Growth-Phase-Dependent Expression of the Pseudomonas putida ...

Vol. 172, No. 12

JOURNAL OF BACTERIOLOGY, Dec. 1990, p. 6651-6660

0021-9193/90/126651-10$02.00/0 Copyright X 1990, American Society for Microbiology

Growth-Phase-Dependent Expression of the Pseudomonas putida TOL Plasmid pWWO Catabolic Genes NICOLE HUGOUVIEUX-COTTE-PATTAT,t THILO KOHLER, MONIQUE REKIK, AND SHIGEAKI HARAYAMA* Department of Medical Biochemistry, Faculty of Medicine, University of Geneva, 1211 Geneva 4, Switzerland Received 25 May 1990/Accepted 7 September 1990

Pseudomonas putida TOL plasmid pWW0 catabolic genes are clustered into two operons. The first, the upper operon, is controlled by the xyLR regulatory gene, whereas the second, the meta operon, is controlled by the xylS regulatory gene. The xylS gene itself is subjected to control by xylR. In this study, we show that the TOL catabolic operons were poorly induced in cells growing at the early-exponential-growth phase but strongly induced in cells at late-exponential-growth phase. We constructed fusions of four TOL promoters, Pm (the promoter of the meta operon), Pu (the promoter of the upper operon), Ps (the promoter of the xylS regulatory gene), and Pr (the promoter of the xylR regulatory gene) with lacZ and examined, in Escherichia coli and P. putda, the expression of these promoters in relation to the growth phase. Expression from Pm, Pu, Ps, and Pr was almost constant if the host cells did not carry either xylS or xyLR. Similarly, expression of Pm and Pu in P. putida in the absence of XylS and XylR was constant during the growth of the cells. XylS-dependent transcription of Pm and XylR-dependent transcription of Ps and Pu, in contrast, varied with the growth phase. This observation suggested that the interaction of XylS and XylR with target promoters or with RNA polymerases was influenced by the growth phase. The nature of the signal which triggers the growth-phasedependent regulation was not clear. A change in the oxygen partial pressure was not responsible for the regulation. E. coli mutants defective in relA, crp, and cya exhibited growth-phase-dependent expression of the TOL catabolic genes, indicating that cyclic AMP and relA-dependent synthesis of ppGpp are not involved in this phenomenon.

The degradative pathway for toluene and xylenes encoded by TOL plasmid pWWO is one of pathways best characterized biochemically and genetically. Studies on the regulation of the catabolic genes on this plasmid have been carried out either in Pseudomonas putida, the organism in which the TOL plasmid was originally discovered, or in Escherichia coli. The model for the regulation of TOL catabolic genes proposed by Worsey and Williams (43) suggested the existence of two catabolic operons (the upper and meta operons) and two regulatory operons (the xylS and xylR operons) (12, 18, 20, 24, 26). The essence of this model is still valid (25, 36, 39). The upper operon encodes three enzymes of the TOL catabolic pathway which convert toluene and xylenes to benzoate and toluates (upper pathway). The expression of the upper operon is strongly induced by upper pathway substrates, namely toluene, xylenes, and (methyl)benzyl alcohols. This induction requires the xylR gene product (12, 26). The meta operon is located some 10 kb downstream from the upper operon and encodes a set of enzymes required for the transformation of benzoate and toluates to Krebs cycle intermediates (the meta pathway). The expression of the meta operon is induced by meta pathway substrates, namely benzoate and toluates. The xylS gene product is required for this induction (12, 24). The meta operon is also induced by the upper pathway substrates, and this induction is triggered by the overproduction of XylS as follows: the XylR protein, by interacting with toluene, xylenes, or (methyl)benzyl alcohols, induces not only the

upper operon but also the xylS gene. As a consequence of the induction, the concentration of the XylS protein reaches a level that causes activation of the meta operon promoter even in the absence of benzoate or toluates (25, 36, 39) (Fig. 1). Although this scheme describes major regulatory features of the TOL genes, there may exist other regulatory factors which influence the expression of the TOL catabolic genes. For example, transcription of the meta operon and of the xylR gene is mediated by RNA polymerase containing the conventional a factor (RpoD), whereas XylR-dependent transcription of the upper and xylS operons is mediated by the RpoN cr factor (11, 28) (Fig. 1). In this article, we demonstrate that induction of the TOL catabolic operons does not occur immediately after the addition of an inducer but only when bacterial cells enter a late-exponential-growth phase. This phenomenon was observed in both P. putida and E. coli. By constructing gene fusions of the TOL promoters with lacZ, we analyzed the expression of these promoters in relation to growth phases. From the analysis, we propose a model for the growth-phase-dependent regulation. MATERIALS AND METHODS Strains and plasmids. Strains and plasmids used are shown in Table 1. A set of isogenic E. coli strains was constructed by P1 transduction as described previously (17). Construction of lacZ translational fusions with the meta operon promoter (Pm), the upper operon promoter (Pu), the xylS operon promoter (Ps), the xylR operon promoter (Pr), and the promoter of the RpoN a factor gene was carried out as follows by using the promoter-probing vector pMLB1034

(38).

* Corresponding author. t Present address: Laboratory of Microbiology, National Institute of Applied Science at Lyon, 69621 Villeurbanne, France.

(i) Construction of Pm-lacZ. The 0.4-kb PstI fragment of TOL plasmid pWWO contains the promoter sequence of the 6651

6652

HUGOUVIEUX-COTTE-PATTAT ET AL.

J. BACTERIOL.

FIG. 1. Regulation of TOL catabolic genes. The TOL catabolic genes are organized into two operons. The first, the upper pathway operon, encodes catabolic enzymes which transform toluene and xylenes to benzoate and toluates. This operon contains five genes, (xylCMABN). The second operon, the meta operon, encodes thirteen genes (xy/XYZLTEGFJQKIH), whose products are responsible for the transformation of benzoate and toluates to Krebs cycle intermediates. Pu and Pm are promoters of the upper and meta pathway operons, respectively. Positive control ($) of the TOL operons are indicated by arrows. Transcription from the Pu promoter is positively regulated by the XylR protein (0), which is activated by toluene, xylenes, or (methyl)benzyl alcohols. This induction requires the RpoN factor. Transcription from the Pm promoter is positively regulated by the XylS protein (0). The XylS protein is activated by benzoate and toluates. Expression from the xyIS promoter (Ps) is also positively controlled by XylR, and this induction also requires the RpoN a factor. Therefore, the effectors of XylR induce the xylS gene, and the overproduction of XylS, in turn, provokes the induction of Pm in the absence of the effectors of XylS.

meta operon (23, 30). This fragment was cloned in the PstI site of the pUC7 plasmid (42), giving plasmid pNC6. The 0.4-kb BamHI fragment of pNC6 containing Pm was excised from pNC6 and reinserted into the BamHI site of the translational lac fusion vector pMLB1034. Light-blue transformants of MC1061 (7) were screened on antibiotic medium 3 (AM3) plates supplemented with 0.04% 5-bromo-4-chloro3-indolyl-j-D-galactoside (X-Gal) and ampicillin (Ap; 100 gxg/ml), and a plasmid isolated from one of these transformants was designated pNC227 (Fig. 2). In this plasmid, the lacZ gene was placed in the correct reading frame after the 90th codon of xylX, the first gene of the meta operon (19). To introduce the Pm-lacZ fusion into P. putida, pNC227 was cleaved with EcoRI and ligated with the EcoRI-cleaved broad-host-range vector pGS72 (37). The pNC227-pGS72 composite plasmid thus constructed was called pGTK727. (ii) Construction of Pu-lacZ. The 0.7-kb SmaI-XhoI fragment containing the upper operon promoter was isolated from pGSH2917 (16) and subcloned into pUC8 to construct pNC192. The 0.7-kb segment containing Pu was then isolated from pNC192 as the 0.7-kb EcoRI-HindIII segment. This segment was cleaved with HaeIII, and the 0.3-kb EcoRI-HaeIII segment thus generated was cloned into pMLB1034 between the EcoRI and SmaI sites. The plasmid pNC243 (Fig. 2) was isolated from, one of the light-blue transformants of MC1061 developed on AM3 plates supple-

TABLE 1. Strains and plasmids used for this study Strains and plasmids

Strain P. putida KT2440 E. coli AB259

A1373 A1376 CA8000 EJ500 GSH2130 GSH3491 LE392 MC1061

S0303 Plasmids pGS72

pGSH2917 pKT570 pMLB1034 pNC103

pNC111 pNC118 pNC192 pNC227 pNC243 pNC6 pNM187 pPL392 pGTK727 pGTK743 pUC7 pUC8 pWWO

Relevant marker(s)

Source or reference

PaWl cured of pWW0

43

HfrH thi-i relA1 spoT) crp rpsL derivative of EJ500 crp::TnS derivative of EJ500 HfrH re/Al F- prototroph Acya derivative of EJ500 rpsE rpsL+ derivative of MC1061 F- hsdRSl4 metB) lacY1 supE44 supF58 galK2 ga/T22 trpRS5 F- araDl39 A(ara-leu)7697 AlacX74 galU ga/K hsdR rpsL F deoR201 relAl metBI spoT1 rpsL254

K. Isono A. lida A. Iida K. Isono A. lida This study This study 29 7 K. Isono

RK2-based broad-host-range vector Lambda pL-based expression vector carrying the 7.5-kb HindIII-XbaI segment containing the Pu promoter pKT231 derivative carrying the 6.5-kb XhoT segment containing two regulatory genes Promoter-probing vector pUC7 derivative carrying the 0.6-kb BglII fragment of pWWO containing Ps and Pr pMBL1034 derivative carrying a Ps-lacZ fusion pMLB1034 derivative carrying a Pr-lacZ fusion pUC8 derivative carrying the 0.7-kb SmaI-XhoT segment of pWWO containing the Pu promoter pMLB1034 derivative carrying a Pm-lacZ fusion pMLB1034 derivative carrying a Pu-lacZ fusion pUC7 derivative carrying the 0.4-kb PsMI segment of pWWO containing the Pm promoter pKT231 derivative carrying Pm xylE and Ps xylS pBR322 derivative carrying Pm xylXYZLTEGFJQKIH and Ps xylS pNC227-pGS72 composite plasmid pNC243-pGS72 composite plasmid Cloning vector Cloning vector Wild-type TOL plasmid

37 16 12 40 This study This study This study This study This study This study This study 31 15 This study This study 42 42 43

GROWTH-PHASE-DEPENDENT EXPRESSION OF THE TOL GENES

VOL. 172, 1990

6653

PstlI

Pstl

BamHI/Bgilll £oIStul >

4

,g~~EcoRI

SmSal

Pstl

/bla

pNC11 8

lacZ

(6.8 kb)

FIG. 2. Structures of pNC plasmids carrying lacZ fusions with TOL promoters. The TOL promoters are indicated by large arrows, while lacZ (the structural gene for P-galactosidase) and bla (the structural gene for P-lactamase) are shown by small arrows.

mented with X-Gal and Ap. The sequencing of the DNA insert in pNC243 confirmed that the hybrid plasmid carries a 340-bp insert containing the Pu promoter; downstream from the Pu promoter, there is an open reading frame preceded by a putative ribosome-binding site (22), and the lacZ gene was present in the correct reading frame after the 20th codon of this open reading frame. The pNC243-pGS72 composite plasmid, named pGTK743, was constructed by the ligation of EcoRI-cleaved pNC243 with EcoRI-cleaved pGS72. (iii) Constructions of Ps-lacZ and Pr-lacZ. The 0.6-kb BglII fragment from plasmid pKT570 (12) contains two promoters, Ps and Pr, divergently transcribing two regulatory genes, xyIS and xyIR. This fragment was inserted into the BamHI site of pUC7 to give pNC103. The 0.6-kb EcoRI fragment from pNC103 was then cloned into the EcoRI site of the pMLB1034 vector. From the sequencing data of this region (24, 26), it was expected that the insertion of this fragment into the EcoRI site of pMLB1034 must give rise to a translational fusion of xyIR and lacZ in one orientation and a translational fusion of xylS and lacZ in the other orientation. Two types of blue colonies were obtained on AM3 plates containing X-Gal and Ap after the transformation of MC1061. Dark-blue clones contained the 0.6-kb EcoRI fragment inserted in one orientation, giving rise to a fusion of lacZ after the 30th codon of xylR. One of these plasmids was named pNC118 (Fig. 2). Light-blue clones contained the 0.6-kb EcoRI fragment in the opposite orientation, giving

rise to a fusion of lacZ after the 12th codon of xylS. One of these plasmids was named pNC111 (Fig. 2). (iv) Construction of an rpoN-lacZ fusion. The 0.5-kb EcoRIAluI fragment of pNTR1 (28) was cloned between the EcoRI and SmaI sites of pMLB1034. According to the DNA sequence of the rpoN gene (27), the resultant plasmid carries the fusion of lacZ after the 27th codon of rpoN. Media and growth conditions. M9 minimal medium, AM3, and L broth have been described previously (16, 29). For anaerobic cultures in L broth, 20 mM KNO3 and 0.2% (wt/vol) glucose were added. As an inducer, 2 mM m-toluate or m-methylbenzyl alcohol was included in the medium. In the case of P. putida, cells were grown either in L broth or in glucose minimal medium at 30°C. Overnight cultures were diluted into fresh medium. When necessary, 2 mM m-methylbenzyl alcohol or m-toluate was added at that time. In the case of E. coli, cells were grown in L broth containing Ap and, if necessary, streptomycin at 25 jig/ml at 30°C. Overnight cultures were diluted by a factor of 100 into fresh L broth and cultivated for about 2 h. When the cell density of the cultures reached 0.06 to 0.1 with A660 samples for time zero were harvested. m-Toluate or m-methylbenzyl alcohol (2 mM) was added when necessary to the cultures at time zero.

DNA sequencing. Sanger dideoxy sequencing was performed with T7 DNA polymerase (Pharmacia, Uppsala, Sweden).

6654


J. BACTEPIOL.

lo'

100[

10.

L-broth ..1-4-- L-broth + mTol

o Q

L-broth

o

M9+ mTol + Mg

l

|mMBA

C1 OO-2 Ao

A-2B2

*~~~~2000~~

~00

00-

200

eU 1000

200 0

0

A-3B3 *

80 60

80 I 60

-

40

4

M 20

2

0 0

1012345~~~~~ 2 3 4 5

Time (hours)

-

_

0 0

2

4

6

Time (hours)

8

FIG. 3. Growth and induction of TOL catabolic genes in P. putida(pWWO). (A-1) Growth of KT2440(pWWO) in L broth measured by A660; (A-2) induction of catechol 2,3-dioxygenase (C230) (milliunits per milligram of protein), the enzyme encoded in the meta operon, in cells grown in L broth; (A-3) induction of benzyl alcohol dehydrogenase (BADH) (milliunits per milligram of protein), the enzyme encoded in the upper operon, in cells grown in L broth; (B-i) growth of KT2440(pWWO) in glucose minimal medium; (B-2) induction of catechol 2,3-dioxygenase (milliunits per milligram of protein) in cells grown in glucose minimal medium; (B-3) induction of benzyl alcohol dehydrogenase (milliunits per milligram of protein) in cells grown in glucose minimal medium. Symbols: A, no inducer; *, 2 mM m-toluate (mTol); K, 2 mM m-methylbenzyl alcohol (mMBA).

Enzyme assays. Methods for preparation of crude extracts (16) and assays for catechol 2,3-dioxygenase (15), benzyl alcohol dehydrogenase (16), P-lactamase (34), and P-galactosidase (32) have been described previously. One unit of enzyme was defined as the amount transforming 1 pmol of substrate per min, except P-galactosidase, for which 1 U was defined by Miller (32) as the amount hydrolyzing 1 nmol of o-nitrophenyl-f-D-galactopyranoside per min. RESULTS Induction of the TOL catabolic operons in P. putida. The induction kinetics of TOL catabolic genes in P. putida KT2440(pWWO) grown in L broth were followed by measuring the activity of catechol 2,3-dioxygenase (the xylE product), which is encoded in the meta operon, and the activity of benzyl alcohol dehydrogenase (the xylB product), which is encoded in the upper operon. As expected, m-methylbenzyl alcohol induced both the upper and meta operons, whereas

m-toluate induced only the meta operon. The induction, however, did not occur immediately after the addition of these inducers but after a latent period extending up to 3 h (Fig. 3). Induction of the catabolic genes was further delayed when the culture was diluted in fresh medium containing the inducers (Fig. 4). This observation demonstrated that the TOL catabolic genes were not fully induced in cells growing exponentially in L broth. When cells were grown in M9 glucose minimal medium containing an inducer (m-toluate or m-methylbenzyl alcohol), induction occurred within 1 h; however, more significant induction occurred 3 h later when the growth phase of the cells became late exponential (Fig. 3). As the cell density of a culture increases, the concentration of nutrients decreases, the oxygen partial pressure decreases, and the pH of the medium becomes more acidic. A factor(s) other than a decrease in the oxygen partial pressure and a change in pH is responsible for the growth-


VOL. 172, 1990

40N

C)

'IC

0

*V4

Time (hours) FIG. 4. Culture of P. putida(pWWO) was either grown continuously in L broth containing 2 mM m-methylbenzyl alcohol (0) or diluted into fresh L broth containing 2 mM m-methylbenzyl alcohol to about 0.075 (A660) (-). The induction of catechol 2,3-dioxygenase activity (C230) and benzyl alcohol dehydrogenase activity (BADH) (both measured in milliunits per milligram of protein) was followed in these cultures.

phase-dependent induction, because a similar delay (4 h) for the induction of the TOL catabolic genes was observed in cells growing anaerobically in L broth containing 20 mM KNO3 and 0.2% glucose and in cells growing at pH 7.5 in L broth containing 100 mM potassium phosphate (data not shown). Expression of the TOL catabolic and regulatory operons in E. coi. The RSF1010-based pNM187 carries xyIE, the structural gene for catechol 2,3-dioxygenase, and xylS, the positive regulator for Pm (31). Expression of the plasmid xylE gene is under the control of Pm. In E. coli LE392 containing pNM187, catechol 2,3-dioxygenase was induced by m-toluate because m-toluate activates XylS, which in turn induces transcription from Pm. The specific activity of catechol 2,3-dioxygenase in the E. coli strain grown in L broth, however, started to increase only 4 h after the addition of 2 mM m-toluate, when the growth phase of the cells became late exponential. The specific activity of catechol 2,3-dioxygenase was constant in cells of LE392(pNM187) during growth in a medium containing no m-toluate, as has been observed by Mermod et al. (31). Therefore, the increase in enzyme activity in cells of LE392(pNM187) grown in L broth

6655

containing 2 mM m-toluate at late-exponential-growth phase was not due to the increase in copy number of pNM187 at this growth stage. We also tested the expression of Pm on the pNM187 plasmid in different E. coli strains bearing global regulatory mutations (crp, cya, and relA). Amino acid starvation provokes the so-called stringent response, which is induced by accumulation of unusual guanine nucleotides, ppGpp and pppGpp. The synthesis of the guanine nucleotides is mediated by the relA product (8). The RelA- mutants are, therefore, defective in the stringent response. The expression of catechol 2,3-dioxygenase from pNM187 was examined in relA mutants CA8000, S0303, and AB259. These mutants showed growth-phase-dependent expression of catechol 2,3-dioxygenase. The data indicated that stringent control is not responsible for growth-phase-dependent regulation. This observation also rules out the possibility that TOL catabolic genes are induced at the late-exponentialgrowth phase when free RNA polymerases become abundant after the arrest of transcription of rRNA and tRNA genes. If the latter hypothesis is the case, the induction of the TOL catabolic genes at the end of the exponential growth phase should be minimal in the E. coli relA mutant, since this mutant continues to transcribe the rRNA and tRNA genes even in the late-exponential-growth phase. Cyclic AMP is an important molecule which modulates catabolite repression (33). We therefore tested the expression of xylE on pNM187 in crp (strains A1373 and AI376) and cya (GSH2130) mutants along with their isogenic parent, EJ500. No significant difference in the induction kinetics of catechol 2,3-dioxygenase was observed among these four strains (data not shown). This result indicated that the expression of TOL catabolic genes requires neither cyclic AMP nor the catabolite gene activator protein positive regulator and therefore that the growth-phase-dependent regulation is not due to the classic catabolite repression mediated by cyclic AMP. Several mechanisms are conceivable for this regulation. For example, the promoters of the upper and meta operons could be regulated directly by the growth phase. Alternatively, expression of the regulatory genes xylS and xyIR may be growth phase dependent. Furthermore, one should consider the possibility that transportation of the inducers into bacterial cells is affected by the growth phase. To examine these possibilities, four TOL promoters (Pu, Pm, Pr, and Ps) were cloned into a promoter probing-vector (pMLB1034) to give plasmids pNC243, pNC227, pNC111, and pNC118, respectively. These plasmids were introduced into E. coli GSH3491, and the plasmid-encoded activities of ,-galactosidase and P-lactamase were examined in this strain. P-Lactamase encoded by these plasmids is expressed constitutively, and its activity reflects the copy number of the plasmids (41). P-Lactamase activity was almost constant during the growth of the cells, except that in some experiments the enzyme activity increased by two to three times at the late-exponential-growth phase. We calculated the ratio of ,-galactosidase activity to P-lactamase activity. Figure 4A shows the change in the ratios during the growth of the bacteria. The ratios for Pm, Ps, and Pr remained practically constant during the growth of the cells, indicating that transcription of Pm, Ps, and Pr in the absence of any positive regulator was essentially growth phase independent. The fluctuations in the ratio of P-galactosidase to P-lactamase values observed in Fig. 5A-1 to 5A-3 were within day-to-day variation. However, the values for Pu at the end of exponential growth were consistently two to three times higher

6656


101

#

20

150

\010 1,

10-2

~. 2

4

Time (hours) 101 O

1,0

B-1 Ps + XylSR

100

0,8

#

0,6

4

-S lo- 1

-0,4S

V~0 10-14

.0; I

4

6

02

10

101

Ir

0o c

0

100

-4 ~

10-

0,06

""

0,04 %--

'0,02 .0

i

C=

0,010

0,008

0,006

8,

0,004 04

0,002 c n

6

'0,8 +

0,6

1

'0,4

'0,2 c 0

B4 Pu+XylSR+mMBA

i

10-1'

-m

2

0O

0

2

t

Time (hours)

Time (hours)

'

'0,08

A-4 Pu_

0

-

J. BACTERIOL.

11

lu-s.

4

Time (hours) than those in the early-exponential-growth phase. The threetimes difference in Pu expression, however, was negligible compared with the difference in Pu expression in the presence of XylR (more than 100 times; see below). The levels of expression of these promoters were markedly different, and

-

2

4

2,0 .1,5

1n .0,5

of 0 .

6

Time (hours)

the

P-galactosidase/03-lactamase ratios of Pu, Ps, Pm, and Pr

about 1, 5, 20, and 5,000, respectively. The activities of Ps, Pu, and Pm were also examined in the presence of pKT570, an RSF1010-based plasmid carrying the regulatory genes xylS and xylR (12). We showed above

were


VOL. 172, 1990

0,10

C-1 Pm + XylSR

101

0,08 +

0

°

100

0,04

10-1

-

0,02 1 0

10-2

2.

4

Time (hours)

6

0

cm

tol 100

CM.

@ 10-1 nIJ;

-"

0

2

4

6

Time (hours) FIG. 5. Activities of the TOL promoters expressed in the E. coli GSH3491. E. coli cells were grown in L broth. Symbols: O, growth curve; *, ratio of P-galactosidase to 1-lactamase activity. 13-Galactosidase is expressed from the TOL promoters cloned in pMLB1034, whereas 13-lactamase is encoded on the vector DNA. (A-1) Expression of Ps in GSH3491(pNC111); (A-2) expression of Pr in GSH3491 (pNC118); (A-3) expression of Pm in GSH3491(pNC227); (A-4) expression of Pu in GSH3491(pNC243); (B-1) expression of Ps in GSH3491(pKT570, pNC111); (B-2) expression of Ps in GSH3491 (pKT570, pNC111) in the presence of 2 mM m-methylbenzyl alcohol (mMBA); (B-3) expression of Pu in GSH3491(pKT570, pNC243); (B4) expression of Pu in GSH3491(pKT570, pNC243) in the presence of 2 mM mMBA; (C-1) expression of Pm in GSH3491(pKT570, pNC227); (C-2) expression of Pm in GSH3491(pKT570, pNC227) in the presence of 2 mM m-toluate (mTol); (C-3) expression of Pm in GSH3491(pKT270, pNC227) in the presence of 2 mM mMBA.

that the copy number of pNM187, another RSF1010-based plasmid, may not change during the growth of E. coli cells. We therefore assume that the copy number of pKT570 is constant in different growth phases. The derivatives of

6657

GSH3491 containing pKT570 in addition to pNC111, pNC227, or pNC243 were constructed, and the ratios (PIgalactosidase/,B-lactamase) were examined in relation to cell growth in L broth. Interestingly, xyIR-dependent transcription of Ps in both the presence and absence of m-methylbenzyl alcohol was strongly growth phase dependent (Fig. 5B). In the absence of any inducer, the expression from Ps in GSH3491(pKTS70, pNCi11) at the early-exponential-growth phase was at the same level as Ps expression in GSH3491 (pNC111). Ps expression in GSH3491(pKT570, pNC111) at the mid- and late-exponential-growth phases was, in contrast, 10 and 100 times higher than Ps expression in GSH3491 (pNC111). The XylR positive regulator synthesized by the high-copy-number pKT570 plasmid therefore induced the Ps promoter even in the absence of its effector. Transcription from Ps in GSH3491(pKT570, pNC111) was further stimulated by m-methylbenzyl alcohol. An apparent induction occurred within 2 h after addition of the inducer, and expression reached a level 1,000-fold higher than that in GSH3491(pNC111). xylR-dependent expression of the Pu promoter also varied with the growth phase. In the absence of m-methylbenzyl alcohol, the P-galactosidase activity in GSH3491(pKT570, pNC243) remained low at the early-exponential-growth phase but increased as the culture moved into the mid- to late-exponential phases. Therefore, XylR induced Pu in the absence of its effector. Unexpectedly, the effect of m-methylbenzyl alcohol on the induction of the Pu promoter at the early-exponential-growth phase was less significant than that on the induction of the Ps promoter (Fig. SB). Expression from the Pm promoter was examined in GSH3491(pKT570, pNC227) (Fig. SC). In the absence of any effector, Pm activity in this strain was similar to that in GSH3491(pNC227). In contrast, the Pm promoter in GSH 3491(pKT570, pNC227) was partially induced within 1 h after the addition of m-toluate; more significant induction, however, occurred at 4 h, when the growth rate of the cells started to decline. The induction of Pm by m-methylbenzyl alcohol followed similar induction kinetics to that of m-toluate, although induction by m-methylbenzyl alcohol occurred indirectly as follows: m-methylbenzyl alcohol activated the xylR product, which induced xylS, and the overproduction of the xylS product then stimulated tran' scription from Pm. Although the absolute ratios (P-galactosidase/p-lactamase) varied significantly in different cultures of the same clone (standard deviations were 30 to 65%), the patterns of the induction curves were reproducible. Expression of Pm and Pu in P. putida in the absence of any positive regulator. Two broad-host-range plasmids, pGTK743 carrying a Pu-lacZ fusion and pGTK727 carrying a Pm-lacZ fusion, were introduced in P. putida KT2440, and the expression of Pu and Pm in the absence of XylS and XylR was examined in this strain. P-Galactosidase activity expressed from Pu and Pm in KT2440 exhibited small fluctuation (1 to 3 U/mg of protein for Pm and 0.1 to 0.4 U/mg of protein for Pu) during growth, which was negligible compared with the growth-phase-dependent variation seen in Fig. 3. Therefore, as expected from the E. coli results, the expression of Pm and Pu in the absence of XylS and XylR was growth phase independent in P. putida. Expression of the rpoN promoter in P. putida. The Pu and Ps promoters are recognized by the RpoN u factor but not by the conventional RpoD a factor (11, 28). It is therefore possible that the growth-phase-dependent expression of Pu and Ps is the consequence of the growth-phase-dependent

6658


expression of RpoN. To examine this possibility, expression of the rpoN promoter was examined in KT2440, by using the translational fusion of rpoN with lacZ. We found that the expression of rpoN in KT2440 was constant during the growth of the cells (data not shown). DISCUSSION Analysis by using the TOL promoter-lacZ fusions revealed some interesting features of the mechanism of their growth-phase-dependent regulation. Transcription from Pm, Pu, Pr, and Ps was not growth phase dependent in the absence of XylS and XylR. However, XylS-dependent transcription of Pm and XylR-dependent transcription of Ps and Pu were growth phase dependent. These observations suggested that the interaction between these positive regulators and the target promoters or RNA polymerases was influenced by the growth phase. It is unlikely that the entry of the inducers of XylS and/or XylR is growth phase dependent because (i) m-methylbenzyl alcohol and m-toluate could induce, albeit at a low efficiency, the target promoters at the early exponential growth phase, and (ii) growth-phase-dependent induction of Pu and Ps was observed even in the absence of m-methylbenzyl alcohol. How could the cooperation between the XylS or XylR positive regulators and their target promoters be influenced by the growth phase? We can conceive of several mechanisms: (i) some negative regulator represses Ps, Pm, and Pu in the early-exponential-growth phase; (ii) some positive regulator in addition to XylS or XylR is necessary for the induction of TOL promoters, and this unidentified positive regulator is only expressed at the late-exponential-growth phase; (iii) there is a growth-phasedependent modification of a DNA structure, e.g., supercoiling, and this change affects the affinity of XylS and XylR for the target promoters or RNA polymerases. We are currently examining the effect of mutations in two component systems (5) and attempting to isolate mutants defective in growthphase-dependent regulation to investigate these possibilities. Since P-galactosidase activities expressed from pNC111 and pNC118 reflected the actual translational activities of xylS and xylR, respectively, we calculated the intracellular concentrations (in moles per liter) of XylS and XylR under the following assumptions: (i) the specific activities of the XylS-LacZ and XylR-LacZ fusions are the same as that of P-galactosidase, namely 300,000 U/mg of protein (32); (ii) 109 bacteria correspond to a cellular volume of 2 x 10-6 liter and to 0.3 mg of protein (1); (iii) the stability of XylS and XylR are similar to that of P-galactosidase; (iv) the degree of expression of ,-galactosidase is proportional to the copy number of its structural gene (the copy numbers of pMLB1034 [a pBR322 derivative] and pKT570 [an RSF1010 derivative] are 60 and 10, respectively [4, 40]). Then the concentration of XylS or XylR expressed from pKT570 in E. coli was calculated to be (specific activity of ,-galactosidase from pNC111 or pNC118)/(specific activity of pure P-galactosidase) x 10-3/(molecular weight of P-galactosidase) x 1/ (volume of cells containing 1 mg of protein) x (copy number of pKT570)/(copy number of pMLB1034) = (specific activity of P-galactosidase from pNC111 or pNC118)/300,000 x 10-3/116,351 x 0.3/(2 x 10-6) x 10/60. The expression of ,-galactosidase from Pr was almost constant during cell growth in a batch culture and about 7,500 U/mg of protein. The intracellular concentration of XylR synthesized from pKT570 was then estimated to be about 5 ,uM. The expression of Ps in GSH3491(pKT570, pNC111) at the earlyexponential-growth phase in the absence of m-methylbenzyl

J. BACTERIOL.

alcohol was equal to that in GSH3491(pNC111) and was about 25 U of P-galactosidase per mg of protein, which corresponded to 0.02 puM. XylS concentration in GSH3491 (pKT570, pNC111) increased to 0.3 RM at the mid-exponential-growth phase and to S puM at the late-exponential-growth phase. In the presence of m-methylbenzyl alcohol, XylS concentration increased from 0.7 pLM (early) to 4 puM (mid) and reached 16 pLM (late). The intracellular concentration of cloned promoters in pMLB1034 could be calculated from the copy number of the plasmid and is about 0.05 p.M. Since the concentration of XylR was about 300 times higher than the concentration of Ps and Pu, only a small fraction of intracellular XylR may be involved in the activation of these promoters. The induction of Pm by m-toluate was dependent on the growth phase in cells containing pKT570 probably because of two reasons: (i) XylS concentration increased at a lateexponential-growth phase (Fig. SB-1); (ii) the activity of XylS on the Pm promoter increased at the late-exponentialgrowth phase. Although the concentration of XylS increased more than 300 times at the late-exponential-growth phase to reach S p.M, this concentration of XylS did not stimulate the transcription of Pm in the absence of m-toluate (Fig. SC-1). In contrast, XylS at 0.02 p.M stimulated Pm in the presence of m-toluate. Therefore, the affinity of XylS to the Pm promoter is strongly elevated by m-toluate. Since m-methylbenzyl alcohol induced the Pm promoter indirectly through an overproduction of XylS, XylS at a concentration of 16 p.M could stimulate the transcription of Pm in the absence of m-toluate. From the copy number of pMLB1034, which may be close to pBR322, the concentration of the plasmid (i.e., the concentration of a promoter on pMLB1034) was calculated to be 0.05 ,uM. The concentration of XylS at the early-exponential-growth phase in E. coli containing pKT570 was about 0.02 ,uM as described above. It is striking that XylS at this concentration could stimulate the transcription of Pm to 1,300 U of ,B-galactosidase per mg of protein even at the early-exponential-growth phase. If most of Pm forms a complex with XylS at the late-exponential-growth phase in the presence of m-toluate to induce P-galactosidase (to 23,000 U/mg of protein), then about 5% (1,300/23,000 x 100) of Pm should form a complex with XylS in the presence of m-toluate at the early-exponential-growth phase. The equilibrium constant, K, of this reaction is then calculated from the equation K = [XylS] [Pm]/[XylS-Pm] = (0.02 ,uM - 0.05 ,uM x 0.05) (0.05 puM x 0.95)/(0.05 puM x 0.05) and was about 0.3 in the early-exponential-growth phase and may further decrease in the late-exponential-growth phase. This result indicated that XylS in the presence of m-toluate exhibits strong affinity for Pm. Growth-phase-dependent regulation has been observed in many genes in many bacteria. Among these are the E. coli mob operon responsible for Microcin B17 production (9); the E. coli genes for glycogen biosynthetic enzymes (35), for phosphoenolpyruvate carboxylase (14), and for acid phosphatase (3, 10); the E. coli bolA gene involved in shape determination (2); the Serratia genes for protease (6) and phospholipase (13); and the Erwinia chrysanthemi genes for pectinases and cellulase (21). We do not find -any common role in the metabolic and physiological functions of these gene products. However, the common occurrence of growth-phase-dependent regulation among different bacterial species suggests that this type of control may have a very important role in bacterial life cycles, especially in survival under starvation conditions. The mechanism(s) and genes governing this growth-phase-dependent regulation remain to


VOL. 172, 1990

be elucidated. It is important to determine whether or not the conclusion of this study, which is that the interaction between positive regulators and promoters is sensitive to the growth phase, could be applied generally to other growthphase-dependent phenomena. Aldea et al. (2) proposed a consensus sequence for the promoters induced during the transition from the exponential to stationary growth phase. However, this consensus sequence was not found in Pm, Pu, Ps,

or

Pr. It is therefore possible that

a

multitude of control

systems exist to regulate various growth-phase-dependent

promoters. ACKNOWLEDGMENTS We thank K. N. Timmis for encouragement in the initial stage of this project, A. lida and K. Isono for strains, E. Falvey for careful reading of the manuscript, and F. Rey for secretarial work. This study was supported by a grant from the Swiss National Foundation to S.H. and by an EMBO long-term fellowship to N.H.-C.-P. LITERATURE CITED 1. Alberts, B., D. Bray, J. Lewis, M. Raff, K. Roberts, and J. D. Watson. 1983. Molecular biology of the cell. Garland Publishing Inc., New York. 2. Aldea, M., T. Garrido, C. Hernandez-Chuco, M. Vicente, and S. R. Kushner. 1989. Induction of a growth-phase-dependent promoter triggers transcription of bMlA, an Escherichia coli morphogene. EMBO J. 8:3923-3931. 3. Atlung, T., A. Nielsen, and F. G. Hansen. 1989. Isolation, characterization and nucleotide sequence of app Y, a regulatory gene for growth-phase-dependent gene expression in Escherichia coli. J. Bacteriol. 171:1683-1691. 4. Barth, P., and N. J. Grinter. 1974. Comparison of the deoxyribonucleic acid molecular weights and homologies of plasmids conferring linked resistance to streptomycin and sulfonamides. J. Bacteriol. 120:618-630. 5. Bourret, R. B., J. F. Hess, K. A. Borkevich, A. A. Pakula, and M. I. Simon. 1989. Protein phosphorylation in chemotaxis and two-component regulatory systems of bacteria. J. Biol. Chem. 264:7085-7088. 6. Braun, V., and G. Schmitz. 1980. Excretion of a protease by Serratia marcescens. Arch. Microbiol. 124:55-61. 7. Casadaban, M., and S. N. Cohen. 1980. Analysis of gene control signals by DNA fusion and cloning in Escherichia coli. J. Mol. Biol. 138:179-207. 8. Cashel, M., and K. E. Rudd. 1987. The stringent response, p. 1410-1438. In F. C. Neidhardt, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium. American Society for Microbiology, Washington, D.C. 9. Connell, N., Z. Han, F. Moreno, and R. Kolter. 1987. An E. coli promoter induced by the cessation of growth. Mol. Microbiol. 1:195-201. 10. Dassa, E., M. Cahu, B. Desjoyaux-Cherel, and P. L. Bocquet. 1982. The acid phosphatase with optimum pH of 2.5 of Escherichia coli. J. Biol. Chem. 257:6669-6676. 11. Dixon, R. 1986. The xylABC promoter from the Pseudomonas putida TOL plasmid is activated by nitrogen regulatory genes in Escherichia coli. Mol. Gen. Genet. 203:129-136. 12. Franklin, F. C. H., P. R. Lehrbach, R. Lurz, B. Rueckert, M. Bagdasarian, and K. N. Timmis. 1983. Localization and functional analysis of transposon mutations in regulatory genes of the TOL catabolic pathway. J. Bacteriol. 154:676-685. 13. Givshov, M., L. Olsen, and S. Molin. 1988. Cloning and expression in Escherichia coli of the gene for extracellular phospholipase Al from Serratia liquefaciens. J. Bacteriol. 170:5855-5862. 14. Goldie, A., and B. Sanwal. 1980. Genetic and physiological characterization of Escherichia coli mutants deficient in phophoenolpyruvate carboxykinase activity. J. Bacteriol. 141: 1115-1121. 15. Harayama, S., P. R. Lehrbach, and K. N. Timmis. 1984.

6659

Transposon mutagenesis analysis of meta cleavage pathway operon genes of the TOL plasmid of Pseudomonas putida mt-2. J. Bacteriol. 160:251-255. 16. Harayama, S., R. A. Leppik, M. Rekik, N. Mermod, P. R. Lehrbach, W. Reineke, and K. N. Timmis. 1986. Gene order of the TOL catabolic plasmid upper pathway operon and oxidation of both toluene and benzyl alcohol by the xylA product. J. Bacteriol. 167:455-461. 17. Harayama, S., E. T. Palva, and G. L. Hazelbauer. 1979. Transposon-insertion mutants of Escherichia coli K12 defective in a component common to galactose and ribose chemotaxis. Mol. Gen. Genet. 171:193-203. 18. Harayama, S., and M. Rekik. 1990. The meta-cleavage operon of TOL degradative plasmid pWWO comprises 13 genes. Mol. Gen. Genet. 221:113-120. 19. Harayama, S., M. Rekik, and K. N. Timmis. 1986. Genetic analysis of a relaxed ring dioxygenase toluate 1,2-dioxygenase, encoded by TOL plasmid pWWO of Pseudomonas putida. Mol. Gen. Genet. 202:226-234. 20. Harayama, S., M. Rekik, M. Wubbolts, K. Rose, R. A. Leppik, and K. N. Timmis. 1989. Characterization of five genes in the upper-pathway operon of TOL plasmid pWWO from Pseudomonas putida and identification of the gene products. J. Bacteriol. 171:5048-5055. 21. Hugouvieux-Cotte-Pattat, N., S. Reverchon, G. Condemine, and J. Robert-Baudoux. 1986. Regulatory mutations affecting the synthesis of pectate lyase in Erwinia chrysanthemi. J. Gen. Microbiol. 132:2099-2106. 22. Inouye, S., A. Nakazawa, and T. Nakazawa. 1984. Nucleotide sequence surrounding the transcription initiation site of xylABC operon on TOL plasmid of Pseudomonas putida. Proc. Natl. Acad. Sci. USA 81:1688-1691. 23. Inouye, S., A. Nakazawa, and T. Nakazawa. 1984. Nucleotide sequence of the promoter of the xylDEFG operon on TOL plasmid of Pseudomonas putida. Gene 29:323-330. 24. Inouye, S., A. Nakazawa, and T. Nakazawa. 1986. Nucleotide sequence of the regulatory gene xylS on the Pseudomonas putida TOL plasmid and identification of the protein product. Gene 44:235-242. 25. Inouye, S., A. Nakazawa, and T. Nakazawa. 1987. Overproduction of the xylS gene product and activation of the xylDLEGF operon on the TOL plasmid. J. Bacteriol. 169:3587-3592. 26. Inouye, S., A. Nakazawa, and T. Nakazawa. 1988. Nucleotide sequence of the regulatory gene xylR of the TOL plasmid from Pseudomonas putida. Gene 66:301-306. 27. Kohler, T., J. M. Cayrol, J.-L. Romas, and S. Harayama. 1989. Nucleotide and deduced amino acid sequence of the rpoN cr-factor of Pseudomonas putida. Nucleic Acids Res. 17:10125. 28. Kohler, T., S. Harayama, J.-L. Ramos, and K. N. Timmis. 1989. Involvement of Pseudomonas putida RpoN cu factor in regulation of various metabolic functions. J. Bacteriol. 171:4326-4333. 29. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 30. Mermod, N., P. R. Lehrbach, W. Reineke, and K. N. Timmis. 1984. Transcription of the TOL plasmid toluate catabolic pathway operon of Pseudomonas putida is determined by a pair of coordinately and positively regulated overlapping promoters. EMBO J. 3:2461-2466. 31. Mermod, N., J. L. Ramos, P. R. Lehrbach, and K. N. Timmis. 1986. Vector for regulated expression of cloned genes in a wide range of gram-negative bacteria. J. Bacteriol. 167:447-454. 32. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 33. Neidhardt, F. C. 1987. Multigene systems and regulons, p. 1313-1317. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium. American Society for Microbiology, Washington, D.C. 34. O'Callaghan, C. H., A. Morris, S. M. Kirby, and A. H. Shingler. 1972. Novel method for detection of ,B-lactamase by using a chromogenic cephalosporin substrate. Antimicrob. Agents Chemother. 1:283-288.

6660


35. Okita, T., R. Rodriguez, and J. Preiss. 1981. Biosynthesis of bacterial glycogen. J. Biol. Chem. 256:6944-6952. 36. Ramos, J.-L., N. Mermod, and K. N. Timmis. 1987. Regulatory circuits controlling transcription of TOL plasmid operon encoding meta-cleavage pathway for degradation of alkylbenzoates by Pseudomonas. Mol. Microbiol. 1:293-300. 37. Servaraj, G., and V. N. Iyer. 1985. A small mobilizable IncP group plasmid vector packageable into bacteriophage X capsids in vitro. Plasmid 13:70-74. 38. Silhavy, T., M. Berman, and L. Enquist. 1984. Experiments with gene fusions. Cold Spring Harbor Laboratory. Cold Spring Harbor, N.Y. 39. Spooner, R. A., M. Bagdasarian, and F. C. H. Franklin. 1987. Activation of the xyIDLEFG promoter of the TOL toluenexylene degradation pathway by overproduction of the xylS

J. BACTERIOL.

regulatory gene product. J. Bacteriol. 169:3581-3586. 40. Stuber, D., and H. Bujard. 1982. Transcription from efficient promoters can interfere with plasmid replication and diminish expression of plasmid specific genes. EMBO J. 1:1399-1404. 41. Uhlin, B. E., and K. Nordstrom. 1978. R plasmid gene dosage effects in Escherchia coli: copy mutants of the R plasmid R1 drd-19. Plasmid 1:1-7. 42. Vieira, J., and J. Messing. 1982. The pUC plasmids, an M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19:259-268. 43. Worsey, M. J., and P. A. Williams. 1975. Regulation of the degradative pathway enzymes coded for by the TOL plasmid (pWWO) from Pseudomonas putida mt-2. J. Bacteriol. 134:757764.