An Essential Role of Domain D in the Hormone-Binding Activity of Human /?1 Thyroid Hormone Nuclear Receptor

Kwang-Huei Lin, Clifford Parkison, Peter McPhie, and Sheue-yann Cheng Laboratory of Molecular Biology National Cancer Institute Laboratory of Biochemistry and Metabolism National Institute of Diabetes and Digestive and Kidney Diseases (P.M.) National Institutes of Health Bethesda, Maryland 20892

By analogy with steroid receptors, human placental thyroid hormone nuclear receptor (hTR/?1) could be divided into four functional domains: A/B (Met1Leu101), C (Cys102-Ala170), D (Thr171-Lys237), and E (Arg238-Asp456). The E domain was thought to bind thyroid hormone. To evaluate whether domain E alone is sufficient to bind T3 or requires the presence of other domains for functional T3-binding activity, a series of deletion mutants was constructed. The mutants were expressed in Escherichia coli, and the expressed proteins were purified. Analysis of the T3-binding affinity and analog specificity of the purified truncated hTR/?1 indicated that domain E alone did not have T3-binding activity. Extension of the amino-terminal sequence of domain E to include part of domain D yielded a mutant (Lys201-Asp456) with a Ka for T3 of 0.5 ± 0.2 x 109 M"1. Further extension to include the entire domain D (Met169-Asp456) yielded a mutant with T3-binding activity with a Ka of 0.8 ± 0.1 x 109 M"1. Further extension of the amino-terminal sequence to include domain C increased the affinity for T3 by nearly 2-fold (Ka = 1.5 ± 0.4 x 109 NT1). The Ka for the wild-type hTR/31 is 1.5 ± 0.2 x 109 M"1. Furthermore, mutant (Met169-Asp456) binds to 3',5',3-triiodo-L-thyropropionic acid, D-T3 , L-T4, and L-T 3 with 307%, 37%, 7%, and 0.1%, respectively, of the activity of L-T3. This order of analog affinity is similar to that of the wild-type hTR/?1. These results indicate that domain D is essential for hormone-binding activity. In contrast, the A/B domain is not required for T3-binding activity. Domain C has modulation activity for domains D and E. Deletion of the last eight carboxyl amino acids completely abolishes the T3-binding activity of the mutant (Met169-Asp456). Thus, domain D is essential for 0888-8809/91/0485-0492$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society

domain E to function as a hormone-binding domain. (Molecular Endocrinology 5: 485-492, 1991)

INTRODUCTION Studies on the regulation of gene expression by thyroid hormone have been facilitated by the isolation of cDNAs for thyroid hormone nuclear receptors (1, 2). Sequence analysis of the cDNA isolated from chicken fibroblasts (3); rat brain (4), liver (5), and pituitary (6); and human placenta (7) and muscle (8) indicated that they contain distinct domains similar to those found in the steroid hormone and retinoic acid receptors (9). These receptors consist of an amino-terminal region which is variable in length and sequence (domain A/B), a highly conserved cysteine-rich domain which is involved in DNA binding (domain C), a hydrophilic hinge region (domain D), and a ligand-binding domain (domain E). The structural requirement of domain C for DNA interaction has been characterized by deletion and mutational analysis (10,11). More recently, the three-dimensional structure of domain C in the glucocorticoid receptor has been determined by nuclear magnetic resonance spectroscopy and distance geometry (12). For the ligand-binding domain, additional functional roles have been identified. In addition to its role in hormone binding, its involvement in dimerization and hormone-relieved transcriptional inactivation have begun to be understood (13, 14). Less characterized are the functional roles of domains A/B and D. However, in the estrogen and progesterone receptors, domain A/B has been suggested to differentially enhance the activation of certain genes (15, 16). In chicken thyroid hormone nuclear receptors, an intact N-terminal region containing domain C is important for nuclear localization (17). It also has been shown that deletion of domain D and the first 12 amino acids of domain E eliminated the hormone-binding activity of the in vitro translational products of chicken thyroid hormone nuclear receptor

485

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

MOL ENDO-1991 486

(17). In these studies, however, whether the loss of hormone-binding activity is due to the deletion of domain D or to the deletion of the first 12 amino acids in domain E is unclear. Recently, we have overexpressed the human placental thyroid hormone nuclear receptor (hTR/31) in E. coli and purified it to apparent homogeneity (18). The purified hTR/31 retains its T3-binding affinity and specificity. By affinity labeling of hTR/31 with underivatized [3',5'125 I]T4, followed by partial tryptic digestion, we found that the T3-binding site was in a 25K domain beginning at Phe240 and ending at Asp456. The predicted hormonebinding domain E begins at Arg238 and ends at Asp456 (19, 20). However, the isolated 25K domain (Phe240Asp456) does not bind T3 (18). We postulated that additional structure beyond domain E is required for hormone-binding activity. Immediately preceding domain E is domain D. Prediction of the secondary structure of domain D showed that it consists of three segments, two helical structures (Asp172-Gln200 and Asp211-Gin230) separated by a random coil (Lys^-Thr 210 ) (21). To define which of these structures is essential for the hormone-binding activity of domain E, we constructed a series of mutants and expressed them in E. coli. With a few exceptions, we purified all of the truncated proteins to apparent homogeneity. We analyzed their T3binding affinity and analog specificity. The present study showed that the second helix and the random coil are essential to generate the T3-binding activity of domain E. However, extension of the N-terminal sequence to include the first helix (Asp172-Gln200) and domain C "fine tunes" the T3-binding affinity to as high as that in native hTR/31. Domain A/B, however, is not required for T3binding activity. Deletion of the last helix formed by Leu449-Asp456 at the carboxyl-terminus abolished the hormone-binding activity of hTR/31. Thus, for functional hormone-binding activity, both domain D and the entire domain E must be present.

RESULTS Expression and Purification of the Truncated hTR/91 To define the minimal structural requirements for functional hormone-binding activity of hTR/31, a series of deletion mutants was constructed. Figure 1 shows the size and location of the deletion in relation to the intact hTR/31. The first and second letters in the name represent the one-letter code for the amino- and carboxylterminal amino acids of the mutants, respectively. The number indicates the first two digits of the mol wt. ED41 lacks the A/B domain. MD32 lacks the A/B and C domains, KD29 and DD28 lack the A/B, C, and part of the D domains. In KD25, only the E domain remains. KP24, KH27, and KP28 have the same amino-terminal sequence as KD29, except 42, 20, and 8 carboxylterminal amino acids were deleted, respectively. These mutants were overexpressed in E. coli using a T7 expression system (18). The level of expression for each mutant is shown by comparing the total lysate (Fig. 2A, lanes 2-10) with that of the control BL21/LysS cells (lane 1). In all cases, 80-85% of the expressed

Vol 5 No. 4

proteins were associated with the insoluble inclusion bodies. The remainder were expressed as a soluble form. The truncated proteins were all purified from the inclusion bodies. Lanes 11-19 in Fig. 2A show the Coomassie blue-stained proteins after purification. Except for KP28, in which one additional protein with a higher mol wt of 30,000 was coexpressed and copurified (lane 13, Fig. 2A), all proteins were greater than 95% pure. The identity of the proteins was confirmed not only by the expected mol wt, but also by Western blotting. For ED41, MD32, KD29, DD28, and KD25, we used monospecific antibody C-91, whose antigenic site is located in Cys441-Asp456 of hTR/31 (22). As shown in lanes 3-7 (Fig. 2B), the purified ED41, MD32, KD29, DD28, and KD25 were recognized by C-91, indicating no proteolytic degradation. For the C-terminal deleted mutants, we used a newly developed monoclonal antibody J53, whose antigenic site is in domain E. The characterization of the monoclonal antibody J53 will be described elsewhere. As shown in lanes 8-10 (Fig. 2B), KP28, KH27, and KP24 were recognized by J53. Taken together, these results indicated that these purified proteins are intact. Binding of T3 and Its Analogs to the Intact hTR/31 and Truncated Proteins Expressed in E. coli To understand the relative importance of domain A/B, domain C, and a different part of domain D on the folding of domain E for full hormone-binding activity, the binding of T3 to the purified truncated proteins was evaluated. Figure 3 shows the representative Scatchard plots of the binding data for intact hTR/31, ED41, MD32, and KD29. The affinity constants from three to six repeated experiments, as analyzed by computer (23), are summarized in Table 1. Results of [125I]T3 binding to other truncated proteins are also shown in Table 1. KD25, which has the sequence of the predicted hormone-binding domain, did not bind T3. Neither did DD28, in which 24 more amino acids were extended beyond the predicted hormone-binding domain at the amino-terminus. However, KD29, in which 10 more amino acids were extended at the amino-terminus, bound to T3. It has a Ka of 0.5 ± 0.2 x 109 M"1. MD32, which contained the entire domain D, bound to T3 with an affinity similar to that of KD29 (Ka = 0.8 ± 0.1 x 109 M~1; P > 0.05). A significant increase in the affinity of binding of T3 to ED41 was observed, in which the entire C domain was included (Ka = 1.5 ± 0.4 x 109 M~1; P < 0.01). The extension of ED41 at the amino-terminus to include the entire A/B domain did not further increase its affinity for T3 binding. The Ka for hTR/31 is 1.5 ± 0.2 x 109 M"1, which is identical to that of ED41 (P > 0.05). These results indicate that the 10 amino acids (Lys201Thr210) are essential to generate the T3-binding activity of domain E. However, for maximal binding activity, the entire domain D and domain C are required. In contrast, domain A/B is not required for T3-binding activity. Since all of the truncated proteins were purified from insoluble inclusion bodies which were denatured and renatured, it is possible that the lack of binding of DD28 and KD25 might be an artifact of renaturation. To eliminate this possibility, we also examined the binding of T3 to the soluble form of truncated proteins ex-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

487

Binding of T3 to Deletion Mutants of hTR/31

50 amino acids

Protein C domain

A/B domain

D domain

E domain

h-TR/?1 102

70

pCJ2

456

238

ED41

pJL08

456 MD32

pJL05

KD29

pJL06

DD28

pJC07

KD25

pCJ4

KP28

pCJ5

KH27

pCJ6

KP24

pCJ7

456

100

169

201

456

211

456

456

235

201

448

201

201

Expression Plasmid

436

414

Fig. 1. Schematic Representation of Various Deletion Mutants of hTR/31 The full-length of hTR01 consists of 456 amino acids. By analogy with steroid receptors, the sequences from 1-101,102-170, 171-237, and 238-456 are designated A/B, C, D, and E domains, respectively, according to the method of Green and Chambon (19). The length of each coding region in the expression vectors is shown. The names of the truncated proteins and their corresponding expression plasmids are shown.

pressed in £. coli. The binding data from three to six repeated experiments were evaluated by computer analysis. The Ka values are summarized in Table 1. Consistent with the results obtained from the purified proteins, only hTR/31-S, ED41-S, MD32-S, and KD29S bound T3. Furthermore, there was no significant difference in Ka values for T3 binding between the soluble form and the purified truncated proteins (P > 0.05). These results indicated that the failure of DD28 and KD25 to bind T3 is not due to an artifact from purification; rather, it is due to the absence of essential structural elements. These data further indicated that the purified proteins retained the native structure and showed that the contaminating E. coli proteins present in the lysate have no effects on the binding of T3 to the soluble form of expressed hTR/31 and its truncated proteins. The relative importance of domains A/B, C, and D to the folding of domain E was also probed by the binding of T3 analogs to the truncated proteins. Panels I-IV in Fig. 4 show the competitive binding of four analogs to hTR/31, ED41, MD32, and KD29. The comparison of relative binding affinity is summarized in Table 2. Even though KD29 and MD32 bound T3 with a lower affinity than that of hTR/31 and ED41, the order of potency in the analog was not significantly different from one another, i.e. 3,3',5-triiodo-L-thyropropionic acid > L-T 3 > D-T3 > L-rT3 (P > 0.05). These results further show that the T3-binding sites in hTR/31, ED41, MD32, and KD29 are not grossly different from one another. To evaluate the amino acids essential for T3 binding in the carboxyl-terminus, a series of deletion mutants of KD29 was prepared by deleting 8, 20, and 42 carboxyl-terminal amino acids to give KP28, KH27, and

KP24, respectively. These mutants were expressed and purified (Fig. 2A). Binding studies showed that deletion of as few as eight amino acids (KP28) resulted in the loss of total T3-binding activity (see Table 1). We also have examined T3-binding activity of the soluble protein (KP28-S). No binding activity was observed. Therefore, the loss of T3-binding activity in the purified KP28 was not due to incorrect refolding after renaturation. These results indicated the requirement of the eight amino acid residues at the carboxyl-terminus for T3-binding activity. Our findings are consistent with the earlier reports by Glass et al. (24) and Darling et al. (25), in which loss of T3-binding activity was observed when 35 and 15 amino acids were deleted from hTR/31, respectively. Binding of T3 to Intact hTR/?1 and Truncated Proteins Synthesized by in Vitro Translation To be certain that the lack of T3-binding activity observed for DD28 and KP28 is not due to a possible alteration in protein folding or modification resulting from the expression of eukaryotic proteins in a prokaryotic system, we further evaluated the binding of T3 to hTR/31 and its truncated proteins synthesized by in vitro translation using rabbit reticulocyte lysate. The plasmids pCJ2, pJL08, pJL05, pJL06, pJC07, and pCJ5 were linearized with EcoRI at the 3' end. Their respective mRNAs were prepared by using T7 polymerase. Examination of the size and purity of mRNAs transcribed from pCJ2, pJL08, pJL05, pJL06, pJC07, and pCJ5 by formaldehyde-agarose gel indicated that in all cases, a single species of mRNA with the expected sizes of 1.40, 1.1, 0.86, 0.77, 0.74, and 0.74, respec-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

Vol 5 No. 4

MOL ENDO-1991 488

A. Coomassie Blue-Stained Proteins

II. ED41

KDal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 97- «fj 684329-

m

0.15

0.15

0.1

0.1

0.05

0.05

——

0.05 0.1 Bound (nM) III. MD32 0.2

0.05 0.1 Bound (nM) 0.25 S. IV. KD29 0.2 '

\

0.15

B. Western Blot 0.1

0.1

KDal 68-

1 2 3 4

56

\

7 8 9 10 0.05

4329-

0.1 0.2 Bound (nM)

""""«•—

18-

Fig. 2. Expression and Purification of the Truncated Proteins, as Analyzed by SDS-Polyacrylamide Gel Electrophoresis (A) and Western Blotting (B) A, The total E. coli lysate (0.2 ml cell suspension) and the purified proteins (1-2 ^g) were loaded onto a 15% SDSpolyacrylamide gel. After electrophoresis, gel was stained with Coomassie blue. The order of lysate in lanes 1-10 is BL21/ LysS, hTR/31, KD25, KP28, KH27, KP24, MD32, KD29, DD28, and ED41. The order of purified proteins in lanes 11-19 is hTR/31, KD25, KP28, KH27, KP24, MD32, KD29, DD28, and ED41. B, After SDS-polyacrylamide gel electrophoresis, the proteins (20-40 ng) were transferred onto nitrocellulose paper and reacted with 3 ^g/ml C-91 (22) (lanes 1-7) or 4 ^g/ml monoclonal antibody J53 (lanes 8-10), respectively. The order of the proteins in lanes 1-10 is BL21 /LysS lysate, hTR/31, KD25, MD32, KD29, DD28, ED41, KP28, KH27, and KP24.

tively, was detected (Fig. 5A). Messenger RNA from hTR/31 was also similarly transcribed from peA101, as described by Weinberger et al. (7). The size of the mRNA transcribed from pCJ2 was identical to that transcribed from peA101 (lanes 1 and 2, Fig. 5A). Using these mRNAs, in vitro translation was carried out using reticulocyte lysate, and the translation products were analyzed by sodium dodecyl sulfate (SDS)-polyacrylamide gel. Figure 5B compares the size and purity of intact and truncated proteins purified from E. coli and synthesized from in vitro translation. Lane 2 shows radioactive hTR/31 translated from mRNA prepared from pCJ2. Two major proteins with mol wt of 55K and 52K, and two minor proteins with mol wt of 43K and 32K were detected. The size and abundance of these four proteins were indistinguishable from those synthesized using plasmid peA101, described by Weinberger et al. (7) (lane 1, Fig. 5B). The 55K protein represents the intact hTR/31, as this protein was recognized by both monospecific antibodies N-98 and C-91 (22). The antigenic sites for N-98 and C-91 are located at the N-

0.1

A

0.2 0.3 0.4 Bound (nM)

Fig. 3. Representative Analyses of the Binding of T3 to the Purified hTR/31 and Its Truncated Proteins Purified proteins (5-16 ng/0.24 ml) were incubated with 0.2 nM [125I]T3 in buffer B in the absence or presence of increasing concentrations of unlabeled T3 for 60 min at room temperature. The free and bound [125I]T3 were separated by G-25 column chromatography, as described in Materials and Methods.

Table 1. Affinity Constants for the Binding of T3 to Intact and Truncated h-TR/?1 No. of K. (xio 9 M- 1 y Experiments Purified protein h-TRj81 ED41 MD32 KD29 DD28m KD25 KP28, KH27, KP24 Soluble form (£. coli lysate) h-TR/31-S ED41-S MD32-S KD29-S DD28-S, KD25-S KP28-S, KH27-S, KP24-S

1.5 ± 0 . 2 1.5 ± 0 . 4 0.8 ± 0 . 1 0.5 ± 0.2 No binding No binding

6 4 4 4 2 2

1.4 ± 0 . 5 1.4 ± 0 . 5 0.7 ± 0.04 0.7 ± 0.05 No binding No binding

6 6 3 3 2 2

a

Mean ± SD. Each experiment was carried out in duplicates. Student's r test was used to compare the statistical difference of the Kas.

terminus and C-terminus, respectively (22). Analysis of the reactivity of 52K and 32K proteins with antibodies N-98 and C-91 indicated that these two proteins most likely represent proteins initiated from methionine at positions 27 and 170, respectively. The identity of the 45K protein is unknown. It could result from a degradation product of 55K and/or 52K protein or could represent product from premature termination. Lane 3 shows that the purified hTR/31 from E. coli has the same mol wt as the in vitro translated 55K protein. These results indicated that pCJ2 was as efficiently and correctly transcribed by T7 polymerase as peA101.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

489

Binding of T3 to Deletion Mutants of hTR/31

II. ED41

A . Size and Purity of mRNA „. Kb

1 2

3

4

5

6 7

9.497.464.42.371.35-





0.24B. Size and Purity of Protein 0 0.1

1

10

00.1

100

1

KDal

10

[Analog], nM

Fig. 4. Competitive Binding of [125I]T3 in the Presence of T3 or Its Analogs [125I]T3 (0.2 nM) was incubated with the purified protein (516 ng/0.24 ml) in the presence of increasing concentrations (0.1-100 nM) of 3,3',5-triiodo-L-thyropropionic acid (O), L-T3 (•), D-T3 (A), L-T4 (A), and i_-rT3 (•). Bound and free [125I]T3 were separated by G-25 column chromatography, as described in Materials and Methods.

Table 2. Comparison of the Relative Binding Affinity of T3 and its Analogs to the Intact and Truncated h-TR/31 Relative Binding Affinity0 Tripro"

h-TR/31 ED41 MD32 KD29

396 342 307 246

± ± ± ±

160° 68C 95C 140c

1 2 3 4 5 6 7 8 9 1011 1213

[Analog], nM

o-T 3

27 27 37 43

±15 ±3 ±16 ±12

r-T 3 "

L-T4

4±2 5±2 7±4 5±2

0. 1 0. 1 0. 1 0. 1

± ± ± ±

0.02 0.03 0.03 0.02

Values are mean + so (n = 4). Binding to T3 is 100. "Tripro, 3,3',5-triiodo-L-thyropropionic acid; r-T3, 3',5',3triiodo-L-thyronine. c Analysis by Student's t test indicated that they are not significantly different from each other (P > 0.05). a

Lanes 4, 6, 8, 10, and 12 show that the major radioactive proteins from in vitro translation had mol wt of 41K, 32K, 29K, 28K, and 28K for ED41, MD32, KD29, DD28, and KP28, respectively. The identities of the minor contaminating proteins were unidentified. The corresponding proteins expressed and purified from E. coli shown in lanes 3 , 5 , 7 , 9 , 1 1 , and 13 were visualized by Coomassie blue staining (Fig. 5B). Comparison of the sizes of proteins in pairs in lanes 2-3, 4-5, 6-7, 8 9,10-11, and 12-13 showed that the mol wt of the in vitro translated intact hTR/31 and its truncated proteins

43-

"""

2918Fig. 5. A, Comparison of the Size and Purity of mRNA Messenger RNA was transcribed from linearized plasmids, as described in Materials and Methods. Messenger RNA (1 jtg/5 MO was loaded onto a 1.2% formaldehyde-agarose gel and stained with ethidium bromide. Lanes 1-7 are mRNA from peA101, pCJ2, pCJ5, pJL05, pJL06, pJC07, and pJL08, respectively. B, Comparison of the purity and size of the intact hTR/31 and its truncated proteins synthesized by in vitro translation and purified from E. coli. Proteins (0.5-1 ^g) from E. coli were loaded onto a 15% SDS-polyacrylamide gel and stained with Coomassie blue: lane 3, hTRj81; lane 5, ED41; lane 7, MD32; lane 9, KD29; lane 11, DD28; and lane 13, KP28. Five microliters of [35S]methionine-labeled in vitro translated lysate were loaded onto the same gel. The in vitro translated proteins were detected by autoradiography. The translated proteins are: lane 1, hTR/31 from peA101 obtained from Weinberger ef a/. (7); lane 2, hTR/31 from pCJ2; lane 4, ED41; lane 6, MD32; lane 8, KD29; lane 10, DD28; and lane 12, KP28.

are identical to those purified from E. coli. These results indicated that the truncated proteins were correctly translated. The in vitro translated lysate containing hTR/31 or its truncated proteins was used directly for binding studies. The results are shown in Table 3. It is clear that DD28 and KP28 failed to bind T3, whereas hTR/31, ED41, MD32, and KD29 did. These results are identical to

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

Vol 5 No. 4

MOL ENDO-1991 490

Table 3. Binding of [125I]T3 to Intact h-TR/31 and its Truncated Proteins Synthesized by in vitro Translation [ 125 I]T 3 Bound (cpm)

h-TR01 ED41 MD32 KD29 DD28 KP28

Total

Nonspecific

Specific

12,935 12,833 11,301 10,058 690 650

567 505 555 650 710 690

12,368 12,328 10,746 9,408 No No

The linearized plasmids corresponding to the intact h-TR/31 and truncated proteins were transcribed in vitro using T7 RNA polymerase. The size and quality of the transcribed mRNA were analyzed in a formaldehyde-agarose gel. A single species of mRNAs with the expected size were seen (Fig. 5A). In vitro translation was carried out using 0.5 ng mRNA. Binding of [125I]T3 to proteins in lysate was performed as described in Materials and Methods. [125I]T3 bound was normalized by the amounts of protein detected in Fig. 5B and the predicted number of methionines for each protein. The data shown are the average of two experiments using a single lot of reticulocyte lysate.

those from proteins synthesized in E. coli. Since the in vitro translation products are heterogeneous (see lanes 1, 2, 4, 6, 8, 10, and 12 of Fig. 5B) and, furthermore, the extremely low amounts of translated proteins precluded the possibility of their isolation and purification, no attempt was made to compare the affinities of the in vitro translated hTR01, ED41, MD32, and KD29. The relative quantitative data did permit us to conclude that the second helix and random coil (Lys201-Gln230) of domain D and the last carboxyl-terminal amino acids are essential for T3 binding to hTR/?1 synthesized in reticulocyte lysate.

DISCUSSION

To determine whether domain A/B, C, or D plays a functional role in expression of the full T3-binding activity of domain E in hTR/?1, the present study systematically deleted the various domains and evaluated the T3binding activity of the resultant truncated proteins purified from E. coli. Removal of domain A/B has no effect on T3-binding activity, as indicated by similar apparent affinity constants in the binding of T3 to intact hTR/31 and ED41. Deletion of domain C (MD32) lowered the Ka by - 5 0 % . However, domain D plays an essential role in T3-binding activity. Without domain D, domain E alone cannot bind T3. These results are entirely consistent with our previous findings. Partial tryptic digestion of intact hTR/?1 yielded a 25K protein. By amino acid sequencing and immunological studies, this 25K protein was shown to be the intact domain E (Phe240Asp456). However, this 25K protein had no T3-binding activity (18). Thus, additional structural elements are required for domain E to have a functional T3-binding activity. To study the requirement of domain D for the T3binding activity of domain E in greater detail, we sys-

tematically deleted domain D in segments. The strategy in the deletion of domain D was based upon the predicted secondary structure. Domain D consists of three segments, two helical structures (Asp172-Gln200 and Asp211-Gin230) separated by a random coil (Lys201Thr210). Deletion of the first helical structure (Asp172Gln200) resulted in the loss of - 5 0 % of the binding affinity found in intact hTR/?1. Deletion of the first helical structure together with the random coil (Asp172-Thr210) resulted in the loss of total T3-binding activity in DD28. Even though we have demonstrated that the renaturation process used in our purification yielded native hormone-binding domain, as judged by the identical hormone-binding affinities between the renatured and soluble forms of receptor proteins (hTR/31 vs. hTR/31 S, ED41 vs. ED41-S, MD32 vs. MD32-S, and KD29 vs. KD29-S; see Table 1), we further evaluated T3 binding using the soluble form of DD28-S. No T3-binding activity was detected. Therefore, the loss of the T3-binding activity in DD28 was not a purification artifact. Rather, it demonstrated the essential role of the 10-amino acid random coil (Lys^-SerlleGlyHisCysProGluProThr210) for the expression of the T3-binding activity of domain E. Furthermore, the loss of T3-binding activity of DD28 is not due to a possible difference in the folding of a eukaryotic protein in E. coli, because similar results were obtained using the proteins synthesized in a mammalian system (see Results). The critical importance of domain D for hormone binding was also shown in another member of the thyroid hormone nuclear receptor family, chicken TRa1, by Horowitz et al. (17). Even though the critical subregion in domain D of chicken TR«1 was not identified, lack of domain D resulted in the total loss of T3-binding activity. Furthermore, it is of interest to point out that the sequence similarities in the D domain hTR/31 to that in rat TR/31, human TR«1, rat TRa1, and chicken TRa1 are 97%, 70%, 70%, and 78%, respectively. Moreover, analysis of the predicted secondary structures indicated that helical, random coil, and helical structures are similarly present in domain D of rat TR/31 and human, rat, and chicken TR«1. The helical, random coil, and helical structures have a certain flexibility. Therefore, it is reasonable to postulate that when domain D is in the "on" conformation, it signals the hormone to bind. When it is in an "off" conformation, the hormone is released. The "on" and "off" conformations could be affected by the structure of domain C as a result of receptor-DNA interaction or as a result of interaction with some cellular affectors. The latter possibility is supported by the recent findings that domain D is exposed, as it is accessible to trypsin (18), and furthermore by the discovery that nuclear proteins that interact with receptors are present in GH3 cells (26) and liver cells (27). The critical importance of the C-terminal amino acids was also demonstrated in the present study. Removal of as few as eight amino acids abolished hormonebinding activity. The loss of hormone-binding activity in KP28, KH27, and KP24 was not due to the lack of domains A/B and C, as it was clearly demonstrated in the present study that domains A/B and C play a minor role in the hormone-binding activity of domain E. Furthermore, Glass et al. (24) and Darling et al. (25) have shown that removal of 35 or 15 amino acids from the carboxyl-terminus resulted in the total loss of T3-binding

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

Binding of T3 to Deletion Mutants of hTR/31

activity. Analysis of the predicted secondary structure showed that the last eight amino acids form the first of a series of three a-helical structures (Leu449-Asp456, Met437-Pro448, and Lys415-His436) in the C-terminal region of domain E. The deletion of this a-helical structure (Leu449-Asp456) could either destabilize the binding domain or lose the amino acids responsible for direct contact with the T3 molecule. The essential role of domain D and the last eight carboxyl amino acids in the hormone-binding activity of hTR/?1 demonstrated in the present study is entirely consistent with the model proposed by Forman and Samuels (14). Based on their earlier observations, in which deletion of domain D of chicken TR«1 led to the loss of total T3-binding activity (17), and analysis of the sequences of receptors for thyroid hormone, retinoic acid, and vitamin D, Forman and Samuels proposed that there were two critical regions in these receptors for ligand binding. Without either region, the receptor lost its ligand-binding activity. One was designated Ligandi, which encompassed nearly the entire domain D and the first 26-30 amino acids of domain E; the other was designated Ligand2, which consisted of the terminal carboxyl 23-28 amino acids. For hTR/31, Ligand! and l_igand2 were proposed to be in the region of Ser178-Phe267 and Ser^-Asp 456 , respectively. Using truncated proteins synthesized in E. coli or rabbit reticulocyte lysate, the present study identified Lys201-Gln230 and Leu449-Asp456 to be critically important for T3 binding. Our detailed analysis not only lent support to the proposed model, but, more importantly, pinpointed the essential amino acids in regions Ligandi and Ligand2. However, an understanding of the details of structural changes leading to the loss of hormone binding due to the deletion of domain D or the last eight carboxyl amino acids will have to await studies using physiochemical and crystallographic analysis.

MATERIALS AND METHODS Materials 125

[3'- l]T3 (2206 Ci/mmol; 1 Ci = 37 Gigabecquerel) was purchased from DuPont-New England Nuclear (Boston, MA). The gene Amp DNA amplification reagent kit was obtained from Perkin-Elmer Cetus (Norwalk, CT). Sea plaque GTG low melting temperature agarose was obtained from FMC (Rockland, ME). Restriction enzymes were purchased from Bethesda Research Laboratory (Gaithersburg, MD). lsopropyl-/3D-thiogalactopyranoside was obtained from Research Organics, Inc. (Cleveland, OH). L-T 3 , o-T3, L-T 4 , L-rT3) 3,3',5-triiodoL-thyropropionic acid, and phenylmethylsulfonyfluoride were purchased from Sigma (St. Louis, MO). Aprotinin and leupeptin were obtained from Boehringer Mannheim (Indianapolis, IN). The in vitro translation kit was purchased from DuPont/New England Nuclear. Riboprobe Gemini II for in vitro transcription was purchased from Promega (Madison, Wl).

491

sized, purified, and used as primers to synthesize the inserts by the PCR. Amplification was carried out using the PerkinElmer Cetus Gene Amp PCR kit. The inserts from PCR were purified and ligated into the vector prepared from pCJ2 by restriction with A/del and EcoRI to form the vectors described in Fig. 1. The nucleotide sequences at the 5' and 3' ends of the inserts in the expression vectors (pJL08, pJL05, pJL06, pJC07, pCJ4, pCJ5, pCJ6, and pCJ7) were confirmed by DNA sequencing using U.S. Biochemical Sequence kit version 2.0 (Cleveland, OH). The 150-200 nucleotides sequenced from each end were the expected correct sequence (20). In addition, the entire clones of pJC07 and pCJ5, which encode mutants DD28 and KP28, respectively, were sequenced. Six primers, three from each direction, were synthesized and purified to sequence both strands. The primers from 5' to 3' were the T7 promoter, CJLseqi (nucleotides 1171-1189), and CJLseq3 (nucleotides 1154-1171). The primers from 3' to 5' were pCJ universal primer (sequence outside of the insert, CAGCGGTGGCAGCAGCCAA), CJLseq2, and CJLseq4 (complementary sequences of nucleotides 1120-1136 and 13831400, respectively). The above numbering of nucleotide sequences was based on the reported sequence by Weinberger et al. (7). The dGTP was first used in the sequencing reaction mixture. When sequence ambiguity occurred from G-C compressions, sequencing was repeated using the dITP mixtures, which gave clearly identifiable sequence. There was no sequence error resulting from PCR. The entire sequence is the same as that reported recently by Sakurai et al. for hTR/31 (20). Expression and Purification of the Truncated hTR/31 Proteins The expression of the truncated hTR/31 proteins in BL21/LysS cells was under the control of T7 promoter (28). BL21/LysS cells containing the plasmid were grown in 1 liter LB broth with 100 ng/vn\ ampicillin and 15 fig/m\ chloramphenicol to an absorption at 600 nm of 0.45-0.55. The expression of proteins was induced with 1 IDM isqpropyl-/?-D-thiogalactopyranoside for 3-6 h at 37 C. The purification of the expressed proteins from the inclusion bodies was carried out similarly, as described by Lin ef al. (18). After extraction of the expressed proteins with 5 M guanidinium chloride, the extracted proteins were renatured in buffer R [50 mM Tris-HCI (pH 8.0), 1 IDM dithiothreitol, 20% glycerol, 1 ng/m\ leupeptin, 0.5 mM phenylmethylsulfonylfluoride, and 20 ng/m\ aprotinin] for 18-20 h. Analysis of the purity of renatured proteins by SDS-polyacrylamide gel electrophoresis indicated that the purity was greater than 95%. These proteins were stored frozen at - 7 0 C. The binding activity is stable for 2-3 months. In Vitro Transcription and Translation of Wild-Type and Mutated hTR01 The in vitro transcription/translation was carried out similarly to that described by Weinberger ef al. (7). The wild-type pCJ2 and its mutants (pJL08, pJL05, pJL06, pJC07, and pCJ5) were linearized with EcoRI, and their mRNAs were prepared using transcription kit Riboprobe Gemini II. The mRNAs synthesized were analyzed on a 1.2% formaldehyde-agarose gel. The in vitro transcribed mRNAs (500 ng) were translated using NEN translation kit. The in vitro translation products were analyzed by a 15% SDS-polyacrylamide gel electrophoresis, followed by autoradiography. The band intensity was quantified by a desitometer. Binding of T3 and Its Analogs to hTR/31 and Truncated hTR/81

Construction of the Expression Plasmids The expression vector (pCJ2), which contains the entire coding sequence of hTR/31 (18), was restricted with A/del and EcoRI. The inserts with various deletions, as shown in Fig. 1, were prepared by polymerase chain reaction (PCR), using peA101 as a template. Oligomers (32-39 bases) containing an A/del site with 15-20 bases of the 5' coding sequence of the mutants or containing an EcoRI site together with the 15-20 bases of the 3' coding sequence of the mutants were synthe-

Frozen hTR/?1 or truncated hTR/31 proteins were thawed and dialyzed against 500 ml buffer B [50 mM Tris (pH 8.0), 0.2 M NaCI, 0.01% Lubrol, 1 mM dithiothreitol, and 10% glycerol] for 3 h at 4 C, with one change of dialysate. The binding of [ 5I] T3 to hTR/81 or truncated hTR/J1 proteins was carried out in the presence or absence of analogs, as described previously (18). Free or [125l]T3-bound hTR/31 was separated on a Sephadex G-25 (fine) column, as previously described (18). Analysis of T3 binding to the in vitro translation products was carried

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

Vol 5 No. 4

MOL ENDO-1991 492

out similarly, except that 5 iA of the translation products were used directly, similarly to that previously described (7). 13.

Acknowledgments Received September 20, 1990. Revision received January 23,1991. Accepted January 31,1991. Address requests for reprints to: Sheue-yann Cheng, Ph.D., Building 37, Room 4B09, National Cancer Institute, Bethesda, Maryland 20892.

REFERENCES 1. DeGroot LJ, Nakai A, Sakurai A, Macchia E 1989 The molecular basis of thyroid hormone action. J Endocrinol Invest 12:843-861 2. Samuels HH, Forman BM, Horowitz ZD, Ye ZH 1989 Regulation of gene expression by thyroid hormone. J Clin Invest 51:623-639 3. Sap J, Munoz A, Damm K, Goldberg Y, Ghysdael J, Leutz A, Beug H, Vennstrom B 1986 The c-ertb-A protein is a high-affinity receptor for thyroid hormone. Nature 324:635-640 4. Thompson CC, Weinberger C, Lebo R, Evans RM 1987 Identification of a novel thyroid hormone receptor expressed in the mammalian central nervous system. Science 237:1610-1614 5. Murray MB, Zilz ND, McCCreary NL, McDonald MJ, Towle HC1988 Isolation and characterization of rat cDNA clones for two distinct thyroid receptors. J Biol Chem 263:12770-12777 6. Hodin RA, Lazar MA, Wintman Bl, Darling DS, Koenig RJ, Larson PR, Moore DD, Chin WW 1989 Identification of a thyroid hormone receptor that is pituitary-specific. Science 244:76-78 7. Weinberger C, Thompson CC, Ong ES, Lebo R, Gruol DJ, Evans RM 1986 The c-erb-k gene encodes a thyroid hormone receptor. Nature 324:641-646 8. Nakai A, Sakurai A, Bell Gl, DeGroot LJ 1988 Characterization of a third human thyroid hormone receptor coexpressed with other thyroid hormone receptors in several tissues. Mol Endocrinol 2:1087-1092 9. Evans RM 1988 The steroid and thyroid hormone receptor superfamily. Science 240:889-895 10. Umesono K, Evans RM 1989 Determinants of target gene specificity for steroid/thyroid hormone receptors. Cell 57:1139-1146 11. Mader S, Kumar V, Verneuil H, Chambon P 1989 Three amino acids of the oestrogen receptor are essential to its ability to distinguish an oestrogen from a glucocorticoidresponsive element. Nature 338:271 -274 12. Hard T, Kellenbach E, Boelens R, Maler BA, Dahlman K, Freedman LP, Carlstedt-Due J, Yamamoto KR, Gustafsson F-A, Kaptein R 1990 Solution structure of the gluco-

14. 15. 16.

17.

18.

19. 20. 21.

22.

23. 24.

25.

26.

27.

28.

corticoid receptor DNA-binding domain. Science 249:157-160 Beato M 1989 Gene regulation by steroid hormones. Cell 56:335-344 Forman BM, Samuels HH 1990 Interactions among a subfamily of nuclear hormone receptors: the regulatory zipper model. Mol Endocrinol 4:1293-1301 Kumar V, Green S, Stack G, Berry M, Jia J-R, Chambon P 1987 Functional domains of the human estrogen receptor. Cell 51:941-951 Tora L, Gronemeyer H, Turcotte B, Gaub M-P, Chambon P1988 The N-terminal region of the chicken progesterone receptor specific target gene activation. Nature 313:185188 Horowitz ZD, Yang C, Forman BM, Casanova J, Samuels HH 1989 Characterization of the domain structure of chick c-enbA by deletion mutation: in vitro translation and cell transfection studies. Mol Endocrinol 3:148-156 Lin KH, Fukuda T, Cheng S-Y 1990 Hormone and DNA binding activity of a purified hormone thyroid nuclear receptor expressed in Escherichia coli. J Biol Chem 265:5161-5165 Green S, Chambon P 1986 A superfamily of potentially oncogenic hormone receptors. Nature 324:615-617 Sakurai A, Nakai A, De Groot LJ 1990 Structural analysis of human thyroid receptor 0 gene. Mol Cell Endocrinol 71:83-91 Gamier J, Osguthorpe DJ, Robson B, 1978 Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J Mol Biol 120:97-120 Fukuda T, Willingham MC, Cheng S-y 1988 Antipeptide antibodies recognize c-erbA and a related protein in human A431 carcinoma cells. Endocrinology 123:26462652 Knott GD 1979 MLAB-a mathematical modeling tool. Comput Progr Biomed 10:271-278 Glass CK, Lipkin SM, Devary OV, Rosenfeld MG 1989 Positive and negative regulation of gene transcription by a retionic acid. Thyroid hormone receptor heterodimer. Cell 59:697-708 Darling DS, Beebe JS, O'Donnel AC, Bumside J, Koenig RJ, Chin WW, The T3 receptor-binding-enhancement factor (BEF) bindins to the hormone binding domain of the T3 receptor and to DNA. 72nd Annual Meeting of The Endocrine Society, Atlanta GA, 1990, p 358 (Abstract) Bumside J, Darling DS, Chin WW 1990 A nuclear receptor that enhances binding of thyroid hormone receptors to thyroid hormone response elements. J Biol Chem 265:2500-2504 Murray MB, Towle HC 1989 Identification of nuclear factors that enhance binding of the thyroid hormone receptor to a thyroid hormone response element. Mol Endocrinol 3:1434-1442 Studier FW, Rosenberg AH, Dunn JJ, Dubendorff JN 1990 Use of T7 RNA polymerase to direct expression of cloned genes. In: Goeddel DV (ed) Methods of Enzymology. Academic Press, San Diego, vol 185:60-89

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 May 2015. at 05:27 For personal use only. No other uses without permission. . All rights reserved.

An essential role of domain D in the hormone-binding activity of human beta 1 thyroid hormone nuclear receptor.

By analogy with steroid receptors, human placental thyroid hormone nuclear receptor (hTR beta 1) could be divided into four functional domains: A/B (M...
1MB Sizes 0 Downloads 0 Views