Badh2, Encoding Betaine Aldehyde Dehydrogenase, Inhibitsthe Biosynthesis of 2-Acetyl-1-Pyrroline, a Major Componentin Rice Fragrance W
Saihua Chen,a Yi Yang,b Weiwei Shi,b Qing Ji,a Fei He,c Ziding Zhang,c Zhukuan Cheng,d
Xiangnong Liu,b and Mingliang Xua,e,1
a National Maize Improvement Center of China, China Agricultural University, Beijing 100193, People’s Republic of Chinab College of Bioscience and Biotechnology, Yangzhou University, Yangzhou, Jiangsu 225009, People’s Republic of Chinac College of Biological Sciences, China Agricultural University, Beijing 100193, People’s Republic of Chinad Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, People’s Republic of Chinae Beijing Key Laboratory of Crop Genetic Improvement, Beijing 100193, People’s Republic of China
In rice (Oryza sativa), the presence of a dominant Badh2 allele encoding betaine aldehyde dehydrogenase (BADH2) inhibits
the synthesis of 2-acetyl-1-pyrroline (2AP), a potent flavor component in rice fragrance. By contrast, its two recessive
alleles, badh2-E2 and badh2-E7, induce 2AP formation. Badh2 was found to be transcribed in all tissues tested except for
roots, and the transcript was detected at higher abundance in young, healthy leaves than in other tissues. Multiple Badh2
transcript lengths were detected, and the complete, full-length Badh2 transcript was much less abundant than partial
Badh2 transcripts. 2AP levels were significantly reduced in cauliflower mosaic virus 35S-driven transgenic lines expressing
the complete, but not the partial, Badh2 coding sequences. In accordance, the intact, full-length BADH2 protein (503
residues) appeared exclusively in nonfragrant transgenic lines and rice varieties. These results indicate that the full-length
BADH2 protein encoded by Badh2 renders rice nonfragrant by inhibiting 2AP biosynthesis. The BADH2 enzyme was
predicted to contain three domains: NAD binding, substrate binding, and oligomerization domains. BADH2 was distributed
throughout the cytoplasm, where it is predicted to catalyze the oxidization of betaine aldehyde, 4-aminobutyraldehyde (AB-
ald), and 3-aminopropionaldehyde. The presence of null badh2 alleles resulted in AB-ald accumulation and enhanced 2AP
biosynthesis. In summary, these data support the hypothesis that BADH2 inhibits 2AP biosynthesis by exhausting AB-ald, a
presumed 2AP precursor.
INTRODUCTION
Fragrant rice (Oryza sativa) is gaining widespread popularity
among consumers worldwide (Bhattacharjee et al., 2002); thus,
its market price is much higher than that of nonfragrant rice (Qiu
and Zhang, 2003). A mixture of 114 different volatile compounds
was detected in the flavor of cooked rice (Yajima et al., 1978).
One of them, 2-acetyl-1-pyrroline (2AP), is a potent flavor com-
ponent with a lower odor threshold that gives both basmati and
jasmine rice their distinctive fragrance (Buttery et al., 1982). 2AP
is found in all parts of plants of fragrant rice varieties except for
the roots (Buttery et al., 1983b). The 2AP level is relatively higher
in the aerial parts of plants than in milled rice grains (Yoshihashi
et al., 1999). In contrast with aromatic double haploid lines, no
2AP is detected in nonscented double haploid lines (Lorieux
et al., 1996). In nature, 2AP is also detected in panda leaves
(Pandunus amaryllifolius) (Buttery et al., 1983a) and is formed
both in baking wheat breads (Schieberle and Grosch, 1991) and
in cocoa fermentation (Romanczyk et al., 1995).
Genetic analysis shows that a single recessive gene (fgr) on
chromosome 8 is associated with rice fragrance and that the
dominant Fgr allele is associated with lack of fragrance (Sood
and Siddiq, 1978; Huang et al., 1994; Jin et al., 2003). A number
of markers were identified that are closely linked to fgr (Ahn et al.,
1992; Causse et al., 1994; Chen et al., 1997; Cho et al., 1998; Jin
et al., 2003). Two restriction fragment length polymorphism
markers, RG1 and RG28, were identified that flank fgr, with the
estimates of genetic distances ranging from 10 centimorgan (cM)
(Causse et al., 1994) to 12 cM (Lorieux et al., 1996) to 25.5 cM
(Cho et al., 1998). After genetic mapping and sequence analysis
of 17 genes in the fgr region, Bradbury et al. (2005) suggested
that a gene encoding putative betaine aldehyde dehydrogenase
(BADH2) is most likely to be the fgr gene, due to its sequence
divergence between fragrant and nonfragrant rice varieties.
Furthermore, the badh2 alleles from fragrant rice varieties all
have common insertions/deletions and single nucleotide poly-
morphisms compared with those from nonfragrant genotypes,
demonstrating a common ancestor for all fragrant genotypes
(Bradbury et al., 2005). In our previous mapping, fgr was local-
ized to a 69-kb region bordered by the markers L02 and L06
(Chen et al., 2006). In addition to Badh2, two other genes were
1 Address correspondence to [email protected] author responsible for distribution of materials integral to thefindings presented in this article in accordance with the policy describedin the Instructions for Authors (www.plantcell.org) is: Mingliang Xu([email protected]).W Online version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.108.058917
The Plant Cell, Vol. 20: 1850–1861, July 2008, www.plantcell.org ª 2008 American Society of Plant Biologists
located in this fgr region: Cah and Mccc2, encoding eukaryotic-
type carbonic anhydrase and 3-methylcrotonyl-CoA carboxyl-
ase b-chain, respectively (Chen et al., 2006).
In cocoa fermentation, 2AP is synthesized from either L-Pro or
L-Orn by Bacillus cereus (Romanczyk et al., 1995). In fragrant
rice, L-Pro is a precursor to 2AP, providing the nitrogen of the
pyrroline ring but not the carbon of the acetyl group (Yoshihashi
et al., 2002). 2AP formation positively correlates with an accu-
mulation of Pro, resulting in the strong aroma of the Thailand
fragrant rice variety Khao Dawk Mali 105, which is grown in the
drought region Tung Kula Rong Hai (Yoshihashi et al., 1999).
Interestingly, Pro has been shown to be involved in osmoregu-
lation and to accumulate in rice plants under diverse types of
stress, including drought (Yang and Kao, 1999). It has been
suggested that 2AP is synthesized via the polyamine pathway,
in which Orn, via 1-pyrroline, is the major source of the nitrogen
in 2AP (A. Vanavichit, T. Yoshihashi, S. Wanchana, S. Areekit,
D. Saengsraku, W. Kamolsukyunyong, J. Lanceras, T. Toojinda,
and S. Tragoonrung, unpublished data). To date, however, the
biosynthetic pathway of 2AP has not been demonstrated clearly.
Therefore, the identification of the fragrant-related gene(s) would
provide valuable insight regarding the mechanism of 2AP bio-
synthesis.
RESULTS
Identification and Cloning of Fgr Candidates
We previously mapped fgr to a 69-kb region flanked by the
markers L02 and L06 (Chen et al., 2006). Here, fgr was further
localized to an interval between L04 and L06 (Figure 1). Se-
quence alignment between the Fgr-containing nonfragrant rice
varieties indica cv 93-11 and japonica cv Nipponbare revealed
considerable variation within this L04/L06 interval, which had a
physical distance of 62 kb in 93-11 but only 35 kb in Nipponbare.
In Nipponbare, a 30-kb DNA segment rich in transposition-
related genes was inserted inside an interval of L02/L03. In 93-
11, a large DNA insert was present between L04 and L05. Three
putative genes in the L04/L06 interval, shared by both 93-11 and
Nipponbare, were considered to be candidates for the Fgr gene
in these nonfragrant varieties. These genes are Cah, Mccc2, and
Badh2, encoding putative eukaryotic-type carbonic anhydrase,
3-methylcrotonyl-CoA carboxylase b-chain, and betaine alde-
hyde dehydrogenase, respectively (Figure 1).
After screening the indica Nanjing11 and japonica Suyunuo
BAC libraries, we obtained two and four positive BAC clones in
the Fgr (Nanjing11; nonfragrant) and fgr (Suyunuo; fragrant)
regions, respectively. Two BAC clones from each library were
sufficient to cover the regions of interest and are shown in Figure
1. Three Fgr candidates and three fgr candidates from these BAC
clones (corresponding to Cah, Mccc2, and Badh2 in each case)
were subcloned into the expression vector pCAMBIA1302
(pCAM). Coding regions for Cah, Mccc2, and Badh2 were
determined to be 2499, 6506, and 5846 bp, respectively. Maps
of the resulting six clones are show in Supplemental Figure 1
online. The inserts of these six Fgr/fgr constructs were se-
quenced and compared with their counterparts in the indica cv
93-11 (http://rise.genomics.org.cn) and the japonica cv Nippon-
bare (http://www.gramene.org). Sequence alterations specific in
the fragrant variety Suyunuo were observed only in the badh2
gene, including an 8-bp deletion (59-GATTATGG-39) and three
single nucleotide polymorphisms in exon 7 (see Supplemental
Figure 2A and Supplemental Data Set 1 online). Furthermore, we
sequenced another 23 fragrant rice varieties at their badh2 loci.
Apart from the above badh2 allele (designated badh2-E7), a
new null badh2 allele (named badh2-E2) with a 7-bp deletion
(59-CGGGCGC-39) in exon 2 was identified (Shi et al., 2008). In
addition, other nucleotide sequence divergences were also
Figure 1. Positive BAC Clones, Predicted Genes, Developed Markers Used for Mapping, and Genomic Sequences in the Fgr Regions for the indica cv
93-11 and the japonica cv Nipponbare.
Similar genomic regions are shaded in gray, while major insertions are shown in white (regions A and B). Dark arrows show three putative genes, Cah,
Mccc2, and Badh2, within the L04/L06 region that are common to both the indica and japonica rice subspecies and, therefore, were considered to be
candidates for the Fgr gene. BAC clones used for subcloning are shown above and below the genomic region.
The Badh2 Gene Inhibits 2AP Synthesis 1851
observed within introns in two fragrant rice varieties, cv
Wuxiangjing9 and Suyunuo, including a TT deletion in intron 2
and a repeated (AT)n insert in intron 4.
An alignment of amino acid sequences of Badh2, badh2-E2,
and badh2-E7 alleles, together with a highly homologous rice
Badh1 gene as reference, is provided in Supplemental Figure 2B
online. The intact BADH2 shows 75.6% identity with the BADH1
amino acid sequence. The truncated BADH2 from badh2-E7
consisted of a presumed peptide with 251 residues. Another
truncated BADH2 from badh2-E2 consisted of a shorter peptide
with 82 residues.
Confirmation of Badh2 Function in Rice Fragrance
Because rice fragrance is a recessive trait, three Fgr constructs
(pCAM-Badh2, pCAM-Cah, and pCAM-Mccc2) were trans-
formed into the fragrant rice recipient Wuxiangjing (japonica,
badh2-E7) and transgenic lines were identified (see Supplemen-
tal Figure 3 online). Details and photographs showing the steps
of the Agrobacterium tumefaciens–mediated rice transformation
procedure are available from the authors upon request. All
30 plantlets regenerated from the calli transformed with pCAM-
Badh2 were confirmed to be positive. Only two and three
transgenic plantlets were obtained from the calli transformed
with pCAM-Cah and pCAM-Mccc2, respectively (see Supple-
mental Figures 3A and 3B online).
The 2AP contents were determined for all regenerated plant-
lets by gas chromatography–mass spectrometry (GC-MS) (see
Supplemental Figure 4 online). Transgenic lines carrying either
the exogenous Cah or Mccc2 gene were fragrant, and their 2AP
levels were as high as those of nontransgenic lines. By contrast,
all Badh2 transgenic lines were nonfragrant with no or very low
2AP content (Table 1; see Supplemental Table 1 online). The
Badh2 transgenic lines were further self-pollinated to generate
T1 progeny, which were analyzed by PCR using Badh2 gene-
specific primers (see Supplemental Figure 5 online). Segregation
of the exogenous Badh2 gene in T1 progeny was consistent with
the 3:1 ratio expected for monogenic inheritance. All T1 progeny
with the exogenous Badh2 genes were nonfragrant with reduced
2AP levels, while individuals without the exogenous Badh2 gene
were fragrant.
The construct pCAM-Badh2 was further transformed into
another four fragrant rice varieties (Zhenxiangjing5, Guangling-
xiangjing, Wuxiangjing9, and Xiangjing111) that harbor the other
null badh2-E2 alleles. As expected, transgenic lines showed
significantly reduced 2AP levels compared with nontransgenic
lines (Table 2; see Supplemental Table 2 online). The data from
the above analyses indicate that the Badh2 gene can compen-
sate for the defective functions of both badh2-E2 and badh2-E7
to render transgenic lines nonfragrant.
Transcription Profiling of the Badh2/badh2-E7 Alleles
In the nonfragrant rice cv Nanjing11, Badh2 transcripts were
detected in all tissues tested except for the roots. The Badh2
transcript was more abundant in the initial fully expanded leaf
than in other tissues (Figure 2). Interestingly, a similar transcrip-
tion pattern with a lower transcription level was observed for the
badh2-E7 allele in the fragrant rice cv Wuxiangjing (Figure 2). This
indicates that sequence alterations specific to the badh2-E7
allele significantly reduce its transcription levels but cause only a
slight change in its transcription pattern.
Multiple Transcription Start Sites Are Present in the
Badh2 Gene
To identify the transcription start point, rapid amplification of
cDNA ends (RACE) analysis was performed on mRNA from
leaves of nonfragrant rice cv Nanjing11. As expected, the
39-RACE product was 1013 bp, the size predicted from the Badh2
Figure 2. Examination of the Badh2/badh2 Transcription Levels in
Various Plant Tissues Using Real-Time RT-PCR.
Initial fully expanded leaf (A), last fully expanded leaf (B), stem (C), young
panicle at the booting stage (D), panicle at the heading stage (E), panicle
at 15 d after the heading stage (F), and roots (G) are shown. The relative
transcription level for each plant tissue was represented as a 2�DCT
value. Error bars represent the SD of transcription levels determined from
the three independent real-time PCRs.
Table 1. 2AP Levels and Their Significant Differences among Three
Kinds of Transgenic Lines and Nontransgenic Lines
Nontransgenic and
Transgenic Lines
No. of
Plantlets
2AP Levels
(ng/g; means
6 SD)
Significancea
At 5% At 1%
Nontransgenic lines 11 21.33 6 7.37 a A
Cah transgenic lines 2 18.94 6 1.32 a A
Mccc2 transgenic lines 4 20.52 6 3.75 a A
Badh2 transgenic lines 11 1.99 6 1.03 b B
a 2AP levels with different letters are significantly different at P < 0.05
(lowercase letter) or at P < 0.01 (uppercase letter).
1852 The Plant Cell
gene sequence (http://www.gramene.org). However, to our sur-
prise, the majority of the 59-RACE products were only ;500 bp,
much shorter than the 800 bp expected from the Badh2 gene
(Figure 3). Sequence analysis of the 59-RACE products showed
that the major transcription start point moved downstream of
the predicted Badh2 start codon, resulting in partial Badh2
transcripts. Moreover, the Badh2 coding sequence (CDS) ap-
peared to have variable transcription start points, since different
leader sequences were observed that corresponded to exon 2,
intron 2, or exon 3. A search for a possible Badh2 cDNA sequence
in the database (www.ncbi.nlm.nih.gov) identified a Badh2 cDNA
clone (National Center for Biotechnology Information accession
number AK060461) with an even shorter coding sequence (1095-
bp CDS) compared with the partial Badh2 cDNA (1182-bp CDS)
identified in this study.
To confirm the presence of multiple Badh2 transcripts, we
designed three forward primers (FP1, FP2, and FP3) and one
common reverse primer (RP) based on the Badh2 predicted
sequence (see Supplemental Figure 6 and Supplemental Table 4
online). Using real-time RT-PCR, the primer pair FP1/RP gener-
ated very small amounts of PCR products, while the other two
primer pairs, FP2/RP and FP3/RP, generated much larger
amounts (Figure 4A). Likewise, no PCR product was observed
in RT-PCR with the primer pair FP1/RP, although strong PCR
bands were amplified using the other two primer pairs, FP2/RP
and FP3/RP (Figure 4B).
The relative abundance of different Badh2 transcripts was also
estimated by RNA gel blot hybridization. Two probes were
designed based on the predicted Badh2 cDNA sequence (see
Supplemental Figure 6 online). Probe A covers the region from
�74 to 244 bp, and this probe hybridizes specifically to the
complete Badh2 transcript. Probe B corresponds to the Badh2
cDNA segment from 367 to 1325 bp, which hybridizes to all
possible Badh2 transcripts. When hybridized to probe A, almost
no signal was detected for all RNA samples (Figure 5). When
hybridized to probe B, however, strong signals were observed in
two nonfragrant rice varieties (Nangjing11 and Nipponbare) but
not in two fragrant rice varieties (Wuxiangjing and Wuxiangjing9)
(Figure 5). The major signal band was estimated to be ;1.5 kb,
corresponding to the size of the partial Badh2 transcript (;100-
bp 59 untranslated region [UTR], 1182-bp CDS, and 210-bp 39
UTR) detected in the RACE reaction.
The data from the above RACE, real-time RT-PCR, RT-PCR,
and RNA gel blot analysis indicated the following: (1) the major
transcription start point was downstream of the predicted Badh2
start codon, resulting in very few complete Badh2 transcripts but
abundant partial Badh2 transcripts; and (2) transcription of the
badh2 gene was severely suppressed in fragrant rice varieties.
The Intact BADH2 Protein Reduces the 2AP Levels
As multiple Badh2 transcripts were present in nonfragrant rice
varieties, we were anxious to know which one in fact is associ-
ated with rice fragrance. Therefore, we overexpressed both the
complete and partial Badh2 coding sequences in the fragrant
rice recipient Wuxiangjing to examine the effect on fragrance.
When driven by the CaMV35S promoter, overexpression of the
predicted complete Badh2 CDS (pCAM35S:GL) rendered trans-
genic lines nonfragrant and caused a reduction in 2AP levels
(Table 3; see Supplemental Table 3 online). CaMV35S-driven
overexpression of the partial Badh2 CDS (pCAM35S:GM,
pCAM35S:GS, pCAM35S:CM, and pCAM35S:CS) had no influ-
ence on rice fragrance (Table 3; see Supplemental Table 3 online).
Additional evidence that BADH2 inhibits the production of
rice fragrance was obtained from our protein gel blot anal-
ysis. The intact 55-kD BADH2 protein with 503 amino acids
was observed in the nonfragrant rice varieties (Nanjing11 and
3037) as well as in nonfragrant transgenic lines harboring
either the native Badh2 gene or the CaMV35S-driven complete
Badh2 CDS (Figure 6). However, no BADH2 protein appeared
Table 2. 2AP Levels in Nontransgenic and Transgenic Plantlets Using Four Fragrant Rice Varieties (badh2-E2) as Recipients
Fragrant Rice
Recipients (badh2-E2)
Transgenic Lines (Badh2þbadh2-E2/badh2-E2) Nontransgenic Lines (badh2-E2/badh2-E2)
No. of
Plantlets
2AP Levels
(ng/g; means 6 SD)
No. of
Plantlets
2AP Levels
(ng/g; means 6 SD)
Xiangjing111 3 5.75 6 0.46 2 23.85 6 3.47
Guanglingxiangjing 2 8.21 6 1.83 2 27.81 6 2.53
Wuxiangjing9 3 9.54 6 1.15 –a –
Zhenxiangjing5 3 7.92 6 2.99 – –
a No data available.
Figure 3. The 59- and 39-RACE Products of the Badh2 Coding Se-
quence.
Badh2 gene-specific primers were designed based on the predicted
Badh2 coding sequence (see Supplemental Table 4 online). mRNA was
isolated from leaf tissues of the nonfragrant rice cv Nanjing11 and used
for RACE reactions. The 59-RACE (A) and 39-RACE (B) products were
analyzed by electrophoresis on 1% agarose gels. M, DL2000 molecular
marker (Tiangen Biotech); 59, 59-RACE product; 39, 39-RACE product.
The Badh2 Gene Inhibits 2AP Synthesis 1853
in the fragrant rice varieties (Wuxiangjing9 and Suyunuo), non-
transgenic lines, or fragrant transgenic lines with the CaMV35S-
driven partial Badh2 CDS (Figure 6). We concluded that the
intact 503–amino acid BADH2 protein encoded by the complete
Badh2 gene inhibits 2AP synthesis and thus renders rice non-
fragrant.
The Predicted Three-Dimensional Structure of BADH2
Screening of the protein database allowed us to identify a human
mitochondrial aldehyde dehydrogenase (ALDH2; Protein Data
Bank code 1o04; x-ray resolution, 1.42 A) as the structural
template for BADH2 (Perez-Miller and Hurley, 2003). The tem-
plate ALDH2 shared 42% sequence identity with BADH2 (see
Supplemental Figure 7 online). The predicted three-dimensional
model of BADH2 could be divided into three domains: a NAD
binding domain (residues 9 to 124 and 152 to 262), an oligomer-
ization domain (residues 129 to 151 and 480 to 486), and a
substrate binding domain (residues 263 to 464) (Figure 7A).
Based on the active sites of ALDH2 (Johansson et al., 1998),
eight residues conserved between ALDH2 and BADH2 were
annotated as the potential active sites in BADH2. Two residues,
Asn-162 and Cys-294, acting in a catalytic role, are responsible
for interacting with the substrate oxygen. Six residues, Tyr-163,
Leu-166, Trp-170, Glu-260, Cys-453, and Trp-459, are predicted
to form the substrate binding pocket (Figure 7B).
Figure 4. Relative Abundance of different Badh2 Transcripts.
(A) Real-time RT-PCR. Tissues A, B, and C represented leaf tissues of Nanjing11 (A), immature seeds 14 d after the flowering of Nanjing11 (B), and leaf
tissues of Wuxiangjing (C). Total RNA from these three tissues was extracted and used for real-time PCR. Error bars represent the SD of the transcript
levels determined from the three independent real-time PCRs.
(B) RT-PCR. 59-RACE products from the leaf tissue of Wuxiangjing were subjected to RT-PCR. Lanes 1, 2, and 3 corresponded to use of the primer pairs
FP1/RP, FP2/RP, and FP3/RP, respectively.
Figure 5. Relative Abundance of Different Badh2 Transcripts Estimated by RNA Gel Blot Hybridization.
(A) Probe A (lanes A) and probe B (lanes B) were synthesized and digoxygenin-labeled with an asymmetry PCR labeling system. Probes A and B were
analyzed by electrophoresis on a 1% agarose gel to check their quantities and qualities.
(B) Total RNA extracted from four rice varieties was analyzed by electrophoresis on a 1.2% formaldehyde agarose gel.
(C) Images of RNA gel blot hybridization using probes A and B. Very weak signals were present in two fragrant rice varieties (lanes 1 and 3). M, RNA
marker.
Lanes 1 to 4 in (B) and (C) corresponded to the rice cv Wuxiangjing (japonica, badh2-E7), Nanjing11 (indica, Badh2), Wuxiangjing9 (japonica, badh2-E2),
and Nipponbare (japonica, Badh2), respectively.
1854 The Plant Cell
The badh2-E7 allele, due to its 8-bp deletion in exon 7,
encodes a presumably truncated BADH2 that lacks 252
C-terminal residues. The missing C-terminal amino acids cover
the entire substrate binding domain and partial oligomerization
domains, thus rendering the truncated BADH2 nonfunctional.
Likewise, the badh2-E2 allele encodes an even shorter truncated
BADH2 (82 residues) that is not predicted to have any function as
an aldehyde dehydrogenase.
Subcellular Localization of BADH2
Because BADH2 is highly expressed during cell division, young
panicles of rice plants were used to determine its subcellular
localization. Immunodetection using an anti-BADH2 antibody
and a fluorescein-conjugated secondary antibody showed
strong fluorescent signals for BADH2 in the inflorescence mer-
istem during cell division. A significant amount of BADH2 signal
was seen in the cytoplasm. By contrast, none of the BADH2
signal was observed to be localized in the nucleus (Figure 8).
In Vitro Expression of the Badh2/badh2 Alleles
We predicted that in vitro expression of the Badh2/badh2-E7
cDNAs would result in the intact 503–amino acid polypeptide
(encoded by the complete Badh2 cDNA), the partial 393–amino
acid polypeptide (encoded by the partial Badh2 cDNA), the
truncated 251–amino acid polypeptide (encoded by the com-
plete badh2-E7 cDNA), and the truncated 141–amino acid poly-
peptide (encoded by the partial badh2-E7 cDNA) BADH2
proteins, respectively. Since each construct expressed a fusion
protein in which the tagged portion (Nus, His, and S tags) was
;61 kD, the fusion proteins were estimated to be 116 kD
(PET43.1a-NFCcDNA), 104 kD (PET43.1a-NFPcDNA), 88 kD
(PET43.1a-FCcDNA), and 76 kD (PET43.1a-FPcDNA), respectively.
We observed the expected fusion proteins to be produced in
Escherichia coli in an isopropylthio-b-D-galactoside–dependent
manner (see Supplemental Figure 8 online), confirming that these
cDNAs could produce the appropriate proteins.
BADH2 Exhibits Aldehyde Dehydrogenase Activity
The predicted three-dimensional structure of BADH2 suggested
that it belongs to the NAD-dependent aldehyde dehydrogenase
family, characterized by the typical Aldedh substrate binding
domain (Steinmetz et al., 1997; http://pfam.sanger.ac.uk/
search/sequence). The BADH enzymes in sugar beet (Beta
vulgaris), spinach (Spinacia oleracea), and oat (Avena sativa) all
showed broad substrate specificities, catalyzing the oxidation
not only of betaine aldehyde (Bet-ald) but also of other
v-aminoaldehydes (Trossat et al., 1997; Incharoensakdi et al.,
2000; Livingstone et al., 2003). To elucidate the role of BADH2 in
2AP biosynthesis, we characterized its aldehyde dehydrogenase
activity as well as its substrate specificity.
The intact BADH2 (503 residues) and partial BADH2 (393
residues) were expressed in E. coli and purified using nickel-
nitrilotriacetic acid agarose (see Supplemental Figure 9 online).
The intact BADH2 showed high Bet-ald dehydrogenase activity,
with a rapid increase in A340 within the first 10 min after the
enzymatic reaction (Figure 9). In addition, the intact BADH2 also
showed strong 4-aminobutyraldehyde (AB-ald) and 3-amino-
propionaldehyde (AP-ald) dehydrogenase activities (Figure 9).
By contrast, the partial BADH2 polypeptide showed only back-
ground activity for all three aldehyde substrates (Figure 9). These
data indicate that the intact BADH2 of rice, like BADH of sugar
beet and spinach, showed high aldehyde dehydrogenase activ-
ity and wide substrate specificities.
Table 3. 2AP Levels and Their Significant Differences among
Nontransgenic Lines and Three Kinds of Transgenic Lines Carrying
Different Badh2 CDS Driven by the CaMV35S Promoter
Exogenous Badh2
CDS
No. of
Plantlets
2AP Levels
(ng/g; means
6 SD)
Significancea
At 5% At 1%
Nontransgenic lines 9 33.07 6 18.35 a A
Partial Badh2 CDS
(1095 bp) 8 27.12 6 8.86 a A
Partial Badh2 CDS
(1182 bp) 7 31.36 6 6.05 a A
Complete Badh2 CDS
(1512 bp) 11 9.12 6 3.20 b B
a 2AP levels with different letters are significantly different at P < 0.05
(lowercase letter) or at P < 0.01 (uppercase letter).
Figure 6. Detection of BADH2 by Protein Gel Blot Hybridization.
Total proteins were extracted from leaf tissues and separated on a 12% SDS-PAGE gel, followed by electroblotting onto a polyvinylidene difluoride
membrane. After being blocked with 1% BSA, the membrane was hybridized first with the polyclonal BADH2 antibody and then with donkey anti-rabbit
antibody. BADH2 bands were observed in the positive control and lanes 2, 4 to 6, and 11 to 13 but not in lanes 1, 3, and 7 to 10. Symbols above the lanes
are as follows: CK, control; F, fragrant variety; N, nonfragrant variety; C, complete Badh2 CDS; P, partial Badh2 CDS; B, native Badh2 gene. The control
(left lane) constitutes the intact BADH2 protein, obtained by digestion of the Nus-BADH2 fusion protein with enterokinase. Lanes 1 to 5, rice varieties
Suyunuo (lane 1; fragrant, badh2-E7), 3037 (lane 2; nonfragrant, Badh2), Xiangjing111 (lane 3; fragrant, badh2-E2), Nanjing11 (lane 4; nonfragrant,
Badh2), and Nipponbare (lane 5; nonfragrant, Badh2); lanes 6 to 10, Transgenic lines containing the overexpression constructs pCAM35S:GL (lane 6;
the CaMV35S-driven complete genomic Badh2 CDS), pCAM35S:GM (lane 7; the CaMV35S-driven partial genomic Badh2 CDS), pCAM35S:GS (lane 8;
the CaMV35S-driven partial genomic Badh2 CDS), pCAM35S:CM (lane 9; the CaMV35S-driven partial Badh2 cDNA), and pCAM35S:CS (lane 10; the
CaMV35S-driven partial Badh2 cDNA); lanes 11 to 13, three transgenic lines containing the construct pCAM-Badh2 (the native Badh2 gene with an
;1.3-kb promoter, an ;5.8-kb coding region, and an ;5.6-kb 39 UTR).
The Badh2 Gene Inhibits 2AP Synthesis 1855
DISCUSSION
We have shown that the badh2 locus of rice constitutes the fgr
gene that determines fragrance. Of three Fgr candidates in the rice
genomic region to which Fgr (corresponding to the lack of fra-
grance) was mapped previously, only Badh2 significantly reduced
2AP content and thus rendered fragrant rice nonfragrant. A non-
functional allele, either badh2-E7 or badh2-E2, is required for rice
fragrance. The presence of the dominant (Fgr) or recessive (fgr)
allele, determining the nonfragrant versus the fragrant trait, re-
spectively, corresponds completely with the presence of the
functional Badh2 or null badh2 allele, respectively.
Compared with the functional Badh2 allele, both nonfunctional
badh2-E2 and badh2-E7 alleles had very low transcription levels,
as revealed by real-time RT-PCR and RNA gel blot analysis. This
indicates that loss-of-function mutations in Badh2 severely sup-
press mRNA transcription. Theoretically, deletions in the Badh2
gene sequence, for example, the observed 7-bp deletion in exon
2 and an 8-bp deletion in exon 7, would cause frame shifts and
result in truncated BADH2 proteins. This was confirmed by in
vitro expression in which the Badh2 cDNA resulted in an intact
503–amino acid BADH and the badh2-E7 cDNA resulted in a
truncated 251–amino acid BADH (see Supplemental Figure 8
online). However, no such truncated BADH2 proteins were
detected in vivo in the fragrant rice cv Wuxiangjing9 and Suyunuo
by protein gel blot hybridization (Figure 6). This suggests that the
major reason that the badh2 alleles are nonfunctional is the
suppression of both transcription and translation, rather than
being a result of the truncation of the BADH2 protein.
We conducted a RACE analysis to determine the transcription
start point and found that the 59-RACE products for both the
Badh2 and badh2 alleles were much shorter than expected.
Subsequent real-time RT-PCR, RT-PCR, and RNA gel blot
analyses all showed that the complete Badh2 transcript was
much less abundant than partial Badh2 transcripts. However,
protein gel blots confirmed that the BADH2 protein had a mass
Figure 7. The Three-Dimensional Structure of BADH2 and the Annotated Active Sites.
(A) Representation of BADH2. Domain coloring is as follows: the oligomerization domain is shown in red, the substrate binding domain is shown in
orange, and the NAD binding domain is shown in green. To position the NAD cofactor in the predicted model, the NAD cofactor from the crystal
structure of ALDH2 is superimposed into the predicted BADH2 model and is indicated by gray spheres.
(B) The annotated active sites of BADH2. Asn-162 and Cys-294 are catalytic residues that are predicted to interact with the substrate oxygen. Tyr-163,
Leu-166, Trp-170, Glu-260, Cys-453, and Trp-459 are predicted to form the substrate binding pocket. The side chain atoms in these highlighted
residues are represented as sticks and are colored as follows: green, carbon; blue, nitrogen; red, oxygen; orange, sulfur.
Images in both panels were prepared using PyMol (http://pymol.sourceforge.net).
Figure 8. Indirect Subcellular Immunodetection of BADH2.
(A) Green signal represents BADH2 detected with fluorescein-conjugated goat anti-rabbit antibody.
(B) Blue signal indicates nuclei stained with 4,6-diamidino-phenylindole (DAPI).
(C) The merged image of BADH2 and DAPI-stained nuclei.
1856 The Plant Cell
of ;55 kD and, therefore, must be encoded by the longest cDNA
we identified. Such intact 503–amino acid BADH2 was observed
only in nonfragrant rice varieties and transgenic lines expressing
the longest Badh2 transcript (Figure 6). Consistent with protein
gel blot analysis, only the intact 503–amino acid BADH2 showed
strong aldehyde dehydrogenase activity and broad substrate
specificities. In other cereal plants, Badh2 genes also have been
reported to encode full-length BADH proteins (wheat [Triticum
aestivum] BAD2, AAL05264; barley [Hordeum vulgare] BAD2,
BAB62846).
To test whether the abundant partial Badh2 transcripts serve
any role in the expression of the intact BADH2 protein, we
transformed fragrant rice with the different Badh2 cDNAs and
their genomic DNA segments under control of the CaMV35S
promoter. The intact 503–amino acid BADH2 protein was pro-
duced by the transgenic lines carrying the CaMV35S-driven
complete Badh2 coding sequence. No short 393–amino acid
BADH2 was detected in the transgenic lines carrying the
CaMV35S-driven partial Badh2 coding sequence. Compared
with the native Badh2 gene, the overexpression of the complete
Badh2 gene resulted in low levels of BADH2 protein (Figure 6).
Analysis of these data indicated that (1) full-length Badh2 tran-
script driven by CaMV35S, although abundant in transgenic
lines, did not result in more BADH2 protein, and (2) although the
partial Badh2 transcript itself cannot be translated into protein,
the presence of abundant partial Badh2 transcripts might lead to
high-efficiency translation of the complete Badh2 transcript.
Because the absence of BADH2 protein results in fragrance,
this suggests that Badh2 is not directly involved in 2AP biosyn-
thesis. Alternative possibilities to explain the effect of BADH2 are
that the BADH2 enzyme is involved in a competing pathway in
which one of the 2AP precursors serves as a BADH2 substrate
(Bradbury et al., 2005) or that BADH2 participates in 2AP catab-
olism. Trossat et al. (1997) showed that the sugar beet BADH
catalyzes the oxidization not only of Bet-ald but also of other
substrates structurally similar to Bet-ald, such as 3-dimethylsul-
foniopropionaldehyde, AP-ald, and AB-ald. AB-ald is known to
be maintained in an equimolar ratio with D-1-pyrroline, an im-
mediate 2AP precursor. On the other hand, AB-ald can be
converted into 4-aminobutyric acid (GABA) (A. Vanavichit,
T. Yoshihashi, S. Wanchana, S. Areekit, D. Saengsraku, W.
Kamolsukyunyong, J. Lanceras, T. Toojinda, and S. Tragoonrung,
unpublished data). It was reported that GABA was found in lower
amounts in leaves of the aromatic isogenic line than the nonar-
omatic counterpart (A. Vanavichit, T. Yoshihashi, S. Wanchana,
S. Areekit, D. Saengsraku, W. Kamolsukyunyong, J. Lanceras,
T. Toojinda, and S. Tragoonrung, unpublished data). Therefore,
AB-ald levels appear to be an important factor regulating the rate
of 2AP biosynthesis. Consumption of AB-ald by converting it into
GABA inhibits 2AP synthesis, whereas the accumulation of AB-
ald results in increased 2AP synthesis.
In this study, the intact BADH2 protein encoded by the
complete Badh2 gene sequence was shown to influence this
critical switch, possibly due to its strong AB-ald dehydrogenase
activity. From its activity in vitro, we conclude that, in nonfragrant
rice, the BADH2 enzyme converts AB-ald into GABA, inhibiting
2AP biosynthesis. In fragrant rice lacking intact BADH2, failure to
convert AB-ald into GABA due to the absence of BADH2 enzy-
matic activity results in AB-ald accumulation, which activates
2AP biosynthesis. Interestingly, the low level of BADH2 detected
in some transgenic lines overexpressing Badh2 might not com-
pletely inhibit the consumption of AB-ald, resulting in a small
quantity of 2AP (see Supplemental Table 3 online). Apart from
BADH2, another homologous protein, BADH1, is encoded by the
rice genome. Although we have not yet investigated the enzy-
matic activity of BADH1, we predict that BADH1 will show a low
affinity for AB-ald but high affinities for other aldehyde sub-
strates. It is unlikely that BADH1 has the same aldehyde dehy-
drogenase activity and substrate specificities as BADH2 in rice,
since BADH2 clearly confers the nonfragrant trait and loss of 2AP
accumulation.
In higher plants, betaine is well known as a nontoxic or
protective cytoplasmic osmolyte, allowing normal growth of
plants in a saline or arid environment. Betaine is synthesized
via a two-step oxidation of choline, and the second step (from
Bet-ald to betaine) is catalyzed by BADH. In barley, two BADH
isozymes (BBD1 and BBD2) were reported, and both of them are
induced to higher levels by salt, drought, and abscisic acid
treatments (Nakamura et al., 2001). The rice BADH2 protein in
this study showed high betaine aldehyde dehydrogenase activ-
ity, suggesting that Badh2 also may play a key role in osmoreg-
ulation in nonfragrant rice. However, the absence of BADH2 in
fragrant rice did not negatively affect normal growth. For exam-
ple, the Thailand fragrant rice variety Khao Dawk Mali 105
Figure 9. Determination of Aldehyde Dehydrogenase Activity of Purified Intact BADH2 and Partial BADH2 Using Various Aldehyde Substrates.
The enzymatic activities were spectrophotometrically assayed by A340 at pH 8.0 at intervals of 0, 10, 20, 30, and 60 min after the initiation of reactions.
Data are means 6 SD from three independent experiments with 1 mM Bet-ald (A), 50 mM AB-ald (B), and 50 mM AP-ald (C).
The Badh2 Gene Inhibits 2AP Synthesis 1857
(badh2-E7/badh2-E7) grows well in the arid region Tung Kula
Rong Hai (Yoshihashi et al., 1999; Bradbury et al., 2005). This
suggests that other Badh genes in the rice genome can com-
pensate for the defective null badh2 alleles and allow for toler-
ance to salinity and drought stresses. Apart from the known
Badh1 gene on chromosome 4, we found at least two other
genes encoding putative BADHs on chromosome 7. The rice
Badh1 gene on chromosome 4 showed high homology with the
Badh1 genes in barley and sorghum (Sorghum bicolor) (Bradbury
et al., 2005). More research is required to elucidate the functions
of these Badh genes in both tolerance to salinity/drought
stresses and the regulation of 2AP biosynthesis.
In summary, the intact 503–amino acid BADH2 encoded by the
complete Badh2 gene inhibits 2AP biosynthesis by converting
AB-ald (a presumed 2AP precursor) to GABA, whereas the
absence of BADH2 due to nonfunctional badh2 alleles results
in AB-ald accumulation and thus turns on the pathway toward
2AP biosynthesis.
METHODS
Construction of Rice BAC Libraries and Identification of Positive
BAC Clones
Both a local Chinese fragrant rice cv, Suyunuo (Oryza sativa japonica), and
a nonfragrant rice cv, Nanjing11 (Oryza sativa indica), were used to
construct BAC libraries. The leaf tissues were collected from etiolated
shoots after a 2-week culture at 258C. The construction and screening of
the BAC library were performed as described by Xu et al. (2001) with some
modifications. Two restriction enzymes, HindIII and BamHI, were used to
partially digest rice genomic DNA. The markers in the fgr region, including
L02, L03, L04, L05, and L06 (see Supplemental Table 4 online), were used
to screen both the Suyunuo and Nanjing11 BAC libraries.
Cloning of the Fgr/fgr Candidates from the Positive BAC Clones
The complete genomic sequences of the Fgr region from both Nippon-
bare and 93-11 were obtained from sequences deposited in the data-
bases (http://rise.genomics.org.cn and http://www.gramene.org). The
restriction sites flanking the three candidate genes (Cah, Mccc2, and
Badh2) were identified and used to subclone the intact Fgr/fgr candidates
from the corresponding BAC clones. For our studies, an intact candidate
gene must contain at least the 1.3-kb promoter, the full coding region, and
the 39 UTR region. For each of the three genes, a DNA fragment with the
intact candidate gene was then subcloned into the expression vector
pCAM (http://www.cambia.org.au/daisy/cambia/home.htm).
Rice Transformation and Complementation Tests
The constructs containing Fgr candidates were introduced into Agrobacte-
rium tumefaciens strain EHA105 (Hood et al., 1993) and then used to infect
the fragrant Wuxiangjing callus. Plantlets regenerated from hygromycin-
resistant calli were transplanted into large culture pots and grown in a culture
room under a 14-h photoperiod with a room temperature of 26 6 28C. Leaf
tissue of each independently regenerated plantlet was harvested to examine
both genotype and fragrance. The marker tagged to each Fgr candidate was
used to identify positive transgenic lines (primers listed in Supplemental
Table 4 online). The fragrance trait for each regenerated plantlet was first
investigated using the smelling method (Chen et al., 2006). Thereafter, an
aliquot of 0.3 g fresh leaf was subjected to 2AP extraction at 658C for 1 h
in a solution consisting of ethanol and 0.459 ng/mL 2,4,6-trimethylpyridine
as an internal standard (Bergman et al., 2000). After centrifugation at
12,000 rpm (relative centrifugal field of 13,400g) for 30 min, the super-
natant was directly used to measure 2AP concentrations with the GC-
MS method (Itani et al., 2004). The smelling method is unreliable due to
its inherent subjectivity; while the GC-MS method can directly estimate
2AP level; therefore, it is more accurate than the smelling method in
evaluating rice fragrance.
Real-Time RT-PCR
Real-time RT-PCR was conducted on both the nonfragrant rice cv
Nanjing11 and the fragrant rice cv Wuxiangjing. Rice tissues, including
the initial fully expanded leaf, the last fully expanded leaf, the stem, young
panicles at the booting stage, panicles at the heading stage, panicles 15 d
after the heading stage, and roots, were all collected for total RNA
extraction. After digestion with RNase-free DNase I, an aliquot of 500 ng
of total RNA was used for one-step real-time RT-PCR using a QuantiTech
SYBR Green RT-PCR kit (Qiagen). Each RNA sample was assayed in
triplicate. For a negative control, total RNA without reverse transcription
was directly used for real-time PCR. With a reference Actin gene, the
relative amount of the Badh2 transcript is presented as 2�DCT according
to the DCT method described in the Real-Time PCR Applications Guide
(Bio-Rad Laboratories; http://www.bio-rad.com). A fluorescence thresh-
old was set up at a value of <0.05 to ensure the amplification in the
logarithmic phase. The DCT value was obtained by subtracting the CT
(threshold cycles) number of the Actin gene from that of the Badh2 gene
(DCT ¼ CTBadh2 � CTActin). The DCT value was converted to the linear
form in terms of 2�DCT for statistical analysis. Analysis of variance was
performed using the SAS system (SAS Institute), and the expression level
of each sample was represented as mean 2�DCT 6 SD.
RACE Reactions and Cloning of cDNAs
Leaf tissues were collected from both the nonfragrant rice cv Nanjing11
and the fragrant rice cv Wuxiangjing. Total RNA was isolated from leaf
tissue with the LiCl method (Sambrook and Russell, 2001), followed by
mRNA isolation using the PolyATract System III kit (Promega). The syn-
thesis of first-strand cDNA (RACE-ready cDNA) was catalyzed by Power-
Script reverse transcriptase (BD SMART RACE cDNA amplification kit; BD
Biosciences). The Badh2-specific primers were designed for the Fgr/fgr
alleles based on the predicted coding sequences and were used for the
amplification of 39- and 59-RACE–ready cDNAs (see Supplemental Table 4
online). The amplicons (RACE products) were then cloned into the vector
pGEM-T (Promega) for sequencing. The sequences of 39- and 59-RACE
products were merged into full-length cDNAs based on the condition that
both RACE products shared the identical overlapping region.
RNA Gel Blot Analysis
Leaf tissues were collected from four rice varieties, Wuxiangjing (japon-
ica, badh2-E7), Nanjing11 (indica, Badh2), Wuxiangjing9 (japonica,
badh2-E2), and Nipponbare (japonica, Badh2), and used for total RNA
extraction with the LiCl method (Sambrook and Russell, 2001). For RNA
gel blot analysis, all RNA samples (;20 mg each) were separated on 1.2%
formaldehyde agarose gels and then blotted onto a nylon membrane
(Amersham Pharmacia Biotech.). The probes were synthesized and
digoxygenin-labeled with an asymmetric PCR labeling system (Zhang
et al., 2005). RNA gel blot hybridization was repeated twice to ensure the
reproducibility of hybridization signals.
Functional Analysis of Badh2 Transcripts
We obtained cDNAs corresponding to the complete Badh2 transcript
(1512-bp CDS) and two partial Badh2 transcripts (1182-bp CDS and 1095-
bp CDS) via RT-PCR. In parallel, we amplified genomic DNA segments
1858 The Plant Cell
corresponding to these three Badh2 transcripts from the pCAM-Badh2
construct. All Badh2 fragments were cloned into the expression vec-
tor pCAM under the control of the CaMV35S promoter. In total, three
genic Badh2 overexpression constructs (pCAM35S:GL, pCAM35S:GM,
and pCAM35S:GS) and two Badh2 cDNA overexpression constructs
(pCAM35S:CM and pCAM35S:CS) were successfully made. Here, L, M,
and S represent the long (1512 bp, complete CDS), medium (1182 bp,
partial CDS), and short (1095 bp, partial CDS) Badh2 transcripts, respec-
tively. These five constructs were separately transformed into the fragrant
rice recipient Wuxiangjing via Agrobacterium. The positive transgenic
plantlets were subjected to 2AP assay with the GC-MS method.
Protein Gel Blot Analysis
Leaf tissues were harvested for total protein extraction using the trichloro-
acetic acid–acetone precipitation method (Damerval et al., 1986). The
quantity of each sample was estimated using a NanoDrop ND-1000
spectrophotometer (NanoDrop Technologies). All protein samples (;30
mg each) were separated by 12% SDS-PAGE and then electroblotted onto
a polyvinylidene difluoride membrane (Amersham Hybone-P). Hybridiza-
tion was performed as described by Burnette (1981) with some modifica-
tions. Polyclonal BADH2 antibody was purified from the antiserum from
rabbits immunized using a BADH2 peptide (antigen, YLAESLDKRQNAPV).
Donkey anti-rabbit antibody was used as the secondary antibody
(ProteinTech Group). The signal appeared after incubating in nitroblue
tetrazolium/5-bromo-4-chloro-3-indolyl phosphate mixture (Amresco)
owing to the alkaline phosphatase in the second antibody.
Subcellular Immunodetection of BADH2
The young panicles (;3 cm in length) were harvested and fixed for 1 h
with 4% (w/v) paraformaldehyde in 20 mL of 13 PBS buffer (137 mM
NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 2 mM KH2PO4, pH 8.0) followed
by three washes in 13 PBS (5 min each). The spikelet was squashed in 13
PBS solution on a microscope slide covered with a cover slip. After
soaking in liquid nitrogen and removing the cover slip, the slide was
dehydrated through an ethanol series (70, 90, and 100%) prior to being
used in immunostaining. Slides were then incubated in a humid chamber
at 378C for 3 h with BADH2 antibody diluted 1:500 in TNB buffer (0.1 M
Tris-HCl, pH 7.5, 0.15 M NaCl, and 0.5% blocking reagent). After three
rounds of washing in PBS, fluorescein-conjugated goat anti-rabbit anti-
body was added to the slides. The slides were counterstained with 4,6-
diamidino-phenylindole in an antifade solution (Vector Laboratories). The
signals were observed using a Leica TCS SP5 confocal system. Captured
images were enhanced and pseudocolored by Adobe Photoshop CS2
software.
Prediction of Three-Dimensional Structure for BADH2
We submitted the BADH2 sequence to the Swiss-Model homology
modeling server (http://swissmodel.expasy.org/SWISS-MODEL.html) to
predict its three-dimensional model using the automatic modeling mode
(Arnold et al., 2006).
In Vitro Expression and Assay of Enzymatic Activity
The complete (C) and partial (P) Badh2 cDNAs from the nonfragrant (NF)
cv Nanjing11 and the fragrant (F) cv Wuxiangjing were amplified via RT-
PCR. The Badh2 cDNAs without any mismatch nucleotides were cloned
into the expression vector PET43.1a (Novagen), resulting in four con-
structs: PET43.1a-NFCcDNA, PET43.1a-FCcDNA, PET43.1a-NFPcDNA,
and PET43.1a-FPcDNA. These four constructs were separately trans-
formed into the Escherichia coli expression strain BL21(DE3), and the
fusion protein containing the Nus flag was expressed by induction with
0.5 mM isopropylthio-b-D-galactoside at 188C overnight (see Supple-
mental Figure 8 online). The induced E. coli cells were collected and lysed
by sonication on ice. The fusion proteins present in the supernatant were
subjected to electrophoresis on 12% SDS-polyacrylamide gels.
The fusion proteins from PET43.1a-NFCcDNA and PET43.1a-
NFPcDNA were purified using nickel-nitrilotriacetic acid agarose affinity
chromatography columns (Qiagen) and assayed for their aldehyde de-
hydrogenase activities. The Bet-ald chloride was purchased from Sigma-
Aldrich. The diethylacetals of AB-ald and AP-ald were obtained from
Sigma-Aldrich and TCI American, respectively. The preparation of three
aldehyde substrates (Bet-ald, AP-ald, and AB-ald) and standard enzy-
matic reactions were conducted as described by Trossat et al. (1997). The
enzymatic activity of BADH2 was measured at 0, 10, 20, 30, and 60 min
following the initiation of reactions by monitoring the increase in A340 with
a spectrophotometer. Three replicates for each reaction were performed.
Accession Numbers
Sequence data from this article can be found in the GenBank/EMBL
databases under the following accession numbers: Badh2 (EU770319),
badh2-E7 (EU770320), badh2-E2 (EU770321), Badh2mRNA (EU770322),
badh2-E7mRNA (EU770323), badh2-E2mRNA (EU770324), 1095-bp
Badh2 mRNA (AK060461), barley BAD1 (BAB62847), rice BAD1
(BAA21098), wheat BAD2 (AAL05264), barley BAD2 (BAB62846), and
rice Actin mRNA (AB047313).
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure 1. Maps of Three Fgr Constructs Derived from
the Nonfragrant cv Nanjing11 and Three fgr Constructs Derived from
the Fragrant cv Suyuno.
Supplemental Figure 2. Characterization of Nucleotide Acid Se-
quences and Amino Acid Sequences of the Badh1, Badh2, badh2-E2,
and badh2-E7 Alleles.
Supplemental Figure 3. Identification of Transgenic Lines Using the
Gene-Tagged Markers.
Supplemental Figure 4. Chromatogram of Rice Leaf Tissue Extracts.
Supplemental Figure 5. Genotypes at the Exogenous Badh2 Gene in
T1 Progeny.
Supplemental Figure 6. Maps of the Three Badh2 Transcripts and
Locations of the Primers and Probes Used for Investigation of Badh2
Transcription.
Supplemental Figure 7. Sequence Alignment between BADH2 and
ALDH2.
Supplemental Figure 8. In Vitro Expression of the Badh2/badh2
cDNAs in E. coli Cells.
Supplemental Figure 9. Purification of the Fusion Proteins Contain-
ing the Intact or Partial BADH2.
Supplemental Table 1. 2AP Levels in Each Nontransgenic Line and
Transgenic Line with Different Exogenous Fgr Candidates.
Supplemental Table 2. 2AP Levels in Transgenic and Nontransgenic
Lines Using the Four Fragrant Rice Recipients Containing the badh2-
E2 Alleles.
Supplemental Table 3. 2AP Levels in Each Nontransgenic Line and
Transgenic Line with the Different Badh2 CDS Driven by the
CaMV35S Promoter.
The Badh2 Gene Inhibits 2AP Synthesis 1859
Supplemental Table 4. List of All Primers Used in This Research.
Supplemental Data Set 1. Sequence Alignment among the Badh2
cDNA and Its Genomic Genes from Various Rice Varieties.
ACKNOWLEDGMENTS
We thank Jiming Jiang and Jason Walling for their critical review of the
manuscript. We also are grateful to Jufei Lu for his assistance in the field
testing, HongMei Wang and Qing Cao for their guidance in tissue
culture, Xin Li and ShuZhu Tang for their help in collection of rice
varieties, and Lu Jiang, ChunXia Liu, and Yan Zhang for their technical
assistance. This research was financially supported by the National
High-Tech 863 Program of the People’s Republic of China (Grant
2002AA224041) and the National Natural Science Foundation of China
(Grant 30771318).
Received February 20, 2008; revised June 4, 2008; accepted June 17,
2008; published July 3, 2008.
REFERENCES
Ahn, S.N., Bollich, C.N., and Tanksley, S.D. (1992). RFLP tagging of a
gene for aroma in rice. Theor. Appl. Genet. 84: 825–828.
Arnold, K., Bordoli, L., Kopp, J., and Schwede, T. (2006). The SWISS-
MODEL workspace: A web-based environment for protein structure
homology modelling. Bioinformatics 22: 195–201.
Bergman, C.J., Delgado, J.T., Bryant, R., Grimm, C., Cadwallader,
K.R., and Webb, B.D. (2000). Rapid gas chromatographic technique
for quantifying 2-acetyl-1-pyrroline and hexanal in rice. Cereal Chem.
77: 454–458.
Bhattacharjee, P., Singhal, R.S., and Kulkarni, P.K. (2002). Basmati
rice: A review. Int. J. Food Sci. Technol. 37: 1–12.
Bradbury, L.M.T., Fitzgerald, T.L., Henry, R.J., Jin, Q., and Waters,
D.L.E. (2005). The gene for fragrance in rice. Plant Biotechnol. J. 3:
363–370.
Burnette, W.N. (1981). ‘Western blotting’: Electrophoretic transfer of
proteins from sodium dodecyl sulfate-polyacrylamide gels to unmod-
ified nitrocellulose and radiographic detection with antibody and
radioiodinated protein A. Anal. Biochem. 112: 195–203.
Buttery, R.G., Juliano, B.O., and Ling, L.C. (1983a). Identification of
rice aroma compound 2-acetyl-1-pyrroline in Panda leaves. Chem.
Ind. (London) 23: 478.
Buttery, R.G., Ling, L.C., and Juliano, B.O. (1982). 2-Acetyl-1-
pyrroline: An important aroma component of cooked rice. Chem.
Ind. (London) 12: 958–959.
Buttery, R.G., Ling, L.C., Juliano, B.O., and Turnbaugh, J.G. (1983b).
Cooked rice aroma and 2-acetyl-1-pyrroline. J. Agric. Food Chem. 31:
823–826.
Causse, M.A., et al. (1994). Saturated molecular map of the rice
genome based on an interspecific backcross population. Genetics
138: 1251–1274.
Chen, S.H., Wu, J., Yang, Y., Shi, W.W., and Xu, M.L. (2006). The fgr
gene responsible for rice fragrance was restricted within 69kb. Plant
Sci. 171: 505–514.
Chen, X., Temnykh, S., Xu, Y., Cho, Y.G., and McCouch, S.R.
(1997). Development of a microsatellite framework map providing
genome-wide coverage in rice (Oryza sativa L). Theor. Appl. Genet.
95: 553–567.
Cho, Y.G., McCouch, S.R., Kuiper, M., Kang, M.R., Pot, J., Groenen,
J.T., and Eun, M.Y. (1998). Integrated map of AFLP, SSLP, and RFLP
markers using a recombinant inbred population of rice (Oryza sativa
L.). Theor. Appl. Genet. 97: 370–380.
Damerval, C., de Vienne, D., Zivy, M., and Thiellement H. (1986).
Technical improvements in two-dimensional electrophoresis increase
the level of genetic variation detected in wheat seedling proteins.
Electrophoresis 7: 52–54.
Hood, E.E., Gelvin, S.B., Melchers, L.S., and Hoekema, A. (1993).
New Agrobacterium helper plasmids for gene transfer to plants.
Transgenic Res. 2: 208–218.
Huang, N., McCouch, S.R., Mew, T., Parco, A., and Guiderdoni, E.
(1994). Development of an RFLP map from a doubled haploid pop-
ulation in rice. Rice Genet. Newsl. 11: 134–137.
Incharoensakdi, A., Matsuda, N., Hibino, T., Meng, Y.L., Ishikawa,
H., Hara, A., Funaguma, T., Takabe, T., and Takabe, T. (2000).
Overproduction of spinach betaine aldehyde dehydrogenase in Esch-
erichia coli. Structural and functional properties of wild-type, mutants
and E. coli enzymes. Eur. J. Biochem. 267: 7015–7023.
Itani, T., Tamaki, M., Hayata, Y., Fushimi, T., and Hashizume, K.
(2004). Variation of 2-acetyl-1-pyrroline concentration in aromatic rice
grains collected in the same region in Japan and factors affecting its
concentration. Plant Prod. Sci. 7: 178–183.
Jin, Q.S., Waters, D., Cordeiro, G.M., Henry, R.J., and Reinke, R.F.
(2003). A single nucleotide polymorphism (SNP) marker linked to the
fragrance gene in rice (Oryza sativa L.). Plant Sci. 165: 359–364.
Johansson, K., El-ahmad, M., Ramaswamy, S., Hjelmqvist, L.,
Jornvall, H., and Eklund, H. (1998). Structure of betaine aldehyde
dehydrogenase at 2.1 A resolution. Protein Sci. 7: 2106–2117.
Livingstone, J.R., Maruo, T., Yoshida, I., Tarui, Y., Hirooka, K.,
Yamamoto, Y., Tsutui, N., and Hirasawa, E. (2003). Purification and
properties of betaine aldehyde dehydrogenase from Avena sativa. J.
Plant Res. 116: 133–140.
Lorieux, M., Petrov, N., Huang, N., Guiderdoni, E., and Ghesquiere,
A. (1996). Aroma in rice: Genetic analysis of a quantitative trait. Theor.
Appl. Genet. 93: 1145–1151.
Nakamura, T., Nomura, M., Mori, H., Jagendorf, A.T., Ueda, A., and
Takabe, T. (2001). An isozyme of betaine aldehyde dehydrogenase in
barley. Plant Cell Physiol. 42: 1088–1092.
Perez-Miller, S.J., and Hurley, T.D. (2003). Coenzyme isomerization is
integral to catalysis in aldehyde dehydrogenase. Biochemistry. 42:
7100–7109.
Qiu, Z.J., and Zhang, Y.S. (2003). Why fragrance rice produced in
Thailand can be sold worldwide? (In Chinese). World Agric. (China) 2:
33–36.
Romanczyk, L.J., McClelland, C.A., Post, L.S., and Aitken, W.M.
(1995). Formation of 2-acetyl-1-pyrroline by several Bacillus cereus
strains isolated from cocoa fermentation boxes. J. Agric. Food Chem.
43: 469–475.
Sambrook, J., and Russell, D.W. (2001). Molecular Cloning: A Labo-
ratory Manual, 3rd ed. (Cold Spring Harbor, NY: Cold Spring Harbor
Laboratory Press).
Schieberle, P., and Grosch, W. (1991). Potent odorants of the wheat
bread crumb. Z. Lebensm. Unters. Forsch. 192: 130–135.
Shi, W.W., Yang, Y., Chen, S.H., and Xu, M.L. (2008). Discovery of a
new fragrance allele and the development of functional markers for
the breeding of fragrant rice varieties. Mol. Breed. http://dx.doi.org/
10.1007/s11032-008-9165-7.
Sood, B.G., and Siddiq, E.A. (1978). A rapid technique for scent
determination in rice. Indian J. Genet. Plant Breed. 38: 268–271.
Steinmetz, C.G., Xie, P., Weiner, H., and Hurley, T.D. (1997). Structure
of mitochondrial aldehyde dehydrogenase: The genetic component of
ethanol aversion. Structure 5: 701–711.
Trossat, C., Rathinasabapathi, B., and Hanson, A.D. (1997).
1860 The Plant Cell
Transgenically expressed betaine aldehyde dehydrogenase effi-
ciently catalyzes oxidation of dimethylsulfoniopropionaldehyde and
v-aminoaldehydes. Plant Physiol. 113: 1457–1461.
Xu, M.L., Song, J.Q., Cheng, Z.K., Jiang, J.M., and Korban, S.S.
(2001). A bacterial artificial chromosome (BAC) library of Malus
floribunda 821 and contig construction for positional cloning of the
apple scab resistance gene Vf. Genome 44: 1104–1113.
Yajima, I., Yanai, T., Nakamura, M., Sakakibara, H., and Habu, T.
(1978). Volatile flavor components of cooked rice kaorimai (scented
rice, O. sativa japonica). J. Agric. Biol. Chem. 43: 2425–2429.
Yang, C.W., and Kao, C.H. (1999). Importance of ornithine-d-amino-
transferase to proline accumulation caused by water stress in de-
tached rice leaves. Plant Growth Regul. 27: 189–192.
Yoshihashi, T., Huong, N.T.T., and Inatomi, H. (2002). Precursors of
2-acetyl-1-pyrroline, a potent flavor compound of an aromatic rice
variety. J. Agric. Food Chem. 50: 2001–2004.
Yoshihashi, T., Huong, N.T.T., and Kabaki, N. (1999). Quality evalu-
ation of Khao Dawk Mali 105, an aromatic rice variety of northeast
Thailand. JIRCAS Working Report 30: 151–160.
Zhang, Z.W., Zhou, Y.M., Zhang, Y., Guo, Y., Tao, S.G., Li, Z., Zhang,
Q., and Cheng, J. (2005). Sensitive detection of SARS coronavirus
RNA by a novel asymmetric multiplex nested RT-PCR amplification
coupled with oligonucleotide microarray hybridization. In Methods
in Molecular Medicine, Vol. 144, Microarrays in Clinical Diagnos-
tics, T. Joos and P. Fortina, eds (Totowa, NJ: Humana Press), pp.
59–78.
The Badh2 Gene Inhibits 2AP Synthesis 1861
DOI 10.1105/tpc.108.058917; originally published online July 3, 2008; 2008;20;1850-1861Plant Cell
Mingliang XuSaihua Chen, Yi Yang, Weiwei Shi, Qing Ji, Fei He, Ziding Zhang, Zhukuan Cheng, Xiangnong Liu and
2-Acetyl-1-Pyrroline, a Major Component in Rice Fragrance, Encoding Betaine Aldehyde Dehydrogenase, Inhibits the Biosynthesis ofBadh2
This information is current as of October 11, 2020
Supplemental Data /content/suppl/2008/07/28/tpc.108.058917.DC1.html
References /content/20/7/1850.full.html#ref-list-1
This article cites 35 articles, 2 of which can be accessed free at:
Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X
eTOCs http://www.plantcell.org/cgi/alerts/ctmain
Sign up for eTOCs at:
CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain
Sign up for CiteTrack Alerts at:
Subscription Information http://www.aspb.org/publications/subscriptions.cfm
is available at:Plant Physiology and The Plant CellSubscription Information for
ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists