Intein Fusion Systems for Protein Purification

Protein purification can be simplified by fusion of the protein of interest to an affinity tag, such as poly-His tag, MBP tag, etc. (LaVallie and McCoy 1995). However, removal of the affinity tag by protease treatment and subsequent purification steps are often costly and time-consuming. The discovery of inteins and the elucidation of protein-splicing mechanisms provided a novel approach to affinity protein purification (Xu et al. 2000). The ability of an intein to catalyze peptide bond cleavage at its termini is the basis for intein-medi-ated affinity protein purification (Fig. 1). An intein fusion system consists of

S. Chong, M.-Q. Xu (e-mail: [email protected])

32 Tozer Road, New England Biolabs, Inc., Beverly, Massachusetts 01915 USA

Nucleic Acids and Molecular Biology, Vol. 16 Marlene Belfort et al. (Eds.) Homing Endonucleases and Inteins © Springer-Verlag Berlin Heidelberg 2005

Intein Affinity
Fig. 1. A schematic comparison between conventional affinity purification and internmediated affinity purification

a modified intein fused to the C-terminus (C-terminal fusion) or N-terminus (N-terminal fusion) of a target protein. The intein is linked to an affinity tag and is modified to catalyze peptide bond cleavage at one or both of its termini. The affinity tag immobilizes the intein fusion precursor on an affinity column and a subsequent intein-catalyzed cleavage step releases the target protein from the column-bound affinity tag. Therefore, protein purification using an intein fusion system can lead to elution of a single target protein species after just one chromatographic step (Fig. 2).

Several intein fusion systems have been developed to ensure that the intein-mediated protein purification is applicable to as many target proteins as possible (Wood et al. 1999; Xu et al. 2000). These systems differ by the modified intein tags (the modified intein+affinity tag) and by whether the intein tags are at the N-terminus or at the C-terminus. Two other important differences

Intein Affinity

Fig. 2. Single column purification of the recombinant protein Hhal methyltransferase from E. coli expressed from a C-terminal intein fusion vector. Samples taken from different steps during the expression and purification procedures were separated by SDS-PAGE and the gel was stained with Coomassie blue. Lane 1 Protein molecular weight standards (kDa, NEB); lane 2 uninduced cell extract; lane 3 induced cell extract; lane 4 flow through from the load. After loading, the column is washed with column buffer until the protein content of the eluate reaches a minimum. Lane 5 Flow through from the quick DTT flush; lanes 69 the first four fractions of the elution after overnight incubation at 4 °C in the presence of DTT; lane 10 a fraction from the SDS elution of the chitin resin

Fig. 2. Single column purification of the recombinant protein Hhal methyltransferase from E. coli expressed from a C-terminal intein fusion vector. Samples taken from different steps during the expression and purification procedures were separated by SDS-PAGE and the gel was stained with Coomassie blue. Lane 1 Protein molecular weight standards (kDa, NEB); lane 2 uninduced cell extract; lane 3 induced cell extract; lane 4 flow through from the load. After loading, the column is washed with column buffer until the protein content of the eluate reaches a minimum. Lane 5 Flow through from the quick DTT flush; lanes 69 the first four fractions of the elution after overnight incubation at 4 °C in the presence of DTT; lane 10 a fraction from the SDS elution of the chitin resin are the residues at the fusion sites that are favorable for intein cleavage and the conditions under which intein cleavage is induced.

2.1 C-Terminal Fusion System

In a C-terminal fusion system, the C-terminal residue of a target protein is fused to the first N-terminal residue (usually a cysteine) of a modified intein. The intein is mutated to catalyze only an NIS acyl rearrangement at the N-terminal cysteine, which leads to a labile thioester bond linkage between the intein and the target protein (Fig. 3). The modified intein is in turn fused at its C-terminus to a small chitin-binding domain (CBD) from Bacillus circulans as an affinity tag, making it possible to purify the fusion precursor in vitro.

Intein Fusion Protein

Fig. 3. Chemical mechanism of intein-catalyzed peptide bond cleavage at the fusion junctions between a target protein and a modified intein. In a C-terminal fusion (left), the first N-terminal residue, a Cys, of the modified intein is linked to the last C-terminal residue of the N-extein, termed X-1 residue, e.g., the last residue of a target protein. The N/S acyl rearrangement generates a thioester linkage which can be cleaved by an exogenous thiol compound, R-SH, resulting a reactive thioester bond at the C-terminus of the target protein. The C-terminal thioester can be hydrolyzed by H20 to generate the native carboxyl end or attacked by a compound containing a reactive thiol group (SH-R*). This latter reaction is the basis of protein ligation and labeling applications. In an N-terminal fusion (right), the last C-terminal residue, an Asn, of the modified intein is linked to the first N-terminal residue of the C-extein, termed X+1 residue, e.g. the first residue of a target protein. Under certain conditions, the modified intein catalyzes succinimide formation to release the target protein from the intein moiety

Fig. 3. Chemical mechanism of intein-catalyzed peptide bond cleavage at the fusion junctions between a target protein and a modified intein. In a C-terminal fusion (left), the first N-terminal residue, a Cys, of the modified intein is linked to the last C-terminal residue of the N-extein, termed X-1 residue, e.g., the last residue of a target protein. The N/S acyl rearrangement generates a thioester linkage which can be cleaved by an exogenous thiol compound, R-SH, resulting a reactive thioester bond at the C-terminus of the target protein. The C-terminal thioester can be hydrolyzed by H20 to generate the native carboxyl end or attacked by a compound containing a reactive thiol group (SH-R*). This latter reaction is the basis of protein ligation and labeling applications. In an N-terminal fusion (right), the last C-terminal residue, an Asn, of the modified intein is linked to the first N-terminal residue of the C-extein, termed X+1 residue, e.g. the first residue of a target protein. Under certain conditions, the modified intein catalyzes succinimide formation to release the target protein from the intein moiety

Addition of a thiol compound cleaves the thioester bond at the intein N-ter-minus by transesterification. This thiol-induced cleavage when performed in vitro on an affinity matrix (e.g., chitin beads) can result in selective elution of only the target protein from the affinity column (Fig. 2).

The 454-residue See VMA intein was the first intein engineered for protein purification (Chong et al. 1997). As a C-terminal fusion system, the last residue, Asn 454, of the See VMA intein was replaced with an Ala, yielding a modified intein lacking splicing and C-terminal cleavage activities but capable of thioester bond formation at its N-terminus. Addition of thiol compounds, such as dithiothreitol (DTT), P-mercaptoethanol or free cysteine, efficiently cleaved the thioester bond linkage, thereby separating the target protein from the intein tag. Mini-inteins, with smaller molecular weights due to their lack of internal endonuclease domains, were later used for protein purification. These mini-inteins include the 198-residue Mxe GyrA intein and the 134-residue Mth RIR1 intein (Evans et al. 1999b; Southworth et al. 1999). Both inteins were modified by changing the last Asn to Ala to block the splicing reaction and allow the thioester formation at the intein N-termini. DTT and 2-mercaptoethanesulfonic acid (MESNA) were both found to induce the intein cleavage efficiently.

In a C-terminal fusion system, protein synthesis and folding of the upstream target protein determine, to a large extent, the expression and solubility of the whole fusion protein. Some target proteins were found to express well as upstream fusion domains, whereas others expressed better when fused downstream of the modified intein (Chong et al. 1998). To facilitate successful expression of a variety of target proteins, inteins were also modified to generate the N-terminal fusion system.

2.2 N-Terminal Fusion System

In an N-terminal fusion system, the N-terminal residue of a target protein is fused to the last C-terminal residue (usually an Asn) of a modified intein. The cleavage of the peptide bond between the intein and the target protein is achieved by the cyclization of the intein C-terminal Asn (Fig. 3). Based on the proposed protein-splicing mechanism (Paulus 2000), the nucleophilic attack of the Asn P-amide N on the carbonyl C of the peptide bond leads to cyclization of the Asn and concomitant peptide bond cleavage. Spontaneous cyclization of Asn residues coupled to peptide cleavage occurs at a low rate in proteins under physiological conditions (Clarke 1987). A unique feature of inteins is to allow efficient peptide bond cleavage at their C-terminal Asn either independently or in the context of protein splicing. In an N-terminal fusion system for protein purification, the inteins are modified to minimize the C-terminal cleavage in vivo, and, when the fusion protein is immobilized on an affinity matrix, the C-terminal cleavage is induced by addition of a thiol compound or a change in pH and/or temperature.

The See VMA intein was also modified to allow inducible C-terminal cleavage (Chong et al. 1998). The first C-extein residue Cys 455 was mutated to Ala to block the splicing reaction. Since the conserved first and last intein residues Cys 1 and Asn 454 were not mutated, the intein retained the cleavage activities at both its termini. The penultimate His 453 of the intein was then replaced by Gin with the purpose of attenuating in vivo cleavage activity. As expected, the See VMA intein containing this double mutation exhibited very limited in vivo cleavage at either terminus. Remarkably, the modified intein in the purified fusion protein catalyzed efficient in vitro cleavage at both termini in the presence of thiols such as DTT or free cysteine. The cleavage reaction was more efficient at 23 °C than at 4 °C and was inhibited at pH 6.0 and below. It seemed that the thiols induced N-terminal cleavage prior to C-ter-minal cleavage. Exactly how C-terminal cleavage was induced by an upstream cleavage event remained unclear. Nevertheless, the modified See VMA intein has been successfully used as an N-terminal fusion system to purify proteins, including proteins whose purification was unsuccessful in a C-terminal fusion system (Chong et al. 1998).

The N-terminal fusion system also includes several modified mini-inteins: Ssp DnaB, Mtu RecA and Mth RIR1 mini-inteins (Wu et al. 1998b; Evans et al. 1999b; Wood et al. 1999). Both Ssp DnaB (154-residue) and Mtu RecA mini-inteins (168-residue) were derived from full-length inteins by deletion of the endonuclease domains, whereas the Mth RIR1 intein is a natural 134-resi-due mini-intein. To achieve C-terminal cleavage, the first N-terminal Cys of all three mini-inteins was replaced by Ala. In the case of Ssp DnaB and Mth RIR1 mini-inteins, this single substitution was sufficient to allow the C-termi-nal cleavage to occur in a pH- and temperature-dependent manner. The cleavage reaction was most favored at pH 6.0-7.0, inhibited at pH 8.5, and was accelerated with increasing temperatures (4-25 °C). For the Mtu RecA mini-in-tein, additional mutations were made by a selection system to achieve efficient C-terminal cleavage (Wood et al. 1999). The cleavage activity of the modified Mtu RecA mini-intein was remarkably pH- and temperature-sensitive. The cleavage rates increased sharply as the pH was reduced (from 8.5 to 6.0) and the temperature was elevated.

2.3 Choosing an Appropriate Intein Fusion System

One reason for the development of different intein fusion systems (Fig. 4) is to accommodate the expression, folding and purification requirement of a variety of recombinant proteins. The C-terminal intein fusion system has the advantage of generating a target protein with an unmodified N-terminus and, for protein labeling and ligation applications, a reactive thioester bond at its

Protein Purification
Fig. 4. Intein fusion constructs for protein purification

C-terminus. For many target proteins, use of a smaller intein, e.g., the Mxe GyrA mini-intein, results in a higher expression level and yield. Nevertheless, the See VMA intein fusion system has been shown to work well in eukaryotes such as insect cells (Pradhan et al. 1999).

The N-terminal intein fusion system has the advantage of generating a target protein with an N-terminal residue other than Met and, in the case of mini-inteins, inducing intein cleavage without using a thiol compound. In addition, the N-terminal intein fusion allows the translation initiation to be optimized. For instance, the See VMA intein fusion system [Fig. 4, construct (5)] encodes the first ten residues from MBP at its N-terminus to ensure a good transla-tional start (Chong et al. 1998). The Mtu RecA mini-intein system [Fig. 4, construct (8)] encodes the entire MBP as an N-terminal domain to enhance protein expression and solubility. The only N-terminal fusion system that uses a thiol compound to induce C-terminal cleavage is that of the modified See VMA intein. However, this system cannot generate a target protein with an N-terminal Cys, as Cys at the fusion site (X+1 position) results in protein splicing rather than the C-terminal cleavage. Other N-terminal fusion systems contain mini-inteins (Fig. 4), which allow induction of the intein C-terminal cleavage by a shift in pH and temperature. Furthermore, these systems are capable of producing target proteins with an N-terminal Cys, useful for protein ligation and other applications.

Fusion of a target protein upstream or downstream of a modified intein tag would change the dynamics of protein synthesis and folding. This may explain the observation that the N- and C-terminal intein fusions often gave quite different expression and purification results for a given target protein (Chong et al. 1998). There are no sufficient data to support one fusion system over the other for all target proteins. Successful expression and purification of a target protein depend not only on its location in the fusion protein, but also on the type of the intein tag. In addition, the residue at the fusion junction between the target protein and the modified intein-tag sometimes plays a critical role.

2.4 Choosing an Appropriate Residue at the Fusion Junction

The active site of an intein is located at its termini. Fusion of a target protein to either terminus of an intein places the N- or C-terminal residue of the target protein adjacent to the catalytic residues of the intein. In an intein fusion protein, the target protein residues adjacent to the intein N-terminal and C-terminal residues are named minus 1 residue (X-1) and plus 1 residue (X+1), respectively (Figs. 3 and 4). The X"1 and X residues exhibit more effect on the intein cleavage efficiency than any other residues further away from the intein cleavage sites. The most favorable and unfavorable residues for intein cleav age in different fusion systems are listed in Table 1. The complete lists including all 20 amino acids for certain intein fusion systems can be found in the original references (see Table 1). It should be noted that these data were generated by examining a particular test protein. Therefore, the conclusion may not be completely applicable to other target proteins. Nevertheless, the effect of some amino acid residues at the fusion sites has been proven consistent in multiple target proteins. For instance, Asp as the X"1 residue invariably results in almost complete in vivo cleavage in a C-terminal fusion system, whereas Pro as the X-1 or X+1 residue leads to inhibition of the intein cleavage in the C-terminal or N-terminal fusion system, respectively. For a target protein with a C-terminal Asp or Pro, the N-terminal fusion system rather than the C-ter-minal fusion system should be used in order to obtain a recombinant protein with a native sequence. If addition of one or more extra residues at the target protein termini is not expected to have an adverse effect on its structure and activity, both N-terminal and C-terminal fusion systems can be used. Addition of one or two residues favorable for intein cleavage (Table 1) can often enhance significantly both the expression level of the fusion protein and the efficiency of intein cleavage.

2.5 Conditions for Intein Cleavage

A unique advantage of the intein fusion system is the separation of a target protein from its affinity tag by inducing intein cleavage under mild conditions. Different intein fusion systems require different cleavage conditions (Table 1). Thiol reagents such as DTT and (less expensive) P-mercaptoethanol can induce efficient cleavage at around pH 8.0. For recombinant proteins that are sensitive to reducing reagents such as DTT, free cysteine can be used, which can attach to the C-terminus of the target protein through a peptide bond after cleavage. This modification of the target protein C-terminus occurs only when the C-terminal fusion system is used. The target protein is not modified by free cysteine when the See VMA intein is used in the N-terminal fusion system. After inducing intein cleavage, DTT also modifies the target protein by forming a thioester bond with the C-terminal residue. At pH 8.0, the DTT tag is not stable in solution and is normally hydrolyzed. For applications such as protein labeling and ligation, a stable and reactive thioester bond at the C-terminus of a target protein is desirable. It was found that MESNA could also form a reactive thioester that allows for more efficient ligation reaction than when DTT was used (Evans and Xu 1999). Since MESNA induces more efficient cleavage by the modified Mxe GyrA mini-intein than the See VMA intein, the Mxe GryA mini-intein and MESNA are recommended for purification of recombinant proteins to be used for protein labeling and ligation.

Table 1. Appropriate residues at the intein fusion sites and the conditions to induce intein cleavage

Intein fusion constructs

Modified intein tags

Residues preferred at fusion sites

Residues to be avoided at fusion sites

Conditions for intein cleavage

C-terminal fusion

See VMA intein CBD Mxe GyrA mini-intein CBD Mth RIR1 mini-intein CBD Ssp DnaE mini-intein (CBD)-GFP

Gly, Ala Met, Tyr, Phe Gly, Ala Gly, Ala

Asp, Pro

Asp, Val, Pro

Addition of DTT, free cysteine, or MESNA, pH 8-8.5 at 4-25 °C

N-terminal fusion

See VMA intein (CBD)

Ala, Met, Gly

Ser, Cys, Pro

Addition of DTT, free cysteine, pH 8-8.5 at 4-25 °C

CBD-Mth RIR1 mini-intein MBP-Mtu RecA mini-intein

Cys, Cys-Arg Cys

Met, Cys

Pro, Lys, Arg, lie, Leu, Asn, Gin Not determined Not determined

pH 6.0 at 4-25 °C

Note: In an intein fusion protein, the target protein residues adjacent to the intein N- and C-terminal residues are named minus 1 residue (X-1) and plus 1 residue (X+1), respectively (Figs. 3 and 4). In the C- and N-terminal fusion systems, residues preferred at fusion sites listed in this table are the minus 1 or plus 1 residues, respectively. The complete lists including all 20 amino acid residues at fusion sites for certain intein fusion systems can be found in the original references (Chong et al. 1997,1998; Southworth et al. 1999; Zhang et al. 2001)

2.6 Intein Fusion Systems for High-Throughput and Large-Scale Applications

Though intein fusion systems have simplified affinity purification, intein fusion proteins have to be expressed in a soluble, correctly folded form in order for intein-mediated purification to be effective. Fusion proteins overexpressed in E. coli sometimes misfold resulting in very low expression or inclusion bodies. It would be advantageous if the level of soluble expression of an intein fusion protein is known before the cell lysis and electrophoretic analyses steps. This is especially desirable if a large number of recombinant proteins are to be expressed and purified. The intein fusion system potentially useful for high-throughput protein purification is a C-terminal fusion system consisting of a modified Ssp DnaE mini-intein whose C-terminus is fused to green fluorescent protein (GFP) as an expression and solubility reporter (Zhang et al. 2001; Fig. 4). A CBD domain is inserted between the N- and C-terminal domains of the Ssp DnaE mini-intein, which is modified to allow thiol-inducible cleavage at its N-terminus. A variety of target proteins were expressed and purified in this intein-GFP fusion system (Zhang et al. 2001). The data suggested a positive linear correlation between GFP fluorescence of induced cultures and the final target protein yields after intein-mediated purification. The intein-GFP system can be used to screen for culture conditions, strain/vector variants, mutations, etc., that improve soluble expression of a particular fusion protein. Once an optimized condition is found, the target protein can be directly purified on a single column. Alternatively, the system can be used to screen for soluble expression of a large number of recombinant proteins for rapid purification in high-throughput experiments.

Since intein fusion systems purify proteins on an affinity matrix and remove the affinity tag in the same chromatographic step, it has the potential to significantly reduce recovery costs for the industrial production of recombinant proteins. Studies have shown that intein fusion proteins remain stable in vivo in E. coli during high-cell-density fermentation and can catalyze efficient peptide bond cleavage (Sharma et al. 2003). It is therefore feasible to use intein fusion systems for industrial-scale production of recombinant proteins.

0 0

Post a comment