with no extra residues (Figure 2b). The genes of the SPI fusion proteins were sequenced and the recombinant plasmids were transformed into E. coli BL21(DE3).

2.3. Expression and isolation of FaeMK, LeuA, and NeoA

Each of the four constructs was expressed under the same set of conditions. Induction at low temperature for a longer time appeared to be the best condition for minimizing premature intein cleavage inside the cells during expression (Figure S5). The fusion proteins were isolated and the desired peptides were purified following a new procedure, as described below (Figure 3a).

FIGURE 3.

SPI fusion system and its use to isolate FaeMK. (a) General isolation procedure for SPI fusion proteins. (b) SDS‐PAGE of fractions from the first Ni‐NTA column during SUMO‐FaeMK‐intein isolation. MW = molecular weight marker. FT = column flow through containing the cellular lysate. 20–500 = elution fractions containing 20–500 mM imidazole. (c) MALDI‐TOF MS of SUMO‐FaeMK and the cleaved intein after incubation in 100 mM DTT

The lysis buffer was modified from previous SUMO fusion isolation procedures ¹⁹ , ²³ to contain 300 mM NaCl (from 150 mM) because the higher salt content better prevented precipitation during isolation. We did not need to use Protease Inhibitor Cocktail tablets during any of the SPI isolations, as the expression system “sandwich” was found to be effective enough at preventing degradation during isolation. A larger volume of buffer during cell lysis was also found to improve solubility of the fusion protein (~15 mL buffer/g cell pellet). A Ni‐NTA affinity column was chosen as the initial chromatographic step because Ni‐NTA resin has a higher binding capacity than chitin resin. ³¹ The elution fractions from this column were examined by both matrix‐assisted laser desorption ionization‐time of flight mass spectrometry (MALDI‐TOF MS) and sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE), which confirmed the presence of each the desired SPI proteins (Figures 3b,c and S5–S7).

After isolation by Ni‐NTA affinity chromatography, the intein was cleaved off SPI fusion proteins by the addition of 100 mM DTT and gentle shaking overnight. We chose to cleave the intein first, since SUMO cleavage would have required an additional buffer exchange step to remove imidazole. Chitin resin was also added directly to the cleavage mixture, allowing for a longer binding time of the cleaved intein. The manufacturer (New England Biolabs) recommended 50 mM DTT and 2 days for cleavage. However, 100 mM DTT was used as we found that this concentration allowed for complete cleavage of the intein after ~15 hr in all four cases (Figures 3c, S6b, and S7b). The mixture was then passed through a fritted column, which retained the chitin resin with the bound intein, while the His₆‐SUMO‐peptide fusion protein was able to pass through the column. Size exclusion chromatography was then used for buffer exchange and removal of DTT and imidazole. This step could be exchanged with dialysis, but we preferred size exclusion because of the faster time and better purification. The SUMO tag was then cleaved by addition of the SUMO protease, and the SUMO tag and protease were removed from the mixture by a second Ni‐NTA affinity column. The peptide obtained in the flow through and wash fractions was further purified by high‐performance liquid chromatography (HPLC) (Figure S8).

The NeoA and FaeMK fusion proteins were purified using the standard method above and resulted in improved yields of 1.4 (threefold) and 2.0 mg/L (10‐fold), respectively (Table 1). However, in the case of LeuA, we noticed that after cleavage of the intein from the His₆‐SUMO‐LeuA‐intein fusion protein, LeuA appeared to be degrading during purification. This may be because LeuA has been reported to have low stability at neutral to high pH. ²⁵ To prevent this degradation, we modified the buffer conditions. Elution of the protein from the first Ni‐NTA column, intein cleavage, and size exclusion chromatography were all performed at pH 6 (Figure S6b). Surprisingly, the intein cleaved readily at pH 6 for all four of the SPI fusion proteins, even though the recommended pH for cleavage is 8. ³² The pH was raised to 7.8 for the proteolysis step because SUMO tag cleavage was incomplete at pH 6. After the second Ni‐NTA column, MALDI‐TOF MS of LeuA showed two distinct m/z peaks, one for the desired peptide at 3938, and another at 4071 (+133 m/z) (Figure S6c). We hypothesized that the peak with the larger m/z value corresponded to the LeuA peptide with a C‐terminal thioester (or ester) of DTT still intact, as isolation at pH 6 may have slowed hydrolysis to the desired carboxylic acid (Figure 2b). Indeed, basification to pH 10 and stirring for 15 min completely hydrolyzed the thioester without significant degradation of the peptide (Figure S6d). The solution was then acidified to pH 4, concentrated, and purified by HPLC (Figure S8) to yield 1.7 mg/L of pure LeuA (17‐fold improvement).

2.4. Modified isolation procedure for LcnA

We initially attempted to isolate LcnA with the standard SPI isolation procedure used for NeoA and FaeMK; however, it consistently resulted in poor yields of the purified peptide. The protein appeared to be aggregating during the chitin column passage and elution was difficult. We designed an alternative isolation procedure that produced good results and may be broadly applicable when chitin affinity chromatography is undesired or unavailable (Figure 4a).

FIGURE 4.

Alternative isolation procedure for LcnA to avoid a chitin affinity column. (a) General isolation procedure for SUMO‐LcnA‐intein fusion protein to avoid the use of chitin resin. (b) SDS‐PAGE of samples during SUMO‐LcnA‐intein isolation. MW = molecular weight marker. Pel = insoluble pellet after centrifugation of the cell lysate. Ni1 FT = flow through of the first Ni‐NTA column. Ni1 20 = 20 mM imidazole wash of first Ni‐NTA column. Ni1 500 = elution fraction of the first Ni‐NTA column containing 500 mM imidazole. DTT = sample after intein cleavage with 100 mM DTT overnight. Dial = sample after first dialysis step. Ni2 FT = flow through of the second Ni‐NTA column. Ni2 20 = 20 mM imidazole wash of second Ni‐NTA column. Ni2 500 = elution fraction of the second Ni‐NTA column containing 500 mM imidazole. (c) MALDI‐TOF MS of LcnA and SUMO tag after SUMO cleavage

Ni‐NTA affinity chromatography was used to initially isolate the His₆‐SUMO‐LcnA‐intein protein (Figure 4b). This was followed by intein cleavage overnight in 100 mM DTT but without chitin resin added. This mixture of His₆‐SUMO‐LcnA and cleaved intein was then dialyzed to remove DTT and imidazole. The His₆‐SUMO‐LcnA protein was then isolated with a second Ni‐NTA column, allowing the intein to separate in the column flow through (Figure 4b). The desired fusion protein was eluted, followed by dialysis to remove the imidazole. The SUMO tag was then cleaved by the addition of SUMO protease, and the protease and SUMO tag were separated from the LcnA peptide with a third Ni‐NTA column. The peptide was then purified with HPLC (Figure S8), resulting in a yield of 0.7 mg/L (>sevenfold improvement), which was lower than the other peptides, but still much better than the initial SUMO fusion isolation procedure which was unable to isolate LcnA in any appreciable amount (Table 1).

2.5. Factors affecting intein cleavage

Potential issues others may encounter with this expression system may be due to difficulty with intein cleavage. In this work, we found that for all four of our SPI fusion proteins, the intein cleaved readily in 100 mM DTT, even at pH 6, which we did not expect. The last residues of LeuA, FaeMK, NeoA, and LcnA are Trp, Arg, Cys, and His, respectively. The intein encoded in the pTXB1 vector is predicted to have reduced cleavage efficiency with Trp and Cys, and His has been shown to cause premature cleavage of the intein during cellular expression. ³² However, these cleavage characteristics are predictions and cleavage efficiency varies with every protein, as demonstrated here.

We did, however, encounter issues with the intein cleaving prematurely. During isolation of FaeMK, LcnA, and NeoA (Figures 3b, 4b, and S7a,b), we observed a small amount of the intein‐less fusion proteins (His₆‐SUMO‐peptide) and their truncation(s) in SDS‐PAGE and MS, in addition to the desired SPI proteins. The presence of the intein‐less fusion protein (His₆‐SUMO‐peptide) and its truncations was most apparent for the LcnA construct (Figure 4b). To determine whether truncations of the His₆‐SUMO‐peptide proteins occurred before or after intein cleavage, the flow through of the first Ni‐NTA column for the His₆‐SUMO‐LcnA‐intein construct was collected. This flow through contained the cell lysate and any prematurely cleaved intein protein, and was then passed over a chitin column. SDS‐PAGE and MALDI‐TOF MS analysis of the proteins retained by the chitin resin indicated the presence of the intein mass only, without any extra N‐terminal peptide residues still attached (Figure 5). This suggests that truncations of the His₆‐SUMO‐LcnA protein only occurred after the intein was prematurely cleaved off in vivo, and that the presence of the intein on the fusion protein C‐terminus prevented this degradation.

FIGURE 5.

Truncations during expression of SUMO‐LcnA‐intein fusions, and intact C‐terminal inteins as protection against these truncations. (a) Chitin column experiment to determine whether a C‐terminal intein protects against degradation of SUMO‐LcnA fusion proteins. The flow through eluted from a Ni‐NTA column that retained and isolated the SUMO‐LcnA‐intein protein was collected. This flow through fraction contained the cellular lysate and free intein proteins that cleaved prematurely during cellular expression. The flow through was then passed over a secondary chitin column, and the elution fractions of this secondary chitin column are shown in SDS‐PAGE to examine the molecular weight of the chitin‐bound intein proteins. MW = protein molecular weight ladder. Ni FT = flow through of first Ni‐NTA column before the chitin column. Chi FT = flow through of chitin column. W1 – W4 = wash fractions of chitin column. Chi bead = chitin resin after four wash fractions. (b) MALDI‐TOF MS of chitin resin bead sample after four wash fractions. The m/z value for the intein is the expected value and does not have additional N‐terminal peptide residues attached

It may be possible that the presence of a C‐terminal intein alone could prevent degradation of peptides fused to its N‐terminus. However, this approach would necessitate that the peptide of interest begins with an N‐terminal Met residue, which limits the universality of the expression system. Furthermore, this system could be potentially implemented with other inteins besides the mutant GyrA intein encoded in the pTXB1 vector. ³³ Other inteins may offer improved cleavage characteristics, varying pH or salt requirements, N‐terminal peptide sequence compatibilities, or temperature constraints. As a result, the identity of the intein chosen may significantly influence the success of the SPI strategy for a particular peptide of interest. Additionally, use of the SPI strategy may not be limited to just improving yields of peptides that undergo degradation. Osunsade et al. used a similar strategy to produce a human histone H1 protein, which had previously been difficult due to the intrinsically disordered and basic C‐terminal domain that led to insolubility or truncation. ³⁴ Taken together, their results and ours suggest a “sandwiching” strategy may be generally effective to prevent degradation of proteins or peptides of varying sizes during heterologous expression.

As this method requires large amounts of reducing agent, disulfide bonds are reduced during purification, and so correct folding and reoxidation would have to occur after purification to reestablish these bonds. However, this is also the case for SUMO fusion protein isolation as the SUMO protease works best with 1 mM DTT added.

In this study, we demonstrated that expression as SPI fusions led to sizable increases in purified peptide yields for FaeMK, LeuA, LcnA and NeoA (Table 1). The inclusion of a C‐terminal intein in this “sandwiched” fusion protein appeared to be protective against degradation, indicating that this system may lead to an increase in yield for peptides undergoing degradation, even without fine‐tuning expression conditions. Both plasmids used in cloning this expression system are commercially available, and the isolation procedures use only common protein purification techniques and materials, making this method potentially very useful for other peptide researchers that encounter degradation during heterologous expression.

3. MATERIALS AND METHODS

3.1. Cloning

The pET‐SUMO expression vector (Invitrogen) encoding neopetrosiamide A fused to the His‐tagged SUMO protein was described previously, ²³ and cloning of genes for FaeMK, LeuA, and LcnA into this vector followed the described protocol (Appendix S1).

Forward and reverse primers containing NdeI and SpeI restriction sites, respectively, were used to amplify the SUMO‐peptide genes from the pET‐SUMO vectors (Table S2). The PCR products were then cloned into the NdeI and SpeI restriction sites of the pTXB1 expression vector (New England BioLabs), thus ensuring that each SUMO‐peptide was in frame with the C‐terminal intein with no extra residues being attached to the peptide before the +1 Cys of the intein. The resulting plasmids were sequenced to confirm that the genes were correct and the gene products were in frame with the intein, and then transformed into E. coli BL21(DE3). We can provide all His₆‐SUMO‐peptide‐intein plasmids produced in this study to others upon request.

3.2. Expression and induction conditions

For overexpression of the SUMO‐peptide fusion proteins, 50 mL of Luria Broth media (LB) (10 g of tryptone, 5 g of yeast extract, 10 g of NaCl) was inoculated with the E. coli transformant and the cells were grown overnight at 37°C with shaking at 225 rpm and kanamycin (50 μg/mL) as a selective pressure. 20 mL of the overnight culture was added to 500 mL of LB media and cells were grown with shaking to an optical density (OD₆₀₀) of 0.5 at 37°C. Isopropyl β‐d‐1‐thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM, and culture growth was continued with shaking for 4 hr at 37°C. Cells were harvested by centrifugation (5,000 g for 10 min at 4°C) and the pellets were stored at −80°C.

For overexpression of the SPI fusion proteins, the same procedure described above was used with the following changes. Ampicillin (100 μg/mL) was used as a selective pressure during cell growth. Cells were grown to an OD₆₀₀ of 0.6, and then the culture was placed on ice for 5 min. IPTG was added to a final concentration of 0.1 mM, and culture growth was continued for 18 hr at 15°C.

3.3. Isolation procedures

Isolation procedures for SUMO‐peptide fusion proteins have been previously described, ¹⁹ , ²³ and the detailed procedure used in this study is included in Appendix S1.

For isolation of SUMO‐FaeMK‐intein or SUMO‐NeoA‐intein proteins, frozen cell pellets were resuspended in ice cold high salt lysis buffer (20 mM Tris–HCl, pH 7.8, 300 mM NaCl) with 5 mM imidazole added and gently vortexed to resuspend. Cells were lysed by sonication while kept on ice, then DNase I (Thermo Scientific, 1 U) was added, and the lysate kept on ice for 15 min. The cellular debris was removed by centrifugation (20,000 × g for 30 min at 4°C), and then the desired fusion protein was isolated from the clarified supernatant by loading on Ni‐NTA resin (Qiagen, 3 mL of resin per 500 mL of cell culture) at 4°C. The resin was washed with 10 column volumes of high salt lysis buffer containing 20 mM imidazole, and then the fusion protein was eluted with 40–500 mM imidazole in high salt lysis buffer. Eluted fractions were analyzed by SDS‐PAGE and/or MALDI‐TOF MS and the samples containing the protein of interest were pooled together. The pooled fractions had 100 mM DTT added and were pH adjusted to 7.8 (if needed), and then pre‐washed chitin resin (New England BioLabs, 5 mL per 500 mL of cell culture) was added. The sample was gently rocked overnight at 4°C to cleave the intein. The sample was then passed through a fritted column, and the chitin resin was washed with five column volumes of lysis buffer (20 mM Tris–HCl, pH 7.8, 150 mM NaCl). The pooled flow through and wash fractions were loaded on a size exclusion column (Sephadex G‐15) and equilibrated in lysis buffer. Fractions containing the SUMO‐peptide fusion protein were pooled. The fusion protein was digested with His‐tagged SUMO protease (McLab, South San Francisco, CA, 2000 units) at 4°C for 4 hr to cleave the SUMO tag. The SUMO tag and protease were then removed with a second Ni‐NTA column at 4°C. The flow through and wash fraction were pooled and then the crude peptide was lyophilized and redissolved in an acetonitrile/water mixture with 0.1% trifluoroacetic acid . The peptide was further purified using HPLC (Appendix S1) and then lyophilized to obtain a final yield of purified peptide.

For isolation of SUMO‐LeuA‐intein, the same procedure as above was followed, except for the following modifications. The Ni‐NTA elution buffers (40–500 mM imidazole) were adjusted to pH 6. Intein cleavage and size exclusion buffers were adjusted to pH 6. After size exclusion, the pH of the SUMO‐LeuA sample was adjusted to pH 7.8, and SUMO protease cleavage was allowed to proceed for 1 hr at 4°C. After the second Ni‐NTA column, the pooled flow through and wash fractions were adjusted to pH 10 and stirred at RT for 15 min to hydrolyze any residual thioester. After 15 min, the pH was adjusted to 4 and the sample was lyophilized before HPLC purification.

For isolation of SUMO‐LcnA‐intein without chitin resin, cells were lysed as described above and the fusion protein was isolated using Ni‐NTA resin with 40–500 mM imidazole in high salt lysis buffer (20 mM Tris–HCl, pH 7.8, 300 mM NaCl). Pooled fractions had 100 mM DTT added and were pH adjusted to 7.8 if needed. The sample was gently rocked overnight at 4°C to cleave the intein. The sample was then dialyzed against high salt lysis buffer for 3 hr at 4°C. The SUMO‐LcnA protein was then isolated with a second Ni‐NTA column. The eluted fractions were then dialyzed against lysis buffer (20 mM Tris–HCl, pH 7.8, 150 mM NaCl) with 1 mM DTT for 3 hr at 4°C to remove the imidazole. The fusion protein was digested with His‐tagged SUMO protease at 4°C for 4 hr to remove the SUMO tag. The SUMO tag and protease were then removed with a third Ni‐NTA column at 4°C, and the peptide was collected in the column flow through and further purified as described above.

AUTHOR CONTRIBUTIONS

Tess Lamer: Conceptualization (equal); data curation (lead); investigation (equal); methodology (lead); writing – original draft (lead); writing – review and editing (equal). Marco J. van Belkum: Conceptualization (equal); investigation (equal); methodology (equal); writing – review and editing (equal). Anjalee Wijewardane: Investigation (supporting); writing – review and editing (supporting). Sorina Chiorean: Investigation (supporting); methodology (supporting); writing – review and editing (equal). Leah A. Martin‐Visscher: Funding acquisition (equal); supervision (supporting); writing – review and editing (equal). John C. Vederas: Conceptualization (equal); funding acquisition (lead); investigation (supporting); methodology (supporting); project administration (lead); resources (lead); supervision (lead); writing – review and editing (equal).

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.

Supporting information

Appendix S1 Supporting Information

Lamer T, van Belkum MJ, Wijewardane A, Chiorean S, Martin‐Visscher LA, Vederas JC. SPI “sandwich”: Combined SUMO‐Peptide‐Intein expression system and isolation procedure for improved stability and yield of peptides. Protein Science. 2022;31(5):e4316. 10.1002/pro.4316