Excretion and uptake of cadaverine by CadB and its physiological functions in Escherichia coli

Authors


E-mail iga16077@p.chiba-u.ac.jp; Tel. (+81) 43 290 2897; Fax (+81) 43 290 2900.

Summary

The functions of the putative cadaverine transport protein CadB were studied in Escherichia coli. CadB had both cadaverine uptake activity, dependent on proton motive force, and cadaverine excretion activity, acting as a cadaverine-lysine antiporter. The Km values for uptake and excretion of cadaverine were 20.8 and 303 µM respectively. Both cadaverine uptake and cadaverine-lysine antiporter activities of CadB were functional in cells. Cell growth of a polyamine-requiring mutant was stimulated slightly at neutral pH by the cadaverine uptake activity and greatly at acidic pH by the cadaverine-lysine antiporter activity. At acidic pH, the operon containing cadB and cadA, encoding lysine decarboxylase, was induced in the presence of lysine. This caused neutralization of the extracellular medium and made possible the production of CO2 and cadaverine and aminopropylcadaverine instead of putrescine and spermidine. The induction of the cadBA operon also generated a proton motive force. When the cadBA operon was not induced, the expression of the speF–potE operon, encoding inducible ornithine decarboxylase and a putrescine-ornithine antiporter, was increased. The results indicate that the cadBA operon plays important roles in cellular regulation at acidic pH.

Introduction

The Escherichia coli cadBA operon and the adjacent cadC gene encode lysine decarboxylase (cadA), a putative cadaverine-lysine antiporter (cadB), and CadC, a positive regulator of cadBA expression (Meng and Bennett, 1992; Watson et al., 1992). Expression of the cadBA operon is induced by acidic pH and lysine (Neely and Olson, 1996). It is thought that production and excretion of cadaverine and the consumption of a proton during the decarboxylation reaction contribute to an increase in pH of the extracellular medium and to cell growth. We also proposed that the cadBA operon functions as a supplier of carbon dioxide (Takayama et al., 1994).

Several secondary transporters that catalyse antiport of a weak acid and its product such as malate and lactate, coupled with decarboxylation, have been reported as secondary proton motive force generation systems (Harold and Maloney, 1996; Lolkema et al., 1998). The basis for energy coupling was first characterized in Oxalobacter by Anantharam et al. (1989). Furthermore, one antiport system for a basic amino acid and its amine product (histidine/histamine), coupled with decarboxylation, has been reported to generate a proton motive force in Lactobacillus buchneri (Molenaar et al., 1993). Thus, it is conceivable that the cadaverine-lysine antiport system, coupled with decarboxylation, also functions as a proton motive force generation system in E. coli.

To understand the role of the products of the cadBA operon, it is important to elucidate the function of the cadaverine transport protein CadB, which has not yet been studied in detail. We recently delineated the structure and function of PotE, which functions as a putrescine-ornithine antiporter and a putrescine transporter and is encoded by the pPT71 (speF–potE) operon consisting of speF, a gene for inducible ornithine decarboxylase, and potE (Kashiwagi et al., 1991; 1992; 1997; 2000). As the amino acid sequences of PotE and CadB show high similarity, CadB was hypothesized to be a cadaverine-lysine antiporter. In this study, we have investigated the properties of cadaverine transport by CadB and its physiological functions in cells, particularly at acidic pH.

Results

Uptake and excretion of cadaverine by CadB

Chromosomal location and schematics of the cadC gene, cadBA operon and speF–potE operon are shown in Fig. 1A. A positive or negative regulator gene for the speF–potE operon is not located close to the operon, and is so far unidentified. Homology between the amino acid sequences of CadB and PotE was compared, and high sequence similarity was observed (30.7% overall identity) especially in the NH2-terminal region (Fig. 1B), suggesting that CadB has structure and function similar to PotE and that it may be involved in both uptake and excretion of cadaverine.

Figure 1.

Schematic representation of the cadC gene, cadBA operon and speF–potE operon (A) and alignment of the amino acid sequence of CadB and PotE (B).
A. Chromosomal location and encoding proteins of the cadC gene, cadBA operon and speF–potE operon are shown.
B. Sequence similarities of CadB and PotE were analysed according to the method of Needleman and Wunsch (1970) using a dnasis program (Hitachi Software Engineering). The identical amino acid residues are shown by black boxes with white lettering, and the similar residues by grey boxes. Transmembrane segments are labelled I to XII.

The cadB gene was cloned into the vector pMW119 under the control of the lacUV promoter. pMW119 and pMWcadB were introduced into MA261cadC::Km, which is deficient in positive regulator CadC, and into JM109. Uptake and excretion of cadaverine by CadB were measured using intact cells of MA261cadC::Km/pMW119 or pMWcadB, and lysine-loaded inside-out membrane vesicles prepared from JM109/pMW119 or pMWcadB respectively. The results are shown in Fig. 2. Uptake of cadaverine by CadB was observed at neutral pH, and uptake was immediately inhibited by CCCP (carbonyl cyanide m-chlorophenylhydrazone), a proton uncoupler, suggesting that uptake is dependent on the proton motive force (Fig. 2A). Putrescine and lysine inhibited cadaverine uptake by CadB about 20% at 25 times higher concentration (250 µM) of substrate (10 µM). Ornithine and agmatine inhibited cadaverine uptake to a lesser extent than did putrescine and lysine, and arginine had no effect on uptake (data not shown). The results suggest that CadB has a high specificity for cadaverine as the uptake substrate. Cadaverine uptake activity was low at acidic pH (data not shown). The Km value for cadaverine was calculated as 20.8 µM from Lineweaver–Burk plots. Excretion was measured by cadaverine uptake into lysine-loaded inside-out membrane vesicles. Inside-out membrane vesicles prepared from JM109/pMWcadB showed high activity compared with the vesicles from JM109/pMW119, and the activity was not inhibited by CCCP, indicating that cadaverine is excreted by CadB through cadaverine-lysine antiporter activity (Fig. 2B). When cadaverine uptake was measured using unloaded inside-out membrane vesicles of JM109/pMWcadB, the activity was less than 5% compared with lysine-loaded inside-out membrane vesicles. Cadaverine uptake activity with inside-out membrane vesicles prepared at pH 7.5 rather than at pH 6.6 was low (data not shown). The Km value for cadaverine was calculated as 303 µM from Lineweaver–Burk plots.

Figure 2.

Cadaverine uptake (A) and excretion (B) by CadB.
A. Cadaverine uptake activity was measured using intact cells of MA261cadC::Km containing pMW119 or pMWcadB with 10 µM [14C]-cadaverine as a substrate as described in Experimental procedures. Where indicated, 40 µM CCCP, 250 µM putrescine or 250 µM lysine was added to the reaction mixture. (▪), MA261cadC::Km/pMWcadB; (□), MA261cadC::Km/pMWcadB+ CCCP; (▵), MA261cadC::Km/pMWcadB + putrescine; (◊), MA261cadC::Km/pMWcadB + lysine; (•), MA261cadC::Km/pMW119. Km and Vmax values of cadaverine uptake by intact cells were measured by changing the concentration of substrates from 5 to 30 µM.
B. Cadaverine excretion activity was measured by cadaverine uptake by lysine-loaded inside-out membrane vesicles of JM109 containing pMW119 or pMWcadB with 100 µM [14C]-cadaverine as a substrate as described in Experimental procedures. Where indicated, 40 µM CCCP was added to the reaction mixture. (▪), JM109/pMWcadB; (□), JM109/pMWcadB + CCCP; (•), JM109/pMW119; (○), JM109/pMW119 + CCCP. Km and Vmax values of cadaverine uptake by lysine-loaded inside-out membrane vesicles were measured by changing the concentration of substrates from 25 µM to 400 µM.
Values are mean ± SE of triplicate determinations.

Function and properties of CadB at neutral pH

CadB catalyses the uptake of cadaverine at neutral pH. Polyamines [putrescine (1,4-diaminobutane), spermidine and spermine] are necessary for normal cell growth (Tabor and Tabor, 1984; Cohen, 1998), and putrescine and spermidine are usually present in E. coli. We therefore examined whether cadaverine (1,5-diaminopentane), taken up by CadB, contributes to cell growth in a polyamine-deficient mutant MA261 similar to the effects of exogenous polyamines in this mutant. The E. coli polyamine-deficient mutant MA261 is deficient in the speB gene encoding agmatine ureohydrolase and the speC gene encoding ornithine decarboxylase (Cunningham-Rundles and Maas, 1975). Thus, MA261 cells cannot synthesize putrescine, although putrescine transported into cells can be converted to spermidine. When MA261 cells were cultured in the presence of cadaverine, cell growth was stimulated compared with cells cultured in the absence of cadaverine, but the effect of cadaverine on cell growth was smaller than that of putrescine (Table 1). The stimulatory effect of cadaverine on the growth of MA261cadC::Km cells, deficient in the cadC gene, a positive regulator of the cadBA operon, became smaller (Table 1). The difference in the stimulation effect of cell growth between putrescine and cadaverine became greater in MA261cadC::Km than in MA261. Cadaverine uptake activity of MA261cadC::Km cells was also low compared with MA261 cells (data not shown). Then, the polyamine content was measured under these conditions (Table 2). Cadaverine and aminopropylcadaverine accumulated in the polyamine-requiring mutant MA261, but there was less accumulation in the CadC-deficient mutant (Table 2). Aminopropylcadaverine can be synthesized from cadaverine by spermidine synthase, like spermidine from putrescine, and it also stimulates cell growth like spermidine (Igarashi et al., 1986). Levels of putrescine and spermidine were decreased slightly after exposure to cadaverine. The results indicate that cadaverine transported by CadB stimulates cell growth of a polyamine-deficient mutant of E. coli by itself or after conversion to aminopropylcadaverine, although the expression of CadB mRNA was low at neutral pH (see Fig. 7). When strain MA261cadB::Km was used instead of MA261cadC::Km, essentially the same results were obtained (data not shown).

Table 1.  Cell growth of E. coli MA261 and MA261cadC::Km at neutral pH.
CellsAdditionGeneration
time (h)
Growth
stimulation (-fold)
  1. Cells were grown in medium A in the absence and presence of cadaverine or putrescine. Cell growth was followed by measuring A540. Values are mean ± SE of four determinations.

MA261
 None5.25 ± 0.591.0
Putrescine (10 µg ml−1)1.68 ± 0.093.1
Putrescine (30 µg ml−1)1.68 ± 0.073.1
Cadaverine (10 µg ml−1)3.11 ± 0.231.7
Cadaverine (30 µg ml−1)3.03 ± 0.151.7
MA261cadC::Km
 None4.92 ± 0.711.0
Putrescine (10 µg ml−1)1.88 ± 0.032.6
Putrescine (30 µg ml−1)1.71 ± 0.022.9
Cadaverine (10 µg ml−1)3.98 ± 0.331.2
Cadaverine (30 µg ml−1)3.90 ± 0.331.3
Table 2.  Polyamine content in E. coli MA261 and MA261cadC::Km cultured at neutral pH.
CellsAdditionPolyamine (nmol mg−1 protein)
PutrescineCadaverineSpermidineAminopropylcadaverine
  1. Cells were cultured for 11 h in the presence and absence of putrescine or cadaverine, and polyamine content was measured as described in Experimental procedures. Values are mean ± SE of triplicate determinations.

  2. ND, not detected.

MA261
 None0.08 ± 0.020.63 ± 0.095.22 ± 0.574.02 ± 0.46
Putrescine (10 µg ml−1)19.8 ± 2.340.37 ± 0.0515.5 ± 1.83ND
Cadaverine (10 µg ml−1)ND8.55 ± 0.973.58 ± 0.5316.4 ± 1.77
MA261cadC::Km
 None0.13 ± 0.030.03 ± 0.016.45 ± 0.744.61 ± 0.52
Putrescine (10 µg ml−1)24.9 ± 3.35ND14.8 ± 1.64ND
Cadaverine (10 µg ml−1)ND1.03 ± 0.184.78 ± 0.628.43 ± 0.95
Figure 7.

Expression of CadB and PotE mRNAs.
A. Expression of CadB mRNA in MA261 cultured at acidic and neutral pH in the presence and absence of lysine or cadaverine. MA261 cells were cultured at pH 5.2 or pH 7.0 as described in the legend to Fig. 5, and dot-blot analysis of CadB mRNA was performed as described in Experimental procedures.
B. Comparison of expression of CadB and PotE mRNAs in E. coli B and BcadC::Km cultured at acidic pH in the presence and absence of ornithine or lysine. Culture of E. coli B and BcadC::Km, and dot-blot analysis of CadB and PotE mRNAs were performed as described in Experimental procedures. CAD, cadaverine.

The characteristics of cadaverine uptake by CadB were examined using right side-out membrane vesicles of JM109/pUCcadB(Fig. 3). Right side-out membrane vesicles of JM109/pUCcadB showed high cadaverine uptake activity compared with the vesicles of JM109/pUC119 (vector) when the vesicles were energized by potassium ascorbate–phenazine methosulphate (Tolner et al., 1995). Uptake of cadaverine by CadB was greatly inhibited by CCCP, which dissipates the proton motive force, valinomycin, which dissipates the membrane potential, and nigericin, which dissipates the transmembrane pH gradient (Fig. 3A), suggesting that cadaverine uptake was dependent on a proton motive force composed of both ΔpH and ΔΨ. This was confirmed by measurement of cadaverine uptake by generating artificial ion gradients. When a proton motive force composed of both ΔpH and ΔΨ was generated, cadaverine uptake was observed. However, cadaverine uptake was greatly diminished when only ΔpH or ΔΨ was generated (Fig. 3B). The pH inside cells was slightly acidified during cadaverine uptake (data not shown). The result suggests that cadaverine may be transported with protons, although the stoichiometry of proton and cadaverine transport is not known.

Figure 3.

Effect of ionophores (A) and artificially imposed ion gradients (B) on cadaverine uptake by right side-out membrane vesicles of JM109/pUCcadB.
A. The effect of various ionophores on cadaverine uptake was measured in the presence of the electron donor system consisting of potassium–ascorbate–phenazine methosulphate. Right side-out membrane vesicles were prepared from JM109/pUCcadB (•, ◆, ▴, ▪) or from JM109/pUC119 (□). Cadaverine uptake activity was measured in the absence (•, □) and presence of 10 µM nigericin (◆), 8 µM valinomycin (▴) or 40 µM CCCP (▪).
B. Cadaverine uptake by right side-out membrane vesicles prepared from JM109/pUCcadB was measured under artificial ion gradients as described in Experimental procedures. •, Δp; ▴, ΔΨ; ◆, ΔpH; ▪, none. Values are mean ± SE of triplicate determinations.

Function and properties of CadB at acidic pH

The growth of MA261 and MA261cadC::Km cells was compared at pH 5.2 in the presence and absence of lysine or cadaverine (Fig. 4A) in a synthetic medium. When MA261 cells were cultured at an initial acidic pH (5.2) in the presence of 5 mM lysine, stimulation of cell growth and neutralization of the medium were observed compared with cells cultured in the absence of lysine. Under this condition, expression of CadB mRNA increased greatly (see Fig. 7). Cadaverine did not support growth of MA261 cells at pH 5.2. Growth of MA261cadC::Km cells was not affected by lysine or cadaverine, and neutralization of medium did not occur (Fig. 4A). Similar results were obtained with MA261cadB::Km (data not shown). When MA261 cells were cultured in the presence of lysine, the ATP content of the cells also increased (Fig. 4B). The contents of polyamines, cadaverine and aminopropylcadaverine in cells and medium were also measured (Table 3). When MA261 was cultured in the presence of lysine or cadaverine, an increase in cadaverine and aminopropylcadaverine was observed, although cadaverine content was much higher with cells cultured in the presence of lysine than in the presence of cadaverine. In addition, the cadaverine content in the medium increased greatly when MA261 cells were cultured with lysine. Almost all lysine added to the medium was converted to cadaverine. As a control, the cadaverine content in MA261cadC::Km cells and in medium was measured after culturing with lysine. The increase in cadaverine in these cells was small, and no cadaverine was excreted into the medium (Table 3). The results indicate that the cadBA operon was induced under acidic conditions in the presence of lysine, promoting decarboxylation of lysine to cadaverine by lysine decarboxylase encoded by cadA and excretion of cadaverine by the cadaverine–lysine antiport activity of CadB, effects that support cell growth. The results also suggest that cadaverine and aminopropylcadaverine are used instead of putrescine and spermidine as polyamines for cell growth, because putrescine and spermidine contents decreased greatly in cells cultured in the presence of lysine.

Figure 4.

Cell growth and ATP content in cells cultured in acidic conditions.
A. MA261 and MA261cadC::Km cells were cultured in medium A, pH 5.2, containing 62 mM KH2PO4 instead of 40 mM K2HPO4 plus 22 mM KH2PO4 in the presence and absence of lysine or cadaverine. Cell growth was followed by measuring A540. After 11 h, pH of the medium was measured. MA261 (▪, ◆, •) and MA261cadC::Km (□, ◊, ○) were cultured in the absence (▪, □) and presence of 5 mM lysine (◆, ◊) or cadaverine (•, ○). Each point represents the mean of duplicate measurements.
B. MA261 cells were cultured at acidic pH as described above in the absence and presence of 5 mM cadaverine or lysine for 11 h, and ATP content of cells was measured as described in Experimental procedures. CAD, cadaverine.
Values are mean ± SE of triplicate determinations.

Table 3.  Polyamine content in E. coli cells and medium cultured at acidic pH.
StrainAdditionPolyamine
PutrescineCadaverineSpermidineAminopropylcadaverine
  1. Cells were cultured in medium A, pH 5.2, containing 62 mM KH2PO4 instead of 40 mM K2HPO4 plus 22 mM KH2PO4 for 11 h in the presence and absence of lysine or cadaverine, and polyamine content was measured as described in Experimental procedures. Values are mean ± SE of triplicate determinations.

  2. ND, not detected.

Cells (nmol mg−1 protein)
 MA261
 None1.82 ± 0.32 4.19 ± 0.5516.1 ± 1.78ND
Lysine (5 mM)ND 192 ± 20.50.29 ± 0.055.05 ± 0.64
Cadaverine (5 mM)0.11 ± 0.0463.7 ± 7.543.60 ± 0.396.03 ± 0.74
 MA261cadC::Km
 None1.83 ± 0.23 23.2 ± 2.5610.9 ± 1.23ND
Lysine (5 mM)0.83 ± 0.09 44.1 ± 5.2311.2 ± 1.33ND
Cadaverine (5 mM)ND 76.3 ± 8.216.10 ± 0.454.90 ± 0.43
Medium µM
 MA261
 NoneND 30.6 ± 4.5NDND
Lysine (5 mM)ND4660 ± 533NDND
Cadaverine (5 mM)ND4750 ± 567NDND
 MA261cadC::Km
 NoneNDNDNDND
Lysine (5 mM)NDNDNDND
Cadaverine (5 mM)ND4890 ± 440NDND

The characteristics of the cadaverine-lysine antiport activity of CadB were studied (Fig. 5). Inside-out membrane vesicles of JM109/pMWcadB were loaded with [14C]-lysine or unlabelled lysine, and the stoichiometry of exchange of lysine with cadaverine was measured by adding unlabelled cadaverine or [14C]-cadaverine respectively. The amount of lysine retained in the membrane vesicles decreased as cadaverine uptake increased. The exchange ratio of lysine with cadaverine was nearly 1:1 (Fig. 5A). To examine whether CadB has not only cadaverine-lysine antiporter activity but also cadaverine–cadaverine and lysine–lysine exchange activity, lysine-loaded and cadaverine-loaded inside-out membrane vesicles were prepared, and cadaverine or lysine uptake activity was measured. As shown in Fig. 5B, significant cadaverine or lysine uptake activity was observed in vesicles loaded with cadaverine or lysine, indicating that CadB has cadaverine–lysine, cadaverine–cadaverine and lysine–lysine antiporter activities. [14C]-Cadaverine uptake by lysine-loaded inside-out membrane vesicles was inhibited strongly by 10-fold excess of cadaverine and lysine (1 mM) of substrate (100 µM), about 20% by putrescine and only slightly (about 5%) by ornithine, agmatine and arginine (data not shown). This suggests that the antiporter activity of CadB has a high substrate specificity for cadaverine and lysine.

Figure 5.

Exchange of cadaverine (CAD) with lysine (Lys) (A), and cadaverine and lysine uptake (B) by inside-out membrane vesicles.
A. To measure the exchange ratio of lysine with cadaverine, inside-out membrane vesicles of JM109/pMWcadB prepared in the absence of amino acid were preloaded with 1 mM [14C]-lysine (○) or unlabeled lysine (•) for 1 h. Exchange activity was then measured by diluting the membrane vesicles by 1:10 and adding 100 µM cadaverine (○) or [14C]-cadaverine (•) respectively.
B. Inside-out membrane vesicles of JM109/pMWcadB were prepared in the absence (○) and presence of 2.5 mM lysine (•) or cadaverine (▴). [14C]-Cadaverine (left) or [14C]-lysine (right) uptake activity was measured as described in Experimental procedures. CAD, cadaverine.
Values are mean ± SE of triplicate determinations.

We next tested whether electrogenic exchange of lysine with cadaverine takes place (Molenaar et al., 1993; Abe et al., 2002). To examine this possibility, right side-out membrane vesicles of JM109/pUCcadB were loaded with cadaverine together with either N-methylglucamine (NMG) salts or potassium salts. Addition of NMG salts or potassium salts together with valinomycin generates a membrane potential, the polarity of which was either interior positive (potassium outside, NMG inside) or interior negative (NMG outside, potassium inside). As shown in Fig. 6, when a membrane potential was not gen-erated (○, ▵), significant cadaverine–lysine exchange activity was observed. When an interior positive membrane potential was generated (•), cadaverine–lysine exchange activity was greatly stimulated, whereas it was inhibited when an interior negative membrane potential was generated (▪). Accordingly, the process of cadaverine excretion with lysine uptake catalysed by CadB will result in the generation of an interior negative membrane potential. When right side-out membrane vesicles were prepared without cadaverine, no significant lysine uptake was observed (□), indicating that this process is dependent on the presence of cadaverine as a countersubstrate.

Figure 6.

Electrogenic exchange of lysine with cadaverine in right side-out membrane vesicles of JM109/pUCcadB. Right side-out membrane vesicles were loaded with 1 mM cadaverine (•, ○, ▵, ▪) or without cadaverine (□) plus 100 mM phosphate as the NMG salt (•, ○, ▵, □) or the potassium salt (▪) and diluted 10-fold by the reaction mixture containing 100 µM [14C]-lysine along with 100 mM sulphate plus 100 mM phosphate as the potassium salt (•) or the NMG salt (○, ▵, ▪, □) in the presence (•, ▵, ▪, □) and absence (○) of 1 µM valinomycin. Values are mean ± SE of triplicate determinations.

Expression of cadBA and speF–potE operons

It has been reported that expression of the cadBA operon is induced by lysine at acidic pH (Neely and Olson, 1996). Expression of the cadBA operon was compared at neutral and acidic pH using MA261 cells cultured in synthetic medium in the presence and absence of lysine or cadaverine (Fig. 7A). Expression of the cadBA operon was greatly stimulated only when cells were cultured at acidic pH in the presence of lysine. Expression of the cadBA operon was low at neutral pH even in the polyamine-requiring mutant MA261. Expression of the cadBA and speF–potE operons was then compared in E. coli B and BcadC::Km cultured at acidic pH in rich medium. In E. coli B, the degree of induction of CadB mRNA by lysine was 3.5-fold. Using BcadC::Km cells, in which cadC, a positive regulator of cadBA expression, is disrupted, there was no induction of CadB mRNA by lysine. Similarly, PotE mRNA could be induced 4.5-fold by ornithine. Although the precise comparison of the degree of expression of two operons is difficult by dot-blot analysis, the level of PotE mRNA in the presence of ornithine seems to be about 25% of the level of CadB mRNA in the presence of lysine. At acidic pH, the level of CadB mRNA may be equal to or greater than that of PotE mRNA regardless of the presence of lysine or ornithine, unless cadC was disrupted. When the cadC gene was disrupted, the level of PotE mRNA was increased fivefold. The results suggest that PotE and ornithine decarboxylase have functions that can compensate for the loss of CadB and lysine decarboxylase.

Discussion

In this study, we have shown that CadB has not only cadaverine uptake activity but also cadaverine-lysine antiport activity. These data were obtained using E. coli MA261 intact cells, which are deficient in polyamine biosynthesis, and right side-out and inside-out membrane vesicles prepared from E. coli JM109 cells. This was because preparation of right side-out and inside-out membrane vesicles of MA261 strain was difficult and more active membrane vesicles could be made from E. coli JM109 cells. At neutral pH, CadB is involved in cadaverine uptake consuming a proton motive force as an energy source (Fig. 8). Cadaverine and aminopropylcadaverine can presumably substitute for the intracellular functions of putrescine and spermidine if these two polyamines are not available. However, expression of the cadBA operon was low at neutral pH. Consequently, the physiological effect of CadB at neutral pH is thought to be small, and the effect would appear only in the specific polyamine-deficient strain. At acidic pH, CadB functions as a cadaverine-lysine antiporter (Fig. 8). The process of cadaverine excretion and lysine uptake catalysed by CadB generates a membrane potential, and decarboxylation of lysine by lysine decarboxylase generates a pH gradient through consumption of a cytoplasmic proton. This proton motive metabolic cycle causes neutralization of acidic conditions and an increase in the level of ATP in cells (Figs 4 and 8). CO2 produced by lysine decarboxylase may also contribute to the various metabolic pathways including nucleotide biosynthesis. Although the surrounding concentration of protons is much higher at acidic pH, CadB functions as a cadaverine-lysine antiporter, which benefits bacterial cell growth. Thus, the cadBA operon plays important roles in cell growth at acidic pH. In this respect, CadB may contribute more than PotE because the cadBA operon seems to be expressed more than the speF–potE operon. This seems rational because lysine in nature is more common than ornithine. However, in a cadC-disrupted mutant, expression of the speF–potE operon is enhanced. These results suggest that the speF–potE operon replaces the function of the cadBA operon at acidic pH or that cadC is a negative regulator of the speF–potE operon. As for the enhancement of expression of the speF–potE operon, we have reported previously that RNase III increases the translational efficiency of iODC-PotE mRNA by cutting the 5′ untranslated region of the mRNA (Kashiwagi et al., 1994). However, it remains to be clarified whether RNase III is activated at acidic pH or not.

Figure 8.

Physiological functions of CadB in E. coli. CadA, lysine decarboxylase; CadB, cadaverine transporter. In neutral conditions, CadB functions as a cadaverine and proton symporter, and cell growth was stimulated by cadaverine. In acidic conditions, CadB functions as an electrogenic cadaverine-lysine antiporter, and pH in the medium is increased by cadaverine. CadA (lysine decarboxylase) generates a pH gradient by consumption of a cytoplasmic proton. This process causes the increase in the level of ATP in cells.

With regard to uptake activity, PotE may be more efficient than CadB, because the Km value of PotE for putrescine is 1.8 µM (Kashiwagi et al., 1997), whereas that of CadB for cadaverine is 20.8 µM. Polyamines (putrescine, spermidine and spermine) are known to stimulate protein synthesis (Yoshida et al., 1999; 2001; 2002; Igarashi and Kashiwagi, 2000). The effective concentrations required for stimulation of protein synthesis are in the order putrescine > spermidine > spermine (Igarashi et al., 1974). Cadaverine and aminopropylcadaverine also stimulate protein synthesis, and their effective concentrations are similar to those of putrescine and spermidine respectively (Kakegawa et al., 1988). Thus, the slower growth of cells with cadaverine than with putrescine in the polyamine-requiring mutant MA261 (Table 1) was probably the result of less accumulation of cadaverine than putrescine in the cells (Table 2). At acidic pH, cadaverine and aminopropylcadaverine accumulate in cells through the function of CadB and lysine decarboxylase, but the synthesis of putrescine and spermidine was inhibited to maintain an optimal concentration of ‘polyamines’ including cadaverine and aminopropylcadaverine.

Polyamine content in cells was regulated by polyamine biosynthesis, degradation and transport. Although polyamine biosynthesis and degradation are well characterized, properties of polyamine transport are not. In E. coli, we reported that there are two ATP-dependent polyamine uptake systems in addition to PotE and CadB (Igarashi and Kashiwagi, 1999), but only a PotE (or CadB)-like protein exists as a polyamine transport system in Rickettsia prowazekii (Andersson et al., 1998). Similarly, only PotE (or CadB)-like proteins have been found thus far as polyamine transport systems in Saccharomyces cerevisiae (Tomitori et al., 1999; 2001). Thus, the secondary polyamine uptake systems such as PotE or CadB may be the major polyamine uptake systems in certain cell types.

Experimental procedures

Bacterial strains and plasmids

A polyamine-requiring mutant, E. coli MA261 (speB speC thr leu ser thi; Cunningham-Rundles and Maas, 1975), was provided by Dr W. K. Maas, New York University School of Medicine. For disruption of the cadC gene, a 3.7 kb SalI fragment containing the cadC gene of Kohara's ♯648 (5G7) λ DNA (Kohara et al., 1987) was inserted into the same restriction site of pUC9 (Gibco BRL) to construct pUCcadC. A 1.3 kb HincII fragment conferring kanamycin (Km) resistance of pUC4K (Amersham Biosciences) was inserted into the SnaBI site of pUCcadC to construct pUCcadC::Km. pUCcadC::Km was then digested with SalI, and the 5 kb fragment containing cadC::Km was introduced into W3110recD::Tn10 to obtain W3110recD::Tn10 cadC::Km. E. coli MA261cadC::Km and E. coli BcadC::Km were constructed by means of P1 transduction, in which W3110recD::Tn10 cadC::Km was used as a donor and MA261 and B (laboratory stock) were used as recipient strains. MA261cadB::Km was also constructed by means of P1 transduction from a cadB gene-disrupted mutant JW4093, which was kindly provided by Drs T. Baba and H. Mori, Institute for Advanced Biosciences, Keio University, and Research and Education Center for Genetic Information, Nara Institute of Science and Technology, Japan.

For the construction of pMWcadB, a polymerase chain reaction (PCR) was performed using 5′-CAGTGGATCCTG GTCAGGAAATAGTTA-3′ and 5′-CCTAGGAATTCAATATTG CAATAACGTTC-3′ as primers and Kohara's ♯648 (5G7) λ DNA as a template. After digesting the PCR product with BamHI and EcoRI, a 1.5 kb fragment containing the cadB gene was inserted into the same restriction sites of pMW119 (Nippon Gene). pUCcadB was constructed by inserting the 1.5 kb BamHI and EcoRI fragments of pMWcadB into the same restriction sites of pUC119 (Takara Shuzo).

Culture conditions

Escherichia coli MA261, MA261cadC::Km and MA261cadC::Km/pMWcadB were cultured in medium A, containing 4 g of glucose, 7 g of K2HPO4, 3 g of KH2PO4, 500 mg of trisodium citrate/2H2O, 1 g of (NH4)2SO4, 100 mg of MgSO4/7H2O, 2 mg of thiamine, 10 mg of biotin and 100 mg each of leucine, threonine, methionine, serine, glycine and ornithine per litre of water, in the presence and absence of putrescine or cadaverine as described previously except that pH in the medium was adjusted by changing the ratio of K2HPO4 and KH2PO4 at the total phosphate concentration of 62 mM (Kashiwagi et al., 1990). E. coli JM109 [recA1 supE44 endA1 gyrA96 relA1 thi Δ(lac-proAB)/F′ΔtraD36 proAB+lacIqlacZΔM15)] (Sambrook et al., 2001a) containing pMWcadB or pUCcadB was cultured in a 19-amino-acid-supplemented medium containing 1% glycerol (Kashiwagi and Igarashi, 1988). For isolation of RNA, E. coli B and BcadC::Km were cultured without shaking in a medium containing 1% tryptone, 0.5% yeast extract, 0.5% NaCl and 50 mM MES-NaOH, pH 5.5, in the presence and absence of 20 mM ornithine or 20 mM lysine. Ampicillin (100 µg ml−1), kanamycin (50 µg ml−1) and/or tetracycline (15 µg ml−1) were added to the medium, if necessary.

Cadaverine uptake by intact cells and by right side-out membrane vesicles

Escherichia coli MA261cadC::Km/pMW119 or pMWcadB cells grown in medium A containing 0.5 mM IPTG were suspended in buffer I containing 0.4% glucose, 62 mM potassium phosphate, pH 7.0, 1.7 mM sodium citrate, 7.6 mM (NH4)2SO4 and 0.41 mM MgSO4 to yield a protein concentration of 0.1 mg ml−1. The cell suspension (0.48 ml) was preincubated at 30°C for 5 min, and the reaction was started by the addition of 0.02 ml of 0.25 mM [14C]-cadaverine (370 MBq mmol−1; Sigma). Thus, the concentration of [14C]-cadaverine in the reaction mixture was 10 µM. After incubation at 30°C for 30 s to 3 min, the cells were collected on cellulose acetate filters (0.45 µm; Advantec Toyo), and the filters were washed three times with a total of 12 ml of buffer I. The radioactivity on the filters was measured with a liquid scintillation spectrometer. Protein content was determined by the method of Lowry et al. (1951).

To determine the energy source for cadaverine uptake, cadaverine uptake activity by right side-out membrane vesicles was measured according to the method of Tolner et al. (1995). E. coli JM109/pUCcadB cells were cultured until A540 = 0.5 in the presence of 0.5 mM IPTG. Right side-out membrane vesicles were prepared by the method of Kaback (1971) and suspended in 0.1 M potassium phosphate, pH 6.6, at a protein concentration of 5–10 mg ml−1. The reaction mixture (0.1 ml) for the uptake by right side-out membrane vesicles contained 10 mM Tris-HCl, pH 7.5, 50 mM potassium phosphate, pH 6.6, 10 mM MgSO4, 50 mM KCl, 20 mM ascorbic acid, 10 mM phenazine methosulphate (PMS) and 100 µg of right side-out membrane vesicle protein. Nigericin (10 µM), valinomycin (8 µM) or CCCP (40 µM) was added to abolish the transmembrane proton gradient (ΔpH), transmembrane electrical potential (ΔΨ) or the proton motive force (Δp) respectively (Tolner et al., 1995). The membrane suspension (0.095 ml) was preincubated at 30°C for 5 min, and the reaction was started by the addition of 5 µl of 1 mM [14C]-cadaverine (1.48 GBq mmol−1). Thus, the concentration of [14C]-cadaverine in the reaction mixture was 50 µM. After incubation at 30°C for 30 s to 3 min, membrane vesicles were collected on cellulose nitrate filters (0.45 µm; Advantec Toyo) and washed with 10 mM Tris-HCl, pH 7.5, and 0.5 M LiCl.

Cadaverine uptake by artificial ion gradients was also measured (Tolner et al., 1995). Right side-out membrane vesicles prepared as described above were resuspended in buffer II [20 mM MES (morpholinoethanesulphonic acid), 100 mM acetic acid, 100 mM potassium hydroxide, 5 mM MgSO4, pH adjusted to 6.0 by H2SO4] to make a protein concentration of ≈ 40 mg ml−1. The reaction mixture (0.1 ml) (buffer II to V) containing 50 µM [14C]-cadaverine (1.48 GBq mmol−1) and 100 µg of right side-out membrane vesicle protein was incubated at 30°C for 10 s to 2 min. Buffers used to generate Δp, ΔΨ and ΔpH were buffer III (120 mM MES, 100 mM methylglucamine, 1 µM valinomycin and 5 mM MgSO4, pH 6.0), buffer IV (20 mM MES, 100 mM acetic acid, 100 mM methylglucamine, 1 µM valinomycin and 5 mM MgSO4, pH 6.0) and buffer V (120 mM MES, 100 mM KOH, 1 µM valinomycin and 5 mM MgSO4, pH 6.0) respectively. Membrane vesicles were then collected on cellulose nitrate filters and washed with 0.1 M KCl.

Preparation of inside-out membrane vesicles and cadaverine or lysine uptake by the vesicles

Cadaverine–lysine, cadaverine–cadaverine and lysine–lysine antiport activities by CadB were estimated by measuring cadaverine or lysine uptake by inside-out membrane vesicles prepared according to the method of Houng et al. (1986). E. coli JM109/pMWcadB cells were cultured until A540 = 0.5 in the presence of 0.5 mM IPTG. Inside-out membrane vesicles were prepared by French press treatment (10 000 p.s.i.) of the cells suspended in a buffer containing 100 mM potassium phosphate buffer, pH 6.6, 10 mM EDTA in the presence and absence of 2.5 mM lysine or cadaverine (Kashiwagi et al., 1992). After removal of unbroken cells and cell debris, membrane vesicles were collected by ultracentrifugation (170 000 g, 1 h). Membrane vesicles were washed with a buffer containing 10 mM Tris-HCl, pH 7.5, 0.14 M KCl, 2 mM 2-mercaptoethanol and 10% glycerol, collected by ultracentrifugation and suspended in the same buffer at the protein concentration of 5–10 mg ml−1. The reaction mixture (0.1 ml) for the uptake by inside-out membrane vesicles contained 10 mM Tris-HCl, pH 8.0, 0.14 M KCl, 100 µM [14C]-cadaverine or [14C]-lysine (1.48 GBq mmol−1) and 100 µg of inside-out membrane vesicle protein. After incubation at 22°C for 10 s to 1 min, membrane vesicles were collected on cellulose nitrate filters and washed twice with a total of 9 ml of washing buffer (10 mM Tris-HCl, pH 8.0, and 0.14 M KCl). The radioactivity on the filters was measured with a liquid scintillation spectrometer.

Assay for electrogenic exchange of lysine with cadaverine

Lysine uptake by cadaverine-loaded right side-out membrane vesicles was measured according to the method of Abe et al. (2002). Right side-out membrane vesicles prepared as described above were incubated at 17°C for 2 h in the loading buffer containing 1 mM cadaverine and either 100 mM N-methylglucamine (NMG) phosphate, pH 7.0, plus 100 mM NMG sulphate, or 100 mM potassium phosphate, pH 7.0, plus 100 mM potassium sulphate. The reaction mixture (0.2 ml) for the lysine uptake by cadaverine-loaded right side-out membrane vesicles containing 100 µM [14C]-lysine (1.48 MBq/mmol−1) and either 100 mM potassium phosphate, pH 7.0, plus 100 mM potassium sulphate, or 100 mM NMG phosphate, pH 7.0, plus 100 mM NMG sulphate was preincubated at 17°C for 1 min in the presence and absence of 1 µM valinomycin. The reaction was started with the addition of 100 µg of cadaverine-NMG-loaded or cadaverine-potassium-loaded right side-out membrane vesicle protein. In this way, it was possible to generate a membrane potential with polarity that was either interior positive (potassium outside, NMG inside) or interior negative (NMG outside, potassium inside). After incubation at 17°C for 10 s to 2 min, membrane vesicles were collected on cellulose nitrate filters and washed with the same buffer as the reaction buffer without [14C]-lysine. The radioactivity on the filters was measured with a liquid scintillation spectrometer.

Dot-blot analysis of mRNA

Total RNA was isolated by the method of Emory and Belasco (1990). Dot-blot analysis of CadB and PotE mRNAs was performed according to the method of Sambrook et al. (2001b). The 1.5 kb BamHI and EcoRI fragment of pMWcadB or the 1.6 kb SphI–BamHI fragment of pUCpotE (Kashiwagi et al., 1997) was labelled with [α-32P]-dCTP using a BcaBESTTM labelling kit (Takara Shuzo) and used as a probe. The radioactivity on the blot was quantified by BAS2000II imaging analyser (Fuji Film).

Measurement of polyamines

Polyamine levels in E. coli were determined by high-pressure liquid chromatography as described previously (Igarashi et al., 1986) after homogenization and extraction of the polyamines with 5% trichloroacetic acid and centrifugation at 27 000 g for 15 min at 4°C. The retention times for putrescine, cadaverine, spermidine, aminopropylcadaverine and spermine were 6.7, 9.3, 13, 18 and 25 min respectively.

Measurement of ATP

ATP levels in E. coli were determined using the luciferase enzyme assay system according to the method of Kimmich et al. (1975). The luminescence was determined by TD-20/20 luminometer (Turner Designs).

Acknowledgements

We thank Drs K. Williams and A. J. Michaels for their help in preparing the manuscript. Thanks are also due to Drs W. K. Maas, New York University School of Medicine, Y. Kohara, National Institute of Genetics, Japan, Drs T. Baba and H. Mori, Institute for Advanced Biosciences, Keio University, and Research and Education Center for Genetic Information, Nara Institute of Science and Technology, Japan, for their generous contributions of E. coli MA261, Kohara's ♯648 (5G7) λ phage and cadB gene-disrupted mutant, JW4093. This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan, and by Terumo Life Science Foundation, Japan.

Ancillary