Tài liệu Báo cáo Y học: Analyses of the CYP11B gene family in the guinea pig suggest the existence of a primordial CYP11B gene with aldosterone synthase activity docx

Analyses of the
CYP11B
gene family in the guinea pig suggest
the existence of a primordial
CYP11B
gene with aldosterone synthase
activity
Hannes E. Bu¨ low
1,
* and Rita Bernhardt
2
1
Max-Delbru
¨
ck-Centrum fu
¨
r Molekulare Medizin, Berlin-Buch, Germany;
2
Universita
¨
t des Saarlandes, FR Biochemie, Saarbru
¨
cken,
Germany
In this study we describe the isolation of three genes of the
CYP11B family of the guinea pig. CYP11B1 codes for the
previously described11b-hydroxylase[Bu
¨
low, H.E., Mo
¨
bius,

K., Ba
¨
hr, V. & Bernhardt, R. (1996) Biochem. Biophys. Res.
Commun. 221, 304–312] while CYP11B2 represents the
aldosterone synthase gene. As no expression for CYP11B3
was detected this gene might represent a pseudogene.
Transient transfection assays show higher substrate speci-
ficity for its proper substrate for CYP11B1 as compared to
CYP11B2, which could account for the zone-specific syn-
thesis of mineralocorticoids and glucocorticoids, respec-
tively. Thus, CYP11B2 displayed a fourfold higher ability to
perform 11b-hydroxylation of androstenedione than
CYP11B1, while this difference is diminished with the size of
the C17 substituent of the substrate. Furthermore, analyses
with the electron transfer protein adrenodoxin indicate dif-
ferential sensitivity of CYP11B1 and CYP11B2 as well as the
three hydroxylation steps catalysed by CYP11B2 to the
availability of reducing equivalents. Together, both mecha-
nisms point to novel protein intrinsic modalities to achieve
tissue-specific production of mineralocorticoids and gluco-
corticoids in the guinea pig. In addition, we conducted
phylogenetic analyses. These experiments suggest that a
common CYP11B ancestor gene that possessed both
11b-hydroxylase and aldosterone synthase activity under-
went a gene duplication event before or shortly after the
mammalian radiation with subsequent independent evolu-
tion of the system in different lines. Thus, a differential
mineralocorticoid and glucocorticoid synthesis might be an
exclusive achievement of mammals.
Keywords: guinea pig; 11b-hydroxylase; aldosterone synth-

ase; phylogeny.
Higher vertebrates regulate vital processes like volume/
electrolyte homeostasis and glucose/lipid metabolism by
means of steroid hormones, namely mineralocorticoids and
glucocorticoids. The biosynthesis of these steroids occurs
primarily in the adrenal cortex within morphologically and
functionally distinct zones. Accordingly, mineralocorticoids
are produced by the outer zona glomerulosa while gluco-
corticoids are formed in the two inner layers of the cortex,
the zonae fasciculata/reticularis. Originating from cholester-
ol they are synthesized by a number of consecutive
oxidations and dehydrogenations where all oxidative reac-
tions are catalysed by enzymes of the cytochrome P450
superfamily [2]. The first and rate-limiting step is the
conversion of cholesterol to pregnenolone by the mitochon-
drial cytochrome P450 side-chain cleavage enzyme (P450
scc
,
CYP11A1). Subsequently, pregnenolone is dehydroge-
nated and oxidized in position 17 and/or 21 to yield
11-deoxycortisol or 11-deoxycorticosterone, respectively.
Both compounds in turn are substrates for the cytochrome
P450 enzymes of the CYP11B subfamiliy, namely the
11b-hydroxylase (CYP11B1) and the aldosterone synthase
(CYP11B2). While CYP11B1 hydroxylates 11-deoxycorti-
sol in position 11 to give cortisol as the major glucocorti-
coid, the closely related aldosterone synthase forms
aldosterone as the major mineralocorticoid by means of
an 11b-hydroxylation and an 18-hydroxylation/oxidation of
11-deoxycorticosterone. Thus, the proteins of the CYP11B

subfamily catalysing the last biosynthetic steps are the key
enzymes for the synthesis of both mineralocorticoids and
glucocorticoids.
From molecular cloning of the corresponding genes and
analyses of the cDNAs it became obvious that the
encoded isoenzymes share a very high degree of similarity
ranging up to 95% on the amino acid level for human
CYP11B1 and CYP11B2 [3]. There are, however, a
number of significant species differences. For example,
humans [3], mice [4], rats [5], and hamsters [6,7], possess at
least two functionally different genes with the encoded
proteins exhibiting different enzymatic activities. While
one protein modifies the steroid entity predominantly in
position 11, the other one is able to hydroxylate and
oxidize position 18 as well. In contrast, cows [8], pigs [9],
sheep [10], and frogs [11] apparently possess only one type
of a bipotent enzyme that is capable of catalysing the
reactions at both positions 11 and 18. Nonetheless, the
production of mineralocorticoids and glucocorticoids is
strictly zone specific in all species. It is, however, unknown
Correspondence to R. Bernhardt, Universita
¨
t des Saarlandes, FR
Biochemie, PO Box 15 11 50, D-66041 Saarbru
¨
cken, Germany.
Fax: + 49 681302 4739, Tel.: + 49 681302 3005,
E-mail:
*Present address, Columbia University, College of Physicians &
Surgeons, New York, NY 10032, USA.

Note: a website is available at
/>(Received 12 April 2002, revised 11 June 2002, accepted 26 June 2002)
Eur. J. Biochem. 269, 3838–3846 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03076.x
which factors convey this specificity and how these similar
but distinct systems evolved.
To investigate the zone-specific synthesis of mineralocor-
ticoids and glucocorticoids and the evolution of the
hormonal system in more detail we chose the guinea pig
as a model. The guinea pig is an interesting species because
its taxonomical position remains controversial [12,13].
These features should provide new insight into the evolution
and function of the hormonal system. We first cloned the
genes of the CYP11B family of the guinea pig. The 11b-
hydroxylase of the guinea pig showed higher substrate
specificity than the aldosterone synthase. In addition, the
aldosterone synthase exhibited unique properties in that
18-hydroxylase activity was strongly dependent on the
presence of high levels of reducing equivalents whereas basic
levels were sufficient for high 11b-hydroxylase activity of
this enzyme. This suggests a new regulatory level in
aldosterone synthesis that together with the higher substrate
specificity of the 11b-hydroxylase could be crucial for the
tissue-specific synthesis of steroid hormones. Phylogenetic
analyses indicate a gene duplication event of a bipotent
CYP11B ancestor gene before the mammalian radiation
with subsequent distinct evolution in different clades. This
indicates that a differential glucocorticoid and mineralocor-
ticoid synthesis is an exclusive property of mammals.
EXPERIMENTAL PROCEDURES
General procedures

Molecular biology procedures were carried out according to
standard protocols [14] unless stated otherwise. Chemicals
and enzymes were purchased from the highest quality
sources commercially available.
Screening of a guinea pig genomic library
A total of 1 · 10
6
clones of a guinea pig genomic library
(Stratagene, #946110) were screened under low stringency
conditions as described for Southern blots using a guinea
pig CYP11B1 full-length probe (1618 bp XbaIfragment
of pHBL5 [1]. Positive clones were purified to homoge-
neity and analysed by Southern blotting using various
restriction endonucleases. Appropriate genomic fragments
were subcloned into pBluescript SK(–) (Stratagene) and
sequenced using gene-specific primers. Furthermore, to
sequence parts not represented by genomic phage clones
genomic fragments were amplified by PCR and sequenced
directly.
RNA preparation
Tissue was homogenized in 6
M
guandinium thiocyanate
and subsequently RNA was purified by centrifugation
through a CsCl gradient [14]. PolyA
+
RNA was isolated by
three rounds of affinity purification on oligodT cellulose
(Stratagene).
RNAse protection analyses

RNAse protection analyses were carried out using a
HybSpeed
TM
RPA Kit (Ambion) according to the
manufacturer’s recommendations. Briefly, specific
32
P
labelled RNA antisense transcripts (corresponding to
nucleotides 1491–1700 in the CYP11B1 cDNA [1] and
nucleotides 1511–1750 in the CYP11B2 cDNA; Fig. 2) were
hybridized with total RNA from different tissues. After
digestion of the reaction mixture with RNAse A/H
protected fragments were separated by PAGE and visual-
ized by autoradiography.
RACE
The cDNA for CYP11B2 of the guinea pig was amplified
and cloned using a MarathonÒ cDNA Amplification Kit
(Clontech) following the supplier’s recommendations. In
brief, after reverse transcription of 1 lg of polyA
+
RNA
and second-strand synthesis an adapter comprising the T7
promoter sequence combined with a NotIandaSmaIsite
was ligated to both ends of the cDNA pool. Using a
combination of a primer complementary to the adapter
(adapter primer: 5¢-CCATCCTAATACGACTCACTA
TAGGGC-3¢) and a gene-specific sense primer (5¢-GCCG
CTCGAGTTTGAGTTAGCCAGAAACTCC-3¢, XhoI
site underlined) or antisense primer (5¢-ATAC
GGGCCC

GACAGTGGTGTGCCTGGGAAC-3¢, Bsp120I site
underlined), respectively, a PCR reaction was carried out
with KlenTaq
TM
(Clontech) under the following conditions:
94 °C 2 min initial denaturation, 94 °C 45 s denaturation,
72 °C 1 min annealing (annealing temperature reduced at
1.4 °C per cycle), 72 °C 3 min polymerization; 10 cycles,
followed by 25 cycles at 94 °C45s,58°C1minand72°C
3 min with a final extension step for 8 min at 72 °C. The
5¢-RACE product was cloned directly into a TAÒ Cloning
vector pCR2.1 (Invitrogen) yielding pCR2.1/HG17 while
3¢-RACE products were inserted by using the XhoIand
NotI sites into pBluescript SK(–) (Stratagene) giving
pBSSK/3¢RACE HG17.
DNA sequencing
DNA sequencing was carried out using a Thermo Sequen-
ase
TM
Cycle Sequencing Kit (Amersham/USB) in combi-
nation with [a-
35
S]dCTP followed by autoradiography with
Hyperfilm
TM
MP (Amersham).
Southern blotting
Genomic DNA was digested with the appropriate enzymes,
extracted twice with phenol/chloroform and precipitated
using EtOH and sodium acetate. After extensive washing

the DNA was redissolved in Tris/EDTA, pH 8.0 and sep-
arated on a 1 · Tris/borate/EDTA, 0.9% agarose gel. After
capillary transfer to Hybond
TM
nylon membranes (Amer-
sham) nucleic acids were UV cross-linked (0.24 JÆcm
)2
).
Prehybridization was performed in 5 · NaCl/Cit, 5 · Den-
hardt’s, 0.5% SDS and 50 lgÆmL
)1
sonicated salmon sperm
DNA for 2 h at 65 °C. [a-
32
P]dCTP labelled DNA probes
( 1 · 10
6
c.p.m.ÆmL
)1
) were hybridized in the same solu-
tion for 16 h. For low stringency hybridization the blot was
washed twice at room temperature in 2 · NaCl/Cit, 0.1%
SDS for 10 min followed by two 30 min washes at 50 °Cin
1 · NaCl/Cit, 0.1% SDS. Autoradiography was carried out
with Hyperfilm
TM
MP (Amersham).
Ó FEBS 2002 Guinea pig CYP11B genes (Eur. J. Biochem. 269) 3839
Construction of expression plasmids
For the construction of pCMV/11B2, pRc/CMV was

digested with Bsp120I, trimmed with Pfu polymerase
(Stratagene) and subsequently digested with NotI. Likewise,
pCR2.1/HG17 was digested with SpeI, trimmed with Pfu
polymerase and digested with NotI to release a fragment
comprising the ORF of the guinea pig CYP11B2. This
fragment was ligated using NotI/blunt into the eukaryotic
expression vector.
Hydroxylation assays
COS-1 cells were maintained as described previously [15].
Transfections were carried out using LipofectAMINE
TM
(Gibco/BRL) according to the manufacturer’s recommen-
dations. One mL of transfection mix contained 2 lgofthe
respective expression construct together with 1 lg pBAdx4
(bovine adrenodoxin; gift of M. R. Waterman, Vanderbilt
University, Nashville, TN, USA) and 6 lL
LipofectAMINE
TM
unless stated otherwise. Twenty-four
h after transfection cells were incubated with appropriate
substrates for 48 h using [1,2-
3
H]cortisol, [
14
C]11-deoxy-
corticosterone or [1,2–
3
H]androstenedione, respectively, as
tracers. Media were extracted and analysed by high
performance TLC as described previously [16].

Phylogenetic analyses
Phylogenetic analyses were conducted using the
PHYLIP
package (Version 3.5c, 1993) [17].
The sequences have been submitted to GenBank under
the accession numbers AF191278, AF191279 (for
CYP11B1), AF191281, AF191280 (for CYP11B2), and
AF191282 (for CYP11B3).
RESULTS
In a previous study we isolated an 11b-hydroxylase of the
guinea pig [1] by screening an adrenal cDNA library with a
PCR amplified orthologous probe. Upon expression, the
isolated cDNA turned out to be a pure 11b-hydroxylase
with no detectable 18-hydroxylation activity suggesting the
existence of additional isoenzymes of the CYP11B subfam-
ily in the guinea pig. To investigate this notion, a Southern
blot was performed utilizing an exon-1-specific probe of
CYP11B1 under low stringency conditions and digesting the
genomic DNA with various restriction endonucleases that
did not cut within exon 1. The result (Fig. 1) strongly
suggested the existence of at least three different genes as
judged from the appearance of three bands if the DNA was,
e.g. digested with EcoRI/EcoRV, XbaI, or XbaI/HindIII.
Although guinea pigs had been sodium depleted to
stimulate the expression of a putative aldosterone synthase
as much as possible [18], repeated screening of the cDNA
library did not result in the identification of any cDNA
other than CYP11B1 (data not shown). Thus, we devised
another strategy for the identification of additional genes of
the CYP11B subfamily in the guinea pig. To this end, a

genomic library was screened under low stringency (see
Experimental procedures) utilizing a full-length guinea pig
CYP11B1 cDNA as a probe. As opposed to a cDNA
library, screening of a genomic library should yield clones in
relation to their abundance in the genome rather than their
relative abundance due to differential expression. Indeed,
this approach lead to the isolation of eight genomic clones
that were classified into three subgroups based on restriction
digests and hybridization experiments (data not shown).
One clone termed kHG13 turned out to represent the
CYP11B1 gene while kHG17 and kHG15 represented
closely related genes of the CYP11B family demonstrated
by similarities of > 75% at the nucleotide level. They were
tentatively named CYP11B2 and CYP11B3, respectively.
To clone the corresponding cDNAs, the RACE tech-
nique was used. PolyA
+
RNA was converted into a double-
stranded cDNA pool and adapters comprising the promoter
sequence of the T7 bacteriophage were ligated to both ends.
The sequences of the T7 promoter are extremely rare in
eukaryotic genomes and thus convey a high degree of
specificity in subsequent PCR reactions. Using a primer
combination of an adapter primer and gene-specific sense or
antisense primers, respectively, we were able to amplify two
overlapping fragments in case of kHG17. Upon sequencing
of these cDNA fragments the complete sequence of the
cDNA of CYP11B2 could be deduced. It comprised
2611 bp and an ORF of 1503 bp coding for a putative
mitochondrial preprotein of 501 amino acids with a

calculated molecular weight of 57.7 kDa (Fig. 2). After
Leu24 a cleavage site for the matrix-associated protease was
predicted resulting in a mature mitochondrial protein of
55 kDa. The deduced amino acid sequence showed 81%
similarity to the guinea pig CYP11B1 and 80% similarity to
the human CYP11B2, respectively (see below). The 3¢-UTR
comprised 1079 bp with a canonical polyadenylation site
16 bp upstream of the polyA tail with no indications for the
existence of alternative poly adenylation sites (Fig. 2).
We next investigated the expression of the CYP11B
genes. A Northern blot probed with a CYP11B2-specific
probe showed a single band of 2.9 kb (data not shown)
which is consistent with the length of the isolated cDNA for
CYP11B2 assuming a polyA tail of  200–300 adenine
residues. To see where the CYP11B genes were expressed
Fig. 1. Southern blot analyses with a CYP11B1 exon 1-specific probe.
Fifteen micrograms of guinea pig genomic DNA was digested with
the indicated endonucleases. After transfer, membranes were probed
under low stringency conditions with an exon 1-specific probe of
CYP11B1 (nucleotides 1–141; see Experimental procedures for
details). Sizes of fragments are indicated on the right.
3840 H. E. Bu
¨
low and R. Bernhardt (Eur. J. Biochem. 269) Ó FEBS 2002
and whether they played a role during postnatal develop-
ment we used a highly sensitive RNAse protection assay
with RNAs from different tissues and developmental stages.
As shown in Fig. 3, expression of both the 11b-hydroxylase
and the aldosterone synthase was exclusively in the adrenal
gland. Moreover, there was no difference in expression

between postnatal day 1 and the adult stages suggesting that
the genes were not differentially regulated during postnatal
development. With respect to kHG15 we were not able to
demonstrate expression of the gene in adult tissues using
RT/PCR with various gene-specific primer combinations
(data not shown). Thus, this clone might represent a
pseudogene of the CYP11B family or a gene that is not
expressed in adult tissues.
To compare the enzymatic activities of CYP11B2 and
CYP11B1, the cDNAs were cloned under the control of a
viral promoter and transiently transfected into COS-1 cells.
Transfected cells were incubated with different substrates
and the resulting metabolites were analysed using TLC. As
seen in Fig. 4A, CYP11B2 converted 11-deoxycorticoster-
one to corticosterone and both 18(OH)-corticosterone and
aldosterone. These results clearly demonstrate that
CYP11B2 is the aldosterone synthase of the guinea pig as
it is capable of modifying position 11 and 18 of the steroid
ring. In contrast, CYP11B1 produced only corticosterone
and traces of 18/19(OH)-deoxycorticosterone, confirming
earlier results [1]. Furthermore, CYP11B2 transfected cells
efficiently converted 11-deoxycortisol to cortisol and fur-
ther to 18(OH)-cortisol (Fig. 4B). As 11b(OH)-androsten-
edione is the major C19 steroid in the guinea pig, we also
used androstenedione as a substrate. Under the experi-
mental conditions large amounts of 11b(OH)-androstendi-
one were synthesized by CYP11B2 in comparison with
CYP11B1 (Fig. 4C). It is noteworthy, that CYP11B2
displayed a higher enzymatic activity than CYP11B1 based
on 11b-hydroxylase activity. These differences were highest

Fig. 2. Sequence of the CYP11B2 cDNA of the guinea pig. The nucleotide sequence and the deduced amino acid sequence are both shown. The ORF
(putative start and stop codon underlined) encodes a mitochondrial preprotein with a calculated molecular mass of 57.7 kDa. An arrowhead
indicates the presumptive cleavage site for the mitochondrial matrix associated protease. Numbers on the left denote amino acids, those on the right
indicate nucleotides. A canonical polyadylation site is shown boldface.
Ó FEBS 2002 Guinea pig CYP11B genes (Eur. J. Biochem. 269) 3841
for androstenedione (fourfold) and lowest for 11-deoxy-
cortisol (Fig. 4). This shows a higher substrate specifity of
CYP11B1 which could be due to differences in the active
centre and/or the entry channel. Moreover, it could be
important for tissue-specific synthesis of glucocorticoids
given the differences in expression levels of the two
enzymes.
We next asked whether other accessory proteins might
contribute to the zone-specific synthesis of steroid hor-
mones. A good candidate is adrenodoxin, an iron sulfur
containing electron donor protein that is required for the
function of mitochondrial cytochrome P450 proteins [19]
and has been shown to interact directly with the cyto-
chromes. To test its significance we carried out an experi-
ment where adrenodoxin was either cotransfected or
omitted. After transfection, cells were incubated with
11-deoxycorticosterone as a substrate. As shown in Fig. 5,
the omission of Adx leads to a sharp decrease in the activity
for the 11b-hydroxylase, CYP11B1. Intriguingly, however,
the 11b-hydroxylase activity of the aldosterone synthase
CYP11B2 was basically unaffected whereas the 18-hydroxy-
lation and oxidation potential were abrogated almost
completely. These results indicate clear structural differences
on the surface of these proteins involved either in glucocor-
ticoid or in mineralocorticoid biosynthesis despite a high

degree of similarity between the two isoenzymes. More
importantly, these results indicate a new level of regulation
for tissue-specific aldosterone synthesis depending on the
availability of reducing equivalents.
One intriguing question is how and when animals
developed a hormonal system that differentially regulated
the control of both electrolyte/volume homeostasis and
glucose metabolism. Knowing when and how differential
synthetic pathways for mineralocorticoids and glucocorti-
coids developed would lead to deeper understanding of
these important evolutionary processes. Because the guinea
pig’s taxonomical classification is controversial [12,13], this
species is extremely interesting in terms of vertebrate
evolution and might provide insight into some aspects of
the evolution of the hormonal system.
Fig. 4. Enzymatic acivities of CYP11B2. COS-1 cells were transfected
with pBAdx4 (bovine adrenodoxin) and the expression plasmid
pCMV5 [1] (CYP11B1), pRc/CYP11B2 (CYP11B2), or pRC/CMV
(mock), respectively. Twenty-four h after transfection cells were
incubatedfor48hwith5l
M
11-deoxycorticosterone (DOC) includ-
ing 4 nCiÆmL
)1
[
14
C]DOC (A), 5 l
M
androstendione including
0.5 lCiÆmL

)1
[
3
H]androstendione (B), or 2.5 l
M
11-deoxycortisol
including 0.5 lCiÆmL [
3
H]11-deoxycorticosterone (C). Subsequently,
steroids were extracted and separated by TLC [16]. In culture medium
incubated substrates served as an additional control (substrate).
Positions of cold standards are denoted on the left. On the right
percentage of total radioactivity or relative activity is given ± SD; data
are from at least two different experiments performed in triplicate.
Fig. 3. Tissue and age-specific RNAse protection assays. Different
amountsoftotalRNAfromdifferenttissuesandstagesasindicated
were hybridized in solution with CYP11B1 and CYP11B2-specific
probes. Both probes were chosen from the 3¢ untranslated regions of
the genes where sequence divergence was maximal between the two
isoenzymes. Following RNAse digestion the probes protected a 210
nucleotide fragment of CYP11B1 (corresponding to nucleotides 1491–
1700 [1]) or a 240 nucleotide fragment for CYP11B2 (corresponding to
nucleotide 1511–1750; Fig. 2), respectively. A control lane without
RNAse (–RNAse) shows the corresponding undigested riboprobes of
242 nucleotides (for CYP11B1) and 272 nucleotides (for CYP11B2). A
molecular size marker is given on the left. Different developmental
stages are denoted on the right: P1, postnatal day 1; adult.
3842 H. E. Bu
¨
low and R. Bernhardt (Eur. J. Biochem. 269) Ó FEBS 2002

To answer the question how these genes might have
evolved we determined the complete genomic structure of
both genes. As shown in Fig. 6, both genes exhibit the
typical organization of the family, characterized by nine
exons and eight introns supporting the idea of a gene
duplication event. It is, however, noteworthy that the exons
are grouped into three clusters comprising exons 1 and 2,
exons 3–5 and exons 6–9, respectively. These clusters are
separated by intron 2 and intron 5, which are not only
larger than any other intron but also show considerable
differences in sequence similarity. For example, an align-
ment [20] of intron sequences requires the introduction of 11
and 14 gaps for intron 2 and 5, respectively, as compared to
1 (intron 7) to 7 (intron 6) for the remaining introns. Finally,
the 3¢ UTRs of the guinea pig CYP11B genes show only
42% similarity whereas the 3¢UTR of the guinea pig
CYP11B2 gene shares up to 72% identity with the
corresponding human homologue indicative of a close
relationship of CYP11B2 sequences between species.
Comparing the protein sequences of the CYP11B family
(Table 1) we were surprised to find that both CYP11B1 and
CYP11B2 of the guinea pig are always more closely related
(or equal) to the CYP11B2 sequences of other species.
Moreover, the guinea pig CYP11B2 is always more similar
to CYP11B proteins of other species than CYP11B1 of the
guinea pig (compare lines A and B). Together these results
suggest that the CYP11B2 genes are the primordial genes
and that a common ancestor containing both enzymatic
activities was duplicated. The resulting two genes subse-
quently evolved to give both different regio-specificities and

differential regulatory circuits.
We next asked when the aforementioned gene duplication
event might have occurred. To this end we conducted
phylogenetic analyses with all known sequences of the
CYP11B family of proteins. Including the sequences of the
guinea pig with its highly controversial taxonomical posi-
tion into this highly homologous family of proteins could
possibly give new insights into both its taxonomical
classification and the evolutionary relationships within this
protein family. Also, these analyses might indicate when the
gene duplication event occurred that subsequently led to
isoenzymes harbouring different enzymatic activities like,
for example, those in humans or to an exclusively differen-
tial regulation like seen in cattle. Amino acid sequences of 16
CYP11B proteins were subjected to phylogenetic analyses
using two fundamentally different methods. The use of
various methods should provide an estimate of methodical
errors. On the one hand, two distance matrix methods, the
UPGMA (unweighted pairgroup method using arithmetic
mean) and the neighbor joining method (reviewed in [21]),
were used. The distance matrices for the calculation of
phylogenetic trees were produced with three different
algorithms for amino acid exchanges, namely the Dayhoff
model [22], Kimura’s model [23] and the categories model
developed by Felsenstein [24]. On the other hand, the
maximum parsimony method as a single character state
algorithm was used to evaluate the phylogenetic relation-
ships between these proteins. This approach assumes the
most probable phylogeny to be the one that requires the
fewest nucleotide exchanges [17]. The frog was used as an

outgroup in all applications and the reliability of a given
topology was assessed using the bootstrapping procedure
[25].
The results are depicted in Fig. 7. No matter which
algorithm was used, the guinea pig sequences were grouped
together with bootstrapping probabilities of at least 98%.
The maximum parsimony method placed the guinea pig
into one group with rodents requiring 1349 nucleotide
exchanges thus supporting monophyly of the rodents.
Within the rodents the branching consistently grouped the
orthologues of rat and mouse together demonstrating the
close relationship between these species. In contrast, both
distance matrix methods showed the guinea pig together
Fig. 5. Influence of bovine adrenodoxin on enzymatic activities. COS-1
cells were transfected with pBAdx4 (bovine adrenodoxin) and the
expression plasmids pCMV5 [1] (CYP11B1), pRc/CYP11B2
(CYP11B2), or pRC/CMV (mock), respectively. Twenty-four hours
after transfection, cells were incubated for 48 h with 5 l
M
11-deoxy-
corticosterone (DOC) including 4 nCiÆmL
)1
[
14
C]DOC. Metabolites
were analysed by TLC. Enzymatic activity is given as relative radio-
activity. White bars represent cotransfection with adrenodoxin (¼ 100)
andblackbarsrepresentnocotransfectionexpressedinrelationto
100 ± SD; n ¼ 9 for each data point.
Fig. 6. Genomic structure of CYP11B1 and CYP11B2 of the guinea pig.

Shown is the complete genomic structure of CYP11B1 and CYP11B2.
Exon and intron boundaries are indicated. Scale bar represents 1000
nucleotides.
Table 1. Pair-wise sequence similarities of CYP11B proteins. Similarities for the CYP11B proteins were determined pair-wise using the
PALIGN
program (PcGENE, Intelligenetics) for the proteins from human (HS), mouse (MM), rat (RN) hamster (MA), guinea pig (CP), sheep (OA), cow
(BT), pig (SS) and frog (RC).
HSB1HSB2MMB1MMB2RNB1RNB2MAB1MAB2CPB1CPB2OAB0BTB0SSB0 RCB0
CP B1 74% 75% 72% 76% 73% 76% 72% 74% – 81% 73% 73% 72% 58% A
CP B2 80% 80% 75% 80% 77% 79% 76% 79% 81% – 77% 77% 77% 58% B
Ó FEBS 2002 Guinea pig CYP11B genes (Eur. J. Biochem. 269) 3843
with artiodactyls and primates, i.e. supported the paraphyly
of the order rodentia. The bootstrapping probabilities were
however, comparatively low (Fig. 7D) using the neighbor
joining approach with 49, 52 or 68% for the categories
model, the Dayhoff model or Kimura’s model, respectively.
Using UPGMA the probabilities were slightly higher;
between 70 and 82%. Moreover, the hamster proteins were
now assigned to their rat and mouse paralogues. Interest-
ingly, Kimura’s model consistently produced the highest
bootstrapping probabilities for a given topology.
DISCUSSION
In this paper, we describe the isolation and characterization
of the CYP11B genes of the guinea pig. In an earlier study
[1] we isolated a single cDNA using an orthologous probe
that had been obtained using degenerated primers to screen
a guinea pig adrenal cDNA library. This cDNA proved to
code for the abundantly expressed 11b-hydroxylase,
CYP11B1, of the guinea pig which exhibited exclusive
11b-hydroxylase activity [1]. Although cloning strategies

using PCR-based approaches with degenerated primers
have often been successful this is sometimes hampered by
large differences in expression levels. Thus, we were unable
to isolate more than one cDNA of the CYP11B family of
the guinea pig either by PCR-based approaches or by
repeated screening of the library under low stringency
conditions, even under conditions in which sodium was
depleted to achieve induction of the aldosterone synthase
gene. The enzymatic activity of the cloned protein, however,
strongly suggested the existence of additional isoenzymes
with different enzymatic activities. To test this hypothesis,
we performed a Southern blot analysis using an exon 1-spe-
cific probe of CYP11B1. This experiment indicated the
existence of at least two additional genes of this family in the
guinea pig. To clone these genes we screened a genomic
library which should circumvent difficulties associated with
largely differing expression levels. This led to the isolation of
two additional genes, tentatively named CYP11B2 and
CYP11B3.ForCYP11B2,wewereabletoisolateacDNA
with a complete ORF. In contrast, no specific transcripts
could be detected for CYP11B3 in sensitive RT-PCR
experiments with RNA from adult animals. Thus, this gene
might be a pseudogene that evolved as a consequence of a
secondary gene duplication event (see below). Similar
observations have also been made in cows that have five
genes of the CYP11B family only two of which are functional
[8]. Alternatively, CYP11B3 might be a developmentally
regulated gene as in rats, in which it is expressed solely during
a few specific days of postnatal development [26].
However, no differential expression was observed for

either CYP11B1 or CYP11B2 during postnatal develop-
ment of the guinea pig using RNAse protection assays.
Instead, we saw exclusive adrenal expression of both genes.
This does, however, not rule out expression in other tissues
at lower levels. For example, expression of CYP11B genes in
the rat has been demonstrated in brain [27] and in the heart
[28] using very sensitive RT-PCR and in situ hybridization
techniques. The physiological significance of this low level
expression remains unclear.
To compare the catalytic activities of the guinea pig
CYP11B isoenzymes we cloned the cDNAs downstream of
a cytomegalovirus promoter to drive expression in COS-1
cells. This system has been proven suitable for the
characterization of enzymes of the steroidogenic pathway
[15]. These analyses demonstrated a potent 18-hydroxyla-
tion and 18-oxidation activity of CYP11B2 using various
substrates thus showing it to be the aldosterone synthase
of the guinea pig. It was interesting to note that CYP11B2
of the guinea pig had a considerably higher enzymatic
activity in terms of 11b-hydroxylated product formed
than the guinea pig CYP11B1. These results are in contrast
with findings in other species. For example, the human
CYP11B1 has a 20-fold higher activity towards 11-deoxy-
cortisol than CYP11B2 [29]. Moreover, the activity of the
guinea pig CYP11B2 compared to CYP11B1 could be
correlated to the size of the C17 substituent of the
substrate. Thus, the differences were most pronounced
with androstenedione and least with 11-deoxycortisol as a
substrate. These results might indicate steric hindrance in
the entry channel of the cytochrome P450 enzymes with

CYP11B1 being more selective for its proper substrate.
This would contribute to the zone-specific synthesis of
glucocorticoids given the vast differences in expression level
of the two genes.
In a second set of experiments, we investigated the
significance of adrenodoxin, an iron sulfur protein that is
essential for electron transfer from adrenodoxin reductase
to mitochondrial cytochrome P450 enzymes [19]. Adreno-
doxin has been shown to significantly increase enzymatic
activities of steroidogenic enzymes in transfection experi-
ments [15]. If this cotransfection was omitted, the two
guinea pig enzymes showed differential properties. The 11b-
hydroxylase activity of CYP11B1 was strongly reduced
(Fig. 5), whereas that of the aldosterone synthase was
basically unaffected, while the 18-hydroxlase and oxidase
activity was also greatly diminished. These differences can
Fig. 7. Phylogenetic analyses. Shown are the phylogenetic trees
obtained by the maximum parsimony method (A), the UPGMA
method (categories model) (B), and the neighbour joining method
(categories model) (C). Numbers represent bootstrapping probabilities
of 1000 replicates. The Table in D gives the bootstrapping probabilities
for monophyly and paraphyly of the order rodentia in detail. Pr,
primates; Ar, artiodactyls; My, myomorphs; Cp, guinea pig.
3844 H. E. Bu
¨
low and R. Bernhardt (Eur. J. Biochem. 269) Ó FEBS 2002
be explained by a lower binding affinity of CYP11B1 to
adrenodoxin when compared with CYP11B2. CYP11B1
has a tryptophan at position 366 while all other proteins of
the CYP11B family including CYP11B2 of the guinea pig

have a basic residue (arginine or lysine, respectively) at the
corresponding position. These two basic residues have been
shown in site-directed mutagenesis experiments to be of
significance to the electrostatic interaction of bovine
CYP11A1 (cytochrome P450
SCC
), a closely related mito-
chondrial protein with adrenodoxin [30]. Accordingly, the
11b-hydroxylase activity of CYP11B2 was hardly affected
by omission of cotransfected adrenodoxin. However, the
subsequent enzymatic reactions involving the 18-hydroxy-
lation and oxidation were severely impaired. This could
indicate altered binding affinities for adrenodoxin after
the 11b-hydroxylation. In this regard it is interesting to
note that in vitro experiments with purified bovine
CYP11B0 indicate a conformational change of the protein
due to rearrangement of the substrate after the first
hydroxylation step [31]. This might account for an altered
binding site for adrenodoxin or modified binding affini-
ties. Also, experiments in a reconstituted system with
bovine CYP11B0 using mutant forms of adrenodoxin that
have increased electron transfer capabilities showed a shift
in the spectrum of products formed towards compounds
modified at position 18 [32]. Interestingly, studies with the
microsomal cytochrome P450 enzyme 17a-hydroxylase/
17,20-lyase demonstrated a dependence of the more electron
consuming lyase reaction on the presence of high concen-
trations of the electron donor protein [33]. Taken together,
our results with the guinea pig suggest a new regulatory level
of aldosterone synthesis by the availability of reducing

equivalents. In the light of results with the bovine and
human enzymes [32] this might be a more commonly used
mechanism which could be crucial for the zone-specific
biosynthesis of mineralocorticoids. In this respect it would
be interesting to see whether expression of adrenodoxin is
differentially regulated.
To investigate the evolution of the CYP11B genes, we
compared protein sequences of the CYP11B family. These
analyses showed that CYP11B2 proteins are more closely
related to each other than CYP11B1 proteins across species.
Furthermore, the 3¢ UTR of the guinea pig CYP11B2
shows considerable similarity to its human counterpart
while exhibiting only low resemblance to its paralogue in the
guinea pig. On a functional level the CYP11B1 proteins also
show greater differences across species. For example, the
hamster CYP11B1 produces  50% of 19-hydroxylated
product besides the 11b-hydroxylated steroid [7]. Moreover,
as mentioned above, human CYP11B1 is a very potent 11b-
hydroxylase as compared with CYP11B2 while in the
guinea pig this situation is reversed. These findings suggest
that CYP11B2 genes, i.e. those encoding bipotent enzymes,
are the primordial genes that underwent a subsequent
duplication and, as a consequence, the CYP11B1 genes
evolved independently in different species within the limits
of functional constraints.
To see how this evolution might have occurred, we
determined the complete genomic structure of the two
functional guinea pig genes. Interestingly, we observed the
greatest differences between these closely related genes within
intron 2 and intron 5. This could indicate frequent recom-

bination between the genes for which close association has
been shown in mice [4] and humans [34]. Indeed, in humans
an unequal crossover between CYP11B genes, fusing the
CYP11B2 gene undercontrol of the CYP11B1 promoter and
vice versa, has been demonstrated to cause glucocorticoid
remediable aldosteronism, an autosomal dominant disorder
leading to severe hypertension [34], or congenital adrenal
hyperplasia [35]. In this context it is interesting to note that
the crucial determinants for regio-specificity have been
shown to reside in exon 5 [29]. Moreover, analysis of
breakpoints in patients suffering from glucocorticoid reme-
diable aldosteronism indicate that important regulatory
elements are contained within intron 2 [36]. Thus, a scenario
is conceivable where recombination between the two genes in
intron 2 and intron 5 eventually lead to both distinct regio-
specificities and/or differential regulation.
We next sought to determine when this gene duplication
event might have occurred. To this end, we conducted
phylogenetic analyses with 16 sequences of the CYP11B
family of proteins. To assess possible methodical problems
we used different algorithms, namely the maximum
parsimony and two distance matrix methods. The maxi-
mum parsimony method consistently grouped the guinea
pig with a bootstrapping probability of 98% into one clade
with rodents thus favouring monophyly of the order
rodentia. This is in contrast with the findings of Graur and
colleagues [12] who postulated paraphyly of the order. Our
results are however, in accordance with the results of
Hasegawa et al. [13] who questioned the paraphyly of the
order rodentia using the same data as Graur. In contrast,

the distance matrix algorithms again placed the guinea pig
together with artiodactyls and primates, thus supporting
paraphyly albeit only with bootstrapping probabilities
between 49 and 82%. Interestingly, the highest values were
obtained when Kimura’s model [23] was used to calculate
the matrices. This might reflect the fact that this model
assumes conservative and nonconservative changes to be
equally likely which could lead to an overestimation of
conservative changes.
According to Frye et al. [37] the ambiguity of the
phylogeny as seen in our analyses with respect to the guinea
pig can be interpreted as insufficient methodology. For
example, the data and the algorithms might not be adequate
to assign a statistically unambiguous topology if the radia-
tion of species has occurred in a sufficiently small time frame.
Thus, the guinea pig presumably branched off at a very early
time point within the mammalian radiation irrespective of
the branching order. Intriguingly, however, the guinea pig
possesses already two functional CYP11B genes as have all
other mammals investigated so far. Theoreticallyit is possible
that the gene duplication occurred in all lines independently.
It seems, however, much more likely that the gene was
duplicated in an ancestor mammal. In this respect, it is
interesting that analyses in the frog as an amphibian gave no
indications of two CYP11B genes [11]. Thus, the differential
mineralocorticoid and glucocorticoid synthesis is presum-
ably an exclusive property of mammals.
ACKNOWLEDGEMENTS
We thank K. Denner, B. Bo
¨

ttner and members of the Bernhardt lab
for helpful discussions and advice. We also thank W. Oelkers and
V. Ba
¨
hr for guinea pig adrenal tissues. This work was supported by the
Deutsche Forschungsgemeinschaft Grant DFG Be 13436-1.
Ó FEBS 2002 Guinea pig CYP11B genes (Eur. J. Biochem. 269) 3845
REFERENCES
1. Bu
¨
low, H.E., Mo
¨
bius, K., Ba
¨
hr,V.&Bernhardt,R.(1996)
Molecular cloning and functional expression of the cytochrome
P450, 11B-hydroxylase of the guinea pig. Biochem. Biophys. Res.
Commun. 221, 304–312.
2. Nelson, D.R., Koymans, L., Kamataki, T., Stegeman, J.J.,
Feyereisen, R., Waxman, D.J., Waterman, M.R., Gotoh, O.,
Coon, M.J., Estabrook, R.W., Gunsalus, I.C. & Nebert, D.W.
(1996) P450 superfamily: update on new sequences, gene mapping,
accessionnumbersandnomenclature.Pharmacogenetics 6, 1–42.
3. Mornet, E., Dupont, J., Vitek, A. & White, P.C. (1989) Char-
acterization of two genes encoding human steroid 11 beta-
hydroxylase (P-450 (11) beta). J. Biol. Chem. 264, 20961–20967.
4. Domalik, L.J., Chaplin, D.D., Kirkman, M.S., Wu, R.C., Liu,
W.W., Howard, T.A., Seldin, M.F. & Parker, K.L. (1991) Dif-
ferent isozymes of mouse 11 beta-hydroxylase produce miner-
alocorticoids and glucocorticoids. Mol. Endocrinol. 5, 1853–1861.

5. Mukai,K.,Imai,M.,Shimada,H.&Ishimura,Y.(1993)Isolation
and characterization of rat CYP11B genes involved in late steps of
mineralo- and glucocorticoid syntheses. J. Biol. Chem. 268, 9130–
9137.
6. LeHoux, J.G., Mason, J.I., Bernard, H., Ducharme, L., LeHoux,
J., Veronneau, S. & Lefebvre, A. (1994) The presence of two
cytochrome P450 aldosterone synthase mRNAs in the hamster
adrenal. J. Steroid Biochem. Mol. Biol. 49, 131–137.
7. Veronneau, S., Bernard, H., Cloutier, M., Courtemanche, J.,
Ducharme, L., Lefebvre, A., Mason, J.I. & LeHoux, J.G. (1996)
The hamster adrenal cytochrome P450C11 has equipotent 11beta-
hydroxylase and 19-hydroxylase activities, but no aldosterone
synthase activity. J. Steroid Biochem. Mol. Biol. 57, 125–139.
8. Kirita, S., Hashimoto, T., Kitajima, M., Honda, S., Morohashi,
K. & Omura, T. (1990) Structural analysis of multiple bovine
P-450 (11 beta) genes and their promoter activities. J. Biochem.
(Tokyo) 108, 1030–1041.
9. Okamoto, M., Nonaka, Y., Ohta, M., Takemori, H., Halder,
S.K., Wang, Z.N., Sun, T., Hatano, O., Takakusu, A. & Mur-
akami, T. (1995) Cytochrome P450 (11 beta): structure-function
relationship of the enzyme and its involvement in blood pressure
regulation. J. Steroid Biochem. Mol. Biol. 53, 89–94.
10. Boon, W.C., Roche, P.J., Butkus, A., McDougall, J.G., Jeyasee-
lan, K. & Coghlan, J.P. (1997) Functional and expression analysis
of ovine steroid 11 beta-hydroxylase (cytochrome P450 11 beta).
Endocr. Res. 23, 325–347.
11. Nonaka, Y., Takemori, H., Halder, S.K., Sun, T., Ohta, M.,
Hatano, O., Takakusu, A. & Okamoto, M. (1995) Frog cyto-
chrome P-450 (11 beta,aldo), a single enzyme involved in the final
steps of glucocorticoid and mineralocorticoid biosynthesis. Eur. J.

Biochem. 229, 249–256.
12. Graur, D., Hide, W.A. & Li, W.H. (1991) Is the guinea-pig a
rodent? [see comments]. Nature 351, 649–652.
13. Hasegawa, M., Cao, Y., Adachi, J. & Yano, T. (1992) Rodent
polyphyly? Nature 355,595.
14. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular
Cloning: A Laboratory Manual,Vol.3,2ndedn.ColdSpring
Harbor Laboratory Press, Cold Spring Harbor, New York.
15. Zuber, M.X., Simpson, E.R. & Waterman, M.R. (1986) Expres-
sion of bovine 17 alpha-hydroxylase cytochrome P-450 cDNA in
nonsteroidogenic (COS-1) cells. Science 234, 1258–1261.
16. Bu
¨
low, H.E., Mo
¨
bius, K., Ba
¨
hr,V.&Bernhardt,R.(1996)
Functional expression of the guinea pig 11b-hydroxylase in COS-1
cells. Endocr Res. 22, 479–484.
17. Felsenstein, J. (1993) PHYLIP (Phylogeny Inference Package),
Version 3.5c. Department of Genetics. University of Washington,
Seattle.
18. Holland, O.B. & Carr, B. (1993) Modulation of aldosterone syn-
thase messenger ribonucleic acid levels by dietary sodium and
potassium and by adrenocorticotropin. Endocrinology 132, 2666–
2673.
19. Vickery, L.E. (1997) Molecular recognition and electron transfer
in mitochondrial steroid hydroxylase systems. Steroids 62,124–
127.

20. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J.
(1990) Basic local alignment search tool. J. Mol Biol. 215, 403–410.
21. Saitou, N. (1996) Reconstruction of gene trees from sequence
data. Methods Enzymol. 266, 427–449.
22. Dayhoff, M.O., Schwartz, R.M. & Orcutt, B.C. (1978) Atlas of
Protein Sequence and Structure. 5. Suppl. 3. (Dahoff, M.O., ed.),
National Biomedical Research Foundation, Washington DC.
23. Kimura, M. (1983) The Neutral Theory of Evolution.Cambridge
University Press, Cambridge, USA
24. Felsenstein, J. (1996) Inferring phylogenies from protein sequences
by parsimony, distance, and likelihood methods. Methods Enzy-
mol. 266, 418–427.
25. Felsenstein, J. (1985) Confidence limits on phylogenies. Evolution
39, 783–791.
26. Mellon, S.H., Bair, S.R. & Monis, H. (1995) P450c11B3 mRNA,
transcribed from a third P450c11 gene, is expressed in a tissue-
specific, developmentally, and hormonally regulated fashion in the
rodent adrenal and encodes a protein with both 11-hydroxylase
and 18-hydroxylase activities. J. Biol. Chem 270, 1643–1649.
27. Erdmann, B., Gerst, H., Lippoldt, A., Bu
¨
low, H., Ganten, D.,
Fuxe, K. & Bernhardt, R. (1996) Expression of cytochrome
P45011B1 mRNA in the brain of normal and hypertensive
transgenic rats. Brain Res. 733, 73–82.
28. Silvestre, J.S., Robert, V., Heymes, C., Aupetit-Faisant, B.,
Mouas, C., Moalic, J.M., Swynghedauw, B. & Delcayre, C. (1998)
Myocardial production of aldosterone and corticosterone in the
rat. Physiological regulation. J. Biol. Chem 273, 4883–4891.
29. Bo

¨
ttner, B., Schrauber, H. & Bernhardt, R. (1996) Engineering a
mineralocorticoid- to a glucocorticoid-synthesizing cytochrome
P450. J. Biol. Chem. 271, 8028–8033.
30. Wada, A. & Waterman, M.R. (1992) Identification by site-direc-
ted mutagenesis of two lysine residues in cholesterol side chain
cleavage cytochrome P450 that are essential for adrenodoxin
binding. J. Biol. Chem 267, 22877–22882.
31.Delorme,C.,Piffeteau,A.,Viger,A.&Marquet,A.(1995)
Inhibition of bovine cytochrome P-450 (11 beta) by
18-unsaturated progesterone derivatives. Eur. J. Biochem. 232,
247–256.
32. Cao, P.R. & Bernhardt, R. (1999) Modulation of aldosterone
biosynthesis by adrenodoxin mutants with different electron
transport efficiencies. Eur. J. Biochem. 265, 152–159.
33. Yanagibashi, K. & Hall, P.F. (1986) Role of electron transport
in the regulation of the lyase activity of C21 side-chain cleavage
P-450 from porcine adrenal and testicular microsomes. J. Biol.
Chem. 261, 8429–8433.
34. Lifton, R.P., Dluhy, R.G., Powers, M., Rich, G.M., Gutkin, M.,
Fallo, F., Gill, J.R. Jr, Feld, L., Ganguly, A., Laidlaw, J.C.& et al.
(1992) Hereditary hypertension caused by chimaeric gene dupli-
cations and ectopic expression of aldosterone synthase. Nature
Gene.t 2, 66–74.
35. Hampf, M., Dao, N.T., Hoan, N.T. & Bernhardt, R. (2001)
Unequal crossing-over between aldosterone synthase and 11beta-
hydroxylase genes causes congenital adrenal hyperplasia. J. Clin.
Endocrinol. Metab. 86, 4445–4452.
36. Pascoe, L. & Curnow, K.M. (1995) Genetic recombination as a
cause of inherited disorders of aldosterone and cortisol biosyn-

thesis and a contributor to genetic variation in blood pressure.
Steroids 60, 22–27.
37. Frye, M.S. & Hedges, S.B. (1995) Monophyly of the order
Rodentia inferred from mitochondrial DNA sequences of the
genes for 12S rRNA, 16S rRNA, and tRNA-valine. Mol. Biol.
Evol 12, 168–176.
3846 H. E. Bu
¨
low and R. Bernhardt (Eur. J. Biochem. 269) Ó FEBS 2002