OUP user menu

Genetic identification of two functional regions in the antitoxin of the parD killer system of plasmid R1

Sandra Santos-Sierra, Consolación Pardo-Abarrio, Rafael Giraldo, Ramón Dı́az-Orejas
DOI: http://dx.doi.org/10.1111/j.1574-6968.2002.tb10995.x 115-119 First published online: 1 January 2002

Abstract

We report the identification and genetic analysis of mutants in the antitoxin of the parD (kis, kid) killer system of plasmid R1. Missense mutants placed at codons 10, 11, 12 and 18 maintained the antitoxin activity of Kis, but not the ability of this protein to co-regulate the parD system together with the Kid toxin. Deletion of the last 33 amino acids of Kis inactivated the antitoxin activity of the protein and reduced substantially, but not completely, its regulatory activity. These results define two functional regions in Kis: an amino-terminal region which is specifically involved in regulation, and a carboxy-terminal region of the protein, which is important both for its regulatory and antitoxin activities.

Keywords
  • parD operon
  • parD regulation
  • Kis functional region
  • Kis, ChpAI and ChpBI relationship
  • Toxin–antitoxin system

1 Introduction

parD is a killer stability system found in the drug resistance factor R1 [1,2] and in plasmid R100 (pem system [3]). As found in other proteic toxin–antitoxin systems [4] the parD system encodes a stable toxin, Kid, and an unstable antitoxin Kis [5]. Kid inhibits initiation of ColE1 replication and Kis neutralizes this inhibition [6]. On its own, the Kis protein is a weak transcriptional repressor of the parD operon but in combination with Kid forms an efficient repressor [7]. Kis and Kid proteins interact, forming a stable complex [6] that is probably required to inhibit the toxicity of Kid and to form an efficient repressor of the system.

The Kis antitoxin shares homology, at its amino- and carboxyl termini, with CcdA, the antidote of the ccd system of plasmid F [8]. However, these antidotes are not functionally interchangeable [8]. As in parD, the toxin and antitoxin proteins of ccd form an efficient repressor of the system [911].

chpA and chpB are homologous systems to parD/pem found in the Escherichia coli chromosome [12]. The antitoxins of these chromosomal systems, ChpAI and ChpBI, can neutralize the toxicity of Kid [13,14]. This effect requires increased gene dosage of the chromosomal systems and is enhanced by missense mutations affecting the amino-terminus of their antitoxins. As in the parD system, the antitoxin and the toxin of the chpA system regulate in a concerted action the chpA promoter [15]. The chpB system is probably regulated in a similar way.

In order to identify the functional regions in the Kis antitoxin, we have isolated a collection of Kis mutants deficient in auto-regulation. The data are discussed in the context of mutants in the antidotes of the chpA, chpB and ccd systems.

2 Materials and methods

2.1 Strains and plasmids

E. coli K12 strain CSH16 (F′lacZ, proA+B+/Δ(lacpro) supE, thi) [16] was used as general host. pBR322 is a multicopy cloning vector conferring resistance to ampicillin and tetracycline [17]. pOM34 bears a transcriptional fusion, made on the R1 vector pJEL245, between portions of the parD system and the lacZ gene [7]. This construction confers resistance to ampicillin. pAB24 and pB24 are similar pBR322-parD recombinants conferring resistance to tetracycline and containing the PstI-EcoRI fragment of plasmid R1 that includes the parD operon. pAB24 contains the wild-type parD system and has been described elsewhere [7]. The parD recombinant pB24 contains the amber mutation kis74 [2] and kid85 (R85W), a missense mutation that results in loss of toxicity but maintains the co-regulatory activity of the toxin (Santos Sierra, unpublished results). pSS100 was constructed inserting a BamHI-EcoRI fragment containing kid in the multicloning site of the pNDM220 vector, using as primers PKIDB(−) (5′-cggggatccgtcaggaggaaatctgac-3′) and PKIDE(+) (5′-caagaattctgttcaagtcagaatagtgga-3′). These primers introduce respectively the BamHI and EcoRI sites at the 5′ and 3′ ends of kid. pNDM220 [18] is an R1 derivative vector conferring resistance to ampicillin and carrying a multicloning site downstream from the pA1/O4/O3 promoter, which is inducible by isopropyl β-d-thiogalactopyranoside (IPTG).

2.2 Growth media

L-Broth (LB) and L-broth agar (LA) were used as liquid and solid medium, respectively [19]. M9 minimal medium was as described by Miller [16]. MacConkey agar supplemented with lactose was used for the screening of parD de-repressed mutants. When necessary antibiotics were included: ampicillin to 50 μg ml−1 and tetracycline to 10 μg ml−1.

2.3 Recombinant DNA procedures

DNA extractions, constructions of plasmids and transformation of cells with plasmids were essentially according to Sambrook et al. [20]. In vitro mutagenesis with hydroxylamine, a mutagen that induces GC-AT transitions due to the hydroxylation of cytosines, was carried out for 35 min at 65°C as described [21]. The DNA sample was extensively dialyzed against 10 mM Tris–HCl pH 8.0, 1 mM EDTA before being introduced into the cells by electroporation using a Bio-Rad MicroPulser. Experimental conditions were basically as recommended by this supplier. DNA sequencing of the mutants was performed by the chain terminator method [22] using an automated DNA sequencer (Perkin-Elmer, ABI Prism, 377 DNA sequencer). β-Galactosidase assays were essentially as described by Miller [16].

3 Results

3.1 Screening of mutants that de-repress the parD promoter

Interactions of Kis and Kid are required to form the repressor of the parD system and to neutralize the toxicity of Kid. Therefore the analysis of mutants of Kis leading to de-regulation of the parD promoter can give information on protein regions involved in regulation and in antitoxicity. To isolate these mutants we used two pBR322-parD recombinants, pAB24 and pB24. pAB24 carries the parD wild-type system (kis, kid). Due to the wild-type kid gene, kis mutations obtained in pAB24 should maintain the antitoxin activity of the protein. pB24 relieves this constraint because it carries a mutation in kid, kid85 (R85W), that inactivates the toxicity of Kid but not its activity as co-repressor. To identify possible parD de-regulated mutants, mutagenized pAB24 or pB24 plasmids were rescued in the E. coli CSH16 (lac, supE) strain containing pOM34, an R1 recombinant with the lacZ gene under control of the parD promoter [7]. Mutants unable to repress the parD promoter will elicit β-galactosidase synthesis in this background and they will give red colonies in MacConkey agar supplemented with lactose [16]. Note that the pB24 construct and their mutants carry, in addition to kid85, the amber mutation kis74 that is silenced by the supE mutation of the host.

3.2 An amino-terminal region of Kis is specifically required for parD regulation

pAB24 or pB24 was mutagenized with hydroxylamine and introduced by electroporation into CSH16/pOM34 selecting colonies in MacConkey agar supplemented with tetracycline and ampicillin. These antibiotics select for the incoming and resident plasmids, respectively. To rescue possible mutants interfering with growth of the host at high temperature, colonies were selected at 30°C. This experiment gave red colonies (de-repressed parD mutants) with a frequency close to 1/500. Among these, 1/3 corresponded to mutations in kis (4/12 isolated in pAB24 and 8/27 in pB24); 2/3 were affected in kid and are not described here. Out of 12 kis mutations analyzed, nine were missense (three isolated in pAB24 and six in pB24) and two were nonsense (isolated in pB24). The nonsense mutations are described in Section 3.3. Missense mutations in kis isolated on pAB24 led to the following amino acid changes: G10S, G10D and P18L. The kis mutations isolated on pB24 led to the following changes: G10R, G10S (isolated twice), G10D, G11S, and S12L. All these mutants changed adjacent amino acids located at the amino-terminus of Kis (residues 10, 11, 12 and 18). Therefore they define a narrow region of the protein involved in regulation (Table 1 and Fig. 1). The de-regulated P18L mutant was previously isolated and characterized [1,7] and is not further analyzed here.

View this table:
Table 1

Kis mutants that lead to de-regulation of the parD system

PlasmidsparD genotypeRelevant amino acid (codon) change in KisbparD de-repression levelc
pBR322100
pAB24kis, kid6
pAB24-Kis10Akis10A, kidG10S (GGC-AGC)117
pAB24-Kis101Akis101A, kidG10D (GGC-GAC)95
pB24kis74, kid855
pB24-Kis74kis74, kid1a73
pB24-Kis102kis102b, kid85G10R (GGC-CGC)96
pB24-Kis11kis11b, kid85G11S (GGC-AGC)83
pB24-Kis101Bkis101Bb, kid85G10D (GGC-GAC)110
pB24-Kis12kis12b, kid85S12L (TCA-TTA)106
pB24-Kis10Bkis10Bb, kid85G10S (GGC-AGC)101
pB24-Kis53kis53b, kid85S53Stop (TCA-TAA)45
pB24-Kis51kis51b, kid85Q51Stop (CAA-TAA)47
  • a kid1 introduces an Ochre Stop mutation (TAA) at the Met start codon (ATG) of kid.

  • bMutants based in pB24 contain, in addition to the relevant amino acid/codon change, the parental amber nonsense mutation kis74 (Trp74Stop, TGC-TAG).

  • c parD promoter activity was measured as % of β-galactosidase produced by CSH16/pOM34 in the presence of the listed plasmids. The value obtained in the presence of pBR322, 1200 Miller units, was taken as 100%. Values are the mean of three independent measures differing less than 7%.

Figure 1

Alignment of amino acid sequences of Kis, ChpAI and ChpBI antitoxins, showing the amino acid substitutions introduced by the missense mutations. The amino acid changes are indicated above the sequence. The black arrow and bar above the sequence represent regions with predicted β-sheet and α-helix structure, respectively. Dots indicate identical residues in the alignment. Sequence alignments and secondary structure predictions have been reported previously [14].

Quantification of β-galactosidase levels modulated by the different missense mutants in the CSH16/pOM34 background (Table 1) indicated that the mutations led to high expression levels of the parD promoter. The parental plasmids, pAB24 and pB24, repressed this promoter efficiently, and a mutation that prevented translation initiation of Kid (kid1, pAB24-Kis74) gave the weak repression that corresponds to Kis alone [7].

The antitoxin capacity of the de-regulated Kis mutants was directly analyzed testing neutralization of growth inhibition promoted by Kid. For this purpose we constructed pSS100, a plasmid that leads to Kid induction and toxicity in the presence of IPTG (see Section 2). The pAB24 or pB24 constructs (parental and mutants) were introduced in CSH16 containing pSS100 and the growth of these strains were analyzed in the presence or absence of IPTG. Compared to cells containing only the pSS100 recombinant, cells containing this plasmid plus the pAB24 or pB24 parental plasmid or their missense mutants were able to grow in the presence of IPTG, that induces the lac promoter and therefore the synthesis of the Kid toxin (Fig. 2). This indicates that the Kis mutants maintained the antitoxin activity.

Figure 2

Antitoxin assay of the de-regulated Kis mutants. Cells of CSH16/pSS100 containing the different parD mutants (Table 1) were grown overnight at 30°C in LA medium and in the presence (+) or absence (−) of IPTG (0.1 mM). Sections A and B correspond to cells containing kis missense and nonsense mutations, respectively. As a control, cells of CSH16 containing the pSS100 recombinant alone were included. The sector orientation is the same in the left and right plates.

The positive but poor growth of cells containing pSS100 and pB24 in the presence of IPTG (Fig. 2) indicates that this recombinant protected against Kid toxicity less efficiently than its de-regulated mutants or than pAB24. This is probably due to a poor suppression of the kis74 amber mutation in CSH16 (supE), resulting in lower levels of the antitoxin. Note that these levels are sufficient to give an efficient repression (Table 1). Increased levels of the antitoxin in the de-repressed mutants result in efficient growth in the presence of Kid.

On the whole, the missense de-regulated mutants obtained in pB24 and pAB24 identified an amino-terminal region of Kis (amino acids 10–18), that is required for parD regulation but that is dispensable for its antitoxin activity.

3.3 A region at the carboxy-terminus of Kis is involved both in parD regulation and in antitoxin activity

In addition to missense mutants at the amino-terminus of kis, two nonsense mutants in this gene were isolated in the pB24 parD kid85 recombinant. The mutations changed codons corresponding to Q51 and S53 of kis into the Ochre stop codon UAA. As this stop codon is not suppressed in the CSH16 (supE) background, the mutations should remove 35 and 33 residues, respectively, from the carboxyl terminus of the protein. β-Galactosidase assays indicated that these mutants substantially reduce but do not abolish the regulatory potential of Kis (Table 1). Compared to the missense mutants, the nonsense mutants lose the antitoxin activity of the protein (Fig. 2). This suggests either that the deletions affect the region of Kis needed to inactivate Kid and to form with this protein an efficient repressor, or that these mutations severely affect the overall three-dimensional fold of Kis. However, since circular dichroism spectroscopy indicates that isolated Kis protein mainly has a flexible coil structure (our unpublished results), the latter hypothesis seems unlikely. As expected, these mutants were not isolated when the parD wild-type recombinant pAB24 was used.

The data indicate the double involvement of the carboxy-terminal region of Kis in regulation and in antitoxicity and are also consistent with the presence of a specific regulatory region at the amino-terminus of this protein.

4 Discussion

The analysis of parD de-regulated mutants obtained in this study identified an amino-terminal region in the Kis antitoxin specifically involved in regulation and a carboxy-terminal region in which the antitoxin and regulatory activities of the protein overlap. The simplest interpretation is that Kis interacts with the Kid toxin via its carboxy-terminal region to form a complex that represses the parD operon and also inactivates the toxicity of Kid. The amino-terminus of Kis would be involved in regulatory interactions (protein–protein or protein–DNA interactions) dispensable for neutralization of the toxin. However, the regulatory potential of this amino-terminal region is low and it is greatly enhanced when the toxin and the antitoxin proteins interact.

The parD system is homologous to chpA and chpB, two toxin–antitoxin systems found in the chromosome of E. coli. We previously reported that the ChpAI and ChpBI antitoxins were able to neutralize the Kid toxin and that this cross-activity required high gene dosage of the chromosomal systems and was enhanced by missense mutations affecting the amino-terminus of the antitoxins [13,14]. These mutations are located in amino acids homologous to the ones affected by the missense mutations described here (Fig. 2). Furthermore, as in the mutations in kis, missense mutations in chpAI or chpBI conserve the specific antitoxin activities of the proteins [13,14]. These data support a similar distribution of functional regions in the three antitoxins and therefore indicate that the amino-terminal regions in the ChpAI and ChpBI antitoxins are specifically involved in regulation. We propose that the ability of the ChpAI and ChpBI mutants to neutralize the toxicity of Kid is due to the maintenance of the basal anti-Kid activity of these antitoxins [13,14], and to an increase in their levels due to de-repression of the chpA or chpB operons. Neutralization of Kid by a promoter-up mutation leading to overexpression of the chpB system [14] is consistent with this proposal. These data make unlikely the alternative hypothesis that the mutations at the amino-terminus of the chromosomal antitoxins define an interface of toxin–antitoxin interaction directly involved in neutralization of the toxin [13].

The conservation of activities and the similar genetic organization found in the antidotes of the parD, chpA and chpB systems strongly suggest a close relationship among them. This also applies to CcdA, the antitoxin of the ccd system. Comparison of hydropathy profiles of Kis and CcdA [8] and genetic analysis on these antitoxins ([11] and this work) indicate that their amino- and carboxy-terminal regions share structural and functional homology. These data support the notion that the parD/chpA/chpB and ccd systems may derive from a common ancestor. A model to explain a possible evolution of the plasmid toxin–antitoxin systems from their chromosomal homologues has been proposed [23].

Acknowledgments

This research has been supported by the Spanish Grant BIO99-0859-CO3-01, the EU projects BIO94-CT98-0106, QLK2-CT-2000-00634 and the concerted action BIO4-CT-0099. R.D.O. also acknowledges a grant from the Comunidad de Madrid (Spain). We acknowledge Ana Marı́a Serrano for technical assistance, Marc Lemonnier for critical reading of the manuscript, Andrew Kralicek for linguistic corrections and Kenn Gerdes for plasmid pNDM220.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
View Abstract