Báo cáo khoa học: Change in structure of the N-terminal region of transthyretin produces change in affinity of transthyretin to T4 and T3 pdf

Change in structure of the N-terminal region
of transthyretin produces change in affinity
of transthyretin to T4 and T3
Porntip Prapunpoj
1
, Ladda Leelawatwatana
1
, Gerhard Schreiber
2
and Samantha J. Richardson
2,3
1 Department of Biochemistry, Faculty of Science, Prince of Songkla University, Hat-Yai, Songkhla, Thailand
2 Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, The University of Melbourne,
Parkville, Victoria, Australia
3 UMR CNRS 5166, Evolution des Re
´
gulations Endocriniennes, Muse
´
um National d’Histoire Naturelle, Paris, France
Transthyretin (TTR) is one of the three thyroid hor-
mone distributor proteins found in the plasma of lar-
ger mammals and was first described in human serum
and cerebrospinal fluid (CSF) in 1942 [1,2]. In humans,
the main sites of TTR synthesis are the liver and chor-
oid plexus. TTR synthesis has been described in all
classes of vertebrates [3]. TTR is composed of four
identical subunits [4] and, in humans, has a molecular
mass of 55 kDa. In most vertebrates, the TTR subunit
consists of 127 amino acid residues [5,6]. The tetramer
has a central channel with two thyroid hormone bind-
ing sites [4]. However, only one binding site of TTR

is occupied by thyroid hormone under physiological
conditions [7,8], due to negative co-operativity [9].
Although all amino acid residues reported to partici-
pate in the binding of thyroid hormones in human
TTR are 100% conserved in other vertebrate TTRs,
the binding affinity to thyroid hormones varies among
animal species: TTR from fish, amphibians, reptiles
and birds binds 3¢,5,3-[l]-triiodothyronine (T3) with
Keywords
N-terminal sequence; protein evolution;
recombinant transthyretin; thyroid hormone-
binding plasma proteins; thyroid hormone
Correspondence
P. Prapunpoj, Department of Biochemistry,
Faculty of Science, Prince of Songkla
University, Hat-Yai, Songkhla 90112,
Thailand
Fax: +66 74 446656
Tel: +66 74 288275
E-mail:
(Received 4 May 2006, revised 14 June
2006, accepted 3 July 2006)
doi:10.1111/j.1742-4658.2006.05404.x
The relationship between the structure of the N-terminal sequence of trans-
thyretin (TTR) and the binding of thyroid hormone was studied. A recom-
binant human TTR and two derivatives of Crocodylus porosus TTRs,
one with the N-terminal sequence replaced by that of human TTR
(human ⁄ crocTTR), the other with the N-terminal segment removed (trun-
cated crocTTR), were synthesized in Pichia pastoris. Subunit mass, native
molecular weight, tetramer formation, cross-reactivity to TTR antibodies

and binding to retinol-binding protein of these recombinant TTRs were
similar to TTRs found in nature. Analysis of the binding affinity to thyroid
hormones of recombinant human TTR showed a dissociation constant (K
d
)
for triiodothyronine (T3) of 53.26 ± 3.97 nm and for thyroxine (T4) of
19.73 ± 0.13 nm. These values are similar to those found for TTR purified
from human serum, and gave a K
d
T3 ⁄ T4 ratio of 2.70. The affinity for T4
of human⁄ crocTTR (K
d
¼ 22.75 ± 1.89 nm) was higher than those of
both human TTR and C. porosus TTR, but the affinity for T3 (K
d
¼
5.40 ± 0.25 nm) was similar to C. porosus TTR, giving a K
d
T3 ⁄ T4 ratio
of 0.24. A similar affinity for both T3 (K
d
¼ 57.78 ± 5.65 nm) and T4
(K
d
¼ 59.72 ± 3.38 nm), with a K
d
T3 ⁄ T4 ratio of 0.97, was observed for
truncated crocTTR. The obtained results strongly confirm the hypothesis
that the unstructured N-terminal region of TTR critically influences the
specificity and affinity of thyroid hormone binding to TTR.

Abbreviations
CSF, cerebrospinal fluid; RBP, retinol-binding protein; T3, 3¢,5,3-[
L]-triiodothyronine; T4, 5¢,3¢,5,3-[L]-tetraiodothyronine; TTR, transthyretin.
FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS 4013
higher affinity, whereas TTR from mammals binds
5¢,3¢,5,3-[l]-tetraiodothyronine (T4) with higher affinity
[10–14]. The affinity to T4 increased while the affinity
to T3 decreased during the evolution of mammalian
TTRs from its ancestors.
TTRs from 20 vertebrate species, including mam-
malian, avian, reptilian, amphibian and fish, were iso-
lated and their cDNAs were cloned and sequenced
[6,15]. Maximum parsimony analysis of the derived
amino acid sequences produced phylogenetic trees of
a structure and branching similar to those found in
phylogenetic trees based on morphology of animals
[14,16]. Comparison of derived amino acid sequences
revealed that the amino acid residues involved in
the binding of TTR to thyroid hormones remained
unchanged during evolution of TTR in vertebrates
[see 4,17]. However, the most marked changes in
TTR during vertebrate evolution are concentrated in
the N-terminal region of the TTR subunit. The N-ter-
minal segment is longer and more hydrophobic in
avian, reptilian, amphibian and fish than in mamma-
lian TTRs. X-ray crystallographic studies revealed
that the four N-terminal regions of TTR are unstruc-
tured, protrude from the protein tetramer and are
located near the entrances to the central channel con-
taining the thyroid hormone binding sites [18,19]. Our

previous work [6,15] has shown a systematic change
during evolution in the N-terminal region of TTR
from longer and more hydrophobic to shorter and
more hydrophilic in character. The affinity of TTR to
T3 and T4 seems also to have changed unidirectionally
during evolution [12]. The question arises as to whether
there is a causal relationship between these two types of
changes. The work reported here is a planned quantita-
tive analysis in vitro of the relationship between the two
types of changes.
In a previous report [14], we demonstrated that the
binding affinity of T3 and T4 to crocTTR changed
when its N-terminal segment was replaced by that of
Xenopus laevis TTR. Here we report the synthesis of
two recombinant TTRs in Pichia pastoris which were
designed to test specifically the relationship between
N-terminal structure of TTR and binding of thyroid
hormones. In one of these TTRs, the crocTTR N-ter-
minal region was replaced by the N-terminal region of
human TTR (whereas crocTTR preferentially binds
T3, human TTR preferentially binds T4). In the other,
the N-terminal region of crocTTR was deleted. The
affinity of T3 and T4 to both TTRs was studied and
the results are a powerful demonstration of the rela-
tionship between the evolution of the primary structure
of the N-terminal regions of TTR and the biological
function of TTR.
Results
Synthesis of recombinant human TTR
In the crocodile, TTR is only synthesized in the liver

during development in the egg and in hatchlings and
only in the choroid plexus of the adult. Thus, it is
difficult to obtain the native crocTTR in sufficient
amounts for thyroid hormone binding assays. To
ensure that recombinant produced in P. pastoris has
similar affinity to thyroid hormones as native TTR,
recombinant human TTR was synthesized in Pichia
was analyzed for affinity to T3 and T4, and affinities
were compared to those of human TTR purified from
serum.
Recombinant human TTR was produced from the
cDNA construct in pPIC3.5 but not in pPIC9 (Fig. 1),
and had a subunit mass of 15 kDa by SDS ⁄ PAGE,
that corresponds to that of TTR purified from human
serum. The N-terminal sequences of recombinant
human TTR was as expected, i.e. G P T G.
Synthesis of recombinant chimeric TTRs
Chimeric TTRs were amplified by PCR using specific
primer pairs to generate new N-terminal sequences
with compatible restriction ends for ligation into the
pPIC9 vector (Fig. 2). The transformation efficiency
(10
3
)10
4
transformants per 1 lg DNA) was greater
than that for recombinant human TTR. The two
recombinant crocodile TTRs had masses of 15 kDa
94
67

45
30
21.1
14.4
M
r
(kDa)
12345678
TTR
Fig. 1. Expression of recombinant human TTR. Pichia transformant
clones containing human TTR cDNA inserted in pPIC 3.5 or pPIC 9
were grown and induced with methanol for 4 days. Supernatant of
the yeast culture was collected and aliquots of 90 lL were ana-
lyzed by SDS ⁄ PAGE (15% resolving gel) and protein bands were
detected by silver staining. Positions of protein markers and TTR
are indicated. Numbers under lanes indicate individual clones with
DNA inserted in pPIC3.5 (lanes 1–4) and pPIC9 (lanes 5–8).
Function of transthyretin N-terminus P. Prapunpoj et al.
4014 FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS
by SDS ⁄ PAGE. The N-terminal sequence of recombin-
ant truncated crocTTR was as expected, i.e. S K C P.
Two additional amino acid residues E and A were
found at the N-terminal sequence of human ⁄ crocTTR,
i.e. E A G P.
Physicochemical properties of the recombinant
TTRs
P. pastoris expression systems have the potential to
perform many of the post-translational modifications
typical for higher eukaryotes. Some of these slightly
differ from those in mammals. For example, carbohy-

drate moieties added to secreted proteins in P. pastoris
are predominantly or entirely composed of mannose
residues. Moreover, some foreign proteins synthesized
in P. pastoris are hyperglycosylated [20]. These mecha-
nisms of post-translation differ between P. pastoris
and higher eukaryotes, and can lead to alterations of
properties and ⁄ or function of the recombinant pro-
teins. To ascertain whether unwanted post-transla-
tional modifications occurred to the recombinant
TTRs in P. pastoris, properties of the proteins were
analyzed.
Binding to RBP
Recombinant human TTR and chimeric TTRs bound
to RBP similarly to previously reported for other
TTRs [13,14] (Fig. 3). This facilitated in the purifica-
tion of TTRs from the P. pastoris culture medium.
Approximately 2 mg of purified recombinant human
TTR and up to 4 mg of purified chimeric TTRs were
obtained from 1 L of P. pastoris culture supernatant
by a single purification step of affinity chromatography
using an RBP-Sepharose column.
A
B
Fig. 2. Expression vectors for chimeric TTR
genes. The expression plasmids for (A)
human ⁄ crocTTR and (B) truncated crocTTR
were constructed in pPIC9. Proteins were
synthesized and secreted using the a-factor
protein presegment of Pichia. 5AOX1, pro-
motor of P. pastoris alcohol oxidase 1 gene;

(3)TT, native transcription termination and
polyadenylation signal of alcohol oxidase 1
gene; 3AOX1, sequence from the alcohol
oxidase 1 gene, 3¢ to the TT sequences;
HIS4; histidinol dehydrogenase gene; Amp,
ampicillin resistance gene, ColE1, Escheri-
chia coli origin of replication; SalI, SalI
restriction site for linearization of the vector.
Numbering of amino acid residues, based
on that for human TTR [31], was provided
underneath the amino acid sequence. (black
shading, fragment of human TTR cDNA and
grey shading, fragment of crocTTR cDNA.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 5 10 15 20 25
Elution volume (ml)
mn 082 ta ecnabrosbA
Distilled water
Fig. 3. Chromatography of recombinant TTR on a human RBP-
Sepharose affinity column. Cultured Pichia producing recombinant
human TTR was collected after induction for 4 days. Two milliliters
of the supernatant were loaded onto a column of human RBP-
Sepharose (1 mL of gel) equilibrated in 0.04
M Tris ⁄ HCl, pH 7.4 buf-

fer containing 0.5
M NaCl. Bound protein was eluted with distilled
water. The chromatographic separation was carried out at 4 °C.
P. Prapunpoj et al. Function of transthyretin N-terminus
FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS 4015
Mobility in SDS/PAGE
The analysis in SDS ⁄ PAGE of purified recombinant
human TTR and chimeric TTRs showed that the
recombinant proteins have the relative mobility similar
to those of TTRs from other vertebrate species
(Fig. 4). The subunit molecular masses determined
from the calibration curve were 17 kDa for all
recombinant TTRs. The subunit masses obtained were
similar to those of native TTRs from human and other
vertebrates [14,21], indicating no aberrant post-transla-
tional modification occurred with these recombinant
TTRs.
Mobility in nondenaturing gels
Most TTRs from vertebrates including humans [22]
and birds [12] migrate faster than albumin during elec-
trophoresis at pH 8.6. Only for some eutherian species,
such as pigs and cattle, TTRs comigrate with albumin
in nondenaturing gels [23,24]. The mobilities of all
recombinant TTRs in nondenaturing gel were greater
than those of albumin and similar to TTR from
human plasma (Fig. 5).
Immunochemical cross-reactivity with antibodies
against TTRs
Immunochemical cross-reactivity of TTRs from several
vertebrates is well known [13,14,25]. Two bands were

observed in the test of the immunochemical cross-reac-
tivity of each the recombinant TTR with antiserum
raised against TTRs purified from serum of human,
chicken and wallaby (Fig. 6). The major protein band,
also detected by staining with Coomassie blue, was
97
66
45
30
20.1
14.4
M
r
(kDa)
3
A
2
1
M
Fig. 4. Analysis by SDS ⁄ PAGE and determination of the size of the
subunit of recombinant TTRs. Aliquots of purified (1) recombinant
human TTR, (2) human ⁄ crocTTR, and (3) truncated crocTTR were
boiled for 30 min in 2.5% 2-mercaptoethanol and 2% SDS, prior to
analysis by SDS ⁄ PAGE. Proteins were stained with Coomassie
blue. The positions of protein markers (M) are indicated. The sizes
of the TTR subunits were obtained by comparison of electrophoretic
mobility with protein markers, plotting the relative mobilities (R
f
)
against the logarithmic values of the markers.

TTR
HSA
3
2
1
HP
Fig. 5. Electrophoretic mobility pattern of recombinant TTRs. The
purified recombinant (1) human TTR, (2) human ⁄ crocTTR, and (3)
truncated crocTTR were separated in a nondenaturing polyacryla-
mide gel (10% resolving, 4% stacking). Protein bands were
visualized by Coomassie blue staining. Human plasma (HP) was
overloaded to show the human TTR (TTR). The position of serum
albumin (HSA) is also indicated.
M
r
(kDa)
14.4
21.1
30
45
67
94
Co
oWest
TTR monomer
TTR dimer
M1
2
312
3

Fig. 6. Cross-reactivity with antiserum against a mixture of TTRs.
The purified recombinant human TTR (1), human ⁄ crocTTR (2), and
truncated crocTTR (3) were analyzed by SDS ⁄ PAGE and electro-
phoretically transferred to nitrocellulose membrane. The protein
bands were stained with Coomassie (Coo) and identified by reac-
tion with antiserum (West). The membrane filter was incubated
with rabbit antiserum against a mixture of human, wallaby and
chicken TTRs (1 : 5000) followed by anti-rabbit immunoglobulin
(1 : 10000) conjugated with horseradish peroxidase. Detection was
carried out by enhanced chemiluminescence. The positions of the
TTR monomer and dimmer are indicated.
Function of transthyretin N-terminus P. Prapunpoj et al.
4016 FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS
found in the same position as that of the subunit
of recombinant TTRs. Another band, barely visible,
migrating more slowly, was found in a position corres-
ponding to a molecular mass of about 30 kDa, consis-
tent with being a dimer of TTR. Such dimers were
always seen as faint bands when denaturation of TTR
was not complete, even under conditions of harsh dena-
turation [13,14].
Native molecular mass of the recombinant TTR
tetramers
TTRs purified from P. pastoris culture supernatant
were analyzed by HPLC on a Bio-Sil SEC250 (Bio-
Rad Laboratories, Inc., Hercules, CA, USA) column
in 50 mm potassium phosphate buffer saline, pH 7.4.
Proteins with known molecular masses were used to
calibrate the column. In comparison to the standard
curve, purified recombinant human TTR showed a

molecular mass of 57 kDa, whereas human ⁄ crocTTR
and truncated crocTTR had molecular masses of 63
and 57 kDa, respectively. These molecular masses were
approximately four times the subunit mass for each
TTR species. This strongly suggested that P. pastoris
folded the TTR subunits correctly and that tetramer
was correctly assembled.
Binding affinities of recombinant TTRs
to thyroid hormones
The dissociation constants of the complex of the
recombinant TTRs with thyroid hormones were deter-
mined using the highly reproducible, rapid and sensi-
tive method developed by Chang et al. [12]. The
experiments were performed in triplicate for each hor-
mone. Binding curves were plotted following the gen-
eral equation according to Scatchard [25].
The K
d
values of chicken TTR derived from the
Scatchard analysis and plots, both for T4 and for T3,
were similar to those previously reported [14] (data not
shown). Recombinant human TTR showed a higher
affinity for T4 than for T3, and values similar to those
reported for TTR from human serum [12]. The K
d
for
T4 of the recombinant human TTR was 19.73 ±
0.13 nm and that for T3 was 53.26 ± 3.97 nm, giving
a K
d

T3 ⁄ K
d
T4 ratio to 2.70 (Fig. 7A, B and G). The
binding capacity of the recombinant TTR derived from
the abscissa intercepts both for T4 and for T3 sugges-
ted a capacity of two molecules of thyroid hormones
per TTR molecule.
In comparison, the recombinant chimeric TTRs, i.e.
human ⁄ crocTTR and truncated crocTTR, possessed
different binding affinities to T3 and T4, as well as K
d
T3 ⁄ K
d
T4 ratios. The human ⁄ crocTTR had K
d
values of
5.40 ± 0.25 nm and 22.75 ± 1.89 nm for T3 and T4,
respectively, providing a K
d
T3 ⁄ K
d
T4 ratio of 0.24
(Fig. 7C, D and G). This ratio was higher than that
reported for Crocodylus porosus TTR [14]. Because the
K
d
for T3 of the human ⁄ crocTTR was not significantly
different from that of C. porosus TTR (7.56 ± 0.84 nm)
[14], the higher K
d

T3 ⁄ K
d
T4 ratio could indicate greater
influence of the N-terminal change on binding to T4
than to T3. The K
d
values for both T3 and T4 of the
truncated crocTTR were similar. Truncated crocTTR
bound to T3 with a K
d
of 57.78 ± 5.65 nm and to T4
with a K
d
of 59.72 ± 3.38 nm (Fig. 7E,F and G), lead-
ing to a K
d
T3 ⁄ K
d
T4 ratio to 0.97. For a summary of
K
d
values and ratios, see Fig. 7G.
Discussion
The binding of thyroid hormones is one of the main
functions of TTR, which is functionally integrated with
albumin and thyroxine-binding globulin as a network
system to ensure the appropriate extracellular and
intracellular distribution of thyroid hormones. In the
network, a deficiency in one component can be com-
pensated for by the other components [6].

The affinities of T3 and T4 for TTRs from verteb-
rate species vary considerably, in that TTRs from fish
[11], amphibians [10,13], reptiles [14] and birds [12]
bind T3 with higher affinity than T4, whereas TTRs
from mammals bind T4 with higher affinity than T3
[12]. Paradoxically, the amino acids in the thyroid hor-
mone binding sites of TTR that are involved with the
interaction of TTR with the thyroid hormones are
100% conserved throughout vertebrate TTRs [26].
Examination of the alignment of TTR amino acid
sequences from 20 vertebrate species revealed that the
region of the subunit which changed in a distinct
and directed manner was the N-terminal region. The
‘N-terminal region’ is defined as the amino acids from
the N-terminus until the Cys residue which is the first
to be unambiguously defined by electron density in
X-ray crystal structures (Cys10 in human TTR) and
considered part of the core structure of TTR. In gen-
eral, the character of the N-terminal region of TTRs
changed from longer (14 amino acids) and more
hydrophobic to shorter (nine amino acids) and more
hydrophilic (Fig. 8). These changes could be correlated
with the change from preferential binding of T3 (N-
terminal region that are longer and more hydrophobic)
to preferential binding of T4 (N-terminal region that
are shorter and more hydrophilic) [12]. Two N-ter-
minal regions of TTR are located around each
entrance to the central channel that contains the two
P. Prapunpoj et al. Function of transthyretin N-terminus
FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS 4017

thyroid hormone binding sites (Fig. 9) [18]. To the best
of our knowledge, the only X-ray crystal structure
demonstrating electron density for the N- and C-ter-
minal regions is that of Hamilton et al. 1993 [18], and
there have not been any direct structural analyses of
the interaction of TTR N-terminal region with T3 or
T4 directly. However, there are several indications in
the literature that thyroid hormones interact with the
N-terminal regions of TTR, e.g. Cheng et al. [27] dem-
onstrated that N-bromoacetyl-l-T4 interacts with
Gly1, Lys9 and Lys15 of human TTR; and the Gly6-
Ser mutant of human TTR has a higher affinity for T4
than wild type TTR [28]. However, the synthesis of
chimeric TTRs allowed to directly testing the hypothe-
sis that the N-terminal regions of TTR affect the
affinities of TTR for thyroid hormones. A previous
study revealed that altering the structure of the N-ter-
minal region influenced the affinity of thyroid hor-
mones for TTR [14]. However, that study used a
chimera of two species of TTR (X. laevis and C. poro-
sus), both of which preferentially bound T3 over T4.
Here, we tested that hypothesis that the structure of
the N-terminal region influence the binding of T3 and
T4 to TTR by comparing the affinities of T3 and T4 to
0
2
4
6
8
10

12
14
0 1020304050607080
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 5 10 15 20 25 30 35 40 45 50
0.0
0.1
0.2
0.3
0.4
0.5
0 5 10 15 20 25
0.0
0.1
0.2
0.3
0.4
0246810121416
0.0
0.5
1.0
1.5
2.0
2.5

3.0
0 5 10 15 20 25 30 35 40 45
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 5 10 15 20 25 30 35 40
C
D
BA
E
F
[bound T3], nM
[bound T4], nM
[bound T3], nM
[bound T3], nM
[bound T4], nM
[bound T4], nM
]3T dnuob-RTT[
]3T eerf[
]4T dnuob-RTT[
]4T e
e
rf[
]3T dnuob-RTT[
]3T eerf[

]3T dnuob-RT
T[
]3T ee
r
f[
]4T dnuob-RTT[
]4T
e
er
f
[


]4T dnuob-RTT[
]
4
T
e
er
f
[






G
Kd for T3 (nM) Kd for T4 (nM) KdT3/KdT4 Reference
Human TTR

53.26
±
3.97 19.73
±
0.13
2.70 This paper
Human/crocTTR
5.40
±
0.25 22.75
±
1.89
0.24 This paper
Truncated crocTTR
57.78
±
5.65 59.72
±
3.38
0.97 This paper
C. porosus TTR
7.56
±
0.84 36.73
±
2.38
0.21
Prapunpoj
et al., 2002
Fig. 7. Binding of recombinant TTRs to thy-

roid hormones. One hundred nanomoles of
recombinant human TTR (A and B),
human ⁄ crocTTR (C and D) and truncated
crocTTR (E and F) were incubated with
125
I-
T3 or
125
I-T4 in the presence of various con-
centrations of unlabeled hormone at 4 °C,
overnight. Free hormone was separated
from the TTR-bound hormone by filtering
the incubation mixture through a layer of
methyl cellulose-coated charcoal under
vacuum. All corrections including those for
nonspecific binding were applied before per-
forming the Scatchard analysis. The plots
for the affinity (K
d
), for T3 and T4 of the pro-
teins were calculated and summarized (G).
Function of transthyretin N-terminus P. Prapunpoj et al.
4018 FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS
a TTR that has higher affinity for T4 (human TTR), a
TTR that has higher affinity for T3 (crocodile TTR),
and chimeric TTR consisting of the human TTR
N-terminal region and the ‘rest of the molecule’ being
crocodile TTR (human ⁄ croc TTR) and crocodile TTR
without N-terminal region (truncated croc TTR).
In vitro synthesis of TTR was desirable for two main

reasons. Firstly, although the similarity of the phylo-
genetic trees based on TTR structure and those based
on morphological structure of animals suggest that the
structure of TTR evolved under functional pressures
providing an advantage in selection, it is difficult to
quantitatively analyze in vivo details of the relationship
between TTR structure and function using genetic
alterations or specific inhibitors. The strong redundan-
cies in the network of reactions involved in determin-
ing thyroid hormone distribution, of which TTR is a
part, renders the system so effective that a deficiency
in one component is readily compensated in vivo by
changes in the rest of the system. Second, the expres-
sion of the TTR gene in some animal species some-
times occurs in only one specific organ. TTR is only
synthesized by the liver (and therefore can be purified
from blood) of crocodiles during development in ovo
and in hatchlings [3]. TTR is synthesized by adult cro-
codile choroid plexus [14], but the volume of CSF
required to be collected to enable purification of suffi-
cient amounts of TTR for thyroid hormone binding
assays renders this unfeasible. Therefore, to obtain suf-
ficient amounts of TTR for further characterization, a
heterologous gene expression system is needed. To
determine the influence of the N-terminal region of
TTR on the binding to thyroid hormones, the con-
struction of chimeric recombinant TTRs containing
variations in structure of the N-terminal region is
desired. The functional properties (i.e. binding of T3
and T4) of such chimeric TTRs thus could be analyzed

for providing insight into the relationship between
structure and function.
Fig. 8. Comparison of amino acid sequence of the N-terminal regions of vertebrate TTRs. Amino acid sequences in the N-terminal regions
from six vertebrate species (human, Homo sapiens; sheep, Ovis aries; Tammar wallaby, Macropus eugenii; grey opossum, Monodelphis
domestica; chicken, Gallus gallus domesticus; salt water crocodile, Crocodylus porosus; Xenopus, Xenopus laevis) are aligned with the
amino acid sequence of human TTR. The sequence is written using the single-letter amino acid abbreviation. Asterisks indicate those resi-
dues in other species with identical amino acids to those in human TTR. Gaps were introduced to aid alignment. Features of secondary
structure of human TTR are indicated above the sequences. Numbering of residues is based on that for human TTR, and -a,-b,-c,-d and -e
were introduced to indicate positions of residues in noneutherians. Double underlining indicates amino acid residues located in the central
thyroid hormone binding site. For sources of the TTR sequences, see [13,16].
Fig. 9. X-ray crystal structure of human TTR dimer. Positions of
N- and C-terminal regions located at the entrances of the central
channel containing the thyroid hormone binding sites are indicated
by pink and white arrows, respectively. Reprinted with permission
of The American Society for Biochemistry and Molecular Biology,
from [24]. Permission conveyed through Copyright Clearance Cen-
tre, Inc.
P. Prapunpoj et al. Function of transthyretin N-terminus
FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS 4019
Recombinant TTRs were secreted into the culture
medium, and purified TTRs migrated in SDS ⁄ PAGE
as a single band with an approximate subunit molecu-
lar mass of 17 kDa, similarly to TTR subunits from
other vertebrate species. Immunochemical cross-reac-
tivity with antibody against other TTRs confirmed that
the bands of 15 kDa were TTRs. The electrophoretic
mobility of recombinant TTRs in nondenaturing poly-
acrylamide gel at pH 8.6 was faster than that of albu-
min, which is a typical characteristic of TTR of most
vertebrates found in nature [21].

Native TTR in plasma exists in the tetrameric form.
The molecular masses of the recombinant proteins
determined by gel filtration analysis were approxi-
mately four times the mass of the subunit determined
by SDS ⁄ PAGE, indicating that the recombinant TTRs
as tetramers. Analysis of binding to retinol-binding
protein and thyroid hormones showed that the recom-
binant TTRs retained its function as binding protein
for retinol-binding protein and thyroid hormones.
In the present study, the K
d
values for T4 and T3 of
all the recombinant TTRs were determined with the
method described by Chang et al. [12]. Recombinant
human TTR bound T4 with higher affinity than T3,
similarly to that previous reported [12]. This confirmed
that the recombinant TTR had folded into its proper
native conformation.
In the present study, we analyzed the binding of T3
and T4 to a chimeric TTR consisting of the N-terminal
regions of crocTTR (which has higher affinity for T3)
and the ‘rest of the molecule’ from human TTR (which
has higher affinity for T4), and to truncated crocTTR
(which lacked N-terminal regions). By this strategy,
the result clearly indicated the involvement of the
N-terminal region of TTR subunits in accessibility of
thyroid hormones to the binding site, as well as the
strength and binding preference of thyroid hormones
to TTR. However, because the truncated crocTTR
(lacking the N-terminal segment) had a K

d
T3 ⁄ K
d
T4
ratio of 1, this indicated that in the absence of N-ter-
minal segment, TTR bound to both T3 and T4 with
the same strength and preference.
Comparison of the data in Fig. 7(G) reveals that,
qualitatively, human ⁄ croc TTR is similar to croc TTR,
in that they both have higher affinity for T3 than for
T4. This could imply that the core of TTR is the main
determinant of ligand affinity and than the N-terminal
region exert a modulatory effect. However, truncated
croc TTR has greatly reduced affinity for both T3 and
T4 compared with either croc TTR or human ⁄ croc
TTR. However, human ⁄ croc TTR has more similar T4
affinity to human TTR than croc TTR does, whereas
the affinity of T3 was not greatly altered, implying that
the N-terminal region exert a greater influence on the
affinity of T4 than on the affinity of T3. This could
imply that the core of TTR has a major influence in
determining the affinity of T3 and the N-terminal
region mainly influences the affinity of T4. To resolve
this more precisely, further chimeric TTRs are required
to be analyzed.
Here, we have demonstrated that the character of
the N-terminal region influences the binding of thyroid
hormones to TTR. Taken together with our previous
report [14], we propose that the N-terminal region has
a role in determining the affinities of T3 and T4 to

TTR.
Experimental procedures
Reagents and chemicals
PCR and plasmid purification kits were from GibcoBRL
(Long Island, NY, USA) and Qiagen (Hilden, Germany).
ABI PRISM dye terminator cycle sequencing ready reaction
kit with AmpliTaq DNA polymerase was from Perkin
Elmer (Wellesley, MA, USA). DNA ligase was purchased
from New England Biolabs (Ipswich, MA, USA). Restric-
tion endonucleases and Taq DNA polymerase were from
Promega (Madison, MI, USA) and Invitrogen (Carlsbad,
CA, USA). Oligonucleotide primers were synthesized by
GibcoBRL and the Bioservice unit, National Science and
Technology Development Agency, Thailand. l-[
125
I]-thyrox-
ine (1.25 CiÆmg
)1
) and l-3,5,3¢-[
125
I]-triiodothyronine
(1.25 CiÆmg
)1
) were purchased from Dupont NEN (Boston,
MA, USA), stored in lead containers and kept in the dark
at 4 °C. Sep-Pak C-18 reversed-phase chromatography car-
tridges were from Waters (Milford, MA, USA). A Bio-Sil
SEC250 column and protein molecular weight markers
were from Bio-Rad (Hercules, CA, USA). All other chemi-
cals used were of analytical grade.

Construction of expression vectors
for recombinant human TTR
Recombinant human TTR was first attempted to be syn-
thesized using the pPIC9 vector, as this vector had been
used for production of all other recombinant (including
truncated and chimeric) TTRs. However, as this gave an
extremely low yield, recombinant human TTR was then
produced using the pPIC3.5 vector. BamH I and EcoRI
sites were introduced by PCR into the human wild-type
TTR cDNA, such that either cleavage by BamH I occurred
immediately before the start codon ATG (for methionine)
in the TTR cDNA presegment, or that cleavage by XhoI
occurred immediately before the codon GGC of the first
amino acid (Gly) of the mature TTR, and that cleavage by
EcoRI occurred immediately after the stop codon TGA of
Function of transthyretin N-terminus P. Prapunpoj et al.
4020 FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS
the cDNA. Primers to generate the compatible restriction
ends for ligation into the pPIC3.5 (using the TTR native
secretion signal) and pPIC9 (using secretion signal of the
a-factor) vectors are presented in Table 1. The PCR prod-
uct with compatible restriction sites of BamHI at 5¢ and
EcoRI at 3¢ ends was ligated to Pichia expression vector
pPIC3.5 so that the newly synthesized TTR was secreted
using its native presegment. The construct with XhoIat5¢
and EcoRI at 3¢ end was ligated into pPIC9, containing the
a-factor peptide secretion signal for the recombinant pro-
tein as described below. The inserted vector was linearized
by digestion with Sal I and used for transformation of
P. pastoris strain GS115 by electroporation.

For both constructs in pPIC3.5 and pPIC9, screening of
recombinant colonies with the methanol utilization positive
(Mut
+
) phenotype and induction synthesis of recombinant
protein in the buffered glycerol-complex medium ⁄ buffered
methanol-complex medium (BMGY ⁄ BMMY) were per-
formed as described in the users’ manual (Invitrogen). The
induction of 50 putative transformants with Mut
+
pheno-
type was carried out in 0.5% methanol at 30 °C for 96 h.
Construction of chimeric TTR cDNAs
The human ⁄ crocTTR cDNA (cDNA that would code for
residue Gly1 to Glu7 of human TTR and residues Ser8 to
Glu127 of crocodile TTR) and truncated crocTTR cDNA
(cDNA that would code for residue Ser8 to Glu127 of cro-
codile TTR) were amplified using C. porosus TTR cDNA
as the template. Pairs of specific primers (Table 1) were
incorporated into the reaction mixture to alter the nucleo-
tide sequence and generate XhoIat5¢ and EcoRI at 3¢ ends
of the TTR template, as previously described [14]. The
constructed cDNAs containing compatible restriction ends
(XhoI and EcoRI ends) were ligated into Pichia expression
vector pPIC9. The sequence Glu-Lys-Arg is necessary for
a-factor peptide release by the KEX2 gene product, and
cleavage by KEX2 occurs between arginine and glutamine
in the sequence Glu-Lys-Arg-Glu-Ala-Glu-Ala. According
to Invitrogen, the sequence of Glu-Ala-Glu-Ala is necessary
for correct cleavage and will be removed during transloca-

tion of the recombinant protein. Thus, the TTR cDNA was
constructed such that the cDNA was immediately in-frame
integrated with the coding portion of Glu-Ala-Glu-Ala.
The vector construct was thereafter linearized with Sal I
and introduced into P. pastoris. Screening of transformants
and induction of synthesis of recombinant proteins were
preformed as described for the recombinant human TTR.
Each transformant showed a similar expression level of
TTR (data not shown). One of each chimeric TTR trans-
formant was used for further experiments.
Purification of recombinant TTR from yeast
culture supernatant
The recombinant TTR was purified from the Pichia culture
either by affinity chromatography using a human retinol-
binding protein-Sepharose-4B as described by Larsson et al.
[29] or by preparative discontinuous nondenaturing-PAGE
using the Bio-Rad Prep Cell (model 491) (10% acrylamide
for the resolving gel and 3% acrylamide for the stacking gel,
and buffering system as recommended by the company).
Determination of the masses of TTR tetramers
by gel filtration
Molecular masses of the recombinant human TTR, and chi-
meric and truncated crocTTRs were estimated by HPLC ⁄ gel-
permeation chromatography using Bio-Sil SEC250 column
(Bio-Rad), equilibrated in 50 mm potassium phosphate
buffer saline, pH 7.4. Aliquots (50 lL) of purified TTR
(2 mgÆmL
)1
) were chromatographed and absorbance was
measured at 280 nm. The column was calibrated with bovine

serum albumin (68 kDa), ovalbumin (45 kDa), chymotrypsi-
nogen (25.5 kDa) and ribonuclease A (13.7 kDa).
Table 1. Oligonucleotides used to generate cDNAs for recombinant human TTR, human ⁄ crocTTR, and truncated crocTTR. PCR was per-
formed with the oligonucleotides listed in a single step to generate cDNAs for recombinant human TTR and truncated crocTTR or in two
steps to generate cDNA for recombinant human ⁄ crocTTR. Nucleotide sequence of human TTR in the primers is underlined, and that of
C. porosus TTR is in bold.
Species of TTR
cDNA encoded Vector
PCR
step Sequence 5¢fi3¢ Sense
Human TTR pPIC3.5 1 AGGATCCAGG
ATGGCTTCTCATCG
AGGAGTGAATTCTCA
TTCCTTGGGATTGG
Sense
Antisense
pPIC9 1 TCTCGAGAAAAGAGAGGCTGAAGCTG
GCCCTACGGGG
AGGAGTGAATTCTCA
TTCCTTGGGATTGG
Sense
Antisense
Human ⁄ crocTTR pPIC9 1 A
ACGGGCACCGGTGAATCCAAATGCC
ACGGAATTCTTATTCTTGTGGATCACTG
Sense
Antisense
2 CTCGAGAAAAGAGAGGCTGAAGCT
GGCCCAACGGGCACCGG
ACGGAATTCTTATTCTTGTGGATCACTG

Sense
Antisense
Truncated crocTTR pPIC9 1 CTCGAGAAAAGATCCAAATGCCCACTTATGG
ACGGAATTCTTATTCTTGTGGATCACTG
Sense
Antisense
P. Prapunpoj et al. Function of transthyretin N-terminus
FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS 4021
Analysis of N-terminal amino acid sequencing
The N-terminal amino acid sequence of recombinant TTRs
were determined using an automatic Edman degradation at
La Trobe University, Australia, and at Bioservice Unit,
National Science and Technology Development Agency,
Thailand.
Purification of radioiodinated thyroid hormones
Commercially available l[
125
I]-T3 and l[
125
I]-T4 were separ-
ated from free
125
I
–
and iodinated hormone degradation
products using a SepPak C-18 cartridge column. Purifica-
tion was checked by thin layer chromatography and ana-
lyzed in an LKB 1270 Rackgamma II counter with a
counting efficiency of 70%, as described previously [12].
Analysis of thyroid hormone binding

to recombinant TTRs
Purified TTR (100 nm) was incubated with T4 or T3
from 0 to 1000 nm, in Irvine buffer in the presence of
tracer amounts of l[
125
I]-T3 or l[
125
I]-T4 at 4 °C, over-
night, as described by Chang et al. [12]. Briefly, a volume
of 0.4 mL of the incubation mixture was transferred to a
vial for total radioactivity determination. Free T4 or T3
in 0.4 mL of the same incubation mixture was separated
from the TTR-bound thyroid hormones within 1 s, by
adsorption to a layer of methyl cellulose coated charcoal
on a glass microfilter under constant vacuum. The filter
was rinsed with 0.4 mL of Irvine buffer, then radioactiv-
ity (corresponding to free thyroid hormone) in the filters
was determined using an LKB 1270 Rackgamma II coun-
ter with a counting efficiency of 70%. For analysis of K
d
values for T3 and T4 of all recombinant TTRs, analyses
for chicken TTR purified from serum were always deter-
mined simultaneously so that the same conformity of the
assay system was confirmed and the K
d
values obtained
could be compared. Nonspecific binding was extrapolated
and other corrections were performed prior analysis by
Scatchard plot [25].
SDS/PAGE

Analysis of proteins under denaturing conditions was per-
formed in SDS ⁄ PAGE slab gels (15% polyacrylamide,
pH 8.6) using a 4% polyacrylamide stacking gel (pH 6.8)
and the discontinuous buffer system of Laemmli and Favre
[30].
Western analysis
Western analysis was performed using a polyclonal anti-
body from a rabbit raised against a mixture of TTRs
purified from human, wallaby and chicken sera as the pri-
mary antibody, as described previously [14]. The membrane
filter was blocked, incubated with rabbit anti(human, wal-
laby and chicken TTR) antiserum (1 : 5000), washed, then
incubated with horseradish-peroxidase-conjugated anti-rab-
bit immunoglobulin (1 : 10 000). Detection was carried out
by enhanced chemiluminescence (Amersham, Pittsburgh,
PA, USA). The filter was exposed to Kodak XAR-5 film
with an intensifying screen at room temperature for 10 min,
and then developed immediately.
Acknowledgements
This research was supported by grants from the Thai-
land Research Fund (TRF), the National Research
Council of Thailand and Prince of Songkla University
(The Excellent Biochemistry Program Fund).
References
1 Kabat EA, Landow H & Moore DH (1942) Electro-
phoretic patterns of concentrated cerebrospinal fluid.
Proc Soc Exp Biol Medical 49, 260–263.
2 Kabat EA, Moore DH & Landow H (1942) An electro-
phoretic study of the protein components in cerebro-
spinal fluid and their relationship to the serum proteins.

J Clin Invest 21, 571–577.
3 Richardson SJ, Monk JA, Shepherdley CA, Ebbesson
LO, Sin F, Power DM, Frappell PB & Ko
¨
hrle & Ren-
free MB (2005) Developmentally regulated thyroid hor-
mone distributor proteins in marsupials, reptile and fish.
Am J Physiol 288, R1264–R1272.
4 Blake CCF, Geisow MJ, Oatley SJ, Re
´
rat B & Re
´
rat C
(1978) Structure of prealbumin: secondary, tertiary and
quaternary interactions determined by Fourier refine-
ment at 1.8 A
˚
. J Mol Biol 121, 339–356.
5 Schreiber G, Prapunpoj P, Chang L, Richardson SJ,
Aldred AR & Munro SLA (1998) Evolution of thyroid
hormone distribution. Clin Expt Pharmacol Physiol 25,
728–732.
6 Schreiber G (2002) The evolutionary and integrative
roles of transthyretin in thyroid hormone homeostasis.
J Endocrinol 175, 61–73.
7 Pages RA, Robbins J & Edelhoch H (1973) Binding of
thyroxine and thyroxine analogs to human serum pre-
albumin. Biochemistry 12, 2773–2779.
8 Nilsson SF, Rask L & Peterson PA (1975) Studies on
thyroid hormone-binding proteins. J Biol Chem 250,

8554–8563.
9 Wojtczak A, Cody V, Luft JR & Paugborn W (1996)
Structure of human transthyretin complexed with thyr-
oxine at 2.0 A
˚
resolution and 3¢,5¢-dinitro-N-acetyl-l-
thyronine at 2.2 A
˚
resolution. Acta Crystallogr D52,
758–765.
Function of transthyretin N-terminus P. Prapunpoj et al.
4022 FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS
10 Yamauchi K, Kasahara T, Hayashi H & Horiuchi R
(1993) Purification and characterization of a 3,5,3 ¢-l-tri-
iodothyronine-specific binding protein from bullfrog
tadpole plasma: a homolog of mammalian transthyretin.
Endocrinology 132, 2254–2261.
11 Santos CRA & Power DM (1999) Identification of
transthyretin in fish (Sparus aurata): cDNA cloning and
characterisation. Endocrinology 140, 2430–2433.
12 Chang L, Munro SLA, Richardson SJ & Schreiber G
(1999) Evolution of thyroid hormone binding by trans-
thyretins in birds and mammals. Eur J Biochem 259,
534–542.
13 Prapunpoj P, Yamauchi K, Nishiyama N, Richardson
SJ & Schreiber G (2000) Evolution of structure, onto-
geny of gene expression, and function of Xenopus laevis
transthyretin. Am J Physiol 279, R2026–R2041.
14 Prapunpoj P, Richardson SJ & Schreiber G (2002) Cro-
codile transthyretin: structure, function, and evolution.

Am J Physiol 283, R885–R896.
15 Richardson SJ (2002) The evolution of transthyretin
synthesis in vertebrate liver, in primitive eukaryotes and
in bacteria. Clin Chem Laboratory Med 40, 1191–1199.
16 Prapunpoj P, Richardson SJ, Fumagalli L & Schreiber
G (2000) The evolution of the thyroid hormone distri-
butor protein transthyretin in the order Insectivora,
class Mammalia. Mol Biol Evol 17, 1199–1209.
17 Blake CC, Geisow MJ, Swan ID, Re
´
rat C & Re
´
rat B
(1974) Structure of human plasma prealbumin at 2.5 A
˚
resolution. A preliminary report on the polypeptide
chain conformation, quaternary structure and thyroxine
binding. J Mol Biol 88, 1–12.
18 Hamilton JA, Steinrauf LK, Braden BC, Liepnieks J,
Benson MD, Holmgren G, Sandgren O & Steen L
(1993) The X-ray crystal structure refinements of normal
human transthyretin and the amyloidogenic Val-30?Met
variant to 1.7-A
˚
resolution. J Biol Chem 268, 2416–
2424.
19 Hamilton JA & Benson MD (2001) Transthyretin: a
review from a structural perspective. Cell Mol Life Sci
58, 1491–1521.
20 Cereghino JL & Cregg JM (2000) Heterologous protein

expression in the methylotrophic yeast Pichia pastoris.
FEMS Microbiol Rev 24, 45–66.
21 Richardson SJ, Bradley AJ, Duan W, Wettenhall REH,
Harms PJ, Babon JJ, Southwell BR, Nicol S, Donnellan
SC & Schreiber G (1994) Evolution of marsupial and
other vertebrate thyroxine-binding plasma proteins. Am
J Physiol 266, R1359–R1360.
22 Seibert FB & Nelson JW (1942) Electrophoretic study
of the blood protein response in tuberculosis. J Biol
Chem 143, 29–38.
23 Farer LS, Robbins J, Blumberg BS & Rall JE (1962)
Thyroxine serum protein complexes in various animals.
Endocrinology 70, 686–696.
24 Refetoff S, Robin NI & Fang VS (1970) Parameters of
thyroid function in serum of 16 selected vertebrate spe-
cies: a study of PBI, serum T
4
, free T
4
, and the pattern
of T
4
and T
3
binding to serum proteins. Endocrinology
86, 793–805.
25 Scatchard G (1949) The attractions of proteins for small
molecules and ions. Ann N Y Acad Sci 51, 660–672.
26 Richardson SJ (2005) Expression of transthyretin in the
choroid plexus: relationship to brain homeostatsis of

thyroid hormones. In The Blood–Cerebrospinal Fluid
Barrier (Zheng, W & Chodobski, A, eds), pp. 275–304.
CRC Press, Boca Raton.
27 Cheng SY, Wilchek M, Cahnmmann HJ & Robbins J
(1977) Affinity labeling of human serum prealbumin
with N-bromoacetyl-l-thyroxine. J Biol Chem 252,
6076–6081.
28 Fitch NJS, Akbari MT & Ramsden DB (1991) An
inherited non-amyloidogenic transthyretin variant,
[Ser6]-TTR, with increased thyroxine-binding affinity,
characterized by DNA sequencing. J Endocrinol 129,
309–313.
29 Larsson M, Pettersson T & Carlstro
¨
m A (1985) Thyroid
hormone binding in serum of 15 vertebrate species: iso-
lation of thyroxine-binding globulin and pre-albumin
analog. Gen Comp Endocrinol 58, 360–375.
30 Laemmli UK &Favre M (1973) Maturation of the head
of bacteriophage T4. J Mol Biol 80 , 575–599.
31 Aldred AR, Prapunpoj P & Schreiber G (1997) Evolu-
tion of shorter and more hydrophilic transthyretin
N-termini by stepwise conversion of exon 2 into intron
1 sequences (shifting the 3¢ splice site of intron 1). Eur J
Biochem 246, 401–409.
P. Prapunpoj et al. Function of transthyretin N-terminus
FEBS Journal 273 (2006) 4013–4023 ª 2006 The Authors Journal compilation ª 2006 FEBS 4023