(Received for publication, January 16, 1996, and in revised form, November 18, 1996)
From the Department of Biochemistry, Indian Institute of Science, Bangalore, 560 012 India
Heme deficiency precipitated by
CoCl2 administration to rats leads to a striking
decrease in the inducibility of CYP2B1/B2 mRNA levels and its
transcription by phenobarbitone (PB), besides decreasing the basal
levels. Exogenous hemin administration counteracts the effects of
CoCl2 administration. The binding of nuclear proteins to
labeled positive cis-acting element (69 to
98 nucleotides) in the
near 5
-upstream region of the gene is inhibited by CoCl2 administration to saline or PB-treated rats, as assessed in gel shift
assays. Administration of exogenous hemin to the animal or addition
in vitro to the extracts is able to overcome the effects of
CoCl2 treatment. The protein mediating this effect has been purified from CoCl2 administered nuclear extracts by
heparin-agarose, positive element oligonucleotide affinity, and heme
affinity column chromatography. This 65-kDa protein manifests very
little binding to the positive element, but in the presence of certain
other nuclear proteins, shows a strong heme-responsive binding. The purified protein binds heme. It is also able to stimulate transcription of a minigene construct of the CYP2B1/B2 gene containing
179 nucleotides of the 5
-upstream region and the I exon in a cell-free system, manifesting heme response. It is concluded that the 65-kDa protein mediates the constitutive requirement of heme for the transcription of CYP2B1/B2 gene.
The transcriptional regulation of the cytochrome P-450 (CYP)
supergene family is of considerable interest. The CYP1A1 gene of rat
liver, induced by the prototype drug 3-methylcholanthrene, has been
studied in detail. It involves interaction of the ligand, 3-methylcholanthrene, with the Ah receptor, translocation of the receptor to the nucleus, and interaction with specific upstream elements (1, 2). The details regarding the mechanism of transcriptional
activation of the CYP2B1/B2 gene (2B1 and B2 are 97% homologous and
hence treated as a unit) by the prototype drug, phenobarbitone
(PB),1 are beginning to emerge. Studies in
this laboratory have led to the identification of a positive cis-acting
element (69 to
98 nt) and a negative cis-acting element (
126 to
160 nt) in the near 5
-upstream region of the CYP2B1/B2 gene (3-5).
The positive element includes the 17-base pair PB-responsive consensus element, referred to as Barbie Box, identified first by Fulco and
co-workers (6-8) in Bacillus megaterium, rat, mice, and
other organisms of barbiturate inducible cytochrome P-450 and other proteins. There have also been other PB-responsive sequences identified in the CYP2B1/B2 gene of the rat. Shephard et al. (9) have identified two sequences, located between
183 to
199 nt and
31 to
72 nt, to be PB-responsive, although these do not include the
"Barbie" elements. In addition, Trottier et al. (10)
have located a sequence between
2155 and
2318 nt to be
PB-responsive and Ramsden et al. (11) have identified an
upstream enhancer as far as
20 kilobases. Studies in this laboratory
have led to the development of a model which proposes that a ~26-kDa
protein interacts with the positive or negative element based on its
phosphorylation status. It is proposed that the binding of the
~26-kDa protein with the positive element in its phosphorylated state
and with the negative element in its dephosphorylated state, within the PB-responsive minimal promoter in the near 5
-upstream region of the
CYP2B1/B2 gene, would determine the inducible and basal states of the
gene, respectively. The inducible state would involve the interaction
of the positive element with the upstream enhancer elements, mediated
through protein-protein interaction (12).
Earlier studies from this laboratory have also implicated that heme, the prosthetic group of cytochrome P-450, is a positive modulator of CYP2B1/B2 gene transcription (13-16). This conclusion is based on the fact that inhibitors of heme biosynthesis block induction of CYP2B1/B2 mRNA by PB and its run-on transcription in isolated nuclei. This inhibition is counteracted by the administration of exogenous hemin. In the present study it has been possible to demonstrate that a 65-kDa nuclear protein mediates the positive modulation of CYP2B1/B2 gene transcription by heme.
Rats (75-80 g) were injected with PB (8 mg/100 g, intraperitoneal), CoCl2 (6 mg/100 g, subcutaneously), hemin (75 µg/100 g, intraperitoneal), or saline in appropriate combinations. CoCl2 in general was injected 30 min before PB or hemin administration and the rats killed after 5 h.
Plasmids UsedThe plasmid pP450e179 containing 360 nt of
the CYP2B2 gene and covering positions 179 to +181 nt (3-5) was used
as a template in transcription reactions. pGEM-3Z vector was used to
generate riboprobes.
Nuclei were obtained from the
livers of treated rats using the citric acid homogenization procedure
(17). The nuclei were suspended in a buffer containing 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 0.5 mM MgCl2, 50 mM sucrose, and 10 mM -mercaptoethanol. NaCl from a stock of 4 M was added drop by drop to a final concentration of 0.45 M and the suspension left on ice for 30 min with
intermittent vortexing. The suspension was centrifuged at 12,000 rpm
for 10 min and the supernatant dialyzed against nuclear suspension
buffer.
Nuclear extract (1-3 µg) was used for binding with the body labeled
positive element (5-GAGGAGTGAATAGCCAAAGCAGGAGGCGTG-3
) covering
98
nt to
69 nt of the CYP2B2 gene annealed to a complementary primer and
extended in presence of [
-32P]dCTP. Poly(dI-dC) was
used as a competitor and samples were analyzed on low ionic strength
9% polyacrylamide gels as described earlier (4, 5). A stock solution
(fresh) of hemin (1 mM) was prepared in 0.1 N
NaOH and then adjusted to pH 8.0 with acid. Appropriate dilutions were
made with 10 mM Tris-HCl buffer (pH 8.0) using this stock
for gel shift and other assays. Approximately 2 µl of hemin solution
is usually added to a reaction mixture of 20 µl.
Polysomal RNA was
isolated from magnesium-precipitated rat liver polysomes using
phenol-chloroform extraction procedure (18). CYP2B1/B2 mRNA was
quantified by RNase protection assay using [-32P]UTP-labeled I exon riboprobe. The labeled RNA
probe was hybridized to 20 µg of polysomal RNA, digested with
ribonuclease A and T1, and then analyzed on polyacrylamide-urea gels
(8% polyacrylamide) followed by autoradiography (5).
This was carried out by the procedure
of Guertin et al. (19) and has been described (3, 4).
Briefly, nuclei were incubated with 200 µCi of
[-32P]UTP and other components including human
placental RNase inhibitor in a total volume of 200 µl for 30 min at
25 °C. The labeled transcripts (107 cpm) were hybridized
to filters containing I, VI, and IX exon of the CYP2B2 gene. The
riboprobes corresponding to these exons were prepared using pGEM-3Z
vector. The filters were washed inclusive of RNase treatment and then
subjected to autoradiography.
The nuclear extracts
prepared for gel shift assays were used after 80%
(NH4)2SO4 fractionation. The
precipitate was dialyzed against nuclear suspension buffer and used.
The reaction in a total volume of 20 µl contained: DNA template (a
minigene construct containing 179 nt of the 5
-upstream region and
the I exon, pP450e179), 25 mM HEPES-KOH (pH 7.0); 50 mM KCl, 6 mM MgCl2, 0.6 mM each of ATP, GTP, CTP, and UTP, glycerol (12% in final
concentration), 30 units of RNase inhibitor and transcription extract
(50-100 µg of protein/ml). The transcription extract prepared as
described was found to be very active and 1-2 µg of protein per
20-µl reaction volume was found adequate. The reaction mixture was
incubated at 30 °C for 45 min. The RNA transcripts were then
isolated after RNase-free DNase and proteinase K treatments followed by
phenol/CHCl3 extraction and ethanol precipitation. The
transcripts were quantified by the RNase protection method described
earlier.
Nuclei were prepared from CoCl2-treated rats and the 0.45 M NaCl extract was dialyzed and passed through a heparin-agarose column in buffer containing 50 mM NaCl, 25 mM Tris-HCl (pH 7.4), and 2 mM EDTA. The column was washed with buffer containing 0.3 M NaCl. Further elution was carried out with buffer containing 0.4 M NaCl and then with 0.5 M NaCl. The peak protein fractions were pooled, dialyzed, and used for gel mobility shift assays. The 0.5 M NaCl eluate was further purified on oligoaffinity column. This was prepared by ligating a multimer of the positive element in pUC 19, amplified by polymerase chain reaction, to CNBr-activated Sepharose 4B (12). The column was washed with the binding buffer (25 mM HEPES-NaOH (pH 7.6), 12.5 mM MgCl2, 1 mM dithiothreitol, and 20% glycerol) and eluted with buffer containing 1 M NaCl. This was loaded onto a hemin-agarose column (in the presence of 1 M NaCl), washed thoroughly, and then eluted with 2 M guanidine-HCl in 50 mM Tris-HCl (pH 8.0). The eluate was extensively dialyzed against buffer (10 mM Tris-HCl (pH 7.4), 1 mM EDTA) with repeated changes. The dialysate was concentrated using membrane filters.
Binding of 65-kDa Protein with HeminThe purified protein
preparation was analyzed on SDS-PAGE (10% gel) and proteins visualized
using silver stain. To examine the heme binding property of the
purified protein, the preparation was incubated with
[59Fe]-hemin (10 µCi/mg, kind gift from National
Institute of Immunology, New Delhi) and then subjected to SDS-PAGE,
with the omission of -mercaptoethanol and the boiling step. The gel
was then subjected to autoradiography. Another approach was to
fractionate the protein-hemin complex on a Sepharose 4B column and
identify the radioactivity peaks corresponding to protein bound and
free hemin.
This was basically carried out by pooling up quadruplicates set up for gel shift assays with the 0.5 M NaCl eluate from the heparin-agarose column and the labeled positive element. The mixture was taken in a precooled microtiter plate and irradiated for 15 min on ice with UV light (254 nm, maximum intensity: 7000 uw/cm2) at a distance of 4 cm from the UV source. The sample was then electrophoresed on SDS-PAGE (10% gel) and autoradiographed. In another approach the gel shift assay mixture (pooled samples) after UV irradiation was analyzed by electrophoresis on a preparative 9% native polyacrylamide gel as described earlier. The wet gel was exposed to x-ray film overnight at 4 °C. Bands (II and III) corresponding to DNA-protein complexes were excised and once again exposed to UV. The complexes were then eluted in Tris-EDTA buffer and precipitated with 10% (w/v) trichloroacetic acid. The pellet was thoroughly washed with 70% ethanol, acetone, and ether to remove the acid. The final pellet was analyzed on SDS-PAGE as described earlier.
Earlier studies from this laboratory have shown that
CoCl2 treatment leads to a striking decrease in the heme
pool (13-16). The present study has also made use of this reagent to
precipitate heme deficiency. Earlier studies on mRNA quantification
and run-on transcription were based on filter hybridization and
measurement of radioactivity. Here these experiments have been repeated
using riboprobes and the sensitive method of RNase protection assay. The results presented in Fig. 1a indicate
that PB treatment leads to a striking increase in CYP2B1/B2 mRNA as
monitored using the exon I riboprobe. CoCl2 treatment leads
to a striking inhibition of PB-mediated induction of CYP2B1/B2
mRNA. Exogenous hemin administration leads to a partial but
significant counteraction of the effects of CoCl2. Hemin
administration at this concentration has no significant effects when
given to saline or PB-treated rats. Administration of CoCl2
to uninduced rats decreases even the basal level of CYP2B1/B2 mRNA
and this can be bought back to the basal level by the administration of
exogenous hemin (Fig. 1b). The results presented in Fig.
2 indicate that a significant increase in run-on
transcription is seen after PB treatment as monitored using exon I, VI,
and IX riboprobes. CoCl2 treatment inhibits a PB-mediated
increase in the transcription of CYP2B1/B2 gene and hemin
administration is able to counteract this effect.
The results presented in Fig. 3 indicate that cell-free
transcription of pP450e179, the minigene construct containing 179 nt
of the 5
-upstream region and exon I, is significantly enhanced after
PB treatment of the animal. The basal as well as induced levels of
transcription are inhibited by CoCl2 treatment. Hemin administration significantly counteracts the effects of
CoCl2 on PB treatment.
Studies from this laboratory have shown that the positive element (69
to
98 nt) generates three complexes (I, II, III) with crude nuclear
extract in gel shift assays and PB treatment of the animal leads to a
significant intensification of the complexes (4, 12). It was,
therefore, of interest to examine whether CoCl2 treatment
would influence binding of nuclear proteins to the positive cis-acting
element. The results presented in Fig. 4 reveal that PB
treatment leads to a significant increase in the intensity of complexes
I, II, and III and this increase in prevented by CoCl2
treatment. Interestingly, hemin administration in vivo
restores binding in extracts prepared from CoCl2 as well as
PB + CoCl2-treated rats. It does not have a significant
effect, when administered to saline or PB-treated rats. The results
presented in Fig. 5 (a, b, and c)
reveal that addition of hemin in vitro (maximal effect
obtained at 10
7 M) can also stimulate binding
of nuclear extracts prepared from CoCl2-treated rats to the
positive cis-acting element. At the level of the crude extract, all
three complexes show increased binding to the positive element in the
presence of heme. Again, it does not have a significant effect when
added to extracts from control (saline) or PB-treated rats. The results
presented in Fig. 5 also reveal that the stimulatory effect on binding
is specific to hemin addition in vitro and is not shared by
equivalent concentrations of protoporphyrin or FeCl3.
The next step was to purify the heme-responsive protein factor in the
nuclear extract. For this purpose, nuclear extracts prepared from
CoCl2-treated rats were fractionated on the heparin-agarose column. The results presented in Fig. 6 reveal that
hemin addition in vitro specifically stimulates binding of
the 0.5 M NaCl eluate from the column to the positive
cis-element. Complex II shows a striking response and complex III shows
a modest response. The 0.4 M NaCl eluate does not show a
heme response in terms of binding. A UV cross-linking analysis of the
0.5 M NaCl eluate gives a clear complex in SDS-PAGE
indicating a molecular mass of ~65 kDa after subtracting the
molecular mass of the oligonucleotide used (Fig. 7a). UV cross-linking analysis of complex II
as such gives an identical pattern (data not presented). A similar
analysis of complex III reveals a lower molecular mass protein in the
range 23-26 kDa (Fig. 7b).
On this basis, the 0.5 M NaCl eluate was loaded onto a
positive element affinity column and the bound proteins were eluted with 1 M NaCl. This was loaded onto the heme affinity
column and the bound proteins were eluted with 2 M
guanidine HCl. The eluate was extensively dialyzed and SDS-PAGE
analysis reveals a single protein band of 65 kDa (Fig.
8). Next, binding studies were carried out with the
purified protein and labeled positive element. The preparation shows
hardly any binding to the positive element, but in the presence of
small amounts of crude nuclear extract (100 ng of protein from
uninduced rats), which by itself does not generate easily detectable
complexes, manifests a striking heme-responsive increase in the
intensity of complex II. Both the purified protein and hemin are
required to elicit maximal response. Complex III shows a modest
response (Fig. 9, a and b). Qualitatively similar results are obtained when the purified protein is
used with nuclear extracts prepared from CoCl2-treated rats (data not presented). Addition of BSA to crude extract does not elicit
a similar response, indicating that the heme-dependent increase in binding to the positive element is specific to the 65-kDa
protein (Fig. 9c). It is interesting to note that complexes II and III but not complex I are manifest with 0.5 M
NaCl-heparin eluate of the CoCl2-treated extract. Once
again, complex II shows maximum response with hemin addition, although
complex III also manifests some response. In addition, with the heparin
column eluate maximum response is seen with 109
M hemin, whereas the optimal hemin concentration for crude
extract as already indicated is 10
7 M. The
purified ~65-kDa protein also shows a maximum response around
10
8 to 10
9 M hemin
concentration for binding to the positive cis-acting element. This
depends on the amount and source of crude nuclear extract added,
uninduced or CoCl2-treated, along with the 65-kDa protein
to elicit the response. It is found that higher concentrations of hemin
(10
6 to 10
5 M) are inhibitory.
Thus, a higher concentration of hemin (10
7 M)
is tolerated at the level of crude extract, since there are other
proteins that may be able to compete for binding. As the protein is
purified the competing proteins are removed and therefore the optimum
hemin concentration to facilitate maximum binding gets lower.
It was then of interest to examine whether the heme-affinity column
purified protein actually binds heme. For this purpose, the preparation
was incubated with [59Fe]-labeled hemin and then analyzed
on SDS-PAGE, omitting the boiling step and addition of
-mercaptoethanol. A radioactive complex with a mobility of ~65 kDa
is clearly seen on the autoradiogram (Fig. 10)
indicating that the protein binds heme. The
protein-[59Fe]-labeled hemin complex was also
fractionated on a Sepharose 4B column, which indicates that the
heme-binding proteins migrates in the region of a BSA standard of mass
of 65 kDa (Fig. 10b).
Finally, the purified preparation was examined for its potential to
manifest heme-responsive stimulation of the transcription of pP450e179
DNA in cell-free extracts. For this purpose, the crude transcription
extract was prepared from nuclei of CoCl2-treated rats. The
results presented in Fig. 11 (a and
b) on the basis of laser densitometric analysis reveal that
addition of hemin (108 M) or the purified
protein (2 µg) to the transcription extract shows about 3- and
4-5-fold stimulation of transcription, respectively. The addition of
the two together shows nearly 12-fold stimulation (Fig.
11a). There is also a significant graded response to the addition of increasing concentrations of the purified protein in the
presence of hemin (Fig. 11b).
The present study clearly reveals that CoCl2 treatment
leads to a striking inhibition of CYP2B/B2 gene transcription that is
overcome by the administration of low concentrations of hemin. The
basal (uninduced) as well as PB-induced CYP2B1/B2 mRNA levels are
also depressed by CoCl2 treatment and counteracted by hemin administration (Fig. 1). Run-on and cell-free transcription studies indicate that CoCl2 inhibits the PB-mediated increase in
transcription and this is counteracted by the administration of
exogenous hemin (Figs. 2 and 3). Cell-free transcription studies have
indicated that the heme response element is within the 179 nt of the
immediate 5-upstream region of the CYP2B1/B2 gene.
Gel shift analysis with the positive element (69 to
98 nt)
correlates well with the inhibitory and counteracting effects of
CoCl2 and hemin, respectively, on CYP2B1/B2 mRNA levels
and transcription status. As already reported (4, 12) crude nuclear extract gives three complexes (I, II, and III) with the positive element. The present study reveals that CoCl2 treatment
leads to a suppression of formation of all the three complexes both in
the presence and absence of PB treatment. At the level of the crude
nuclear extract isolated from CoCl2-treated rats, all the three complexes intensify on hemin addition (Fig. 4). The effect is
specific for hemin, since protoporphyrin or iron fail to elicit a
similar effect (Fig. 5). After fractionation on the heparin-agarose column, with the 0.5 M NaCl eluate, complex II shows a
striking increase in intensity to hemin addition and complex III shows a modest increase in intensity (Fig. 6). UV cross-linking analysis indicates that complex III consists of the ~26-kDa protein (Fig. 7).
This is in agreement with the results reported from this laboratory (12), where the ~26-kDa protein purified from the 0.4 M
NaCl eluate of the heparin-agarose column on the positive element
oligonucleotide affinity column, has been shown to give rise to complex
III with the positive element in gel shift analysis. The complex is
faint, but in the presence of low amounts of crude nuclear extract or on concentrations of the purified fraction all the three complexes (I,
II, and III) are manifest. The present study also reveals that complex
II consists of the ~65-kDa protein, although traces of lower
molecular weight proteins are also seen on UV cross-linking analysis.
This is corraborated by the findings that the 0.5 M NaCl
eluate of the CoCl2-treated nuclear extracts from the
heparin-agarose column gives a striking response to hemin addition in
terms of generating complex II with the positive element (Fig. 6) and
that a ~65-kDa heme-binding protein can be purified from this
fraction using oligonucleotide and heme affinity columns (Figs. 8 and
10). As already indicated, the 0.5 M NaCl eluate from the
heparin column also manifests heme-dependent generation of
complex III, although complex II is dominant. It may be pointed that
traces of low molecular mass proteins can be seen in the purified
65-kDa preparation. Once again binding of the purified ~65-kDa
protein to the positive element is very poor and significant binding in
the presence of hemin to the element is seen when small amounts of
crude nuclear extract are added (Fig. 9). Thus, protein-protein
interaction is essential for the binding of ~26- and ~65-kDa
proteins to the positive element. UV cross-linking analysis of
complexes II and III perhaps reveal only the proteins that make primary
contact with positive elements and the other proteins involved in
protein-protein interaction need to be identified.
In a recent study, it has been shown in this laboratory (12) that the
~26-kDa protein shows greater affinity to the positive (69 to
98
nt) and negative (
126 to
166 nt) elements in its phosphorylated and
dephosphorylated states, respectively, providing a basis for a
differential interaction of the protein with the two elements in
induced and uninduced states. While, the role of heme in the
interaction of proteins to negative element is yet to be investigated,
it is clear that interaction of the ~65-kDa protein with the positive
element is heme dependent. Since, heme depletion leads to a decrease in
basal as well as PB-induced transcription of the CYP2B1/B2 gene that is
counteracted by the exogenous administration of hemin, it can be
concluded that heme requirement is constitutive.
Earlier studies from this laboratory (3) have shown by Southwestern
blot analysis that a ~94-kDa protein shows up when the entire 179 nt
of the 5-upstream region of the CYP2B1/B2 gene is used as a probe. PB
treatment significantly enhances the intensity of this complex and this
increase is blocked by CoCl2 treatment of the animal. Thus,
there are at least three proteins with molecular mass values of ~94,
~65, and ~26 kDa, whose interaction with the positive element and
its neighborhood is influenced by the availability of heme and this in
turn may influence the interaction of this protein assembly with the
far upstream enhancer proposed by Ramsden et al. (11) as
well as the initiation complex, leading to productive transcription of
the CYP2B1/B2 gene. The ~65-kDa protein is perhaps the dominant
protein that interacts with heme and facilitates interaction with other
proteins and positive element. It needs to be studied whether heme has
a direct role in the interaction of the ~26- and ~94-kDa proteins
with cis-acting DNA elements. Fig. 12 depicts the model
envisaged for the heme-dependent interaction of the
proteins with the positive element. The ~26-kDa protein has already
been shown to stimulate transcription of the minigene construct
containing the positive element (pP450e179) in cell-free extracts (12).
In the present study, the ~65-kDa protein has been shown to stimulate
transcription of the pP450e179 DNA in cell-free extracts prepared from
CoCl2-treated rat liver nuclei in a heme dependent fashion.
It, therefore, appears likely that the two proteins are positive
regulators of the gene with binding sites on the positive element,
although the possible interaction between the two needs to be
studied.