More than 200 protein splicing elements, termed inteins, have been identified from the genome of eubacteria, eukaryotes, and archaea (12, 17). A naturally occurring trans-splicing intein, which consists of a 123-amino-acid N-terminal splicing domain and a separate 36-amino-acid C-terminal splicing domain, was discovered to be in the DnaE gene of Synechocystis sp. strain PCC6803 (3, 10, 24, 27, 28). An intein-based split-gene technique, termed intein-mediated protein complementation (IPC), has been utilized to reconstitute the activity of a transgenic product with the goal of limiting the spread of transgenes from genetically modified plants to weedy relatives (6, 7, 13, 16, 18, 26). Theoretically, the split-transgene methodology is very straightforward, but it is not trivial to find a site to split a protein and subsequently have the truncated protein fragments form an active complex (9, 20, 22, 25).

In this study, a G2 EPSPS linker-scanning library was constructed using a transposon-based linker-scanning method (3, 4). Linker scanning was carried out according to instructions provided by the supplier (New England Biolabs). The plasmids used in this study are listed in Table S1 in the supplemental material. Escherichia coli EPSPS gene mutant strain ER2799 did not grow on M9 minimal medium unless complemented with an active EPSPS protein (3, 21) and was utilized in the assay for a functional EPSPS protein. To identify potential split sites within G2 EPSPS, the linker-scanning library was transformed into mutant strain ER2799, and active EPSPS proteins were identified by the ability to restore growth of the mutant cells on minimal medium. Twelve unique sites that allowed growth with 5 amino acid residues inserted were identified, suggesting that these positions were potential sites to divide G2 EPSPS. Seven sites that did not tolerate the 5-amino-acid insertion were found (Fig. 1) (see Table S2 in the supplemental material).

The suitabilities of these 12 positions within the G2 EPSPS protein for IPC were explored using two modified plasmids, pKEB12(E) and pMEB2(B), which contained genes encoding C-terminal and N-terminal splicing domains of strain PCC6803 DnaE intein, respectively (3, 8). The G2 EPSPS gene was divided into two fragments and placed into separate modified plasmids for expression. The plasmids encoding the C-terminal strain PCC6803 DnaE intein splicing domain fused to the C terminus of the gene encoding G2 EPSPS were created by inserting the appropriate G2 EPSPS gene fragments into the EcoRI-to-PstI sites in pKEB12(E). The complementary plasmids that encode the N-terminal intein splicing domain fused to an N-terminal fragment of G2 EPSPS were generated by inserting the appropriate portions of the G2 EPSPS gene into the BamHI-to-XhoI sites in pMEB2(B). The most successful complementary plasmids used G2 EPSPS that was split between amino acid residues F295 and T296, termed pMEPS_N295I_N and pKEPS_C296I_C, respectively (see Fig. S2a and S2b in the supplemental material). Subsequently, the intein splicing domain was deleted from pKEPS_C296I_C to generate a control plasmid, termed pKEPS_C296I_C(−), and a plasmid harboring the complete G2 EPSPS gene was also constructed with pMEB2(B) to yield pMEPS. Cotransformation of E. coli ER2799 cells with the two plasmids pMEPS_N295I_N and pKEPS_C296I_C rescued cell growth on M9 minimal plates supplemented with 50 mM glyphosate and 0.3 mM isopropyl-1-thio-d-galactopyranoside (IPTG). Transformation of the cells with plasmids pMEPS_N295I_N and pKEPS_C296I_C(−), containing both EPSPS protein fragments but lacking the C-terminal 36 amino acids of strain PCC6803 DnaE intein did not permit cell growth in M9 minimal medium (see Fig.S3 in the supplemental material).

Western blot analysis was performed using crude cell extracts and antibodies against the N- or C-terminal peptides of G2 EPSPS. The reaction was visualized with a BCIP (5-bromo-4-chloro-3-indolylphosphate)/nitroblue tetrazolium color development substrate (Promega). Full-length EPSPS, about 47 kDa, indicating trans-splicing activity, was not detected when ER2799 was cotransformed with pMEPS_N295I_N and pKEPS_C296I_C, which express residues 1 to 295 of EPSPS fused to the N-terminal splicing domain of strain PCC6803 DnaE intein (EPSPS_N295-I_N) and the C-terminal domain of strain PCC6803 DnaE intein fused to residues 296 to 445 of EPSPS (I_C-EPSPS_C296), respectively. But the two unreacted EPSPS_N295-I_N and I_C-EPSPS_C296 fragments were present (Fig. 2). These results indicated that the G2 EPSPS protein split between F295 and T296 reconstitutes enzyme activity by IPC in E. coli. In order to further evaluate the effectiveness of IPC in this research system, kinetic characterization of crude proteins extracted from ER2799 containing different plasmids was undertaken using the malachite green dye assay method as previously described (2, 5, 11, 14). The standard reaction was carried out at 28°C in a final volume of 100 μl containing 50 mM HEPES (pH 7.0), 1 mM shikimate-3-phosphate, 1 mM phosphoenolpyruvate (PEP), and 5 μl crude extracts. As can be seen in Table 1, there are no significant kinetic differences between the intein-reconstructed G2 EPSPS and the intact G2 EPSPS, suggesting that the EPSPS fragments were brought together by the intein splicing domains to generate a fully active G2 EPSPS (Table 1).