Pathways and genes involved in steroid hormone metabolism in male pigs: A review and update

Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

Contents lists available at ScienceDirect

Journal of Steroid Biochemistry and Molecular Biology journal homepage: www.elsevier.com/locate/jsbmb

Pathways and genes involved in steroid hormone metabolism in male pigs: A review and update Annie Robic a,b,∗ , Thomas Faraut a,b , Armelle Prunier c,d a

UMR444, Génétique Cellulaire, INRA, CS 52627, 31326 Castanet Tolosan, France UMR444, Génétique Cellulaire, Université de Toulouse, INP, ENVT, 31076 Toulouse, France UMR1348-PEGASE, INRA, 35590 Saint-Gilles, France d UMR1348-PEGASE, Agrocampus Ouest, 35000 Rennes, France b c

a r t i c l e

i n f o

Article history: Received 15 July 2013 Received in revised form 19 September 2013 Accepted 4 November 2013 Keywords: Steroidogenesis Boar Hormones 5␣-Reduction CYP11 AKR1C

a b s t r a c t This paper reviews state-of-the-art knowledge on steroid biosynthesis pathways in the pig and provides an updated characterization of the porcine genes involved in these pathways with particular focus on androgens, estrogens, and 16-androstenes. At least 21 different enzymes appear to be involved in these pathways in porcine tissues together with at least five cofactors. Until now, data on several porcine genes were scarce or confusing. We characterized the complete genomic and transcript sequences of the single porcine CYP11B gene. We analyzed the porcine AKR1 gene cluster and identified four AKR1C, one AKR1C like genes and one AKR1E2 gene. We provide evidence that porcine AKR1C genes are not orthologous to human AKR1C. A new nomenclature is thus needed for this gene family in the pig. Thirty-two genes are now described: transcript (30 + 2 characterized in this study) and genomic (complete: 18 + 1 and partial: 12 + 1) sequences are identified. However, despite increasing knowledge on steroid metabolism in the pig, there is still no explanation of why porcine testes can produce androstenone and epiandrosterone, but not dihydrotestosterone (DHT), which is also a reduced steroid. Crown Copyright © 2013 Published by Elsevier Ltd. All rights reserved.

Abbreviations: aa, amino acid; S, sulphate; 5␣-R, 5-␣ reductase; Cytb5, cytochrome b5; Cytb5-red, cytochrome b5 reductase; HSD, hydroxysteroid dehydrogenase; P450scc, cytochrome P450 side chain cleavage encoded by porcine CYP11A1 gene; P450aro, P450 aromatase encoded by one of the three porcine CYP19A genes; P450c11, enzyme encoded by porcine CYP11B gene; P450c17, enzyme encoded by porcine CYP17A1 gene; P450c21, 21 steroid hydroxylase enzyme encoded by porcine CYP21 gene; StAR, steroidogenic acute regulatory encoded by porcine STAR gene; Pregnenolone, 5-pregnen-3␤-ol-20-one; 17OH-pregnenolone, 17-hydroxy pregnenolone; Progesterone, 4-pregnen-3,20-dione; 17OH- progesterone, 17-hydroxy progesterone; 20-OH-progesterone, 20␣progesterone or 4-pregnen-20-␣-ol-3-one; DOC, 11deoxycorticosterone or 21-hydroxyprogesterone (21-hydroxy-4-pregnene-3,20-dione); 4-AD, androstenedione or 4-Androstene-3,17-dione; 5-ADiol, androstenediol or 5-Androstene-3,17-diol; AD, androstanedione or 5␣-androstan-3,17-dione; Androstanediol, 5␣-androstan-3,17-diol; 11-OH-4-AD, 11␤ hydroxy-androstenedione or 11␤-hydroxyandrost-4-ene3,17-dione; 19-OH-4-AD, 19␤ hydroxy-androstenedione or 19␤ hydroxyandrost-4-ene-3,17-dione; DHEA, dehydroepiandrosterone or 3␤-hydroxyandrost-5-en-17-one; EpiA, epiandrosterone or 3␤-hydroxy-5␣-androstan-17-one; Androstenone, 4-androstene -3-one; Androstadienone,  4,16-androstadien-3-one; Androstadienol,  4,16androstadien-3-ol; Androsterone, 3␣-hydroxy-5␣-androstan-17-one; Adrenosterone, androst-4-ene-3,11,17-trione or 11-oxoandrostenedione; Testosterone, 17␤-hydroxy5alpha-androst-1-en-3-one; DHT, dihydrotestosterone or 17␤-hydroxy-5␣-androstan-3-one; 11-OH-Testo, 11␤-hydroxy testosterone or 11␤,17␤-dihydroxy-4-androsten3-one; 19-OH-Testo, 19␤-hydroxy testosterone or 17␤,19-dihydroxyandrost-4-en-3-one; 19-norTesto, 19-nortestosterone (or nandrolone) or 17␤-hydroxyestra-4-en-3-one; 11-K-Testo, 11-ketotestosterone or 17-Hydroxyandrost-4-ene-3,11-dione; 11-K-DHT, 11-keto dihydrotestosterone; 11-OH-DHT, 11␤-hydroxy dihydrotestosterone; Estrone, 3-hydroxyestra-1,3,5(10)-triene-17-one; Estradiol, 17␤-estra-1,3,5(10)-triene-3,17-diol. ∗ Corresponding author at: UMR444, Génétique Cellulaire, INRA, CS 52627, 31326 Castanet Tolosan, France. Tel.: +33 561285121; fax: +33 561285308. E-mail addresses: [email protected] (A. Robic), [email protected] (T. Faraut), [email protected] (A. Prunier). 0960-0760/$ – see front matter. Crown Copyright © 2013 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jsbmb.2013.11.001

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

1. Introduction Even though it is commonly accepted and reported that the synthesis of androstenone is the major feature [1–3], testicular production of steroids has many other remarkable features in pigs [4]. The boar testis produces high amounts of unconjugated and conjugated estrogens as well as 16-unsaturated steroids [2,5,6]. Their production, as indicated by their plasma levels, increases regularly during sexual maturation in parallel to testosterone [7,8]. Active steroidogenesis is not limited to pubertal and mature animals. During the first month after birth, the peak in steroid production reaches similar levels to those observed in mature animals [9–11]. This pattern of steroids parallels that of the development of Leydig cells in the testes [11–13]. After some weeks of relatively low production, steroid production again increases to reach maximum values around puberty, i.e. between five and eight months of age. Circulating steroids with this typical pattern of production are testosterone, estrone, estradiol, epiandrosterone (3␤ androsterone = EpiA), DHEA (dehydroepiandrosterone), androstenone and 19-nortestosterone (19-norTesto). Estradiol secreted by the testes plays an active role in the control of male sexual behavior [14]. Large quantities of estrogens are also excreted in boar semen and may influence female reproductive function by stimulating uterine contractions and the time of ovulation [14–16]. No circulating EpiA or 19-norTesto is observed in men. Very limited knowledge is available on the role of these two androgens in the physiology of the boar apart from the fact that these circulating compounds are produced by the pig testis [17,18]. Among 16-unsaturated steroids (16-androstene steroids), androstenone (5 ␣-androst-16-en-3-one) is of particular importance in the pig. It is excreted in saliva and acts as a pheromone that stimulates sexual behavior in females during estrus [19]. Androstenone is stored in the fat tissue because of its lipophilic properties. A high concentration of androstenone produces a pronounced urine-like flavor in meat, also known as boar taint, which is disliked by consumers and hence poses a problem of meat quality [20]. This is the main reason for the castration of male pigs reared for meat production. This practice is very widespread in European countries [21] but generates severe physiological and behavioral signs of pain [22]. Therefore, surgical castration in male pigs is highly debated in Europe and, finding solutions to rear entire males without boar taint is of primary interest. A better understanding of the biological mechanisms involved in the production of androstenone should help to address this problem. Taking into account the particularities of steroid production and their possible consequences for meat quality, our study focused on steroid synthesis in the boar testis. Our primary objective was to describe all possible steroid synthesis pathways in the pig, to update the pathways originally described in 1986 [1] with particular focus on 16-androstenes and a comparative analysis with human. Our second objective was to complete the characterization of porcine genes involved in steroid synthesis, especially genes CYP11B and AKR1.

2. Materials and methods Information on human genes was obtained from Ensembl (human release 72; http://www.ensembl.org/) and NCBI (Nov. 2012, human annotation release 104; http://www.ncbi.nlm.nih. gov/). The human reference transcripts were used in a BLAST procedure against the pig databases RefSeq RNA or/and NonRefSeq RNA or/and EST at NCBI to characterize the porcine transcripts. To compile all the transcripts documented in porcine databases, a BLASTN procedure was applied using each reference transcript listed in Table 1 against EST and transcript

45

Table 1 Available data on porcine genes involved in steroid pathways. Gene

Main transcript

Genomic sequences

CYP11A1

NM 214427 (X13768)

CYP11B CYP17

D38590 (1) KF314683 NM 214428 (M63507) (2) AK343431 NM 214429

NW 003537258 & NW 003538139 KF314688 NW 003536297 (3)

CYP19A1 CYP19A2 CYP19A3 CYP21

HSD17B2

NM 214430 NM 214431 NM 214433 (M83939) (2) AK343420 NM 001004049 (AF232699) NM 214248 (AF414424) NM 213913 (AF374414) NM 001128472 (EF581989) (2) EU429459 NM 001167649 (AB529535)

HSD17B3 HSD17B4 HSD17B7 HSD17B12 SRD5A1 SRD5A2

NM 001244790 (AK238558) NM 214306 (X78201) NM 001185137 (AB529536) XM 003353892 (5) AK396540 XM 003134156 (AK350108) NM 213988 (AF008440)

STAR

NM 213755 (AY800265)

AKR1C-pig1

DQ474065 (NM 001044569) (6) AK391133 KF314684 6 (7) NM 001044570 (DQ474066)(8) NM 001044618 (DQ474067)(8) NM 001038626 (DQ474068)(8) NM 001044568 (DQ474064)(8)

HSD3B1 HSD11B1 HSD11B2 HSD17B1

AKR1CLP AKR1C-pig3 AKR1C-pig4 AKR1C-pig6 AKR1E2 CYB5A CYB5B CYB5R1 CYB5R3 FDX1 FDXR POR

NM 001001770 (AF016388) NM 001159592 (AY609739) NM 001243918 (AK231742) XM 003125982 (AK392329) NM 214065 (2) AY610208 + AK343486 (9) NM 001244727 (AK236711) NM 001129959 (L33893)

NW 003538376 & NW 003540886 (4) NW 003609189 (SSC1) NW 003610615 (SSC7) (3) NW NW NW NW

003534677 (SSC4) (3) 003539307 003610403 (SSC6) (3) 003611499 (SSC12) (3)

NW NW NW NW NW NW NW NW NW NW

003610375 (SSC6) & 003537949 003611244 (SSC10) 003534350 (SSC2) 003534660 (SSC4) (3) 003609538 (SSC2) 003301585 (SSC16) (3) 003537808 & 003609920 (SSC3) 003612200 (SSC15) (3)

CU972427 (SSC10) (3) CU972427 (SSC10) (3) CU972427 (SSC10) (3) CU972427 (SSC10) (3) CU972427 (SSC10) (3) CU972427 (SSC10) (3) NW NW NW NW NW

003609242 (SSC1) (3) 003534899 (SSC6) (3) 003611243 (SSC10) (3) 003610195 (SSC5) (3) 003535539 (SSC9)

NW 003611480 (SSC12) (3) NW 003538795

When the transcript of reference proposed by the NCBI is not correct, we proposed another sequence and we have striped its name. (1) This sequence is not correct: see Section 3.2.2.2. (2) The sequence proposed as the reference transcript does not contain 5 NC and/or 3 NC sequences. (3) This gene was fully sequenced. (4) Redundant sequences. (5) The genomic sequence of this gene was not complete and the proposed annotation was incorrect; as a result, the predicted transcript included several errors. (6) This sequence is not correct: see Section 3.2.3.4. (7) The sequence of this non-coding transcript was probably incomplete in 5 and in 3 . (8) See Section 3.2.3.4. (9) Assembly of the two transcripts.

databases for porcine species (update 2012-10-01). Each possible transcript was compared to the reference transcript by analyzing BLAST results using a dedicated python script to detect new transcripts. The corresponding porcine genomic sequences were retrieved from the pig database using the NCBI nucleotide megablast search tool (genome reference Sus scrofa 10.2). Each new transcript was examined to determine the exact exon/intron structure. To identify potential exons in a genomic sequence or to determine the exon/intron structure of a gene, the transcript sequence was projected on the genomic sequence using (1) Sim4 (http://pbil.univ-lyon1.fr/members/duret/cours/inserm210604/ exercise4/sim4.html) and (2) GeneSeqer (http://www.plantgdb. org/cgi-bin/GeneSeqer/index.cgi). Annotations of the alternative splicing pattern of human genes were obtained through the

46

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

ASPicDB database designed to provide access to the functional annotation of predicted isoforms (http://t.caspur.it/ASPicDB/; update: 2012-01-31) [23].cDNA preparations and sequences of the PCR fragments were performed as described previously [24]. We used lalign (http://www.ch.embnet.org/software/LALIGN form.html) to determine percentage identity between two sequences (default options were used: local alignment, Blosum50, opening gap penalty: −14, extending gap penalty: −4). Protein sequences were analyzed using InterProScan web tools http://www. ebi.ac.uk/Tools/InterProScan/index.html. The PAML package and, in particular, the codeml module were used to determine the dS and dN of AKR1C genes. Phylogenetic analyses were performed with SeaView [25]. In SeaView, we used Muscle to perform the multiple sequence alignment. We chose to compute phylogenetic trees using a distance method with the BioNJ algorithm and the evolutionary distance provided by the Kimura two-parameter model. A total of 100 bootstraps were performed to estimate branch support. We used Genomicus to identify the ancestral gene in a multigenic family [26]. This tool enables a gene of interest to be displayed in its genomic context in parallel to the genomic context of all corresponding orthologous and paralogous copies in all the other sequenced metazoan genomes. Genomicus also stores and displays the predicted ancestral genome structure in all the ancestral species within the phylogenetic range of interest. Genomicus is available at http://www.dyogen.ens.fr/genomicus-70.01/cgi-bin/search.pl.

3. Results and discussion 3.1. Recapitulation of steroid synthesis pathways in steroidogenic tissues All steroid hormones are synthesized from cholesterol and consequently have closely related structures. Cholesterol possesses 27 carbons (C27) arranged in four rings, and a 6-carbon side chain. The first step in steroidogenesis consists in the cleavage of the side chain to release pregnenolone (C21). Pregnenolone can be converted either into progesterone or 17␣-hydroxyprogesterone (17-OH-progesterone), which are also C21 steroids. Both molecules are precursors for 16-androstene steroids (C19) and for the three major groups of steroid hormones, corticoids (C21), androgens (C19), and estrogens (C18), which are common to all mammals. The type of steroids produced by a particular gland is determined by the transcriptional activities of the genes that code for steroidogenic enzymes and by the availability of their substrates or cofactors [27]. All potential pathways concerning steroid biosynthesis in the pig are depicted in Fig. 1. This figure is based on the pathways proposed by Brooks and Pearson in 1986 in the pig [1], on the list of steroids that have been demonstrated to be secreted by the pig testis [4], and augmented by pathways present in humans [27–30]. Pathways including 19␤-hydroxy testosterone (19-OH-Testo) and 19␤-hydroxy androstenedione (19-OH-4-AD) were reported by Brooks and Pearson [1] in the pig and these steroids were observed in pig testes [31]. These products were probably intermediate products for the synthesis of estrogens or 19-norTesto [4,18,32]. Two pathways proposed by Brooks and Pearson [1] are assumed to transform pregnenolone into 16-androstene steroids with androstadienol/androstadienone and progesterone/11deoxycorticosterone/androstadienone as intermediate products. The conversion of pregnenolone (via androstadienol/androstadienone) is predominant compared to the conversion of progesterone [33,34]. Nevertheless, the conversion of progesterone into androstadienone was demonstrated in the pig testis and adrenal gland by Ahmad and Gower in 1968 [33] but these authors did not characterize any intermediate compounds, in particular 11-deoxycorticosterone (DOC) unlike Brooks and Pearson [1]. To our knowledge, no available data

demonstrate that DOC is an intermediary compound. Nevertheless, we decided to include this pathway and the direct conversion of progesterone into androstadienone proposed by Raeside et al. [4] in Fig. 1. An alternative pathway for the conversion of progesterone into corticosterone, via the synthesis of 11-hydroxy progesterone, was also proposed by Brooks and Pearson [1]. This pathway would need 11-hydroxylase (P450c11␤/CYP11B) and 21-hydroxylase (P450c21/CYP21). Progesterone originates from the conversion of pregnenolone and is located in the microsomal compartment. Unlike P450c21, P450c11␤ is located only in mitochondria [35]. This pathway, which has not been reported in humans, has never been confirmed in the pig. We consequently decided not to include it in Fig. 1. Pathways around DHEA and androstenedione (4-AD) have been reported to have special features regarding the porcine CYP17 enzyme (P450c17) [29]. The conversion of 17OHprogesterone > 4-AD is not possible in humans (see Section 3.2.2.3). The synthesis of androsterone has been reported in boar testes [18]. In humans, androstanedione (AD) is an intermediate product of the biosynthesis of DHT (dihydrotestosterone) by an intracrine pathway [30,36] while androsterone is an intermediate product by a “backdoor pathway”[36,37]. These pathways are included in Fig. 1 even though they have not been fully demonstrated in the pig (see also Sections 3.2.2.3 and 3.2.4). In boars, the testis is able to produce 11␤-hydroxy androgens (11␤-hydroxy androstenedione = 11-OH-4-AD and 11␤-hydroxy testosterone = 11-OH-Testo) [38] and the adrenal gland is able to produce adrenosterone (11-oxoandrostenedione) [39]. In humans, 11␤-hydroxy androgens are known to be produced by adrenal glands [40] from 4-AD. In humans, these 11␤-hydroxy androgens are transformed in reduced products (11-ketodihydrotestosterone (11-K-DHT) and 11␤-hydroxy dihydrotestosterone (11-OH-DHT) which have a higher androgenic activity [40,41]. In the pig, the pathways to produce 11␤-hydroxy androgens are not known. In Fig. 1, we show the pathways described in humans but these may be different in the pig. 3.2. Genes involved Several authors analyzed the expression of the genes involved in steroid metabolism [12,42–44] but a full description of these genes is still lacking. On the other hand, the complete pig genome is available [45] making it possible to perform an exhaustive examination of all genes possibly involved in steroid metabolism. All potential pathways concerning steroid biosynthesis in the pig are shown in Fig. 1 and, as far as possible, we identified the enzyme(s) involved in each pathway. When several genes are able to code for isoenzymes, it is very difficult to assign an isoform to a pathway. The isoform involved depends on the tissue considered. For that reason, Fig. 1 only includes generic names. A sequence of the porcine reference transcript is already available for the majority of genes encoding these proteins. Nevertheless we performed a systematic analysis of available ESTs and transcripts (Supplementary Table 1) to confirm the completeness of the reference transcript and to detect new transcript isoforms. We found very little information on transcripts for the genes CYP11B, HSD17B3, SRD5A1 and SRD5A2. Alternative transcripts were compiled for each of the 23 genes without a recent duplication in the porcine genome (Supplementary Table 1). The availability of a genomic sequence was checked (Table 1) and the structure of each gene was compared to the corresponding human gene. No obvious difference in exon/intron structure (number of exons, size of coding exons) between the two species was identified (data not shown). With knowledge of the gene structure, traces of alternative splicing events were characterized for 12 porcine genes among the 23

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

47

Fig. 1. Pathways involved in steroid biosynthesis in pigs. All biosynthesis pathways concerning steroid hormones and pheromones are shown in a generic pig cell. All circulating compounds and all important steroids (DHEA, EpiA, 4-AD, testosterone, 19-norTesto, estradiol, estrone, aldosterone, cortisol, androstenone, progesterone, and DHT) are in a gray box. Some metabolites that have never been depicted in pig cells are in dashed frames. As far as possible, we have indicated the enzyme involved in each pathway. When many enzymes are involved, we use only the generic term “17␤HSD”, “AKR1C” and “5␣-R”. Minor pathways are represented by arrows with finer lines than other pathways. Names of the proteins involved are given above or below the arrow indicating the pathway (between the two arrows when the same enzyme is involved in both pathways). Dashed arrows represent the flux of metabolites within the cell. Different types of arrows are used to distinguish pathways: arrow with an empty tip: pathway proposed in pig but never demonstrated; arrow with a double tip: pathway reported in humans but with a very minor activity; arrow with a triple tip: pathway never reported in humans; arrow with a vertical line at each end: pathway reported only in humans (peripheral tissues); “back-door pathway”: alternative pathway for DHT biosynthesis described by Fukami et al. [37]

tested (Supplementary Table 1). For some genes, new investigation was undertaken. The following presentation is organized according to the gene involved and grouped according to the family of enzymes. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsbmb. 2013.11.001. 3.2.1. STAR Steroidogenic endocrine tissues including adrenal glands and the gonads respond to external stimuli with a rapid surge in steroid hormone production. This acute steroidogenic response is mainly controlled by the steroidogenic acute regulatory (StAR) enzyme, a rapidly synthesized labile phosphoprotein encoded by the STAR gene [46,47]. This protein stimulates the movement of cholesterol from the outer mitochondrial membrane to the inner mitochondrial membrane [48] where it becomes available for the synthesis of pregnenolone. Few specific data are available for this enzyme in the pig, but we found no reason to believe that human data are not transferable to swine. The genomic sequence of porcine STAR is available (Table 1), and the size of the porcine gene is similar to that of the corresponding human gene (Table 2).

3.2.2. Cytochrome P450 Cytochrome P450 is a generic term for a group of oxidative enzymes. The genes coding for these proteins are formally termed CYP genes. The enzymes are membrane-bound proteins associated with either mitochondrial membranes, and in this case are termed “type 1”, or with the endoplasmic reticulum (microsomal) and in this case are termed “type 2” [48]. All P450-mediated hydroxylation and carbon–carbon bond cleavage reactions are mechanistically and physiologically irreversible [28]. 3.2.2.1. CYP11A1. The cytochrome P450 side-chain cleavage enzyme (P450scc) is encoded by the CYP11A1 gene and is present in the inner mitochondrial membrane of all steroidogenic cells [35,46]. It catalyzes the conversion of pregnenolone into progesterone by hydroxylation, and the cleavage of the 6-carbon side chain. Few porcine specific data are available for this enzyme, but we found no reason to believe that the function or substrates of this enzyme differ between pigs and humans. Transcript and genomic sequences have already been reported in pigs [24]. 3.2.2.2. CYP11B. In humans, the final steps of corticoid and aldosterone synthesis are catalyzed by 11 ␤-hydroxylase

48

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

Table 2 Genomic size of human and porcine genes involved in steroid biosynthesis. Protein or enzyme

Human gene

Size (kb)

Pig gene

Human gene P450scc P450c11 P450c17 P450aro P450c21 3␤HSD 11␤HSD1 11␤HSD2 17␤HSD1 17␤HSD2 17␤HSD3 17␤HSD4 17␤HSD7 17␤HSD12 5␣ reductase 1 5␣ reductase 2 StAR

CYP11A1 CYP11B1/CYP11B2 CYP17A1 CYP19A1 CYP21A2 HSD3B1/HSD3B2 HSD11B1 HSD11B2 HSD17B1 HSD17B2 HSD17B3 HSD17B4 HSD17B7 HSD17B12 SRD5A1 SRD5A2 STAR

AKR1C AKR1C AKR1C AKR1C AKR1C AKR1C AKR1C AKR1C AKR1C ? Cytochrome b5A Cytochrome b5B Cytb5-red 1 Cytb5-red 3 Ferredoxin Ferredoxin reductase P450 oxydoreductase

AKR1C1 AKR1C2 AKR1C3 AKR1C4

AKR1E2 CYB5A CYB5B CYB5R1 CYB5R3 FDX1 FDXR POR

29.98 6.59/7.28 7.0 90.7 3.41 7.85/8.10 48.75 6.44 3.24 63.3 66.85 89.9 21.78 175 36.22 59.6 7.62

Pig gene 5.3 (1) 6.43 3.11 8.61 5.92 3.13

22.53 25.97 7.94

CYP11A1 CYP11B (2) CYP17 CYP19A1/CYP19A2/CYP19A3 (2) CYP21 HSD3B1 (or HSD3B) (2) HSD11B1 HSD11B2 HSD17B1 HSD17B2 HSD17B3 HSD17B4 HSD17B7 HSD17B12 SRD5A1 SRD5A2 (or ST5AR2) STAR

20.03 16.25 13.89 22.11

21.81 38.72 41.7 5.41 26.67 35.0 10.54 87.6

14.64 9.7 27.9 13.42 12.7 35.6 35.5 5.57 24.9 10.73

AKR1C-pig1 AKR1C-pig3 AKR1C-pig4 AKR1C-pig6 AKR1E2 CYB5A CYB5B CYB5R1 CYB5R3 FDX1 FDXR POR

(1) The genomic sequence was determined in the present study (KF314688) but did not contain a 5 NC sequence. (2) Human and porcine genes are not orthologous.

(CYP11B1 ⇒ P450c11␤) and aldosterone synthase (CYP11B2 ⇒ P450c11AS), respectively [49]. CYP11B genes have been studied in many other species. In the rat, mouse, guinea pig, baboon and hamster, two types of CYP11B transcripts have also been described [50,51]. Similarly to human, the two types of transcripts are encoded by two different genes (CYP11B1 and CYP11B2) and are involved in the synthesis of either mineralo- or glucocorticoids. In contrast, in the pig, only one enzyme has been characterized [51,52]. In 1995, a sequence (D38590) was proposed for the main transcript of the CYP11B porcine gene [52]. In the pig, we were only able to identify three ESTs as corresponding to this gene, i.e. BI341717 and CV872108, which contain the 5 coding region, and AW353722, which contains the 3 coding region (even though this sequence is deposited as a cDNA sequence, it contains an intronic sequence). Using PCR on cDNA from porcine adrenal glands, we tried to amplify three different forms of transcripts in the 5 region of the CYP11B gene. PCR products were obtained with one upper primer defined in the 3 part of the “three exons 1” (chosen on CV872108, BI341717 and D38590 sequences) and a lower primer defined in exon 7 or in exon 6 (only chosen on D38590). No amplification was observed when the upper primer was chosen in D38590 or in BI341717 (adult adrenal glands, testes, ovaries and liver). Several fragments were sequenced but the conservation of the 5 end suggests the existence of a single gene. To characterize the last exons, a PCR fragment was observed between exon 5 (or exon 6 or exon 7) and exon 9 (4 possible primers designed on AW353722 or D38590). Amplifications were observed only when the lower primer was based on AW353722. Once again, it was not possible to amplify a fragment equivalent to D38590. In total, the seven exons of CYP11B were characterized in the pig

(GenBank KF314683). Many differences were found in AW353722 compared to D38590 but most were located in the first two exons. The method used to sequence the 5 and the 3 ends of D38590transcript [52] is probably responsible for these errors. As many differences were identified in exon 1 and exon 2 (between KF314683 and D38590), PCR amplification was performed between these two exons on genomic DNA to sequence the corresponding intron. Only one PCR product was obtained, and only when primers were designed using KF314683. In the amplified fragment, we found partial sequences of exon 1 and exon 2 and a 354 bp intron (387 bp in human CYP11B1 and CYP11B2). It is interesting to note that, in humans, the introns 1–2 from CYP11B1 and CYP11B2 differed by 29 nucleotides. Only primers designed on KF314683 were used to sequence other introns. We finally sequenced the entire coding sequence of CYP11B in the pig (GenBank KF314688). Despite further investigation of CYP11B, no proof of the existence of two different genes in pigs was found. Moreover, no genomic sequence related to a second CYP11B gene was included in Sus scrofa10.2. Despite some errors in the sequence of the porcine transcript, the previously performed phylogenetic analysis of CYP11B [53] was correct. In conclusion, a single CYP11B gene appears to exist in the pig in contrast to in other mammals. In these mammal species, the duplication occurred independently after speciation [53]. The genomic sequence of porcine CYP11B is available (Table 1), and the porcine gene (5.3 kb) is smaller than the human genes (Table 2). Like in humans, the porcine protein encoded by CYP11B (P450c11) does not only catalyze corticoid and aldosterone production. Indeed, the boar testis is able to produce 11␤-hydroxy

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

androgens [38], compounds that are also produced in humans [40]. Even if only one P450c11 enzyme is available in the pig, we found no difference between the porcine and human pathways involving these two enzymes. 3.2.2.3. CYP17. In humans, the enzyme encoded by CYP17A1 (P450c17) has two activities: 17␣-hydroxylase (pregnenolone → 17OH-pregnegnolone) and C17,20 lyase (17OHpregnenolone → DHEA). The 17␣-hydroxylase activity requires only P450 reductase (gene POR), unlike C17,20 lyase activity, which requires both P450 reductase and cytochrome b5 (genes CYB5A and CYB5B) [54]. In the porcine species, the enzyme P450c17 encoded by the CYP17 gene has three activities: 17␣-hydroxylase, C17,20 lyase and andien-␤ synthase activity (pregnenolone  androstenone) [29,55]. Concerning the 17␣-hydroxylase activity, porcine P450c17 is able to convert pregnenolone and progesterone in 17 hydroxylated products [29]. Regarding the C17,20 lyase activity, the human P450c17 does not accept substrates with a 4–5 bond: therefore, in contrast to the situation in pigs, the conversion17OH-progesterone > 4-AD is not possible in humans [29]. The complex andien ␤ synthase system includes P450c17, microsomal cytochrome b5, and cytochrome b5 reductase (gene: CYB5R) in the pig (Table 3) [55,56]. In boar testes, in the absence of cytochrome b5 (Cytb5), the biosynthesis of androgens is favored [55,57]. In other cases, when Cytb5 is present in sufficient quantity, androstadienol becomes the major product [55,58]. Andien-␤ synthase activity and C17,20 lyase activity are involved in the 16-androstene and androgens biosynthesis pathways, respectively. Andien-␤ synthase activity is not only found in the pig: in humans, androstadienol and androstadienone are also produced by the testis [55,59]. The genomic sequence of porcine CYP17 is available (Table 1), and the porcine gene is smaller than the human gene (Table 2). 3.2.2.4. CYP19A. Cytochrome P450 aromatase (P450aro), the product of the CYP19A gene, is a major enzyme in the estrogen biosynthetic pathway [60]. Mammals typically have a single CYP19A gene but pigs, peccaries, and other Suinae have two or more genes originating from duplication of a common ancestor [61]. Even though CYP19A was triplicated in the porcine genome [62], only sequences related to CYP19A1 and CYP19A2 are included in the Sus scrofa 10.2 draft sequences (Table 1). Small genomic sequences around exons 4, 5, 6, 7, 8, 9 and 10 are available for CYP19A1 (regrouped in AH006582, in this file CYP19A1 is incorrectly identified as CYP19A3) and around exons 7, 8, and 9 for CYP19A2 (regrouped in AH006583). We characterized intron 7–8 in the three CYP19A genes (GenBank KF314685, KF314686, and KF314687) and demonstrated the existence of three porcine CYP19A genes. Several authors designed primers to amplify only one of the CYP19A genes [42,43] but none stated exactly which CYP19A gene was amplified. Contrary to humans, three isoenzymes P450aro are available in the pig. Corbin et al. studied porcine P450aro and found that the gonadal and placental isoforms have different affinities with testosterone but not with 4-AD [63]. Moreover, the detection of 19-OH-4-AD and 19-OH-Testo in pig testes [31,63] suggests the existence of particular pathways involving porcine P450aro. We can propose that through specific features acquired during evolution (affinity/substrate, intermediate product, etc.) [61,64–66], the three CYP19A genes present in the pig genome contribute probably to the high production of estrogens by the boar testis, as well as to the production of 19-norTesto. 3.2.2.5. CYP21. The human gene CYP21A2 codes for a 21 steroid hydroxylase (P450c21). This enzyme catalyzes the synthesis of DOC from progesterone in human adrenal glands [28]. Analysis of the genomic context using genomicus, did not reveal any duplication

49

of CYP21 in humans, or in the mouse, rat, dog or pig. Nevertheless the name of the functional gene is CYP21A1 in rodents and CYP21A2 in humans and CYP21 in the pig. Few specific data are available for this enzyme in the pig, but we found no reason to believe that the human data are not transferable to swine. The genomic sequence of porcine CYP21 is available (Table 1), and the porcine gene is smaller than the human gene (Table 2). 3.2.3. Hydroxysteroid dehydrogenases (HSD) Whereas most steroidogenic reactions catalyzed by P450 enzymes are due to the action of a single form of P450, each of the reactions catalyzed by HSDs can be catalyzed by at least two, often very different, isozymes. HSD catalyzed reactions are mechanistically reversible and can run in either direction under certain conditions in vitro, but in vivo, each HSD drives a steroid flux predominantly in either the oxidative or reductive mode [67]. Dehydrogenases use NAD+ as their cofactor to oxidize hydroxysteroids to ketosteroids, and the reductases mainly use NADPH to reduce ketosteroids to hydroxysteroids. HSDs include two 3␤HSDs, two 11␤-HSDs, a series of 17␤-HSDs (encoded by HSD17B and AKR1 genes) and 3␣-HSDs (encoded by the AKR1 genes). 3.2.3.1. HSD3B. In humans, the 3␤-hydroxysteroid dehydrogenase/5-4 isomerase (3␤-HSD) isoenzymes are responsible for the oxidation and isomerization of 5–3␤-hydroxysteroid precursors into 4-ketosteroids, thus catalyzing an essential step in the synthesis of all classes of active steroid hormones [68,69]. 3␤HSD is encoded by two distinct genes in humans, HSD3B1 and HSD3B2. Both isoenzymes oxidize 3␤-hydroxy- to 3-keto-steroids (and also the reverse reaction) and isomerize the double bond between C5 and C6 to between C4 and C5. These reactions allow the production of 3-keto, 4-steroids, for example progesterone from pregnenolone and 4-AD from DHEA. In turn, the 3-keto, 4structure allows these steroids to be substrates for other enzymes, leading to the production of potent androgens, estrogens or corticoids. Hebert and Cooke [70] proposed that the conversion of androstadienol to androstadienone could be catalyzed by a 3␤hydrosteroid dehydrogenase-isomerase that is different from the enzyme involved in the 4-AD biosynthesis. Their hypothesis based on kinetic arguments has never been confirmed. Even if only one 3␤-HSD enzyme exists in the pig, we found no difference between humans and pigs for this pathway. Multiple isoforms of 3␤-HSD have been characterized in humans but only the two isoforms from the genes located on HSA1 were involved in the metabolism of sex and adrenal steroids. Human HSD3B1 (ENSG00000203857, NM 000862) and HSD3B2 (ENSG00000203859; NG 013349) genes have a similar structure. Their duplication could be recent, because the size of the intronic sequences is similar. Contrary to the human and rodent species where the duplication occurred independently after speciation (Genomicus), no duplication of the ancestral HSD3B gene has been identified in the pig [71]. Analysis of the genomic context with Genomicus showed that HSD3B2 could be the ancestral form. Even if in the pig, no duplication of the ancestral gene has been observed, the unique HSD3B gene was named HSD3B1, but no orthologous relation exists between human and porcine HSD3B1. The complete genomic sequence of porcine HSD3B1 is available (Table 1), and the porcine gene is a little longer than the human genes (Table 2). 3.2.3.2. HSD11B. The interconversion of biologically active cortisol and inactive cortisone (ketosteroid) in peripheral tissues is controlled by two independent 11␤-hydroxysteroid dehydrogenase enzymes, termed type I (HSD11B1 ⇒ 11␤-HSD1) and type II (HSD11B2 ⇒ 11␤-HSD2). Both enzymes have an oxidase and a reductase activity, depending on whether NADP+ (oxidation) or NADPH (reduction) is available as a cofactor [28]. Regarding human

50

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

Table 3 Comparison of human and pig steroid pathways involving P450c17. Enzymatic activity/pathway 17␣ hydroxylase Pregnenolone >17OH-pregnenolone C17,20 Lyase 17OH-pregnenolone > DHEA 17OH-progesterone > 4-AD Andien ˇ synthase Pregnenolone  androstadienone

Human

Pig

P450c17 + POR

P450c17 + POR

P450c17 + POR + cytb5 Reaction not possible

P450c17 + POR + Cytb5 P450c17 + POR + Cytb5

Very minor activity

P450c17 + Cytb5-red + Cytb5

steroidogenic tissues, 11␤-HSD1 is present in the testes and 11␤HSD2 is present in the placenta [28]. The two genes, HSD11B1 and HSD11B2, coding for the two enzymes are not located on the same human chromosome and do not appear to be the result of a recent duplication. HSD11B1 and HSD11B2 have also been characterized in porcine species [72]. Few porcine specific data are available for these enzymes, but we found no reason to believe that the function or substrates of these enzymes differ in the pig and in humans. The complete genomic sequence is available only for porcine HSD11B2 (Table 1); the porcine gene is a little smaller than the human gene (Table 2). 3.2.3.3. HSD17B. Some enzymes of the 17␤-hydroxysteroid dehydrogenase (17␤-HSDs) family are important for steroidogenesis, as they regulate the availability of both androgens and estrogens by interconversion of active and inactive forms of steroids [73]. Some of these enzymes preferentially trigger oxidase activity, whereas others preferentially trigger reductase activity. They catalyze many reactions including interconversion between estrone and estradiol, 4-AD and testosterone, and conversion of 5␣-androstanedione (5␣-AD) to DHT. To date, 14 different 17␤-HSDs have been identified in humans [28]. These enzymes can be classified in two families; most belong to the short-chain dehydrogenase/reductase (SDR) family (17␤-HSD1 for example), while others belong to the aldo-keto reductase (AKR) family (17␤-HSD5 for example). Among the many forms of 17␤-HSDs, at least three play an important role in the biosynthesis of potent gonadal steroids, types 1, 2 and 3 [74]. Even though 17␤-HSD3 [75,76] is often reported to be responsible for the conversion of 4-AD into testosterone, aldo-keto reductase C3 also plays a role [75,77]. Porcine 17␤-HSD3 is able to convert 4-AD into testosterone in vitro [78]. 17␤-HSD7 has been shown to be involved in both the biosynthesis of cholesterol and the reduction of estrone [74]. 17␤-HSD4 is the only 17␤-HSD located in the peroxisome that was shown to have multifunctional properties, including estradiol oxidation [79]. 17␤-HSD12 is mainly involved in the metabolism of fatty acids but it has also been suggested to play a major role in converting estrone into estradiol in postmenopausal women [80]. The real involvement of 17␤-HSD7, 17␤-HSD12, and 17␤-HSD4 in gonadal steroidogenesis is probably minor but nevertheless cannot be excluded [30]. For that reason, we considered that 17␤-HSD1 (gene: HSDB1), 17␤-HSD2 (HSDB2), 17␤-HSD3 (HSD17B3), 17␤-HSD4 (HSD17B4), 17␤-HSD7 (HSD17B7) and 17␤-HSD12 (HSD17B12) may be involved in the regulation of steroidogenesis in the pig testis (Fig. 1). Only the complete genomic sequences of porcine HSD17B1 and HSD17B7 are available (Table 1); these genes are similar in size to the equivalent human genes (Table 2). 3.2.3.4. AKR1 genes. Among the AKR enzyme family, the human AKR1C (3␣-HSD subfamily) includes AKR1C1, AKR1C2, AKR1C3 and AKR1C4, all of which cause ketosteroid reductase activity but to varying degrees [28,81]. In humans, AKR1C enzymes are involved in the synthesis of steroids at the adrenal and testicular levels and each enzyme displays a specific tissue distribution and repertoire of catalytic reductions [28,82,83]. AKR1C3 is also known as

17␤-HSD5. In humans, four genes in this family (in order 5 > 3 : AKR1C1, AKR1C2 (antisense), AKR1C3, AKR1C4), were included in a cluster with AKR1E2 (first position) and AKR1CL1 (AKR1C-like1, pseudogene, position 5, antisense). Although AKR1E2 is included in this cluster, the function of the protein is unknown [84]. AKR1C genes result from recent duplications and the four AKR1C genes have a similar structure (they code for a protein of 323 amino acids (aa)). As very few data are available for the pig and those in databases are contradictory, we performed new analyses. In the pig, six transcripts were identified as reference transcripts (Table 4): DQ474064 (NM 001044568), DQ474065 (NM 001044569), DQ474066 (NM 001044570), DQ474067 (NM 001044618), DQ474068 (NM 001038626), and DQ494489 (NM 001123075). After examining and comparing these six transcripts, we decided that DQ494489 and DQ474068 are in fact the same transcript. Regarding DQ494489, no sequence was found in databases to confirm the size of the 3 untranslated region. Among porcine reference genomic sequences, NW 003611309 and NW 003611308 contain genes related to this family. NW 003611309 derives from several BACs sequences but the most important is CH242-23K3. It was simpler to go back to the sequence of this original BAC (CU972427) and to examine this sequence as it has been incorporated in NW 003611309. The sequence of this BAC was, in fact, in two segments. Using BLAST, Sim4, and GeneSeqer, we successively found (1) a gene coding for a cDNA similar to AK391133 (named pig-1), (2) a gene without a previously known transcript (pig-2), (3) a gene coding for DQ474066 (pig-3), (4) a gene coding for DQ474067 (pig-4), and (5) a gene coding for DQ474064 (pig-5), in the large fragment. After a gap, in the small fragment, we found a gene coding for DQ474068 (pig6). For the pig-1 gene, except for one nucleotide, this sequence, found in the BAC, encoded a putative protein product of 319 aa resembling AK391133 or AK233685. On the same strand, a sequence sharing similarities with the AKR1C family was found in position 2 (pig-2). We did not find a transcript or EST sequence for this gene in existing databases. Sim4 and GeneSeqer proposed several identifications for genomic sequences that could be exons. We chose several primer pairs to characterize the transcript by amplification of cDNA from porcine adrenal glands, and the fragments obtained were sequenced (GenBank KF614684). Three exons were characterized and the first two contained several stop codons. For the gene in position 3 (pig-3) in the large fragment, we found 7 coding exons corresponding to the transcript DQ474066. For the pig-4 gene, we found 12 exons, the last nine being coding exons (like the human AKR1C4 gene), corresponding to transcript DQ474067. For the gene in position 5 on the other strand (pig-5), we characterized 10 exons corresponding to the transcript DQ474064. Only the first nine were coding exons and this porcine protein contained 288 aa. The size of these coding exons (except exon 4) was similar to the size of the corresponding exons in the human AKR1E2 gene (Table 4). For the gene located in the small fragment (pig-6), we identified 9 coding exons and this gene coded for a protein of 322 aa (transcript DQ474068).

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

51

Table 4 Characterization of the porcine genes in the AKR1 cluster. Name

pig-1 pig-2 pig-3 pig-4 pig-5 gap pig-6

Location in CU972427

Transcript sequence

Protein

Similarities with human proteins InterProScan analysis

Phylogenetic analysis Proposed name

4 > 19 kb 40 > 44 kb 54 > 64 kb 74 > 102 kb 104 < 118 kb

AK391133 KF314684 DQ474066 (NM 001044570) DQ474067 (NM 001044618) DQ474064 (NM 001044568)

319 aa None 271 aa 324 aa 288 aa

AKR1C1 or AKR1C3 or AKR1C4 Pseudogene AKR1C2 AKR1C1 or AKR1C3 or AKR1C4

AKR1C-pig1(1) AKR1CLP AKR1C-pig3 (1) AKR1C-pig4 (1) AKR1E2 (2)

156 < 139 kb

DQ474068 (NM 001038626)

322 aa

AKR1C1 or AKR1C3 or AKR1C4

AKR1C-pig6 (1)

(1) The human gene is not orthologous to the porcine gene. (2) The porcine gene pig-5 is orthologous to the human AKR1E2 gene.

In this AKR1 gene cluster, six genes were characterized as coding for five putative porcine proteins (Table 4). The pig-2 gene presented more similarities with the pig-1 gene than others (Supplementary Table 2A) but it is probably not coding. Human proteins AKR1C1 and AKR1C2 closely resembled each other (97% similarity) as did pig-1 and pig-3 (90% similarity). Human AKR1C3 and AKR1C4 presented more similarities (85% similarity) than pig-4 and pig-6 (68%) (Supplementary Table 2B). A phylogenetic analysis was performed with human and porcine proteins (Fig. 2). As pig-5 and human AKR1E2 were located on a separate branch, we suggest that AKR1E2 and pig-5 genes result from duplication of an ancestral AKR1 gene. This duplication occurred before speciation (pig-5 and human AKR1E2 are orthologous) contrary to the differentiation of the four AKR1C genes, which appear to result from duplications occurring after speciation. These results are based on proteins and reflect only non-synonymous mutations in genes. The number of non-synonymous substitutions per non-synonymous site (dN) and the number of synonymous substitutions per synonymous site (dS) were calculated (Supplementary Table 3) and compared [85]. Levels of synonymous mutations were higher in the pig than in humans for these genes. It is interesting to note that observed dS confirmed that the duplications of AKR1C genes occurred earlier in the phylogenetic tree (built with the molecular clock hypothesis) in the pig than in humans (Fig. 2). Analysis of the genomic context in other mammal species (human/mouse/rat) with genomicus confirmed this hypothesis (data not shown): the four AKR1C porcine genes (pig-1, pig-3, pig-4, pig-6) are not orthologous to the four AKR1C human genes.

Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsbmb. 2013.11.001. The InterProScan web site was used to analyze porcine and human AKR1C proteins (Table 4): the protein encoded by pig-3 appeared to be similar to an AKR1C2 protein. For pig-1, pig-4, and pig-6, this analysis confirmed that they were AKR1C proteins. Thereby, we observed that two independent evolutionary ways led to similarities in the function of proteins encoded by pig-3 and human AKR1C2 genes. Other evidence confirming the hypothesis of a convergence of independent evolutionary ways was the presence of two 5 NC exons for pig-4 and human AKR1C4 genes. Consequently, we propose a new name for each porcine AKR1C gene based on the phylogenetic results (Table 4). According to our phylogenetic analysis, the four porcine AKR1C genes are not orthologous to the four human AKR1C genes and it does not make sense to use the same nomenclature. Moreover, a single sequence was previously attributed to the porcine AKR1C3 gene (DQ474067) and to the porcine AKR1C1 (NM 001044618). InterproScan analyses of porcine and human AKR1E2 proteins showed that their structure resembles that of an AKR1C protein. Nevertheless, no information is available on their physiological involvement in steroid biosynthesis pathways. Pig-2 presented the best homologies with pig-1 and human AKR1C2/AKR1C1 (Supplementary Table 2A) but only three exons have been characterized to date. Like the transcript associated with the human AKR1CL1 gene (NR 027916), the porcine pig-2 transcript included several stop codons and is probably a non-coding transcript. Pig-2 is probably an ancestral

Fig. 2. Phylogenetic analysis of human and porcine proteins encoded by genes in theAK1R cluster. Phylogenic trees of human and porcine AKR1 proteins were computed with SeaView. Alignment of multiple sequences was performed using Muscle. The tree was built with the bioNJ method (distance of Kimura) using sequences derived from the following transcript files: AKR1C1-hu (NM 001353), AKR1C2-hu (NM 001354), AKR1C3-hu (NM 001253908), AKR1C4-hu (NM 001818), AKR1E2-hu (NM 001040177), pig-1 (AK391133), pig-3 (DQ474066), pig-4 (DQ474067), pig-5 (DQ474064), and pig-6 (DQ494489). Proteins were truncated to conserve only “core sequences” (see Supplementary Table 2). As the pig-2 gene was a pseudogene, it was not possible to include the pig-2 protein in the phylogenic tree.

52

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

gene so it can actually be considered as a pseudogene. We propose to name it AKR1Clike pseudogene (AKR1CLP). The sequence of the transcript (GenBank KF314684) is included in the Sus scrofa 10.2 draft sequence (except for one insertion/deletion at the end of the sequence). In NW 003611309, between pig-1 (LOC733634) and pig-3 (LOC100519510) a pseudogene (LOC100623947) was identified by GNOMON that included eight exons. The first exon characterized in KF314684 (uncomplete in 5 ) was compatible with exon 5 of LOC100623947. The second exon characterized in KF314684 was compatible with exon 6 of LOC100623947 but extended in 5 . The third exon characterized in KF314684 (uncomplete in 3 ) was compatible with exon 7 of LOC100623947. Our characterization of this pseudogene was performed independently of the automated computational analysis proposed by Gnomon. The last four exons proposed by Gnomon (LOC100623947) were identical to those proposed by GeneSeqer (which proposed only these four exons). In the present study, the transcription of these four exons was tested but only three were found. Moreover, the experimental evidence (sequencing of a fragment of cDNA amplified by PCR) is in contradiction with the common definition of exon 6 provided by GeneSeqer, Sim4 and Gnomon. In addition, supporting evidence for the automated annotation performed by Gnomon only includes similarity to one protein. In the present study, the existence of one transcript was proved and the sequence of the end of this transcript is provided. Genomic sequences of porcine AKR1 genes are available. The sizes of porcine and human AKR1C genes are very different (Table 2) confirming that these genes result from several duplications that occurred after speciation. According to the genomic sequence of CU972427, AKR1C-pig1, (AKR1CLpseudo), AKR1C-pig3, AKR1C-pig4, AKR1E2 (antisense) and AKR1C-pig6 (antisense but after the gap) should be successively depicted. In contrast to their organization in humans, the five genes are not on the same strand. This divergence confirms that porcine and human AKR1C genes are not orthologous. Porcine AKR1C genes characterization have already been performed [86]] but before publication of the pig genome sequence [45] and changes in human references. Our examination of the sequence (Sus scrofa 10.2) of the region containing the AKR1 cluster indicated that NW 003611309 and NW 003611308 contain genes from this family but also redundant sequences. A comparison of sequences showed that this assembly probably involves several assembly errors. Therefore, it is likely that the AKR1 cluster is not duplicated in the porcine genome. 3.2.4. SRD5A 5␣-Reductases (5␣-R) are important beyond the context of male genital differentiation and androgen action because these enzymes reduce a variety of steroids in catabolic pathways: they allow the reduction of the 4–5 bond of 3-oxo (3-keto), 4,5 C19/C21 steroids [87]. Three isoenzymes of 5␣-R are known to exist (5␣-R1-3), encoded by three distinct genes, respectively SRD5A1, SRD5A2, and SRD5A3 located on separate chromosomes 5, 2 and 4, respectively, in humans [88,89]. Only 5␣-R1 and 5␣-R2 appear to be involved in steroid metabolism, at least in humans. Indeed, in humans, loss-of-function mutations in SRD5A3 are associated with mental retardation and a glycosylation defect rather than a detectable deficiency in steroid metabolism and sexual differentiation, suggesting that the major biochemical function of this enzyme is not in androgen metabolism (reported by [88,90]). The abundant expression of type 1 isoform in the liver and its high affinity for C21steroids suggest that it plays a primary role in degrading circulating C21 steroids for excretion in urine [28,91]. In humans, the 5␣-R2 enzyme has mainly been observed in male reproductive tissues including seminal vesicles, epididymis, and prostate [92] and its essential function in male differentiation is firmly established: a mutation in SRD5A2 leads to a virilization deficit [93]. Nevertheless,

when a SRD5A2 gene deficiency exists, SRD5A1 hyperactivity is observed [94]. In humans, SRD5A1 appear to be involved in 5␣ reduction of steroids but to a lesser extent than SRD5A2 [89]. Androstadienol and androstadienone can be produced by human testes [59,95] but this 16-androstene pathway does not go any further due to a lack of 5␣-R activity. Even if androstadienone is not reduced to androstenone in the human testis, Dufort et al. [96] demonstrated that conversion is possible in vitro with cells expressing the SRD5A2 gene. Human and boar testes have different 5␣ reducing abilities. Boar testes are able to produce 5␣reduced androgens (androsterone, epiA and androstenone) but the synthesis of DHT has never been described in porcine testes [4]. In humans, 5␣-R possesses a lower affinity for testosterone than 4-AD (reported by Luu-The [36]) and hence testosterone is a poor substrate for 5␣-reduction in tissues where 4-AD is available. Current knowledge of porcine 5␣-R enzymes [97] is still not sufficient to explain why porcine testicular tissues can produce androstenone and EpiA but not DHT. According to Smals and Weusten [98] “the presence of androstenone and androstenol in human blood and urine indicates that both 16-androstenes must be synthesized in a 5␣-R rich compartment outside the testis or ovary (skin?)”. In humans, the 5␣-reduction of testosterone to DHT occurs mainly in target tissues [30]. In the human prostate, the pathway “formation of DHT from 4-AD via AD” appears to be active only in CRPC (castration resistant prostate cancer) when the SRD5A2 gene is inactive and is replaced by the SRD5A1 gene [90]. This pathway was described as an intracrine pathway to produce DHT in human peripheral tissues [30,36,99]. To our knowledge, this pathway and the steroid metabolism of peripheral tissues have not been explored in the pig. Since there is no obvious reason to discard it, we hypothesize it exists as a minor pathway in the pig (Fig. 1). The complete genomic sequence of porcine SRD5A1 is available (Table 1); this gene is smaller in pigs than in humans (Table 2). The genomic sequence of porcine SRD5A2 is partially available but it is important to note that only two sequences related to transcripts of SRD5A2 are available in databases (Supplementary Table 1). 3.2.5. Co-factors All microsomal cytochrome P450 enzymes including those involved in steroidogenesis (P450c17, P450c21 and P450aro) receive electrons from the cytochrome P450 reductase (gene POR), a membrane-bound protein [28]. Mitochondrial cytochrome P450 enzymes (P450c11␤/AS, P450scc) receive electrons from ferredoxin (gene FDX1) and regeneration is assumed to be catalyzed by ferredoxin reductase (gene FDRX) [28]. In addition, microsomal Cytb5 (genes: CYB5A and CYB5B) and NADPH cytochrome b5 reductase (gene: CYB5R) are able to modulate P450c17 activities [55,100]. Two isoforms of Cytb5 have been identified (genes: CYB5A and CYB5B) in the porcine testis that could have selective effects on P450c17 activities (Table 3) [101]. Gray and Squires [43] suggest that cytochrome b5 reductase (Cytb5-red) (encoded by genes CYB5R1 and CYB5R3) belongs to the andien ␤ synthase complex (Table 3). The complete genomic sequences of porcine CYB5A, CYB5B, CYB5R1, CYB5R3, and FDXR are available (Table 1); these genes are smaller in pigs than in humans except FDXR, which is similar in size (Table 2). 3.2.6. Availability of sequences In the present work, the sequence of the main transcript for 32 genes was analyzed and several transcripts were rectified (Table 1), mainly non-coding extremities. The analysis of AKR1 genes led us to rename their transcripts (Table 4). Knowledge of all transcripts and alternative transcripts of porcine genes is not at the same level (Supplementary Table 1). For example, very little information is

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

available concerning the transcripts of genes CYP11B, HSD17B3, SRD5A1 and SRD5A2 in the porcine species. The same is true of genes CYP11B2, HSD17B3 and SRD5A2 in humans. Moreover caution is advised in interpreting alternative splicing information (Supplementary Table 1). Most alternative splicing events are characterized by only one sequence in both porcine and human species. Nevertheless, concerning the porcine genes CYP11A1, CYP21A2, HSD11B1 and HSD11B2, at least two sequence records appear to support the same splicing event. Very few data are available for SRD5A2 in either in the pig or in humans. However alternative transcripts have been observed in both species. It is interesting to note that the alternative splicing events differ greatly between species. For example, in a gene with a very high level of alternative splicing in humans (HSD17B4, 115 alternative transcripts recorded by ASPicDB), no splicing event has been reported in pigs (Supplementary Table 1). Moreover, for CYP21A2, retention of introns appears to be the most frequent splicing event in the pig but not in humans. Genomic sequences have been reported in Sus scrofa 10.2 for 18 (complete sequence) and 13 (partial sequence) genes, respectively (Table 1). The present study provides the complete sequence of the CYP11B gene. Comparison of the size of orthologous genes showed that they are very similar (r2 = 0.97, n = 13 pairs) but with a tendency to a smaller size in the pig than in humans. Moreover, three porcine genes (AKR1E2, SRD5A1 and CYB5B) are considerably smaller in the pig than in humans (respectively −40%, −28% and −15%), suggesting possible errors in the current pig genome assembly or in the description of the main transcripts. 4. Conclusion The existing description and characterization of genes encoding enzymes or cofactors involved in steroid biosynthesis in the pig were improved thanks to the present work. Transcript data are available for all porcine genes. Genomic data are also available for the great majority of porcine genes. No sequence of CYP11B was included in Sus-scrofa 10.2 but the present work allows us to propose a complete sequence for this small gene. For genes with recent duplications in the porcine genome and that are still organized in clusters, the assembly of the corresponding region is problematic and was probably incorrect, as was the case in the example studied here. Our work, combined with a review of the literature, shows that the porcine genome includes only one CYP11B gene (two in humans), one HSD3B gene (two in humans) but three CYP19A genes (one in humans). Very few specific traits of steroid biosynthesis appear to be related to the number of gene copies. Nevertheless, the presence of three CYP19A genes in the genome may contribute to the high production of estrogens by the boar testis. Finally, the testis of the boar is unusual among mammals regarding the release of high amounts of 5␣-reduced steroids in the blood. Current knowledge on the porcine 5␣-R enzymes is still not sufficient to explain why porcine testicular tissues can produce androstenone and EpiA but not DHT. Acknowledgements We would like to thank the team running the genomic platform of the Génopole Toulouse Midi-Pyrénées (http://genopole-toulouse.prd.fr/index.php?lang=fr) for their contribution to sequence generation. We also thank Emmanuelle Permal for her help and expertise regarding the PALM package. References [1] R.I. Brooks, A.M. Pearson, Steroid hormone pathways in the pig, with special emphasis on boar odor: a review, J. Anim. Sci. 62 (1986) 632–645.

53

[2] A. Robic, C. Larzul, M. Bonneau, Genetic and metabolic aspects of androstenone and skatole deposition in pig adipose tissue: a review, Genet. Sel. Evol. 40 (2008) 129–143 (Open Access publication). [3] G. Zamaratskaia, E.J. Squires, Biochemical, nutritional and genetic effects on boar taint in entire male pigs, Animal 3 (2009) 1508–1521. [4] J.I. Raeside, H.L. Christie, R.L. Renaud, P.A. Sinclair, The boar testis: the most versatile steroid producing organ known, Soc. Reprod. Fertil. Suppl. 62 (2006) 85–97. [5] S. Carreau, R.A. Hess, Oestrogens and spermatogenesis, Philos. Trans. R. Soc. Lond. B Biol. Sci. 365 (2010) 1517–1535. [6] R. Claus, Physiological role of seminal components in the reproductive tract of the female pig, J. Reprod. Fertil. Suppl. 40 (1990) 117–131. [7] O. Andresen, Concentrations of fat and plasma 5␣-androstenone and plasma testosterone in boars selected for rate of body weight gain and thickness of back fat during growth, sexual maturation and after mating, J. Reprod. Fertil. 48 (1976) 51–59. [8] F. Schwarzenberger, G.S. Toole, H.L. Christie, J.I. Raeside, Plasma levels of several androgens and estrogens from birth to puberty in male domestic pigs, Acta Endocrinol. (Copenh.) 128 (1993) 173–177. [9] R.D. Allrich, R.K. Christenson, J.J. Ford, D.R. Zimmerman, Pubertal development of the boar: testosterone, estradiol-17␤, cortisol and LH concentrations before and after castration at various ages, J. Anim. Sci. 55 (1982) 1139–1146. [10] P.A. Sinclair, E.J. Squires, J.I. Raeside, Early postnatal plasma concentrations of testicular steroid hormones, pubertal development, and carcass leanness as potential indicators of boar taint in market weight intact male pigs, J. Anim. Sci. 79 (2001) 1868–1876. [11] L.R. Franca, V.A. Silva, H. Chiarini-Garcia Jr., S.K. Garcia, L. Debeljuk, Cell proliferation and hormonal changes during postnatal development of the testis in the pig, Biol. Reprod. 63 (2000) 1629–1636. [12] I. Choi, J.Y. Kim, E.J. Lee, Y.Y. Kim, C.S. Chung, J. Chang, N.J. Choi, H.J. Chung, K.H. Lee, Ontogeny of expression and localization of steroidogenic enzymes in the neonatal and prepubertal pig testes, J. Androl. 30 (2009) 57–74. [13] T. Berger, L. Kentfield, J.F. Roser, A. Conley, Stimulation of Sertoli cell proliferation: defining the response interval to an inhibitor of estrogen synthesis in the boar, Reproduction 143 (2012) 523–529. [14] S. Zdunczyk, T. Janowski, A. Ras, W. Baranski, Concentrations of oestrogens in blood plasma and seminal plasma of boars during the postpuberal period, Pol. J. Vet. Sci. 14 (2011) 539–544. [15] R. Claus, C. Hoang-Vu, F. Ellendorff, H.D. Meyer, D. Schopper, U. Weiler, Seminal oestrogens in the boar: origin and functions in the sow, J. Steroid. Biochem. 27 (1987) 331–335. [16] P.H. Hemsworth, A.J. Tilbrook, Sexual behavior of male pigs, Horm. Behav. 52 (2007) 39–44. [17] J.I. Raeside, R.L. Renaud, D.E. Marshall, Identification of 5 alpha-androstane3beta,17 beta-diol and 3 beta-hydroxy-5 alpha-androstan-17-one sulfates as quantitatively significant secretory products of porcine Leydig cells and their presence in testicular venous blood, J. Steroid. Biochem. Mol. Biol. 42 (1992) 113–120. [18] A. Ruokonen, R. Vihko, Steroid metabolism in testis tissue: concentrations of unconjugated and sulfated neutral steroids in boar testis, J. Steroid. Biochem. 5 (1974) 33–38. [19] G.C. Perry, R.L.S. Patterson, H.J.H. MackFie, C.G. Stinson, Pig courtship behaviour: pheromonal property of androstenone steroids in the male submaxillary secretion, Anim. Prod. 331 (1980) 191–199. [20] K. Lundstrom, K.R. Matthews, J.E. Haugen, Pig meat quality from entire males, Animal 3 (2009) 1497–1507. [21] B. Fredriksen, I.F.M. Font, K. Lundstrom, W. Migdal, A. Prunier, F.A. Tuyttens, M. Bonneau, Practice on castration of piglets in Europe, Animal 3 (2009) 1480–1487. [22] A. Prunier, A.M. Mounier, M. Hay, Effects of castration, tooth resection, or tail docking on plasma metabolites and stress hormones in young pigs, J. Anim. Sci. 83 (2005) 216–222. [23] T. Castrignano, R. Rizzi, I.G. Talamo, P.D. De Meo, A. Anselmo, P. Bonizzoni, G. Pesole, ASPIC: a web resource for alternative splicing prediction and transcript isoforms characterization, Nucleic Acids Res. 34 (2006) W440–W443. [24] A. Robic, G. Le Mignon, K. Feve, C. Larzul, J. Riquet, New investigations around CYP11A1 and its possible involvement in an androstenone QTL characterised in Large White pigs, Genet. Sel. Evol. 43 (2011) 15. [25] M. Gouy, S. Guindon, O. Gascuel, SeaView version 4: a multiplatform graphical user interface for sequence alignment and phylogenetic tree building, Mol. Biol. Evol. 27 (2010) 221–224. [26] A. Louis, M. Muffato, H. Roest Crollius, Genomicus: five genome browsers for comparative genomics in eukaryota, Nucleic Acids Res. 41 (2012) D700–D705. [27] W.L. Miller, Molecular biology of steroid hormone synthesis, Endocr. Rev. 9 (1988) 295–318. [28] W.L. Miller, R.J. Auchus, The molecular biology, biochemistry, and physiology of human steroidogenesis and its disorders, Endocr. Rev. 32 (2011) 81–151. [29] A.A. Gilep, T.A. Sushko, S.A. Usanov, At the crossroads of steroid hormone biosynthesis: the role, substrate specificity and evolutionary development of CYP17, Biochim. Biophys. Acta 1814 (2011) 200–209. [30] V. Luu-The, F. Labrie, The intracrine sex steroid biosynthesis pathways, Prog. Brain Res. 181 (2010) 177–192. [31] J.I. Raeside, R.L. Renaud, R.M. Friendship, M.W. Khalil, Secretion of 19-hydroxyandrostenedione and 19-hydroxytestosterone by porcine Leydig cells in vitro and in vivo, J. Endocrinol. 137 (1993) 281–289.

54

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55

[32] S. Poelmans, K. De Wasch, H. Noppe, N. Van Hoof, S. Van Cruchten, et al., Endogenous occurrence of some anabolic steroids in swine matrices, Food Addit. Contam. 22 (2005) 808–815. [33] N. Ahmad, D.B. Gower, The biosynthesis of some androst-16-enes from C21 and C19 steroids in boar testicular and adrenal tissue, Biochem. J. 108 (1968) 233–241. [34] I. Louveau, M. Bonneau, D.B. Gower, Biosynthesis of 16-androstene steroids and testosterone by porcine testis tissue in vitro: effect of age and relationships with fat 5 alpha-androstenone levels in vivo, Acta Endocrinol. (Copenh.) 125 (1991) 526–531. [35] F. Mitani, T. Shimizu, R. Ueno, Y. Ishimura, S. Izumi, N. Komatsu, K. Watanabe, Cytochrome P-45011 beta and P-450scc in adrenal cortex: zonal distribution and intramitochondrial localization by the horseradish peroxidase-labeled antibody method, J. Histochem. Cytochem. 30 (1982) 1066–1074. [36] V. Luu, TheAssessment of steroidogenesis and steroidogenic enzyme functions, J. Steroid Biochem. Mol. Biol. 137 (2013) 176–182. [37] M. Fukami, K. Homma, T. Hasegawa, T. Ogata, Backdoor pathway for dihydrotestosterone biosynthesis: implications for normal and abnormal human sex development, Dev. Dyn. 242 (2013) 320–329. [38] J.I. Raeside, R.L. Renaud, M.W. Khalil, Formation of C19 11-hydroxysteroids by porcine Leydig cells, Biochem. Cell Biol. 70 (1992) 174–176. [39] M. Holzbauer, H.M. Newport, Adrenal secretion rates and adrenal tissue concentrations of pregnenolone, progesterone, 11 beta OH-androstenedione and some other steroids in young pigs and dogs, J. Physiol. 200 (1969) 821–848. [40] A.C. Swart, L. Schloms, K.H. Storbeck, L.M. Bloem, T.D. Toit, J.L. Quanson, W.E. Rainey, P. Swart, 11beta-Hydroxyandrostenedione, the product of androstenedione metabolism in the adrenal, is metabolized in LNCaP cells by 5alpha-reductase yielding 11beta-hydroxy-5alpha-androstanedione, J. Steroid. Biochem. Mol. Biol. 138 (2013) 132–142. [41] K.H. Storbeck, L.M. Bloem, D. Africander, L. Schloms, P. Swart, A.C. Swart, 11beta-Hydroxydihydrotestosterone and 11-ketodihydrotestosterone, novel C19 steroids with androgenic activity: A putative role in castration resistant prostate cancer? Mol. Cell Endocrinol. 377 (2013) 135–146. [42] E. Grindflek, I. Berget, M. Moe, P. Oeth, S. Lien, Transcript profiling of candidate genes in testis of pigs exhibiting large differences in androstenone levels, BMC Genet. 11 (2010) 4. [43] M.A. Gray, E.J. Squires, Effects of nuclear receptor transactivation on steroid hormone synthesis and gene expression in porcine Leydig cells, J. Steroid. Biochem. Mol. Biol. 133 (2013) 93–100. [44] S. Lervik, K. von Krogh, C. Karlsson, I. Olsaker, O. Andresen, E. Dahl, S. Verhaegen, E. Ropstad, Steroidogenesis in primary cultures of neonatal porcine Leydig cells from Duroc and Norwegian Landrace breeds, Theriogenology 76 (2011) 1058–1069. [45] M.A. Groenen, A.L. Archibald, H. Uenishi, C.K. Tuggle, Y. Takeuchi, et al., Analyses of pig genomes provide insight into porcine demography and evolution, Nature 491 (2012) 393–398. [46] W.L. Miller, H.S. Bose, Early steps in steroidogenesis: intracellular cholesterol trafficking, J. Lipid Res. (2011). [47] P.R. Manna, M.T. Dyson, D.M. Stocco, Regulation of the steroidogenic acute regulatory protein gene expression: present and future perspectives, Mol. Hum. Reprod. 15 (2009) 321–333. [48] V. Papadopoulos, W.L. Miller, Role of mitochondria in steroidogenesis, Best Pract. Res. Clin. Endocrinol. Metab. 26 (2012) 771–790. [49] E. Mornet, J. Dupont, A. Vitek, P.C. White, Characterization of two genes encoding human steroid 11 beta-hydroxylase (P-450(11) beta), J. Biol. Chem. 264 (1989) 20961–20967. [50] M. Lisurek, R. Bernhardt, Modulation of aldosterone and cortisol synthesis on the molecular level, Mol. Cell Endocrinol. 215 (2004) 149–159. [51] M. Okamoto, Y. Nonaka, H. Takemori, J. Doi, Molecular identity and gene expression of aldosterone synthase cytochrome P450, Biochem. Biophys. Res. Commun. 338 (2005) 325–330. [52] T. Sun, Y. Zhao, Y. Nonaka, M. Okamoto, Cloning and expression of cytochrome P450(11 beta) of porcine adrenal cortex, J. Steroid Biochem. Mol. Biol. 52 (1995) 227–232. [53] H.E. Bulow, R. Bernhardt, Analyses of the CYP11B gene family in the guinea pig suggest the existence of a primordial CYP11B gene with aldosterone synthase activity, Eur. J. Biochem. 269 (2002) 3838–3846. [54] W.E. Rainey, Y. Nakamura, Regulation of the adrenal androgen biosynthesis, J. Steroid Biochem. Mol. Biol. 108 (2008) 281–286. [55] P. Soucy, L. Lacoste, V. Luu-The, Assessment of porcine and human 16-enesynthase, a third activity of P450c17, in the formation of an androstenol precursor. Role of recombinant cytochrome b5 and P450 reductase, Eur. J. Biochem. 270 (2003) 1349–1355. [56] W.J. Meadus, J.I. Mason, E.J. Squires, Cytochrome P450c17 from porcine and bovine adrenal catalyses the formation of 5,16-androstadien-3 beta-ol from pregnenolone in the presence of cytochrome b5, J. Steroid Biochem. Mol. Biol. 46 (1993) 565–572. [57] S.M. Davis, E.J. Squires, Association of cytochrome b5 with 16-androstene steroid synthesis in the testis and accumulation in the fat of male pigs, J. Anim. Sci. 77 (1999) 1230–1235. [58] J. Lavallee, G.M. Cooke, 17-Hydroxylase and andien-beta synthetase activities in immature pig testis microsomal fraction: kinetic studies of the pregnenolone binding site and possible intermediates of the reactions, J. Steroid Biochem. Mol. Biol. 46 (1993) 73–83.

[59] J.J. Weusten, G. Legemaat, M.P. van der Wouw, A.G. Smals, P.W. Kloppenborg, T. Benraad, The mechanism of the synthesis of 16-androstenes in human testicular homogenates, J. Steroid Biochem. 32 (1989) 689–694. [60] E.A. Gaucher, L.G. Graddy, T. Li, R.C. Simmen, F.A. Simmen, D.R. Schreiber, D.A. Liberles, C.M. Janis, S.A. Benner, The planetary biology of cytochrome P450 aromatases, BMC Biol. 2 (2004) 19. [61] A.J. Conley, C.J. Corbin, A.L. Hughes, Adaptive evolution of mammalian aromatases: lessons from Suiformes, J. Exp. Zool. A Ecol. Genet. Physiol. 311 (2009) 346–357. [62] L.G. Graddy, A.A. Kowalski, F.A. Simmen, S.L. Davis, W.W. Baumgartner, R.C. Simmen, Multiple isoforms of porcine aromatase are encoded by three distinct genes, J. Steroid Biochem. Mol. Biol. 73 (2000) 49–57. [63] C.J. Corbin, J.M. Trant, K.W. Walters, A.J. Conley, Changes in testosterone metabolism associated with the evolution of placental and gonadal isozymes of porcine aromatase cytochrome P450, Endocrinology 140 (1999) 5202–5210. [64] C.J. Corbin, T. Berger, J.J. Ford, C.E. Roselli, W. Sienkiewicz, et al., Porcine hypothalamic aromatase cytochrome P450: isoform characterization, sexdependent activity, regional expression, and regulation by enzyme inhibition in neonatal boars, Biol. Reprod. 81 (2009) 388–395. [65] C.J. Corbin, A.L. Hughes, J.R. Heffelfinger, T. Berger, T.B. Waltzek, et al., Evolution of suiform aromatases: ancestral duplication with conservation of tissue-specific expression in the collared peccary (Pecari tayassu), J. Mol. Evol. 65 (2007) 403–412. [66] C.J. Corbin, J.M. Trant, A.J. Conley, Porcine gonadal and placental isozymes of aromatase cytochrome P450: sub-cellular distribution and support by NADPH-cytochrome P450 reductase, Mol. Cell Endocrinol. 172 (2001) 115–124. [67] A.K. Agarwal, R.J. Auchus, Minireview: cellular redox state regulates hydroxysteroid dehydrogenase activity and intracellular hormone potency, Endocrinology 146 (2005) 2531–2538. [68] J. Simard, M.L. Ricketts, S. Gingras, P. Soucy, F.A. Feltus, M.H. Melner, Molecular biology of the 3beta-hydroxysteroid dehydrogenase/delta5-delta4 isomerase gene family, Endocr. Rev. 26 (2005) 525–582. [69] M.K. Rasmussen, B. Ekstrand, G. Zamaratskaia, Regulation of 3betahydroxysteroid dehydrogenase/delta5-delta4 isomerase: a review, Int. J. Mol. Sci. 14 (2013) 17926–17942. [70] P. Hebert, G.M. Cooke, Kinetic evidence for separate 3 beta-hydroxysteroid dehydrogenase-isomerases in androgen and 16-androstene biosynthetic pathways in the pig testis, J. Steroid Biochem. Mol. Biol. 42 (1992) 901–910. [71] A. von Teichman, H. Joerg, P. Werner, B. Brenig, G. Stranzinger, cDNA cloning and physical mapping of porcine 3 beta-hydroxysteroid dehydrogenase/Delta 5-Delta 4 isomerase, Anim. Genet. 32 (2001) 298–302. [72] V. Sharp, L.M. Thurston, R.C. Fowkes, A.E. Michael, Expression and activities of 11betaHSD enzymes in the testes and reproductive tracts of sexually immature male pigs, J. Steroid Biochem. Mol. Biol. 115 (2009) 98–106. [73] M.E. Baker, Evolution of 17beta-hydroxysteroid dehydrogenases and their role in androgen, estrogen and retinoid action, Mol. Cell Endocrinol. 171 (2001) 211–215. [74] T. Saloniemi, H. Jokela, L. Strauss, P. Pakarinen, M. Poutanen, The diversity of sex steroid action: novel functions of hydroxysteroid (17beta) dehydrogenases as revealed by genetically modified mouse models, J. Endocrinol. 212 (2012) 27–40. [75] K. Blouin, A. Veilleux, V. Luu-The, A. Tchernof, Androgen metabolism in adipose tissue: recent advances, Mol. Cell Endocrinol. 301 (2009) 97–103. [76] J. Beuten, J.A. Gelfond, J.L. Franke, K.S. Weldon, A.C. Crandall, T.L. Johnson-Pais, I.M. Thompson, R.J. Leach, Single and multigenic analysis of the association between variants in 12 steroid hormone metabolism genes and risk of prostate cancer, Cancer Epidemiol. Biomarkers Prev. 18 (2009) 1869–1880. [77] T.M. Penning, M.C. Byrns, Steroid hormone transforming aldo-keto reductases and cancer, Ann. N. Y. Acad. Sci. 1155 (2009) 33–42. [78] S. Ohno, Y. Honda, Y. Nakajima, S. Nakajin, Sequence of pig 17betahydroxysteroid dehydrogenase Type3 cDNA and its expression in mammalian cells, J. Steroid Biochem. Mol. Biol. 101 (2006) 145–150. [79] R. Breitling, Z. Marijanovic, D. Perovic, J. Adamski, Evolution of 17beta-HSD type 4, a multifunctional protein of beta-oxidation, Mol. Cell Endocrinol. 171 (2001) 205–210. [80] V. Luu-The, P. Tremblay, F. Labrie, Characterization of type 12 17betahydroxysteroid dehydrogenase, an isoform of type 3 17beta-hydroxysteroid dehydrogenase responsible for estradiol formation in women, Mol. Endocrinol. 20 (2006) 437–443. [81] K. Blouin, M. Nadeau, J. Mailloux, M. Daris, S. Lebel, V. Luu-The, A. Tchernof, Pathways of adipose tissue androgen metabolism in women: depot differences and modulation by adipogenesis, Am. J. Physiol. Endocrinol. Metab. 296 (2009) E244–E255. [82] X.G. Hui, J. Akahira, T. Suzuki, M. Nio, Y. Nakamura, H. Suzuki, W.E. Rainey, H. Sasano, Development of the human adrenal zona reticularis: morphometric and immunohistochemical studies from birth to adolescence, J. Endocrinol. 203 (2009) 241–252. [83] C.J. Petry, K.K. Ong, D.L. Wingate, F. de Zegher, L. Ibanez, D.B. Dunger, Lack of association between common polymorphisms in the 17beta-hydroxysteroid dehydrogenase type V gene (HSD17B5) and precocious pubarche, J. Steroid Biochem. Mol. Biol. 105 (2007) 176–180. [84] T.M. Penning, J.E. Drury, Human aldo-keto reductases: function, gene regulation, and single nucleotide polymorphisms, Arch. Biochem. Biophys. 464 (2007) 241–250.

A. Robic et al. / Journal of Steroid Biochemistry & Molecular Biology 140 (2014) 44–55 [85] M. Nei, T. Gojobori, Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions, Mol. Biol. Evol. 3 (1986) 418–426. [86] D.J. Nonneman, T.H. Wise, J.J. Ford, L.A. Kuehn, G.A. Rohrer, Characterization of the aldo-keto reductase 1C gene cluster on pig chromosome 10: possible associations with reproductive traits, BMC Vet. Res. 2 (2006) 28. [87] J.D. Wilson, The role of 5alpha-reduction in steroid hormone physiology, Reprod. Fertil. Dev. 13 (2001) 673–678. [88] A.M. Traish, 5␣-Reductases in human physiology: an unfolding story, Endocr. Pract. 18 (2012) 965–975. [89] F. Azzouni, A. Godoy, Y. Li, J. Mohler, The 5␣-reductase isozyme family: a review of basic biology and their role in human diseases, Adv. Urol. 2012 (2012) 530121. [90] N. Sharifi, Minireview: androgen metabolism in castration-resistant prostate cancer, Mol. Endocrinol. 27 (2013) 708–714. [91] A.E. Thigpen, R.I. Silver, J.M. Guileyardo, M.L. Casey, J.D. McConnell, D.W. Russell, Tissue distribution and ontogeny of steroid 5 alpha-reductase isozyme expression, J. Clin. Invest. 92 (1993) 903–910. [92] K. Normington, D.W. Russell, Tissue distribution and kinetic characteristics of rat steroid 5 alpha-reductase isozymes. Evidence for distinct physiological functions, J. Biol. Chem. 267 (1992) 19548–19554. [93] O. Hiort, S.M. Schutt, M. Bals-Pratsch, P.M. Holterhus, C. Marschke, D. Struve, A novel homozygous disruptive mutation in the SRD5A2-gene in a partially virilized patient with 5alpha-reductase deficiency, Int. J. Androl. 25 (2002) 55–58.

55

[94] S. Thiele, U. Hoppe, P.M. Holterhus, O. Hiort, Isoenzyme type 1 of 5alphareductase is abundantly transcribed in normal human genital skin fibroblasts and may play an important role in masculinization of 5alpha-reductase type 2 deficient males, Eur. J. Endocrinol. 152 (2005) 875–880. [95] T.K. Kwan, M.A. Kraevskaya, H.L. Makin, D.J. Trafford, D.B. Gower, Use of gas chromatographic-mass spectrometric techniques in studies of androst-16ene and androgen biosynthesis in human testis; cytosolic specific binding of 5alpha-androst-16-en-3-one, J. Steroid Biochem. Mol. Biol. 60 (1997) 137–146. [96] I. Dufort, P. Soucy, L. Lacoste, V. Luu-The, Comparative biosynthetic pathway of androstenol and androgens, J. Steroid Biochem. Mol. Biol. 77 (2001) 223–227. [97] G.M. Cooke, F. Pothier, B.D. Murphy, The effects of progesterone, 4,16androstadien-3-one and MK-434 on the kinetics of pig testis microsomal testosterone-4-ene-5alpha-reductase activity, J. Steroid Biochem. Mol. Biol. 60 (1997) 353–359. [98] A.G. Smals, J.J. Weusten, 16-Ene-steroids in the human testis, J. Steroid Biochem. Mol. Biol. 40 (1991) 587–592. [99] J.F. Yarrow, S.C. McCoy, S.E. Borst, Intracrine and myotrophic roles of 5alphareductase and androgens: a review, Med. Sci. Sports Exerc. 44 (2012) 818–826. [100] K.H. Storbeck, A.C. Swart, P. Goosen, P. Swart, Cytochrome b(5): novel roles in steroidogenesis, Mol. Cell Endocrinol. 371 (2012) 97–99. [101] M.J. Billen, E.J. Squires, The role of porcine cytochrome b5A and cytochrome b5B in the regulation of cytochrome P45017A1 activities, J. Steroid Biochem. Mol. Biol. 113 (2009) 98–104.