Regulation of the Cell Expansion Gene RHD3 during Arabidopsis Development

12 Pages·2002·0.67 MB·English

_________________________________________________________________________________________________________
This article is published in Plant Physiology Online, Plant Physiology Preview Section, which publishes manuscripts accepted for
publication after they have been edited and the authors have corrected proofs, but before the final, complete issue is published
online. Early posting of articles reduces normal time to publication by several weeks.
_________________________________________________________________________________________________________
Regulation of the Cell Expansion Gene RHD3 during
1
Arabidopsis Development
Haiyang Wang2, Myeong Min Lee, and John W. Schiefelbein*
Department of Molecular, Cellular, and Developmental Biology, University of Michigan, Ann Arbor,
Michigan 48109–1048
The RHD3 (ROOT HAIR DEFECTIVE3) gene encodes a putative GTP-binding protein required for appropriate cell
enlargementinArabidopsis.ToobtaininsightintothemechanismsofRHD3regulation,weconductedamoleculargenetic
dissection of RHD3 gene expression and function. Gene fusion and complementation studies show that the RHD3 gene is
highlyexpressedthroughoutArabidopsisdevelopmentandiscontrolledbytwomajorregulatoryregions.Oneregulatory
region is located between 21,500 and 2600 bp upstream of the RHD3 gene and is required for vascular tissue expression.
Theotherregionisintragenicallylocatedandincludesthe558-bpfirstintron,whichisresponsibleforhigh-levelexpression
of RHD3 throughout the plant. The presence and location of this intron is essential for gene function because constructs
lacking this intron or constructs with the intron in an abnormal position are unable to functionally complement the rhd3
mutations.WealsoanalyzedtheroleofotherRHDgenesandtheplanthormonesauxinandethyleneinRHD3regulation,
andwedeterminedthattheseactdownstreamorindependentlyfromtheRHD3pathway.Thisstudyshowsthatmultiple
levels of regulation are employed to ensure the appropriate expression of RHD3 throughout Arabidopsis development.
The proper development of plant shape is at least opmental stages along the root (Dolan et al., 1993;
partly dependent on regulated cell expansion Schiefelbein et al., 1997).
(Steeves and Sussex, 1989). Although plant hor- In previous studies, several genetic loci involved
mones, cytoskeletal components, and wall-loosening in Arabidopsis root epidermal cell morphogenesis
proteins are known to influence this process (Evans, have been identified. For example, the RHD (ROOT
1984; Abeles et al., 1992; Cosgrove, 1997), the molec- HAIR DEFECTIVE) 1-4 genes (Schiefelbein and
ular mechanisms that regulate their function and the Somerville, 1990) and the COW1 gene (Grierson et
underlying genetic control of cell morphogenesis re- al., 1997) are required for normal root hair growth,
main largely unknown. and the RHD6 gene is required for root hair initia-
The Arabidopsis root epidermis has been ex- tion (Masucci and Schiefelbein, 1994). The RHD3
ploitedforstudyingthemolecularandcellularbasis gene is unique because it is required for regulated
ofplantcellmorphogenesisbecauseofitssimplicity cell expansion throughout root epidermis develop-
and accessibility (Schiefelbein et al., 1997; Paquette ment because root epidermal cells of rhd3 mutants
and Benfey, 2001). During the development of the are reduced in size and abnormal in shape. In addi-
root epidermis, the individual cells undergo a large tion, the RHD3 gene appears to act as a general
increase in cell size and they become highly elon- regulator of cell enlargement in Arabidopsis be-
gated, indicating that the orientation and extent of cause rhd3 mutations cause a reduction in cell size
cell expansion is regulated during cell differentia- throughout the plant and because RHD3 transcripts
tion. Epidermal cells are organized into columns accumulate in all major plant organs (Wang et al.,
(files) along the length of the root, with newly 1997; Galway et al., 1997). The RHD3 gene has been
formed cells located near the apex in the meristem- showntoencodeanovelproteinwithputativeGTP-
aticregionandthe“older”cellslocatedfartherfrom binding sites, and RHD3-like genes exist in diverse
the apex. Thus, each cell in a file is more develop- organisms, suggesting a common cellular function
mentally advanced than the one below it, which in eukaryotes (Wang et al., 1997).
provides an opportunity to easily study all devel- To gain insight into the molecular regulation of
RHD3, we analyzed RHD3 expression by in situ hy-
1ThisworkwassupportedbytheU.S.NationalScienceFoun- bridization, RHD3 promoter-GUS reporter gene fu-
dation (grant no. IBN–9316409) and by the U.S. Department of sions, and molecular complementation experiments.
Agriculture(grantnos.97–35304–4580and01–35304–10134). These studies document the critical role of a 59-
2Presentaddress:DepartmentofMolecularCellandDevelop- non-transcribed region and the large first intron in
mentalBiology,YaleUniversity,NewHaven,CT06520–8104.
regulating RHD3 gene expression and function. In
*Corresponding author; e-mail [email protected]; fax
addition, we analyzed the potential role of other
734–647–0884.
RHD genes and plant hormones in regulating the
Article,publicationdate,andcitationinformationcanbefound
atwww.plantphysiol.org/cgi/doi/10.1104/pp.002675. RHD3 gene.
PlantPhysiology,June2002,Vol.D1o2w9,nplopa.d1e–d1 2fr,owmw own .Jpalnaunatpryh y1s3i,o 2l.0o1rg9 -© P2u0b0li2shAemd ebryi cwawnwS.opclaienttyphoyfsPiolal.notrgB iologists 1of12
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
Wangetal.
RESULTS thestigmaticpapillaecellsandantherofflowers(Fig.
1, S–U). In addition, GUS expression was detected
RHD3 Is Expressed throughout the Arabidopsis Plant
throughout the developing embryo of these
To study the expression pattern of the RHD3 gene, RHD3::GUS lines (Fig. 1W). These results suggest
wefirstemployedwhole-mountinsituRNAhybrid- thattheRHD3geneisexpressedthroughouttheAra-
izationexperimentswithRHD3probesonArabidop- bidopsis plant and therefore may participate in a
sisseedlingroots.RHD3transcriptsweredetectedin fundamental cellular process.
the elongation zone of wild-type seedling roots (Fig.
1, A and B). This result is consistent with the postu- Identification of RHD3 Regulatory Regions
lated RHD3 role in regulating cell enlargement. The
rhd3-3 mutant, which bears a nonsense mutation, To study the regulation of RHD3, a series of dele-
tion constructs of the X1 RHD3::GUS fusion were
exhibits a greatly reduced transcript level in a
generated(Fig.2).TheH2constructisessentiallythe
northern-blot analysis (Wang et al., 1997) and very
same as X1 except that it lacks 100 bp at the 59 end
weak in situ RNA hybridization signal (Fig. 1C). No
(Fig. 2). Transgenic H2 plants exhibit the same GUS
signal was detected when the hybridization probe
expression pattern as construct X1 (data not shown),
was excluded (Fig. 1D).
suggesting that the distal 100-bp upstream sequence
To examine RHD3 expression in greater detail, we
constructed and analyzed a RHD3 promoter-b- is not required for regulating RHD3 expression.
The RHD3::GUS construct H1 is similar to H2 ex-
glucuronidase (GUS) reporter gene fusion. The con-
cept that it is a transcriptional fusion and lacks the
struction of this fusion was based on results from
entire558-bpfirstintronofRHD3aswellas73bpof
studies to define the DNA sequences necessary for
thefirstexonthatflanksthisfirstintron(Fig.2).GUS
RHD3genefunction.Inpreviouslypublishedwork,a
expression in H1 transgenic plants was found to be
construct shown in Figure 2 (designated 12A) con-
taining 1.6 kb of 59-non-coding sequence, the first restrictedtothevasculartissuesoftheroots(Fig.1,I
and J), and discontinuous weak expression in the
exon, and the first intron of the RHD3 gene fused to
vascular tissue of the cotyledon (Fig. 1L), leaf (Fig.
therestofRHD3cDNAwasfoundtorescuetherhd3
1N), stem (Fig. 1P), pedicel (Fig. 1R), and the flower
mutant root hair phenotype in transformed green
sepals (Fig. 1V), and no detectable expression in the
callustissue(Wangetal.,1997).Inthepresentstudy,
developing embryo (Fig. 1X). In addition, quantita-
we extended these findings by generating stable
tive GUS assays showed that the average GUS ac-
transgenic plants that contain the 12A construct in
tivity in the seedlings roots from the H1 lines is less
the rhd3-2 (Columbia background) or rhd3-3 allele
than 1% of the activity present in the H2 lines
(Nossen background). Like the transformed callus (29,300versus210pmolmethylumbelliferonemin21
tissue, these transgenic plants produce root hairs mg protein21). These results indicate that the DNA
withnormallengthandmorphology(Fig.3,A–C).In
segment containing the proximal 73-bp first exon
addition, these plants possess leaves that are compa-
sequence and the first intron is required for the
rable in size to the wild type (Fig. 3, D–F). Further-
appropriate qualitative and quantitative expression
more, their root epidermal cell length and root
of the RHD3 gene.
growth rate were indistinguishable from the wild
Two additional constructs, B2 and B4, were gener-
type (Table I). Together, these results show that the ated that contain approximately 600 bp of the 59
DNAsequencesfusedtotheRHD3cDNAinthe12A
non-transcribedsequenceplus90bpofthedistalfirst
construct are sufficient to rescue the rhd3 mutant
exon sequence transcriptionally fused to the GUS
defects.
reportergene(Fig.2).TheB2constructhasthisRHD3
Based on these results, we used the same genomic DNA segment in the correct orientation, and B4 has
RHD3 DNA sequences (1.6 kb of 59-non-coding se- this in the reverse orientation (as a negative control),
quence,thefirstexon,andthefirstintron)togenerate with respect to the GUS gene-coding region (Fig. 2).
a translational fusion to the GUS reporter gene (des- No GUS staining was observed in transgenic plants
ignated construct X1; Fig. 2). Twelve independent harboringeitheroftheseconstructs(datanotshown),
transgenic lines bearing this construct were obtained suggesting that these sequences are not sufficient to
and analyzed, and the pattern of GUS accumulation drive gene expression at a detectable level. Compar-
was similar in all lines. These RHD3 regulatory se- ing the expression patterns of constructs B2 and H1
quences direct high levels of GUS expression leads to the conclusion that sequences located be-
throughout the root, including the root elongation tween 2600 bp and 21,500 bp are responsible for
zone (Fig. 1, E–H). When the transgenic X1 plants vascular tissue expression of RHD3.
were incubated in GUS substrates for an extended
periodoftime(8–16h),stainedcellsweredetectedin
The First Intron Is Required to Confer Appropriate
all plant tissues. For example, GUS expression was
RHD3 Expression
observed in the cotyledons (Fig. 1K) and hypocotyl
(datanotshown)ofseedlings,theleaf(Fig.1M),stem Thedramaticallydifferentexpressionpatternscon-
(Fig.1O),andpedicel(Fig.1Q)ofmatureplants,and ferred by constructs H2 and H1 indicate that the
2of12 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org Plant Physiol. Vol. 129, 2002
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
RegulationoftheRHD3Gene
Figure 1. Expression of the Arabidopsis RHD3 gene. A through D, In situ RNA hybridization; E through X, RHD3
promoter::GUSreportergeneexpression.AandB,Four-day-oldwild-typeroothybridizedwithRHD3antisenseRNAprobe
showsRHD3transcriptaccumulationintherootelongationzone.BarinA5100mm.BarinB550mm.C,Four-day-old
rhd3-3mutantroothybridizedwithRHD3antisenseRNAprobeshowsreducedRHD3transcriptaccumulation.Bar5100
mm.D,Four-day-oldwild-typerootincubatedwithoutRNAprobeshowsnohybridizationsignal.Bar5100mm.EandF,
Four-day-old root from transgenic line harboring the X1RHD3::GUSconstruct incubated for 30 min in X-Gluc substrate
showsexpressionintheelongationzone.BarinE5100mm.BarinF550mm.GandH,Four-day-oldrootapex(G)and
maturerootsegment(H)fromtransgeniclineharboringtheX1RHD3::GUSconstructincubatedfor4hinX-Glucsubstrate
shows expression throughout the root. Bars 5 100 mm. I and J, Four-day-old root elongation zone (I) and mature root
segment (J) from transgenic line harboring the H1 RHD3::GUSconstruct incubated for 16 h in X-Gluc substrate shows
preferentialexpressioninthevasculartissue.BarinI550mm.BarinJ5100mm.KandL,Six-day-oldseedlingcotyledons
fromtransgeniclinesharboringtheX1(K)orH1(L)RHD3::GUSconstructsincubatedfor16hinX-Glucsubstrateshows
expressionthroughoutthecotyledon(K)orinvasculartissue(L).Bars5500mm.MandN,Three-week-oldrosetteleaves
fromtransgeniclinesharboringtheX1(M)orH1(N)RHD3::GUSconstructsincubatedfor16hinX-Glucsubstrateshows
expressionthroughouttheleaf(M)orinvasculartissue(N).Bars5100mm.OandP,Five-week-oldstemsfromtransgenic
lines harboring the X1 (O) or H1 (P) RHD3::GUSconstructs incubated for 16 h in X-Gluc substrate shows expression
throughoutthestem(O)orinvasculartissue(P).BarinO5200mm.BarinP5100mm.QandR,Five-week-oldpedicels
fromtransgeniclinesharboringtheX1(Q)orH1(R)RHD3::GUSconstructsincubatedfor16hinX-Glucsubstrateshows
expressionthroughoutthepedicel(Q)orinparticularcellfiles(R).BarinQ5200mm.BarinR5100mm.SthroughU,
Five-week-oldtransgeniclinesharboringtheX1RHD3::GUSconstructsshowinganimmatureflower(S),matureflower(T),
orstigmatictissuefrommatureflower(U)incubatedfor16hinX-Glucsubstrate.BarsinSandT5800mm.BarinU5
200mm.V,Five-week-oldsepaltissuefromflowersoftransgeniclineharboringtheH1RHD3::GUSconstructincubatedfor
16hinX-Glucsubstrateshowsexpressioninvasculartissue.Bars5100mm.WandX,Embryosatcurledcotyledonstage
fromtransgeniclinesharboringtheX1(W)orH1(X)RHD3::GUSconstructsincubatedfor16hinX-Glucsubstrateshows
expressionthroughouttheembryo(W)orabsent(X).Bars5100mm.
Plant Physiol. Vol. 129, 2002 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org 3of12
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
Wangetal.
influence gene transcription, these constructs would
beexpectedtoconferahighlevelofGUSexpression.
However, transgenic lines bearing these constructs
failed to exhibit a detectable difference in GUS ex-
pression (assessed by histochemical staining) com-
pared with lines with the intron-less control con-
structs (data not shown).
One possible explanation for the results above is
that the first RHD3 intron regulates expression only
whenitislocatedwithinthetranscribedportionofa
gene. To test this, GUS constructs were generated
withthefirstintron(initscorrectorientation)within
thetranscribednon-codingregiondownstreamofthe
1.5-kb native RHD3 promoter and the 35S constitu-
Figure2. StructureofRHD3genefusions.The12Aconstructrepre- tive promoter (constructs H1Ic and 35Sic, respec-
sentsanin-framefusionofthe59endoftheRHD3gene(including tively;Fig.4).Surprisingly,transgenicplantsbearing
first intron) to the RHD3cDNA. The other constructs shown are these constructs actually exhibited a reduction in
RHD3::GUSreportergenefusionsthatinclude(X1andH2)ordonot GUS expression, compared with plants containing
include(H1,B2,andB4)thefirstRHD3intron.DetailsoftheDNA
the intron-less control constructs (Fig. 5, B–D). Re-
constructionsaregivenin“MaterialsandMethods.”
verse transcriptase-PCR experiments indicated that
the intron was spliced properly from the H1Ic tran-
631-bpintragenicsequenceplaysanimportantrolein script (data not shown). Together, comparing the
regulating RHD3 expression. To determine whether effects of the H1Ic and H2 constructs, the first RHD3
it is the 73-bp proximal first exon segment or the intron apparently affects gene expression only when
558-bp intron sequence that possesses a regulatory located at its normal position.
function, a construct (designated H2-I) was gener-
atedthatisidenticaltoH2exceptthatitlacksthefirst
intron (Fig. 4). The GUS gene expression pattern in
RHD3 Gene Function Is Dependent on the
H2-I transgenic plants is similar to H1 (Fig. 5A; data
Presence of the First Intron
not shown), indicating that the 73-bp proximal first
exon sequence does not have a major effect on the Although the above results show that the first in-
reporter gene expression; thus, the RHD3 expression tron is required for appropriate RHD3 gene expres-
pattern is dependent on the first intron sequence. sion,thesignificanceofthisintronforRHD3function
The effect of the intron sequence was explored by wasnotclear.Asdescribed,the12ARHD3construct,
conducting RNA-blot (northern) analysis with the which contains the RHD3 promoter and first intron
H2, H2-I, and H1 lines. Using a GUS-specific probe, fused to the RHD3 cDNA, can fully complement the
we discovered an abundant RNA of the appropriate rhd3 mutant phenotype (Fig. 3; Table I). A construct
sizeinseedlingRNAfromtheH2,butnohybridizing was generated (designated 12A-I) that contains the
band was detected in the H2-I or H1 lines (Fig. 6). sameupstreamregulatorysequencesasthe12Acon-
This indicates that the RHD3 intron is required for struct fused to the RHD3 cDNA, except that the first
proper gene transcription or RNA stability. intron was excluded (Fig. 7A). This construct was
To test the possibility that the first intron may introduced into the rhd3-2 and rhd3-3 mutant lines,
possess a transcriptional enhancer-like element, a and the resulting transgenic plants exhibited either
GUS reporter fusion was constructed with the first the rhd3 mutant phenotype or a phenotype interme-
intronsequenceplacedupstreamofthe1.5-kbRHD3 diate between the rhd3 mutant and wild type (Fig. 7,
native promoter (construct H1Ia; Fig. 4) and up- B–G; Table I). The inability of this construct to com-
stream of the 0.6-kb promoter fragment in the B2 plementtherhd3mutantshowsthatthefirstintronis
construct (construct B2I; Fig. 4). In addition, the required for RHD3 gene function.
RHD3 first intron was placed upstream of the heter- To test whether increased transcription of the
ologous cauliflower mosaic virus 35S minimal pro- RHD3 gene could substitute for the presence of the
moter(containing46-bppromotersequencesandun- first intron, a construct was generated (35S::RHD3,
abletoinducedetectableexpressionbyitself)andthe Fig. 4) in which the RHD3 cDNA was placed under
cauliflower mosaic virus 35S constitutive promoter control of the constitutive 35S promoter. After trans-
fused to the GUS reporter (constructs 35SMI and formationofrhd3-3mutantplants,thistransgenedid
35Sia,respectively;Fig.4).Furthermore,GUSfusions not rescue the rhd3 mutant phenotype (Fig. 5E), al-
were made with the intron placed in an inverse ori- thoughapartiallycomplementedphenotypewasob-
entation upstream of the 1.5-kb RHD3 native pro- served in three of 14 lines (Fig. 5F). The results from
moter and the 35S constitutive promoter (constructs a northern-blot analysis showed that the failed
H1Ib and 35Sib, respectively; Fig. 4). If the RHD3 complementation was not due to transgene inactiva-
intron sequence acts as an enhancer-like element to tion (e.g. cosuppression); in fact, the overall level of
4of12 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org Plant Physiol. Vol. 129, 2002
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
RegulationoftheRHD3Gene
Figure3. Complementationoftherhd3mutant
phenotypebythe12ARHD3gene::cDNAcon-
struct.AthroughC,Four-day-oldseedlingroots.
Bars 5 300 mm. D through F, Rosettes of
4-week-oldplants.Bar52cm.AandD,Wild
type (Columbia ecotype). B and E, Transgenic
rhd3-2harboring the 12A construct. C and F,
rhd3-2mutant.
RHD3 RNA is much greater in the 35S-RHD3 lines struct (Fig. 5C). Taken together, these data indicate
than in wild-type plants (Fig. 8). This indicates that that the first intron is necessary for appropriate
the high level of gene transcription directed by the RHD3 expression and function, and it does not act
35S promoter is not sufficient to ensure RHD3 gene properly when placed in abnormal locations.
function.
In a related experiment, we sought to determine
Root Hair Development Genes and Plant Hormones Do
whether the 35S promoter might complement the
Not Affect RHD3 Gene Expression
rhd3 mutant if the first intron is included in the
transcribed(non-translated)regionintheproperori- Toattempttodefinefactorsthatregulatetheactiv-
entation (construct 35SI::RHD3, Fig. 4). However, ity of the RHD3 promoter and/or first intron, we
when introduced into rhd3 mutant plants, this trans- examined the possibility that RHD3 expression may
gene did not fully complement the mutant pheno- beregulatedbyitsownproteinorbyotherRHDgene
type (Fig. 5, G and H), although reverse transcrip- products.Thus,thefull-lengthRHD3promoter::GUS
tase-PCR analysis indicated that the intron was reporter gene constructs (X1 and H2) were intro-
spliced properly from the 35SI::RHD3 transcript duced into the rhd3 mutant or into the rhd1, rhd2,
(data not shown). This result is consistent with the rhd4,andrhd6mutants.GUSactivityineachofthese
reporter gene expression studies with the 35SIc con- lines was found to be indistinguishable from the
TableI. EffectofRHD3geneconstructsonrootgrowthandcellelongation
RootGrowthRate EpidermisCellLength
Genotype
mmd21a %Wildtype mmcell21a %Wildtype
Columbiab 5.360.8 100 207636 100
rhd3-2(Col) 2.660.3 48 92613 45
rhd3-2,12A 5.260.8 98 222636 107
rhd3-2,12A-I NDc – 94615 45
Nossenb 5.760.5 100 166629 100
rhd3-3(No) 2.860.4 48 85612 51
rhd3-3,12A 5.660.5 98 184633 111
rhd3-3,12A-I ND – 89611 54
aValuesgivenrepresentthemean6SD. bWildtype. cND,Notdetermined.
Plant Physiol. Vol. 129, 2002 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org 5of12
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
Wangetal.
Figure4. StructureofRHD3genefusionscon-
tainingorlackingthefirstintron.TheH2-Icon-
struct contains the same upstream regulatory
sequences as H2 except that the first intron is
absent. For other constructs, a PCR fragment
containingthefirstintronwasaddedtotheup-
stream or in the transcribed regions of various
GUSreporterorRHD3cDNAgenefusions.The
orientationoftheintronsequencesisindicated
by the arrows. The 11 symbol indicates the
transcription initiation site. Details of the clon-
ingproceduresaredescribedin“Materialsand
Methods.”
wild-typeRHD3::GUSplants,basedonhistochemical toethylenebiosynthesis),or5mmaminoethoxyvinyl-
analysisinroots(datanotshown).Thisindicatesthat Gly (an ethylene biosynthesis inhibitor), and then
these RHD genes do not regulate the activity of the tested for GUS activity by histochemical staining.
RHD3 promoter or first intron. In addition, double- Although these hormone and inhibitor treatments
mutant combinations between rhd3 and rhd1 (Schief- significantly altered root morphogenesis, they did
elbein and Somerville, 1990), rhd3 and rhd4 (Schiefel- not cause a detectable change in the RHD3::GUS
beinandSomerville,1990),andrhd3andrhd6(Fig.9, expression (data not shown).
A–D) display an additive phenotype, which further InasecondtestoftheeffectofhormonesonRHD3
supports the notion that RHD3 acts in a pathway gene expression, RHD3::GUS activity was examined
separate from these genes during root hair in various hormone-related mutant backgrounds.
development. Thesemutantsincludedtheaux1(Pickettetal.,1990),
To determine whether the plant hormones auxin axr2 (Wilson et al., 1990), eto1 (Guzman and Ecker,
and ethylene may affect RHD3 expression, three sets 1990), ein2 (Guzman and Ecker, 1990), and ctr1
of experiments were conducted. First, transgenic (Kieberetal.,1993).AsshowninFigure9,Ethrough
seedlings bearing the full-length RHD3::GUS con- I, each of these mutations affects the development of
structs (X1 and H2) were transferred from standard the root and/or root hairs of Arabidopsis. Despite
growth media to media supplemented with 30 and these effects, none of these mutations caused an ob-
300 nm indole-3-acetic acid (an auxin), 5 and 50 mm servable effect on the expression of the RHD3::GUS
1-aminocyclopropane-1-carboxylic acid (a precursor X1orH2constructs(datanotshown),indicatingthat
6of12 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org Plant Physiol. Vol. 129, 2002
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
RegulationoftheRHD3Gene
Figure 5. Effects of the RHD3 first intron on
gene expression. A through D, Five-day-old
seedlingrootsharboringtheH2-I(A),pBI121(B)
35SIc (C), or H1Ic (D) construct and incubated
withX-glucfor16h.Bars5100mm.Ethrough
H, Root phenotype of 5-day-old rhd3-3seed-
lingsharboringthe35S::RHD3(EandF)orthe
35SI::RHD3(G and H) constructs. Bars 5 200
mm. E and G, Mutant phenotypes. F and H,
Partiallyrescuedphenotypes.
thesehormone-relatedgenesdonotdirectlyregulate functional importance of this intron is demonstrated
the RHD3 promoter or first intron activity. by complementation analyses in which a construct
In a third test of the relationship between RHD3 containing the first intron (12A) is able to fully com-
and plant hormones, double mutants were con- plementtherhd3mutant,whereasaconstructlacking
structed and analyzed that contained the rhd3 and the first intron (12A-I) cannot (Fig. 7). Thus, the first
either the eto1, ctr1 , oraxr2 mutations. Because each intron is required for the high-level expression and
of these mutants exhibit a cell size defect in the root appropriate function of RHD3.
epidermis (Table II), we were able to use this char- This research has also yielded clues regarding the
acter to determine the genetic relationships between mechanism responsible for the effect of the RHD3
the rhd3 and these mutations. As shown in Table II first intron. First, gene fusion experiments show that
andFigure9,JthroughL,therhd3ctr1,rhd3eto1,and theintroncannotsignificantlyincreasereportergene
rhd3 axr2 double mutants each display a phenotype expression when placed upstream of the RHD3 na-
indicative of an additive interaction of the single tive promoter or upstream of the heterologous 35S
mutations, consistent with the possibility that the
genes act in separate pathways.
DISCUSSION
We have conducted a detailed study of the expres-
sion and regulation of the Arabidopsis RHD3 gene,
which encodes a putative GTP-binding protein re-
quiredfornormalcellenlargementthroughoutplant
development.Oneofthemajorfindingsisthecritical
role of the first intron of RHD3 for appropriate gene
expression and function. This intron is notable be-
cause of its relatively large size (558 bp) and its
location adjacent to the translation start codon (in
codon no. 2). These features initially suggested a
possible regulatory function for the first intron se-
quences. Comparison of reporter gene expression
patterns between constructs containing this region
(H2)andlackingthisregion(H1)showsaqualitative
Figure 6. RNA-blot analysis ofGUSreporter expression. RNA was
andquantitativedifferenceintheexpressionpattern.
isolatedfrom7-d-oldseedlingsfromtheH1,H2,andH2-Ilines,and
In the absence of the intron, gene expression is re-
the blot was probed with a GUS-specific DNA fragment. Approxi-
duced by two orders of magnitude (as determined mately 20 mg of total RNA was loaded per lane. The blot was
both by RNA-blot hybridization and reporter en- subsequently hybridized with an 18S rRNA probe as a loading
zymeactivity)andisrestrictedtovasculartissue.The control.
Plant Physiol. Vol. 129, 2002 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org 7of12
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
Wangetal.
Figure 7. The first intron of theRHD3gene is
required for complete complementation of the
rhd3 mutant phenotype. A, RHD3 construct
mapsshowingthe12Aconstruct(includingthe
firstRHD3intron)andthe12A-Iconstruct(lack-
ingthefirstintron).BthroughG,Phenotypeof
4-day-oldseedlingrootsandroothairs.Bars5
300 mm. B, Wild type (no ecotype). C,rhd3-3
mutant. D, Transgenicrhd3-3seedling harbor-
ingthe12Aconstruct.EthroughG,Independent
transgenicrhd3-3seedlingsharboringthe12A-I
construct showing the variation in the partial
complementationphenotypes.
promoter, whether in the correct or the reverse ori- An interesting possibility raised by our results is
entation. This shows that the RHD3 first intron does that the position or context of the RHD3 intron is
notexhibitclassictranscriptionalenhancer-likeprop- critical for its ability to direct appropriate gene ex-
erties. Second, the RHD3 first intron is unable to pression. Although several plant introns have been
direct a high level of reporter gene expression or shown previously to influence gene expression (Cal-
35S::RHD3 cDNA expression when placed in the 59- lis et al., 1987; Dean et al., 1989; Vasil et al., 1989; Fu
transcribed (non-coding) region, which implies that et al., 1995; Bolle et al., 1996; Rose and Last, 1997),
its normal location within the translated region is only a few studies provide evidence for the impor-
criticalforitseffectonRHD3geneexpression.Third, tance of the intron position or context (Callis et al.,
the 35S promoter is unable to drive appropriate ex- 1987; Luehrsen and Walbot, 1991; Norris et al., 1993;
pression of the RHD3 cDNA to enable complete Jeon et al., 2000), and these did not assess the func-
complementationoftherhd3mutant,despitethefact tional importance of the intron position by conduct-
thatthe35S::RHD3rhd3plantsaccumulateagreater ingmutantcomplementationassaysasinthepresent
amount of RHD3 RNA than do wild-type plants study. However, the position of an intron within a
(Fig. 8). This result implies that the RHD3 first in- Xenopus laevis pre-mRNA transcribed in vivo has
tron does not simply act by increasing the RHD3 been shown to determine the translational efficiency
transcript level; rather, it likely plays a regulatory ofthematuremRNAinthecytoplasm(Matsumotoet
role in RHD3 mRNA metabolism. Taken together, al., 1998). Also, studies from several species indicate
our results suggest that the RHD3 first intron acts a connection between mRNA maturation (including
posttranscriptionally to regulate the production of 59 capping, 39 polyadenylation, and export to the
mature RHD3 mRNA. cytosol) and splicing (Bentley, 1999). Together, these
8of12 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org Plant Physiol. Vol. 129, 2002
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
RegulationoftheRHD3Gene
tissues. These results are consistent with our previ-
ous data from a northern-blot analysis, which
showedthatRHD3transcriptsaccumulateinallplant
organs (Wang et al., 1997). Because rhd3 mutants do
not exhibit a lethal phenotype (even rhd3 mutants
bearing a nonsense mutation, like the rhd3-3 allele,
are viable), the RHD3-associated cellular process is
not essential or it is partially carried out by a redun-
dant pathway. One possibility, suggested in earlier
studies(Galwayetal.,1997;Wangetal.,1997),isthat
RHD3 may control vacuole biogenesis, which is crit-
ical for cell enlargement. Further studies to deter-
mine the cellular location of the RHD3 product and
RHD3-interacting proteins are likely to uncover new
insights into the molecular basis of RHD3 function
and plant cell morphogenesis.
Figure8. RNA-blotanalysisofRHD3expression.RNAwasisolated
from7-d-oldseedlingsfromtwowild-typelines(Rschew[RLD]and MATERIALS AND METHODS
Columbia), fourrhd3mutants (rhd3-1,rhd3-2rhd3-4, andrhd3-3),
Plant Materials and Growth Conditions
andtwoindependenttransgeniclinesbearingthe35S-RHD3cDNA
constructintherhd3-3mutantbackground(35S-Aand35S-B).Ap-
The rhd3 mutant alleles were described previously
proximately20mgoftotalRNAwasloadedperlane;the18SrRNA
(SchiefelbeinandSomerville,1990;Wangetal.,1997).The
probeusedasaloadingcontrolisshown.
source of the hormone-related mutants (aux1-7, axr1-3,
axr2, eto1-1, ein2-1, and ctr1-2) was previously given (Ma-
succiandSchiefelbein,1996).Doublemutants(rhd3-1axr2,
suggest that the particular pathway of mRNA matu-
rhd3-1eto1-1,andrhd3-1ctr1-2)wereconstructedbycross-
ration,whichisinfluencedbyintronposition,maybe
ing single mutant plants, examining the F progeny for
a key determinant of gene expression. Further stud- 2
putative double-mutant phenotypes, and confirming the
iesofRNAsplicingandaccumulationwillbeneeded
double mutants in subsequent generations.
to precisely define the mechanism of action of the
Unlessotherwisedescribed,Arabidopsisseedsweresur-
RHD3 first intron.
facesterilizedandgrownonagarose-solidifiedmediumin
In addition to the first intron, another important
vertically oriented petri plates as described previously
RHD3 regulatory sequence was identified using the
(Schiefelbein and Somerville, 1990). Media containing
reporter gene fusions. This region is located in the
upstream sequences between 21,500 and 2600 bp indole-3-acetic acid (Sigma, St. Louis), 1-aminocyclo-
propane-1-carboxylic acid (Sigma), or aminoethoxyvinyl-
and is required to confer gene expression in the
Gly (MAAG Agrochemicals, Vero Beach, FL) were
vascular tissue of various Arabidopsis organs. Pro-
prepared as described (Masucci and Schiefelbein, 1994;
moter constructs that lack this region and the intron
Wang et al., 1997).
areunabletodirectdetectablelevelsofreportergene
expression.
Microscopy and Expression Assays
A second goal of the present study was to define
therelationshipbetweentheRHD3geneproductand
Root growth rate and epidermal cell length were deter-
otherpossibleregulatorsofcellexpansion.Usingthe
mined as previously described (Wang et al., 1997) from at
RHD3::GUS reporter gene and double-mutant tests,
least 10 independent lines. The histochemical staining of
we have demonstrated that other root hair develop-
plants containing the GUS reporter gene was performed
ment genes (RHD1, RHD2, RHD4, and RHD6), the
essentially as described (Masucci et al., 1996). The
plant hormones auxin and ethylene, and hormone-
RHD3::GUS (X1 and H2) were introduced into mutant
related genes (AUX1, AXR2, ETO1, EIN2, and CTR1)
backgroundsbygeneticcrosses,andhistochemicalstaining
do not significantly affect RHD3 gene expression.
wasconductedwithF seedlingsexhibitingthemutantand
2
These data suggest that RHD3 functions in an inde-
wild-type phenotypes. Quantitative assay of the GUS ac-
pendent pathway from these loci and hormones in
tivity was performed essentially as described by Jefferson
the control of plant cell morphogenesis.
etal.(1987).Whole-mountinsituhybridizationwithyoung
ThisRHD3expressionstudysupportstheviewthat
Arabidopsis seedlings was performed as previously de-
the RHD3 protein is involved in a fundamental cel-
scribed(Masuccietal.,1996).RNA-blot(northern)analyses
lular process that is required for regulated cell en-
were conducted as described (Wang et al., 1997).
largement to occur during plant development.
Whole-mount in situ hybridization and RHD3
Constructs and Plant Transformations
promoter::GUS reporter gene fusion studies show
thattheRHD3geneishighlyexpressedinmostplant The 12A construct was described by Wang et al. (1997).
tissues,includingtheembryo,flower,shoot,androot Forthe12A-Iconstruction,theBamHI-SacI(fillingtheSacI
Plant Physiol. Vol. 129, 2002 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org 9of12
Copyright © 2002 American Society of Plant Biologists. All rights reserved.
Wangetal.
Figure9. Rootphenotypesof5-d-oldseedlings
fromrhd3mutantanddoublemutants.A,Wild
type (Columbia ecotype). B,rhd3-1mutant. C,
rhd6-1mutant. D, rhd3-1rhd6-1double mu-
tant. E, aux1-7 mutant. F, axr2-1 mutant. G,
eto1-1mutant. H,ctr1-2mutant. I,ein2-1mu-
tant.J,rhd3-1axr2-1doublemutant.K,rhd3-1
eto1-1double mutant. L,rhd3-1ctr1-2double
mutant.Bars5400mm.
59-protruding end) fragment containing the GUS gene- BamHI site of the pBI101.2 vector to generate transcrip-
codingregioninthepBI101.2vectorwasreplacedwiththe tional fusions, with the B2 construct having the promoter
BamHI-EcoRV fragment from the C3I clone of the RHD3 sequences in the correct orientation and B4 in the reverse
cDNA. Next, an XbaI-XhoI fragment containing the puta- orientation.
tive RHD3 promoter (1.6-kb 59 non-coding sequences and For construction of the H2-I clone, the T3 primer and
80-bp59-untranslatedsequences)wassubcloned(byfilling another primer (59 GTGGATCCCCCATTATCGCTCAAC-
theXhoI39-protrudingend)fromthegenomicclonepR5.2 GAAACGC 39) were used to amplify the RHD3 upstream
(Wangetal.,1997)intotheXbaIandSmaIsitesintheabove sequencesandtheentirefirstexonusingthegenomicclone
modified pBI101.2 vector. pR5.2 as the template. The PCR products were subcloned
FortheX1construct,aXbaI-ClaIfragmentcontainingthe into the HindIII-BamHI sites of the pBI101.2 vector as a
1.6-kb 59-non-coding sequences, the first exon, the first HindIII-BamHIfragmenttogenerateatranslationalfusion.
intron, and three nucleotides of the second exon was sub- Two other primers (59 TGGAAGCTTGTAATTTATATA-
cloned (by filling in the Cla I 39-protruding end) into the TCCATCTCTTC 39 and 59 CCTTCGAACTGCAACGTCC-
XbaI and SmaI sites of the pBI101.2 vector to generate a
translational fusion. The H2 construct was generated the
TableII. Effectofmutationsonrootepidermiscelllength
same way as X1 except a HindIII-ClaI promoter fragment
wassubcloned,whichlacks100bpfromthenon-coding59 Genotype CellLengtha
end of the XbaI-ClaI fragment. mmcell21
The H1 construct was generated by subcloning the Columbia 207636
HindIII-XhoIfragmentcontainingthe1.5-kb59-non-coding rhd3-1 86619
sequencesplusabout80bpof59-untranslatedsequencesof eto1–1 5967
thefirstexonfromtheplasmidpR5.2intotheHindIIIand rhd3-1eto1–1 3264
SalI sites of the pBI101.2 vector to generate transcriptional ctr1–2 3766
rhd3-1ctr1–2 2665
fusions.FortheB2andB4constructs,a690-bpBamHI-BglII
fragment containing about 600 bp of 59-non-coding se- axr2 169623
rhd3-1axr2 6468
quencesand90bpof59-untranslatedsequencesofthefirst
exon from the plasmid pR5.2 was subcloned into the aMean6SDforroothair-bearingepidermalcells.
10of12 Downloaded from on January 13, 2019 - Published by www.plantphysiol.org Plant Physiol. Vol. 129, 2002
Copyright © 2002 American Society of Plant Biologists. All rights reserved.