Full text of DOI:10.1007/BF02900592

proteins in eukaryotes. Knol et al. used RNAMOT pattern
search program to screen GenBankTM, sequence-tagged
site, and EST data bases for SECIS element and found
four new selenoproteins[4]. Gladyshav et al. designed a
SECIS Search program to identify SECIS element in nu-
cleotide sequences, leading to the discovery of two new
selenoproteins[5]. With the success of human genome
mapping and the enrichment of biological information[6],
it becomes possible to predict or discover new selenopro-
teins using some computer programs.
In the present work the RNA Draw program was
used to compare 46 genes from 12 selenoproteins and
some genes with UGA code or non-UGA termination in
the proteins. The results demonstrated that the identifica-
tion of selenoproteins by SECIS element was possible
with RNA Draw. Thus, it is valuable to use the present
program to search the potential new selenoprotein in gene
bank due to the completeness of human genome plan.
Identification of selenocyst-
eine insertion sequence
(SECIS) element in eukaryotic
selenoproteins by RNA Draw
program
XU Huibi1, HUANG Kaixun1, QU Xianghu1,
GAO Zhonghong1, LIU Qiong1, CHEN Runsheng2
& XUAN Zhenyu2
1. Department of Chemistry, Huazhong University of Science and Tech-
nology, Wuhan 430074, China;
2. Department of Protein Engineering, Institute of Biophysics, Chinese
Academy of Sciences, Beijing 100101, China
Abstract
The computer program RNA Draw was used to
1
Materials and methods
identify the secondary structures in the 3Ąuntranslated re-
gions (3ĄUTRs) of the mRNAs from 46 eukaryotic seleno-
proteins among 7 species. The program found one or two
possible SECIS elements in these selenoproteins. The SECIS
element consists of a stem-loop or hairpin structure with
three conserved sequences of AUGA—(A)AA—GA. SECIS
element was not found by the RNA Draw program in ran-
domly selected non-selenoproteins. The results showed that
SECIS element is the unique character of the genes of eu-
karyotic selenoproteins. Thus it is possible to use RNA Draw
to search the SECIS elements in gene bank for potential new
selenoproteins.
(ν) Genes and their sources. The genes of 46 eu-
karyotic selenoproteins and 26 non-selenoproteins of 9
species were from the original references and GenBank.
(ξ) Methods. The program of RNA Draw V1.1 b2
(Ole Matzura, 1995)[9] was used to predict the secondary
structures of the above genes in the 3Ąuntranslated re-
gions. Energy calculation was conducted at 37ć using
the energy model from Zuker[10]. All the secondary struc-
tures folded following the rule of free energy minimiza-
tion. The SECIS elements in the genes were searched
based on the SECIS structures from human or rat 5Ą-Dĉꢀ
and cGPx. The computer processing procedures were only
carried out in the 3Ą-UTR due to the fact that SECIS in
eukaryotic selenoproteins has not yet been reported out-
side 3Ą-UTR. The input of the sequences of 3Ą-UTR to the
computer was performed once for the number of nucleo-
tides smaller than 1000 and twice for the nucleotide num-
ber bigger than 1000. The two-step input was conducted
with the folding bases of 200 across each other.
Keywords: selenoprotein, eukaryotes, 3Ąuntranslated region (3Ą
UTR), selenocysteine insertion sequence (SECIS), RNA Draw pro-
gram.
The essential trace element selenium exerts its bio-
logical function mainly through various selenoenzymes
and selenoproteins[1]. Selenium in these enzymes and pro-
teins is in the form of selenocysteine (Sec), which is the
active center of selenoenzymes[2]. So far, it has been clear
that translating UGA code to Sec residue in eukaryotes
requires a unique stem-loop structure in the 3ĄUTR of the
mRNA. This structure is called the Sec insertion sequence
(SECIS) element. The function of SECIS guiding the in-
sertion of Sec into proteins was shown in the gene expres-
sion of cellular glutathion peroxidase (cGPX) and typeĉ
2
Results and discussion
(ν) SECIS elements from the genes of GPx. RNA
folding analysis was performed in the 3ĄUTRs of 5 cGPx
genes, 4 phGPx genes, 4 pGPx genes, 2 giGPx genes and
the GPx gene from Schistosoma mansoni. A stem-loop
structure similar to the reported rat 5Ą-DI SECIS and hu-
man cGPx SECIS was found in all cGPx genes, human
pGPx gene, mouse giGPx gene and the GPx gene from
Schistosoma mansoni[3,7,8] (fig. 1). One possible SECIS
stem-loop structure was found in all phGPx genes, 3 other
pGPx genes and human giGPx gene. This possible SECIS
structure has similarity to cGPx SECIS, which contains
the conserved base sequences of AUGA, AAA and GA,
and the similar conserved positions of AUGA and GA in
iodothyronine 5Ądeiodinase (5Ą-DI)[3].
Fourteen selenoenzymes and selenoproteins from
seven species have been characterized in eukaryotes. They
are four types of glutathione peroxidases (GPX), three
types of iodothyronine 5Ądeiodinases (types ĉ, Ċ and
ċ; expressed as 5Ą-Dĉ, 5Ą-DĊ and 5Ą-Dċ), three types
of thioredoxin reductases (TR), selenoprotein P, seleno-
protein W, selenophosphate synthetase 2 (SPS2) and the
15 ku selenoprotein. Increasing evidences demonstrate
that there are possibilities of some undiscovered seleno-
Chinese Science Bulletin Vol. 46 No. 1 January 2001
1159

NOTES
SECIS. The contrast between phGPx and cGPx is that 3 or
2 continuous A is not situated in the top loop but in the
on the catalytic activity of the enzyme, we found two pos-
sible SECIS stem-loop structures in the 3ĄUTRs of rat
and human SPS2 genes using the RNA Draw program.
The first one belongs to type ĉ SECIS and the second
below bulge at 5Ąarm of phGPx. The structure character of
cGPx belongs to typeĉ SECIS and that of phGPx be-
longs to type Ċ SECIS (see fig. 1).
one belongs to type Ċ SECIS. Both AAA top loops con-
sist of 14 nucleotides. Both distances between AUGA and
AAA contain 13 nucleotides.
Guimaraes et al. reported in 1997 that the 3ĄUTR in
SPS2 gene could enable the insertion of 75Se into SPS2
protein and increase the expression of SPS2 by more than
20 times. This implicates the existence of SECIS in 3Ą
UTR.
(ρ) SECIS elements in other selenoproteins. Se-
lenoprotein P, selenoprotein W and the 15 ku selenopro-
tein were discovered in recent years. Similar results were
obtained in the 3ĄUTR of the above genes through the
RNA folding analysis.
(ς) SECIS secondary structure model for the genes
of eukaryotic selenoproteins. In conclusion, a stable
secondary structure was formed after the process of RNA
folding program, although the nucleotide sequences were
significantly different in the SECIS elements of different
eukaryotic selenoprotein genes. Two types of SECIS ele-
ments were concluded according to the positions of the
conserved nucleotides in the stem-loop structures (see fig.
1). Type ĉ SECIS element has the top loop consisting of
9 to 16 nucleotides with three continuous A inside the
loop. AUGA and GA locate in the upstream of the top
loop at 5Ąarm and 3Ąarm respectively, forming non-Wat-
son-Crick duplex. The distance between AAA and AUGA
is generally from 9 to 12 nucleotides for different SECIS
Fig. 1. Types ĉ and Ċ SECIS models for eukaryotic selenoprotein
genes.
(ξꢁ SECIS elements from deiodinase genes. Sear-
ching the secondary structures of the 3ĄUTRs from 15
deiodinases demonstrated a stem-loop structure in all 5ĄD
ĉ genes and bullfrog 5ĄDċ genes. This stem-loop
structure is similar to the reported SECIS element of rat 5Ą
Dĉ[3], which belongs to type ĉ SECIS (fig. 1). Never-
theless, bullfrog 5ĄDċ SECIS is different from typical
type ĉ SECIS, the unpaired AUGA in the upstream of 5Ą
arm is replaced by the paired AUG. A type Ċ SECIS
elements. Type Ċ SECIS element has a top loop and an
upstream bulge loop. The conserved AAA or AA does not
situated in the top loop but in the bulge loop at 5Ąarm. The
(A)AA bulge loop is connected with the top loop through
3 to 4 base pairs.
stem-loop structure was found in the 3ĄUTR of 5ĄDċ
genes from human beings, rat, chicken and African toad.
(ο) SECIS elements from TR genes. The RNA
Draw program was used to predict the secondary struc-
(σ) The genes from non-selenoproteins. The pro-
gram of RNA Draw was used to predict the secondary
tures of 3ĄUTR in 3 sequenced TR genes (human, rat and
structure of the 3ĄUTRs in the genes of seven non-se-
lenoproteins containing UGA as a terminating code, the
gene of esGPx with UGA and UAA as terminating codes
and the gene of hydrogen peroxidase. The seven
non-selenoprotein genes were ADH gene, AGT gene, du-
plex RNA adenosine deaminase gene, PGIS gene, PGHS
gene, human TXAS gene and mouse PGI2 gene. No typi-
cal SECIS stem-loop structure was found in the 3ĄUTRs
of the above genes. There were even no AUGA sequence
cow TR genes). A type Ċ SECIS stem-loop structure
was found, which has a top loop containing 7ü8 nucleo-
tides above the conserved CAA bulge loop. The two loops
were connected by three base pairs. The distance between
CAA and the conserved AUGA was 12 nucleotides.
(π) SECIS elements from SPS2 genes. Selenoen-
zyme SPS2 was recently isolated from mouse embryo
cells or human activated CD8+T cells. The above two
genes of SPS2 code for two selenoproteins with 452
amino acids and 448 amino acids respectively. Both se-
lenoproteins contain a UGA code for Sec (located respec-
tively at position 63 and position 60) and use the second
UGA as a terminating code. Although it was reported that
the Sec in SPS2 from influenza hemophilus had no effect
in the 3ĄUTRs of rat and mouse ADH genes, human
TXAS gene, rat PGIS gene, human PGHS gene, etc.
These results further showed the important roles of SECIS
element in the genes of selenoproteins.
1160
Chinese Science Bulletin Vol. 46 No. 1 January 2001

2. Stadtman, T. C., Selenocysteine, Ann. Rev. Biochem., 1996, 65:
83.
3. Berry, M. J., Bann, L., Chen, Y. et al., Recognition of UGA as a
3
Conclusions
SECIS element containing the specific conserved
selenocysteine codon in type ĉ deiodinase requires sequences in
sequences is indeed unique for the genes of selenoproteins.
SECIS could be used as a probe to predict and search the
gene of a new selenoprotein in the complicated genome.
One example is the discovery of the first selenoprotein in
virusüüthe human dermatotropic poxvirus (MCV) se-
lenoprotein, which was found by Berry in 1998 using the
the 3Ąuntranslated region, Nature, 1991, 353: 273.
4. Lescure, A., Gautheret, D., Carbon, P. et al., Novel selenocys-
teines identified in silico and in vivo by using a conserved RNA
structural motif, J. Biol. Chem., 1999, 274: 38147.
5. Kryukov, G. V., Kryukov, V. M., Gladyshev, V. N., New mammal-
ian selenocysteine-containing proteins identified with an algo-
rithm that searches for selenocysteine insertion sequence elements,
J. Biol. Chem., 1999, 274: 33888.
above method[12]
.
The character of SECIS element could also be used
to search for the new selenoprotein in some already se-
quenced eukaryotic protein genes. The good example is
thioredoxin reductase, which had been characterized and
sequenced early but was only identified as a selenoen-
zyme in 1996[11]. Thus it is reasonable to use SECIS as a
probe to screen more selenoprotein genes in gene bank.
The program of RNA Draw is widely used by mo-
lecular biologists to analyze the folding structure of RNA.
The present work demonstrated that this program could
also be used to identify the two types of SECIS elements,
which provided the possibilities for using the program to
search for a new selenoprotein in gene bank.
6. Chen Runsheng, Bioinformation, Acta Biophysics (in Chinese),
1999(1): 5.
7. Qu Xianghu, Huang Kaixun, Gao Zhonghong et al., The seleno-
cysteine insertion sequence element of gultathione peroxidase,
Chinese Journal of Biochemistry Molecular Biology (in Chinese),
1999, 15(2): 274.
8. Qu Xianghu, Huang Kaixun, Xu Huibi, The selenocysteine inser-
tion sequence element of deiodinase, J. Wuhan University (Natu-
ral Science) (in Chinese), 1999, 6: 1.
9. Schuster, P., Fantana, W., Stadler, P. F. et al., From sequence to
shape and back: a case study in RNA secondary structures, Proc.
R. Soc. London, Ser B, 1994, 255: 279.
10. Zuker, M., On finding all suboptimal folding of an RNA molecule,
Science, 1989, 244(4900): 48.
11. Gladyshev, V. N., Jeang, K. T., Stadtman, T. C., Selenocysteine,
identified as the penultimate C-terminal residue in human T-cell
thioredoxin reductase, corresponds to TGA in the human placental
gene, Proc. Natl. Acad. Sci. USA, 1996, 93: 6146.
References
1. Xu Huibi, Huang Kaixun, The Chemistry and Biochemistry of
Selenium and Their Application in Life Science, Wuhan: The
Printing House of Huazhong University of Science and Technol-
12. Shisler, J. L., Senkevich, T. G., Berry, M. J. et al., Ultravio-
let-induced cell death blocked by a selenoprotein from a human
dermatotropic poxvirus, Science, 1998, 279: 102.
(Received December 27, 2000)
ogy, 1994, 104ü105.
Chinese Science Bulletin Vol. 46 No. 1 January 2001
1161