ELSEVIER

MICROBIOLOGY LETTERS

FEMS Microbiology Letters I36 (1996) I29- 136

Suppression of the thermosensitive replication phenotype of the derivative plasmid of ~19789: :Tn 5.52 in Staphylococcus aureus

may involve integration of the plasmid into the host chromosome

Muhammad Sohail, Keith G.H. Dyke *

Microbiology Unit, Department of Biochemistry, Soulh Parks Road, Unirersig of O_Fford, Oxford OX1 3Qll, fJK

Received 8 November 1995; accepted 3 December 1995

Abstract

Plasmid-chromosome co-integration was found to be the mechanism of choice to overcome thermosensitivity of

replication of the plasmid pS1 in PS8Od and RN4220 strains of Sruphylococccrs UUWUS. The integration of the plasmid was sometimes accompanied by deletion of a specific section of the plasmid pS1 in PSSOd. Growth of bacteriophage on strains containing the integrated plasmid and the subsequent use of the phage in transduction gave transductants containing plasmids

that had regained their replication thermosensitivity. These plasmids had not acquired any detectable chromosomal DNA. The 16-kb EcoRI fragment of the PS80d chromosome that hybridizes to pS1 is the target for recombination in many cases, but apparently other sites are also used. This fragment contains sequence homologous to parts of the transposon Tn552 and

it is probable that site-specific recombination is involved in the integration. The possible mechanisms for the integrations and

the deletions are discussed.

Kep~ords: Staphylococcus aureus; pI9789::Tn552; Thermosensitivity; Plasmid-chromosome integration

1. Introduction to nalidixic acid and rifampicin; [4]), not all chromo-

Staphylococcus aureus can be a problem organ-

ism in nosocomial and community-acquired infec- tions because it often possesses multiple antimicro- bial resistances. The mechanisms of homologous recombination, site-specific recombination and trans-

position seem to have played a key role in the evolution of these multiresistant S. aureus [l-3].

Although some chromosomally specified resistance determinants may arise by mutation (e.g. resistance

some-borne resistances are due to mutation. Evi-

dence suggests that some, originally known to be

present on plasmids, are now accumulating on the chromosome. In the majority of recently isolated multiresistant S. aureus strains, the determinants for

resistance to erythromycin, tetracycline, strepto- mycin, spectinomycin. penicillin via a p-lactamase, etc., and to cadmium and mercury are found on the chromosome. Evidence from some more recent Aus- tralian epidemic staphylococcal strains suggests that the chromosomally located determinants for tetracy- cline, penicillin and mercury resistances are homolo- gous to determinants which were plasmid-borne in

_ Corresponding author. Tel.: +44 (1865) 275 293; Fax: +44

(1865) 275 297; E-mail: kdyke@molbiol.ox.ac.uk.

0378.1097/96/$12.00 0 1996 Federation of European Microbiological Societies. All rights reserved

SSDI 0378- 1097(95)00488-2

130 M. Sohuil, K.G.H. D~kr / FEM.5 Microbio/oCq Lettrm 136 (IYY6I /2Y-136

isolates prior to 1970 151. Although accumulation of the resistance determinants on the chromosome is

disadvantageous since the genes occur in a low copy number in the population and are less easily transfer-

able to other bacteria, such accumulation provides a greater stability to the resistance determinants in the

absence of selective pressure. It is thus of clinical

importance to study the mechanisms involved in the

accumulation of resistance determinants on the chro-

mosome and hence plasmid integration into the chro-

mosome. A thermosensitive derivative of the large, natu-

rally occurring staphylococcal plasmid

pI9789::Tn_552, called pS1. was used in this study.

This plasmid carries genes that confer resistance to

metal ions such as cadmium, mercury, arsenate and also to penicillins. For a detailed account of the

plasmid pI9789::Tn5.52 and its derivatives see [6,7].

2. Materials and methods

2.1. Bacterial strains, growth conditions, phage

propagation and transduction

Bacterial strains used here and some of their characteristics are listed in Table 1. S. aureus strains

containing a plasmid with thermosensitive replica- tion were grown and maintained for routine use at 30°C on CY (yeast extract, 10 g l- ’ ; casein hydro-

lysate, 10 g I-‘; and MgSO,, 0.25 g l- ‘1, unless otherwise stated. Long-term storage was at -70°C

in broth containing 25% (v/v) glycerol. Cadmium

acetate was used at 26.6 wg ml ’ Phage propaga- tion and transductions were done according to [8].

Table I Bacterial strains

2.2. DNA preparation and manipulation procedures

Large-scale plasmid DNA from S. aureus was

prepared according to [9]. Small-scale plasmid prepa- ration was carried out by the alkaline lysis method

[lo]. 7 ~1 of lysostaphin (10 mg ml-’ stock) was

added per 100 ~1 of the cell suspension before the

addition of SDS/NaOH lysis solution and the mix-

ture was incubated at 42°C for 5-10 min. The

phenol/chloroform purification step was replaced with two cleaning steps with StrataClean Resin’“’

from Stratagene which was used according to the

manufacturer’s instructions. Total DNA from S. aureus was prepared as fol-

lows. Cells were harvested from 1.5 ml of an

overnight culture and resuspended in 400 ~1 of TE (IO mM Tris . HCl pH 8.0. I mM EDTA) containing

heat-treated RNase A at 10 pg mll ’ 6 ~1 of lysostaphin (10 mg mll’ stock) was added and the

mixture was incubated at 42°C for 7 min. This was followed by addition of 20 ~1 of 10% SDS (v/v). 2

~1 of proteinase K (10 mg ml ’ stock) and incuba-

tion at 37°C for 7 min. 70 ~1 of 5 M NaCl and 50 ~1 of CTAB (lo%, w/vl/NaCl (0.7 M) solution were

then added and the mixture was incubated at 65°C for 3 min. After incubation, the sample was cooled

slowly and extracted twice with phenol/chloro-

form/isoamyl alcohol (25/24/l, v/v/v) and once with chloroform/isoamyl alcohol (24/ 1, v/v). DNA was precipitated with 0.6 volumes of isopropanol at

room temperature for 15 min, pelleted by centrifuga- tion at 12 000 X g for 10 min and washed with 70%

ethanol (v/v). The DNA pellet was suspended in 50 ~1 of sterile distilled water.

Restriction enzyme digestions and agarose gel

Strain Remarks Reference

s. aldrl?llS

PS80

PS8Od

1054

RN4220

E. coli

JM107

Chromosome carries genes for /3-lactamase production

P-Lactamase negative derivative of PS9789

Propagating strain of S. ~~ureu.s phages 53 and 80

Restriction minus derivative of NCTC 8325 strain

Primary host for pBR322- and pOX7-based clones

1161 [I II iI41

1121

1101

M. Sohail, K.G.H. Dyke/ FEMS Microbiology Letters 136 (1996) 129-136 131

electrophoresis, etc. were carried out according to [IO]. For Southern hybridization, DNA was blotted onto a nylon membrane [lo]. The labelling of the probes and the detection of hybridization were car- ried out using the ECL’” (Amersham) labelling/de- tection systems according to the manufacturer’s in- structions.

2.3. Determination of thermosensitivity suppression

The 5. aureus strain containing a thermosensitive plasmid was grown at 30°C for 18 h in 10 ml CY broth in a 50-ml flask. Optical density of the culture was measured at 675 nm and lo6 cfu were plated onto CY-Cd (26.6 pg ml-‘) agar plates. The plates were dried with lids off at 42°C for 30 min. The lids were then replaced and plates were incubated for another 17.5 h at 42°C. The number of cadmium-re- sistant colonies appearing on the plates was used as a measure of the thermostability of replication of the plasmid DNA. All experiments were performed in triplicate and repeated at least three times.

3. Results and discussion

3. I. The suppression of thermosensitivity of the plas- mid pS1 is strain-specific

The increase in the suppression frequency of pS1 in the presence of the transposon Tn551 on the chromosome of PS80d (i.e. in PS80d::Tn551) sug- gested that the composition of the host chromosome can influence the suppression and that there was some interaction between the host chromosome and the plasmid [8]. The effect of the host strain on the thermosensitivity suppression was thus investigated. Two strains of 5. aureus, PS80d [l I] and RN4220 [ 121 were used to study the suppression. The plasmid pS1 was transferred from PS80d to RN4220 using phage transduction. Two phages, phage 80 and phage 53 of the International Phage Typing System of 5. aureus [13] were used. The strain PS80d is suscepti- ble to phage 80 only and RN4220 to phage 53 only. Another strain 1054 [14] of S. aureus is susceptible to both phage 80 and phage 53 and so served as a ‘bridging’ or ‘intermediate’ host strain in the transfer of the plasmid from PS80d to RN4220. Phage 80

was grown on PS8Od(pSl) and the phage used to transduce pS1 to 1054 to give rise to 1054(pS 1). The plasmid pS1 was then transduced to RN4220 using phage 53 grown on 1054(pSl) to obtain RN422O(pSl). The presence of the correct trans- duced plasmid in RN4220 was verified by restriction enzyme analysis and testing for the expression of the phenotypic traits (e.g. /3-lactamase production, resis- tance to cadmium and mercury ions) conferred by the plasmid. The thermosensitivity suppression of the plasmid pS1 was then determined in both strains and the results show that the suppression frequency is influenced by the host strain. It is higher in PS80d (0.33 X 10m4 per bacterium) than in RN4220 (0.04 X 10m4 per bacterium). Twenty independently iso- lated thermostable colonies from each of the PS8Od(pSl) and RN422O(pSl) types were subjected to the procedure for preparation of plasmid DNA. No plasmid DNA was detected in any of these colonies and so it was assumed that at least the cadmium gene and probably the whole plasmid had integrated into the host chromosome. This assumption was investi- gated using bacteriophage transductions and South- em hybridization.

3.2. Evidence of integration from phage transduc- tions and restriction endonuclease mapping

On excision of the integrated plasmid from the chromosome, covalent linkage between the plasmid DNA and some chromosomal DNA may be expected (e.g. see [ 153) and in the case of phenotypically cryptic DNA, restriction enzymes could be used to detect this linkage. The strain PS8Od was chosen for further work because of the higher frequency of the suppression of thermosensitivity obtained in this strain.

Phage 80 was grown separately on 20 of the thermostable colonies of PS8Od(pSl) type and then used to transduce cadmium resistance to a new PS80d host. Two transductants were picked from each transduction and they were all found to contain a plasmid. Analysis of the uncut plasmid DNA re- vealed predominantly three sizes in the population of 40 plasmids. A total of 12 transductants (four from each of the three sets of the plasmids) were picked and their plasmid DNA analysed after digestion with EcoRI and BgZII. This allowed the grouping of the

plasmids into three types: Type I indistinguishable striction enzymes known to cut within this fragment

from PSI: Type II had a deletion of about 3.2 kb in were chosen. The restriction enzyme maps of the two

the large 10.2-kb ELwRI fragment (EcoA fragment): types of the plasmids thus obtained are given in Fig.

and Type III had a deletion of about 6.5 kb in the I. For both types, the deletions have removed most

10.2-kb EcvRI fragment of’ pS 1 (see Fig. I ). How- of the invertible region, itlr,. and appear to start at or

ever, no non-psi DNA was detected in these three near one W.S site. For Type II, the 3.2-kb deletion

types. The positions of the deletion in the Type II has deleted DNA between the A~c.1 site (2.05) and

and III plasmids were investigated more precisely. the MKI (5.7) that is situated in the MO region of

The deletions found in Type II and III plasmids were Tn552. For Type III, the 6.5-kb deletion has re-

shown to be within the EcoA fragment and so re- mo\,ed DNA between the Ac,c,l site (2.05) and the

pS1 and Type I

Type III PI

I 3 7

E’ EV Sn, 3.lli 35 ,

7 i 1 :\_____________--________._______;I

4 rrs

3. Deletion

m irw

(-INegative orientation of inv

(+)Positive orientation of inv

Fig. I. Restriction enqme map5 4~owing dclrtion\ L\ ithtn the EcoA Ira~mcnr of the Types II and 111 plauni&. The restriction ewymr map

of the Type I plasmid is indi\tingui\hahle rrom that of pS I and i\ no1 shown separately. The region tlanletl h! w.$ Gtes (invertible region or-

irzd on the plasmid PSI is capable of inverting and can occur in either of the two orientations. Key to restriction enLyme name\: A. Awl.

B. &III: C. Clol: EI. f?oRI: M. Mvcl: N. N~kl: PI. PI ul: PII. Pi 011: Sl. So/l: Sn. .S~~~tBI. Various co-ordinate\ are shown. Note that the

Au.1 \ites at co-ordinate 2.05 i\ generated when the i/rl is in the negatl\c orientation.

M. Sohail. K.G.H. D,vkr / FEMS Micmhiolog~ Letters 136 (IYY6) 129-136 133

proximal Sal1 site (9.15) situated outside the trans- poson Tn.552, resulting in the deletion of the whole

of Tn552. Although, inr was apparently involved in all of the deletions, none of the deletions were a precise deletion of the ins from the plasmid. Also,

the deletion in the Type III plasmids is not a precise deletion of Tn552, since the CM site (8.8) of pS1 is

absent and it is not within Tn552. Deletions were reported to occur in the plasmid

pE194 when chromosomally integrated pE194 is ex- cised from the chromosome [15]. The deletions recorded for pS1 during this study could be regarded

as a result of imprecise excision during transduction.

(4

2

However, the deletions in pS1 occur only when the

plasmid is transduced from a thermostable derivative but not when pS1 was transduced from cells grown

at 30°C. All three types of plasmid were found to be

thermosensitive for replication and so it was con-

cluded that the suppression was not due to the reversion of the mutation conferring thermosensitiv-

ity on the plasmid. The thermosensitivity suppression

frequency of the Type I plasmids was approx. 0.35 X IO-” per bacterium, which is similar to that for

the plasmid pS I, and that of Types II and III was

approx. 0.1 X IO-’ per bacterium. However, one

(B)

Fig. 2. (A) Agarose gel stained with ethidium bromide. Total DNA was digested with EcoRI and electrophoresed on a 0.7% agarose gel for

14 h. The DNA was transferred to a nylon membrane and hybridized to the pS I probe (plus pUC I8 to illuminate the Raoul ladder; pUC I8

does not hybridize to PS80d, RN4220 or pS I DNA; data not shown). (B) An autolumigraph of the membrane showing hybridization of the

plasmid pS I to the total DNA isolated from PSBo(pS 1) thermostable colonies. The plasmids pS I, Type II and Type III are also shown. For

the plasmid pS I, in addition to the four EcoRI bands, there are fainter hybridizing bands at about 13.5 kb and 7 kb that do not correspond to

any band seen in the corresponding lane on the ethidium bromide stained photograph. These bands have been ignored as either artifacts or minor contaminants.

134 M. Sohail. K. G. H. Dyke / FEMS Microbinlogy Letters 136 f 19961 12% 136

Type 11 multimer

EcoRI

b X Recombination EcoRl

4 PS80 chromosome

16kb

I

Chromosome of the Type II parent

Fig. 3. A possible route for the generation of the cadmium-resistant thermostable bacteria giving rise to some Type II plasmids on

transduction. The drawings are not to scale. See text for details.

M. Sohail, K.G.H. Dyke/FEMS Microbiology Letters 136 (1996) 129-136 135

example of the Type I showed a higher suppression frequency (approx. 1 .O X 1 O-4 ), although indistin- guishable from other members of the group on re- striction analysis.

3.3, Evidence of plasmid integration from the South-

em hybridization analysis

The chromosomes of the bacteria with integrated plasmids, which gave rise to the two types of dele- tions in the plasmid pS1 on transduction, were anal- ysed using Southern hybridization in order to dis- cover whether the deletions found in the two types of transductant plasmids produced during transduction were the consequence of integration/excision into/from the chromosome. The plasmid pS1 was digested with EcoRI and then used to probe total DNA derived from three thermostable colonies that gave rise to Type II plasmids and three that gave rise to Type III plasmids (Fig. 2).

The pS1 probe hybridizes to a single approx. 16-kb EcoRI fragment of the PSSOd chromosome. Analysis of the PSSOd chromosome using more spe- cific probes derived from within the transposon Tn.552 (e.g. binL, res and orf 271) showed that this 16-kb fragment contains most of Tn552 (data not presented). The chromosome of the P-lactamase-pro- ducing S. aureus strain PS9789, which has Tn552 on its chromosome and which gave rise to the p- lactamase-negative derivative PS80d [ 11,16,17], also showed a hybridization pattern similar to that of PSSOd. However, the chromosomal DNA of strain RN4220 did not show any homology to the pS1 probe (not shown).

The thermostable colonies that gave rise to the Type II plasmids all contain the 16-kb hybridizing chromosomal fragment of PS80d and all possess the four EcoRI bands corresponding to those found in the Type II plasmid, thereby suggesting that the deletion found in the plasmids occurred at or before integration. All’three also show bands at 13.5 kb and 9.7 kb which could be the sum of the 16-kb fragment from the chromosome and 7.0 kb from the plasmid. Thus, a possible explanation is that growth at 42°C first caused or selected a deletion of 3.2 kb from within the 10.2-kb EcoRI fragment of pS1 to give the 7.0-kb piece. The integration may be unstable and reversible, thereby producing a mixture of bacte-

ria, some with an intact 16-kb fragment and the others without. There is a slightly lower intensity of hybridization of the 7.0-kb fragment as compared to the 7.9-kb and the 6.5-kb fragments, which is consis- tent with this hypothesis (Fig. 3).

The hybridization patterns from the total DNA of the three thermostable colonies that gave rise to the Type III transductant plasmids are very similar and indicate recombination between the plasmid pS 1 and the PSSOd chromosome. Two EcoRI fragments, one each from the chromosome (16 kb) and the plasmid pS1 (10.2 kb) are replaced by two EcoRI fragments of 11 kb and 5.7 kb. The two new fragments may be a result of the recombination between PSI and the chromosome accompanied by a deletion in both the plasmid and the chromosome, since 16 kb + 10.2 kb (= 26.2 kb) generate 11 kb + 5.7 kb (= 16.7 kb). Moreover, the Type III plasmid has neither the 1 1-kb

nor the 5.7-kb fragment, but does contain a 3.6-kb fragment (Fig. 2).

Hybridization data and the ancestry of the strain PS80d [11,16,17] suggest the presence of a Tn552- like structure on the 16-kb EcoRI fragment of the chromosome which is important in the ‘plasmid integration’ by providing a region of homology be- tween the plasmid and the chromosome. The genera- tion of the P-lactamase-negative phenotype in PSSOd may have been the result of a point mutation or a small deletion rather then deletion of the whole or most of the bla region from the chromosome as was previously suggested [ 111.

Overall, the results can be summarized as below: 1. Plasmid-chromosome co-integration seems to be

the most common mechanism to overcome ther- mosensitivity in both S. aureus PS8Od and RN4220. Such events may also have occurred in the natural populations of bacteria under condi- tions of stress and starvation or otherwise to promote integration of plasmids carrying genes for antibiotic resistances into the host chromo- some.

2. The deletions observed in the transductant plas- mids are not phage-mediated and occur on inte- gration of the plasmid into the chromosome. The deletions always reside in the EcoA fragment of pS1 and one end of the deletions is near the res site, suggesting that this site is probably involved in forming co-integrates with the host chromo-

some. The EcoA fragment of PSI seems to be

contain important DNA sequences which provide

regions for recombination with other genomes. This fragment also contains a recombination site

rsY789 at which it can form stable co-integrates

with another plasmid pox1054 [IS] and a site which is a ‘hot spot’ for site- and orientation- specific transposition of Tn5.51 from the chromo-

some [8]. The functional analysis of such regions

may help in the long run to design drugs/meth- ods to control the spread of resistance determi- nants on introduction of a new antibiotic.

Acknowledgements

M.S. was the recipient of the S&T Scholarship

rom ;;ay-89/G(5)3065) f the Government of Pak-

References

[II

[21

[31

[41

151

Sayre. P. and Miller. R. (1091) Bacterial mobile genetic

elements: importance in assessing the environmental fate of

genetically engineered sequences. Plasmid 26, I5 I - I7 I.

Lyon, B.R. and Skurray, R.A. (1987) Antimicrobial resis-

tance of Sruphylococcw ~I~IWUS: genetic basis. Microbial.

Rev. 51, 88-134.

Cohen. S.N. (1976) Transposable genetic elements and plas-

mid evolution. Nature 263, 73 l-738.

Saunders. J.R. (1984) Genetics and evolution of antibiotic

resistance. Br. Med. Bull. 40, 54-60.

Skurray, R.A.. Rouch. D.A.. Lyon, B.R., Gillespie. M.T.,

Tennent, J.M., Bryne, M.E.. Messerotti, L.J. and May. J.W.

(1988) Multiresistant Sruphdococcus UUMUS: genetics and

evolution of epidemic Australian strains. J. Antimicrob.

Chemother. 21 (Suppl. C). 19-38.

[h] Rowland. S.-J. and Dyke, K.G.H. (1989) Characterization of

the staphylococcal p-lactamase transposon Tn552. EMBO J.

8. 2761-2773.

[7] Rowland. S.-J. and Dyke, K.G.H. (1990) Tn552. a novel

transposable clement from Stcll~h~lococrus UY~US. Mol. Mi-

crohiol. 4. 961-975.

[8] Sohail. M.. Oldridge, M. and Dyke. K.G.H. (1995) Interac-

tion of the chromosomal Tn5.51 with two thermosensitive

derivatives. pS I and p-1D. of the plasmid ~19789 in .Stcqh\~-

loc~oc~c~us LIUT~I~.\. FEMS Microbial. Lett. 127. I65- 170.

[9] Novick, R.P. and Bouanchaud, D. (I 97 I ) Extrachromosomal

nature of drug resistance in Stryhylococcus CIUWU.F. Annals

N.Y. Acad. Sci. 182. 279-294.

[IO] Sambrook, J., Fritach, E.F. and Maniatis. T. (1989) Molecu-

lar Cloning: A Laboratory Manual. Cold Spring Harbor

Laboratory. Cold Spring Harbor, NY.

[I I] Johnston, L.H. and Dyke, K.G.H. (1971) Stability of penicil-

linase plasmids in Stclph~kjc,occ.lrs nureus. J. Bacterial. 107.

68-73.

[I21 Kreiswirth. B.H., LBfdahl, S., Betley. M.J., O’Reilly, M..

Schleivert. P.M., Bergdoll. M.S. and Novick. R.P. (1983)

The toxic shock syndrome exotoxin is not detectably trans-

mitted by a prophage. Nature (Land.) 305. 709-7 12.

[ 131 Kloos. W.E. and Schleifer. K.H. (1986) Staphylococcus. In:

Bergey’s Manual of Systematic Bacteriology (Williams. S.T..

Sharpe, M.E. and Holt. J.G., Eds.). Vol. 2. pp. 1013-1035.

Williams & Wilkins, Baltimore, MD.

[IA] Dyke, K.G.H. and Richmond, M.H. (1967) Occurrence of

various types of penicillinase plasmids among hospital

staphylococci. J. Clin. Pathol. 20. 75-79.

[IS] Conrad. B.. Bashkirov, V.I. and Hofemeister, J. (1992) Im-

precise excision of plasmid pEl94 from the chromosomes of

Btrcillus suhti/i.\ pEl94 insertion strains. J. Bacterial. 174,

6997-7002.

[ 161 Asheshov. E.H. (I 969) The genetics of penicillinase produc-

tion in Srcrphylococcus curreu.s strain PSSO. J. Gen. Micro-

biol. 59, 289-301.

[ 171 Johnston, L.H. and Dyke, K.G.H. (1974) Staphylococcal

penicillinase plasmids: studies on the reversion of a tempera-

ture-sensitive replication mutant to temperature stability. J.

Gen. Microbial. 82, 309-3 17.

[ 181 Sohail, M. and Dyke. K.G.H. (1995) Sites for co-integration

of large staphylococcal plasmids. Gene 162, 63-68.