3. & &(
Department of Biochemical Science and Technology College of Life Science
National Taiwan University Doctoral Dissertation
4% +" )#/ -* Study on the Growth and the Hyper-accumulation of
Nicotine in Hairy Roots of Nicotiana tabacum
$27
Jung-Hao Wang
!58,6
!5810'
Advisors: Kung-Ta Lee, Ph.D.
Chi-Te Liu, Ph.D.
104 6
June 2015
Ñn íŊÉŦ(~ ǑP(2 GDžŧŞvǏO2
íCġÜlŊÉǑ@ĩÜlǏSÉ}JR Ǒ
I(Ǐť¯Lja¿ÕơƟªWÌÔÉ(Ő=<h£ıŌǏ°(ŏNJa
¿}±Ċǎa¿>ơļ|P;ĒPH<UŧŞ ıŌǑ >·
ĀÆ£é(ǏƋƀŁa¿ÏťƱñM 1 bëŐǏa¿pæĚü
ďČ<UŧŞ`*ǏH$Ă ´UNǑÜl£éǏSÉťƱñ a¿M<nþx£é1 äǏ!HXL< GvSıŌŧŞ§Ä ŜŠǑSÉťƱña¿}ƋƀŁa¿õÈjCġŧŞǑÜŧŞÓ -Va¿ M1 ÀŚǑCġ×ß ǏSÉťƱña¿f dõÈǏ$ĂH)ƵŘyĘ ĵƉǑŅìĀÆǏSÉ
PqP! NCġ×qPƂǏ$Ă1 (~ÏŀÜĨMďŕǑ
>·yĘįÐľƻǏSÉPƲôƙ}qPƤŃYŬM`ô
õÈ}ŻŹǏM ćãU%/`*ǑńĎưǀĄƚĽÖĥ,ėǏ.
)õÈǏQ+$zŒ ǑyĘŅ©ǏSÉNjƭǒŎƊǒ6Ǝǒ ŨźĺdǑ) >·ƗƗƅƅ $ 4µbǑ!1SÉ ǒDŽǒ łƨǒ8ĹǒĿƩ{ǏHŧŞ$Oľç}O pǑÏ ) 4¢r
Ɛ½ǑSÉPuńŷ<yĘŃ:ÀŚǑSÉPř0ƳyĘŀÜǒ
ÛďƏúĞǏ8ċÜlǁíŀÜ}ăƒǑSÉāèĆů¼ťƱña¿ǒ ƋƀŁa¿ǒƕĠƕĠĦa¿ǒǍƦƪa¿ǒ±Ċǎa¿ǒ¸ƺǂa¿ǒƔŋ lja¿ďŕǏHyĘ$ĂƑåǑSÉƕĠĦa¿-ªāű5Ɠë Į}ƿňƮ¬ŝàĘgĘ5iǑSÉǍƦƪa¿-ÛyĘőĻ
ëĮǑSÉƋdƶa¿ĥÆúĞǏĨM )î nş5Ɠ}¨
ĕ5ƓďŕǑSÉŸƇQa¿ÍŰ<Ē}ÔÉyĘƉǑSÉ- ĿğħƠőĻ¼¨ĕ56}Ôɪx¡úĢǑA1SÉ:Ǐě Ĉ÷ðcēǑ7FǏSÉ:}NjƭõÈǑPż³ ´
N'uß Ǐ¾\ÝÄ GË ǑÉÉ )Ǒ
Ĥşİ ƟªİĤŲƷƮSƍö-ÃÎņŭǏ¸Bnİ-ŧŞ
Ǒ¶ǏĤşİĝ?û-u}]WÌÔɪÒæÃǏø+Bn-öƟ ªWÌÔɪǑés)<nĤşİƛ=ǒ-uóŤ}WÌÔɪÒæZ^
ÂŽQáXǑoŧŞÜl Ǐ)çBƚĽǒ$SƍƚĽİĤŲƷƮA4Ǐ 8ċƚĽ3wWÌÔɪŖIJŴr#ŧŞōģǏï¦ŗ,ŧŞĤşİ-u }WÌÔɪÒæZ^ÂǏkĤşİ«BǑ-uóŤ®5Ǐ) ųƞnİĤŲƷƮA4 ZTL-DNA °rolļ|Ǐ5 rolArolBrolC
; rolDǑéstXļ|<nƃĤşİĭ@ǏE)ê}ĭ@Ó ÞÈĤşİģü vƢđĜŽáXǑoĩŧŞ Ǐ)ŲƷƮ 5Ƹ ĸ °rolļ|ǏĨçBrol śƝŲƷƮƛĮĤşİǑ)=NǏŲ ƷƮśrolB ÓrolC (ǏƈƆ=İĨĴ=İŢǑ[śrolA ÓrolD (2
NĪǑ¶Ǐäś rolB Ó rolC ŲƷƮSƍv&Ĥşİö-ī
ƥİÊǏŇK u(mćÔ ęVǑ[rolA ÓrolD (ĹĝNĪǑ
)Ú# rolB ;rolC hóŤ-u|6ǑƣäƹŶnjčDŵĔ5ƓǏś
rolB Ó rolC Ĥşİ ƖÛ¨ŪĒ}VƘ56 ļ|N¡ij3Đ näŮ-ģŲƷƮSƍ-@ĤşİǑ)!=Nƥİ ĤşİǏù
]ƖÛ¨ŪĒN¡Ǒ)!ī äŮ-ģŲƷƮSƍö-Ĥş İ}ƚĽ-ݨĕĶũǏ=NƖÛ¨ŪĒN¡ŔyĤşİ
¡ÒæǑĸƖÛ¨ŪĒǏ)çBROSƍƯ<śƝrolB ÓrolC Ĥ şİƍǏ=Nś¹ rolB Ó rolC (īĐ¡VƘ56Ǒƣ¶)Ü
Ĥşİ-uóŤ ǏƖÛ¨ŪĒ}VƘ56ţ¡å |
6ǑWÌÔɪóŤbpǏ)=NÙ^ķ²9ƚĽĤşİţŖ IJŴ¡ ºnƚĽ-İǏ[ŇŖIJŴţ¡}"ƚƽĝ Ǐijò
Ĥşİ WÌÔÉƬ+$ýeÍ]ǐŖIJŴÔÉƬ8ċžę
ÇÁĒĤşİ ¡NǏą@ŖIJŴÒæǑ¶Ǐ)=N-uī?
ûƚĽĤşİţī ŖIJŴǏijòĤşİ-u}ŖIJŴÒæ+$
|6óŤǑ[ŖIJŴĤşİ ÒæâƼƫčű¤ÅƜƬ
ƛĮǏĬšÍØäƼƫčƬƁTZƴLJÿ«|6 ; N¡Ǐ
[k{ƄƿʼnƌdžNſļ¨Ūǃ} ſļƌdžƘÎǃNǑ)=
NÍŰ ĉ»ƾ²]WÌÔɪĤşİb ǐĖ¡Ĥşİ-uûŢǑ
)Ú#Ǐ¥_+8ƑƃĤşİWÌÔɪ-öŧŞ}«BǑ
İĤŲƷƮǒƚĽǒĤşİǒrolļ|ǒWÌÔɪǒŖIJŴǑ
Abstract
Hairy root, which resulted from T-DNA transformation of Agrobacterium rhizogenes, is widely used in studying root biology. It is also applied in producing diverse plant secondary metabolites due to its fast-growth and metabolite- accumulating abilities. However, the regulatory mechanisms of hairy root initiation, growth, and metabolite accumulation are largely unknown. To expand the applicability of hairy roots, we used Nicotiana tabacum L. var Wisconsin 38, its pathogen A. rhizogenes A4, and its well-known metabolite nicotine as a study model to unveil the mechanisms that regulate hairy root growth and secondary metabolite accumulation. In the part of growth regulation, we focused on four rol genes, including rolA, B, C, and D, which are located on TL-DNA of A. rhizogenes A4.
These rol genes are known to participate in rooting; however, the means by which the rol genes contribute to the initiation and the maintenance of hairy roots remain unknown. In this study, we knocked-out these rol genes in A. rhizogenes A4 respectively, and used for inducing hairy roots. We found that A. rhizogenes lacking rolB or rolC induced hairy roots with less rooting ability than wild-type A.
rhizogenes, whereas lacking rolA or rolD showed no significant differences.
Moreover, tobacco hairy roots lacking either rolB or rolC exhibited fewer branch roots and lost their growth ability after long-term subculture than wild-type-induced hairy roots, whereas lacking of rolA or rolD did not show significant differences. We considered rolB and rolC involved mainly in the regulation of hairy root growth. Our microarray analysis revealed that the expression of several groups of genes encoding lipid transfer proteins (LTP) and reactive oxygen species (ROS)-related genes was significantly suppressed in rolB- or rolC- deficient hairy roots. We also found that
hairy root clones that exhibited greater branching also had higher levels of RolB, RolC, and the microarray-identified LTP genes. In addition, we compared the transcriptomic difference between hairy roots and un-infected intact roots by microarray, and the expression levels of the above mentioned LTP-encoding genes were dramatically higher in the hairy root. Moreover, ROS staining showed that ROS level were lower in rolB- or rolC- deficient hairy roots. We therefore suggest that up- regulating LTP and increasing the level of ROS are important for hairy root growth.
In the part of secondary metabolite regulation, we found that tobacco hairy roots accumulate much more nicotine than the intact roots, and the nicotine contents were positively correlated with the amount of another metabolite anabasine, indicating hairy roots had higher secondary metabolic flux. By real-time PCR analysis, hairy roots had more abundant expression of genes encoding enzymes in nicotine biosynthetic pathway and storage transporters, indicating the accumulation of nicotine in hairy roots is via transcriptional regulation. Moreover, hairy roots with a higher growth rate had greater nicotine content, suggesting that growth and nicotine production are regulated synchronically. Nicotine up-regulation in hairy roots was regulated by ethylene response factor (ERF)189 and ERF199 to activate the key enzymes putrescine N-methyltransferase and N-methylputrescine oxidase with a jasmonic acid (JA)-independent signal. However, the possible regulator has not been identified. These findings indicate high secondary metabolites accumulated hairy root clones can be simply selected by measuring their growth rate, which expand the hairy root researches and applications in secondary metabolites.
Key words: Agrobacterium rhizogenes, Nicotiana tabacum, hairy root, rol genes, secondary metabolites, nicotine.
Contents
ūÉ i
lƧiii
Abstract v Contents vii
Contents of Tables xiii
Contents of Figures xv
Abbreviations xvii Chapter 1: Introduction 1. Agrobacterium 1 1.1. T-DNA transferring mechanism (infection mechanism) 2
1.1.1. Attachment 2 1.1.2. T-DNA processing and exporting from bacterium 4
1.1.3. Nuclear targeting and chromosomal integrating 4
1.2. Crown gall 6
1.3. Opine 8 1.4. Hairy root 8
2. Genes on Agrobacterium rhizogenes T-DNA 9
2.1. rolA 12 2.1.1. rolA affects plant morphogenesis 12
2.1.2. rolA and plant hormone 13
2.1.3. rolA promoter 13
2.1.4. RolA biological functions 14
2.2. rolB 15
2.2.1. rolB affects plant morphogenesis 16
2.2.2. rolB and auxin 16
2.2.3. rolB promoter 18
2.2.4. RolB biological function 20
2.3. rolC 20 2.3.1. rolC affects plant morphogenesis 21
2.3.2. rolC and auxin 22
2.3.3. rolC and cytokinin 22
2.3.4. rolC and gibberellins 23
2.3.5. rolC promoter 23
2.4. rolD 24 2.5. orf3n 25 2.6. orf8 25 2.7. orf13 and orf14 27 2.8. orf13a 29 3. Plant secondary metabolites 29
3.1. Secondary metabolites in hairy root 31
3.2. rolA affects secondary metabolites 32
3.3. rolB affects secondary metabolites 33
3.4. rolC affects secondary metabolites 34
4. Nicotine regulatory mechanism 36
5. Objectives 39
Chapter 2: Materials and Methods
1. General DNA manipulation 41
1.1. DNA quantification 41
1.2. Plasmid DNA extraction from E. coli 41 1.3. Polymerase chain reaction (PCR) 42
1.4. DNA purification 43
1.5. Cloning by restriction-ligation method 44
1.6. Cloning by Gateway system 45
1.7. Total DNA isolation from A. rhizogenes 45 1.8. Plasmid DNA extraction from yeast 46
2. General RNA manipulation 47
2.1. RNA quantification and quality control 47
2.2. RNA extraction from plant tissue 47
2.3. First-strand cDNA synthesis 48
2.4. Quantification reverse transcription polymerase chain reaction 49
3. General protein manipulation 50
3.1. Protein quantification 50
3.2. Total protein extraction from yeast 50
3.3. Total protein extraction from plant tissue 51
4. Microorganisms transformation 51
4.1. E. coli transformation by heat shock 51
4.2. A. rhizogenes transformation by electroporation 52
4.3. Yeast transformation by lithium acetate (LiAc) mediated method 53
5. Individual rol genes deficient strains generation 54
5.1. Generation of individual rol genes deficient strains 54
5.2. Confirmation using Southern blot 56
6. Transcriptional and translational fusion 57
6.1. Generation of destination vector pGWYFP 57
6.2. Expression clones establishment 58
7. Hairy root induction 59
7.1. N. tabacum W38 growing 59
7.2. Hairy root induction 60
7.3. Hairy root confirmation 61
8. Microarray assay and data analysis 62
9. Quantification of alkaloids 63
10. Yeast two-hybrid 64
10.1. Bait protein construction 64
10.2. Poly A+ RNA purification 64
10.3. cDNA synthesis and amplification 65
10.4. Gal4 DNA activation domain fused cDNA library of hairy root 66
11. Phosphatase activity assay 66
Chapter 3: Results and Discussion Transcriptomic analysis reveals that ROS and genes encoding LTPs are associated with tobacco hairy root growth and branch development Results 1. rol genes deficient A. rhizogenes mutants grew faster than wild-type. 69
2. ΔrolB and ΔrolC A. rhizogenes mutants have decreased hairy root induction
ability 71 3. Aberrant hairy roots induced by rol-deficient A. rhizogenes mutants 73
4. Microarray data analysis 78 5. The expression levels of genes encoding LTPs were related to hairy root
growth 83 6. ROS accumulate in hairy roots but decrease when either rolB or rolC is
knocked out 85
Discussion 87
Fast-growing Nicotiana tabacum hairy roots accumulate more nicotine than slow-growing hairy roots due to systematic up-regulation of nicotine biosynthetic
genes Results
1. Nicotine accumulated in N. tabacum hairy roots 92 2. Positive corrections were found between the contents of nicotine and
nornicotine and between the contents of nicotine and anabasine in hairy roots 93 3. Transcripts of the nicotine biosynthetic gene were up-regulated in hairy roots
94 4. Growth rate and nicotine content were positively correlated in hairy roots 97 5. The jasmonic acid pathway is not the activator of nicotine biosynthesis in
hairy roots 99
Discussion 100
Chapter 4: Perspectives
1. Functional exploration of LTPs 107
2. Exploring the biochemical function of RolB and RolC proteins 108
3. The mechanism of secondary metabolite accumulation in hairy roots 113
Chapter 5: Conclusion 117
References 119 Appendix 155
Contents of Tables
Table 1-1. Secondary metabolites accumulations in hairy root tissues. 31
Table 1-2. Secondary metabolites affected by rol genes. 32
Table 3-1. GO results for transcripts down-regulated in HRΔrolB compared with HRWT. 80 Table 3-2. GO results for transcripts down-regulated in HRΔrolC compared with HRWT. 81 Table 3-3. GO results for transcripts up-regulated in HRWT compared with tobacco intact roots. 82
Table 3-4. The expression levels of rolB/C and LTPs in hairy root clones with different branch root densities. 84
Supplementary Table S1. Reference genes for qRT-PCR. 155
Supplementary Table S2. PCR conditions. 156
Supplementary Table S3. Primer used in this study. 158
Contents of Figures
Figure 1-1 The rol genes and the other loci on A. rhizogenes A4 TL-DNA. 10
Figure 1-2 The overview of rolB-mediated plant responses. 15
Figure 1-3 Biosynthetic pathway, transportation, and storage of nicotine. 37
Figure 2-1 Mechanism of homologous recombination. 55
Figure 2-2 pGWYFP vector for rol gene complement. 58
Figure 2-3 Inducing tobacco hairy root. 61
Figure 3-1 Southern blot confirmation of respective rol gene deficient strains. 70
Figure 3-2 The growth curves of wild-type and respective rol genes deficient A. rhizogenes. 70 Figure 3-3 The day of the first root emergence post infection (DREPI). 72
Figure 3-4 Primary root number per leaf disc (R/L ratio). 73
Figure 3-5 Morphology of hairy roots at 18 days post-subculture. 74
Figure 3-6 Population distribution of the different hairy root architecture parameters. 76 Figure 3-7 Box plot analysis of hairy root architecture. 77
Figure 3-8 ROS content of HRWT, HRΔrolB, and HRΔrolC. 86 Figure 3-9 Nicotine contents in intact roots, excised roots, and hairy roots. 92
Figure 3-10 The correlation between the contents of (A) nicotine and nornicotine and (B) nicotine and anabasine in tobacco hairy roots. 94
Figure 3-11 Transcripts analysis of the nicotine biosynthetic genes. 96 Figure 3-12 The relationship between (A) nicotine content and growth, (B) the expression levels of rolC and growth, and (C) expression levels of rolB and rolC. 98
Figure 3-13 The transcript levels of AP2/ERFs in intact roots and hairy root clones 9, 22, and 3. 100 Figure 4-1 Sequence clustering of LTPs identified from microarray. 108 Figure 4-2 RolC has transcriptional activity in yeast. 112
Abbreviations
A. belladonna Atropa belladonna
A. rhizogenes Agrobacterium rhizogenes
A. rubi Agrobacterium rubi
A. tumefaciens Agrobacterium tumefaciens
A. vitis Agrobacterium vitis
aa amino acid
ABA abscisic acid
AbA aureobasidin A
ACC 1-aminocyclopropane-1-carboxylic acid
AD activation domain
adc arginine decarboxylase
ADE2 gene encoding Phosphoribosylaminoimidazole
carboxylase, involving in adenine synthesis
ANOVA analysis of variance
ao aspartate oxidase
AP2 domain APETALA2 domain
At Arabidopsis thaliana
AUR1-C inositol phosphorylceramide synthase
BD binding domain
BiFC bimolecular complementation
bp base pair
bZIP basic lucine zipper
C. roseus Catharanthus roseus
CaMV Cauliflower mosaic virus promoter
cDNA complementary DNA
CDPK; CPK calcium-dependent protein kinase
chv chromosomal virulence genes
DAD diode array detector
DF deleting fragment
Dof DNA binding with one finger
dpi day post induction
dps day post subculture
DMSO dimethyl sulfoxide
DREPI the day of the first root emergence post induction
dsDNA double-stranded DNA
EDTA Ethylenediaminetetraacetic acid
ef-1α elongation factor 1α
ERF ethylene response factor
GA gibberellic acid
GAL4 galactose induced gene
GST glutathione transferase
HA hemaglutinin
HIS3 gene encoding Imidazoleglycerol-phosphate
dehydratase, involving in histindine synthesis
HRΔrolA hairy root induced by ΔrolA A. rhizogenes A4
HPLC high-performance liquid chromatography
IAA indole-3-acetic acid
iaaH gene encoding indole acetamide hydrolase
iaaM gene encoding tryptophan 2-monooxygenase
IAM indole-3-acetamide
ILA indole-3-lactate
ipt isopentenyl transferase
IPTG Isopropyl β-D-1-thiogalactopyranoside
JA jasmonic acid
JAs jasmonic acid and its derivates
JAT1 jasmonate-inducible alkaloid transporter 1
JAZ JASMONATE-ZIM DOMAIN
k. daigremontiana kalanchoë daigremontiana
L25 L25 ribosomal protein
LB left border
LB medium Luria-Bertani medium
LRD first-order lateral root densities per centimeter of
main root
LRN first-order lateral root numbers per main root
LTP lipid transfer protein
LPS lipopolysaccharide
M. sexta Manduca sexta
mate multi-drug and toxic compound extrusion
meJA methyl-jasmonic acid
MEL1 gene encoding alpha-galactosidase
mpo methylputrescine oxidase
MRL main root length
MS Murashige and Skoog medium
N. tabacum Nicotiana tabacum
Nt Nicotiana tabacum
NtBBF1 Nicotiana tabacum rolB domain B factor 1
NtBRF1 Nicotiana tabacum rol binding factor 1
Ntcp-23 Nicotiana tabacum circadian protein 23
Ntubc2 Nicotiana tabacum ubiquitin-conjugating enzyme E2
ODC onithine decarboxylase
ORF open reading frame
P. ginseng Panax ginseng
phi-2 phosphate-induced gene 2
pmt putrescine N-methyltransferase
pNPP para-nitrophenylphosphatase
pp2A protein phosphatase 2A
PR pathogen-related
qpt quinolinic acid phosphoribosyl transferase
qs quinolinic acid synthase
R. cordifolia Rubia cordifolia
rat resistant to Agrobacterium transformation
RB right border
Ri plasmid root-inducing plasmid
RL ratio primary root numbers per leaf disc on 21 day post
induction
rol root loci
ROS reactive oxygen species
SA salicylic acid
samdc S-adenosylmethionine decarboxylase
sams S-adenosylmethionine synthase
SD medium synthetic defined medium
SOEing PCR Splicing by overlap extension polymerase chain
reaction
spds spermidine synthase
ssDNA single-stranded DNA
SSH suppression substrative hybridization
T-DNA transfer-DNA
tac-9 tobacco actin 9
Taq Thermus aquaticus
TC transcriptional fusion
Ti plasmid tumor-inducing plasmid
TL translational fusion
TLRL total first-order lateral root lengths per main root
tubA1 α-tubulin
uidA, GUS β-D-glucuronidase
UPP unipolar polysaccharide
UTR untranslated region
V. amurensis Vitis amurensis
vir virulence
WT wild-type
YFP yellow fluorescent protein
Chapter 1: Introduction
1. Agrobacterium
Agrobacterium, which belongs to the family Rhizobiaeae, is a genus of gram negative soil bacteria. Some species of Agrobacterium are plant pathogens which can cause a variety of neoplasm symptoms after infecting plant. For example, A.
tumefaciens and A. vitis cause crown gall disease, A. rubi causes cane gall disease, and A. rhizogenes causes hairy root disease. It is believed that these disease symptoms are caused by inter-kingdom gene transfer from Agrobacterium to host plant. These transferred genes were located on the transferred DNA (T-DNA), which is harbored in the tumor-inducing plasmid (pTi) of A. tumefaciens and A. vitis, or the root-inducing plasmid (pRi) of A. rhizogenes.
pTi and pRi are approximate 200-300 kbp in size. In addition to T-DNA, these plasmids encode genes with function in plasmid conjugation, T-DNA processing and transfer, and opine catabolism. T-DNA is the region of approximate 10-30 kbp flanked by two border sequences, which are called left border (LB) and right border (RB). In general, the border sequences are 25 bp in length with directly repeated orientation (Yadav et al. 1982; Zambryski et al. 1982). The RB is essential for tumorigenesis and rhizogenesis, and it directs T-DNA replication for transfer from RB to LB (Shaw et al. 1984; Wang et al. 1984). In contrast, the LB is dispensable for tumorigenesis and rhizogenesis; therefore, some of the transformation might lose the sequences near the left border and some would carry plasmid sequences out of the LB (Joos, H. 1983). In addition, some of bacterial chromosome sequences could be transferred into plant genome (Ulker et al. 2008). These Agrobacterium-mediated DNA transfer events promote inter-kingdom gene flow.
Due to the ability of gene transfer, Agrobacterium species are widely applied in plant genetic transformation via binary vector strategy to study gene function, promoter trapping, and metabolic engineering (Hoekema et al. 1983). Moreover, Agrobacterium could transfer gene to many other eukaryotic cells, including fungi and animal cells (Lacroix et al. 2006). DNA transferring mechanism and the characterizations of hairy root are described in the following sections.
1.1. T-DNA transferring mechanism (infection mechanism)
There are three major steps occurring during T-DNA transfer process. First, the bacteria attach to the host cell. Second, the vir operon of Agrobacterium is activated by plant-derived phenolic compounds, and the T-DNA is processed and transferred to plant cell by Vir proteins. Finally, the T-DNA is transferred into plant nucleus and integrated into chromosome DNA with the aid of host plant proteins. These processes have been reviewed in many articles (winans 1992; Costantino et al. 1994;
Ziemienowicz 2001; Tzfira and Citovsky 2002; Cascales and Christie 2003; Brencic and Winans 2005; Chen 2005; Christie et al. 2005; McCullen and Binns 2006;
Rodriguez-Navarro et al. 2007; Gelvin 2010a, b; Pitzschke and Hirt 2010). The brief introduction of T-DNA transportation from Agrobacterium to plant cell is presented below.
1.1.1. Attachment
Microorganisms attach to their host plants before symbiosis or pathogenesis. In A. tumefaciens, several attachment strategies have been identified: flagellum- dependent attachment, pili attachment, and unipolar polysaccharide (UPP) adhesin (reviewed by Heindl et al. 2014). Before attachment, Agrobacterium forms biofilm with a hydrated macromolecular matrix that consists of exopolysaccharides,
extracellular DNA, and proteins (Costerton et al. 1995; Flemming and Wingender 2010). To date, six biofilm-forming polysaccharide species have been identified in A.
tumefaciens, these are UPP adhesin, cellulose, succinoglycan, cyclic β-1,2-glucans, β-1,3-glucans, and membrane lipopolysaccharides (LPS). UPP mediates polar attachment of A. tumefaciens to plants or other surface (reviewed by Matthysse 2014). Presence of bacterial cellulose enhances the attachment, and the lose of cellulose shows a slightly reduced ability of tumor formation (Matthysse et al 1981;
Matthysse, 1983; Mattysse et al. 2005). Cyclic β-1,2-glucans is required for biofilm formation and virulence (Dougals et al. 1982; Xu et al. 2012). Succinoglycan and β-1,3-glucans have little effects on biofilm formation and tumor formation (Tomlinson et al. 2010; Ruffing and Chen 2012; Xu et al. 2012). No conclusive result to date has demonstrated the LPS affects biofilm formation and bacteria-plant attachment. Overall, some forward genetic approaches have indicated kinds of polysaccharides participate in attachment, but the detail mechanism of bacterial attachment remains unclear.
On the other side, plant surface molecules provide recognition for pathogen. A group of Arabidopsis resistant to Agrobacterium transformation (rat) mutants were isolated, and some of them were poorly bound with A. tumefaciens (Zhu et al. 2003).
A well-characterized mutant, rat1, which showed a reduced expression level of the cell wall arabinogalactan protein, decreases the binding of bacteria (Zhu et al. 2003;
Gaspar et al. 2004). Further studies are needed to unveil the interaction between Agrobacterium and host cells.
1.1.2. T-DNA processing and exporting from bacterium
Plant signals initiate the T-DNA processing step. Naturally, phenolic compounds activate the VirA/VirG two component system, which is constitutively expressed at a basal level and is highly induced in a positive feedback manner by plant signals (Stachel et al. 1985; Winans et al. 1988). The VirA is a membrane bound sensor kinase, which perceives phenols, aldose monosaccharides, low pH, and low phosphatase. Upon receiving plant signal, VirA auto-phosphorylates itself, and then the phosphate group is transferred to cytoplasmic regulator VirG. The phosphorylated VirG binds at vir boxes of vir promoter and activates transcription to initiate T-DNA processing (reviewed by Palmer et al. 2004; Brencic and Winans 2005).
After production of Vir proteins, Agrobacterium generates single stranded T- DNAs (Stachel and Nester 1986). In A. tumefaciens, VirD2 cleaves and covalently binds to the 5’ end of T-DNA right border to form an immature T-complex, which is VirD2-ssT-DNA complex. This complex would be recruited and translocated into plant cells by bacterial type IV secretion system, which is composed of 11 different VirB proteins and the VirD4 (reviewed by Chen et al. 2005). In addition to VirD2- ssT-DNA complex, other Vir factors, including VirE2, VirE3, VirF, and VirD3, from Agrobacterium are translocated into plant cells during infection (Vergunst et al. 2000;
Vergunst et al 2005).
1.1.3. Nuclear targeting and chromosomal integrating
After VirD2-ssT-DNA imported into the plant cell, VirE2 coats the single stranded T-DNA to form a mature T-complex, which is VirD2/VirE2/T-DNA complex. This VirD2/VirE2/T-DNA complex is a highly ordered structure to facilitate transport through the nucleopore complex (Duckely and Hohn 2003). VirD2 and
VirE2 occupy the surface of single stranded T-DNA to resist nuclease, and both VirD2 and VirE2 contain nucleus localization signals for plant importin recognition. The mature T-complex is transported into host nucleus with the aid of plant alpha type importin (Koncz et al. 1989; Tinland et al. 1995; Deng et al. 1998; Bakó et al. 2003;
Bhattacharjee et al. 2008). VirE2 has been suggested to have an additional function as a transmembrane DNA transporter to translocate actively the ssDNA into host cell (Dumas et al. 2001; Duckely et al. 2005; Grange et al. 2008). In addition, VirE3 is suggested to interact with VirE2 and plant importin to help nucleus translocation of mature T-complex (Lacroix et al. 2005).
In plant nucleus, VirF activates the host proteasome machinery to degrade the proteins surrounding the T-DNA to release the T-DNA from mature T-complex. The T-DNA is subsequently for chromosomal integration (Schrammeijer et al. 2001;
Tzfira et al. 2004). The T-DNA integration is completed by illegitimate recombination (Gheysen et al. 1991). There is little known about the mechanisms and the proteins involved in this process. Some histone proteins as well as histone modifying proteins were proposed to be capable of interacting with T-DNA and therefore T-DNA could be targeted to chromosome (Zhu et al. 2003; Crane and Gelvin, 2007).
Overexpressing Arabidopsis histone HAT1 enhances the transformation efficiency in Arabidopsis and rice (Yi et al. 2002; Yi et al. 2006; Zheng et al. 2009). Besides, proteins participating in recombination and DNA repair are essential for T-DNA integration (Sonti et al. 1995; Nam et al. 1998). The detail about how these genes involved in T-DNA integrating should be further elucidated.
Although the studies of T-DNA processing and transportation mechanisms are focused on A. tumefaciens, it is believed that A. rhizogenes shares the same infection
mechanism with A. tumefaciens. However, A. rhizogenes does not encode VirD2 and VirE2. Instead, full-length and split C-terminal region of GALLS proteins, encoded by a GALLS gene, are the functional homologs to VirD2 and VirE2 respectively (Hodges et al. 2006; Hodges et al. 2009).
1.2. Crown gall
Agrobacterium tumefaciens T-DNA expression in transformed plant results in crown gall formation. Genes located on the T-DNA are so-called oncogenes, including the most common six ones, gene 5, iaaM, iaaH, ipt, gene 6a, and gene 6b.
A. tumefaciens losing iaaM or iaaH produces tumors with differentiated shoots, and the tumor has reduced auxin levels (Garfinkel and Nester 1980; Akiyoshi et al.
1983). iaaM and iaaH encode a tryptophan monooxygenase and an indole-3- acetamide hydrolase, respectively. These two enzymes convert tryptophan to auxin indole-3-acetic acid (IAA) (Schröder et al. 1984; Thomashow et al. 1984;
Thomashow et al. 1986; van Onckelen et al. 1986). The expression of these two genes results in more than 10-fold greater free form IAA level in crown gall tumors than in the periphery (Weiler and Spanier 1981; Veselov et al. 2003). The auxin overproduction is considered to promote tumorigenesis and maintain tumor morphology.
Mutation of ipt causes tumor with small, rooty morphology and a decreased level of zeatin-type cytokinins (Garfinkel and Nester 1980; Akiyoshi et al. 1983). ipt encodes an isopentenyl transferase, which condenses a molecule of adenosine monophosphate with an isoprenoid unit, the rate-limiting step of cytokinin synthesis (Astot et al. 2000). Therefore, cytokinin levels in crown gall are 100-fold higher than those in periphery (Weiler and Spanier 1981). Bedises, A. tumefaciens encodes an
additional ipt-homologous gene, tzs, located on vir operon. It allows bacteria to synthesize cytokinin after activation of VirA/G two component system, and the expression of tzs enhances virulence (Gaudin et al. 1994), which indicates that cytokinin production is closely related to the tumorigenesis.
Gene 6a encodes an opine permease with little effects on tumorigenesis (Messens et al. 1985). The gene 6b of octopine-type strains is capable of inducing small tumors on Kalanchoë; however, it showed no effects on tobacco (Garfinkel et al. 1981). In gene 6b-transgenic plants, several abnormal growth phenotypes were found, including tubular leaves, thicker roots, and ectopic shoots (Tinland et al. 1992;
Wabiko and Minemura 1996; Grémillon et al. 2004). The detailed biochemical and cellular function of gene 6a and 6b in tumorigenesis should be further elucidated.
Mutation of gene 5 has no phenotypic effect on tumor. However, co-mutation of gene 5 with iaaM or iaaH causes more shoots than iaaM or iaaH mutant (Leemans et al. 1982). Gene 5 encodes an enzyme converting tryptophan into indole-3-lactate (ILA), which competes with IAA for the auxin-binding proteins (Körber et al. 1991;
Sprunck et al. 1995). The expression of gene 5 was positively regulated by auxin, but negatively regulated by ILA (Korbor et al. 1991). These indicate gene 5 play a non- essential role in tumorigenesis.
It is believed that hormone regulation causes crown gall formation. The stem inoculated with A. tumefaciens produces high levels of auxin and cytokinin followed by increases in ethylene and abscisic acid (Veselov et al. 2003). The ethylene insensitive tomato mutant Never ripe generates smaller crown gall tumor than wild- type tomato, and the tumor from mutant contains 50-fold lower ethylene level (Aloni et al. 1998). These indicated the tumorigenesis is regulated by multiple endogenous
hormone biosynthesis, and iaaM/iaaH and ipt from A. tumefaciens contribute dominantly to the hormone re-balance.
1.3. Opine
Besides neoplastic growth related genes, T-DNA encodes opine biosynthetic genes that allow host plant to produce opine in crown galls or hairy roots. Opines are low molecular weight compounds consisting of nitrogen and carbon, and serve as nutrient sources for Agrobacterium (Hong et al. 1997). Nowadays, over 30 kinds of opines have been found. Each Agrobacterium strain produces one specific opine compound; therefore, the types of opine have been used for classification of Agrobacterium (Petit et al. 1983). Each set of opine catabolism gene is found in non- T-DNA region of pTi/pRi in corresponding strain. Opine can be utilized by only a little groups of soil organisms, which offers a competitive advantage to Agrobacterium (Wilson et al. 1995).
1.4. Hairy root
Unlike crown gall, there has not been a generally agreed mechanism of hairy root formation. The expression of A. rhizogenes T-DNA genes in infected plant results in adventitious root disease syndrome (Chilton et al. 1982; Tepfer 1984;
Cardarelli et al. 1987). The hairy roots can emerge from most types of plant tissues, such as shoots, roots, and calli. The root emergence from plant cells is involved with cell re-programming process, which most be likely caused by hormone re-balance.
Some strains of A. rhizogenes encode iaaM/iaaH functional homologs aux1/aux2 (Gaudin and Jouanin; 1995), but these auxin biosynthetic genes are not essential for hairy root formation. Instead, rol genes are sufficient for hairy root formation.
However, there are many controversial results in these oncogenes encoded by A.
rhizogenes. Due to the highly potential applicability of hairy root, we aimed to understand the mechanism of the formation of hairy roots. The advantages of hairy root in biotechnology will be briefly described in next paragraph, and the studies about oncogenes encoded by A. rhizogenes will be stated in the next section.
Hairy root is characterized by high growth rate and genetic stability, and it could maintain these characteristics in hormone-free medium (Benvenuto et al. 1983).
Unlike tumor induced by A. tumefaciens, hairy root is only composed of transformed cells (Bercetche et al. 1987). Hairy root is widely applied in root biology because it can harbor interested nucleotide fragment homogeneously with root physiology.
Furthermore, hairy root provides a route to plant genetic engineering for producing heterologous proteins and for secondary metabolites. For the past three decades, many researches have proposed hairy root could accumulate much higher plant secondary metabolite levels compared with intact plant tissues. Furthermore, the secondary metabolites production could be enhanced via expressing structural genes of synthetic pathway, down-regulating competitive pathways, controlling the environmental factors such as light and sucrose, or treating with plant hormone such as methyl jasmonate (MeJA) or salicylic acid (SA) (reviewed by Giri and Narasu 2000; Srivastava and Srivastava 2007; Mehrota et al. 2010; Zhou et al. 2011).
Furthermore, hairy root culture could be scaled up to industrial level (Mehrotra et al.
2008; Baque et al. 2011). These advantages of hairy root make it is getting attention.
2. Genes on Agrobacterium T-DNA
After T-DNA integrated into plant chromosome, the expression of T-DNA gene would cause the neoplastic effects on plant, such as tumorigenesis in the case of A.
tumefaciens and rhizogenesis in the case of A. rhizogenes. The well-studied gene loci
are opine synthesis genes and hormone-related genes functionally homologs with the above-mentioned ones among Agrobacterium species. In addition to these genes, there are some unique genes in A. rhizogenes, including rol genes and other T-DNA loci. These genes affect plant in a wide diversity and result in rhizogenesis.
White and coworkers generated several deletions and transposon insertion- mutations on the TL-DNA of Ri plasmid A4, and they found that four of the potential 18 open reading frames on the TL-DNA affect the induction of roots on Kalanchoë daigremontiana. Therefore, these four loci were then named root locus A-D (rolA-D) genes (1985). Slightom and coworkers demonstrated these genes are corresponding to open reading frame (ORF) 10, 11, 12, and 15, respectively (1986). The diagram of genes on TL-DNA is shown in Figure 1-1. Many researchers have been interested in how these gene affect plant due to their extremely strong effects on plant growth, hormone balance, and metabolic flux.
Figure 1-1 The rol genes and the other loci on TL-DNA of A. rhizogenes strain A4.
The open reading frames were predicted by ORF finder of VectorNTI 10 (Life technologies) and the results were checked by nucleotide database of NCBI website.
A. tumefaciens harboring each one of rolA, rolB, or rolC is sufficient to induce root formation on tobacco leaves, and rolB or rolC alone is able to stimulate rooting on K. daigremontiana leaves (Spena et al. 1987; Vilaine et al. 1987). However, each pairwise combination of rolA, B, and C genes driven by their own promoters showed more efficient rooting abilities than any single gene, and three genes all together could promote the root production with the greatest efficiency (Spena et al. 1987).
These different degrees in root-promoting abilities suggested that these rol genes have different biological functions and act synergistically in hairy root formation.
The root emerging from leave cell is certainly caused by cell re-differentiation.
Auxin is the first considered hormone participating in root growing. Compared with intact roots, tobacco hairy roots accumulate approximately 2.5-fold higher auxin concentration (Spanò et al. 1988). Moreover, Lotus comiculatus hairy roots showed 100- to 1000-fold increase in auxin sensitivity compared with intact roots (Shen et al.
1988). By measuring the transmembrane potential differences between single rol gene transformed tobacco mesophyll protoplast and non-transformed cells, Maurel and colleagues discovered that rolB transformed cells could increase the auxin sensitivity up to 10000-fold, while the rolA up to 1000-fold, and the rolC up to 10-fold; whereas the TL-DNA transformed cells only raise the sensitivity to 30-fold (1991). These data demonstrate that hairy root growth closely relates to auxin responses with the fact that hairy root accumulates higher auxin and has enhanced auxin perception at the same time. In addition, rol genes act in a synergistic manner in rhizogenesis, but they did not act synergistically in raising auxin sensitivities.
In addition to promoting abnormal growth of plants, rol genes have strong effects on stimulating secondary metabolite accumulations. These indicated rol genes
alter plant physiology in a diverse range. I will introduce the biochemical and genetical studies of rol genes in plant in the introduction section 2.1~2.4, and the effects of rol genes on secondary metabolites in the next introduction section.
Besides rol genes, there are many other predicted ORFs on T-DNA. In transposon-mutagenesis experiment, these ORFs did not affect rooting (White et al.
1985). However, plant transformed with these genes would show diversified morphologies and altered hormone sensitivity. More details about these genes, including orf3n, orf8, orf13, orf13a, and orf14, will be introduced in the introduction section 2.5~2.9.
2.1. rolA
rolA encodes a small protein of approximate 11 kDa molecular mass.
Oligonucleotide sequences of rolA gene in all type of Ri plasmids share high homology (Nilsson and Olsson 1997). rolA was initially demonstrated as a gene related to rhizogenesis; however, later research discovered that rolA is a minor factor in rooting. In addition to rhizogenesis, rolA has been proposed to stimulate secondary metabolites in many types of plant tissues. The detailed descriptions are presented below.
2.1.1. rolA affects plant morphogenesis
A. tumefaciens harboring rolA with its promoter was able to induce rooting on tobacco leaf discs (Spena et al. 1987; Vilaine et al. 1987), but not on Kalanchoë leaves (Spena et al. 1987). rolA-deficient A. rhizogenes induced thicker and more curled hairy roots (White et al. 1985). The phenotypic changes in rolA-transgenic tobacco include highly wrinkled leaves, shorter internodes, and more condensed inflorescences with larger flowers (Schmülling et al. 1988; Sinkar et al. 1988); by
contrast, it was also reported to have smaller flowers with lower male fertility (Sun et al. 1991; Martin-Tanguy et al. 1993; Michael and Spena 1995). rolA-transgenic tomato had longer internodes, a smaller root system, smaller wrinkled leaves, smaller flowers, and lower pollen germination rate (van Altvorst et al. 1992).
2.1.2. rolA and plant hormone
rolA could increase the sensitivity to auxin in transgenic plant (Maurel et al.
1991; Vansuyt et al. 1992). Besides, rolA-expressing tobacco showed a similar phenotype with wild-type plant treated with gibberellic acid (GA) biosynthesis inhibitor; however, treating GA with rolA-transgenic tobacco only partially restored the phenotypic change (Dehio et al. 1993). Mortiz and Schmülling discovered that two active GAs, GA1 and GA20, were reduced in rolA-transgenic tobacco plant, and the precursors GA53 and GA19 were accumulated, indicating blocking GA synthetic pathway partially explains the phenotypic change caused by rolA (1998). Other hormone levels were measured in rolA-transgenic tobacco (Dehio et al. 1993), but no conclusive result could be proposed from the above reports.
2.1.3. rolA promoter
Transformation of rolA along with its 473-nucleotide upstream sequence, which is similar to some upstream sequences of auxin-responsible genes, was sufficient to cause the phenotypic change in tobacco (Carneiro and Vilaine 1993). In the same report, they demonstrated that stem had the most abundant rolA mRNA level, which was 5-fold and 50-fold higher than those in leaf and in root, respectively. rolA transcripts containing a 5’-untranslated region (5’-UTR), which would be spliced in Arabidopsis, was proposed to be an indispensable fragment to rolA expression, and it might act as a cis-acting regulatory factor (Magrelli et al. 1994). Pandolfini and
coworkers found that rolA mRNA could be transcribed in bacteria. However, the rolA transcripts were abolished while the 5‘-UTR was deleted (2000). This 5’-UTR has been proposed as a bacterial promoter. In 1996, Guivarc’h and coworkers expressed rolA driven by its 477 bp or 366 bp long upstream fragments, and they discovered the longer promoter would induce wrinkled leaves and short internodes in transgenic tobacco, whereas the shorter promoter only cause a dwarf phenotype with normal leaves (1996). In summary, rolA driven by its own promoter expresses in both prokaryotic and eukaryotic cells under the regulation of 5’-UTR, and the tissue- specific activation pattern is regulated by its 477 bp long promoter sequence.
2.1.4. RolA biological functions
Through sequence analysis and structure modeling, RolA was proposed to be a DNA- binding protein owing to the fact that it is a alkaline protein structurally homologous with papillmavirus E2 DNA-binding protein (Levesque et al. 1988;
Rigden and Carneiro 1999). However, the RolA-GUS transgenic tobacco cells showed the lowest GUS activity in nucleus and the highest in plasma membrane system (Vilaine et al. 1998). There is no transmembrane signals in RolA protein, which indicated RolA is a non-integrated membrane protein. Moreover, RolA might expresses in not only plant but also bacterial cell (Guivarc’h et al., 1996). These studies propose that RolA might possess multiple functions. From these evidence, we could hypothesize that RolA is a membrane-associated protein in plant cell, and it might be a transcriptional factor in bacterial cell. Combining these evidence, we proposed a possible role of RolA in plant. RolA is associated with plasma membrane in ground state, and it could be translocated into nucleus to regulate transcription process via unknown signaling stimulation. Nevertheless, RolA is still an functionally
known protein with the positive effects on rhizogenesis, development, and hormone homeostasis.
2.2. rolB
rolB encodes a protein of 259 amino acids in A. rhizogenes strain A4. It is discovered in all types of Ri plasmids. rolB is the most well-studied gene among the rol genes; however, rather conflicting results have been obtained. Recently, studies of rolB have been focused on its ability to induce secondary metabolites in plants. The way how rolB affects plant is illustrated in Figure 1-2, and the details are presented as follows.
Figure 1-2 The overview of rolB-mediated plant responses. rolB can be activated by auxin-dependent signals or auxin-independent transcript factors BBF1 and RBF1.
The RolB expression would inhibit the growth of R. cordifolia calli and induce tobacco leaf necrosis. The RolB can interact with tobacco 14-3-3 protein which resulting in nucleus translocation; however, abolishing the interaction by several point mutations only make little damage on the rooting. Besides, RolB can enhance the secondary metabolites and increase the auxin perceptions in transformants.
Increasing the auxin perceptions promotes plant de novo pluripotent meristem formation, resulting in rhizogenesis and other organogenesis.
2.2.1. rolB affects plant morphogenesis
rolB is the earliest research of interest among the rol genes. By deleting rolB from A. rhizogenes A4, the ability to induce hairy root on K. daigremontiana was abolished (White et al. 1985). Besides, rolB alone was capable of inducing tobacco rhizogenesis with almost the same efficiency as wild-type A. rhizogenes (Cardarelli et al. 1987; Spena et al. 1987). These reports indicated that rolB was the most important gene in rooting among the rol genes. Altamura and coworkers reported that rolB could strongly promote meristem formation by bypassing the regulatory factors of all types of organs (1994; 1998), and the results suggested that rolB regulates plant re- differentiation toward rooting.
rolB-transgenic tobacco presented smaller leaves with lower length-to-width ratio and highly branched plentiful roots (Cardarelli et al. 1987). Tobacco expressing rolB driven by Cauliflower mosaic virus (CaMV) 35S promoter showed bigger flower and early necrotic leaf (Schmülling et al. 1988). rolB-transgenic tomato showed reduced internode length and apical dominance with smaller flowers, lower pollen viability, and smaller fruits (van Altvirst et al. 1992; Arshad et al. 2014). On the other hand, rolB induces apical dominance in rose (van der Salm et al. 1997).
These indicate rolB functions in a species-dependent and tissue-dependent manner.
2.2.2. rolB and auxin
The phenotypic change in rolB-transgenic plants suggest that rolB is an-auxin responsive protein which mediates auxin signaling. Membrane potential measurement showed that rolB-transformed cells could increase auxin sensitivity up to 10000-fold (Maurel et al., 1991), and the polarization of auxin in rolB-expressing protoplast
could be blocked by a larger number of anti-auxin binding protein antibodies (Venis et al. 1992).
RolB was at first characterized as a glucosidase that hydrolyses indole glucosides in vitro, and rolB-transformed plant would increase auxin sensitivity by increasing IAA directly (Estruch et al. 1991c). However, two independent research groups invalidated the hypothesis later. Nilsson and coworkers demonstrated that wild-type and rolB-transformed tobacco showed the same contents of free IAA, and they had the same capacity of hydrolyzing IAA conjugates (1993a). Increasing auxin sensitivity in rolB-transformed tobacco was independent to intracellular auxin concentration because neither the accumulation nor the metabolism of endogenous auxin was affected; instead, rolB might increase the auxin perception (Delbarre et al.
1994; Maurel et al. 1994). This hypothesis was consistent with the experiment that the plasma membranes of rolB-transformed tobacco cells had additional auxin binding ability (Filippini et al. 1994).
There is other evidence supporting that rolB has a close connection with auxin.
Expressing rolB in tomato ovary by the tissue-specific promoter results in fruit parthenocarpy (Carmi et al. 2003), which has the similar effect with accumulating auxin in ovary by expressing bacterial auxin synthetic gene iaaH (indole acetamide hydrolase) driven by the same ovary-specific promoter along with treating its substrate (Szechtman et al. 1997). Expression of rolB in tobacco anther cells reduces stamen elongation and delays dehiscence (Cecchetti et al. 2004), which is considered as the result of lacking auxin polar transport system (Okada et al. 1991). In 1994, Altamura and coworkers demonstrated rolB promotes de novo primordia formation from tobacco thin cell layer in not only root but also flower (1994), which is
consistent with the reports that auxin controls the development of tobacco cells toward roots and flowers (Smulders et al. 1988; Smulders et al. 1990). In addition, auxin plays a crucial role in floral meristem formation and subsequent flower primordia formation (reviewed by Cheng and Zhao 2007). All in all, rolB has the auxin-like effects on plant fruit, ovary, and flower development, which supports the concept that rolB enhances the auxin perception in transformants.
Besides auxin, rolB was reported to have the correlation with cytokinin in promoting shoot formation from thin cell layers (Altamura et al. 1998). However, little connection between rolB and cytokinin has been proposed.
2.2.3. rolB promoter
rolB and rolC share a bidirectional promoter. Respective rolB and rolC native promoters drive uidA (β-D-glucuronidase) showed a similar expression pattern in shoot phloem but distinguishably in roots. rolB promoter activity shows mainly in the root primordia, including both primary and lateral primordia, and root cap, whereas rolC promoter activity does in phloem and in the apical meristems (Schmülling et al.
1989). The expression pattern of rolB indicates it has a close relation to cell differentiation and proliferation in the root. Overall, these two genes share a bidirectional promoter but they are regulated distinguishably.
rolB seems to be an auxin-regulated gene. In tobacco mesophyll protoplast, the expression level of rolB could be stimulated 20- to 100-fold by auxin treatment, whereas rolC expression increases only 5-fold (Maurel et al. 1990). In the same report, they discovered that treating exogenous auxin makes rolB express not only in root primordia but also in root vascular tissue and pericycle cells. Maurel and coworkers proposed the full activation of rolB by auxin is 12 to 18 hours after
treating, indicating that rolB belongs to an auxin late responsive gene. On the other hand, rolB could increase auxin perception within 8 hours after auxin treatment (1994). We can hypothesize the following two points. First, a low level of rolB expression is enough to increase auxin sensitivity in plant, and second, the activation of rolB might not only participate in amplifying auxin signals but also regulate other physiological behaviors independent to auxin signals.
There is much other evidence supporting that rolB is responsive to auxin but regulates not only auxin-related physiology. An auxin antagonist oligogalacturonide polymer is capable of inhibiting the rhizogenesis of rolB, and this effect disappears while rolB is driven by tetracycline-inducible promoter (Bellincampi et al. 1996).
Expressing rolB under control of its native promoter resulted in root or flower primordia formation, which is similar to treating exogenous auxin, whereas expression under CaMV 35S promoter in Hieracium piloselloides resulted in multi- potency (Koltunow et al. 2001). These phenomena showed the activation and the function of rolB have a close relationship with auxin.
Chimeric fusion of uidA with different lengths of upstream non-coding sequence of rolB shows that a 1185 bp length promoter region triggers the highest GUS activity (Capone et al. 1991; Capone et al. 1994). However, the 623 bp length promoter sequence drives a comparable activity. In addition, they identified five cis-elements, including regions -623 to -341, -341 to -306, -216 to -158, and the other two within regions about 70 and 80 bp around the CAAT and the TATA box, and they are so named as domain A-E, respectively. De Paolis and coworkers isolated a protein, which binds to the ACTTTA motif within domain B of rolB promoter via a single zinc finger structure. This protein was designated NtBBF1, representing N. tabacum rolB
domain B factor 1 (1996). NtBBF1 is essential for tissue-specific expression of rolB (Baumann et al. 1999). However, NtBBF1 is not an auxin-regulated gene, which indicates rolB is regulated at least by an unknown factor related to auxin and NtBBF1. In addition, another trans-acting element NtRBF1 (N. tabacum rol binding factor 1) can bind to -533 to -530 region of rolB promoter in non-meristem cells, and there is no differences between the concentrations of NtRBF1 in rolB-transformed and non-transformed tobacco plants (Filetici et al. 1997). Collectively, rolB is an auxin-inducible gene which increases auxin perception, but rolB can also be activated by an auxin-independent pathway and regulates auxin-independent responses in plants.
2.2.4. RolB biological function
Protein crude extract from RolB-expressing Escherichia coli has higher phosphatase activity than the extract from empty plasmid transformed E. coli, and the phosphatase is inhibited by tyrosine phosphatase inhibitor (Filippini et al. 1996).
Moriuchi and colleagues reported that RolB was a nucleus-localized protein that could interact with tobacco 14-3-3 κI, κII, ωI, ωII, ωIII, and ε (2004). In the same report, they generated a series of point mutations in RolB, and some of them could abolish the interaction; however, these point mutation reduce, but not abolish, root induction ability. The findings indicated that the physical interaction of RolB and 14-3-3 proteins are not essential for rhizogenesis.
2.3. rolC
In A. rhizogenes A4, rolC is a 543 bp gene which encodes a protein of approximate 20 kDa molecular mass. RolC is proved to be a cytosolic protein via ultra-centrifugation combining with specific antibody detection (Estruch et al.,
1991b). rolC plays an important role in promoting rooting, increasing branching, and stimulating secondary metabolites. rolC can strongly promote rooting and secondary metabolite production; therefore, it has been studied for a long time. However, until now, there is no conclusion about biological function of rolC. The followings summarize the studies about rolC.
2.3.1. rolC affects plant morphogenesis
rolC affects plant morphology in many aspects. rolC could induce root in tobacco (Spena et al. 1987; Schmülling et al. 1988), belladonna (Bonhomme et al.
2000a), carnation (Casanova et al. 2004), trifoliate orange (Kaneyoshi and Kobayashi 1999), and persimmon (Koshita et al. 2002). rolC induces not only roots but also adventitious shoot in carnation (Casanova et al. 2003; Casanova et al. 2004) and ginseng (Gorpenchenko et al. 2006). Besides, rolC induces somatic embryogenesis in ginseng (Gorpenchenko et al. 2006). In tobacco, rolC induces more abundant branch roots than rolA- or rolB-expressing hairy roots (Schmülling et al. 1988). Furthermore, rolC plays a role in hairy root elongation (White et al. 1985). These results suggest rolC exhibits cytokinin- and auxin-like activities, and it might induce the formation of pluripotent meristematic cells as rolB.
rolC-transgenic tobacco reduces apical dominance, and the plant appears dwarfism with shorter internodes, smaller leaves with narrow shapes, early flowering with smaller size, lower pollen viability and seed production (Schmülling et al. 1988;
Oono et al. 1990; Nilsson et al. 1993b; Scorza et al. 1994; Kaneyoshi and Kobayashi 1999; Koshita et al. 2002).
2.3.2. rolC and auxin
rolC was characterized as an auxin-related gene due to its ability to induce rooting and increase auxin sensitivity (Spena et al. 1987; Smulders et al. 1988;
Casanova et al. 2003). However, IAA content in the rolC-transformed plants showed no differences (Nilsson et al. 1993b; Schmülling et al. 1993; Casanova et al. 2004) or even decreased (Nilsson et al. 1996).
rolC has a cell-autonomous behavior. rolC induces rooting from the transformed cell, but the neighboring untransformed cells are not affected (Schmülling et al.
1988). This indicates RolC is neither a mobile nor a diffusible factor in rhizogenesis.
Besides, rolC transformation does not alter the growth habit of original tissues (Estruch et al. 1991b). In summary, rolC promotes de novo pluripotent meristem formation by its protein expression.
2.3.3. rolC and cytokinin
rolC reduces apical dominance and enhances lateral shoot development, and these were suggested to be the effects of cytokinin (Schmülling et al. 1988). Estruch and coworkers demonstrated that RolC expressed in E. coli had an in vitro β- glucosidase activity that releases the active free form cytokinin by cleaving glucosidic conjugates directly (Estruch et al. 1991a). However, the level of glucosidic conjugated cytokinin in vivo was not altered by expressing rolC, and free cytokinin levels in the plants were either the same or even lower (Nilsson et al. 1993b;
Schmülling et al. 1993; Faiss et al. 1996; Nilsson et al. 1996). Furthermore, the fraction of around 20 kDa isolated by size exclusive gel chromatography from rolC- transformed Panax ginseng extract showed no β-glucosidase activity (Bulgakov et al.
2002a). Transforming with rolC or ipt, a cytokinin biosynthetic gene encoded by
Agrobacterium, shows different phenotypes in root system and leaf color (Schmülling et al. 1988; Fladung 1990; Beinsberger et al. 1991; Schmülling et al. 1993; Faiss et al.
1996). Taken together, the phenotypic alternations caused by rolC are not directly related to cytokinins.
2.3.4. rolC and gibberellins
rolC-transgenic plant showed reduced size with shorter internodes, which are GA-reduced-like effects. Tobacco transformed with 35S-rolC showed less GA1 and GA3 but higher GA19, so rolC might decrease active form of GA by blocking GA biosynthetic pathway (Nilsson et al. 1993; Schmülling et al. 1993). However, exogenous GA3 application only restore the morphological change on internode length (Schmülling et al. 1993). We can conclude GA content alternation is one of the effects resulted from rolC expression.
2.3.5. rolC promoter
rolC promoter is identified to be specifically activated in companion cells (Nilsson et al. 1996). In rolC-transgenic tobacco, rolC expresses mainly in the vascular tissues of both root and stem (Spena et al. 1987; Schmülling et al. 1988;
Schmülling et al. 1989). The phloem-specific cis-acting element was found within -1 to -153 bp region of rolC promoter (Sugaya and Uchimiya 1992). However, rolC expression level in leaves was as high as that in roots while plant transformed with the entire T-DNA (Durandtardif et al. 1985; Leach and Aoyagi 1991). It hinted that rolC may be regulated by other genes localized on T-DNA.
The promoter region driving rolC during somatic embryogenesis is around -255 bp upstream of the transcriptional start site (Fujii and Uchimiya 1991; Fujii et al.
1994). It is shown that a sucrose-responsive cis-element locates between -135 and -94
bp of rolC promoter (Yokoyama et al. 1984). When sucrose is present in culture media, rolC could be activated in the whole transgenic plant (Nilsson et al. 1996).
However, sucrose is responsiveness and phloem-specific expression share the same cis-acting element, which indicates the two phenomena are linked. High concentration of sucrose usually present in phloem of roots and stems. Therefore, the fact that the rolC mainly expresses in these two tissues is reasonable.
2.4. rolD
rolD is only found in the agropine-type Ri plasmid TL-DNA region (Christey 2001). rolD-transgenic tobacco displayed early flowering and reduced rooting.
Moreover, rolD alone cannot induce rooting (Mauro et al. 1996). rolD expresses mainly in elongating and expanding tissues in adult plants with temporal regulation (Trovato et al. 1997).
rolD promoter, as well as rolB promoter, contains an auxin-responsive cis- element with a zinc finger binding element. rolD is also a late auxin-induced gene.
However, the rolB expression level increases with treatment of raised IAA concentration, whereas the rolD reaches the maximum level at approximate 1 μM of exogenous IAA and then decreases while increasing IAA concentration (Mauro et al.
2002).
In vitro enzyme reaction shows that RolD is an ornithine cyclodeaminase (OCD), which converts ornithine to proline. The result was consistent with the fact that higher proline content was detected in rolD-expressing flower (Trovato et al.
2001). There is no any other gene encoding OCD found in plants or in A. rhizogenes.
According to the research, the phenotypic change in rolD-transgenic plant might be
related to either higher proline accumulation or decreasing pool of ornithine, limiting the polyamine biosynthesis.
rolD is the only rol gene which has not been reported to affect plant secondary metabolites; however, rolD-transgenic tomato accumulated pathogen-related protein 1 (PR-1), whose expression frequently accompanies with secondary metabolites accumulation (Bettini et al. 2003). Besides, rolD-transformed tomato showed increased number of inflorescences and higher fruit yield. It is the only gene which seems to increase plant reproductivity among the rol genes.
2.5. orf3n
There are several genes other than rol genes on Agrobacterium rhizogenes A4 TL-DNA, and these genes are named orf1-18 (Slightom et al. 1986; Figure 2). In A.
rhizogenes strain HRI, the orf3 homologous gene is slightly larger than the gene on pRiA4, and it was designated orf3n (Lemcke and Schmülling 1998). Compared with wild-type, tobacco expressing 35S-orf3n showed shorter internode length, different leaf morphology with necrosis on leaf tip, delaying flowering, and lower density of inflorescences. Besides, orf3n repressed shoot formation from callus, which indicates that orf3n decreased the sensitivity to cytokinin. orf3n was therefore suggested to suppress the dedifferentiation of tissues, which may favor the hairy root formation and maintenance (Lemcke and Schmülling 1998).
2.6. orf8
orf8 is the longest ORF on TL-DNA. It encodes a protein of 780 amino acids. In addition to rol genes, orf8 is the most studied ORF on TL-DNA. The ORF8 N- terminal domain shows homology to the RolB, and the C-terminus shows a significant similarity to the iaaM-encoded protein (Levesque et al., 1988), which can
