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: ritabern@mx.uni-saarland.de *Present address, Columbia University, College of Physicians & Surgeons, New York, NY 10032, USA. Note: a website is available at http://www.uni-saarland.de/fak8/bernhardt. (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 . Analyses of the CYP11B gene family in the guinea pig suggest the existence of a primordial CYP11B gene with aldosterone synthase activity Hannes. 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.