(Received for publication, January 21, 1997)
From the Departments of Biochemistry and Radiation
Oncology, The University of Arizona Health Science Center, Tucson,
Arizona 85724 and the § Center for Pharmaceutical
Biotechnology and the Department of Medicinal Chemistry and
Pharmacognosy, University of Illinois,
Chicago, Illinois 60607
Multidrug transporters pump structurally dissimilar toxic molecules out of cells. It is not known, however, if detoxification is the primary physiological function of these transporters. The chromosomal organization of the gene encoding the Bacillus subtilis multidrug transporter Blt suggests a specific function for this protein; it forms a single operon with another gene, bltD, whose protein product is identified here as a spermine/spermidine acetyltransferase, an enzyme catalyzing a key step in spermidine degradation. Overexpression of the Blt transporter in B. subtilis leads not only to the multidrug-resistance phenotype but also to the efflux of large amounts of spermidine into the medium; this efflux is supressed by an inhibitor of Blt, reserpine. Taken together, these results strongly suggest that the natural function of the Blt transporter is the efflux of spermidine, whereas multiple drugs may be recognized by Blt merely opportunistically.
Multidrug transporters, first discovered in mammalian cells (P-glycoprotein) (1) and later found in yeast (2) and many species of bacteria (3-5), pump structurally dissimilar toxic molecules, including many anticancer, antifungal, and antibacterial agents, out of cells. The ability of each multidrug transporter to recognize dozens of compounds that have no apparent structural consensus seemingly contradicts basic dogmas of biochemistry and remains enigmatic.
Another unresolved question associated with multidrug transporters concerns their normal physiological functions. Because overexpression of these proteins causes an increase in cellular resistance to toxins, it is commonly believed that they have evolved to protect cells from diverse environmental toxic molecules. It is possible, however, that each multidrug transporter has a specific, presently unidentified natural substrate, e.g. some cellular metabolite, whereas multiple drugs are effluxed by them merely opportunistically.
This latter hypothesis is indirectly supported by the analysis of the primary structures of multidrug transporters. More than two dozen known transporters capable of effluxing multiple drugs share no common sequence motif and belong to at least four distinct superfamilies of membrane proteins (1-5). Furthermore, many multidrug transporters display strong homology to substrate-specific transporters. For example, human P-glycoprotein, the product of the MDR1 gene, is highly homologous to the protein MDR2, which does not efflux drugs but is a specific phosphatidylcholine flippase (6). Similarly, the bacterial multidrug transporters Bmr and Blt of Bacillus subtilis and NorA of Staphylococcus aureus are highly homologous to the efflux transporters whose only known substrate is tetracycline (7-10).
It has recently been demonstrated that P-glycoprotein transports not only drugs but also structurally diverse natural lipids across the membrane (11). Here we demonstrate that the activity of the B. subtilis multidrug transporter Blt leads to the efflux of a natural cellular constituent, the polyamine spermidine. Moreover, the chromosomal organization of the Blt transporter gene, namely its association, within the same operon, with the gene encoding a spermidine-acetylating enzyme, strongly suggests that the efflux of spermidine is the primary physiological function of this protein.
The bltD
gene (8) was amplified by polymerase chain reaction using B. subtilis DNA as a template and cloned into the vector pTrc99A
(Pharmacia Biotech Inc.) so that its expression was under the control
of the vector trc promoter and ribosome binding site. The
expression of BltD was induced in JM109 or CAG2242 E. coli cells transformed with this construct by incubating them for 3 h
with 1 mM
isopropyl-1-thio--D-galactopyranoside. Cells resuspended in 50 mM Tris-HCl, pH 7.5, were then lysed in a French
press (10,000 psi) and centrifuged at 100,000 × g for
2 h. BltD comprised approximately 10% of proteins in the
supernatant.
The protein extract of the BltD-expressing E. coli and a similarly prepared extract of control JM109 or CAG2242 E. coli cells were incubated in a 10-fold volume of 50 mM Hepes buffer, pH 7.5, containing 1 µCi/ml [14C]acetyl-CoA (Amersham Corp.) and 1 mM prospective BltD substrates, which were purchased from Sigma. After 15 min of incubation the samples were applied to a cellulose TLC plate and chromatographed in a n-butanol/pyridine/water (1:1:1) system. The position of the label was then revealed by autoradiography.
Alternatively, 0.01 µCi of [14C]spermine or [14C]spermidine (Amersham Corp.) was incubated for 15 min in 10 µl of 50 mM Hepes buffer, pH 7.5, with 1 µl of the extracts of the CAG2242 E. coli either expressing or not expressing BltD and in either the presence or the absence of 100 µM acetyl-CoA. After incubation the reaction mixtures were chromatographed on a cellulose plate in n-butanol/acetic acid/water (12: 3: 5) and autoradiographed.
Analysis of Intracellular Polyamines and Measurement of Spermidine EffluxCells were treated with lysozyme (10 mg/ml) and then sonicated to disruption in 0.1 N HCl. Lysates were then adjusted to 0.2 N perchloric acid and clarified by centrifugation. Polyamines were then separated by HPLC1 and detected fluorimetrically as described elsewhere (12). To assess the polyamine efflux, B. subtilis cells were grown to A600 0.5, pelleted and resuspended to A600 2 in MG1 medium (13). After 40 min of incubation at 37 °C with intensive aeration, cells were removed by centrifugation, and the medium was filtered to remove remaining bacteria, supplemented with 0.2 N HClO4, and analyzed for polyamine content by HPLC (12).
The B. subtilis multidrug transporter Blt, a member of the major facilitator superfamily of transporters, effluxes from cells a number of toxic molecules, such as fluoroquinolone antibiotics, ethidium, rhodamine, tetraphenylphosphonium, acridine dyes, puromycin, doxorubicin, etc. (8). The only common features of these compounds are their moderate hydrophobicity and positive charge (the only electronegative substrate of Blt, fluoroquinolone antibiotics, may acquire a positive charge by chelating divalent cations (14)).
We have shown previously that the blt gene is cotranscribed with another gene, bltD, which is located downstream of blt (8). The putative protein product of this gene, BltD, displays strong sequence similarity to a number of substrate-specific acetyltransferases that act by transferring acetyl groups from acetyl-CoA to the amino groups of either antibiotics, specific proteins, or polyamines (8). The fact that the blt and bltD genes comprise a single operon strongly suggests that the physiological functions of their protein products are closely related. Therefore, to identify the natural function of the Blt transporter, we attempted to determine the substrate specificity of the bltD-encoded enzyme.
BltD was expressed in E. coli, and a number of potential substrates were tested for the ability to accept a radioactive label from [14C]acetyl-CoA in the presence of BltD-containing E. coli extract. The transfer of the label from [14C]acetyl-CoA to aminoglycoside antibiotics (gentamycin, kanamycin, and sisomycin), some of the Blt-transported drugs (acriflavine and rhodamine 6G), polyamines (putrescine, cadaverine, spermidine, spermine, and norspermine), and miscellaneous organic amines (aniline, ethanolamine, benzylamine, lysine, p-aminobenzoic acid, and choline) was monitored by following a change in the chromatographic mobility of the label. Of all the substrates tested, only two polyamines, spermidine and spermine, served as acceptors of the label.
To verify that BltD is indeed a spermine/spermidine acetyltransferase
and not an activator of the endogenous E. coli enzyme, we
expressed the protein in the E. coli strain CAG2242 (15) that is deficient in polyamine acetylation (16). Fig. 1
demonstrates that [14C]spermine changes its
chromatographic mobility after incubation with acetyl-CoA and
BltD-containing CAG2242 extract. A similar result was obtained with
[14C]spermidine.
The finding that BltD is a spermine/spermidine acetyltransferase was confirmed by HPLC analysis of the reaction products. The position of the HPLC peak corresponding to the product of spermidine acetylation by BltD coincided with the peak position of the N1-acetylspermidine but not of N8-acetylspermidine standard. Acetylation of spermine resulted in a product not detectable by the fluorimetric technique used in HPLC (12), suggesting modification of both its N1 and N12 primary amino groups. The product of [14C]spermine acetylation eluted in HPLC as a single peak; upon acid hydrolysis (6 M HCl at 110 °C for 18 h) it was converted back to spermine. Finally, N1-acetylspermine incorporated an equimolar amount of labeled acetyl groups when incubated with BltD and [14C]acetyl-CoA. All these data indicate that BltD converts spermine into N1,N12-acetylspermine.
BltD exhibits a high affinity for polyamines; the Km is 200 µM for spermidine and 67 µM for spermine, which makes it a more potent acetyltransferase than the E. coli spermidine acetyltransferase, whose Km for spermidine is at a millimolar level (17). More detailed description of the enzymatic properties of BltD will be published elsewhere.2
Spermidine comprises 90-95% of polyamines in B. subtilis,
whereas spermine is present only in trace amounts (Fig.
2 and Ref. 18). Under standard cultivation conditions
neither Blt nor BltD are expressed in the control B. subtilis strain BD170 (thr-5, trpC2) (8) and no
products of polyamine acetylation can be detected (Fig. 2). We have
shown previously that the acfA mutation (19) changes the
sequence of the blt-bltD operon promoter making it
constitutively active (8). This mutation was introduced into BD170
cells resulting in the BD170/acfA strain (8). HPLC analysis
of polyamines in this strain revealed the presence of a significant
amount of N1-acetylspermidine (Fig. 2),
which was apparently due to the expression of BltD. In animal cells,
polyamine acetylation is considered the rate-limiting step in polyamine
degradation and results in depletion of the intracellular polyamine
pools (20). Indeed, the amount of spermidine in the
acfA-carrying cells was significantly reduced as compared
with control cells; whereas BD170 cells contained 59 nmol of
spermidine/mg of protein, BD170/acfA cells contained only 31 nmol of spermidine/mg of protein.
The acfA mutation leads to the expression of not only BltD
but also Blt. It was, therefore, tempting to speculate that spermidine is a substrate of not only BltD but also of the Blt transporter and
that the Blt-mediated efflux of spermidine contributes to the reduction
of the intracellular spermidine content in these cells. To verify this
hypothesis, control BD170 cells and BD170/acfA cells were
incubated for 40 min in a minimal medium, and the polyamines released
into the medium were quantified by HPLC analysis. The results of these
measurements supported our hypothesis that BD170/acfA cells
efflux spermidine into the medium because these cells released four
times more of this polyamine than did the BD170 cells (Fig. 3A). After incubation, as much as 24% of the
total spermidine in the BD170/acfA culture was detected in
the medium, whereas in the control BD170 culture this value was only
4%. Furthermore, reserpine, an inhibitor of the Blt transporter (8)
dramatically reduced spermidine release from BD170/acfA
cells to control levels while having no effect on spermidine release
from control BD170 cells (Fig. 3A).
A similar experiment was performed with a different pair of strains: BD224 (recombination-deficient strain isogenic to BD170) and the Blt-expressing strain BD224/pBLT (8), in which Blt is expressed from a plasmid vector. No BltD expression occurs in either of these strains. Nevertheless, the intracellular pool of spermidine was reduced in the Blt-expressing strain by more than a half as compared with the control strain (32 versus 79 nmol of spermidine/mg of protein). Accordingly, the amount of spermidine released into the medium was almost five times higher for the Blt-expressing strain than for the control strain, which again was dramatically inhibited by reserpine (Fig. 3B).
The results presented here demonstrate that the activity of the B. subtilis multidrug transporter Blt causes the release of spermidine from bacteria. The fact that Blt is encoded within the same operon as BltD, which, as we showed here, is a specific enzyme acetylating spermidine, strongly suggests that spermidine is not simply one of the many possible substrates of Blt but rather that Blt has evolved as a specific spermidine-efflux protein. Apparently, B. subtilis uses the blt-bltD operon to down-regulate the amount of spermidine in the cell through two distinct mechanisms: efflux and chemical modification. The physiological conditions under which the expression of this operon is activated remain to be identified.
The finding that spermidine appears to be a physiological substrate of Blt raises the intriguing question as to why Blt would also efflux multiple drugs. It is possible that to create an efficient transporter of spermidine, a relatively hydrophobic and positively charged but otherwise shapeless flexible molecule, evolution had to come up with a very nonselective substrate recognition site. As a result, Blt can opportunistically recognize and transport those multiple drugs that, like spermidine, are moderately hydrophobic and positively charged. It remains to be seen if all other multidrug transporters have specific, as yet unidentified, physiological substrates.
We are grateful to Drs. A. S. Mankin, T. Tenson, and A. V. Chervonsky for critical reading of the manuscript.