Molecular Genetics and Genomics

, Volume 274, Issue 2, pp 180–188

Bacterial luciferase activity and the intracellular redox pool in Escherichia coli

Authors

  • K. Koga
    • Institute of Molecular and Cellular BiosciencesThe University of Tokyo
  • T. Harada
    • Institute of Molecular and Cellular BiosciencesThe University of Tokyo
  • H. Shimizu
    • Institute of Molecular and Cellular BiosciencesThe University of Tokyo
    • Institute of Molecular and Cellular BiosciencesThe University of Tokyo
Original Paper

DOI: 10.1007/s00438-005-0008-5

Cite this article as:
Koga, K., Harada, T., Shimizu, H. et al. Mol Genet Genomics (2005) 274: 180. doi:10.1007/s00438-005-0008-5

Abstract

In this study, we analyzed the activity of a bacterial luciferase (LuxAB of Vibrio fischeri) expressed under the control of a consensus-type promoter, lacUV5, in Escherichia coli, and found that activity declines abruptly upon entry into the stationary growth phase. Since this decline was reproducibly observed in strains cultured in various growth media, we refer to this phenomenon as ADLA (Abrupt Decline of Luciferase Activity) and define the time point when activity begins to decline as T0. Because the levels of luciferase proteins (LuxA and LuxB) remained constant before and after T0, ADLA cannot be due to the repression of luciferase gene expression. Further analyses suggested that a decline in the supply of intracellular reducing power for luciferase was responsible for ADLA. We also found that ADLA was alleviated or did not occur in several mutants deficient in nucleoid proteins, suggesting that ADLA is a genetically controlled process involved in intracellular redox flow.

Keywords

Escherichia coliGrowth phase transitionLuciferaseNucleoid proteinsRedox regulation

Introduction

Under conditions of nutrient starvation, Escherichia coli enter into a physiological state termed the stationary phase. Although they are not considered as differentiated, stationary-phase E. coli cells have many properties in common with dormant spores, which are characterized by the ability to survive prolonged periods of starvation and by strong resistance to multiple-stresses (Hengge-Aronis 1993). The transition from logarithmic growth to stationary phase in E. coli batch cultures is an important and complex program for differentiation, during which numerous changes were observed in every subcellular compartment—the outer and inner membranes, the periplasm and the cytoplasm (Huisman et al. 1996). The copy number of chromosomal DNA in the cell decreases because initiation of chromosomal DNA replication is arrested while cell division continues (Nanninga 2001). The nucleoid is condensed as a result of changes in the major protein components that protect its DNA against oxidative stress and silence growth-related genes (Eisenstark et al. 1996; Azam et al. 1999; Ishihama 1999). 70S ribosomes are converted into the dimerized 100S storage form by RMF (Ribosome Modulation Factor), which binds to 50S ribosomal subunits (Ishihama 1999). In addition, RNA polymerase containing σS is responsible for the gene expression involved in general stress resistance (Hengge-Aronis 1996, 2000, 2002).

While the promoter regulation activated during the onset of stationary phase has drawn a great deal of attention, the activities of the usual consensus promoters devoid of known specific regulation were not clearly described during and after growth-phase transition. To approach this issue, we designed a method for monitoring consensus promoter activity. Since the product of the most versatile reporter gene in E. coli, lacZ, is known to be extremely stable in vivo (Strange 1966), we concluded that lacZ would not be suitable for our purpose. During the growth phase transition, the cessation of growth should be accompanied by arrest of net protein synthesis, and the specific activity of β-galactosidase would not reflect the rate of de novo protein synthesis. We therefore chose the bacterial luciferase encoded by the genes luxCDABE of Vibrio fischeri as our reporter. Using this system, we found that luciferase activity began to decline during the transition into stationary phase, although the level of luciferase proteins remained constant. We suggest that this decline in luciferase activity was the result of a decrease in the supply of reducing equivalents for the bacterial luciferase, which in turn was shown to be under the control of nucleoid proteins such as H-NS and HU. The mechanism underlying ADLA (Abrupt Decline of Luciferase Activity) and its significance for the physiology of the stationary phase are discussed.

Materials and methods

Bacterial strains and plasmids

The strains used in this study are listed in Table 1. Plasmid pLXUV5 was constructed by introducing a 207-bp EcoRI lacUV5(L8) promoter fragment (Kajitani and Ishihama 1983) into the EcoRI site of a pUCD615 plasmid containing the V. fischeriluxCDABE genes. A multiple cloning site is present ∼600 bp upstream of the start codon of luxC (Rogowsky 1987). Using the primer pair 5′-TCGGGTACCCGATTCATTAATGCAGCTGG-3′ and 5′-AGCGGATCCGCTCACAATTCCACAC-3′, a lacUV5(L8) promoter fragment was amplified from pLXUV5 and inserted into the KpnI/BamHI site of pSP- luc+NF (Promega) using the attached restriction sites (underlined), to generate pSPUV5. pSTV28-GROSL was constructed by introducing an 8.1 kb EcoRI fragment of pOF12 (Fayet 1986) containing groES and groEL genes into the EcoRI site of pSTV28 (TaKaRa). For the construction of the Fre [NAD(P)H flavin oxidoreductase; see below], overexpression plasmid for the in vitro luciferase assay (pFreEx), the fre ORF was amplified (with the primer pair 5′-TGTCATATGACAACCTTAAGCTGTAAAG-3′ and 5′-GAGGATCCTCAGATAAATGCAAACGCATC-3′) from E. coli genomic DNA, and cloned into the NdeI/BamHI sites of pET20b(+) (Novagen) using the attached restriction sites. The luxB ORF was amplified with the primers 5′-ATACATATGTATGAAATTTGGATTATTTTTTC-3′ and 5′-TAACTCGAGTTATGGTAAATTCATTTCGATTTTTTG-3′, from pLXUV5, and cloned into the XhoI/BamHI sites of pET15b (Novagen) using the attached restriction sites, to yield pET15b- luxB. PCR was performed with ExTaq DNA polymerase (TaKaRa), and the nucleotide sequences of the amplified products were confirmed by sequencing.
Table 1

Strains used in this study

Strain

Relevant genotype

Reference

KT1008

[MC4100 (F-: araD139D(argF-lacU169 rpsL150 (Strr) relA1flbB5301deoC1ptsF2thiAtonA21) but relA+]

Tanaka et al. (1997)

KT1008H

KT1008 hns::Km

Yamada et al. (1991), this study

KK1001

KT1008 himA::Tc (CK8 × YK2674)

This study

KK1002

KT1008 hupA::Km (CK8 × YK1130)

This study

KK1003

KT1008 hupB::Cm (CK8 × YK1220)

This study

KK1004

KT1008 lacZΔ145 fis-767 (Kmr) (CK8 × MDW246)

This study

YK2674

himA::Tc trpC

Kano and Imamoto (1990)

YK1130

hupA::Km trpC

Wada et al. (1988)

YK1220

hupB::Cm trpC

Wada et al. (1988)

MDW246

MG1655 lacZΔ145 fis-767 (Kmr)

Weinreich and Reznikoff (1992)

Media

Escherichia coli strains were cultivated in LB (1% bactotryptone, 0.5% yeast extract, 1% NaCl), M9 medium (Miller 1972) supplemented with glucose (final concentration 0.2%) and casamino acids (0.2%), or MOPS minimal medium (Neidhardt 1974) supplemented with glucose (0.4%), K2HPO4 (5 mM) and casamino acids (0.05%) in shaken flasks at 30°C.

Measurement of luciferase activity

Luciferase activity is expressed in relative luminescence units (RLUs), based on 10-s measurements in the raw data measurement mode with a LUMAT LB9507 (Berthold, Bad Wildbad, Germany) luminometer. When cell cultures were used as samples (defined as in vivo activity), a 0.1-ml aliquot of the culture was placed in a test tube, and the RLUs for the bacterial luciferase were measured immediately. For the firefly luciferase, 10 μl of a 20-μg/ml solution of beetle luciferin (Promega, Tokyo, Japan) was automatically injected into the sample by the luminometer prior to measurement. When cell lysates were used as samples (defined as in vitro activity), procedures were based on the so-called coupled assay (Meighen et al. 1993). Aliquots (10 μl) of 0.2 mg/ml FMN (Wako) in 2 mM potassium phosphate buffer (pH 7), 10 μl of 4 mg/ml NADPH (Wako) in the same buffer, and 2 μl of cell lysate containing overexpressed Fre were added to the sample (5–25 μl), and the volume was adjusted to 200 μl with 2 mM potassium phosphate buffer (pH 7). The RLUs were measured by the luminometer after automatic injection of 100 μl of 0.001% n-caprinaldehyde (Sigma) in water. Specific luciferase activities in vivo and in vitro were calculated as RLU/OD600 and RLU/total protein (mg), respectively. Samples for the in vitro luciferase assay were prepared as follows. Cultures equivalent to 0.5 OD600 units were sampled, and cells were collected by centrifugation. After resuspension in 0.5 ml of 2 mM potassium phosphate buffer (pH 7), cells were disrupted with sonication, and the supernatant was recovered after centrifugation (17,000×g for 10 min). Cell lysates were prepared from the Fre-overexpressing strain follows. BL21(DE3)(pLysS) (Novagen) carrying pFreEx was grown in 10 ml of LB with ampicillin (50 μg/ml) and chloramphenicol (40 μg/ml) at 27°C. When the OD600 value reached about 0.6, IPTG (isopropyl thio-D-galactopyranoside) was added (final concentration 1 mM), and incubation was continued for 2 h under the same conditions. Cells were collected by centrifugation, resuspended in 0.5 ml of PBS (137 mM NaCl, 8.1 mM Na2HPO4·12 H2O, 2.68 mM KCl, 1.47 mM KH2PO4), and disrupted by sonication. The supernatant was used as the Fre-containing lysate after centrifugation at 17,000×g for 30 min.

Purification of LuxB and preparation of antiserum

BL21(DE3)(pLysS) (Novagen) carrying pET15b- luxB was grown in 200 ml of LB with ampicillin (50 μg/ml) and chloramphenicol (40 μg/ml) at 30°C, and 6×His-tagged LuxB was induced by the addition of 1 mM IPTG when the OD600 value reached about 0.6. After cultivation for 2 h at 25°C, cells were collected by centrifugation, resuspended in 10 ml of PBS, disrupted by sonication, and the overproduced protein was recovered in the pellet as inclusion bodies after centrifugation. The pellets were suspended in lysis buffer [100 mM Na2HPO4, 10 mM TRIS-HCl (pH 8) and 6 M urea] and loaded onto a column containing Ni-NTA Agarose (Qiagen) for affinity chromatography. After washing the column with wash buffer [100 mM Na2HPO4, 10 mM TRIS-HCl (pH 6.3) and 6 M urea], the 6×His-tagged LuxB was eluted with elution buffer [100 mM Na2HPO4, 10 mM TRIS-HCl (pH 4.5) and 6 M urea] and dialyzed against PBS. About 2 mg of protein was purified from 200 ml of culture. Rabbit antiserum against the purified His-tagged LuxB was produced by Sawady Technology Co. Ltd.

Western analysis

Western blotting analysis was performed as described previously (Tanaka 1993) using rabbit anti-LuxA antiserum (a gift from Carl H. Johnson) or rabbit anti-LuxB antiserum (this study).

Results

Monitoring of bacterial luciferase activity during growth phase transition

To analyze the activity of consensus promoters devoid of specific regulation, we constructed the plasmid pLXUV5 which carries a transcriptional fusion gene (P lacUV5- lux) composed of the lacUV5(L8) promoter (Dickson et al. 1977), an E. coli consensus-type promoter, and the V. fischeri luciferase genes (luxCDABE) as reporters. Using this system, it is possible to measure the luciferase activity of E. coli without cell disruption or the addition of a substrate (long chain aldehyde; RCHO), because the genes luxC, D and E encode enzymes for the biosynthesis of adequate levels of substrate (Rogowsky 1987).

By monitoring the luciferase activity of E. coli cells harboring pLXUV5 we found that the activity declined abruptly during the transition from the vegetative to the stationary phase (Fig. 1). The decline was observed not only in minimal medium (M9) but also in rich medium (LB; Fig. 1), and in some cases growth was also arrested by carbon, nitrogen, phosphate or sulfur starvation (data not shown). Furthermore, similar observations were also made on several well-known laboratory strains, such as MC4100, CSH26 and W3110 (data not shown). Therefore, this is not a specific response that is restricted to any specific strain or due to depletion of any specific nutrient. It is assumed to be the general response that is associated with the transition from the vegetative growth phase to the stationary phase. Thus, we designated this phenomenon as ADLA (Abrupt Decrease of Luciferase Activity) and defined the time point when luciferase activity begins to drop as T0.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig1_HTML.gif
Fig. 1

Time-course of the specific activity of luciferase expressed by the P lacUV5 -lux fusion gene. KT1008 harboring pLXUV5 was grown in LB (A) or M9 (Miller 1972) (B), and luciferase specific activity (closed triangles) and the cell density (open circles) were monitored at the indicated times

Next, we examined whether or not ADLA resulted from changes in lux gene expression. The amounts of LuxA and LuxB proteins were monitored by Western analysis before and after T0. These tests revealed that the level of neither protein decreased after T0 (Fig. 2). Thus, ADLA is not caused by a rapid decrease in the amounts of luciferase proteins available. We therefore concluded that luciferase activity is somehow repressed at the enzyme level. The half-life of the luciferase proteins increased from 10 min at T−0.25 to more than 30 min at T0.5 (Fig. 3), and stabilization after T0 compensates for the cessation of de novo synthesis, resulting in no net change in luciferase protein levels.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig2_HTML.gif
Fig. 2

Levels of LuxA and LuxB proteins during the transition to stationary phase. A KT1008 (pLXUV5) was grown in MOPS minimal medium (Neidhardt 1974) supplemented with 0.4% glucose, 5 mM K2HPO4, and 0.05% casamino acids. The times at which Lux protein levels were measured are indicated as 1–4. Luciferase specific activity (closed triangles) and cell density (open circles) were measured at the indicated times. B Monitoring of LuxA and LuxB protein levels by Western analysis. Lanes correspond to the time points of sampling indicated in A

https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig3_HTML.gif
Fig. 3

Stability of the LuxA and LuxB proteins before and after T0. KT1008 (pLXUV5) was grown in MOPS minimal medium supplemented as in the legend to Fig. 2, and the levels of LuxA and LuxB proteins were monitored by Western analysis after the addition of chloramphenicol (100 μg/ml) at OD600 0.8 (i.e. ∼15 min before T0) (A), or 30 min after T0 (B). Protein levels were quantified from images scanned using NIH image software (http://rsb.info.nih.gov/nih-image/), and are shown as percentages relative to that at the time when chloramphenicol was added. Open circles LuxA; open triangles LuxB

ADLA is caused by a decrease in the availability of flavin mononucleotide

Bioluminescence produced by V. fischeri luciferase (Engebrecht et al. 1983; Meighen and Dunlap 1993; Zaslaver et al. 2004) arises from the oxidation of RCHO and a reduced flavin mononucleotide (FMNH2) by molecular oxygen, yielding the corresponding fatty acid (RCOOH), an oxidized flavin mononucleotide (FMN), water and photons as follows:
$${\text{RCHO}} + {\text{FMNH}}_2 + {\text{O}}_2 \to {\text{RCOOH}} + {\text{FMN}} + {\text{H}}_2 {\text{O}} + {\text{h}}\nu $$
Thus, based on the above equation, we tested four hypotheses for the occurrence of ADLA: (1) limitation of RCHO, (2) inactivation of luciferase, (3) depletion of O2, and (4) limitation of redox supply through FMNH2.
  1. 1.

    Limitation of the substrate. We analyzed luciferase activity in the presence or absence of the cell-permeable substrate n-caprinaldehyde, in order to examine whether ADLA is caused by RCHO limitation. ADLA was observed irrespective of the addition of the aldehyde, although the activity reached after the rapid decline was a little higher in the presence of the exogenous substrate (Fig. 4). This indicated that RCHO was not the factor limiting the luciferase activity in ADLA.

     
  2. 2.

    Inactivation of luciferase. To examine this possibility, we prepared cell lysates from cultures 15 min before T0 (T−0.25) and 15 min after T0 (T0.25), and compared the luciferase-specific activity in the presence of excess substrate. Luciferase-specific activity remained unchanged before and after T0, while the light emission from the cells at T0.25 decreased to 1% of that at T−0.25 (Fig. 5). Thus, luciferase itself was shown to be sufficiently active, and ADLA cannot be due to luciferase inactivation. We also examined the effects of overexpression of GroE and low temperature on ADLA, as the luciferase activity in vivo may depend on stabilization of the protein by GroE (Escher et al. 1991; Escher and Szalay 1993). In case of the GroE overproduction, ADLA was still observed, although the light emission was enhanced before and after T0, hence indicating that the protein stability was not a critical determinant of ADLA (Fig. 6a). However, we found that ADLA was significantly alleviated at 18°C (Fig. 6b), and the implication of this observation is discussed below.

     
  3. 3.
    Depletion of O2. In contrast to bacterial luciferase, luciferase from fireflies (Luc) requires O2, luciferin and ATP for light emission, but does not require FMNH2 (Lundin 2000):
    $${\text{D - luciferin}} + {\text{ATP}} + {\text{O}}_2 \to {\text{oxyluciferin}} + {\text{AMP}} + {\text{PPi}} + {\text{CO}}_2 + {\text{h}}\nu $$
     
Based on a luc reporter plasmid (pSP- luc+NF), pSPUV5 which carries a transcriptional fusion (P lacUV5 -luc) comprising the lacUV5 promoter and luc was constructed. Time-course analysis of light emission was performed as for pLXUV5, but ADLA was not observed at all when this plasmid was used (Fig. 7). Thus because firefly luciferase, which also requires O2 for the reaction, could emit light continually after T0, ADLA is not caused by the depletion of intracellular O2.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig4_HTML.gif
Fig. 4

Effects of substrate addition (n-caprinaldehyde) to the culture medium. KT1008 (pLXUV5) was grown in MOPS minimal medium supplemented as described in the legend to Fig. 2, and luciferase specific activity (circles) and OD600 (triangles) were measured in cultures with (open symbols) or without (closed symbols) added n-caprinaldehyde

https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig5_HTML.gif
Fig. 5

Comparison of luciferase specific activities in vivo (A) and in vitro (B). Cells were sampled at OD600 0.8 (about 15 min before T0) and 15 min after T0 as in Fig. 3, and the specific activities were determined as described in the Materials and methods. Values with standard deviations are the means of three independent experiments

https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig6_HTML.gif
Fig. 6

Effects of GroE overexpression (A) and low temperature (B) on the time course of luciferase specific activity. E. coli strains were grown in MOPS minimal medium supplemented as described in the legend to Fig. 2, and luciferase specific activity (solid lines) and cell density (dotted lines) were monitored periodically. The strains used were: A KT1008 (pLXUV5) (triangles) and KT1008 (pLXUV5 and pSTV28-GROEL) (circles); B KT1008 (pLXUV5) grown at 18°C (squares), 25°C (triangles) or 30°C (circles)

https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig7_HTML.gif
Fig. 7

Time course analysis of firefly luciferase activity. KT1008 (pSPUV5) was grown in MOPS minimal medium supplemented as described in the legend to Fig. 2, and luciferase activity (filled triangles) and cell density (open circles) were monitored periodically

Based on the above experimental results, the fourth possibility, the depletion of redox supply through FMNH2, was strongly suggested to be the reason for ADLA. This suggestion does not contradict the results of the firefly luciferase experiment, because FMNH2 is not required by the firefly enzyme.

ADLA is alleviated in mutants deficient in nucleoid proteins such as H-NS and HU

Considering that ADLA is related to the physiology of stationary phase, we examined the effects on ADLA of mutations in genes such as rpoS (Lange and Hengge-Aronis 1991) and rsd (Jishage and Ishihama 1999), and genes involved in the stringent response (Xiao et al. 1991), which are known to affect stationary-phase gene expression. Among those tested, only the hns mutation (Yoshida et al. 1993) was found to strongly alleviate ADLA: light emission remained at a high level at least during the early stationary phase (Fig. 8a and data not shown). Since the hns gene product is one of the bacterial nucleoid proteins, H-NS, several mutants deficient in other nucleoid proteins were also examined for ADLA. It emerged that a mutation in hupA (encoding a α subunit of HU) alleviated ADLA as strongly as loss of the hns function (Fig. 8a), whereas mutations in hupB, himA and fis (which encode a β subunit of HU, and α subunit of IHF and Fis, respectively) had only a moderate effect on ADLA (Fig. 8b). A mutation in dps (encoding Dps; Almiron et al. 1992) had no effect on ADLA (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0008-5/MediaObjects/438_2005_8_Fig8_HTML.gif
Fig. 8

Time course analyses of the luciferase specific activity in various nucleoid mutants. Nucleoid mutant strains were transformed with pLXUV5, and grown in MOPS minimal medium supplemented as described in the legend to Fig. 2. Luciferase specific activity (solid lines) and cell density (dotted lines) were monitored periodically for: A KT1008 (wild-type, bars), KT1008H (hns::Km, circles), and KK1002 (hupA::Cm, triangles); B KT1008 (wild-type, bars), KK1001 (himA::Tc, circles), KK1003 (hupB::Cm, squares) and KK1004 (fis::Km, triangles)

Discussion

In this study, we found that light emission from E. coli cells expressing bacterial luciferase abruptly declines during the transition to stationary phase. Of the requirements for bacterial luciferase activity, long-chain aldehydes (Fig. 4) and O2 (Fig. 7) were shown not to be the limiting factors of light emission, and the luciferase itself retained enzyme activity (Fig. 5). Thus, depletion of reduced flavin mononucleotide was strongly suggested to be responsible for ADLA.

It is well known that FMNH2 is the direct electron donor for the bacterial luciferase reaction (Engebrecht et al. 1983; Meighen and Dunlap 1993; Zaslaver et al. 2004). Because FMN is reduced by oxidation of the pyridine nucleotides NADH and NADPH by NAD(P)H flavin oxidoreductase (Spyrou et al. 1991; Fontecave et al. 1994), depletion of the FMNH2 pool might be caused by depletion of the pool of NAD(P)H or by the repression of NAD(P)H flavin oxidoreductase activity. In previous studies, it was reported that the fre gene encodes an NAD(P)H flavin oxidoreductase in E. coli. However, since the fre mutation did not affect either luciferase activity during the vegetative phase or ADLA (our unpublished observations), this gene product cannot be the major enzyme responsible for the reduction of FMN in vivo. It has also been reported that luciferase itself can function as an NAD(P)H flavin oxidoreductase (Imlay 1991). Given that there are multiple activities that reduce FMN, it is unlikely that all these activities are inhibited at once at T0. Therefore, we speculate that ADLA is caused by depletion of the pool of reducing power stored in NAD(P)H. Assuming that the reducing power stored in NAD(P)H rapidly declines after T0, various biosynthetic pathways will be inhibited, and there should be no further increase in cell mass should occur. We observed that the rate of increase of optical density (OD600) of the culture was slowed or arrested around T0, which is consistent with the hypothesized decline in the NAD(P)H pool. Although there are many possible explanations for the decline of the NAD(P)H pool, we currently have no further indications as to which might be the most important.

In minimal media where growth cessation is triggered by carbon starvation, depletion of the levels of reducing equivalents generated by fermentation or respiration is unavoidable. However, in rich media not all essential nutrients required for growth are depleted before the cessation of growth; nevertheless, ADLA is observed during entry into the stationary phase. Thus, we suggest that ADLA does not result from passive depletion of reducing power as a result of nutrient shortage, but rather is a programmed process for differentiation into stationary-phase cells. Moreover, we found that ADLA was not observed, or was alleviated, in several mutants deficient in nucleoid proteins, which implies underlying genetic control and is consistent with this idea.

ADLA was not observed or was alleviated in several mutants deficient in individual nucleoid proteins, and although the underlying mechanism is unknown at present, it is likely that some gene expression controlled by these nucleoid proteins is involved in ADLA. Of the seven nucleoid proteins examined, dps is expressed only in stationary-phase cells (Azam et al. 1999), and loss of this gene had no effect on ADLA. Furthermore, of the two genes whose gene products comprise HU, hupB—which is predominantly expressed in the stationary phase (Claret and Rouviere-Yaniv 1997)—had less effect on ADLA than hupA. Thus, gene expression during the vegetative phase appears to be important for ADLA. Presumably, appropriate gene expression prior to T0 is essential for the preparation of the machinery involved in ADLA. Alternatively, it is possible that specific gene expression at T0 triggers ADLA by some concerted action of nucleoid proteins, in particular HupA and H-NS. It was reported that some genes for some catabolic enzymes involved in NAD(P)H production are under the control of H-NS (Laurent-Winter et al. 1997). It is known that H-NS affects the gene expression profiles induced by various environmental changes (Williams and Rimsky 1997; Hommais et al. 2001; Schröder and Wagner 2002), and thus might function as a triggering factor for ADLA. If this is the case, the findings made in this study might reveal a new type of cellular regulation involved in the global control of genes for metabolic and nucleoid proteins.

At low temperature (as low as 18°C), we unexpectedly found a significant alleviation of ADLA (Fig. 6b). This observation probably reflects the effect of low temperature on the intracellular redox pool, since the specific activity of luciferase, and hence its stability, was not the direct cause of ADLA (Fig. 5). Here, it should be noted that ADLA was also alleviated similarly in an hns mutant (Fig. 8). Previously, it was reported that H-NS was required for the acclimation to low temperatures (Dersch et al. 1994), and can act as a temperature sensor in vitro (Amit et al. 2003). Thus, the effects of H-NS loss and low temperatures on ADLA might have a common mechanistic basis.

It was suggested that reactive oxygen species originating from the respiratory electron transfer chain are deleterious for stationary-phase cells (Woodmansee and Imlay 2002). In particular, flavin nucleotides are suspected to transfer electrons to molecular oxygen and free ferric ions to generate reactive oxygen species (Woodmansee and Imlay 2002). Given that intracellularly generated reactive oxygen species are deleterious for stationary-phase cells, a genetically controlled decline of the NAD(P)H redox pool in stationary-phase cells is very likely, and must be favorable for long-term survival.

The reporter system involving the bacterial luciferase genes was widely used to monitor gene expression in vivo. However, as shown in this study, the luciferase activity could be strongly affected by not only gene expression but also by the intracellular redox pool and the O2 concentration. This point should be always taken into account in evaluating the results obtained using these systems.

Acknowledgements

We thank Drs. Clarence I. Kado (U.C. Davis) and Takashi Yura for providing plasmids, Drs. Carl H. Johnson (Vanderbilt University) and Hideo Iwasaki (Nagoya University) for the antiserum against Vibrio LuxA, and Drs. Yasunobu Kano (Kyoto Pharmaceutical University), William S. Reznikoff (University of Wisconsin-Madison), Chiharu Ueguchi (Nagoya University) and Yasuhiko Sekine [Rikkyo (St. Paul) University] for E. coli strains. We also thank Dr. Hideo Takahashi for discussions. This study was supported in part by CREST of JST (to K.T.). The authors declare that this work was carried out in compliance with the current laws governing genetic experimentation in Japan.

Copyright information

© Springer-Verlag 2005