Plant Molecular Biology37: 385–389, 1998. 385c 1998Kluwer Academic Publishers. Printed in Belgium.

Short communication

Identification of new early auxin markers in tobacco by mRNA differentialdisplay

Camille Roux, Jurg Bilang, Benjamin H. Theunissen and Catherine Perrot-Rechenmann�

Institut des Sciences Vegetales, CNRS, Batiment 23, Avenue de la Terrasse, F-91198 Gif sur Yvette, France(�author for correspondence)

Received 16 July 1997; accepted in revised form 28 December 1997

Key words:auxin, differential display, tobacco seedlings, auxin-regulated gene expression

Abstract

Differential display of mRNA has been improved by developing a two-step PCR amplification procedure. Themodified differential display has been applied to identify early alterations of mRNA expression in response toauxin treatment of tobacco seedlings. This approach has led to the isolation of four fragments corresponding toauxin-up-regulated mRNAs. One, namedGO15-13, shows significant homology with the 30 end of the codingregion of the soybeanSAUR X10A[10]. The three other fragments present no homology with sequences availablein the databases and constitute potential new early auxin markers.

The plant growth factor auxin is implicated in manyphysiological processes controlling plant growth anddevelopment [3]. The molecular mechanisms of auxinaction remain largely unknown. One of the strategiesto investigate auxin signal transduction pathways is thestudy of early auxin-responsive genes. Using elong-ating tissues or dividing cells as starting materials,several families of auxin-responsive genes have beenidentified by differential screening [1, 4, 11]. Someare rapidly and selectively induced by auxin, amongwhich are theSAURandGH3 genes, and members oftheIAA/AUXgene family which respond specifically toauxin stimulation within less than 30 min. To diversifythe early molecular markers of auxin action in tobacco,we have looked for changes in gene expression inducedby auxin using mRNA differential display. This meth-od, first described by Liang and Pardee [8], seems tobe currently the most sensitive and powerful methodto detect variations in mRNA expression between twopopulations [12].

The nucleotide sequence data reported will appear in the EMBL,GenBank and DDBJ Nucleotide Sequence Databases under theaccession numbers AF030070 (GO15-21), AF030071 (GO15-13),AF030072 (AK2-10) and AF030073 (AK3-2).

To assess the overall performance of differen-tial display, preliminary experiments were performedusing total RNA extracted from whole etiolatedNico-tiana tabacumseedlings treated or not with auxin.Using the conventional differential display protocoldescribed by Liang et al. [7], with a PCR amplificationof 40 cycles, the resulting bands were difficult to dis-cern due to a strong background that partly masked theamplified products (Figure 1A). In order to reduce thisbackground, we have optimised the differential displayprocedure by developing a two-step PCR amplificationof 25 cycles each (Figure 1B). The first set of amplific-ations was performed in the absence of radiolabelleddNTP and was followed by a second phase of ampli-fication in the presence of�-33P-dATP, starting withone tenth of the first amplification reaction medium. Inall cases, the PCR were duplicated to ensure the repro-ducibility and significance of the patterns obtained on6% polyacrylamide gel. The two-step amplificationprotocol lead to a reliable profile with less backgroundand easily detectable bands.

Differential display has been applied to intact eti-olated tobacco seedlings grown in Gamborg’s liquidmedium for 7 days and treated or not with 1�m NAA(Naphthalene acetic acid) for 30 min. These condi-

386

Figure 1. Comparison of amplification patterns obtained with twodifferent protocols of differential display RT-PCR. A. The amplific-ation has been realised for 40 consecutive cycles as described byLianget al. [7]. B. A two-step procedure of 25 cycles each has beenperformed: the first 25 cycles in the absence of radiolabelled dATP,the second 25 cycles in the presence of�-33P-dATP. Seven day-oldetiolated seedlings ofNicotiana tabacumL. cv Xanthi were treated(+) or not (�) with 1 �M NAA for 30 min in the dark. Total RNAwere extracted and treated with DNase RQ1 (Promega). cDNA weresynthetized as described in [5] from 1�g of total RNA in presenceof 1 �M of the anchored primer, T12MG in the presented example.One tenth of each cDNA was PCR-amplified in presence of 10�Mof each dNTP, 3�M of T12MG, 225 nM of arbitrary primer (O15),0.25 U of Taq DNA polymerase (Appligene) with its own buffercontaining 1.5 mM MgCl2, in the presence or not of 75 kBq of�-33P-dATP. After denaturation at 92�C for 3 min, each cycle con-sisted of 92�C for 45 s, 40�C for 2 min and 70�C for 1 min. Thelast cycle was followed by a 5 min extension step at 70�C. Onesixth of the PCR was mixed with formamide loading buffer and afterdenaturation at 80�C for 2 min, the samples were resolved on a(6% acrylamide; 7 M urea) gel. The arrow indicates a differentiallydisplayed band (GO15) present in auxin treated samples.

tions should allow us to identify differences in mRNAexpression appearing within a short time, in responseto a rather low concentration of auxin (to avoid stressresponses). On this mRNA, 120 combinations ofprimers have been used: three poly(A)-anchored oli-gonucleotides (T12MA, T12MC and T12MG whereM was A, C or G) were used in combination with 40arbitrary decamers (Operon Technologies, Alameda,CA). The number of observed bands varied with thecombination of primers providing an average numberof 65 bands with sizes ranging from 150 to 500 bp on astandard profile resolved in a 6% polyacrylamide gel.Thus, 28 displayed bands were found to be specificallyexpressed after auxin treatment of the seedlings,amongthe approximately 7800 bands observed. Among these28 bands, 23 have been successfully re-amplified fromexcised bands and differential expression has beeninvestigated for 15 using northern blot analysis. ThePCR amplification products were labelled using ran-dom priming in the presence of�-32P-dATP and werehybridized on 4�g of poly(A) RNA extracted from7-day-old etiolated seedlings treated or not for 2 hwith 1 �m NAA. Most of the fragments have givenmultiple hybridization signals revealing that the ini-tial band contained overlapping DNA fragments. Onlythree PCR products, resulting from the amplificationwith the combination of primers T12MG-O15 (50-TGGCGTCCTT), T12MA-K2 (50-GTCTCCGCAA)and T12MA-K3 (50-CCAGCTTAGG), have exhibiteddetectable differences of hybridization between themRNA from auxin-treated and mock-treated samplesas shown in Figures 2A, 3B and 4B, respectively.

To facilitate the re-amplification and furthercloning of the PCR products, the DNA frag-ments were re-amplified using a combination ofmodified anchored and arbitary primers containingBamHI sites (the anchored primer was BT10: 50-TCAGCAGGATCC(T)10 and the arbitrary primerswere BO15: 50-GACGGATCCTGGCGTCCTT; BK2:50-GACGGATCCGTCTCCGCAA and BK3: 50-GACGGATCCTGGCGTCCTT). After a denaturationperiod at 92�C for 3 min, the amplifications wereperformed for 5 cycles at 92�C for 45 s, 40�C for2 min and 70�C for 90 s followed with 30 cyclesat 92 �C for 1 min; 57�C for 2 min and 70�C for2 min and an additional extension period at 70�Cfor 5 min. The re-amplified products were digestedwith BamHI, purified from 1.5% agarose gels and thenligated intoBamHI-linearized and dephosphorylatedBlueScript plasmid. This procedure has been applieddirectly to the first amplification products or to cDNA

387

Figure 2. A. Northern blot analysis of the GO15 band obtained bydifferential display. PolyA RNA were extracted from dark-grownseedlings mock-treated (�) or treated (+) with 1 �M NAA for 2 hand 4�g were separated onto formaldehyde gel and transferred ontonylon membrane. Hybridization was performed using the T12MG-O15 re-amplified crude products as probe. B. Hybridizations of theGO15-21, GO15-13inserts andubiquitin (as a loading control) with4 �g poly(A) RNA from mock-treated (�) or treated (+) samples.The probes were radiolabelled with 1.8 MBq of�-32P-dATP. Hybrid-izations were done at 60�C for 12 h in 7% SDS, 5�SSC, 10�Den-hardt’s solution, 100�g/ml sonicated, denaturated salmon spermDNA. The membranes were successively washed in 4� SSC/1%SDS; 2� SSC/1% SDS and 1� SSC/0.5% SDS for 30 min eachand exposed to Kodak X-Omat film. C. Sequence alignment of theinsertGO15-13andSAUR 10XA[10]. The framed portion corres-pond to the initial primer O15. The grey portions corresponds to thenucleotides identical between the two sequences. The stop codon ofSAUR 10XAis in bold characters.

probes recovered from the northern blot using affinity-capturing method [6]. The choice of non-orientatedligation was intentional to recover all the fragmentswhatever the primers’ end. DNA sequencing was per-formed in an ABI 373A automated DNA sequencer(Applied Biosystems) and sequence analysis was car-ried out using the GCG software (Genetics ComputerGroup, Madison, WI) and the BLAST or FASTA net-work services.

Figure 3. A. Portion of the amplification patterns obtained by dif-ferential display with the oligonucleotides T12MA and K2 (tripletband: AK2). B. Hybridization of the crude re-amplified productswith 4�g poly(A) RNA from mock-treated (�) or auxin-treated (+)samples. Conditions of labelling, hybridization and loading controls(not shown) were as described in Figure 2.

The amplification pattern obtained with the com-bination of T12MG and O15 oligonucleotides hasrevealed a differential display band named GO15 (Fig-ure 1). The hybridization of the re-amplified PCRproducts, shown in Figure 2A, gave complex pat-terns revealing at least three mRNAs constitutivelyexpressed and one mRNA of about 2.2 kb specificallyexpressed in auxin-treated tobacco seedlings. cDNAfragments recovered from the northern blot using affin-ity capturing were re-amplified in the presence of BT10and BO15 primers and cloned as described above.Ten independent clones have been sequenced. Afteraffinity capturing, four distinct sequences have beenfound: two contained the BO15 and BT10 primers andthe others ended with two BT10 primers. The insertsGO15-13andGO15-21hybridized with mRNA of 0.6and 2.2 kb, respectively, both exhibiting differentialexpression between auxin-treated and mock-treatedseedlings (Figure 2B). The insertGO15-21correspon-ded to the differential signal observed after hybridiza-tion with the T12MG/O15 re-amplified products (Fig-ure 2A). The sequence analysis and comparison withdatabases did not reveal any homology ofGO15-21with known sequences. The insertGO15-13hybridizedwith a low-abundant mRNA which was not detected onthe first northern blot as it was masked by the hybrid-ization signal of another fragment to a constitutivelyexpressed mRNA of approximately the same size. TheGO15-13sequence presents 74% identity on a lengthof 62 bases with the sequence ofSAUR X10Aisolatedfrom Glycine max[10]. The identities are located atthe 30 end of the coding frame of the soybeanSAURwhereas its 30 non-coding region does not show signi-ficant homology with the tobaccoGO15-13sequence(Figure 2C).

The combination of primers T12MA and K2 haslead to a pattern which showed a band (AK2) present

388

Figure 4. A. Portion of amplification patterns obtained by differentialdisplay with the oligonucleotides T12MA and K3 (AK3 band). B.Hybridization of the crude re-amplified products with 4�g poly(A)RNA from mock-treated (�) or auxin-treated (+) seedlings. Con-ditions of labelling, hybridization and loading controls (not shown)were as described in Figure 2.

only in the treated samples (Figure 3A). This band,located at the bottom of the denaturating gel, appearedin fact to be a set of three bands with nearly the sameintensity. It has been suggested that such bands couldbe the two denatured strands of one fragment andmolecules with or without an additional A, known to beadded by theTaqpolymerase at the 30 end [2]. The PCRreamplified products hybridized strongly with a mRNAof about 1.5 kb whose expression level is stronglyenhanced after treatment with 1�m NAA (Figure 3B).After the direct cloning of the re-amplified PCR frag-ments, the sequencing of 18 independent clones hasrevealed that 8 types of sequences were present withinthe BT10-BK2 products, with no homology betweenthem. One of them, theAK2-10insert, hybridized witha mRNA of 1.5 kb expressed in auxin-treated seedlings.Clones of similar sequence have been identified, theirsizes ranging from 167 to 172 bp due to a variable num-ber of A at the 30 end of the sequence. This confirms thehypothesis formulated about the observed triplet bands.A search in data banks did not reveal any homology ofthis sequence with known sequences. Hybridization ofthe labelledAK2-10insert, on a northern blot, alloweddetection of a mRNA of about 1.5 kb present only inauxin-treated samples (data not shown).

The differential display patterns obtained with theprimers T12MA and K3 and northern blot hybridiz-ation with the crude re-amplified products, from theAK3 band, are presented Figure 4A and 4B. The dif-ferential expression was confirmed at least for onefragment but again, a complex pattern was obtained.After cloning, the sequencing of the resulting cloneshas revealed 6 distinct sequences with no homologybetween them. The insertAK3-2 hybridized with anmRNA of about 1 kb expressed in auxin-treated seed-lings and corresponded to the previous observation

whereas the other inserts have given no differen-tial signal (data not shown). The sequence of insertAK3-2 presents no significant homology with knownsequences available in the databases.

We have improved the differential display proced-ure by developinga two-step amplification PCR (2�25cycles) allowing a significant reduction in the back-ground of the patterns and an increase in the resolutionof the gels. In addition, we have used modified oli-gonucleotides, including aBamHIsite, to facilitate there-amplification and the cloning of the band of interest.The number of tested combinations (120) should besufficient to screen all the mRNA populations as it hasbeen theoretically calculated by Liang and Pardee [8]and by Bauer and Strauss [2], but this evaluation isstill questionable. The differential display is first ofall dependent on the primers used and not only on themRNA expression level. Therefore, differential dis-play does not allow the systematic screening of differ-ences between two populations, but is still the methodof choice for the cloning of differentially expressedmRNAs as it identifies mRNA independent of pre-valence [12]. The relatively low occurence of differ-entially displayed fragments suggests that few mRNAare differentially expressed between etiolated seedlingstreated or not with 1�m NAA for 30 min. Amongthe differentially displayed fragments, a vast major-ity of re-amplified products hybridized with numerousmRNAs which were expressed at the same level inauxin-treated and mock-treated samples. These bandscould be false positives and could also mask differ-ential expression of lower abundant mRNAs as wehave found in the case of the insertGO15-13. Theselection of interesting fragments becomes hazardouson the basis of northern blot analysis. To confirmthe differential display obtained on the polyacrylam-ide gel remains an unsolved problem of dd-RT-PCR,even though numerous technical tricks have been pro-posed since 1992 [9]. However, this differential dis-play approach has permitted isolation of 4 sequencescorresponding to mRNA whose expression levels areenhanced in whole tobacco seedlings within 30 minof NAA treatment at 1�m. The sequence analysisof the GO15-13insert suggests that the correspond-ing mRNA could be related to theSAURgene family.The isolation and characterization of the correspondingfull-length cDNA will allow verification of this pre-liminary result. Considering the relatively short sizeof the three other cloned fragments (172 to 324 nt),the majority of the sequences could correspond to 30-untranslated regions. These regions are known to be

389

poorly-conserved between homologous genes in con-trast with coding regions which reveal significant iden-tities. TheGO15-21, AK2-10andAK3-2 inserts showno homology with known sequences available in data-bases. The isolation of the corresponding full-lengthcDNAs and expression studies will permit further char-acterization of these potentially novel auxin-regulatedmRNAs.

Acknowledgements

We thank Dr Spencer Brown for critical reading of themanuscript and Dr Ulla Bonas for kindly providing thecollection of decamer oligonucleotides (Operon Tech-nologies). This paper is based in part on a doctoralthesis by C.R. supported by a MESR (Ministere del’Enseignement Superieur et de la Recherche) grant.J.B. was a recipient of the Roche Research Foundationand the Fonds National Suisse.

References

1. Abel S, Theologis A: Early genes and auxin action. PlantPhysiol 111: 9–17 (1996).

2. Bauer B, Muller H, Reich J, Riedel H, Ahrenkiel V, Warthoe P,Strausss M: Identification of differentially expressed mRNA

species by an improved display technique (DDRT-PCR). NuclAcids Res 21: 4272–4280 (1993).

3. Davies P: Plant Hormone Physiology, Biochemistry andMolecular Biology, 2nd ed. Kluwer Academic Publishers,Dordrecht (1995).

4. Hagen G: The control of gene expression by auxin. In: Dav-ies P (ed) Plant Hormone Physiology, Biochemistry andMolecular Biology, pp. 228–245. Kluwer Academic Publish-ers, Dordrecht (1995).

5. Leblanc N, Roux C, Pradier JM, Perrot-Rechenmann C: Char-acterization of two cDNAs encoding auxin-binding proteins inNicotiana tabacum. Plant Mol Biol 33: 679–689 (1997).

6. Li F, Barnathan E, Kariko K: Rapid method for screeningand cloning cDNAs generated in differential mRNA display:application of Northern blot for affinity capturing of cDNAs.Nucl Acids Res 22: 1764–1765 (1994).

7. Liang P, Averboukh L, Pardee AB: Distribution and clon-ing of eucaryotic mRNAs by means of differential display:refinements and optimization. Nucl Acids Res 21: 3269–3275(1993).

8. Liang P, Pardee AB: Differential display of eukaryotic messen-ger RNA by means of the polymerase chain reaction. Science257: 967–971 (1992).

9. Liang P, Pardee AB: Recent advances in differential display.Curr Opin Immunol 7: 274–280 (1995).

10. McClure B, Hagen G, Brown C, Gee M, Guifoyle T: Tran-scription, organization, and sequence of an auxin-regulatedgene cluster in soybean. Plant Cell 1: 229–239 (1989).

11. Sitbon F, Perrot-Rechenmann C: Expression of auxin-regulatedgenes. Physiol Plant 100: 443–455 (1997).

12. Wan JS, Sharp SJ, Poirier GMC, Wagaman PC, Chambers J,Pyati J, Hom YL, Galindo JE, Huvar A, Peterson PA, JacksonMR, Erlander MG: Cloning differentially expressed mRNAs.Nature Biotechnol 14: 1685–1691 (1996).