Molecular and Cellular Endocrinology 402 (2015) 107–112

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / m c e

The variable region of iodothyronine deiodinases directs their catalytic properties and subcellular localization Aurora Olvera, Arturo Mendoza, Patricia Villalobos, Lidia Mayorga-Martínez, Aurea Orozco *, Carlos Valverde-R Instituto de Neurobiología, Universidad Nacional Autónoma de México (UNAM), Boulevard Juriquilla 3001, Juriquilla, Querétaro 76230, Mexico

A R T I C L E

I N F O

Article history: Received 7 October 2014 Received in revised form 8 January 2015 Accepted 8 January 2015 Available online 12 January 2015 Keywords: Deiodinase catalytic selectivity Deiodinase subcellular localization Shark deiodinase

A B S T R A C T

The stereospecific removal of iodine from thyroid hormones is an essential first step for T3 action and is catalyzed by three different deiodinases: D2 and D3 remove iodine only from the outer or inner ring, respectively, whereas D1 catalyzes both pathways. We used in silico predictions from vertebrate deiodinase sequences to identify two domains: the N-terminal variable region (VR) containing the transmembrane, hinge and linker domains, and the conserved or globular region (CR). Given the high sequence and structural identity of the CR among paralogs as well as of the VR among orthologs but not paralogs, we hypothesized that both the catalytic properties and the subcellular localization rely on the VR. We used shark D2 and D3 as templates to build the chimeric enzymes D2VR/D3CR and D3VR/D2CR. Biochemical characterization revealed that D3VR/D2CR has inner-ring deiodination activity and T3 as preferred substrate, whereas D2VR/D3CR showed no deiodinating activity. Also, D2VR/D3CR and D3VR/D2CR reside in the endoplasmic reticulum and plasmatic membrane, respectively, as do their D2 and D3 wild-type counterparts. We conclude that the VR determines the subcellular localization and is critical in defining the catalytic properties and activity of thyroid hormone deiodinases. © 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Thyroid hormone (TH) deiodination is the critical first step of thyroid hormone action and is catalyzed by a family of selenoenzymes known as deiodinases (Ds). It is well accepted that deiodinase activity determines the relative concentrations of bioactive vs. inactive intracellular THs, thus playing a key role in homeostasis of the vertebrate thyroidal system (reviewed by Bianco, 2011; Gereben et al., 2008b). Three Ds have been identified: D1, a multifunctional enzyme that can cleave iodine from the outer as well as the inner ring of an iodothyronine, thereby activating or inactivating the TH; D2 which functions predominantly as an activating deiodinase via outer ring deiodination (ORD), and D3 which only inactivates TH through inner ring deiodination (IRD) (Bianco and Larsen, 2005; Gereben et al., 2008a). Consequently, the selectivity and stereo-specificity of iodine removal from the thyroid hormone molecule determine the function of the deiodinase isotype. Interestingly, the three deiodinases show considerable similarity and share a common structural organization, suggesting that they may have diverged from a common ancestral gene (Darras and Van Herck,

* Corresponding author. Instituto de Neurobiología, UNAM, Boulevard Juriquilla 3001, Querétaro, Qro. 76230, México. Tel.: +52 (442) 238 1068; fax: +52 (442) 238 10 38. E-mail addresses: [email protected], [email protected] (A. Orozco). http://dx.doi.org/10.1016/j.mce.2015.01.011 0303-7207/© 2015 Elsevier Ireland Ltd. All rights reserved.

2012; Laudet, 2011; Orozco et al., 2012; Valverde-R et al., 2004). In this context, the in silico modeling of human Ds suggests that the molecular conformation of the three paralogs consists of four functional domains known as: transmembranal (TM), hinge (H), linker (L), and catalytic or globular (G) (Callebaut et al., 2003). Furthermore, we recently reported that when the amino acid sequences of most expressed vertebrate deiodinases are aligned, two distinct major regions are revealed, which we have identified as the variable region (VR) and conserved region (CR) (Orozco et al., 2012). The VR includes the TM, H and L domains and is highly variable among paralogs, presenting only 20% molecular identity; however, this region is relatively conserved among orthologs (percent identity: D1, 50%; D2, 55% and D3, 60%). The CR, which comprises only the G domain, is very similar among the 3 paralogs (60% molecular identity) as well as among orthologs (percent identity: D1, 75%; D2, 79% and D3, 73%). As previously mentioned, the function of each deiodinase isotype is the stereo-specific removal of an iodine atom from the thyroid hormone. It would be reasonable to propose that, since the catalytic reaction takes place in the CR where the active site resides, this region would, at least in part, determine the stereo-specificity of TH iodine removal, but this has not been demonstrated to date. An argument against this proposal is the high sequence conservation of this region among deiodinase paralogs. Moreover, mammalian D paralogs show a different subcellular localization: D2 is inserted into the endoplasmic reticulum membrane, whereas D1 and D3 are

108

A. Olvera et al./Molecular and Cellular Endocrinology 402 (2015) 107–112

that contained NcoI and NotI sites, respectively, for site-directed cloning into a pXENEX1 vector (see discussion later).

plasma membrane residents (Baqui et al., 2000, 2003; Friesema et al., 2006). No structure–function studies are available regarding the domains of Ds that direct subcellular localization or confer catalytic properties. Here, we provide experimental evidence showing that both stereo-selectivity and subcellular localization of Ds are influenced, to some extent, by the VR molecular structure.

2.3. Construction of deiodinase vectors 2.3.1. Chimeras The chimeras were generated by PCR by fusing the entire VR of sD2 to the CR of sD3 and vice versa, the VR of sD3 to the CR of sD2. The resulting products were named D2VR/D3CR and D3VR/D2CR, respectively. To this end, internal hybrid primers (Supplement Table S1) were prepared to amplify those segments that correspond to the VR and CR of each enzyme as illustrated in Fig. 1b. VR and CR fragments form sD2 and sD3 were independently amplified (PCR), and the corresponding fragments (sD2 + sD3 and sD3 + sD2) were hybridized in two, sequential PCR reactions in which the first protocol was designed to amplify the template and did not include primers, whereas the second protocol included the corresponding external primers and extended the complete chimeras. In all cases, the entire chimeric cDNA sequences were amplified using sense and antisense primers that contained the NcoI and NotI sites, respectively, for site directed cloning into a pXENEX1 vector (see discussion later) and verified by sequencing.

2. Materials and methods 2.1. Animals Sharks (Chiloscyllium punctatum) ranging from 25 to 50 cm in total length were obtained from a commercial fish collector and held at 25 ± 1 °C in a seawater aquarium under a 12-h:12-h L:D photoperiod. Sharks were anesthetized in MS-222 (0.5 g L−1) prior to sacrifice, and the livers were immediately dissected and quick frozen. All animal procedures were approved by the institutional animal care committee. 2.2. Cloning and sequencing of shark D2 Shark D2 (sD2) was cloned as previously described for shark D3 (sD3) (Martínez et al., 2008). Briefly, 10 μg of total RNA from shark liver was reverse transcribed (SuperScriptTM ll, RNase reverse transcriptase; Invitrogen, Carlsbad CA) with an oligo(dT) primer. Degenerate primers recognizing the sequence coding for the active site of deiodinases (Supplement Table S1) were used in touchdown PCRs (Platinum Taq DNA Polymerase, Invitrogen, Carlsbad CA), and the amplicon (130 bp) was subcloned into a pGEM-T vector (Promega; Madison, WI) and sequenced. The open reading frame (ORF) of sD2 was amplified by 3′ and 5′ rapid amplifications of cDNA ends (RACEs) in a series of nested PCRs using specific primers designed based on the initial 130 bp fragment, as previously described (Martínez et al., 2008; Orozco et al., 2002, 2003). The amplified products from both 3′ and 5’ RACE were cloned and sequenced. In order to express a functional protein, the ORF of sD2 was linked to the selenocysteine insertion sequence (SECIS) from killifish D2 using a hybrid pair of oligonucleotides (Supplement Table S1), as previously described (Orozco et al., 2002). The entire cDNA sequence that included the SECIS was amplified using sense and antisense primers

2.3.2. GFP-tagged deiodinases ORFs for D3, D2, D2VR/D3CR or D3VR/D2CR were subcloned using EcoRI/BamHI sites into a pEGFPN1 vector [Clontech (Mountain View, CA)]. In all cases, the selenocysteine coding codon was replaced by the cysteine codon to allow expression of GFP C-terminal GFPtagged deiodinases. Constructs were confirmed by sequencing. 2.4. Expression of recombinant shark deiodinases As previously described, cDNA constructs of the native sD2 and sD3, as well as the chimeras D2VR/D3CR and D3VR/D2CR, were digested and ligated into a vector (pXENEX1) designed to express RNA in Xenopus oocytes (Jeziorski et al., 1998). The cDNAs were verified by sequencing and linearized with HindIII [New England Biolabs (Ipswich, MA)], purified, and used to transcribe capped RNA using T7 polymerase. Stage V–VI oocytes were removed from Xenopus under MS-222 anesthesia and treated as previously described

a 11

D2 D3

*

81 81

D2 D3

* *

**

*

*

*

*

*

* *

*

80 80

* *

**

*

*

***

* *

*

* ** **

**

* *** *** *

*

** * *

* * * * * ** * *

* * * *160* 160

*

*

*

*

*

*

*

*

****

**

*

*

*

*

* * **

*

*240 240

V A S L P A F C R T A E E Y Q A G A D F L L V Y I D E A H P S D G W A A T - S P F Q L P R H R T L A E RC S A A S L L L R H F P V P P Q C P M A R L K S F Q R V A T Q Y A D I A D F L L I Y I E E A H P S D G W V S T D A P Y N I P R H R S L E D RL K A A S L I D K E - - - S P G C L

241 241

D2 D3

*

V W R S F L L D A AK P V R P G G Q A P N P R V V R L G E R Q G E A E R L Q R V G G L A E C R L L D F AS P G R P L V V N F G S A T U P P F I W Y G Q K L D F F K S A H V G S P A P N T E V V Q L Q D Q R - - - - - - - - - - - - - K V R L L D Y S R G A R P L V L N F G S C T U P P F

161 161

D2 D3

*

M G L R S V D L L I K L Q I L P G F F S N C L F L A V Y D S V V L L K Q A V S R L S G P A A G - - D G V P - - - - - - - R M L T V E G M Q T M D L L K - E I V AI L I L L P R F F I T A F M L W L L D I L C I R K R L L P K R R E Q T E G S A D D P P L C V S D T N R M F T L E S V K A

* * **

* **

*

** *

* **

* **

**

**

*

316 316

* * **

V V A D C M D N N S N V A Y G V S F E RF C I V Q D Q K I V Y L G G K G P F F Y S L K E V R A W L E Q T P F - - - - - - - - - - - V V A D T M D N S S N S A Y G A Y F E RL Y V L R D Q K V V Y Q G G R G P E G Y K I S E L R L W L E Q Y K S Q S H N S S T V L I E V

b

1

D2 +1

D3 263

118-119 VR

CR

+1

D2VR/D3CR

2 D2 VR

D3 CR

269

113-114 VR

CR

D3VR/D2CR D3 VR

D2 CR

Fig. 1. Construction of shark deiodinase chimeras. (a) Full amino acid sequence of shark D2 (Accession number KP455983) and D3 (Accession number ABX60542.1). The invariant amino acids are indicated with an asterisk. Light green and light red highlight the VR while dark green and dark red highlight the CR. (b) Schematic representation of VR and CR from D2 and D3 with their corresponding amino acid numbering; D2VR/D3CR and D3VR/D2CR chimeras were constructed by exchanging the entire VR of D2 with the CR of D3, and vice versa. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

A. Olvera et al./Molecular and Cellular Endocrinology 402 (2015) 107–112

(Jeziorski et al., 1998). Oocytes (n = 30) were injected with 50 nL (100 ng/oocyte) of RNAs encoding the recombinant shark (rs) D2 and rsD3, as well as the chimeric D2VR/D3CR and D3VR/D2CR, and incubated for 3 days; uninjected oocytes were used as a negative control. Crude oocytes were homogenized in 10 volumes of buffer (0.25 M sucrose, 10 mM Hepes, 1 mM DTT, pH 7), snap frozen in aliquots, and stored at -80 °C until deiodinase activity was measured as will be described later. Expression efficiency of the different deiodinase constructs was indirectly evaluated by fluorescence microscopy (see discussion later). All studied constructs yielded GFPtagged deiodinases. 2.5. Deiodination assays As previously described (Orozco et al., 2000), all enzyme activity assays were carried out in duplicate. In all cases the total volume of the reaction mixture was 100 μL and contained the radiolabeled substrate 125I-T4 or 125I-T3 (specific activities: 1200 μCi/μg; Perkin-Elmer Life Sciences, Inc. Boston MA) plus the corresponding non-radioactive thyronine (Sigma Chemical Co., St. Louis, MO) and DTT (Calbiochem, La Jolla, CA) as will be indicated later. Radiolabeled thyronines were purified prior to use by means of a SEPPACK C18 cartridge from Millipore (Waters Chromatography, Boston, MA). D3 activity (IRD) was measured using a descending paper chromatography system as previously described (Fenton et al., 1997). The assay mixture contained 2 nM 125I-T3, 50 mM DTT and unlabeled T3. D2 activity (ORD) was measured by a modification of the radiolabeled iodine release assay, as previously described (Orozco et al., 1997). The assay mixture contained 0.5 nM 125I-T4 and 10 mM DTT. Released acid-soluble 125I was isolated by chromatography on Dowex 50W-X2 columns. In all cases, conversion of 125I-T4 to 125I-T3 for ORD, or 125I-T3 to 125I-3,3’-T2 for IRD was corrected for the non-enzymatic deiodination observed in blanks incubated in the absence of protein (buffer blank). Protein concentrations were determined using the BioRad assay reagents (Richmond, CA) and BSA as standard. The kinetic characterization of rsD2 or the chimeric Ds included the following parameters: protein concentration (1–200 mg), pH (5–9, at intervals of 0.5), cofactor (DTT) concentration (1–100 mM), as well as incubation time (1–3 h) and temperature (4–45 °C). Kinetic constants were determined by non-linear fit of the raw data to the Michaelis–Menten equation using the GraphPad Prism 6 software. 2.6. Fluorescent confocal microscopy Cell treatment for confocal microscopy was performed as briefly described here. One day prior to transfection 1 × 105 GH3 cells were plated on a FluoroDish with a poly-D lysine coated cover glass bottom and cultured in F-12K medium supplemented with 15% horse serum and 1.5% fetal calf serum. Cells were transfected with 2 μg of the plasmid encoding GFP-tagged deiodinases D2, D3, D2VR/D3CR or D3VR/D2CR using the lipofectamine 2000 reagent, which yields 2–5% of highly (confocal microscopy suitable) GFP-positive cells. Twentyfour hours after transfections, the cells were washed with 1× PBS and incubated for 30 min with 2 drops/mL of Nuc Blue ready probes for nucleus, and either ER-tracker (1 μM) or Cell Mask (5 μg/mL) for endoplasmic reticulum or plasma membrane staining, respectively [Molecular Probes by Life Technologies (Eugene, OR)]. After washing the cells 3 times with 1× PBS, live cell confocal microscopy was performed using a 510 LSM (Zeiss) instrument, and images were processed using ImageJ software from NIH. 3. Results and discussion We previously reported that when the amino acid sequences of vertebrate deiodinases are aligned, two major regions can be iden-

109

tified: VR and CR (Orozco et al., 2012). We noted that the CR or globular domain that encompasses the catalytic region is highly conserved and shares the same molecular structure among the three paralogs as well as among species; however, the three enzymes differ in subcellular localization, biochemical features and the ability to perform ORD or IRD (Vivek-Sagar et al., 2007, 2008). On the other hand, the VR is the most variable among the three paralogs, but it is conserved among orthologs. Thus, we hypothesized that the biochemical features, stereo-selectivity and subcellular localization of Ds depend, significantly, on the VR conformation. To test these hypotheses we used D2 and D3 since they exclusively catalyze ORD or IRD, respectively; also, we extended our previous work in nonmammalian Ds by using shark enzymes. We had previously reported the cloning and characterization of shark D3 (Martínez et al., 2008; GeneBank accession number: ABX60542.1); in the present study, we report the cloning and characterization of a shark enzyme with molecular and biochemical properties that correspond to those of D2 (Fig. 1a). Indeed, the amino acid sequence of the cloned cDNA (accession number KP455983) exhibits a 79.35% identity with other representative vertebrate D2s (H. sapiens, M. musculus, G. gallus, X. laevis, F. heteroclitus). The initial kinetic characterization of the recombinant enzyme included the determination of optimal incubation time, temperature and pH as well as protein and cofactor (DTT) concentrations (data not shown). Kinetic parameters were determined and are summarized in Table 1. The expressed enzyme exhibits an apparent Km value for T4 in the nanomolar range. Together, the results obtained show that the cloned enzyme indeed corresponds to a D2. We then tested the substrate preference for both recombinant deiodinases (D2 and D3) in order to verify thyroid hormone specificity between shark paralogs. As expected, T4 is the preferred substrate for rsD2, whereas rsD3 preferentially deiodinates T3 (Fig. 2). We proceeded to test if the VR molecular conformation was involved in conferring stereo-selectivity of iodine removal. Two deiodinase chimeras were constructed by attaching the entire VR of rsD2 to the CR of rsD3 (chimera D2VR/D3CR) and vice versa (chimera D3VR/D2CR) (Fig. 1b), and ORD and IRD activities were determined for both recombinant deiodinases as well as for the two chimeras. As shown in Fig. 3, rsD3 clearly catalyzes IRD, and its activity is comparable to that measured in rat placenta, a tissue that only and highly expresses this deiodinase isotype. In contrast, ORD activity was only present in rsD2. Only one of the chimeric proteins was functional, D3VR/D2CR; this chimera presented a low but significant IRD activity, as compared to the rsD3. Indeed, despite possessing the catalytic domain proper for ORD, and in agreement with our hypothesis, D3VR/D2CR catalyzes T3-IRD but not T4ORD. In contrast, the opposite chimera, D2VR/D3CR, which contains the entire VR from D2 and the catalytic domain from D3, lost its ability to remove iodine; thus, it did not catalyze either ORD or IRD. A low translation efficiency could not explain this result since this chimera showed the highest expression in GH3 cells (see discussion later). These findings on the one hand did not entirely support our hypothesis since the chimeric protein did not perform ORD. It is interesting to note that, in evolutionary terms, amino acid conservation is higher for D2 (75%) than for D3 (69%) (Orozco et al., 2012). Thus, D2 has been under stronger selective pressure, which could be interpreted as ORD being less permissive to amino acid

Table 1 Kinetic parameters for ORD of rsD2. Substrate

Km (nM)

Vmax (pmol/mg protein/hr)

Vmax/Km

3,5-T2 T3 T4 rT3

359 44 0.9 11

3414 496 33 161

10 11 4 15

110

A. Olvera et al./Molecular and Cellular Endocrinology 402 (2015) 107–112

rsD3

%Deiodination

rsD2 100

100

75

75

Control T3 rT3 T4

50

50

25

25

0

0 Control

T3

rT3

T4

Control

T3

rT3

T4

Fig. 2. Substrate preference for rsD2 and rsD3. Competitive deiodination was evaluated under the following assay conditions: 0.5 nM 125T4 and 10 mM DTT or 2 nM 125T3 and 50 mM DTT for rsD2 ORD or rsD3 IRD determination, respectively. The protein concentration used was approximately 10 μg/tube, and incubations were carried out for 1 h at 37 °C. ORD and IRD assays included 100 nM unlabeled T3, rT3 or T4. Results are expressed as % deiodination in the presence of the unlabeled substrate relative to that in the control incubation without unlabeled substrate.

IRD

5.0X102 0 0

D3VR/ D2CR

1.3X1055 3.0X1022 2.0X1022 1.0X1022 0 0

D2VR/ D3CR

Control Placenta rsD3

R

Control Placenta rsD2

1.6X1055

D3VR/ D2CR

D2VR/ D3CR R

4.5X1033 1.0X1022

1.8X1055

R

fmol [125I]-3,3’T2/mg.hr

5.0X1033

R

fmol released 125I/mg.hr

ORD 5.5X1033

Fig. 3. The VR molecular conformation is involved in conferring stereo-selectivity to iodine removal. IRD activity was detected in rat placenta crude homogenates, rsD3 and D3VR/D2CR, while ORD activity was only present in rsD2. Control groups (uninjected oocytes) and D2VR/D3CR showed no activity. Deiodination assays were performed as follows: 0.5 nM 125T4 and 10 mM DTT or 2 nM 125T3 and 50 mM DTT for rsD2 ORD or rsD3 IRD determination, respectively. The protein concentration used was approximately 10 μg/tube, and incubations were carried out for 1 h at 37 °C. Results are expressed as the enzyme specific activity: fmol released 125I/mg.hr (ORD) and fmol [125I]-3,3′-T2/mg.hr) (IRD).

et al., 2012). In light of the recent accomplishments regarding the partial crystallographic resolution of a D3 (Schweizer et al., 2014), the identification of the exact domain in the VR that confers the catalytic selectivity of Ds could be approached in the near future through in silico molecular modeling together with structural– functional data.

D3VR/D2CR 100 Control T3

%Deiodination

changes and the subsequent interaction between them. IRD appears to be a more plastic system. On the other hand, the D2VR/D3CR results did highlight that, for the active conformation of a deiodinase, a specific interaction between VR and CR is essential. Indeed, the inactivity of D2VR/D3CR, as well as the very low activity of the D3VR/ D2CR chimeric deiodinase, could reflect an inefficient intermolecular interaction between key amino acids of the two regions. In this context, mouse D2-D2 dimers were found to be formed by interactions between the TM and the G domains, and modeling these interactions revealed a region of negative electrostatic potential around the active sites that could help attract the hormone (Vivek-Sagar et al., 2007). Thus, altering VR–CR interactions could disrupt this electrostatic environment, resulting in a reduced or lost capacity to catalyze deiodination. Furthermore, the expression of the D2 globular domain alone results in an inactive deiodinase; however, when co-expressed with a full-length, inactive D2 (SeCys replaced with Ala), the dimer regains activity (Vivek-Sagar et al., 2007, 2008). Thus, even when the globular (catalytic) domain is intact, the VR conformation appears to be essential for a functional deiodinase. Moreover, substrate preference is also influenced by the VR, since D3VR/D2CR preferentially deiodinates the inner ring of T3. We then analyzed substrate preference in this chimera. Fig. 4 shows that D3VR/D2CR has greater specificity for T3 IRD, resembling a D3 paralog. Overall, these findings are consistent with our proposal that the molecular features of VRs that are conserved among orthologs strongly influence the catalytic properties of Ds (Orozco

75

rT3 T4

50 25 0 Control

T3

rT3

T4

Fig. 4. The VR molecular conformation is involved in conferring substrate preference on D3VR/D2CR. Despite having the catalytic region of D2, T3 is the preferred substrate of D3VR/D2CR; thus, it resembles a D3 enzyme. Deiodination assays were performed using 2.5 nM of the various unlabeled THs, and otherwise as in Fig. 2.

A. Olvera et al./Molecular and Cellular Endocrinology 402 (2015) 107–112

ER-tracker

ER-tracker

Cell mask

Cell mask

D2-gfp

VRD2/CRD3-gfp

D3-gfp

VRD3/CRD2-gfp

111

Nuc-blue

Overlay

Fig. 5. The VR of shark Ds determines their subcellular localization. GH3 cells were transfected (Invitrogene Lipofectamine) with plasmids encoding D3, D3VR/D2CR, D2, and D2VR/D3CR. In each case, sequences coding for GFP were added at the C-terminus. Cells were treated with: Cell Mask for plasma membrane staining or ER-Tracker for endoplasmic reticulum staining. Cell nuclei were stained with Nuc Blue. Overlay of the three channels is depicted in last row; the scale bar corresponds to 5 μm. The fraction of GFP that does not overlap with the markers could represent overexpressed protein not yet inserted.

Due to the fact that the TM domain is located in the VR, we asked whether this region could also determine subcellular localization. As other vertebrate Ds (Gereben et al., 2008a), rsD2 and rsD3 insert into endoplasmic reticulum and plasma membrane, respectively (Fig. 5). Interestingly, the D3VR/D2CR chimera was found mostly in the plasma membrane resembling rsD3 localization, whereas D2VR/ D3CR was found essentially to be an endoplasmic reticulum resident like rsD2. The fraction of GFP that does not overlap with the markers could represent overexpressed protein not yet inserted. These results are comparable to those by Vivek-Sagar et al., (2007, 2008) who showed that deletion of the mouse D2 TM domain resulted in an inactive and cytosol-resident deiodinase. Together, these results provide further evidence that the VR not only contains the subcellular localization signal, but is also necessary for the various deiodinase isotypes to adopt an active conformation, which in turn could impact the stereo-specificity of iodine removal. Crystallographic data of full-length deiodinases would further enlighten the functional role of the different domains of this singular set of enzymes and their corresponding intra-domain interactions. Acknowledgements Aurora Olvera is a doctoral student from Programa de Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México (UNAM) and received fellowship 245201 from CONACYT. We acknowledge the technical support of Edith Garay, Elsa Nydia Hernández Ríos and Michael Jeziorski, and thank Dorothy Pless for

critically reviewing the manuscript as well as Leonor Casanova and Ramón Martínez Olvera. This study was supported by grants from PAPIIT IN201614 and CONACyT 166357.

Appendix: Supplementary material Supplementary data to this article can be found online at doi:10.1016/j.mce.2015.01.011.

References Baqui, M.M.A., Gereben, B., Harney, J.W., Larsen, P.R., Bianco, A.C., 2000. Distinct subcellular localization of transiently expressed types 1 and 2 iodothyronine deiodinases as determined by immunofluorescence confocal microscopy. Endocrinology 141 (11), 4309–4312. doi:10.1210/en.141.11.4309. Baqui, M.M.A., Botero, D., Gereben, B., Curcuio, C., Harney, J.W., Salvatore, D., et al., 2003. Human type 3 iodothyronine selenodeiodinase is located in the plasma membrane and undergoes rapid internalization to endosomes. J. Biol. Chem. 278, 1206–1211. doi:10.1074/jbc.M210266200. Bianco, A.C., 2011. Minireview: cracking the metabolic code for thyroid hormone signaling. Endocrinology 152, 3306–3311. doi:10.1210/en.2011.1104. Bianco, A.C., Larsen, P.R., 2005. Cellular and structural biology of the deiodinases. Thyroid 15 (8), 777–786. doi:10.1089/thy.2005.15.777. Callebaut, I., Curcio-Morelli, C., Mornon, J.P., Gereben, B., Buettner, C., Huang, S., et al., 2003. The iodothyronine selenodeiodinases are thioredoxin-fold family proteins containing a glycoside hydrolase clan GH-A-like structure. J. Biol. Chem. 278, 36887–36896. doi:10.1074/jbc.M305725200. Darras, V.M., Van Herck, S.L.J., 2012. Iodothyronine deiodinase structure and function: from ascidians to humans. J. Endocrinol. 215 (2), 189–206. doi:10.1530/JOE-120204.

112

A. Olvera et al./Molecular and Cellular Endocrinology 402 (2015) 107–112

Fenton, B., Orozco, A., Valverde-R, C., 1997. Kinetic characterization of skin inner-ring deiodinative pathway and its correlation with circulating levels of reverse tri-iodothyronine in developing rainbow trout. J. Endocrinol. 154 (3), 547–554. doi:10.1677/joe.0.1540547. Friesema, E.C., Kuiper, G.G., Jansen, J., Visser, T.J., Kester, M.H., 2006. Thyroid hormone transport by the human monocarboxylate transporter 8 and its rate-limiting role in intracellular metabolism. Mol. Endocrinol. 20 (11), 2761–2772. doi:10.1210/ me.2005-0256. Gereben, B., Zavacki, A.M., Ribich, S., Kim, B.W., Huang, S.A., Simonides, W.S., et al., 2008a. Cellular and molecular basis of deiodinase-regulated thyroid hormone signaling. Endocr. Rev. 29, 898–938. doi:10.1210/er.2008-0019. Gereben, B., Zeöld, A., Dentice, M., Salvatore, D., Bianco, A.C., 2008b. Activation and inactivation of thyroid hormone by deiodinases: local action with general consequences. Cell. Mol. Life Sci. 65, 570–590. doi:10.1007/s00018-0077396-0. Jeziorski, M.C., Greenberg, R.M., Clark, K.S., Anderson, P.A., 1998. Cloning and functional expression of a voltage-gated calcium channel a1 subunit from jellyfish. J. Biol. Chem. 273 (35), 22792–22799. doi:10.1074/jbc.273.35.22792. Laudet, V., 2011. The origins and evolution of vertebrate metamorphosis. Curr. Biol. 21 (18), R726–R737. doi:10.1016/j.cub.2011.07.030. Martínez, M.L., Orozco, A., Villalobos, P., Valverde-R, C., 2008. Cloning and characterization of a type 3 iodothyronine deiodinase (D3) in the liver of the chondrichtyan Chiloscyllium punctatum. Gen. Comp. Endocrinol. 156, 464–469. doi:10.1016/j.ygcen.2008.02.012. Orozco, A., Silva, J.E., Valverde-R, C., 1997. Rainbow trout liver expresses two iodothyronine phenolic ring deiodinase pathways with the characteristics of mammalian types I and II 5′-deiodinases. Endocrinology 138, 254–258. doi:10.1210/endo.138.1.4878.

Orozco, A., Linser, P., Valverde-R, C., 2000. Kinetic characterization of outer-ring deiodinase activity (ORD) in the liver, gill and retina of the killifish Fundulus heteroclitus. Comp. Biochem. Physiol. 126, 283–290. doi:10.1016/S03050491(00)00186-3. Orozco, A., Jeziorski, M.C., Linser, P.J., Greenberg, R.M., Valverde-R, C., 2002. Cloning of the gene and complete cDNA encoding a type 2 deiodinase in Fundulus heteroclitus. Gen. Comp. Endocrinol. 128, 162–167. doi:10.1016/S00166480(02)00071-0. Orozco, A., Villalobos, P., Jeziorski, M.C., Valverde-R, C., 2003. The liver of Fundulus heteroclitus expresses deiodinase type 1 mRNA. Gen. Comp. Endocrinol. 130, 84–91. doi:10.1016/S0016-6480(02)00570-1. Orozco, A., Valverde-R, C., Olvera, A., García-G, C., 2012. Iodothyronine deiodinases: a functional and evolutionary perspective. J. Endocrinol. 215, 207–219. doi:10.1530/JOE-12-0258. Schweizer, U., Schlicker, C., Braun, D., Köhrle, J., Steegborn, C., 2014. Crystal structure of mammalian selenocysteine-dependent iodothyronine deiodinase suggests a peroxiredoxin-like catalytic mechanism. Proc. Natl. Acad. Sci. U.S.A. 111, 10526–10531. doi:10.1073/pnas.1323873111. Valverde-R, C., Orozco, A., Becerra, A., Jeziorski, M.C., Villalobos, P., Solis-S, J.C., 2004. Halometabolites and cellular dehalogenase systems. An evolutionary perspective. Int. Rev. Cytol. 234, 143–199. doi:10.1016/S0074-7696(04)34004-0. Vivek-Sagar, G.D., Gereben, B., Callebaut, I., Mornon, J.P., Zeöld, A., Da Silva, W.S., et al., 2007. Ubiquitination-induced conformational change within the deiodinase dimmer is a switch regulating enzyme activity. Mol. Cell. Biol. 27, 4774–4783. doi:10.1128/MCB.00283-07. Vivek-Sagar, G.D., Gereben, B., Callebaut, I., Mornon, J.P., Zeöld, A., Curcio-Morelli, C., et al., 2008. The thyroid hormone-inactivating deiodinase functions as a homodimer. Mol. Endocrinol. 22 (6), 1382–1393. doi:10.1210/me.2007-0490.