JOURNAL OF BACRIOLOGY, Nov. 1992, p. 7026-70320021-9193/92/217026-07$02.00/0Copyright © 1992, American Society for Microbiology

Vol. 174, No. 21

The Phenotypes of Temperature-Sensitive Mini-RK2Replicons Carrying Mutations in the Replication Control Gene

trfA Are Suppressed Nonspecifically by Intrageniccop Mutations

KARE HAUGAN, PONNIAH KARUNAKARAN, JANET MARTHA BLATNY, AND SVEIN VALLA*UNIGEN Centerfor Molecular Biology, The University of Trondheim, N-7005 Trondheim, Nonvay

Received 26 May 1992/Accepted 1 September 1992

The minimal replicon of the broad-host-range plasmid RK2 consists of the origin of vegetative replication(oriW and a gene (trjA) encoding an essential replication protein that binds to short repeats in oriV. We reporthere the results of a DNA sequence analysis of seven unique mutants that are temperature sensitive forreplication in Escherichia coil. The mutations (designated rts) were distributed throughout 40% of thedownstream part of the trjA gene. Spontaneous revertants of the ts mutants were isolated, and further analysisof four such revertants demonstrated that the new phenotypes resulted from intragenic second-site copy up(cop) mutations. Subcloning experiments showed that all tested intragenic combinations of rts and copmutations resulted in elimination or strong reduction of the temperature sensitivity of replication. Thissuppression was also observed under conditions where the mutant TrfA protein was provided in trans withrespect to oriV, indicating that the reduction in temperature sensitivity could not be a TrfA protein dosageeffect. The phenotypes of two of the cop mutants in Pseudomonas aeruginosa were analyzed; the resultsdemonstrated that the mutants were either not functional or poorly functional in this host. The Yts mutantplasmids were also reduced in their ability to replicate in P. aeruginosa, and the intragenic cop mutations didnot improve the fumctionality of these mutants. The significance of the results is discussed in relation to currentmodels of the mechanism of action of the TrfA protein.

RK2 is a 60-kb self-transmissible plasmid belonging toincompatibility group IncP1. These plasmids are particularlywell known for their ability to replicate in a large number ofbacterial species (18). The minimal RK2 replicon consists oftwo genetic elements, the origin of vegetative replication(oriV) and a gene encoding an initiator protein (TrfA) thatbinds to short repeated sequences (iterons) present in theorigin (9, 10). The trfA gene contains two in-frame transla-tional initiation signals, resulting in expression of two formsof the TrfA protein (33 and 44 kDa) (6, 15). TrfA-33 has beenshown to be sufficient for replication in some hosts, includ-ing Escherichia coli (15), whereas TrfA-44 is required forefficient replication in Pseudomonas aeruginosa (3). Theseobservations may imply that the two TrfA proteins are partof an adaptation to replication in different hosts.The copy number of RK2 replicons in the cells may to

some extent be limited by the concentration of the TrfAprotein; when the concentration was increased by a factor of2 to 3 over a certain basal level, a copy number increase ofabout 30% was observed (4). Further increases of the TrfAconcentration (up to 170-fold) had no additional effect on thecopy number. Certain mutations in the trfA gene, on theother hand, have been found to result in up to a 23-fold-higher copy number (5). The TrfA protein therefore actsboth as a positive factor required for replication and as anegative regulator. The mechanism by which RK2 utilizesthe copy number regulation potential of the TrfA protein invivo is not yet fully understood, but a model involvingTrfA-mediated inhibition of replication by intermolecularcoupling of origins has been suggested (7).We recently described the isolation of 14 temperature-

* Corresponding author.

sensitive mutants in the trfA gene of minimal RK2 replicons,and the phenotypes of these mutants were characterized inE. coli and to some extent in P. aeruginosa (20). Thetemperature requirements for replication of these mutantsvaried over a wide range, and they were also sensitive to themutant TrfA protein dosage. In addition, experiments withP. aeruginosa indicated that the phenotypes were stronglydependent on unidentified host-specific parameters in theintracellular environment.

In the present report we describe the results of a molecularanalysis of the temperature-sensitive mutants. The datashow that the mutants represent seven different mutationsand that these mutations are distributed both upstream anddownstream of the previously reported mutations givingrise to a copy up phenotype. Molecular characterizationof spontaneous and in vitro-constructed revertants of thetemperature-sensitive mutants demonstrated that copy upmutations act as nonspecific intragenic suppressors of tem-perature-sensitive mutants. These observations are not im-mediately obvious from the model for copy number regula-tion by intermolecular coupling described above.

MATERIALS AND METHODSBacterial strains and plasmids. The bacterial strains and

plasmids used in this study are listed in Table 1.Growth of bacteria, conjugative matings, and standard

molecular biology techniques. E. coli cells were grown in Lbroth or on L agar (11), and P. aeruginosa was grown in M9medium (8). Matings were performed on membranes at 30°Cwith S17.1 containing the relevant plasmids as the donorstrain, and P. aeruginosa transconjugants were selected at23°C on M9 agar medium supplemented with 100 ,ug ofcarbenicillin per ml. Transformations were performed by the

7026

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MUTATIONS IN THE REPLICATION CONTROL GENE trfA OF RK2 7027

TABLE 1. Bacteria and plasmids used in this study'

Bacterial strain or plasmid Properties Source or reference

E. coliDH5a endA41 hsdR17 supE44 thi-1 X- recAl gyrA96 reL41 AlacU169 (+80dlacZAM15) Bethesda Research

LaboratoriesHB101 proA2 recA13 ara-14 lacYl hsdS20 (rB- mB-) galK2 xyl-5 mtl-l rpsI20 supE44 11BL21 hsdS gal rB mB 17S17.1 RP4 2-Tc::Mu-Km::Tn7pro res mod' 16

P. aeruginosa leu-38 hsdR 14PA01161

PlasmidspUC9 Apr 21pRD110-34 pBR322 replicon where an EcoRI-PstI fragment was substituted with the trfA gene 5

from plasmid RK2, TcrpBK3 Expression vector for trfA-33, Apr 7pBK20 pUC19 with 5 oniV iterons from RK2 cloned in polylinker, Apr 7pTJS65 pUC8 where the 0.7-kb oriV fragment from RK2 was cloned in the polylinker, Apr 14pSV16 RK2 replicon, requires trfA in trans, Apr Kmr 20pSV6 RK2 replicon, Apr Cmr 20pFFl Derivative of pSV6 where oniT was inserted from RK2, ApT Cm' 5pTJS42 RK2 replicon containing 5 iterons from onV, Tcr 13pRD110-34rts213S trfA mutant, isolated from pRD110-34 20pRD110-34rts247C trfA mutant, isolated from pRD110-34 20pRD110-34rts313L trfA mutant, isolated from pRD110-34 20pSV6rts230V trfA mutant, previously designated pSV6ts54 20pSV6rts3l6V trfA mutant, previously designated pSV6ts9O 20pSV6rts332F trfA mutant, previously designated pSV6ts114 20pSV6rts374A trfA mutant, previously designated pSV6ts97 20pSV6rts316Vcopl71W trfA double mutant, isolated as spontaneous revertant of pSV6rts3l6V This studypSV6rts316Vcopl73K trfA double mutant, isolated as spontaneous revertant of pSV6rts316V This studypSV6rts374Acop203L trfA double mutant, isolated as spontaneous revertant of pSV6rts374A This studypSV6rts316Vcop254D trfA double mutant, isolated as spontaneous revertant of pSV6rts316V This study

a Apr, ampicillin resistance; Cmr, chloramphenicol resistance; Kmr, kanamycin resistance; TcT, tetracycline resistance. All trfA mutant plasmids not listed inthe table were constructed in vitro as described in Materials and Methods.

method of Chung et al. (2), and transformants were selectedat 23°C when temperature-sensitive plasmid mutants wereused. Preparation of plasmid DNA, restriction endonucleasedigestions, agarose gel electrophoresis, filling in of 5' over-hangs with Klenow DNA polymerase, removal of 3' over-hangs with T4 DNA polymerase, random priming, andligations were performed by standard protocols (11).

Isolation of revertants, localization of mutations, and DNAsequencing. Spontaneous revertants of temperature-sensitivetrfA mutants were isolated by plating HB101 cells containingthe temperature-sensitive mutants on L agar (106 to 108 cellsper plate) supplemented with 30 ,ug of chloramphenicol perml and 200 ,ug of ampicillin per ml and then incubating at42°C. Before DNA sequencing, mutations were localized bysubstituting each of three different parts of the wild-type trfAgene (in pFF1 or pRD110-34) with the corresponding frag-ment from the mutant gene. The three fragments weregenerated by digestion of the relevant plasmids with EcoRI-SfiI, SfiI-NdeI, and NdeI-PstI, respectively. Fragmentswere isolated from low-melting-point agarose gels, and liga-tions of fragments were performed directly by mixing ali-quots of the relevant melted gel pieces. To simplify subclon-ing involving NdeI, we used a pRD110-34 derivative inwhich the NdeI site in the vector part was removed by afill-in reaction with Klenow DNA polymerase. The trfArestriction fragment responsible for the temperature-sensi-tive or copy up phenotype was identified either by plating oftransformants under selective conditions at 23 and 42°C(temperature-sensitive mutants) or by comparing the inten-

sity of plasmid bands on agarose gels (copy up mutants).Mutants in pRD110-34 derivatives were identified aftertransformation with pSV16.

In vitro-constructed double mutants were made by sub-cloning the relevant parts of the trfA gene with the restrictionendonucleases described above. In cases where these en-zymes did not separate the two mutations, PflMI was usedas an alternative. To simplify clonings with this enzyme, thetwo PflMI sites in the vector part of pRDllO-34 were firsteliminated by deleting the 49-bp PflMI fragment. The 3'overhangs were removed before circularization with theexonuclease activity of T4 DNA polymerase. All mutationswere sequenced in the relevant mutant derivatives of pFF1or pRD110-34 by the chain termination method (12). Com-mercially available primers were used for sequencing reac-tions initiated in the vector part. Four oligonucleotides weresynthesized for sequencing reactions starting within the trfAgene. The nucleotide sequences of these primers were 5'-TGCGAGCTGAAATAGTC-3', 5'-AGGAAATCGTCGTGCTG-3', 5'-TATCGAACAAGGAAAGC-3', and 5'-TATGACGACCAAGAAGC-3'.

Plasmid copy number determinations. Copy numbers weremeasured by the method of Durland and Helinski (4), withthe following exceptions. Plasmid-containing DH5a cellswere washed in 0.9% NaCl-0.1% NaN3 and then filteredonto a nylon membrane (Amersham, Hybond-N) with aBio-Rad Bio-Dot cell. The DNA was fixed to the membranewith shortwave UV light, and the membranes were sub-jected to hybridization with the 710-bp 32P-labelled oniV

VOL. 174, 1992

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

7028 HAUGAN ET AL.

Ts mutant (rts)

ATGTr fA-44

Revertant (cop)

(I)

I

230VGTC 313L TTG 316VGTC

213STCT ,J 247CTGC 2FTTC 3744ATGA &-

Q.

ATG

T r f A--33 DI171WACG 173KAA 203LCT 254DIGAC

FIG. 1. Map of rts and cop mutations. The mutant number designations correspond to the amino acid residues affected by the mutations;the first methionine (from the amino-terminal end) in the 44-kDa protein represents residue 1. The uppercase letters after the numbers indicatethe new amino acid resulting from the mutation, and the three-letter indices indicate the new DNA sequence at the corresponding codon. Thesymbol A is used to indicate a deletion, and the hatched box ( ) shows the region where cop mutants have been reported previously (5).

EcoRI-BamHI fragment of plasmid pTJS65 as a probe. Theprobe was labelled by random priming. Hybridization wascarried out at 65°C in 0.5 M sodium phosphate buffer-7%sodium dodecyl sulfate (SDS)-1 mM Na2EDTA (pH 7.2).The membranes were washed in sodium phosphate buffer(pH 7.2) at 65°C (four times in 40 mM buffer-1% SDS andthen twice in 20 mM buffer-0.5% SDS). After exposure onX-ray film, the dot intensities were quantified with an LKB2202 Ultrascan densitometer.

In vitro replication assays. The 33-kDa wild-type andcopl71W TrfA proteins were purified from BL21(pBK3)cells to approximately 50% purity as described by Perri et al.(9) down to (and including) the heparin-Sepharose CL 6Bstep. The in vitro replication assays were performed asdescribed by Kittel and Helinski (7).

RESULTS

DNA sequence analysis of the temperature-sensitive mu-tants. The 14 temperature-sensitive mutants with mutationsin the trfA gene were analyzed at the molecular level (Fig. 1).Only seven of the mutants turned out to be unique, and thesewere distributed from the codons representing amino acids213 through 374, which is close to the 3' end of the gene. Sixof the seven mutations were caused by 1-bp substitutions,whereas one mutation was caused by an out-of-frame 3-bpdeletion. This deletion resulted in loss of an aspartic acidresidue at position 374, eight amino acids from the carboxy-terminal end of the protein. Among the six 1-bp substitutionmutants, five resulted in replacement of uncharged aminoacids (at positions 213, 230, 313, 316, and 332). One mutationresulted in substitution of a positively charged amino acid(arginine at position 247) with cysteine; this residue islocalized only three amino acids from the previously de-scribed copy up mutation cop250V (5). It was also interest-ing to note that the mutation in the codon representing aminoacid 213 resulted in substitution of a proline residue withserine and that three of the other mutations affected aminoacids localized very close to proline residues (mutations313L, 316V, and 332F).

Intragenic suppression of the temperature-sensitive mutantsby copy up mutations. As described in a previous paper (20),the temperature-sensitive mutants reverted spontaneously atlow but easily detectable frequencies. Such revertants havenow been studied in more detail, and in these studies theTrfA protein was expressed either in cis (from plasmid pSV6or pFF1) or in trans (from the pBR322 derivative pRD110-34) with respect to oriV. Figure 2 illustrates the nature ofthese plasmids in more detail and also includes data on

plasmid pSV16, which was used in the experiments with trfAgene expression in trans.

Preliminary analysis of the revertant cells indicated thatthey frequently appeared to contain high-copy-number plas-mids. Subcloning experiments with a revertant isolated frompSV6rts3l6V confirmed that the new phenotype was causedby a mutation in the trfA gene, and nucleotide sequenceanalysis showed that this mutation resulted in substitution ofthe glycine residue at position 254 with aspartic acid (Fig. 1).Surprisingly, this particular mutation is identical to thepreviously described copy up mutation cop254D (5). Thedata on revertant pSV6rts3l6V thus gave the first indicationthat copy up mutations may act as intragenic suppressors oftemperature-sensitive mutations.An obvious question related to the observations described

above is whether the suppression is a general property ofcop254D or a specific property of the particular temperature-sensitive mutant rts316V. To analyze this problem, weconstructed double mutants (in plasmid pFF1) consisting ofcop254D and each of the four temperature-sensitive mutantsrts213S, rts313L, rts332F, and rts374A. Analysis of thephenotypes of these four double mutants showed that theycould replicate at 42°C, indicating that the ability to suppresstemperature-sensitive mutants is a general property ofcop254D. It was particularly interesting to note that thesuppression must be very efficient, since the two tempera-ture-sensitive mutants rts213S and rts313L could not beestablished in cis (in pSV6/pFF1) even at low temperature.To analyze the efficiency of the suppression in more detail,

we compared the copy numbers of the five different doublemutants with the copy number of wild-type pFF1 andpFFlcop254D. The analysis was performed both directly bymeasurements of copy numbers and indirectly by analyzingthe levels of ampicillin resistance expressed by the cellscontaining the different plasmids (19). The results of theseexperiments (Fig. 3) demonstrate that the copy numbers ofthe five double mutants are higher than the copy number ofthe wild type but lower than that of cop254D. The individualdouble mutants varied significantly in their copy numbers,and there appeared to be a reasonably good correlationbetween the copy numbers measured directly and the levelsof ampicillin resistance expressed by the cells.The observation that cop254D appears to suppress tem-

perature-sensitive mutants regardless of their location in thetrfA gene also raises the question of whether this property isassociated with copy up mutants in general or whether it isrestricted to cop254D. To study this problem we isolated andsequenced two new spontaneous revertants of rts316V andone revertant of rts374A. The results of these experiments

i-fa

J. BACTERIOL.

I

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MUTATIONS IN THE REPLICATION CONTROL GENE trfA OF RK2 7029

FIG. 2. Map of plasmids used for characterization of trfA mutantphenotypes. pFF1 is an RK2 replicon used for expression of trfA incis, and pRD110-34 (a derivative of pBR322) was used for expres-sion of trfA in trans. pSV16 is an RK2 replicon lacking the trfA gene,and this plasmid was used as the target plasmid for the TrfA protein

Pstl expressed in trans. Plasmid pSV6, which was used in some of theexperiments, is the same as pFF1, except that pSV6 lacks oriT.pRD110-34 (the vector part) contains an NdeI site and two PflMIsites not shown. As described in Materials and Methods, these sites

) Ndel were eliminated to simplify the experimental work.

(Fig. 1) showed that the two rts316V revertant phenotypeswere caused by 1-bp substitutions affecting amino acidresidues 171 (alanine substituted by tryptophan) and 173(glutamine substituted with lysine), respectively. Similarly,

SflI reversion of rts374A was caused by a 1-bp substitutionresulting in replacement of the glutamine residue at position203 with leucine. To analyze the nature of these newmutations in more detail, they were subcloned into thewild-type trfA gene of pFF1 such that the mutations respon-sible for the temperature-sensitive phenotypes were elimi-nated. The copy numbers of these new plasmid constructswere determined (Table 2). All of the revertants (copl71W,copl73K, and cop2O3L) are present in the cells in elevatedcopy numbers (similar to those in cop254D), indicating thatthe generality of the suppression effect is not restricted tocop254D. Further evidence for this conclusion was obtainedby constructing the double mutant rts247Ccopl71W. Tem-perature-sensitive mutant rts247C alone is poorly functionalin vivo and could not be established in cis (in pSV6 or pFF1).The double mutant, on the other hand, was fully functionalin cis at 30°C but not 37C (data not shown).

PstI 25

NdeI 6

(I, ~~~~~~~~~~~~~EE E

15-4CL~~~~~~~~~~~~~~~0~~~~~~~~~~~~~~~

2CL0~~~~~~~~~~~~~~~~

5-

mutans 213S 332F 374A 313L 316V mutations

tion an4ysis;1, max rMum aleels of ampicil resistanceeexFIG. 3. Determination of relative copy numbers (wild type = 1)of pFFlcop254D and double mutants of cop254D and different rtsmutants. OJ, copy numbers determined by DNA-DNA hybridiza-tion analysis; M, maxiimum levels of ampicillin resistance ex-pressed by the cells. These levels were determined by platingapproximately 200 cells on agar medium containing different con-centrations of the antibiotic (500-p,g/ml increase in the concentra-tions in each tested step). All measurements were performed twiceor more, and deviations were as indicated. Incubations of cells wereperformed at 37°C.

EcoRl

EcoRI

VOL. 174, 1992

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

7030 HAUGAN ET AL.

TABLE 2. Relative copy numbers of spontaneouspFF1 cop mutantse

cop mutant Relative Maximum ampicilincopy no. tolerance (mg/ml)

Wild type 1 0.5copl71W 24.0 ± 4.0 4.5copl73K 18.3 ± 0.4 5.0cop2O3L 19.3 ± 4.6 5.0cop254D 23.6 ± 2.0 5.0

a The deviations in the copy number determinations are based on two ormore independent experiments. Maximum ampicillin tolerance levels weredetermined as described in the legend to Fig. 3.

We showed previously that the temperature sensitivity oftrfA mutants was strongly reduced relative to expressionfrom pSV6 when the mutant TrfA protein was expressedfrom pRD110-34 (20). The reason for this seemed to be thatpRD110-34 derivatives express about 10 times more TrfAprotein than do pSV6 derivatives, since we are not aware ofany reason why expression in cis or trans in itself shouldaffect the phenotypes. HB101 was used as the host in theseexperiments, and we have since found that at least twophenotypes can be generated from cells containing pSV16and a given pRD110-34 trfA temperature-sensitive mutant.Analysis of the plasmid content of cells displaying the lowestlevel of temperature sensitivity (like those reported previ-ously) showed that these cells contained new plasmid struc-tures, possibly cointegrates of pRD110-34 and pSV16. Suchcells were apparently easily enriched in growing cell culturescontaining trfA temperature-sensitive mutants. We havefound that these enrichments can be avoided if DH5Sa is usedas host for the temperature-sensitive mutants; Table 3 showsthe effect on phenotypes as a function of the plasmid used forexpression of the TrfA mutant protein. All four rts mutantstested are less temperature sensitive in the trans test system(pRD110-34 plus pSV16) than are the corresponding mutantsexpressed in cis (pFF1). The reductions in temperature

TABLE 3. Phenotypes of rts and rts cop mutants in the trfA geneof plasmid pFFi or pRD110-34"

Ampicillin Growth at:Plasmid genotype concn

(ji.g/ml) 23°C 30°C 37°C 42°CpFFl/pRD110-34 (wild type) 20 + + + +pFFlrts23V 20 + - - -pRD110-34rts230V 20 + + + -pFFlrts316V 20 + + - -pRD110-34rts316V 20 + + + -pFFlrts332F 20 + + + +pRD110-34rts332F 20 + + + +pFFlrts374A 20 + + + -pRD110-34rts374A 20 + + + +

pRD110-34rts213S 200 + - - -pRD110-34rts213Scop254D 200 (+) + + +pRD110-34rts247C 200 + - - -pRD110-34rts247Ccopl71W 200 + + + -

I E. col DH5a was used as the host in all the experiments, and cellscontaining pRD110-34 or its mutant derivatives (as indicated) also containedpSV16. The cells were grown in liquid medium at 23'C in the presence of 100pg of ampicillin per ml, diluted, and plated on agar medium (about 200 cellsper plate) containing 20 or 200 1ig of ampicillin per ml. The plates wereincubated overnight (2 days for incubations at 23'C) at the temperaturesindicated and then inspected for growth. (+) indicates poor growth.

0

E

0

0

._

00

0.

100 I

50+

total protein (pg)

2

FIG. 4. Iteron inhibition of in vitro pTJS42 replication as a

function of increasing concentrations of the TrfA wild-type (x) or

Cop171W (0) protein. Replication was assayed by measuring theincorporation of 3H-labelled dlTP. Iterons were added in the formof plasmid pBK20 in a molar ratio of 3:1 with respect to the iteronsin plasmid pTJS42. Corresponding amounts of plasmid pUC9 DNAwere added as controls (no iterons added). Protein concentrationsare given as total protein, since the two TrfA proteins were notpurified to homogeneity (see Materials and Methods).

sensitivity are less significant than those reported previouslyfor the same mutants (20), but we have since found that thephenotypes in DH5a and HB101 are the same if the selectionof more resistant derivatives (see above) in the latter strain isavoided.

It seemed that there might be a relationship between thesuppression of the temperature-sensitive phenotypes byintragenic copy up mutations and the effects on phenotypesof the TrfA protein dosage. To analyze this possibility, wecompared the phenotypes in a situation where the differentTrfA mutant proteins were expressed in trans (frompRD110-34) instead of in cis (from pSV6/pFF1 derivatives).This difference made it possible to eliminate the probableside effect of increased copy numbers leading to increasedtrfA gene expression. The two double mutants tested aremuch less temperature sensitive than the corresponding rtsmutants alone (Table 3). On the basis of these experiments,the possibility that differences in intracellular levels of theTrfA mutant proteins explain the suppression of rts mutantphenotypes by intragenic copy up mutations could be ex-cluded.

Analysis of the properties of the TrfA Copl71W protein.The copy up mutants previously reported (5) were all clus-tered in a fairly small region of the trfA gene (Fig. 1),whereas the three new copy up mutations described herewere localized significantly upstream of this cluster. Thedifference in the localization of the mutations might meanthat the new mutant TrfA proteins belong to a class that isfunctionally different from that of the copy up mutantsdescribed by Durland et al. (5). To study this problem, weanalyzed the properties of the TrfA protein produced bycopl71W in an in vitro replication assay (Fig. 4). The

x Noiterons0 added

x

0

0 o+5 iterons

0

x +5 iterons

J. BAcTERiOL.

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MUTATIONS IN THE REPLICATION CONTROL GENE trfA OF RK2 7031

TABLE 4. Carbenicillin tolerance of pFFl rts and cop mutants inP. aeruginosa at 42°C

Growth in the presence of carbenicillin attrfA mutant the following concn (jg/ml):

100 800 1,600

Wild type + + +rs247C + -rts313L + -rts332F + +rts374A + +copl71W + -rts247Ccopl71W + -

a Cells were plated on agar medium (approximately 200 cells per plate)containing the carbenicillin concentrations indicated. After incubation over-night, the plates were inspected for growth. Plasmid-free cells did not grow inthe presence of 100 p.g of carbenicillin per ml. The mutants cop254D, rts213S,rts213Scop254D, rts313Lcop254D, rts316Vcop254D, rts332Fcop254D, andrts374Acop254D could not be transferred to and stably maintained in P.aenuginosa, presumably because they were not functional in this host.

on V-containing plasmid pTJS42 replicated efficiently in thepresence of a heterologous plasmid (pUC9), whereas theaddition of oniV iterons (in plasmid pBK20) to the assaysystem led to a strong inhibition of replication, as describedby Kittel and Helinski (7). In contrast, when the in vitroreplication of pTJS42 was mediated by the Cop171W pro-tein, the iteron inhibition effect was much weaker. The newlocalizations of the cop mutations described here thereforedo not appear to reflect new properties at the functionallevel.

Analysis of functionality and suppression of temperature-sensitive mutants in P. aeruginosa. In our previous paper weshowed that the pFFl temperature-sensitive mutants testedwere functional even at 42°C in P. aeruginosa, although thecarbenicillin resistance levels expressed by the plasmidswere somewhat reduced relative to the level expressed bythe pFF1 wild type (20). The DNA sequence analysis de-scribed in this report showed that we had access to fivemutants that were different from the two previously tested inP. aeruginosa. Three of these (rts213S, rts247C, andrts313L) were nonfunctional in cis in E. coli but could beestablished in strain S17.1, which expresses wild-type TrfAfrom the chromosome (16). This strain can also be used forconjugative mobilization of RK2 replicons, and we appliedthis method to transfer the five temperature-sensitive mu-tants to P. aeruginosa. Characterization of the phenotypesof the corresponding transconjugants showed that four of thefive mutants replicated at 42°C in this host (Table 4). The twomutants that could be established in cis in E. coli (ts332Fand rts374A) expressed a level of carbenicillin resistance inP. aeruginosa at 42°C that was somewhat reduced relative tothat of the corresponding wild-type plasmid. Two of theremaining mutants (rts247C and rts313L) expressed lowerlevels of carbenicillin resistance under these conditions, andone (rts213S) was not functional.The reduced functionality of the temperature-sensitive

mutants in P. aeruginosa at 42°C raised the question whetherthe copy up mutations will restore functionality, as they doin E. coli. To test this, we first transferred pFF1cop171Wand pFFlcop254D. pFF1cop171W could be established in P.aeruginosa, but the level of carbenicillin resistance ex-pressed by the transconjugant was reduced compared withthat of the wild type (Table 4). The other tested mutant,pFFlcop254D, could not be stably established in P. aerug-inosa. These experiments thus indicated that copy up muta-

tions in E. coli do not have a similar phenotype in P.aeruginosa. In agreement with this conclusion, we alsofound that the reduced functionality of pFFlrts247C was notsuppressed by the copl71W mutation in P. aeruginosa(Table 4). In addition, we found that the double mutants ofpFFl containing the cop254D mutation and the es mutations213S, 313L, 316V, 332F, and 374A were nonfunctional in thishost.

DISCUSSION

The results of the molecular analysis of the temperature-sensitive mutants showed that the mutations causing thesephenotypes were distributed throughout about 40% of thegene. Three of the mutations were localized upstream, andfour were localized downstream, of the copy up cluster ofmutations described previously (5). These results may indi-cate that the temperature-sensitive mutations represent twofunctional domains in the TrfA protein, but it seems equallypossible that the protein is folded in such a way that the tworegions are kept close together in the three-dimensionalstructure. The identification of new copy up mutationsupstream of all of the rts mutations also suggests thatfunctional domains cannot be easily identified simply on thebasis of the linear distribution map of the different kinds ofmutations.The observation that the temperature-sensitive mutant

phenotypes could be suppressed nonspecifically by intra-genic copy up mutations was quite unexpected, although itwas observed previously that cop mutants may suppress(intragenically) a temperature-sensitive mutation in a proteininvolved in replication of plasmid pSC101 (1). The intermo-lecular coupling model proposed for RK2 replicons does notseem to predict the results presented here, since cop mutantsshould apparently (according to the model) affect only theupper level of plasmid copies per cell. This does not meanthat the model is incorrect, and our data may rather beinterpreted to mean that the model does not give a completedescription of the biological effects of cop mutations. Themajor problem is to understand how the suppression effectcan be nonspecific with respect to the nature of the two kindsof mutations present in the same gene. Increases in TrfAprotein dosages alone were shown not to be sufficient toexplain the observed phenotypes, but one might assume thatthe copy up mutations lead to an increase in the number oforigins (compared with that in the wild type) during growthof the cells at the permissive temperature. It seems possiblethat after transfer to the nonpermissive temperature thepreexisting high number of origins has an effect similar tothat of a high dose of the TrfA protein, thus enhancing therequired TrfA-oniV interactions. The advantage of thismodel is that it explains the nonspecificity of the suppres-sion, whereas the major weakness is that it is difficult to seehow the suppression effect can be retained permanentlyunder nonpermissive conditions.

Formally, it also seems possible that the cop mutationslead to enhanced stability of the different rts mutant pro-teins, although it is difficult to see why so many different copmutations should lead to protein stabilization. It is alsoknown that cop mutations do not lead to elevated levels ofTrfA protein in the cells (5), indicating that TrfA proteinturnover is not affected by the cop mutations.Another model that might explain the observed suppres-

sion effect is based on the assumption that the temperature-sensitive mutations cause intermolecular coupling that isstronger (at the nonpermissive temperature) than that of the

VOL. 174, 1992

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

7032 HAUGAN ET AL.

wild-type protein, thereby blocking replication even at verylow copy numbers. According to this model, the suppressingeffect of the copy up mutations is due to their reduction ofthis strong coupling induced by the first mutation. The invitro data obtained with the copl71W protein are also inagreement with this model, which therefore also is consis-tent with the intermolecular coupling model. One weaknesswith this explanation is that one would have to assume thatall of the rts mutants tested for suppression (six of seven)cause stronger intermolecular coupling at the nonpermissivetemperature.

It is known that several host factors are required for RK2replication (10), and one may therefore imagine that thereare as yet unknown interactions that are relevant to thesuppression effect. In support of this is the fact that copmutants in E. coli may be nonfunctional or poorly functionalin P. aeruginosa. This observation is difficult to explain ifone assumes that the proposed intermolecular interactionsalone are sufficient to regulate the plasmid copy number. Webelieve that the best way to answer the questions raised bythe results presented in this report is to purify the differentmutant proteins and analyze their properties in vitro. Suchstudies are therefore now in progress in our laboratory.

ACKNOWLEDGMENTS

We thank Donald R. Helinski, Barbara Kittell, and Joan Lin forhelp with the in vitro replication assays.This work was supported by a grant from The Norwegian Re-

search Council for Science and the Humanities.

REFERENCES1. Armstrong, K. A., R Acosta, E. Ledner, Y. Machida, M.

Pancotto, M. McCormick, H. Ohtsubo, and E. Ohtsubo. 1984. A37 x 103 molecular weight plasmid-encoded protein is requiredfor replication and copy number control in the plasmid pSC101and its temperature-sensitive derivative pHSl. J. Mol. Biol.175:331-348.

2. Chung, C. T., S. L. Niemela, and R. H. Miller. 1989. One-steppreparation of competent Escherichia coli: transformation andstorage of bacterial cells in the same solution. Proc. Natl. Acad.Sci. USA 86:2172-2175.

3. Durland, R. H., and D. R. Helinski. 1987. The sequence encod-ing the 43-kilodalton trfA protein is required for efficient repli-cation or maintenance of minimal RK2 replicons in Pseudomo-nas aeruginosa. Plasmid 18:164-169.

4. Durland, R. H., and D. R. Helinskd. 1990. Replication of thebroad-host-range plasmid RK2: direct measurement of intracel-lular concentrations of the essential TrfA replication proteinsand their effect on plasmid copy number. J. Bacteriol. 172:3849-3858.

5. Durland, R. H., A. Toukdarian, F. Fang, and D. R. Helinski.1990. Mutations in the trfA replication gene of the broad-host-range plasmid RK2 result in elevated plasmid copy numbers. J.Bacteriol. 172:3859-3867.

6. Figurski, D. H., and D. R. Helinski. 1979. Replication of anorigin-containing derivative of plasmid RK2 dependent on aplasmid function provided in trans. Proc. Natl. Acad. Sci. USA76:1648-1652.

7. Kittel, B. L., and D. R. Helinski. 1991. Iteron inhibition ofplasmid RK2 in vitro: evidence for intermolecular coupling ofreplication origins as a mechanism for RK2 replication control.Proc. Natl. Acad. Sci. USA 88:1389-1393.

8. Maniatis, T., E. F. Fritsch, and J. Sambrook 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

9. Perri, S., D. R. Helinski, and A. Toukdarian. 1991. Interactionsof plasmid-encoded replication initiation proteins with the originof DNA replication in the broad host range plasmid RK2. J.Biol. Chem. 266:12536-12543.

10. Pinkney, M., R. Diaz, E. Lanka, and C. M. Thomas. 1988.Replication of mini RK2 plasmid in extracts of Escherichia colirequires plasmid-encoded protein TrfA and host-encoded pro-teins DnaA, B, G, DNA gyrase and DNA polymerase III. J.Mol. Biol. 203:927-938.

11. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring HarborLaboratory Press, Cold Spring Harbor, N.Y.

12. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci.USA 74:5463-5467.

13. Schmidhauser, T. J., M. Filutowicz, and D. R. Helinski. 1983.Replication of derivatives of the broad host range plasmid RK2in two distantly related bacteria. Plasmid 9:325-330.

14. Schmidhauser, T. J., and D. R. Helinski. 1985. Regions ofbroad-host-range plasmid RK2 involved in replication and sta-ble maintenance in nine species of gram-negative bacteria. J.Bacteriol. 164:446-455.

15. Shingler, V., and C. M. Thomas. 1984. Analysis of the trfAregion of broad host-range plasmid RK2 by transposon muta-genesis and identification of polypeptide products. J. Mol. Biol.175:229-249.

16. Simon, R, U. Priefer, and A. Ptihler. 1983. A broad host rangemobilization system for in vivo genetic engineering: transposonmutagenesis in Gram negative bacteria. Bio/Technology 1:784-791.

17. Studier, F. W., and B. A. Moffatt. 1986. Use of bacteriophage T7RNA polymerase to direct selective high-level expression ofcloned genes. J. Mol. Biol. 189:113-130.

18. Thomas, C. M., and D. R. Helinski. 1989. Vegetative replicationand stable inheritance of IncP plasmids. In C. M. Thomas (ed.),Promiscuous plasmids of Gram-negative bacteria, p. 1-25. Ac-ademic Press, Inc., New York.

19. Uhlin, B. E., and K. Nordstrom. 1977. R plasmid gene dosageeffects in Escherichia coli K-12: copy mutants of the R plasmidRldrd-19. Plasmid 1:1-7.

20. Valia, S., K Haugan, R H. Durland, and D. R. Helinsld. 1991.Isolation and properties of temperature-sensitive mutants of thetrfA gene of the broad host range plasmid RK2. Plasmid25:131-136.

21. Vieira, J., and J. Messing. 1982. The pUC plasmids, anM13mp7-derived system for insertion mutagenesis and sequenc-ing with synthetic universal primers. Gene 19:259-268.

J. BACTERIOL.

on Decem

ber 9, 2020 by guesthttp://jb.asm

.org/D

ownloaded from