Part I: Critical Appraisal of the Evidence for the Presence, Biosynthesis and Uptake of Steroids

Steroids 77 (2012) 1450–1468

Contents lists available at SciVerse ScienceDirect

Steroids

journal homepage: www.elsevier.com/locate/steroids

Review Do mollusks use vertebrate sex steroids as reproductive hormones? Part I: Critical appraisal of the evidence for the presence, biosynthesis and uptake of steroids

Alexander P. Scott

Centre for Environment, Fisheries and Aquaculture Science, Barrack Road, Weymouth, Dorset DT4 8UB, UK article info abstract

Article history: The consensus view is that vertebrate-type steroids are present in mollusks and perform hormonal roles Received 12 June 2012 which are similar to those that they play in vertebrates. Although vertebrate steroids can be measured in Received in revised form 16 August 2012 molluscan tissues, a key question is ‘Are they formed endogenously or they are picked up from their envi- Accepted 21 August 2012 ronment?’. The present review concludes that there is no convincing evidence for biosynthesis of verte- Available online 31 August 2012 brate steroids by mollusks. Furthermore, the ‘mollusk’ genome does not contain the genes for key enzymes that are necessary to transform cholesterol in progressive steps into vertebrate-type steroids; Keywords: nor does the mollusk genome contain genes for functioning classical nuclear steroid receptors. On the Vertebrate steroid other hand, there is very strong evidence that mollusks are able to absorb vertebrate steroids from the Invertebrate steroid Mollusc environment; and are able to store some of them (by conjugating them to fatty acids) for weeks to Biosynthesis months. It is notable that the three steroids that have been proposed as functional hormones in mollusks Receptor (i.e. progesterone, testosterone and 17b-estradiol) are the same as those of humans. Since humans (and Tissue concentrations indeed all vertebrates) continuously excrete steroids not just via urine and feces, but via their body surface (and, in ﬁsh, via the gills), it is impossible to rule out contamination as the sole reason for the presence of vertebrate steroids in mollusks (even in animals kept under supposedly ‘clean laboratory conditions’). Essentially, the presence of vertebrate steroids in mollusks cannot be taken as reliable evidence of either endogenous biosynthesis or of an endocrine role. Crown Copyright Ó 2012 Published by Elsevier Inc. All rights reserved.

Contents

1. General introduction ...... 1451 2. Occurrence of steroids in molluscan tissues...... 1451 2.1. Seasonal changes...... 1451 2.2. Sex differences...... 1451 2.3. Evidence that contaminants affect steroid concentrations ...... 1451 2.4. Tissue differences ...... 1454 2.5. Studies not included in Table 1 ...... 1454 3. What might explain the presence of vertebrate steroids in mollusks? ...... 1454 4. Could there be problems with the measurement procedures? ...... 1455 4.1. Lack of specificity ...... 1455 5. Do the animals biosynthesize the steroids themselves? ...... 1455 5.1. Evidence for a vertebrate-type biosynthetic pathway ...... 1455 5.2. Possible problems with steroid identification ...... 1455 5.3. Incubation procedures ...... 1456 5.4. The problems of low (and unreported) yields ...... 1456 5.5. Do mollusks have cholesterol? ...... 1456 5.6. The first key step in steroid biosynthesis – the cleavage of the side chain of cholesterol ...... 1456

5.7. D5-3b-HSD activity (column 2 in Table 2)...... 1458 5.8. C17,20-lyase (column 3 in Table 2)...... 1458 5.9. 17b-hydroxysteroid dehydrogenase (17b-HSD) activity (column 4 in Table 2) ...... 1459

E-mail address: [email protected]

5.10. 5a-reductase (column 6 in Table 2)...... 1459 5.11. CYP19 (aromatase; column 5 in Table 2)...... 1459 5.12. A strong streak of anthropomorphism in mollusk studies ...... 1459 5.13. Interim conclusion ...... 1459 6. The animals pick up the steroids from the environment?...... 1460 6.1. Possible sources of steroids in the environment ...... 1460 6.2. Experimental evidence that mollusks can take up and store exogenous vertebrate steroids ...... 1460 6.3. What is the point of esterification? ...... 1461 6.4. What can explain the presence of steroids in laboratory-reared animals? ...... 1461 6.5. Looking for steroids that are unlikely to have been formed endogenously...... 1461 7. The evidence for steroid receptors in mollusks ...... 1462 7.1. Presence of classical nuclear receptors ...... 1462 7.2. Another class of nuclear receptor?...... 1462 7.3. Could it be a ‘non-classical’ membrane-bound receptor?...... 1464 7.4. Could there be binding proteins that are not receptors? ...... 1464 7.5. Does the presence of a receptor necessarily mean the ligand has to be endogenous?...... 1464 7.6. Is a receptor necessary to elicit effects? ...... 1464 8. Conclusion ...... 1464 9. Future directions? ...... 1465 Acknowledgements ...... 1465 References ...... 1465

1. General introduction tion. An indication is also given of the magnitude of these changes – as the fact that steroid concentrations change (with season in The question as to whether vertebrate steroids act as hormones particular) is frequently cited as evidence that the steroids perhaps in mollusks is a very important one. Many compounds that behave function as reproductive hormones as they do in vertebrates. like vertebrate estrogens are present in the environment (so-called ‘endocrine disrupters’), and if they act in the same way in mollusks as they do in vertebrates [1] then there is genuine cause for con- 2.1. Seasonal changes cern. The first reports of the existence of vertebrate-type steroids in mollusks appeared in the 1950s [2,3]. Since then, all but a hand- One argument for the involvement of vertebrate steroids in ful of the 200+ scientific papers and reviews that have been pub- mollusk reproduction is that, in many studies (column 7 in Table 1), lished in this area have a positive message (i.e. they conclude the concentration of steroids (and this mostly refers to free ste- either that mollusks contain vertebrate steroids, are able to biosyn- roids) varies significantly with the time of year. Furthermore, these thesize them de novo, appear to contain steroid receptor-like bind- changes have been claimed to match reproductive events. How- ing activity or respond in one way or another when exposed to ever, neither the size of the change nor any apparent association vertebrate steroids). The sheer ‘weight of evidence’ would seem, of a peak with reproductive events can be considered as incontro- on the surface, to make it an ‘open and shut’ case that vertebrate vertible evidence that the animals use those steroids as hormones. steroids are an important component of molluscan endocrinology. Equally plausible alternative hypotheses are, firstly, that the con- However, if one looks beyond the headline claims (i.e. essentially centrations of the steroids just reflect the levels of steroids that what is written in the titles and abstracts of many of the papers), are available to be absorbed from the water (see Section 6) or sec- a different story emerges – one in which most, if not all, of the po- ondly, any association is purely by chance. sitive evidence can be seen to be rather weak (in that the data are open to alternative interpretations). This review paper deals with the strength of the evidence for the presence of steroids in mol- 2.2. Sex differences lusks, for their biosynthesis and for the presence of steroid receptors. Another review paper [4] deals with the strength of evidence Very little difference has been found between steroid concen- for the biological actions of vertebrate steroids on mollusks. Most trations in males and females (column 8 in Table 1). However, this of the literature that has been reviewed was obtained from the ref- neither proves nor disproves that T and E2 are hormones in mol- erence lists of previous reviews and key papers in the field. More lusks. In most teleost fishes, both steroids are found in the blood recent papers were picked up by searching Scopus and Web of Sci- plasma of females [5]; and in a few species, both steroids also occur ence for papers that cited certain key papers (plus keywords such in male plasma [6]. as ‘mollusk’ and ‘steroid’). A final search was made in March, 2012.

2. Occurrence of steroids in molluscan tissues 2.3. Evidence that contaminants affect steroid concentrations

There are numerous (>50) publications that report the presence A few experiments have been carried out to determine whether of typical vertebrate steroids such as testosterone (T), 17b-estra- contaminants, especially tributyl tin (TBT), affect steroid concen- diol (E2) and progesterone (P) in a wide range of mollusks (Table trations (column 9 in Table 1). The effects have in most cases been 1). In Table 1, concentrations are all shown as an approximate minimal. Other studies have involved sampling (or deliberately range (or in some cases a single value); are quoted as pg gÀ1 wet placing) animals at contaminated or pristine sites. Although these weight of tissue; and are shown for free as well as esterified (see have tended to yield larger differences in steroid concentrations Section 6) steroids, when both have been measured. Mode of assay than purely laboratory experiments, it is impossible to rule out is briefly indicated. The final three columns indicate whether the that the differences were due to the amounts of steroids that the steroid concentrations were found to differ by season, sex or pollu- animals might have taken up from the environment (see Section 6). Table 1 1452 Steroid concentrations that have been reported in molluscan tissues.

Species Source Tissue Steroid Assay Levels (pg g À1 ) Seasonality/stage differences? Sex differences? Contamination effects? Giant ramshorn snail, Lab. Whole tissue T (ester) RIA 800 (2000–30,000) 4-fold Yes for esterified T and E2 TBT for 100 days reduced esterified T and E2 Marisa cornuarietis Stock E2 (ester) 100 (6000–90,000) (males 3–5 times higher) in females only [75] Cuttlefish, Sepia officinalis Wild Hemolymph T RIA 150 – No – [40] Estrogen 10 ? Tissue T 10,000–20,000 – Dogwhelk Nucella lapillus Wild/ Whole tissue T (ester) RIA 1000 (10,000) – – Slight elevation of free by TBT [81] tank (female only) E2 1000 2-fold elevation by TBT + CPA Dogwhelk Nucella lapillus Wild/ Whole tissue of T RIA 1000–7000 – – Slight effect of TBT [117] tank females E2 20–500 P 1000 Dogwhelks Nucella Wild/ Whole tissue of T RIA 600–1200 – – 2-fold difference between imposex stages 0 lapillus and Hinia tank females only and 4 reticulata [118] Eastern mud snail, Wild Gonad/viscera T (ester) RIA 25,000 (100,000) 4-fold (fem) to 12-fold (male), 10-fold higher in females – Ilyanassa obsoleta but same pattern in both sexes at dormant phase ..Sot/Seod 7(02 1450–1468 (2012) 77 Steroids / Scott A.P. [119] E2 (ester) 500 (3000) 10-fold No Eastern mud snail, Wild Whole body T RIA 2000–45,000 (highest levels in April and 10-fold No – Ilyanassa obsoleta November coincide with low [68,69] esterification) Eastern mud snail, Wild/ Whole body T (ester) RIA 2500 (25,000) – – Slight increase in free T after 3 months of Ilyanassa obsoleta tank TBT [68,78] Garden snail, Helix Culture Gonad T RIA and 8000–28,000 3-fold higher in juveniles – – aspersa [82] AdGC–MS 2500 3-fold higher in juveniles P 3000 2-fold higher in adults Hemolymph T 300 No Ad 20 No P 1000–6000 5-fold higher in adults Giant African land snail, Lab Hemolymph T RIA 30–700 – Yes – Achatina fulica [17] E2 0–2000 Yes P 440–8000 Yes Ad 0–100 Yes Cortisol 300–2000 No Grooved carpet shell, Culture Whole tissue T RIA 120,000 – – After 5 weeks in TBT-contaminated harbor, T Ruditapes decussatus E2 5000–20,000 rose and E2 fell [120] Grooved carpet shell, Wild/ Whole tissue T RIA 100–800 – – Increased by TBT exposure Ruditapes decussatus tank minus digestive E2 <0.2 [121] gland Grooved carpet shell, Wild GonadE2 T RIA 10–240 40–400 20-fold 8-fold Slight – Ruditapes decussatus P 200–2500 15-fold Japanese[122] scallop, Wild Gonad E2 EC 1500–4500 3-fold No – Paticopecten yessoensis E1 <500 [56] Manilla carpet shell, Wild Whole body T RIA 100–200 2-fold Slight – Tapes philippinarum E2 100–300 3-fold [123] P 600–1600 3-fold Table 1 (continued)

Species Source Tissue Steroid Assay Levels (pg g À1 ) Seasonality/stage differences? Sex differences? Contamination effects?

Mussel, Mytilus edulis Wild Mantle + gonad T RIA 1400–40,000 40-fold in males 10-fold in males at late – [42] stage E2 4500 Only present in late stage No E1 35,000 Only present in late stages No

Mussel, Mytilus edulis Wild Whole body P GC–MS 5,000–40,000 10-fold No – [124] Gonad RIA 1,500–4,000 3-fold Mussel, Mytilus edulis Wild Whole body T GC–MS 200–600 3-fold No - [125] E2 20–40 2-fold P 500–5000 10-fold E1 60 No change Ad 1000–3500 3.5-fold Mussel, Mytilus edulis [8] Wild Gonad T (ester) RIA 3500 (4000–12000) – – 3-fold increase in esteriﬁed steroids in E2 (ester) 1000 (4000–11000) gonads after exposure to North Sea Oil for Peripheral tissue T (ester) 150 (2000–5000) 3 weeks E2 (ester) 20 (2000) Mussel, Mytilus Wild Whole tissue E2 (ester) RIA 800 (10,000) – – Not affected by addition of T to water for

galloprovincalis [73] 5 days 1450–1468 (2012) 77 Steroids / Scott A.P. Mussel, Mytilus Wild Whole tissue T (ester) RIA 1000 (1500) – – Not affected by addition of E2 to water for galloprovincalis [126] 5 days Mussel, Mytilus Wild Mantle + gonad E2 RIA <10–100 (but from pooled samples only) 10-fold – – galloprovincalis [127] Mussel, Mytilus spp. [25] Wild Digestive gland P GC–MS <400–7000 >15-fold No – Preg 500–9000 18-fold No T, E1, E2, All <400 – No Ad, DHT, DHA Octopus, Octopus vulgaris Wild Reproductive T RIA/EIA 2500–5000 – – – [9] tissue of males and 200–1000 only HPLC 1000–5000 Octopus, Octopus vulgaris Wild/ Ovary E2 RIA 25–200 4-fold elevation of both –– [128] Tank steroids in May only (i.e. no cycle) Paciﬁc oyster, Crassostrea Culture Gonad E2 RIA 0–1500 >20-fold 3-fold in females – gigas [50] E1 0–300 >20-fold Peppery furrow shell, Wild Gonad T EIA 60–240 4-fold Yes Yes T higher in males at two sites and P Scrobicularia plana E2 100–300 3-fold No higher in females at other site [129] P 200–2800 5-fold Yes Purple dye murex, Bolinus Wild Digestive T RIA 800 – No E2 levels much lower at TBT-contaminated brandaris [130] gland + gonad E2 <5–250 site Rockshell, Thais clavigera Wild Testis and ovary T EIA & 1000 – No No [7,24] Culture Gonad E2 ECGC–MS 400–1100 No 2-fold higher in females – Scallop, Patinopecten[50] Wild Gonad E1 200 No only Soft-shellyessoensis clam, Mya Wild Gonad T EIA & 40 Slight (2-fold) Slightly higher in females – arenaria [131] E2 GC–MS 300 Soft-shell clam, Mya Wild Gonad P EIA & 4000 Slight (1.5-fold) No – arenaria [132] GC–MS 1453

(continued on next page) 1454 A.P. Scott / Steroids 77 (2012) 1450–1468

2.4. Tissue differences

Very few studies seem to have examined tissue distribution of steroids. In the rockshell, Thais clavigera [7], the gonads had ﬁve times higher concentrations of T and E2 than in the remaining tissues. In the mussel, Mytilus edulis [8], the concentrations of free (but not of esteriﬁed) T and E2 were 18- to 91-fold, respectively, higher in the gonads than in the remaining tissue. In the octopus, Octopus vulgaris [9], steroid levels were also more abundant in reproductive tissue (i.e. testis vas deferens, seminal vesicle and prostate) than in non-reproductive tissues. Steroid concentrations in hemolymph (the mollusk equivalent of ‘plasma’) have in general been found to be much lower than those that can be extracted from No effect of NP on E2 amounts tissues (see Table 1). It would be tempting to conclude that if steroid concentrations in gonads are higher than those in other tissues it means either that they are synthesized by the gonads [8] or that they must have a role in reproduction. However, a plausible alter-

nt ‘Free’ steroid that is present in a non-conjugated form; Ad, native explanation is that steroids are preferentially accumulated by the gonads because of their generally higher content of fat and protein (to which steroids could be non-speciﬁcally absorbed).

one; RIA, radioimmunoassay; T, testosterone; TBT, tributyltin. 2.5. Studies not included in Table 1

There are at least seven studies on the New Zealand mudsnail, Potamopyrgus antipodarum that have not been included in Table 1, because steroid concentrations in these studies were only quoted À1 À1 À1

-estradiol; EC, electron capture detector on HPLC; EIA, enzyme immunoassay; GC–MS, gas as fmol individual or pg individual (and not pg g tissue as b in Table 1) and, also, the authors in most cases only measured total (i.e. free + ester) steroid concentrations [10–16]. There are, surprisingly, only two studies involving in vitro release of steroids by gonad tissues from mollusks. These are not included in Table 1 either. In the giant African land snail, Achatina fulica [17], steroids were found to be released by ovotestis tissue. More P and T release was found by male phase gonads than by female phase gonads and vice versa for E2. No cortisol release was found in vitro (despite evidence for its presence in hemolymph). In O. vulgaris [18], ovarian follicles and spermatozoa were both shown to release immunoactive T (10-fold), E2 (2-fold) and ) Seasonality/stage differences? Sex differences? Contamination effects?

1 P (4-fold) in vitro, over a period of 90 min, in response to stimula- À -dihydrotestosterone; E1, estrone; E2, 17

a tion with a peptide similar to vertebrate gonadotropin-releasing hormone. Finally there are several early studies in which steroids were putatively identiﬁed by bioassay or chromatography coupled with microchemistry. For example, in the periwinkle, Littorina littorea, compounds that had estrogenic and androgenic effects in mam- mals were extracted from the gonads [3]. In the slug, Arion ater rufus [19], the presence of ‘estrogens’ was revealed by TLC and microchemistry in spermatheca that had been incubated in vitro. There were no signs of precursors such as T and androstenedione E2 (ester) 300 (8000) T (ester) RIA 100 (4000) – – Highest dose of ﬂuoxetine doubled E2 E2 (ester) 60 (8000)

-dihydroandrostenedione; DHT, 5 (Ad), however and authors suggested that these steroids may have a already been present in the tissues (i.e. not synthesized de novo). Another study identified 11-ketotestosterone (KT; an androgen in teleosts) by the same methods in the same species [20]. However, later studies showed no evidence for biosynthesis of this steroid in Whole tissue T (ester) RIA 120 (7000) – – Esterified T reduced by NP treatment Whole body (excl. gills) the banana slug, Ariolimax californicus [21]. In the scallop, Pecten maximus [22], P was tentatively identified by its position on TLC. Wild/ tank Wild/ tank Wild Gonad P EIA <600–5000 5-foldNone of these No procedures are – specific enough for any of the identi- fications to be accepted as ‘definitive’. ‘Ester’, amount of steroid that is conjugated to fatty acids (and can be converted to free steroid by base hydrolysis); numbers not in brackets represe Mya ) [135] [134] Dreissena Dreissena 3. What might explain the presence of vertebrate steroids in

[133] mollusks? continued ( As we have seen, there are a surprisingly large number of stud- polymorpha polymorpha arenaria ies that report the presence of vertebrate-type (predominantly hu- Zebra mussel, Soft-shell clam, Zebra mussel, Species Source Tissue Steroid Assay Levels (pg g androstenedione; CPA, cyproterone acetate; DHA, 5 Abbreviations used: chromatography coupled with mass spectrometry; HPLC, high performance liquid chromatography; NP, nonylphenol; P, progesterone; Preg, pregnenol Table 1 man) steroids in the tissues of mollusks. To explain the presence of A.P. Scott / Steroids 77 (2012) 1450–1468 1455 these steroids in mollusk tissues, there are essentially three A. Part of a speciﬁc pathway involved in the formation of the hypotheses – none of them mutually exclusive: same steroids that are found in vertebrates? B. Part of a speciﬁc pathway involved in the formation of ste- A. Problems exist with the measurement procedures. roid hormones that are peculiar to the mollusk in question? B. The animals biosynthesize the steroids themselves. C. Part of a pathway involved in the ‘metabolism’ of steroids? C. The animals sequester the steroids from the environment If the answer is metabolism, then a further question that may be asked is whether the substrate that is transformed by the mollusk 4. Could there be problems with the measurement procedures? in question is:

4.1. Lack of specificity a) an endogenous substrate for which there is a specific enzyme, The main methodology that has been used to measure steroids b) an exogenous substrate that is transformed by an enzyme is immunoassay. Immunoassays are notoriously liable to interfer- that has evolved to deal with a different substrate ence termed ‘matrix effects’ in which compounds that have been altogether, extracted along with the steroids affect the affinity of the antibody c) an exogenous substrate that is transformed by an enzyme and thus alter the overall binding to the labeled ligand (and hence that has evolved to deal with that specific substrate the apparent concentration of steroid). Also, no steroid antibodies (because, perhaps, that substrate has been encountered in are 100% specific and some compounds (usually closely related ste- the environment for millions of years). roids; and conjugates of the steroid) are able to displace the labeled steroid to the same or a lesser extent. Very few of the published pa- One further question [27] (especially when working with filter pers in the mollusk field include any attempt to even partially feeders) is whether any of the transformations might have been characterize the purported steroids that were being measured. caused by the algae and bacteria that might have been mixed in One paper that did [23] found that most of the E2 measured by with the tissue. radioimmunoassay (RIA) in tissue extracts of M. edulis, ran in the In biosynthetic studies on mollusks (Table 2), there has been at void volume of a high performance liquid chromatography (HPLC) least one positive report (defined as a reported yield of >=2% and column and that only 10% actually ran in the expected elution po- shown in bold type in the table) that indicates the presence in mol- sition of E2. The latter peak, however, was definitively identified as lusks of all steps of the pathway necessary for the conversion of E2 by mass spectrometry. Another study [24] showed that, when pregnenolone (Preg) to P, T and E2. However, these positive obser- enzyme immunoassay (EIA) was used to measure T in crude tissue vations, especially those on two of the key steps in the cycle (i.e. extracts, it gave at least 10-fold higher readings than gas chroma- those catalyzed by D5-3b-hydroxysteroid dehydrogenase [D5-3b- tography coupled with mass spectrometry (GC–MS). However after HSD] and C17,20-lyase) have all been on different species. Despite carrying out several purification steps, the EIA gave the same read- this, it has been concluded that the whole vertebrate pathway ing as GC–MS. probably exists in mollusks [27,29], but is, for some unknown rea- In addition to the possibility of matrix effects and cross-reac- son (that, it is suggested, will be revealed by more research), diffi- tion, there are plenty of opportunities for human error in assay cult to demonstrate in an individual species. However, this measurements. However, without access to data sheets or labora- conclusion ignores the existence of the far higher proportion of tory workbooks, it is impossible to establish that such errors might studies that have yielded negative, or weakly positive, results or might not have taken place. There are enough studies on steroid (shown in normal type in Table 2) and secondly, it assumes that measurement in mollusks, however, that have been repeated and these few positive observations were robust. This is not possible independently verified and have used highly specific detection to establish without independent verification. All sorts of things methods (e.g. GC–MS), that one must accept that vertebrate ste- can go wrong with biosynthetic studies that might yield a false po- roids can definitely (though, it must be stressed, not always [25] sitive result (e.g. contamination, misidentification of compounds, be detected in molluscan tissues. Also, it seems very clear that sub- mislabeling and miscalculation). stantial amounts of T and E2 in particular are present in the form of In view of the importance of being able to demonstrate all the fatty acid esters (in those species that have been investigated). steps of the cycle (using quantitative yield data) within a single species, it is astonishing that, in the last 30 years, only one group [30] has carried out any experiments using either Preg, P or 17- 5. Do the animals biosynthesize the steroids themselves? hydroxypregn-4-ene-3,20-dione (17-P) as precursors. That single study showed that the animals were unable to transform P 5.1. Evidence for a vertebrate-type biosynthetic pathway in vivo. All other biosynthetic studies that have been carried out since 1981 have been concerned purely with conversions of andro- The pathway that leads to the biosynthesis of steroids such as P, gens or estrogens (that, as will be made clearer, provide no clues at T and E2 in humans is well-known and well-characterized (Fig. 1). all as to whether the steroid biosynthesis pathway in mollusks is The conventional way to demonstrate the existence of steroid bio- the same as that in vertebrates). synthetic pathways is to incubate tissue explants with radioactive steroid precursors. Although there have not been a huge number of 5.2. Possible problems with steroid identification such studies in mollusks (Table 2), they are often cited in support of the claim that mollusks are able to make their own vertebrate- The methodology that investigators have used for identifying type steroids. So how robust is the evidence? radioactive steroids (column 7 in Table 2) would seem in most There is little doubt that mollusk tissues will transform steroids cases to be robust. All have used a combination of chromatography in vitro. None of the studies in Table 2 have reported zero transfor- (coupled with microchemistry) and recrystallization to determine mation of added precursors. This is not surprising, as steroids are the identity of the steroids (Table 2). However, it should be cau- also transformed by organisms as diverse as algae, bacteria, other tioned that such steps are not necessarily foolproof. In a study on invertebrates and higher plants [26–28]. What we need to ask is the scallop, Placopecten magellanicus [31], Preg was 34% converted whether, in mollusks, any such transformations are: to a steroid that behaved like androstenedione (Ad) in six different 1456 A.P. Scott / Steroids 77 (2012) 1450–1468

choice of procedure might or might not have affected the results of any of the studies.

5.4. The problems of low (and unreported) yields

Not all studies have quoted yields (indicated by ‘?’ in Table 2). In some that have, the conversion rates for critical points in the pathway are so low (<0.01–0.1%) that the question has already been asked [29,32] as to whether this could possibly result in production of reported concentrations of vertebrate-type steroids in mollusks. In A. californicus, especially [21], the two key steps (transformation

of 3b-D5 ? D4 steroids and of C21 ? C19 steroids), both had <0.04 yield (which would in effect mean that 1 mg of Preg would be needed to produce 1 ng of Ad! One of the most often-quoted studies is that on the sea hare, Aplysia depilans [33,34]. This study showed that the whole range of human steroids (including cortisol) appeared to be formed from 14C-labeled acetate, cholesterol, Preg and P by incubation of acetone-dried powder of gonads and hepatopancreas. Problems with this study are that the methods are not described in enough detail to even know whether the authors were using fresh tissue or acetone-dried powder for the incubations, the yields were not reported and identiﬁcation was by recrystallization only. Results in that paper, furthermore, indicate the presence of an impossible product – namely ‘conjugated P’. Conjugation requires the presence of a reactive hydroxyl group (see Section 6) and P has no hydroxyl groups in its structure.

5.5. Do mollusks have cholesterol?

The raw material from which vertebrate steroids are made is cholesterol. This is found in mollusks in quantities that are more than sufficient to act as a precursor for steroid synthesis if so required [35]. The only debate is about which species are able to syn- Fig. 1. Biosynthetic pathway of the main vertebrate steroids that are discussed in thesize their own cholesterol de novo or have to rely on obtaining it this review – starting from pregnenolone (that is formed directly from cholesterol). from their diet (a question that does not particularly matter to us, Steroids: Preg, pregnenolone; P, progesterone; 17-Preg, 17-hydroxypregnenolone; 17-P, 17-hydroxyprogesterone; DHEA, dehydroepiandrosterone; Ad, androstenedi- as either source could be used for steroid biosynthesis). In the one; T, Testosterone; E1, estrone; E2, 17b-estradiol. Enzymes: D5-3b-HSD, D5-D4- Rayed Mediterranean limpet, Patella caerulea, and turban shell, + isomerase) or 3b-hydroxy-D5-steroid:NAD 3-oxidoreductase); 17-OHase, 17- Monodonta turbinata, it was shown [36] that a high proportion of hydroxylase; 17b-HSD, 17b-hydroxysteroid dehydrogenase; CYP19, aromatase. the injected 14C-acetate was converted into cholesterol. This was Note that 17-hydroxylase and 17,20-lyase activities are both found together on the one enzyme, CYP17A1. also shown in Viviparus fasciatus and L. littorea [37]; A. californicus [38] and A. ater rufus [37]. However, similar experiments with the whelk, Buccinum undatum [39], and cuttlefish, Sepia officinalis [40] chromatographic systems. However, when it came to the recrystal- showed zero incorporation. It was suggested that the ability to syn- lization step, only 0.014% of this activity appeared to behave like thesize cholesterol depended on whether the animals were carni- Ad. The rest of the activity could not be identified. vores or not [39]. Even the technique of ‘recrystallization to constant specific It has been speculated that mollusks may use a sterol other than activity’ is not without its potential problems. If the standard ste- cholesterol as the starting material for steroid synthesis [35]. roids are allowed to precipitate (rather than crystallize) this can Although several other sterols are known to occur in mollusks give the appearance of homogeneity between labeled compound [35,41], there is no evidence yet that they are able to be converted and standard. Most steroid chemists will claim indignantly that to pregnenolone (i.e. it is still a matter of speculation). they know what they are doing when carrying out crystallizations (and this is almost certainly true). However, very few people pro- 5.6. The first key step in steroid biosynthesis – the cleavage of the side vide evidence that it was actually crystals rather than precipitates chain of cholesterol that they produced during their procedures. A key reaction in the synthesis of vertebrate steroids is the cleavage of the side chain of cholesterol (Fig. 2) between C-20 5.3. Incubation procedures and C-22, starting with the addition of an oxygen atom on C20, to form Preg (a 21 carbon compound). The enzyme that performs In terms of the methodology used to produce the radioactive this reaction is now commonly referred to as CYP11A1 (with steroids, there are big differences between studies and it is difficult ‘CYP’ standing for cytochrome P450). It is difficult, if not impossi- to tell what influence this had on the types or yields of steroids that ble, to demonstrate the activity of this enzyme by adding radioac- were formed. Steroid precursors have been variously incubated tive cholesterol to tissue explants in vitro because cholesterol using live animals, tissue fragments, minced tissue, thawed tissue, (unlike the other steroids) will not diffuse readily into the cells acetone-dried tissue and microsome preparations. Since no proce- and also needs to be actively transported into the mitochondria dure is accepted as a standard, it is impossible to determine how (where the enzyme is located in vertebrates). This is perhaps the A.P. Scott / Steroids 77 (2012) 1450–1468 1457

Table 2 The evidence for biosynthesis of vertebrate steroids by mollusks (using 3H- or 14C-labeled steroid precursors).

Species D5-3b-HSD 17-Hydroxylase 17b-HSD Aromatase (CYP 19) Other enzymes Methodology plus 17,20-lyase Eastern oyster, T ? Ad, 14%; Sperm; XT Crassostrea virginica E2 ? E1, 93% [136] Atlantic sea scallop Preg ? 17-P or Ad, no 17-P ? Ad, 0.04% Ad ? T, no Gonad or Placopecten conversion (gonad, but not conversion hepatopancreas magellanicus [31] hepatopancreas) (minced tissue); PC, TLC, XT Banana slug, Ariolimax Preg ? P, <0.04%; Preg ? 17-preg, Ad ? T, no No conversion of Ad to any Ad ? 5a-reduced Ovotestis californicus [21] DHEA ? Ad, 0.02% 0.01%; 17- conversion; 3a- estrogens androgens, 8%; (mitochondrial- P ? DHEA, 0.1%; A(5a) ? 3a,17b- T ? 5a-reduced microsome fraction); DHEA ? A, 0.02% A(5a), 16% androgens, 60% PC, TLC, DF, XT Mussel, Mytilus edulis Preg ? P, 0.4%; 17- 17-P ? A, 0.2% Ad ? T, 2%; Ad ? 5 a -DHT, 0.4% Male and female [42,43] Preg ? 17-P, 0.25% E1 ? E2, 50% gonads (homogenate); TLC, DF, XT Common periwinkle, P ? 17-P, no P ? 20b-P, 8% Male and female Littorina littorea conversion gonads (homogenate); [137] PC, DF, XT Common slipper shell, DHEA ? Ad,<0.1% 17-P ? Ad,<0.1% A ? T, 20%; No conversion of Ad to any Male-phase ovotestis Crepidula fornicata T ? Ad, 5% estrogens (teased tissue); PC, DF, [138] XT Common whelk, P ? 17-P, Ad and T, Hepatopancreas; TLC, Buccinum undatum some, but ‘very low’ PC, MC (failed XT) [139] (?%) Sea hare, Aplysia Preg ? P, yes (?%) Preg ? DHEA, yes Preg and P ? T, Gonads and depilans [33,34] (?%); P ? 17-P and yes (?%) hepatopancreas Ad, yes (?%) (acetone-dried powder); TLC, XT Cuttleﬁsh, Sepia Preg ? P 17-Preg ? 17- P ? 17-P, 0.5%; 17- Ad ? T, 1.5% T ? E2, 3% (ovaries only) Testes and ovaries ofﬁcinalis [40] P; 17-DHEA ? Ad, 0.5% Preg and (teased tissue); PC, Preg ? DHEA, 4– TLC, DF, XT 10% Edible snail, Helix Preg ? P, 2–4% by Ovotestis pomatia [47] ovotestis; 0.1–0.6% by (homogenate); TLC, dorsal body; DHEA ? A, DF, XT 2–6% Great pond snail, Preg ? P, 7–14% P ? 17-P or Ad, no Ovotestis Lymnaea stagnalis conversion (homogenate of frozen [48] tissue); TLC, DF, XT Garden snail, Helix T ? Ad, > 50%, Ad ? E3 (but not E1 or E2), Ad ? 5a-reduced Gonads (homogenate aspersa [82] (predominant in 0.8% androgens,>30% of frozen tissue); TLC, juveniles) (predominant in DF, XT adults) Common periwinkle Ad ? E1 and E2,<0.01% T ? 5a-DHT, ‘good Hepatopancreas Littorina littorea yields’ including gonads [140] (microsomes);TLC Sea slug, Clione P ? 17-P or Ad, no Ad ? T, no No conversion of Ad to any P ? P(5a) and 3a- Whole animals; TLC, antarctica [30] conversion conversion estrogens P(5a), 53%; DF, XT Ad ? A(5a) and 3a- A(5a), 45%

Scallop Patinopecten [1b-3H]Ad ? H2O. 3.2% yessoensis [50] Grooved carpet shell, T ? Ad, ?% T ? E1 and E2, ?% T ? 5a-reduced Hepatopancreas Ruditapes decussates (dominant androgens, ?% (microsome prep. of [121] metabolite) frozen tissue); TLC Mud snail, Ilyanassa T ? 5a-DHT, ?% Intact animals TLC obsoleta [141]

Purple dye murex, [1b-3H]Ad ? H2O, activity Bolinus brandaris detected [130] Paciﬁc oyster, Ad ? steroids with same Gonads (homogenate) Crassostrea gigas mobility as E1 and E2 on HPLC, TLC [55] TLC, <0.1%;+ve; [1b-

3H]Ad ? H2O Eastern oyster, T ? Ad; E2 ? E1 Sperm PC, DF, XT Crassostrea virginica

(continued on next page) 1458 A.P. Scott / Steroids 77 (2012) 1450–1468

Table 2 (continued)

Species D5-3b-HSD 17-Hydroxylase 17b-HSD Aromatase (CYP 19) Other enzymes Methodology plus 17,20-lyase [74] Giant ramshorn snail, Ad ? T; T ? Ad Ad ? 5a-DHA and 5a- Gonad/ Marisa cornuarietis DHT hepatopancreas [142,143] (microsome preparation); TLC, HPLC, GC–MS Banded dye-murex, Ad ? T, 8% Gonad/ Hexaplex trunculus hepatopancreas [144] (microsomes);HPLC Purple dye-murex, Ad ? T, 10% in Ad ? 5a-reduced Bolinus brandaris females,0%in androgens, 80% in [144] males) females and 98% in males

NB, conversions in bold type indicate those with ‘appreciable’ (>2%) yields. Abbreviations used: 17-P, 17-hydroxyprogesterone, 17-hydroxypregn-4-ene-3,20-dione; 20b-P, 20b-hydroxypregn-4-ene-3-one; 3a,17b-A(5a), 5a-androstane-3a,17b-diol; 3a-A(5a), 3a-hydroxy-5a-androstan-17-one; 3a-P(5a), 3a-hydroxy-5a-pregnan-20-one; A, androstenedione, androst-4-ene-3,17-dione; DF, Derivative formation (microchemistry); DHEA, dihydroepiandrosterone, 3b-hydroxyandrost-5-en-17-one; E1, 17b-estradiol, (17b)-estra-1,3,5(10)-triene-3,17-diol; E2, estrone, 3-hydroxyestra- 1,3,5(10)-trien-17-one; E3, estriol, estra-1,3,5(10)-triene-3,16 [a], 17 [b]-triol; HPLC, high performance liquid chromatography; P(5a), 5a-pregnane-3,20-dione; P, progesterone, pregn-4-ene-3,17-dione; PC, paper chromatography; Preg, pregnenolone, 3b-hydroxypregn-5-en-20-one; T, testosterone, 17b-hydroxyandrost-4-en-3-one; TLC, thin layer chromatography; XT, re-crystallization to constant speciﬁc activity.

explanation of why M. edulis failed to convert cholesterol to preg- against the vertebrate enzyme [50]. Since in situ staining proce- nenolone [42,43] or why S. officinalis [40] apparently did so, but dures can never be 100% specific (plus, in all cases, the data pro- with very low yields (<0.01%). However, a more likely explanation vided are qualitative), this type of evidence cannot be accepted is that this enzyme does not exist in mollusks. Despite a report that as firm proof of the specific presence of this enzyme. mRNA extracted from Mytilus spp. was able to hybridize with a probe based on human CYP11A1 [44] and another report that an 5.8. C17,20-lyase (column 3 in Table 2) antibody against rat CYP11A1 was able to bind a protein in a particular cell-type of the digestive gland of the mussel, Mytilus gallo- There are in fact two key events associated with this enzyme. provincialis [45], genomic studies suggest that the gene for The first involves the insertion of an oxygen at the C-17 position CYP11A1 only evolved in the vertebrate line [46]. It is of course of P or Preg to form 17-P and 17-hydroxypregnenolone (17-Preg), possible that some other enzyme may have subsumed the role of respectively (Fig. 1); and the second involves the removal of the CYP11A1 in mollusks. However, one has to bear in mind that the side chain (C-20 and C-21) to form Ad or DHEA, respectively. The genes for most of the other critical events in steroid synthesis also insertion of an oxygen atom on the C-17 is an essential step, as this appear to be missing [46] from the genomes of invertebrates that atom will become the 17-oxo group of Ad or DHEA. Without an are not in the direct vertebrate line. Thus what are the chances that oxygen atom, the side-chain cannot be cleaved to form an andro- a substitute for CYP11A1 would evolve in mollusks if they did not gen (either by an enzyme or by chemical means). It has for a long have the enzymes to further transform Preg into P, T and eventu- time been thought that there was only one enzyme in vertebrates ally E2 (unless, although unlikely, a whole suite of enzymes to per- that performed these combined actions, and that the formation of form these activities evolved independently)? 17-hydroxylated C21 steroids such as cortisol and 17,20b-P was due to selective inhibition of the lyase activity. However, an en- 5.7. D5-3b-HSD activity (column 2 in Table 2) zyme has recently been characterized, from a fish, that specifically

Cholesterol is unsaturated (i.e. there is a carbon double bond) between C-5 and C-6 (Fig. 2). It also has a hydroxyl group at position C-3. The four-ring skeleton is unaffected when cholesterol is converted to Preg. However, when Preg is converted to P (or dehydroepiandrosterone [DHEA] to Ad), the hydroxyl group is converted to an oxo group (by removal of two hydrogen atoms) and the position of unsaturation moves to between C-4 and C-5; form- ing what are often referred to as ‘D4-3-one’ (or 4-pregnene) steroids. Most of the vertebrate steroids that are used as hormones have this conﬁguration – such as P, T, KT, 17,20b-dihydroxy- pregn-4-en-3-one (17,20b-P), cortisol and corticosterone. Two studies on two different mollusk species [47,48] have reported relatively high levels of this enzyme activity. Six other studies

(see Table 2), however, have reported very low or no activity. D5- 3b-HSD activity been histochemically demonstrated in tissues of maturing Pacific oyster, Crassostrea gigas (specifically in what are described as ‘elongated epithelioid tissue adjacent to adductor muscle and visceral ganglia’ [49]; and in isolated cells of the testis of O. vulgaris [9]. Scattered single cells of the ovaries of the scallop, Fig. 2. Basic structure of cholesterol, with the positions of the carbon atoms Pecten yessoensis, have also been immunostained with an antibody numbered. The same numbering system is used for all the vertebrate steroids. A.P. Scott / Steroids 77 (2012) 1450–1468 1459 carries out 17-hydroxylation of P without any subsequent cleavage specific the activity is (i.e. could the weakly positive results have to form Ad [51]. It is likely that it is this enzyme that is operative other explanations such as misidentification of the products, con- when vertebrates make C21 hormones such as cortisol and tamination and actions of unrelated metabolic enzymes?). Using 17,20b-P. For the purposes of this review, evidence for the forma- antibodies against mammalian CYP19, immunohistological stain- tion of 17-P and 17-Preg is grouped with evidence for the forma- ing has been demonstrated in scattered single cells of the ovaries tion of Ad and DHEA from either P, 17-P, Preg or 17-Preg). There of P. yessoensis [50] and in scattered cells of the testis of the same is only one study in mollusks [40] that has reported good yields species [56]. However, non-specific cross-reactivity cannot be ru- for this activity. The other nine papers (column 3 in Table 2) have led out. The most powerful argument against the presence of reported very low or negative yields. CYP19 in mollusks is that the gene only first appeared in a direct ancestor of the chordates [57,58]. 5.9. 17b-hydroxysteroid dehydrogenase (17b-HSD) activity (column 4 in Table 2) 5.12. A strong streak of anthropomorphism in mollusk studies

Although the evidence for the existence of D5-3b-HSD, 17- hydroxylase and C17,20-lyase activity in molluscan tissues is weak, As pointed out by Markov and colleagues [46], all the biosyn- the existence of 17b-HSD activity (i.e. the conversion of Ad ? T and thetic studies that have been carried out in mollusks have looked estrone [E1] ? E2, or vice versa) is very strong (13 studies). The specifically for transformations that are essentially part of the hu- yields are high, and the observations have been repeated and inde- man steroid biosynthetic pathway (i.e. asking ‘can Preg be con- pendently verified in a few mollusk species. The existence of this verted to P; can P be converted to 17-P; can 17-P be converted to activity, however, does not necessarily mean that mollusks make Ad; can Ad be converted to T; and can T be converted to E2?’). vertebrate steroids, or that the enzyme is the same enzyme as used Any other products of the incubations have (with a few exceptions) by vertebrates. Such transformations (that are assisted by the co- been ignored. This is perhaps not surprising, as it is rather difficult enzyme nicotinamide Adenine dinucleotide) could be (and proba- and expensive to identify a steroid for which one does not already bly are) incidental. The reaction is a relatively simple oxidoreduc- have some sort of an idea of what it might be. The normal identi- tion reaction (either adding two hydrogens or removing two fication procedures for steroids (thin layer chromatography, micro- hydrogens from an exposed oxo group) similar to the conversion chemical conversion and recrystallization to constant specific of pyruvate to lactate (and vice versa) and there are many such en- activity) all rely on having synthetic standard steroids with which zymes that catalyze this type of reaction in many biochemical to compare the radioactive steroids that are formed in the incuba- pathways in both the animal and plant kingdoms [52]. HSDs de- tions. Unsurprisingly, those that are commercially available are al- rived from bacteria, for example, are used widely in the laboratory most all variants of known vertebrate steroids. An example of what as reagents for steroid conversions, such as preparing 17,20b-P happens when an incubation does not generate immediately rec- from 17-P [53]. ognizable steroids is illustrated by the study on P. magellanicus [31]. In some experiments within this study, the gonads and hepa- 5.10. 5a-reductase (column 6 in Table 2) topancreas of this animal converted up to 70% of added radiolabeled Preg and 17-P into several metabolites. However, David Quite a few mollusks (nine studies) appear able to reduce the Idler, who was arguably the greatest steroid chemist of his time double bond between C-4 and C-5 of Ad, T or P and convert them (he discovered both 17,20b-P and 11-KT in teleost fish) was unable to 5a-reduced steroids (see also [28]). Although 5a-reduction of to identify a single one of these (apart from a trace of Ad)! the A ring occurs in many vertebrates (and is important in the hu- The same criticism of latent anthropomorphism also applies to man male for the production DHT from T), it does not only affect studies in which steroids have been measured in mollusks (Table 1). steroids. The bile acid-derived compounds that lampreys use as The vast majority of studies appear to have been concerned with pheromones are also 5a-reduced [54] and these are ‘sterols’ as op- concentrations of what are essentially human steroids (i.e. T, E2 posed to ‘steroids’. Like the presence of 17b-HSD activity, the pres- and/or P). Why not 11-KT and 17,20b-P, that are, respectively, ence of 5a-reductase activity does not in any way imply that the main androgen and the main progestin in teleosts [59]?Or mollusks make vertebrate steroids, nor that such steroids are their why not 15a-hydroxprogesterone that is the dominant progestin normal substrate. in the blood plasma of mature males of the sea lamprey, Petromy- zon marinus [60]? And when measuring vertebrate stress steroids [17], why choose cortisol (the human stress steroid) rather than 5.11. CYP19 (aromatase; column 5 in Table 2) corticosterone (which is the stress steroid in many other vertebrates as diverse as rodents and amphibians)? E2 is made from T, and E1 from Ad, by a rather complex reaction that involves an enzyme called aromatase. There are two ways to demonstrate aromatase activity in tissues. One is to add radiolabeled Ad or T to the tissues and then establish whether they have 5.13. Interim conclusion been transformed to radiolabeled E1 and/or E2. The other is to add T that has been labeled with tritium at the C-1 position and then to It is clear that mollusks are able to carry out transformations on measure how much is converted into water (‘released’) during the vertebrate steroids. However there is no solid evidence that any of reaction. Using the first procedure, three studies found no conver- these transformations indicate the presence of a pathway involved sion of androgens to estrogens, three found <0.01% conversion, one in the formation of vertebrate-type steroids; nor indeed of path- found 3% conversion, and one found apparent conversion of Ad to ways involved in the formation of ‘mollusk-type’ steroids. The estriol but not to E1 or E2 (see Table 2 for references). All ‘tritium transformations can all be interpreted as ‘metabolism’ – though release’ assays have given positive signals that have been described whether any of the transformations are ‘deliberate’ (i.e. part of a as close to background measurements [28].InC. gigas, the aroma- pathway specifically geared to the metabolism of vertebrate ste- tase activity was noted to be only 0.1% of that found in a similar roids of exogenous origin) or ‘incidental’ (i.e. the steroids are acted amount of bovine tissue [55]. With most studies suggesting that on non-specifically by general metabolic enzymes) is as yet aromatase activity is either absent or very low, one wonders how uncertain. 1460 A.P. Scott / Steroids 77 (2012) 1450–1468

6. The animals pick up the steroids from the environment? their levels of free E2 from 0.3 to 2.5 ng gÀ1 and of esterified E2 from <1 to 17 ng gÀ1; the zebra mussel, Dreissena polymorpha 6.1. Possible sources of steroids in the environment [72], in which 14C-E2 that had been added to the water was shown to bioaccumulate over a period of 13 days by a factor of 840 in If one accepts the evidence that T, E2 and P definitely are pres- males and 580 in females, and when the animals were transferred ent in mollusks, then the only other possible source is from the to freshwater, the activity was retained (with no apparent loss of environment (i.e. water, sediment and food). Two major inputs of activity) for at least 10 days, and also, when the tissues were finally vertebrate steroids into the environment (especially in rivers) are extracted, the esterified E2 was found not to have been metabo- sewage treatments works [61] and farm animal effluents [62]. lized in any way; M. galloprovincalis [73] in which it was shown However, any free-living wild vertebrate is a potential source of that, when T was added to water for 5 days, it was taken up in sub- steroids. For example, in fish, there is clear evidence that there is stantial amounts – and esterified T increased in a dose-dependent continuous traffic of steroids across the gills – as well as frequent manner, while the levels of free steroid were relatively unaffected; release via the urine and feces [63–65]. Overall, the majority of ver- the eastern oyster, Crassostrea virginica [74] and the Giant rams- tebrate steroids probably find their way into the environment via horn (Apple) snail, Marisa cornuarietis [75], in which it was shown urine and feces. Although, when they are excreted via these routes, that levels of esterified (but not of free) E2 also increased in a dose- they are mainly present as ‘biologically inactive’ glucuronide or dependent manner; M. edulis again [76], in which 14C-E2 was sulfate conjugates, many organisms appear to contain enzymes shown to reach a bioconcentration factor of 2500 after 13 days that can readily convert them back into ‘free’ steroids. Of particular and have a half-life of depuration of 8.6 days, and all the activity relevance to this review, one of the most abundant sources of such was all associated with a single peak on HPLC that, after base enzymes (which are frequently used in the laboratory for hydroly- hydrolysis, was reconverted to pure E2. In this last study, when sis of vertebrate steroid conjugates) is mollusks! Preparations the animals were exposed to 14C-E1, they also readily absorbed this made from the hepatopancreas of the edible snail, Helix pomatia steroid (with a bioconcentration factor of 1000 after 8 days). How- and the keyhole limpet, Patella vulgata, are sold commercially by ever, when the esters were subjected to base hydrolysis, the ste- Sigma–Aldrich Inc. under the headings of ‘glucuronidase’ and ‘sul- roid that was released was 14C-E2. fatase’. In other words, even if the steroids in the environment are The above-mentioned studies on D. polymorpha [72] and M. present as conjugates, some, possibly all, mollusks have a strong edulis [76] are particularly interesting, because, in the first case, capability of converting them back to free steroids. Bacteria in sew- the authors found from sampling of animals in the wild, that age treatment works are also known to convert glucuronidated although the sediment in which the animals lived contained far estrogens back into their free and biologically active forms [66]. more E1 than E2, only esterified E2 could be detected in tissues; Since every vertebrate is a fairly continuous emitter of steroids and, in the second case, the authors found, as just mentioned, that (whether free or conjugated), it is unlikely that there are many if they added 14C-E1, it was converted to and then stored as 14C-E2. parts of the aquatic environment which do not contain some ver- These results indicate that there is something about E1 that pre- tebrate steroids (whether free or conjugated). This point has been vents it being conjugated to fatty acids. The only difference be- made previously in relation to chemical pollution [1] i.e. that there tween E1 and E2 is that the latter has a C-17b-hydroxyl group is probably nowhere on earth that has not been affected by hu- rather than a C-17-oxo group. In order for a fatty acid to be com- mans. Essentially, any statement to the effect that ‘the animals bined with a steroid, a molecule of water must be formed and it were collected from an unpolluted site’ can probably be described is a pre-requisite that the steroid has a hydrogen atom to donate as ‘wishful thinking’. to the reaction (Fig. 3). Because the @O group on position 17 of E1 does not have a hydrogen atom, conjugation is impossible. Ste- roids such as P and Ad that do not have any hydroxyl groups at all 6.2. Experimental evidence that mollusks can take up and store can also not be conjugated. In theory, E1 could be conjugated via its exogenous vertebrate steroids 3-hydroxyl group. However, this is an unusual group in that it is attached to an aromatic ring; and the fact that no conjugated E1 In laboratory experiments, it has been shown (and indepen- was found in the study on M. edulis indicates that it does not take dently verified) that mollusks have a remarkable ability to absorb part in conjugation reactions, in this species at least. Other steroid and retain at least two vertebrate steroids (namely T and E2). Fur- hydroxyl groups that are known to be unreactive to conjugating thermore it has been discovered [67] that a key part of the mech- enzymes in vertebrates (and the same would probably be the case anism that allows them to retain these steroids is by conjugating in mollusks) are those at position C-11b of cortisol and at position them to fatty acids (by a process referred to as ‘esterification’; C-17 of 17-P. However, the C-3b-hydroxyl group of DHEA is one Fig. 3). The initial study on the eastern mud snail, Ilynassa obsoleta that is reactive in mollusks, as this steroid was readily esterified [67] showed that it took one animal only eight hours to remove in vitro by microsome preparations of digestive glands and gonads 80% of 14C-labeled T from 3 ml solution. When the animal was of C. virginica [74]. Incidentally, in M. cornuarietis [75] it was shown transferred to a clean solution, hardly any of the label was released over 90 min. The activity was found to be associated with apolar (i.e. lipid-like) compounds. It was further shown that these compounds could be made in vitro and that their production could be enhanced by addition of palmitoyl coenzyme A. Subsequent studies [68,69] showed that exposure of I. obsoleta to T via injection or water had hardly any effect on free T levels, but had a dose- dependent effect on concentrations present as fatty acid esters. Other studies that have demonstrated the ability of mollusks to Fig. 3. The mechanism of the esterification reaction. The steroid atoms are shown absorb steroids have been made using: C. gigas [70], in which indi- in bold type. Note that the steroid must have a hydroxyl (also known as an ‘alcohol’ viduals were shown to bioaccumulate 14C-E2 from water, with a group) in order to participate in the reaction. Also note that some steroid hydroxyl groups cannot participate in conjugation reactions at all, due to steric hindrance bioaccumulation index (the ratio of activity in 1 g tissue to activity (e.g. that attached to C-17 of 17-hydroxypregn-4-ene-3,20-dione) and, for others, in 1 ml water) of 30 after 48 h; M. edulis [71], in which individuals there might not be an appropriate enzyme (e.g. the C-3 hydroxyl group of 17b- 1 that had been exposed to 200 ng LÀ of E2 for 10 days increased estradiol and estrone in mollusks [76]). A.P. Scott / Steroids 77 (2012) 1450–1468 1461 that rates of esterification of E2 and T were correlated, supporting the animals have to be maintained by humans. Humans are not the existence of a single enzyme that in effect works on an identi- inconsiderable synthesizers of P, T and E2 – the three steroids that cal C-17b-hydroxyl group in both steroids. have been mainly studied in mollusks. It is well known that humans Working on the basis that T that might be responsible for stimu- (like all vertebrates) excrete steroids via the urine and feces. How- lating penis development in female mollusks (a hypothesis ex- ever it is not so well known that they also excrete steroids via saliva, plained in detail by Matthiessen and Gibbs [77]), several sweat and skin [83–85]. Even if mollusks (and their food) are at all investigators have looked at whether the way in which TBT exerts times handled with gloves, it still has to be taken into account that its effect on penis formation is perhaps by inhibition of the esterifi- humans are continuously shedding their skin and hairs. In fact, hu- cation of T – thereby leading to an increase in the levels of free T. man ‘skin scales’ (potentially contaminated with steroids) are a ma- However, most studies, either in vivo [68,75,78] or in vitro [79] show jor constituent of airborne dust in most buildings [86]. Even though only small effects of TBT on esterification. Only one in vitro study on the amounts of dust and associated steroids entering rearing tanks I. obsoleta [80] reported relatively pronounced effects of TBT. might be relatively low, the proven ability of mollusks to bioaccu- There is further (albeit circumstantial) evidence from field stud- mulate vertebrate steroids would probably ensure a build up of ies that indicates that mollusks tend to pick up steroids from the measurable quantities over the many months that mollusk test spe- environment. For example, in P. antipodarum that had been caged cies are normally kept in the laboratory. Another obvious potential in rivers, steroid concentrations in individuals always increased source of steroids for mollusks is their food – especially if the food with time, irrespective of the external conditions. There was a 5- takes the form of ‘organic lettuce’ that may well have been fertilized fold increase in total (i.e. free and esterified) amounts of T and P with animal (vertebrate) manure. in animals caged downstream of a STW for 21 days [15]. In another As confirmation that steroids are present on human skin, volun- study [14], animals were also placed in cages upstream and down- teers at the Centre for Environment, Fisheries and Aquaculture Sci- stream of STWs and total levels of E2, T and P again increased ence were asked to dip one of their arms (up to the elbow) into a markedly (5- to 10-fold) with time at all downstream sites (and plastic bag containing either 1 L or 1.5 L of water for 1 min. The to a lesser extent upstream). The same increase with time, but with water was then extracted and assayed for cortisol, T and E2. In no link to the conditions being tested (in that case metal contam- the first trial with 29 participants, the amount of cortisol extracted ination), was noted in yet a third study by the same group [11]. from the 1 min water samples ranged from 2000 to 120,000 pg and Although these increases could possibly have been due to stimula- T ranged from 800 to 6000 pg. In the second trial, with 38 partici- tion of endogenous steroid synthesis by unknown factors in the pants, E2 ranged from 80 to 540 pg, T from 580 to 3400 pg, P from river water, the only common factor in all the experiments was 1500 to 4900 pg and cortisol from 3400 to 32,000 pg. The data ‘time of exposure’. from these trials are still being analyzed (and thus only reported here in summary). They are mentioned ahead of full publication 6.3. What is the point of esterification? purely to make the point that the steroids that we make are not confined within our bodies or totally excreted via either our urine The question everyone who works in this field has not yet been or feces. Some of these steroids appear on our body surfaces – and able to answer is ‘what is the relevance of conjugation of T and E2 to we need to take this into account when thinking about likely fatty acids by mollusks?’. These two steroids are known to be avail- sources of contamination in the laboratory. able exogenously in the environment and both possess the same The hypothesis outlined above that mollusks pick up human/ reactive C-17b-hydroxyl group. For all we know at the moment, mammalian steroids as a result of steroid contamination obviously they may just be two among many other compounds with hydroxyl relies upon those steroids being relatively stable in the environ- groups that are capable of being linked to fatty acid groups; and we ment. In fact, in comparison to many other types of organic com- only know about these two compounds because they are the only pounds that are handled in the laboratory (e.g. peptides and ones that anyone has so far been interested in measuring. In fact, prostaglandins), steroids are very stable indeed. In a paper that is in the pioneering study on I. obsoleta [67], it was shown that, after quoted as evidence that algae make vertebrate steroids [87], the base hydrolysis of apolar material extracted from the animals, most impressive fact is not the identity of any of the metabolites although T was identified among the products, there was a large ex- (there was, incidentally, no evidence of any 17-hydroxylation or cess of free fatty acid (indicating that there were esters of other formation of androgens), but the fact that in all 12 species that compounds in the total mix). If one could get some handle on the were examined, the bulk of the P that was added as a precursor identity of the other compounds that are stored as esters in mol- was recovered intact (i.e. not broken down or metabolized in any lusks, it might give some clue to the function of this mechanism. way) after having been stirred in an aqueous solution, under lighted conditions for 14 days at 24 °C! 6.4. What can explain the presence of steroids in laboratory-reared Having said all the above, it has to be stressed that there is, at the animals? moment, no actual proof of steroid transfer between humans and laboratory stocks of mollusks. It is purely a hypothesis. However, Those researchers that have published the clearest evidence for as a potential explanation for the presence of human steroids in the uptake of vertebrate steroids from the water [10,72,76] freely mollusks, it has more plausibility than the hypothesis that mollusks acknowledge the likely impact of uptake on steroid concentrations have evolved their own alternative biosynthetic pathway for mak- in molluscan tissues, but appear to be reluctant to abandon the ing human-like steroids. Whatever the merits of either hypothesis, concept that the same steroids are also produced endogenously. until cross-contamination has been excluded experimentally in the In order to defend the endogenous origin of steroids in animals kept laboratory, it cannot be assumed that the presence of steroids in in the laboratory, arguments are used such as: even though col- supposedly ‘clean’ laboratory stock animals necessarily means that lected in the wild, the animals had been kept in the laboratory for the animals must have been able to make the steroids themselves. several months in artificial sea- or fresh-water [81]; the animals had been hatched and grown entirely under clean laboratory condi- 6.5. Looking for steroids that are unlikely to have been formed tions [75]; or the species was land-living [17,82]. Essentially, where endogenously could the steroids have come from if the water was clean (or there was no water) and there were thus no upstream effluents or fish to One possible way to examine the problem of potential cross- worry about? The answer may lie (it is suggested) in the fact that contamination in laboratory stocks would be to look for the pres- 1462 A.P. Scott / Steroids 77 (2012) 1450–1468 ence of steroids that one would not expect the mollusks to be able aromatase (CYP19) and the cholesterol side chain cleavage enzyme to make themselves. A potential candidate would be cortisol (that, (CYP11A1) only evolved in the immediate ancestors of the verte- as shown above, is found in substantial levels on human skin). brates [46,58]. How come then that the ancestral steroid nuclear However, it would first need to be shown that mollusks have the receptor was an estrogen receptor? This anomaly has already been ability to bioconcentrate and accumulate this steroid in the same noted by Baker [95,96], who has suggested that, while the primitive way as they do with T and E2. Another candidate would be EE2 nER may well have been able to bind to estrogens, it probably, at that (which is a part of ‘the pill’ and does not occur in nature at all). time in evolution, had a ligand that was not actually E2 (as it is to- In fact, there are two studies that have already convincingly dem- day). This is entirely plausible. One interesting fact about the verte- onstrated the presence of EE2 in the flesh of wild-caught mollusks brate nER (of which the human b form has received the most [7,88]. In the second of these studies, five mollusk species (includ- attention) is its relative promiscuity compared to the other steroid ing the abalone, Haliotis diversicolor supertexta) from a bay in China nuclear receptors. There are estimated to be hundreds, probably were analyzed and, while E2 levels were below the levels of detec- thousands, of compounds in the environment (both natural and tion, EE2 levels were an astonishingly high 80,000–130,000 pg gÀ1 man-made) that are able to activate the human nERb (admittedly of dry weight of EE2 in all five species. Subject to independent mostly with many times less potency than E2). Such compounds verification, these findings are, at the very least, strong positive (examples are alkyphenols, BPA, genistein, some DDT metabolites) evidence that at least some (if not all) of the steroids that have are very diverse in their structure and most of them look nothing like been measured in mollusks are of exogenous origin. typical estrogens [97,98]. Quoting from Barnes [97]: ‘in general, pla- nar compounds with a phenolic character and two oxygen-containing moieties approximately 11–12 Å apart can fit into the binding 7. The evidence for steroid receptors in mollusks pocket’ of the nER. Thus while Baker has suggested (see above) that the primitive ligand might have been an androgen or a sterol derived 7.1. Presence of classical nuclear receptors from cholesterol, the present author suggests that it might not even have been a steroid at all. However, this is pure speculation. Most, if not all, studies carried out on steroid receptors in mol- This brief (and rather oversimplified) potted history of the evo- lusks to date (Table 3) have been based on the assumption that, if lution of nuclear steroid receptors does have important implica- such receptors exist, they are so-called ‘classical’ nuclear receptors. tions for the interpretation of receptor studies in mollusks. There These are proteins that have a characteristic structure – a highly are a few studies that have reported the presence of steroid bind- conserved DNA-binding domain and a moderately conserved li- ing in mollusk tissues (Table 3). Also, in some of these studies, the gand-binding domain – that makes them easy to recognize within binding appears to do everything that a classical nuclear receptor the genomes of animals; and analysis of proteins with such struc- would be expected to do [99,100], including binding to DNA. How- tures in the genomes of a variety of animals [89] suggests that ever, if the genomic studies are to be believed, this cannot be so – there were 25 ancestral nuclear receptors in Urbilateria (the hypo- because the nAR and nPR should not be present in mollusks and thetical ancestor of all animals with bilateral symmetry). These 25 the nER is unresponsive to estrogens. So what explanations could receptors have diversified (variously being replicated and lost) in there be? the animal kingdom to serve as mediators for a large array of different compounds including vertebrate steroids. In relation to ste- 7.2. Another class of nuclear receptor? roids, the present thinking is that one of these ancestral nuclear receptors was and always has been a receptor that was able to bind This is a distinct possibility. In humans, a different ancestral nu- to estrogens [90,91]. The ‘classical’ nuclear receptors for proges- clear receptor lines has given rise to a protein that is able to bind a tins, androgens and other vertebrate steroids, however, are be- wide range of compounds including progestins, corticosteroids, lieved to evolved from the nER only after the divergence of the bile acids and E2 [101]. In response to these compounds, this bind- vertebrates [57,92]. ing protein, called the Pregnane X Receptor, activates genes specif- If this scenario is correct, then one would expect to find the ically involved in their metabolism and excretion (i.e. as is gene for nER in mollusks, but no genes for nuclear progestin recep- discussed in Section 7.5, it appears to have a role as a ‘sensor’ tor (nPR) or nuclear androgen receptor (nAR). The facts do indeed rather than a ‘receptor’). appear to match the expectation. There is a gene in all mollusks The ability of a binding protein to bind to DNA–cellulose is usu- that have so far been studies that is the undoubted analogue of ver- ally accepted as proof that that particular protein is a nuclear tebrate nERs (publications listed in Table 3) and no-one has yet receptor [102]. Since the binding proteins for P and E2 in O. vulgaris identified any genes that match nuclear receptors for other verte- bind to DNA–cellulose [99,100], this does indeed suggest that there brate steroids. However, despite the presence of nER-like proteins may be another class of nuclear receptors in mollusks that are able in mollusks, the ligand-binding domains appear to be unable to to bind steroids. However, one must probably be a bit cautious in bind to estrogens (i.e. they are structurally like nERs but are not the interpretation of such data. The procedure for characterizing functional nERs) [93,94]. The mollusk nER still binds to DNA and binding proteins in tissues is very crude. The tissues are essentially activates genes, but because it does this without the need for any mixed with a buffer and then turned into a ‘cold soup’ that is then ligand at all, it has been termed a ‘constitutive receptor’. The theory separated, purely by centrifugation steps, into what are termed the that the ancestral steroid receptor was an ‘estrogen’ receptor has ‘cytosolic’, membrane’ and ‘nuclear’ fractions. It is these relatively gained support from the demonstration that the earthworm (not crude fractions that are tested for binding and put through in the line of evolution of vertebrates or mollusks) has an nER that DNA–cellulose columns. As one can imagine, the binding proteins is able to bind to estrogens [90]. In other words, the nER in mol- form a tiny proportion of the total organic content of these fractions, lusks is more likely to have ‘lost its ability to recognize estrogens’ and who really knows what interactions there might be with other rather than ‘not evolved the ability to recognize estrogens’. compounds that might give the appearance of a protein being able Alert readers will almost certainly have picked on an anomaly. to specifically bind to DNA. Another problem that one must watch The author has argued above that there is no firm evidence (only out for is that a binding protein might have the appearance of being ‘poor’ and at best ‘circumstantial’ evidence) that mollusks biosyn- soluble (i.e. it is present in the cytosol fraction), but on closer thesize vertebrate steroids. Furthermore, the genomic evidence examination, turns out to be ‘membrane-bound’. This was found shows that two critical enzymes necessary to make estrogens, to be the case for the putative ‘androgen receptor’ of the sea A.P. Scott / Steroids 77 (2012) 1450–1468 1463

Table 3 Evidence for presence of steroid receptors in mollusks.

Species (author) Hormone Type of evidence (and data Location and approximate Other findings Comments provided) amounts Giant ramshorn E2 Molecular biology: Cloning In all tissues tested, but slightly Zero binding to E2 snail, Marisa of protein similar to more in penis and sheath (not high cornuarietis classical vertebrate nER in gonads) [145] Giant ramshorn E2 and T Binding activity: specificity Testis and ovary cytosols Testis and ovary had similar binding snail, Marisa only cornuarietis [75] Dogwhelk, Nucella E2 Molecular biology: Cloning 3-fold increase in expression in They comment: structural lapillus [146] of protein similar to ovary after exposure to 0.25% (but similarity to Thais clavigera nER classical vertebrate nER not 1%) effluent suggest it does not bind E2 Eastern mudsnail, E2, T Molecular biology: Cloning Gonad-viscera complex Changes (up to 5-fold) noted by time No gene similar to vertebrate Ilyanassa of protein similar to of year and reproductive stage nAR obsoleta [119] classical vertebrate nER Freshwater E2 Immunoreactive bands on Three bands (55, 75 and 165 kDa) mussel, Elliptio PAGE using human anti- in foot, hepatopancreas and gonads complanata nERb [147] Freshwater E2 Binding activity: affinity, Testis and ovary cytosol Less at one contaminated site than Used E2-BSA-fluorescein mussel, Elliptio capacity, specificity two others complanata [148,149] Marine snail, Thais E2 Molecular biology: Cloning Four times more in ovary and Zero binding to E2 clavigera [94] of protein similar to testis; also present in ganglia classical vertebrate nER Mussel, Mytilus E2 Immunoreactive bands on Hemocytes Involvement of nuclear receptors edulis [104] PAGE using human anti- in mediating rapid effects not nER3a and anti-nERb discussed Mussel, Mytilus E2 Molecular biology: PCR Gonads Recorded large changes in response Not dose-responsive to E2; cf. edulis [150] quantification of mussel to estrogen exposure (more negative response to estrogens in nER mRNA responsive in Feb than Apr) [151] Mussel, Mytilus E2 Molecular biology: Cloning Gonads Not affected by exposure to E2 for edulis [151] of protein similar to ten days classical vertebrate nER Mussel, Mytilus E2 Immunoreactive bands on Positive staining of proteins Also showed that NO release by They suggest: there may two edulis [152] PAGE using human anti- extracted from membranes and pedal ganglia in vitro could be types of receptor (nER and nERb cytosol of pedal ganglia stimulated by E2 coupled to BSA membrane-bound ER) Mussel, Mytilus E2 Molecular biology: Cloning Identity to human nER almost edulis [153] of protein identical to certainly due to contamination human nER Octopus, Octopus E2, T and Binding activity: Affinity, Only P and T binding in testis and Not much binding in non- vulgaris male P capacity, specificity seminal vesicle; P, T and E2 in vas reproductive tissues [9] deferens an prostate Octopus, Octopus P Binding activity: Affinity, Anti-chicken nPR stained band on They comment: characteristics vulgaris female capacity, specificity, DNA- PAGE (70 kDa) and ovarian tissues are those of classical vertebrate [154] binding nPR Octopus, Octopus E2 Binding activity: Affinity, In ovary and oviduct (in cytosol but Anti-human nER stained band on vulgaris female capacity, specificity, DNA- not nucleus); none in PAGE (mass 70 kDa) and in nuclei of [99] binding hepatopancrease or hemolymph ovarian follicle cells Octopus, Octopus E2 Molecular biology: Cloning Extracted from ovary and oviduct Zero binding to E2 vulgaris [93] of protein similar to classical vertebrate nER Pacific oyster, E2 Molecular biology: Cloning In all tissues tested, but most in Zero binding to E2; anti-oyster nER Crassostrea of protein similar to ovary, testis, labial palp and gills immunostained nuclei of follicle cells gigas [155] classical vertebrate nER and oocytes in ovary Scallop, E2 Molecular biology: Cloning They were worried about Placopecten of protein similar to contamination (close to sequence magellanicus classical vertebrate nER of rainbow trout nER) [156] E2 Binding activity: affinity, Testis and ovary (cytosol and Nuclear binding lower in spent than They comment: less specific than capacity, specificity nucleus) ripe females vertebrate nERs

Abbreviations used: E2, 17b-estradiol; kDa, kilodalton; nAR, ‘classical’ nuclear androgen receptor; nER, ‘classical’ nuclear estrogen receptor; nPR, ‘classical’ nuclear progesterone receptor; P, Progesterone; T, Testosterone. lamprey, P. marinus [103]. In this primitive vertebrate, binding sol fraction of the testes. Furthermore it binds to DNA–cellulose. activity for Ad and T is present in impressive quantities in the cyto- However, on closer examination, the activity is found to be part of 1464 A.P. Scott / Steroids 77 (2012) 1450–1468 a neutrally-buoyant lipid complex that had a mass in excess of Any enzyme that is able to modify a steroid should in theory be >1000 kDa. Since the receptor is not actually dissolved, this calls able to bind that steroid temporarily (in order to effect the modifi- into question the relevance of binding to DNA–cellulose. Is it genu- cation). However, although the present author has heard this men- ine or an artefact? Yet another point to bear in mind is that many tioned as a possible explanation for binding activity, he has not people use properties such as ‘affinity’, ‘binding capacity’, ‘half-life found an example to quote. of association’, ‘half-life of dissociation’ and ‘specificity’ in order to argue the point as to whether the binding is ‘typical’ of a nuclear 7.5. Does the presence of a receptor necessarily mean the ligand has to receptor. However, as discussed by Bryan and colleagues [103], be endogenous? there is far too much overlap between the properties of nuclear receptors and other types of binding protein (see below) to be able Even if, after further research, it is found that there actually is a to reach any such conclusions. Only knowledge of the structure of receptor for a steroid in mollusks, the question that then needs to be the binding moiety, plus exhaustive testing of its function, can pos- answered: ‘is this fortuitous (i.e. the receptor evolved to mediate sibly answer the question as to whether it is a receptor or not. the action of an entirely different compound but, by reason of shape and size, the steroid ‘fits’ the binding site) or by design (i.e. the 7.3. Could it be a ‘non-classical’ membrane-bound receptor? receptor specifically evolved to mediate the action of that steroid)?’. If it by design, then yet a further question is ‘does that necessarily The concept of membrane-bound steroid receptors has already mean that the steroid has to be of endogenous origin?’. The ances- been introduced. This is an ever-expanding field in vertebrates, and tors of protochordates and vertebrates appear to have been the first some researchers have already suggested that this may be the way animals to have the capability to synthesize steroids (as opposed to in which steroids exert their effects in mollusks [27]. However, this sterols), and these evolved about 500 million years ago [58]. This is at the moment speculation. Certainly, the fact that isolated cells means that steroids have likely been around in the aquatic environ- from some mollusks show very rapid responses to E2 in vitro [104] ment for a considerable part of mollusk evolution. It is pure specu- seems to imply the existence of a membrane-bound receptor that lation, but if there was perhaps something about the shape of some recognizes this steroid. However, it is unclear whether this repre- steroids (e.g. E2) that made them interfere (even weakly) with some sents the presence of a specific receptor for E2, a receptor for some other types of receptor in mollusks, then this might have provided other compound with which it cross-reacts (e.g. [105]) or indeed, sufficient evolutionary pressure for the development of binding any receptor at all. proteins that would recognize them in order to facilitate their deac- At the moment, there are three ‘main players’ as membrane- tivation. In fact, Baker [113] has hypothesized that it might have bound steroid receptors in vertebrates [106–108]. These are: the been the necessity to recognize and deactivate xenobiotics that en- membrane progestin receptor (mPR; of which there are five vari- abled the evolution of receptors; and has pointed out that some ants, labeled from a to c); the progesterone membrane receptor receptors appear to have retained that role – e.g. the aryl hydrocar- component (PGMRC; of which there are two variants, labeled 1 bon receptor, the main ligand(s) for which appear to be toxic com- and 2); and the membrane estrogen receptor (termed GPR30). pounds that the animals have absorbed or ingested and the main However, there are more possibilities (see [108]). function of which is to activate P450 enzymes in the liver to metab- Anyone setting out to identify membrane-bound receptors in olize those compounds. In relation to E2, the only effect that has so mollusks should be aware that, even within the vertebrate field, far been properly verified in mollusks (Mytilus spp.) is its ability to there is presently a fierce debate about whether the mPRs and very rapidly (<2 min) cause a decrease in lysosomal membrane sta- PGMCRs actually are progestin receptors; and also whether GPR30 bility [104,114]. As pointed out by Canesi and coworkers [115], this really is an estrogen receptor [107,108]! It could ultimately prove is a classic ‘immune response’ rather than an ‘endocrine response’, unrewarding looking for homologous sequences of one or other of thus strengthening the hypothesis that, as far as some mollusks are these putative membrane receptors in mollusks if one then finds concerned, E2 is more likely to be a xenobiotic ‘inflammatory agent’ out that the original protein was not a steroid receptor after all. rather than an endogenous hormone. The gene for PGMRC1, for example, is likely to be present in mollusks, as it has an ancient lineage, and its protein product has been purified 7.6. Is a receptor necessary to elicit effects? from another invertebrate, the rotifer, Brachionus manjavacas [109]. Although it was one of several proteins pulled out of a crude rotifer The fact that steroids are taken up by mollusks, and are then extract using a progesterone-labeled affinity reagent, its presence esterified, means that biochemical and physiological changes take may be fortuitous, because, as stated by two experts in the field of place in the organisms, whether or not they are taken up by a sys- membrane-bound receptors [106,107] ‘it has yet to be demonstrated tem that recognizes them specifically. For all we know at the mo- that PGRMC1 exhibits specific progesterone binding’! ment, mollusks might just treat the steroids in the water as a convenient source of soluble food. When steroids such as T and E2 are added to the water, the animals presumably need to mobi- 7.4. Could there be binding proteins that are not receptors? lize (and probably even manufacture fatty acids). They also need to activate the conjugating enzymes. These steps all require energy In vertebrates, there are a well-known class of proteins, the ‘ste- usage and re-allocation of resources. This makes it possible that roid-binding globulins’ (SBGs) that circulate in the plasma [110] any effects that might be noted when these steroids are added to and are also expressed in the testes (androgen-binding protein). the water might be due to these sorts of changes rather than to SBGs that bind T and E2 with high affinity and specificity are found any receptor activation. in most vertebrates that have been studied. These proteins (the functions of which are still debated) have binding properties that are very similar to those of nuclear receptors. 8. Conclusion Albumin [111] and even egg yolk protein [112] are able to bind to steroids (the latter with very high capacity). However, these pro- Despite many studies failing to take account of the possibility of teins do so with considerably lower affinities than those for nuclear non-specific assay interference, vertebrate-type steroids can receptors or SBGs (and are thus unlikely to explain the steroid undoubtedly be detected in molluscan tissues. Furthermore, in binding activity of O. vulgaris tissue extracts, for example). some studies, it has been claimed that the concentrations of these A.P. Scott / Steroids 77 (2012) 1450–1468 1465 steroids are related to changes in the reproductive cycle, are higher dently repeated. One referee of the present review suggested the in one sex than another or are modified by presumed ‘endocrine use of deuterated water (as opposed to 3Hor14C precursors), so disrupters’. However, despite studies starting over fifty years ago, that any labeled steroids (if they were made) could be simulta- no-one has come up with clear evidence that any mollusk can actu- neously detected and definitively identified by mass spectrometry. ally make vertebrate steroids de novo (although there is little doubt The fact that the tissues of some molluscs have steroid binding that mollusks can metabolize vertebrate steroids – e.g. saturate the activity is still something that needs to be investigated, even if this A ring or convert E1 to E2). Individual steps of the biosynthetic binding turns out to have more to do with the sensing of ‘xenobi- pathway have been demonstrated in some species, but yields have otics’ rather than reception of ‘hormones’. If such binding is medi- in most cases been zero or very, very low (and have, furthermore, ated via non-nuclear mechanisms, knowledge of their identity and never been independently verified). Crucially, the genes for key en- function would be a useful addition to the ongoing debate on the zymes involved in steroid biosynthesis in vertebrates are missing role of such receptors in the physiology and control of hormone- from the genomes of mollusks. This means that, if mollusks do sensitive cancers in humans. make T, E2 and P, they will have to do so via an independently Further investigations are also obviously required on the ability evolved set of enzymes. Unless there is something remarkable of mollusks to absorb, esterify and ‘store’ steroids such as T and E2, about these particular steroids for the whole animal kingdom (or even if these investigations cannot prove or disprove that these indeed most life forms – because there are even claims that plants steroids act as hormones in these animals. Apart from helping to can synthesize mammalian steroids [116], then the development reveal the physiological importance of this mechanism for mol- of independent pathways making exactly the same steroidal end- lusks, it is important to know how mollusks deal with the many products is very unlikely to have occurred. other steroids that they are likely to encounter in their environ-

Furthermore, there is still no firm evidence for the presence of spe- ment (such as cortisol, P and EE2). The possibility that humans cific functional receptors for vertebrate steroids; nor firm evidence might be the source of steroids found in laboratory stocks of mol- that vertebrate steroids have endocrine effects on mollusks [4]. lusks is obviously also something that needs to be investigated in In contrast to the weak evidence for biosynthesis and function- more detail. ality of vertebrate steroids in mollusks, there is strong evidence that mollusks can readily pick up vertebrate steroids from the Acknowledgements environment. If the steroid has a reactive hydroxyl group (e.g. the C-17b-hydroxyl group of both T and E2), then mollusks can None of the opinions expressed in this review are intended as readily conjugate it to a fatty acid. The resulting esters can be re- personal criticisms of any individual or organization. The author tained in the animals for weeks (possibly months). It is pointed has experience working with steroids and steroid binding moieties out, for the first time, that even though humans (and indeed all in fish and cyclostomes, but not directly in mollusks. The author vertebrates) mainly excrete their steroids in the form of glucuro- especially thanks Dr. Alan Pickering and Dr. Christina Lye for refer- nide and sulfate conjugates, this does not necessarily mean that eeing the manuscript (and providing many useful suggestions) they cannot be accessed by mollusks, as mollusks are one of the prior to its submission. The author also expresses gratitude to Dr. richest sources of commercially available glucuronidase and sulfa- Mike Roberts of the Chemicals and Nanotechnology Division, Defra, tase enzymes. UK, for encouraging the project and providing the necessary finan- When one bears in mind that the steroids that everyone has cial assistance; and to Paul Haywood of Cefas, Weymouth for pro- been measuring in mollusks are the same as those made by hu- viding efficient administration. mans and livestock, one is presented with a far more plausible explanation for the presence of steroids in mollusks (i.e. contamination). The only observation that still appears be preventing the References abandonment of the ‘endogenous origin’ hypothesis is that steroids have been detected in the flesh of animals that have been sampled [1] Sumpter JP, Johnson AC. Lessons from endocrine disruption and their application to other issues concerning trace organics in the aquatic from what the authors have presumed to be ‘unpolluted rivers’ or environment. Environ Sci Technol 2005;39:4321–32. ‘clean laboratory conditions’. However, the present review pre- [2] Hagerman DD, Wellington FM, Villee C. Estrogens in marine invertebrates. sents evidence (some of it new) that everything we, as humans, Biol Bull 1957;112:180–3. [3] Rohlack S. Über das vorkommen von sexualhormonen bei der Meeresshnecke touch and everywhere we go, we leave traces of ourselves (and Littorina littorea L.. Z Vergl Physiol 1959;42:161–80. the steroids we have formed). It is concluded that mollusks would [4] Scott AP. Do mollusks use vertebrate sex steroids as reproductive hormones? have to be reared from eggs in completely sterile conditions with II: Unconvincing evidence that vertebrate steroids have biological effects in molluscs. Steroids, in press. their food, water and air being completely devoid of any traces of [5] Scott AP, Bye VJ, Baynes SM. Seasonal variations in sex steroids of female vertebrate steroids (whether in free or conjugated form) before rainbow trout (Salmo gairdneri R.). J Fish Biol 1980;17:587–92. the presence of steroids in mollusk tissues could be accepted as [6] Scott AP, Robinson CD. Fish vitellogenin as a biological effect marker of oestrogenic endocrine disruption in the open sea. In: Payne AIL, Cotter JR, evidence that they are of endogenous origin. Potter ECE, editors. Advances in Fisheries Science. Fifty Years on from Beverton and Holt. Oxford: Blackwell Publishing; 2008. p. 472–90. [7] Lu M, Horiguchi T, Shiraishi H, Shibata Y, Abo M, Okubo A, et al. Identification and quantitation of steroid hormones in marine gastropods by GC/MS. 9. Future directions? Bunseki Kagaku 2001;50:247–55. [8] Lavado R, Janer G, Porte C. Steroid levels and steroid metabolism in the mussel The main thrust of this review has been to point out that there Mytilus edulis: the modulating effect of dispersed crude oil and alkylphenols. is no firm proof (i.e. no incontrovertible evidence) that mollusks Aquat Toxicol 2006;78S:565–72. [9] D’Aniello A, Di Cosmo A, Di Cristo C, Assisi L, Botte V, Di Fiore MM. Occurrence have an endocrine system based on vertebrate steroids. Some peo- of sex steroid hormones and their binding proteins in Octopus vulgaris Lam. ple will undoubtedly argue that the debate cannot be concluded Biochem Biophys Res Commun 1996;227:782–8. until it can be definitively proved that mollusks do not have the [10] Gust M, Mouthon J, Queau H, Dussart C, Buronfosse T, Garric J. Natural variability and response interpretation of fecundity, vertebrate-like sex- ability to make vertebrate steroids. In order to do this, it would steroid levels and energy status in the New Zealand mudsnail Potamopyrgus probably be necessary to revisit one or two of the species that have antipodarum (Gray). Gen Comp Endocr 2011;172:243–50. already been reported to carry out critical transformations of [11] Gust M, Buronfosse T, Geffard O, Coquerry M, Mons R, Abbaci K, et al. Comprehensive biological effects of a complex field poly-metallic pollution radioactive vertebrate steroid precursors in good yields (in bold gradient on the New Zealand mudsnail Potampyrgus antipodarum (Gray). type in Table 2) in order to see whether the results can be indepen- Aquat Toxicol 2011;101:100–8. 1466 A.P. Scott / Steroids 77 (2012) 1450–1468

[12] Gust M, Buronfosse T, André C, Mons R, Gagné F, Garric J. Is exposure [42] De Longcamp D, Lubet P, Drosdowsky M. The in vitro biosynthesis of steroids temperature a confounding factor for the assessment of reproductive by the gonad of the mussel (Mytilus edulis). Gen Comp Endocr parameters of New Zealand mudsnails Potamopyrgus antipodarum (Gray)? 1974;22:116–27. Aquat Toxicol 2011;101:396–404. [43] De Longcamp D, Drosdowsky M, Lubet P. Biosynthèse des stéroïdes chez les [13] Gust M, Garric J, Giamberini L, Mons R, Abbaci K, Garnier F, et al. Sensitivity of mollusques. Mise en évidence d’une 17b-hydroxystéroïde déshydrogénase New Zealand mudsnail Potamopyrgus antipodarum (Gray) to a specific dans les gonades de Mytilus edulis L. (mollusque bivalve). CR Acad Sc Paris aromatase inhibitor. Chemosphere 2010;79:47–53. 1970;271:1564–6. [14] Gust M, Buronfosse T, Geffard O, Mons R, Queau H, Mouthon J, et al. In situ [44] Wootton AN, Herring C, Spry JA, Wiseman A, Livingstone DR, Goldfarb PS. biomonitoring of freshwater quality using the New Zealand mudsnail Evidence for the existence of cytochrome P450 gene families (CYP1A, 3A, 4A, Potamopyrgus antipodarum (Gray) exposed to waste water treatment plant 11A) and modulation of gene expression (CYP1A) in the mussel Mytilus spp.. (WWTP effluent discharges. Water Res 2010;44:4517–28. Mar Environ Res 1995;39:21–6. [15] Gust M, Vulliet E, Giroud B, Garnier F, Couturier S, Garric J, et al. Development, [45] Martínez AA, Muñoz YR, Serrano FSF, García PM. Immunolocalization of validation and comparison of LC–MS/MS and RIA methods for quantification cholesterol side chain cleavage enzyme (P450scc) in Mytilus galloprovincialis of vertebrate-like sex-steroids in prosobranch molluscs. J Chromatogr B and its induction by nutritional levels. J Comp Physiol 2008;178:647–54. 2010;878:1487–92. [46] Markov GV, Tavares R, Dauphin-Villemant C, Demeneix BA, Baker ME, Laudet [16] Gust M, Buronfosse T, Giamberini L, Ramil M, Mons R, Garric J. Effects of V. Independent elaboration of steroid hormone signaling pathways in fluoxetine on the reproduction of two prosobranch mollusks: Potamopyrgus metazoans. PNAS 2009;106:11913–8. antipodarum and Valvata piscinalis. Environ Pollut 2009;157:423–9. [47] Krusch B, Schoenmakers HJN, Voogt PA, Nolte A. Steroid synthesizing capacity [17] Bose R, Majumdar C, Bhattacharya S. Steroids in Achatina fulica (Bowdich): of the dorsal body of Helix pomatia L. (Gastropoda) – an in vitro study. Comp steroid profile in haemolymph and in vitro release of steroids from Biochem Physiol B 1979;64. endogenous precursors by ovotestis and albumen gland. Comp Biochem [48] De Jong-Brink M, Schot LPC, Schoenmakers HJN, Bergamin-Sassen MJM. A Physiol C 1997;116:179–82. biochemical and quantitative electron microscopic study on steroidogenesis [18] Kanda A, Takahashi T, Satake H, Minakata H. Molecular and functional in ovotestis and digestive gland of the pulmonate snail Lymnaea stagnalis. Gen characterization of a novel gonadotropin-releasing hormone receptor isolated Comp Endocr 1981;45:30–8. from the common octopus (Octopus vulgaris). Biochem J 2006;395:125–35. [49] Mori K, Tamate H, Imai T. Presence of 17b-hydroxysteroid dehydrogenase [19] Gottfried H, Dorfman RI, Wall PE. Steroids of invertebrates: production of activity in the tissues of maturing oocytes. Tohoku J Agric Res oestrogens by an accessory reproductive tissue of the slug Arion ater rufus 1965;16:147–57. (Linn.). Nature 1967;215:409–10. [50] Matsumoto T, Osada M, Osawa Y, Mori K. Gonadal estrogen profile and [20] Gottfried H, Lusis O. Steroids of invertebrates: the in vitro production of 11- immunohistochemical localization of steroidogenic enzymes in the oyster ketotestosterone and other steroids by the eggs of the slug, Arion ater rufus and scallop during sexual maturation. Comp Biochem Physiol B (Linn.). Nature 1966;212:1488–9. 1997;118:811–7. [21] Gottfried H, Dorfman RI. V. The in vitro biosynthesis of steroids by the male- [51] Zhou L-Y, Wang D-S, Kobayashi T, Yano A, Paul-Prasanth B, Suzuki A, et al. A phase ovotestis of the slug (Ariolimax californicus). Gen Comp Endocr novel type of P450c17 lacking the lyase activity is responsible for C21-steroid 1970;15:120–38. biosynthesis in the fish ovary and head kidney. Endocrinology [22] Saliot A, Barbier M. Sur l’isolement de la progestérone et de quelques 2007;148:4282–91. cétostéroïdes de la partie femelle des gonades de la coquille Saint-Jacques [52] Baker ME. Evolution of 17b-hydroxysteroid dehydrogenases and their role in Pecten maximus. Biochimie 1971;53:265–6. androgen, estrogen and retinoid actions. Mol Cell Endocrinol 2001;171:211–5. [23] Zhu W, Mantione K, Jones D, Salamon E, Cho JJ, Cadet P, et al. The presence of [53] Scott AP, Sheldrick EL, Flint APF. Measurement of 17a,20b-dihydroxy-4- 17-b estradiol in Mytilus edulis gonadal tissues: evidence for estradiol pregnen-3-one in plasma of trout (Salmo gairdneri Richardson): seasonal isoforms. Neuroendocrinol Lett 2003;24:137–40. changes and response to salmon pituitary extract. Gen Comp Endocr [24] Lu M, Horiguchi T, Shiraishi H, Shibata Y, Abo M, Okubo A, et al. 1982;46:444–51. Determination of testosterone in an individual shell of Thais clavigera by [54] Li W, Scott AP, Siefkes MJ, Yan H, Liu Q, Yun S, et al. A novel bile acid functions ELISA. Bunseki Kagaku 2002;51:21–7. as a sex pheromone in the sea lamprey. Science 2002;296:138–41. [25] Dévier M-H, Labadie P, Togola A, Budzinski H. Simple methodology coupling [55] Le Curieux-Belfond O, Moslemi S, Mathieu M, Séralini GE. Androgen microwave-assisted extraction to SPE/GC/MS for the analysis of natural metabolism in oyster Crassostrea gigas: evidence for 17b-HSD activities and steroids in biological tissues: application to the monitoring of endogenous characterization of an aromatase-like activity inhibited by pharmacological steroids in marine mussels Mytilus sp.. Anal Chim Acta 2010;657:28–35. compounds and a marine pollutant. J Steroid Biochem 2001;78:359–66. [26] Lafont R. Reverse endocrinology, or ‘‘hormones seeking functions’’. Insect [56] Osada M, Tawarayama H, Mori K. Estrogen synthesis in relation to gonadal Biochem 1991;21:697–721. development of Japanese scallop, Patinopecten yessoensis: gonadal profile and [27] Janer G, Porte C. Sex steroids and potential mechanisms of non-genomic immunolocalization of P450 aromatase and estrogen. Comp Biochem Physiol endocrine disruption in invertebrates. Ecotoxicology 2007;16:145–60. B 2004;139:123–8. [28] Fernandes D, Loi B, Porte C. Biosynthesis and metabolism of steroids in [57] Reitzel AM, Tarrant AM. Correlated evolution of androgen receptor and molluscs. J Steroid Biochem 2010;127:189–95. aromatase revisited. Mol Biol Evol 2010;27:2211–5. [29] Lehoux JG, Sandor T. The occurrence of steroids and steroid metabolizing [58] Callard GV, Tarrant AM, Novillo A, Yacci P, Ciaccia L, Vajda S, et al. enzyme systems in invertebrates. A review. Steroids 1970;16:141–71. Evolutionary origins of the estrogen signaling system: insights from [30] Hines GA, Bryan PJ, Wasson KM, McLintock JB, Watts SA. Sex steroid amphioxus. J Steroid Biochem 2011;127:176–88. metabolism in the Antarctic pteropod Clione antarctica (Mollusca: [59] Fostier A, Jalabert B, Billard R, Breton B, Zohar Y. The gonadal steroids. In: Gastropoda). Invertebr Biol 1996;115:113–9. Hoar WS, Randall DJ, Donaldson EM, editors. Fish Physiology Volume IX [31] Idler DR, Sangalang GB, Kanazawa A. Steroid desmolase in gonads of a marine Reproduction Part A Endocrine Tissues and Hormones. New York: Academic invertebrate, Placopecten magellanicus Gmelin. Gen Comp Endocr Press; 1983. p. 277–372. 1969;12:222–30. [60] Bryan MB, Scott AP, Cerny´ I, Young B, Li W. 15a-Hydroxyprogesterone in male [32] Sandor T. Steroids in invertebrates. In: Clark Jr WH, Adams TS, editors. sea lampreys, Petromyzon marinus L.. Steroids 2004;69:473–81. Advances in Invertebrate Reproduction. Amsterdam: Elsevier, North Holland [61] Purdom CE, Hardiman PA, Bye VJ, Eno NC, Tyler CR, Sumpter JP. Estrogenic Inc.; 1981. p. 81–96. effects of effluents from sewage treatment works. Chem Ecol 1994;8:275–85. [33] Lupo di Prisco C, Dessi’ Fulgieri F. Alternative pathways of steroid [62] Matthiessen P, Arnold D, Johnson AC, Pepper TJ, Pottinger TG, Pulman KGT. biosynthesis in gonads and hepatopancreas of Aplysia depilans. Comp Contamination of headwater streams in the United Kingdom by oestrogenic Biochem Physiol B 1975;50:191–5. hormones from livestock farms. Sci Tot Environ 2006;367:616–30. [34] Lupo di Prisco C, Dessi’ Fulgieri F. Identification and biosynthesis of steroids in [63] Scott AP, Ellis T. Measurement of fish steroids in water – a review. Gen Comp the marine mollusc Aplysia depilans. Comp Biochem Physiol B 1973;45:303–10. Endocr 2007;153:392–400. [35] Idler DR, Wiseman P. Molluscan sterols: a review. J Fish Res Board Can [64] Ellis T, James JD, Scott AP. Branchial release of free cortisol and melatonin by 1972;29:385–98. rainbow trout. J Fish Biol 2005;67:535–40. [36] Voogt PA. Investigations of the capacity of synthesizing 3b-sterols in [65] Scott AP, Pinillos M, Huertas M. The rate of uptake of sex steroids from water mollusca. III: The biosynthesis of 3b-sterols in some archeogastropods.. by tench Tinca tinca L. is influenced by their affinity for sex steroid binding Arch Int Phys Biochem 1968;76:721–30. protein in plasma. J Fish Biol 2005;67:182–200. [37] Voogt PA. Investigations of the capacity of synthesizing 3b-sterols in [66] Panter GH, Thompson RS, Beresford N, Sumpter JP. Transformation of a non- mollusca. IV: The biosynthesis of 3b-sterols in some mesogastropods. Comp oestrogenic steroid metabolite to an oestrogenically active substance by Biochem Physiol 1969;31:37–46. minimal bacterial activity. Chemosphere 1999;38:3579–96. [38] Voogt PA. Biosynthesis of 3b-sterols in a snail, Arion rufus L from 1-14C- [67] Gooding MP, LeBlanc GA. Biotransformation and disposition of testosterone acetate. Arch Int Phys Bioch 1967;75:492–9. in the Eastern Mud Snial Ilyanassa obsoleta. Gen Comp Endocr [39] Voogt PA. Investigations of the capacity of synthesizing 3b-sterols in 2001;122:172–80. mollusca. I: Absence of 3b-sterol synthesis in a whelk, Buccinum undatum [68] LeBlanc GA, Gooding MP, Sternberg RM. Testosterone-fatty acid L.. Arch Int Phys Bioch 1968;75:809–15. esterification: a unique target for the endocrine toxicity of tributyltin to [40] Carreau S, Drosdowsky M. The in vitro biosynthesis of steroids by the gonad of gastropods. Integr Comp Biol 2005;45:81–7. the cuttlefish (Sepia officinalis). Gen Comp Endocr 1977;33:554–65. [69] Gooding MP, LeBlanc GA. Seasonal variation in the regulation of testosterone [41] Fagerlund UHM, Idler DR. Marine sterols. IX: Biosynthesis of 21- levels in the eastern mud snail, Ilyanassa obsoleta. Invertebr Biol methylenecholesterol in clams. Can J Biochem Phys 1961;39:1347–55. 2004;123:237–43. A.P. Scott / Steroids 77 (2012) 1450–1468 1467

[70] Le Curieux-Belfond O, Fievet B, Séralini GE, Mathieu M. Short-term [102] Thomas P, Nuclear and membrane steroid receptors and their functions in bioaccumulation, circulation and metabolism of estradiol-17b in the oyster teleost gonads. In: Norberg B, Kjesbu OS, Taranger GL, Andersson E, Crassostrea gigas. J Exp Mar Biol Ecol 2005;325:125–33. Stefansson SO, editors. Proceedings of the 6th international symposium on [71] Puinean A-M, Rotchell JM. Vitellogenin gene expression as a biomarker of the reproductive physiology of fish, institute of marine research and endocrine disruption in the invertebrate, Mytilus edulis. Mar Environ Res University of Bergen, Bergen 2000, Norway; 4–9 July 1999. p. 149–56. 2006;62:S211–4. [103] Bryan MB, Scott AP, Li W. The sea lamprey (Petromyzon marinus) has a [72] Peck MR, Labadie P, Minier C, Hill EM. Profiles of environmental and receptor for androstenedione. Biol Reprod 2007;77:689–96. endogenous estrogens in the zebra mussel Dreissena polymorpha. [104] Canesi L, Ciacci C, Betti M, Lorusso LC, Marchi B, Burattini S, et al. Rapid effects Chemosphere 2007;69:1–8. of 17b-estradiol cell signaling and function of Mytilus hemocytes. Gen Comp [73] Fernandes D, Navarro JC, Riva C, Bordonali S, Porte C. Does exposure to Endocr 2004;136:58–71. testosterone significantly alter endogenous metabolism in the marine mussel [105] Nadal A, Ropero AB, Laribi O, Maillet M, Fuentes E, Soria B. Nongenomic Mytilus galloprovincialis. Aquat Toxicol 2010;100:313–20. actions of estrogens and xenoestrogens by binding at a plasma membrane [74] Janer G, Mesia-Vela S, Porte C, Kauffman FC. Esterification of vertebrate-type receptor unrelated to estrogen receptor a and estrogen receptor b. PNAS steroids in the Eastern oyster (Crassostrea virginica). Steroids 2004;69:129–36. 2000;97:11603–8. [75] Janer G, Lyssimachou A, Bachmann J, Oehlmann J, Schulte-Oehlmann U, Porte [106] Cahill MA. Progesterone receptor membrane component. 1: An integrative C. Sexual dimorphism in esterified steroid levels in the gastropod Marisa review. J Steroid Biochem 2007;105:16–36. cornuarietis: the effect of xenoandrogenic compounds. Steroids [107] Thomas P. Rapid steroid hormone actions initiated at the cell surface and the 2006;71:435–44. receptors that mediate them with an emphasis on recent progress in fish [76] Labadie P, Peck MR, Minier C, Hill EM. Identification of the steroid fatty acid models. Gen Comp Endocr 2012;175:367–83. ester conjugates formed in vivo in Mytilus edulis as a result of exposure to [108] Thomas P. Characteristics of membrane progestin receptor alpha (mPRa) and estrogens. Steroids 2007;72:41–9. progesterone membrane receptor component one (PGMRC1) and their roles [77] Matthiessen P, Gibbs PE. Critical appraisal of the evidence for tributyltin- in mediating rapid progestin actions. Front Neuroendocrinol 2008;29: mediated endocrine disruption in mollusks. Environ Toxicol Chem 292–312. 1998;17:37–43. [109] Stout EP, La Clair JJ, Snell TW, Shearer TL, Kubanek J. Conservation of [78] Gooding MP, Wilson VS, Folmar LC, Marcovich DT, LeBlanc GA. The biocide progesterone hormone function in invertebrate reproduction. PNAS tributyltin reduces the accumulation of testosterone as fatty acids esters in 2010;107:11859–64. the mud snail (Ilyanassa obsoleta). Environ Health Persp 2003;111:426–30. [110] Bobe J, Guigen Y, Fostier A. Diversity and biological significance of sex [79] Janer G, Sternberg RM, LeBlanc GA, Porte C. Testosterone conjugating hormone-binding globulin in fish, an evolutionary perspective. Mol Cell activities in invertebrates: are they target for endocrine disruptors? Aquat Endocrinol 2010;316:66–78. Toxicol 2005;71:273–82. [111] Baker ME. Albumin, steroid hormones and the origin of vertebrates. J [80] Sternberg RM, LeBlanc GA. Kinetic characterization of the inhibition of acyl Endocrinol 2002;175:121–7. coenzyme A: steroid acyltransferases by tributyltin in the eastern mud snail [112] Yoshikuni M, Shibata N, Nagahama Y. Specific binding of [3H]17a,20b- (Ilyanassa obsoleta). Aquat Toxicol 2006;78:233–42. dihydroxy-4-pregnen-3-one to oocyte cortices of rainbow trout [81] Santos MM, Castro LFC, Vieira MN, Micael J, Morabito R, Massanisso P, et al. (Oncorhynchus mykiss). Fish Physiol Biochem 1993;11:15–24. New insights into the mechanism of imposex induction in the dogwhelk [113] Baker ME. Xenobiotics and the evolution of multicellular animals: emergence Nucella lapillus. Comp Biochem Physiol C 2005;141:101–9. and diversification of ligand-activated transcription factors. Integr Comp Biol [82] Le Guellec D, Thiard M-C, Remy-Martin JP, Deray A, Gomot L, Adessi GL. In 2005;45:172–8. vitro metabolism of androstenedione and identification of endogenous [114] Moore MN, Lowe DM, Fieth EM. Responses of lysosomes in the digestive cells steroids in Helix aspersa. Gen Comp Endocr 1987;66:425–33. of the common mussel, Mytilus edulis, to sex steroids and cortisol. Cell Tissue

[83] Tóth I, Faredin J. Steroids excreted by human skin. I: C19 steroids in axillary Res 1978;188:1–9. hair. Acta Med Hung 1983;40:139–43. [115] Canesi L, Lorusso LC, Ciacci C, Betti M, Zampini M, Gallo G. Environmental

[84] Tóth I, Faredin J. Steroids excreted by human skin. II: C19 steroid sulphates in estrogens can affect the function of mussel hemocytes through rapid human axillary sweat. Acta Med Hung 1983;42:21–8. modulation of kinase pathways. Gen Comp Endocr 2004;138:58–69. [85] Feek U, Brachfield K, Krause W. Steroid hormone concentration in human skin [116] Janeczko A, Skoczowski A. Mammalian sex hormones in plants. Folia surface lipids. Acta Endocrinol 1987;116:S143–4. Histochem Cytobiol 2005;43:71–9. [86] Clark RP. Skin scales among airborne particles. J Hyg-Camb 1974;72:47–51. [117] Spooner N, Gibbs PE, Bryan GW, Goad LJ. The effect of tributyltin upon steroid [87] Pollio A, Pinto G, Della Greca M, De Maio A, Fiorentino A, Previtera L. titres in female dogwhelk, Nucella lapillus, and the development of imposex. Progesterone bioconversion by microalgal cultures. Phytochemistry Mar Environ Res 1991;32:37–49. 1994;37:1269–72. [118] Bettin C, Oehlmann J, Stroben E. TBT-induced imposex in marine [88] Liu Y, Guan Y, Mizuno T, Tsuno H, Zhu W. A pretreatment method for GC–MS neogastropods is mediated by an increasing androgen level. Helgolander determination of endocrine disrupting chemicals in mollusk tissues. Meeresun 1996;50:299–317. Chromatographia 2009;69:65–71. [119] Sternberg RM, Hotchkiss AK, LeBlanc GA. The contribution of steroidal [89] Bertrand S, Brunet FG, Escriva H, Parmentier G, Laudet V, Robinson-Rechavi androgens and estrogens to reproductive maturation of the eastern mud snail M. Evolutionary genomics of nuclear receptors: from twenty-five ancestral Ilyanassa obsoleta. Gen Comp Endocr 2008;156:15–26. genes to derived endocrine systems. Mol Evol Biol 2004;21:1923–37. [120] Morcillo Y, Porte C. Evidence of endocrine disruption in clams – Ruditapes [90] Keay J, Thornton JW. Hormone-activated estrogen receptors in annelid decussata – transplanted to a tributyltin-polluted environment. Environ invertebrates: implications for evolution and endocrine disruption. Pollut 2000;107:47–52. Endocrinology 2009;150:1731–8. [121] Morcillo Y, Ronis MJJ, Porte C. Effects of tributyltin on the phase I testosterone [91] Thornton JW, Need E, Crews D. Resurrecting the ancestral steroid receptor: metabolism and steroid titres of the clam Ruditapes decussata. Aquat Toxicol ancient origin of estrogen signaling. Science 2003;301:1714–7. 1998;42:1–13. [92] Thornton JW. Evolution of vertebrate steroid receptors from an ancestral [122] Ketata I, Guermazi F, Rebai T, Hamza-Chaffai A. Variation of steroid estrogen receptor by ligand exploitation and serial genome expansions. PNAS concentrations during the reproductive cycle of the clam Ruditapes 2001;98:5671–6. decussatus: a one year study in the gulf of Gabès area. Comp Biochem [93] Keay J, Bridgham JT, Thornton JW. The Octopus vulgaris estrogen receptor is a Physiol A 2007;147:424–31. constitutive transcriptional activator: evolutionary and functional [123] Negrato E, Marin MG, Bertotto D, Matozzo V, Poltronieri C, Simontacchi C. Sex implications. Endocrinology 2006;147:3861–9. steroids in Tapes philippinarum (Adams and Reeve 1850) during gametogenic [94] Kajiwara M, Kuraku S, Kurokawa T, Kato K, Toda S, Hirose H, et al. Tissue cycle: preliminary results. Fresen Environ Bull 2008;17:1466–70. preferential expression of estrogen receptor gene in the marine snail, Thais [124] Reis-Henriques MA, Coimbra J. Variations in the levels of progesterone in clavigera. Gen Comp Endocr 2006;148:315–26. Mytilus edulis during the annual reproductive cycle. Comp Biochem Physiol A [95] Baker ME. Co-evolution of steroidogenic and steroid-inactivating enzymes 1990;95:343–8. and adrenal and sex steroid receptors. Mol Cell Endocrinol 2004;215:55–62. [125] Reis-Henriques MA, Le Guellec D, Remy-Martin JP, Adessi GL. Studies of [96] Baker ME. Adrenal and sex steroid receptor evolution: environmental endogenous steroids from the marine mollusc Mytilus edulis L. by gas implications. J Mol Endocrinol 2001;26:119–25. chromatography and mass spectrometry. Comp Biochem Physiol B [97] Barnes S. Oestrogens and their promiscuous receptors: confronting reality. 1990;95:303–9. Biochem Soc Trans 2001;29:231–6. [126] Janer G, Lavado R, Thibaut R, Porte C. Effects of 17b-estradiol exposure in the [98] Katzenellenbogen JA, Muthyala R. Interactions of exogenous endocrine active mussel Mytilus galloprovincialis: a possible regulating role for steroid substances with nuclear receptors. Pure Appl Chem 2003;75:1797–817. acyltransferases. Aquat Toxicol 2005;75:32–42. [99] Di Cosmo A, Di Cristo C, Paolucci M. A estradiol-17b receptor in the [127] Kaloyianni M, Stamatiou R, Dailianis S. Zinc and 17b-estradiol induce reproductive system of the female of Octopus vulgaris: characterization and modification in Na+/H+ exchanger and pyruvate kinase activity through immunolocalization. Mol Reprod Dev 2002;61:367–75. protein kinase C in isolated mantle/gonad cells of Mytilus galloprovincialis. [100] Di Cosmo A, Paolucci M, Di Cristo C, Botte V, Ciarcia G. Progesterone receptor Comp Biochem Physiol C 2005;141:257–66. in the reproductive system of the female of Octopus vulgaris: characterization [128] Di Cosmo A, Di Cristo C, Paolucci M. Sex steroid hormone fluctuations and and immunlocalization. Mol Reprod Dev 1998;50:451–60. morphological changes of the reproductive system of the female of Octopus [101] Xue Y, Moore LB, Orans J, Peng L, Bencharit S, Kliewer SA, et al. Crystal vulgaris throughout the annual cycle. J Exp Zool 2001;289:33–47. structure of the pregnane X receptor-estradiol complex provides insights into [129] Mouneyrac C, Linot S, Amiard J-C, Amiard-Triquet C, Métais I, Durou C, et al. endobiotic recognition. Mol Endocrinol 2007;21:1028–38. Biological indices, energy reserves, steroid hormones and sexual maturity in 1468 A.P. Scott / Steroids 77 (2012) 1450–1468

the infaunal bivalve Scrobicularia plana from three sites differing by their [144] Lysimachou A, Rámon M, Porte C. Comparative study on the metabolism of level of contamination. Gen Comp Endocr 2008;157:133–41. the androgen precursor androstenedione in two gastropod species: in vitro [130] Morcillo Y, Porte C. Evidence of endocrine disruption in the imposex-affected alterations by TBT and TPT. Comp Biochem Physiol C 2009;149:409–13. gastropod Bolinus brandaris. Environ Res 1999;81:349–54. [145] Bannister R, Beresford N, May D, Routledge EJ, Jobling S, Rand-Weaver M. [131] Gauthier-Clerc S, Pellerin J, Amiard JC. Estradiol-17b and testosterone Novel estrogen receptor-related transcripts in Marisa cornuarietis:a concentrations in male and female Mya arenaria (Mollusca bivalvia) during freshwater snail with reported sensitivity to estrogenic chemicals. Environ the reproductive cycle. Gen Comp Endocr 2006;145:133–9. Sci Technol 2007;41:2643–50. [132] Siah A, Pellerin J, Benosman A, Gagné J-P, Amiard J-C. Seasonal gonad [146] Castro LFC, Melo C, Guillot R, Mendes I, Queirós S, Lima D, et al. The estrogen progesterone pattern in the soft-shell clam Mya arenaria. Comp Biochem receptor of the gastropod Nucella lapillus: modulation following exposure to Physiol A 2002;132:499–511. an estrogenic effluent? Aquat Toxicol 2007;84:465–8. [133] Siah A, Pellerin J, Amiard J-C, Pelletier E, Viglino L. Delayed gametogenesis [147] Won S-J, Novillo A, Custodia M, Rie MT, Fitzgerald K, Osada M, et al. The and progesterone levels in soft-shell clams (Mya arenaria) in relation to freshwater mussel (Elliptio complanata) as a sentinel species: vitellogenin and in situ contamination to organotins and heavy metals in the St. Lawrence steroid receptors. Integr Comp Biol 2005;45:72–80. River (Canada). Comp Biochem Physiol C 2003;135:145–56. [148] Gagné F, Blaise C, Pellerin J, Pelletier E, Douville M, Gauthier-Clerc S, et al. Sex [134] Lazzara R, Blázquez M, Porte C, Barata C. Low environmental levels of alteration in soft-shell clams (Mya arenaria) in an intertidal zone of the Saint fluoxetine induce spawning and changes in endogenous levels in the zebra Lawrence River (Quebec, Canada). Comp Biochem Physiol C 2003;134:189–98. mussel Dreissena polymorpha. Aquat Toxicol 2012;106–107:123–30. [149] Gagné F, Blaise C, Salazar M, Salazar S, Hansen PD. Evaluation of estrogenic [135] Riva C, Porte C, Binelli A, Provini A. Evaluation of 4-nonylphenol in vivo effects of municipal effluents to the freshwater mussel Elliptio complanata. exposure in Dreissena polymorpha: bioaccumulation, steroid levels and Comp Biochem Physiol C 2001;128:213–25. oxidative stress. Comp Biochem Physiol C 2010;152:175–81. [150] Ciocan CM, Cuberro-Leon E, Puinean A-M, Hill EM, Minier C, Osada M, et al. [136] Hathaway RR. Conversion of estradiol-17b by sperm preparations of sea Effects of estrogen exposure in mussels, Mytilus edulis, at different stages of urchins and oysters. Gen Comp Endocr 1965;1965:501–8. gametogenesis. Environ Pollut 2010;158:2977–84. [137] Lehoux JG, Williams EE. Metabolism of progesterone by gonadal tissue of [151] Puinean A-M, Labadie P, Hill EM, Osada M, Kishida M, Nakao R, et al. Littorina littorea (L.) (Prosobranchia, Gastropoda). J Endocrinol 1971;51: Laboratory exposure to 17b-estradiol fails to induce vitellogenin and 411–2. estrogen receptor gene expression in the marine invertebrate Mytilus edulis. [138] Bardon O, Lubet P, Drosdowsky M. Biosynthèse de stéroïdes chez un mollusque Aquat Toxicol 2006;79:376–83. gastéropode marin, Crepidula fornicata. Steroidologia 1971;2:366–77. [152] Stefano GB, Cadet P, Mantione K, Cho JJ, Jones D, Zhu W. Estrogen signaling at [139] Teshima S-I, Kanazawa A. Metabolism of progesterone in the hepatopancreas the cell surface coupled to nitric oxide release in Mytilus edulis nervous of whelk, Buccinum undatum. Bull Jpn Soc Sci Fish 1971;37:529–33. system. Endocrinology 2003;144:1234–40. [140] Ronis MJJ, Mason AZ. The metabolism of testosterone by the periwinkle [153] Stefano GB, Zhu W, Mantione K, Jones D, Salamon E, Cho JJ, et al. 17-b (Littorina littorea) in vitro and in vivo: effects of tributyl tin. Mar Environ Res estradiol down regulates ganglionic microglial cells via nitric oxide release: 1996;42:161–6. presence of an estrogen receptor b transcript. Neuroendocrinol Lett [141] Oberdörster E, Rittschof D, McClellan-Green P. Testosterone metabolism in 2003;24:130–6. imposex and normal Ilyanassa obsoleta: comparison of field and TBTA Cl- [154] Di Cristo C, Paolucci M, Di Cosmo A. Progesterone affects vitellogenesis in induced imposex. Mar Pollut Bull 1998;36:144–51. Octopus vulgaris. Open Zool L 2008;1:29–36. [142] Janer G, LeBlanc GA, Porte C. A comparative study on androgen metabolism in [155] Matsumoto T, Nakamura AM, Mori K, Akiyama I, Hirose H, Takahashi Y. three invertebrate species. Gen Comp Endocr 2005;143:211–21. Oyster estrogen receptor: cDNA cloning and immunolocalization. Gen Comp [143] Janer G, LeBlanc GA, Porte C. Androgen metabolism in invertebrates and its Endocr 2007;151:195–201. modulation by xenoandrogens. A comparative study. Ann NY Acad Sci [156] Croll RP, Wang C. Possible roles of sex steroids in the control of reproduction 2005;1040:354–6. in bivalve molluscs. Aquaculture 2007;272:76–86.