Two Amino Acid Residues of Transposase Contributing to Differential Transposability of IS1 Elements in Escherichia coli

Abstract

Escherichia coli W3110 contains four types of IS1 elements in the chromosome. Using an insertion element entrapping system, we collected 116 IS1 plasmid insertion mutants, which resulted from a minimum of 26 independent IS1 insertion events. All of them had insertions of IS1 of the IS1A (IS1E and IS1G) type. Inspection of the transposase sequences of the four IS1 types and the IS1 of the resistance plasmid R100 showed that two amino acid residues, His-193 and Leu-217 of transposase, might contribute to differential transposability of IS1 elements in W3110. The two amino acid residues of the transposase in IS1A (IS1E and IS1G) were altered separately by site-directed mutagenesis, and each mutant was found to mediate transposition at a frequency about 30-fold lower than that of IS1A (IS1E and IS1G). Thus, the assumption that His-193 and Leu-217 of transposase contribute to differential transposability of IS1 elements in W3110 was confirmed.

Escherichia coli K-12 strains contain 6 to 10 copies of insertion sequence 1 (IS1) in their chromosomes (10, 15). In W3110, seven IS1 copies have been identified, and six were mapped at loci is1A to is1F at 0.4, 6.3, 6.5, 22.3, 75.6, and 97.5 min, respectively (1, 23, 24). These six IS1 copies belong to four sequence types. IS1A and IS1E have identical nucleotide sequences, as do IS1B and IS1C. IS1B (IS1C) and IS1D each have 9 nucleotide (nt) substitutions, and IS1F has 73 nt substitutions (24). The seventh IS1 copy was initially thought to be located upstream of the citrate-dependent iron(III) transport fec genes at 7 min in a non-W3110 strain (25, 27). Since nucleotide sequences of this IS1 copy and its flanking host regions are identical to those of IS1F and its flanking host regions in W3110, and the fec genes are in fact located at 97.5′ in W3110 (18, 19), it has been suggested that IS1F and its adjacent fec genes have undergone a DNA rearrangement(s), resulting in translocation at map position 7′ in this non-W3110 strain (5). An IS1 copy later identified by Zuber and Schumann (27) at 49.6′ in W3110, therefore, should be the seventh instead of the eighth IS1 copy. Only its terminal 21-bp sequence as well as the neighboring 120-bp host sequence was reported.

In our previous study (5), we PCR amplified this seventh IS1 copy in W3110, using primers homologous to the first 16-nt sequence of IS1A (IS1E) and the host sequence adjacent to the end of the seventh IS1 copy. We showed that it has a nucleotide sequence identical to that of IS1A (IS1E) and named it IS1G. A plasmid system carrying sucrose-sensitive sacR and sacB genes from Bacillus subtilis (7) was used to isolate IS1 plasmid insertion mutants in W3110. Ninety-four IS1 insertion mutants were isolated from six independent subcultures, and detailed restriction mapping revealed that all had insertions of IS1 of the IS1A (IS1E and IS1G) type. In this study, we isolated 22 IS1 plasmid insertion mutants by the same method and confirmed that all were insertions of IS1 of the IS1A (IS1E and IS1G) type. The deduced transposase sequence of IS1A (IS1E and IS1G) was compared with those of the other three IS1 types and the IS1 of the resistance plasmid R100, IS1R (16). Two amino acid residues seemed important for the transposition of IS1 elements in W3110. The two amino acid residues in IS1A (IS1E and IS1G) were altered separately by site-directed mutagenesis, and each mutant was shown to mediate transposition at a frequency about 30-fold lower than that of IS1A (IS1E and IS1G).

Collection and analysis of plasmid IS1 insertion mutants.

In our previous study, we collected 94 IS1 plasmid insertion mutants from six independent subcultures which were mapped into five restriction fragments (BamHI-EcoRI, EcoRI-DraI, DraI-DraI, DraI-HindIII, and HindIII-BstEII) in the target 2.6-kb sacR and sacB region (Fig. 1). In order to understand the nature of the 94 IS1 insertions, we further mapped them with HincII and randomly picked six mutants each from the 28 DraI-HindIII insertions in subculture 3, the 18 HincII-DraI insertions in subculture 4, and the 10 DraI-HindIII insertions in subculture 10 for IS1 junction sequence determination. The result shows that five of the six insertions in subculture 4 and all six insertions in both subcultures 3 and 10 had the same insertion locations (data not shown), indicating that most of the insertions in these regions in the three subcultures were probably siblings. Thus, we counted all IS1 insertions in one restriction fragment in a subculture as a single independent event.

FIG. 1.

Distribution of the 116 IS1 insertions isolated from eight independent subcultures in the 2.6-kb sacR-and-sacB-containing BamHI-PstI DNA fragment. The top line represents the 2.6-kb fragment, which is further divided into eight restriction fragments according to nucleotide sequence (22). For the indicated subcultures, numbers in parentheses above each restriction fragment indicate how many IS1 insertions were detected in that fragment. Insertions from subcultures 1, 3, 4, 6, 7, and 10 were isolated and mapped previously (5), but in this study they were also mapped with HincII. An insertion in subculture 6 was found mapped at a specific HincII recognition site and is indicated by an asterisk. Insertions from subcultures 11 and 12 were isolated and mapped in this study.

To increase the sampling size resulting from IS1-independent insertions, two more independent subcultures were collected and analyzed according to the method used in our previous study (5). Briefly, a pACYC177 derivative carrying a sucrose-sensitive sacB gene and the regulatory sequence sacR from B. subtilis (pS177; Table 1) was transformed into W3110, and serial dilutions of cultures of the transformant were plated onto a sucrose-containing plate. Plasmid DNA of all survivor colonies on one dilution plate was analyzed by restriction mapping and Southern hybridization, with radiolabelled IS1 DNA fragment as the probe. By this method we found 7 IS1 and 15 IS5 plasmid insertion mutants from the 33 survivor colonies in one subculture (subculture 11) and 15 IS1 and one IS30 plasmid insertion mutants and 19 plasmid deletion mutants from the 41 survivor colonies in the other subculture (subculture 12). The rest of the mutants did not have sizable changes in their plasmids. Detailed restriction mapping indicated that the 22 IS1 insertions were insertions of IS1 of the IS1A (IS1E and IS1G) type, as in the previous study. Together with the result obtained in the previous study, this result indicated that, of the 338 mutant plasmids analyzed, 135 (40%) had mutations without any sizable change, 22 (7%) had deletions, and 180 (53%) had insertions, 116 (34%) of which had insertions of chromosomal IS1 of the IS1A (IS1E and IS1G) type, 53 (16%) of which had insertions of chromosomal IS5, and 11 (3%) of which had insertions of chromosomal IS30. The 22 deletions were not mediated by IS1 elements that had been transposed onto the plasmid, as evidenced by Southern hybridization with the IS1 probe, despite the fact that IS1-mediated adjacent deletions have been reported to occur at a high frequency (17, 21). Figure 1 illustrates the distribution of the 116 IS1 insertions. A minimum of 26 independent IS1 insertions were analyzed based on their locations in different restriction fragments in mutants of a single subculture or on their isolation from different subcultures.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant characteristics	Reference or source
E. coli strains
W3110	K12 F− λ−, prototrophic	ATCC 27325, 24
DH1	F−supE44− λ− gyrA96 recA1 relA1 endA1 thi-1 hsdR17	8
HB101	F−hsdS20 leuB6 recA13 ara-14 proA2 lacY1 thi-1 galK2 rpsL20 xyl-5 mtl-1 sup44 λ	2
Plasmids
pS177	pACYC177::2.6-kb BamHI-PstI fragment containing sacR and sacB, Kmr	5
pCJ105	pOX38 harboring a Cmr marker	11, 12
pJCA1	pUC18::849-bp BamHI-HindIII fragment containing IS1G and the 81-bp host sequence adjacent to IRR, Apr	5
pJCA2	pJCA1 carrying Leu-217→Gln, Apr	This study
pJCA3	pJCA1 carrying His-193→Tyr, Apr	This study
pJCA4	pJCA1 carrying A6C→A7C, Apr	This study
pJCA5	pJCA1 carrying A6C→A7C and Leu-217→Gln, Apr	This study
pJCA6	pJCA1 carrying A6C→A7C and His-193→Tyr, Apr	This study
pOK12	Medium-copy-number vector, Kmr	26
pJCA11	pOK12::849-bp BamHI-HindIII fragment of pJCA1, Kmr	This study
pJCA12	pJCA1 carrying Leu-217→Gln, Kmr	This study
pJCA13	pJCA1 carrying His-193→Tyr, Kmr	This study
pJCA14	pJCA1 carrying A6→A7C, Kmr	This study
pJCA15	pJCA1 carrying A6C→A7C and Leu-217→Gln, Kmr	This study
pJCA16	pJCA1 carrying A6C→A7C and His-193→Tyr, Kmr	This study

Sequence comparison of the four IS1 types and the IS1 of the plasmid R100.

An IS1 found in the resistance plasmid R100, called IS1R, contains seven base substitutions compared with the IS1A (IS1E and IS1G) sequence (16, 24). Genetic studies with IS1R and IS1K, an IS1 of the IS1A (IS1E and IS1G) type, indicated that both code for two consecutive partly overlapping open reading frames, insA and insB′, and that the functional InsAB′ transposase protein is generated by a −1 frameshifting mechanism at an A₆C motif located at the 3′ end of the upstream insA frame (13, 20). Compared with the IS1A (IS1E and IS1G) sequence, IS1F contains 73 nt mismatches, including a G-to-A transition at nt 542, generating nonsense codon TAG and leading to premature termination of the InsAB′ transposase protein (24). Despite this, an IS1 of the IS1F type (called IS1T) was found to be capable of transposition in a supE44 background and likely to be involved in an insertion mutation in lacZ (4, 14). On the other hand, IS1B (IS1C) and IS1D each have only nine mismatches, and their transposition activities have not been detected previously or in our studies.

The nine mismatches in IS1B (IS1C) lead to six conservative changes (Tyr-24, Tyr-69, Arg-82, His-83, Gly-113, and Tyr-114) and three amino acid changes in the transposase sequence, Leu-81→Phe, Ser-130→Arg, and Leu-217→Gln. The nine mismatches in IS1D include a G-to-T substitution at bp 50, five conservative changes (Tyr-24, Arg-82, His-83, Gly-113, and Tyr-114), and three amino acid changes in the transposase sequence (Leu-81→Phe, Ser-130→Arg, and His-193→Tyr). The G-to-A substitution at bp 50 is located between the Shine-Dalgarno sequence and the start codon for the transposase and is unlikely to play a role in IS1 transposability. IS1R, on the other hand, contains seven mismatches, leading to five conservative changes (Thr-69, Arg-82, His-83, Gly-113, and Thr-114) and two amino acid changes (Leu-81→Phe and Ser-130→Arg) in the transposase sequence (Fig. 2). Since IS1R transposase has been previously shown to be functional (20), and because changes of Leu-81→Phe and Ser-130→Arg are observed in IS1R, IS1B (IS1C), and IS1D, we hypothesized that these two changes have no effect on IS1 transposability and that the Leu-217→Gln change in IS1B (IS1C) and the His-193→Tyr change in IS1D separately contribute to the loss or reduction of transposability in W3110. The fact that IS1F also contains the Ser-130→Arg change, a conservative change in Leu-217, and no change in Leu-81 and His-193 further supports our hypothesis.

FIG. 2.

(A) Alignment of the transposase amino acid sequences of IS1A (IS1E and IS1G), IS1B (IS1C), IS1D, IS1R, and IS1F (IS1T). An amber at position 164 in IS1F (IS1T), which could be suppressed in a supE44 background and result in a transposase of 232 amino acids, is indicated by an asterisk. (B) Structures of IS1A (IS1E and IS1G), IS1B (IS1C), IS1D, and IS1R and positions of the primers used in constructing the transposase mutants. The top two horizontal lines represent the two open reading frames insA and insB′, encoded by the four IS1s, with the frames indicated in parentheses. The −1 frameshift A₆C motif for generation of the InsAB′ transposase protein is shown as a hatched box. The four IS1 elements (768 bp) are shown as horizontal lines, with filled triangles indicating the two terminal inverted repeats, IRL and IRR. For clarity, restriction enzyme recognition sites and the −1 frameshift A₆C motif, present in all four IS1 elements, are shown only in IS1A (IS1E and IS1G). The number in parentheses above each restriction enzyme recognition site is the IS1 coordinate for that site. For IS1B (IS1C), IS1D, and IS1R, only bases different from IS1A (IS1E and IS1G) are shown. Vertical arrows indicate amino acid residues that are altered by the base substitutions. Primers are represented by horizontal arrows below their regions, according to the sequences. Primers IS1DLO and IS1DRO were designed according to the host sequences flanking IS1D, which are not shown. IS1F (IS1T) has 73 bases different from IS1A (IS1E and IS1G) and is not shown.

Generation of transposase mutants of IS1A (IS1E and IS1G).

To prove our hypothesis, we planned to construct two IS1A (IS1E and IS1G) transposase mutants carrying either the Leu-217→Gln mutation or the His-193→Tyr mutation and to assay for their in vivo transposition activities. A mutagenic primer was specifically designed to construct the Leu-217→Gln transposase mutant, which has a T-to-A mutation at bp 704 in the IS1A (IS1E and IS1G) sequence. Primer IS1Leu217 (5′GAGGACTTTGTCATGCTGCTCCACCGATTTTGA3′) is complementary to bp 688 to 718 of the IS1A (IS1E and IS1G) sequence, including a Tth111I recognition site (underlined), except that trinucleotide GAG was added 5′ to the Tth111I recognition site and the T-to-A mutation at bp 704 was incorporated into the center of the sequence (boldface). A pUC18 derivative carrying the IS1A (IS1E and IS1G) sequence (pJCA1; Table 1) was used as the template for PCR with primers IS1Leu217 and IS1L (5′GGTGATGCTGCCAACT3′), which is identical to the first 16 bases of the IS1A (IS1E and IS1G) sequence (Fig. 2). The 282-bp MluI-Tth111I fragment of the resulting 0.8-kb PCR fragment was cloned into pJCA1, generating plasmid pJCA2. DNA sequencing was carried out to confirm that pJCA2 carries only the T-to-A mutation at bp 704 in the IS1A (IS1E and IS1G) sequence (the Leu-217→Gln change in the InsAB′ transposase).

To construct the His-193→Tyr transposase mutant, we used a somewhat different strategy. Genomic DNA of W3110 was used as the template, and two primers (IS1DLO [5′TCGCATGGACAATACG3′] and IS1DRO [5′AAGCGTACGTATTGCA3′]) identical to the outside host sequences adjacent to the left terminal inverted repeat (IRL) and the right terminal inverted repeat (IRR) of IS1D were used for PCR amplification of IS1D (24). The 267-bp BspEI-Tth111I fragment of the resulting 0.8-kb PCR fragment was cloned into pJCA1, generating plasmid pJCA3. DNA sequencing was carried out to confirm that pJCA3 carries only the His-193→Tyr change in the sequence containing IS1A (IS1E and IS1G).

Transposition of IS1 occurs naturally at a very low frequency. It has been demonstrated that insertion of an A residue within the frameshift A₆C motif or replacement of A₆C with GA₂GA₃C, which would fuse the insA and insB′ frames and lead to constitutive production of the InsAB′ transposase, would increase the transposition frequency 40- to 100-fold (20, 21). To increase the assay sensitivity, a transposase-constitutive A₇C mutant of pJCA1 was first constructed by the overlap extension technique with two pairs of PCR primers (9). Primer IS1L (5′GGTGATGCTGCCAACT3′) is identical to the first 16 bases of IRL, and primer A7L (5′ACGTCACTTAAAAAAACTCAGGCCG3′) is identical to bp 298 to 321 of the IS1A (IS1E and IS1G) sequence except that an A (boldface) is introduced into the motif containing A₆C. Primer A7R (5′CGGCCTGAGTTTTTTTAAGTGACGT3′) is complementary to A7L, and primer IS1R (5′GGTAATGACTCCAACT3′) is complementary to the last 16 bp of the IS1A (IS1E and IS1G) sequence. With pJCA1 as the template, two separate PCRs were conducted with primers IS1L and A7R and primers A7L and IS1R, generating two PCR fragments with overlapping ends. The two fragments were gel purified and pooled together as a template for PCR with primers IS1L and IS1R. The 148-bp PstI-BstEII fragment of the resulting 0.8-kb PCR fragment was cloned into pJCA1, generating pJCA4. DNA sequencing was carried out to confirm that pJCA4 carries only the A₇C mutation in the IS1A (IS1E and IS1G) sequence. The 386-bp BstEII-Tth111I DNA fragments of pJCA2 and pJCA3 were separately cloned into pJCA4 to yield pJCA5 and pJCA6. Plasmids pJCA5 and pJCA6 carry the Leu-217→Gln mutation and the His-193→Tyr mutation separately in a transposase-constitutive background, i.e., A₇C; this was also confirmed by DNA sequencing. Figure 2 depicts the overall structures of IS1A (IS1E and IS1G), IS1B (IS1C), IS1D, and IS1R and the positions of the primers used in constructing the transposase mutants, whereas Table 1 describes the characteristics of the transposase mutants constructed as well as other plasmids and E. coli strains used in this study.

Assay of transposition activity of the transposase mutants.

Transposition activities of the transposase mutants were determined by the standard mating out assay, which measures the cointegration frequency between an IS1 donor plasmid and a target plasmid (3, 6, 11). A conjugative plasmid, pCJ105, carrying a chloramphenicol resistance gene was used as the target replicon, and the plasmids carrying mutations in the IS1A (IS1E and IS1G) sequence constructed as described above were the IS1 donor plasmids. The IS1 donor plasmids were transformed into DH1 cells (streptomycin sensitive) harboring pCJ105, and the transformants were grown overnight in Luria broth supplemented with antibiotics at 30°C. The culture was diluted 200-fold in fresh medium without antibiotics, and growth was continued until it reached an A₄₅₀ of 0.7 to serve as the donor cells. Fresh-grown HB101 cells (streptomycin resistant) at the same density were mixed with an equal volume of the donor cells and incubated at 37°C without shaking for 2 more hours to allow conjugation to occur. The mating mixture was plated out onto selective plates. Initially, we used plasmids pJCA1 to pJCA6 as IS1 donor plasmids, and the mating mixtures were plated on plates containing ampicillin (50 μg/ml), chloramphenicol (25 μg/ml), and streptomycin (200 μg/ml) for selection of cointegrate-containing recipient cells and on plates containing chloramphenicol and streptomycin for selection of pCJ105-containing and cointegrate-containing recipient cells. Transposition frequency was calculated as the number of cointegrate-containing recipient cells divided by the number of pCJ105-containing and cointegrate-containing recipient cells to determine the transposition activity of the IS1 element in the donor plasmid. The experiment failed at the beginning, due to the many background colonies on the ampicillin-, chloramphenicol-, and streptomycin-containing plates. We then recloned the 0.8-kb IS1-containing BamHI-HindIII fragment of pJCA1 through pJCA6 into plasmid pOK12 (Km^r), generating plasmids pJCA11 through pJCA16 (Table 1). Selection for cointegrate-containing recipient cells was carried out on plates containing kanamycin (50 μg/ml), chloramphenicol, and streptomycin, and we did not observe background colonies. The results of three independent mating out assays are shown in Table 2.

TABLE 2.

Transposition frequencies of the IS1 transposase mutants in three mating out assays

Plasmid	IS1 in plasmid	Transposition frequency (107)a for indicated expt
		1	2	3
pOK12	None	0	0	0
pJCA11	IS1G	1.45	1.85	2.23
pJCA12	IS1G with Leu-217→Gln	1.84	2.21	2.48
pJCA13	IS1G with His-193→Tyr	1.50	2.79	2.68
pJCA14	IS1G with A7C	49.3	41.3	47.9
pJCA15	IS1G with A7C, Leu-217→Gln	1.39	2.10	2.62
pJCA16	IS1G with A7C, His-193→Tyr	1.14	2.71	2.49

^aTransposition frequency was calculated as the number of recipient cells resistant to kanamycin, chloramphenicol, and streptomycin divided by the number of recipient cells resistant to chloramphenicol and streptomycin in the mating mixture (see text). 

The results show that the transposase-constitutive construct of IS1A (IS1E and IS1G) demonstrated an increase in transposition activity of 21- to 34-fold, indicating that the sensitivity of the mating out assay can be increased about 30-fold with the transposase-constitutive construct. This result is consistent with the results of others (20, 21). The transposase-constitutive constructs of the Leu-217→Gln mutant and the His-193→Tyr mutant demonstrated transposition activities 15- to 43-fold lower than the activity of the IS1A (IS1E and IS1G) transposase-constitutive construct, indicating that the two mutations reduced the transposition activity about 30-fold. The fact that there was essentially no difference in transposition activities between the two mutants and their transposase-constitutive constructs also suggests that Leu-217→Gln and His-193→Tyr both abolish most, if not all, of the transposition activity of the IS1A (IS1E and IS1G) transposase. The fact that both mutants still showed low levels of transposition activities, similar to those of IS1A (IS1E and IS1G), is probably due to trans-complementation by the functional transposases of the chromosomal IS1 elements in DH1 cells. This does not conflict with the result that trans-complementation by IS1A (IS1E and IS1G) did not occur during our isolation of plasmid insertion mutants, because two separate IS1-mediated events were involved (i.e., cointegration between two replicons versus transposition from chromosome to plasmids).

Although studies have shown that transcription activity of the external host sequences could modulate IS1 activity (3), our results indicate that the Leu-217→Gln change in IS1B (IS1C) and the His-193→Tyr change in IS1D contribute to our failure to detect transposition of IS1B (IS1C) and IS1D in W3110. Serre et al. (21) found that His-200, Arg-203, and Tyr-231 residues of the IS1 transposase are important for activity. These three residues are conserved in the transposase sequences of the four IS1 types and IS1R. Together, our results and those of Serre et al. indicate the importance of amino residues in the C-terminal part of the transposase for activity.