It was evident at the start that inteins had chemically reactive amino acids at both splice junctions - amino acids normally found in enzyme active sites.
This led to the hypothesis that folding of the intein brought the conserved residues at both splice junctions into close proximity to form the protein-splicing active site and also brought together the exteins for ligation. We still do not know whether there is a single protein-splicing active site or the relationship between the potentially different active sites for each step in the protein-splicing pathway. All experiments with split inteins fail to detect cleavage in single precursor fragments, confirming that the both ends of the intein are necessary for activity at either splice junction (Lew et al. 1998; Mills et al. 1998; Shingledecker et al. 1998; Southworth et al. 1998; Wu et al. 1998a,b; Martin et al. 2001; Nichols et al. 2003).
Although we know the major players in the splicing reaction (the nucle-ophiles), the assisting groups have yet to be fully identified, partly because different amino acids perform similar functions in different inteins. The block B histidine activates the amino-terminal splice junction (Kawasaki et al. 1997) and the penultimate histidine activates the carboxy-terminal splice junction (Cooper et al. 1993; Chong et al. 1996, Chong et al. 1998; Xu and Perler 1996; Kawasaki et al. 1997; Wang and Liu 1997; Chen et al. 2000), although the block B histidine may sometimes play a role in carboxy-terminal cleavage (South-worth et al. 2000). Mutagenesis and structural data suggest that block F plays a role in activating the carboxy-terminal splice junction (Ghosh et al. 2001; Ding et al. 2003).
The protein-splicing mechanism was deciphered using a combination of techniques. It did not take long before each group was mutating the intein amino-terminal residue, the conserved histidines, the carboxy-terminal as-paragine and the +1 residue (Davis et al. 1992; Hirata and Anraku 1992; Hodges et al. 1992; Cooper et al. 1993; Chong et al. 1996; Xu and Perler 1996; Kawasaki et al. 1997; Wang and Liu 1997; Evans et al. 1999; Mathys et al. 1999; Chen et al. 2000). Results varied slightly with each system. Some inteins were able to splice if conservative mutations (serine, threonine, cysteine) were made at the carboxy-terminal side of either splice junction, but usually not efficiently. Mutation of the Tli Pol-2 intein Serl to threonine in the natural extein resulted in about 10% DNA polymerase activity compared to wild type, while mutation to cysteine yielded only minor amounts of spliced DNA polymerase and mutation to alanine blocked splicing (Hodges et al. 1992). Mutation of Cysl to Serl or Cys+1 to Ser+1 in the See VMA intein blocked splicing in yeast cells when assayed by Western blot, although enough spliced product was made to rescue growth (Hirata and Anraku 1992; Cooper et al. 1993). This highlights a significant difference between a biochemical and a biological assay: cells usually require far fewer properly spliced molecules than needed for detection in an in vitro assay. Conservative mutation of the carboxy-terminal asparagine blocked splicing.
Many protein-splicing mechanisms were proposed, which served as guides for experimentation. All of the early models were discarded when the organization of the branched intermediate was determined (Xu et al. 1994), because they were inconsistent with this intermediate. The first proposed model for protein splicing consisted of three steps (Wallace 1993; Xu et al. 1993): an N/ O acyl shift of both Serl and Ser+1, followed by attack on the upstream ester by the downstream primary amine, yielding ligated exteins with the carboxy terminus of the intein connected to the side chain of Ser+1. These models did not make strong predictions for how the branch was resolved, although hydrolysis was suggested. Also in 1993, Cooper et al. suggested a protein-splicing mechanism based on their mutational analysis of the See VMA intein (Cooper et al. 1993). According to this model, the peptide bond between the intein and the C-extein is cleaved during cyclization of the intein carboxy-terminal asparagine (succinimide formation). The resultant amino terminus of the C-extein cleaves the peptide bond at the amino-terminal splice junction, releasing the intein and ligating the exteins. In 1994 another mechanism involving two cyclizations of the intein carboxy-terminal asparagine was suggested (Clarke 1994).
In 1993, Xu et al. developed an in vitro protein-splicing system where the intein in the DNA polymerase gene from the extreme thermophilic archaeon Pyrococcus sp. strain GB-D (Psp-GBD Pol) was cloned between the Escherichia coli maltose-binding protein and the delta Sal fragment of Dirofilaria immi-tis paramyosin, for the first time allowing purification of an active precursor (Xu et al. 1993). This in vitro splicing system was critical to defining the protein-splicing mechanism (Xu et al. 1993,1994; Xu and Perler 1996) in combination with studies of model peptides (Shao et al. 1995, 1996). The key discovery was a branched intermediate, initially identified as a slowly migrating protein in sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS-PAGE). Protein sequencing revealed two amino termini (the intein and the N-extein). The branched intermediate reverted to a linear molecule at pH 10. It was alkali labile, yielding the N-extein and a fusion of the intein with the C-extein. Mutation of either Serl or Ser+1 prevented branch formation. These data suggested an ester linkage between the N-extein and the side chain of Ser+1. The observation that mutation of Serl also blocked branched intermediate formation suggested that an acyl shift of this residue to form a linear ester intermediate might be necessary for branch formation, which was later verified (Shao et al. 1996; Xu and Perler 1996). Mutagenesis data indicated that the intein carboxy-terminal asparagine was necessary for resolution of the branched intermediate. A less common form of side-chain cyclization of asparagine was proposed. Analysis of the carboxy-terminal peptide produced by cyanogen bromide cleavage at an engineered methionine in the Psp-GBD Pol intein proved that the intein ended in a cyclized form of aspar-
agine (a carboxy-terminal succinimide; Xu et al. 1994; Shao et al. 1995). Similar experiments performed with the See VMA intein confirmed this mechanism (Chong et al. 1996,1998).
These data support the accepted four-step protein-splicing pathway: (1) an acyl shift of Serl or Cysl yields an ester or thioester linkage at the amino-ter-minal splice junction; (2) this (thio)ester linkage is cleaved and the branched intermediate is formed by transesterification resulting from attack by the C-extein Ser+1, Cys+1 or Thr+1; (3) the branch is resolved by cyclization of the conserved asparagine, which cleaves the carboxy-terminal splice junction and releases the ligated exteins; and (4) a peptide bond is formed between the ex-teins by a spontaneous acyl shift of the +1 residue. The details of this process are described by Mills and Paulus (this Vol.).
Was this article helpful?