Involvement of the xyloglucan endotransglycosylase/hydrolases encoded by celery XTH1 and Arabidopsis XTH33 in the phloem response to aphids

Involvement of the xyloglucanendotransglycosylase/hydrolases encoded by celery XTH1and Arabidopsis XTH33 in the phloem response to aphids

FANCHON DIVOL1,2, FRANÇOISE VILAINE1, SANDRA THIBIVILLIERS1,4, CHANTAL KUSIAK1,MARIE HELENE SAUGE3 & SYLVIE DINANT1

1Laboratoire de Biologie Cellulaire UR501, Institut National de la Recherche Agronomique (INRA), Versailles F-78026,France, 2Laboratori CSIC-IRTA de Genética Molecular Vegetal, Barcelona 08034, Spain, 3Plantes et Systèmes de culturehorticoles, UR115, INRA, Avignon F-84914, France and 4Department of Plant Microbiology and Pathology, University ofMissouri, Columbia, MO 65211, USA

ABSTRACT

During infestation, phloem-feeding insects induce tran-scriptional reprogramming in plants that may lead to pro-tection. Transcripts of the celery XTH1 gene, encoding axyloglucan endotransglycosylase/hydrolase (XTH), werepreviously found to accumulate systemically in celery(Apium graveolens) phloem, following infestation with thegeneralist aphid Myzus persicae. XTH1 induction was spe-cific to the phloem but was not correlated with an increasein xyloglucan endotransglycosylase (XET) activity in thephloem. XTH1 is homologous to the Arabidopsis thalianaXTH33 gene. XTH33 expression was investigated followingM. persicae infestation. The pattern of XTH33 expression istightly controlled during development and indicates a pos-sible role in cell expansion. An xth33 mutant was assayedfor preference assay with M. persicae. Aphids settled pref-erentially on the mutant rather than on the wild type. Thissuggests that XTH33 is involved in protecting plants againstaphids; therefore, that cell wall modification can alter thepreference of aphids for a particular plant. Nevertheless,the ectopic expression of XTH33 in phloem tissue was notsufficient to confer protection, demonstrating that modify-ing the expression of this single gene does not readily alterplant–aphid interactions.

Key-words: Apium graveolens; Arabidopsis thaliana; Myzuspersicae, cell wall; compatible interaction; plant–aphidinteraction.

INTRODUCTION

The phloem is a highly specialized tissue with major physi-ological functions, such as photoassimilate translocation,carbon and nitrogen partitioning (Sjölund 1997; Oparka &Turgeon 1999), and long-distance interorgan signalling indevelopmental, physiological processes and responses tobiotic stresses (Ruiz-Medrano, Xoconostle-Cazares &Lucas 2001; van Bel 2003; van Bel & Gaupels 2004). As the

phloem provides a high-speed pathway between distantorgans, it is used by plant pathogens, such as viruses, viroids,phytoplasms and phloem bacteria, for efficient propagationwithin the plant. The high sugar and free amino acid con-centrations of the phloem sap make it the targeted diet ofnumerous piercing-sucking insects, such as whiteflies,cicadels and aphids (Gatehouse 2002; Douglas 2006).

Aphids, the largest group of phloem feeders, can havemajor effects on carbon assimilation and assimilate parti-tioning, severely impairing plant growth and development(Girousse et al. 2003, 2005; Macedo et al. 2003). Aphidscause damage to cells encountered during the progressionof their stylets towards the phloem sieve elements (Miles1999). Salivary secretions emitted by aphids during feedingare believed to limit plant response to wounding (Miles1999; Tjallingii 2006) and to prevent the plugging of sieveelements (Will & van Bel 2006). In response, the plantsperceive and react to aphid infestation by a number ofmetabolic and physiological modifications limiting thedamage caused by sustained insect feeding. Recent tran-script profiling analyses have greatly advanced our under-standing of plant defence mechanisms against aphids(Moran & Thompson 2001; Moran et al. 2002; Heidel &Baldwin 2004; Voelckel, Weisser & Baldwin 2004; Zhu-Salzman et al. 2004; Divol et al. 2005; Qubbaj, Reineke &Zebitz 2005; de Vos et al. 2005; Park, Huang & Ayoubi2006). These studies demonstrate extensive reprogram-ming of transcription in plants following aphid attack. Thegenes affected encode proteins involved in known defencepathways, oxidative stress responses, plant structure andmetabolism (Kaloshian & Walling 2005; Thompson &Goggin 2006). It is tempting to speculate that these genesare involved in limiting aphid infestation, but effects ontranscript accumulation may not necessarily be correlatedwith changes in accumulation of the corresponding proteinor its activity. Moreover, changes in transcript levels may bean indirect consequence of the primary plant response toaphids, and the down-regulation of gene expression doesnot necessarily affect plant–aphid interactions.

As aphids use sieve elements for feeding, the phloemresponse to aphids is probably a key element in the

Correspondence: Sylvie Dinant. Fax: 33 1 30 83 30 99; e-mail:[email protected]

Plant, Cell and Environment (2007) 30, 187–201 doi: 10.1111/j.1365-3040.2006.01618.x

Journal compilation © 2007 Blackwell Publishing LtdNo claim to original French government works 187

mailto:[email protected]

establishment of plant defences. But the mechanismsinvolved in this tissue are difficult to observe and may bemasked by the responses of other tissues, because thephloem accounts for only a small fraction of the volume ofplant organs (Sjölund 1997). An alternative approach is toanalyse plant responses in phloem-enriched tissues (Asanoet al. 2002; Nakazono et al. 2003; Vilaine et al. 2003) orphloem sap (Walz et al. 2004; Giavalisco et al. 2006). As thephloem response may differ from the whole plant reaction,we previously analysed the transcriptional response toinfestation with the generalist peach-potato aphid (Myzuspersicae) in celery (Apium graveolens) phloem. Celery wasstudied because the phloem is readily separated from othertissues by peeling in this species. Many of the genes display-ing changes in transcript accumulation encoded proteinsinvolved in cell wall remodelling (Divol et al. 2005), such asxyloglucan endotransglycosylase/hydrolase (XTH), pectinesterase, pectin acetyl esterase, expansin and cellulose syn-thase. Deregulation of the expression of genes involved inwater transport in the cells was also observed, suggestingthat both events might reflect specific regulatory modifica-tions to the phloem cell wall, affecting local water retrievalin sieve elements (Divol et al. 2005), compensating for thevolume of sap taken up by aphids. Alternatively, cell wallmodifications may deter phytophagous insects both locallyand systemically, by strengthening barriers against probingand feeding (Thompson & Goggin 2006), as observed inplant defences against chewing insects or a range of patho-gens (Esquerre-Tugaye, Boudart & Dumas 2000; Vorwerk,Somerville & Somerville 2004).

A main class of genes deregulated following aphid infes-tation in celery (Divol et al. 2005) encoded XTH. XTHscatalyse the breaking and reformation of xyloglucancrosslinks and are involved in the transfer of xyloglucanchains on hemicelluloses (Campbell & Braam 1999). Xylo-glucans bind non-covalently to cellulose, coating and crosslinking adjacent cellulose microfibrils. The resulting exten-sive xyloglucan-cellulose network is the major tension-bearing structure in the primary wall. These enzymes areconsidered essential for cell wall remodelling, including cellwall architecture, strength and extensibility (Fry 1989; Fry1992; Nishitani & Tominaga 1992). The role of XTHs in cellexpansion is well documented (Burstin 2000; Hyodo et al.2003; Vissenberg et al. 2003; Matsui et al. 2005; Romo et al.2005; Shin et al. 2005; Vissenberg et al. 2005; Wu et al. 2005),although an hydrolase activity, enabling the release of oli-gosaccharides, has also been described (Fanutti, Gidley &Reid 1993; de Silva et al. 1993; Tabuchi et al. 2001).

Celery XTH1 displayed the largest systemic increase intranscript accumulation of any of the cell wall genes, par-ticularly at 7 d post infestation (Divol et al. 2005). To get aninsight in the function of XTH1 in response to M. persicaeinfestation, the specificity of expression of XTH1 was inves-tigated, either at a tissue level or in response to wounding.The transcriptional regulation of XTH1 appeared specific.A parallel study was engaged on a closely related XTHgene in Arabidopsis thaliana, a species that is also sus-ceptible to M. persicae infestation. The specificity of the

transcriptional regulation of this gene was analysed toevaluate whether it had retained the same function regard-ing aphid infestation. The response to aphid infestation ofone Arabidopsis mutant in which this XTH gene was dis-rupted by T-DNA insertion was analysed. This study alsoconsidered whether plant–aphid interactions in transgenicplants could be modified by the ectopic expression of thisgene in the phloem.

MATERIALS AND METHODS

Plant and aphid material

Seeds of celery (Apium graveolens var. dulce cv. Vertd’Elne) were allowed to germinate for 1 week on wetWhatman paper at 20 °C, in a growth room with a 16 hlight/8 h dark cycle and 70% relative humidity. They werethen grown in soil in the greenhouse under long-day con-ditions (14 h day/10 h night cycle) with differences betweenday and night temperatures (20 °C day/15 °C night cycle).

Seeds of Arabidopsis xth33 mutant lines (#Salk_072153,#Salk_120473) were ordered from the Salk Institute (LaJolla, CA, USA). Seeds of Arabidopsis pp2-A1 line(#Salk_080914.54.75) were provided by Castelain (unpub-lished results). These mutants were obtained in aColumbia-0 (Col-0) background. Seeds of Arabidopsis(wild-type Col-0, xth33 mutant line or transformedpXTH33-GUS lines) were sterilized and sown on Murash-ige and Skoog medium (MS) plates. After 2 d at 4 °C, theseedlings were grown in a growth chamber (200 mE m-2 s-1,16 h day/8 h night, 20 °C, 70% relative humidity). Primarytransformants, their progeny and Salk mutants wereselected on MS plates containing the appropriate antibiotic,either kanamycin (50 mg L-1) or hygromycin (15 mg L-1).The plants were then allowed to acclimate in the green-house, where they were grown in soil under long-day con-ditions (14 h day, 10 h night, 20 °C). The seeds of self-fertilized plants were collected and germinated on MSplates supplemented with kanamycin, for genetic analysisbased on the segregation of kanamycin resistance and theselection of homozygous plants.

The peach-potato aphid Myzus persicae (Sulzer) strainSUS (provided by R. Delorme, INRA, Versailles, France)was maintained on broad bean grown in soil in a growthchamber.

Molecular characterization of the Arabidopsismutant line

An Arabidopsis xth33 mutant line (lines #Salk_072153)was identified by Basic Local Alignment Search Tool(BLAST) search, using the At1g10550 sequence, on theArabidopsis mutant database available at the Salk Institute(http://signal.salk.edu/cgi-bin/tdnaexpress/). The mutantwas generated by Agrobacterium-mediated T-DNA inser-tion in the Col-0 ecotype (Alonso et al. 2003). The annota-tion in the Salk Institute Genome Analysis Laboratory(SIGnAL) database from a single-pass sequence recovered

188 F. Divol et al.

Journal compilation © 2007 Blackwell Publishing LtdNo claim to original French government works, Plant, Cell and Environment, 30, 187–201

http://signal.salk.edu/cgi-bin/tdnaexpress

from the left border of the T-DNA indicates that the#Salk_072153 insertion was located in the third exon ofAt1g10550. The insertion was confirmed by Southern blot-ting, using EcoRI or BamHI for the digestion of genomicDNA, which did not reveal any large deletion in this region.Mutant plants homozygous for the T-DNA insertion wereidentified among the progeny by PCR, using the specificprimers (5′ATGAAGATTATGTGGGAAACAGC3′ and5′ATGACTTTGTATCTTGGCTTATCACTG3′). For thepp2a1 mutant, homozygous progeny plants (#Salk_080914.54.75) were used.The insertion in this line lies in thepromoter region of At4g19840 and does not compromisePP2-A1 expression.

Aphid infestation, choice test and wounding

For the assay on celery, the plants were infested by a releaseof 200–300 aphids per plant onto the mature leaves of adultcelery plants, as described (Divol et al. 2005). Control andinfested plants were grown in separate insect-proof cages inthe same growth chamber. Plant response was analysed 3 or7 d post infestation (3 or 7 dpi) on individual plants, and thetissues studied were isolated from the petioles of extendedleaves sampled at about the same time of day (1000–1100 h). Phloem strands were isolated from petioles asdescribed by Daie (1987), and xylem strands and paren-chyma were also isolated from the same plants.

For the assay on Arabidopsis, aphids grown on broadbean were applied directly onto the rosette leaves of 6- to8-week-old plants grown in the greenhouse.The plants wereanalysed 48 and 72 h, and 7 d post infestation. An in vitrobioassay was used to evaluate the response of Arabidopsisgenotypes to aphid infestation, as described by Cabrera yPoch, Ponz & Fereres (1998). Briefly, three sterilized seedsfor each genotype (wild-type Arabidopsis Col-0 and mutantline) were evenly distributed on a Petri dish containing MSmedium. After 2 d at 4 °C, the plates were transferred to agrowth chamber for 3 weeks (200 mE m-2 s-1, 16 h day/8 hnight, 20 °C, 70% relative humidity).At this stage, the plantswere infested with M. persicae: on average, 20 apterousadult aphids were released in the Petri dish, on a sterileWhatman paper placed at an equal distance from all plants.The number of adult aphids and newborn larvae present oneach wild-type and mutant plant was recorded after 24 h,estimating preference for a given genotype (choice test).The experiments were repeated for each combination ofmutant and wild-type lines: 43 times for mutant line xth33-1,74 for pPP2-XTH (including 49 and 25 times for pBIN19and pCAMBIA derivatives, respectively) and 10 times forthe mutant line pp2-a1.

For wounding experiments, several leaflets per leaf werewounded on celery plants, as previously described (Pautot,Holzer & Walling 1991). Phloem from wounded plants wassampled 3 or 7 d after wounding.

XTH sequences

Celery expressed sequence tag (EST) identified by thelarge-scale sequencing of several celery phloem cDNA

libraries (Vilaine et al. 2003; Divol et al. 2005) were anno-tated as putative XTH genes if a significant probabilityvalue (E-score < 10-5) was obtained with other XTHsin a BLAST search of available sequence databases (http://www.ncbi.nlm.nih.gov/BLAST/). BLAST (Altschul et al.1997) and FASTA (Genetics Computer Group (GCG),Madison, WI, USA) analysis identified six distinct contigs,on the basis of local alignment, for the eight ESTs. The BIPclones (Table 1) were generated using the pCRII-TOPOvector (Invitrogen, Cergy Pontoise, France), whereas thePA cDNA clones were constructed with the pBluescriptvector (Stratagene, La Jolla, CA, USA). The full-lengthXTH1 cDNA was obtained by primer extension, usingcelery phloem cDNA libraries (Vilaine et al. 2003); analiquot of the phage PA cDNA library was used for PCR,using XTH-specific primers complementary to the 5′ or 3′borders of the initial EST (CN254566) and polylinkerprimers (M13 forward or reverse) complementary tosequences in the polylinker of the cloning vector pBlue-script (Stratagene). Amplification products were intro-duced into the pCRII-TOPO vector (Invitrogen) andsequenced. This made it possible to extend the XTHsequence gradually to cover the full-length open readingframe (ORF) and the 5′ and 3′ untranslated regions (UTRs)(GenBank accession number DQ204724).

Arabidopsis XTH genes were analysed using the geneannotations proposed by Yokoyama & Nishitani (2001).The Arabidopsis Genome Initiative identification numbersof these various XTH genes are as follows: XTH1(At4g13080), XTH2 (At4g13090), XTH3 (At3g25050),XTH4 (At2g06850), XTH5 (At5g13870), XTH6(At5g65730), XTH7 (At4g37800), XTH8 (At1g11545),XTH9 (At4g03210), XTH10 (At2g14620), XTH11(At3g48580), XTH12 (At5g57530), XTH13 (At5g57540),XTH14 (At4g25820), XTH15 (At4g14130), XTH16(At3g23730), XTH17 (At1g65310), XH18 (At4g30280),XTH19 (At4g30290), XTH20 (At5g48070), XTH21(At2g18800), XTH22 (At5g57560), XTH23 (At4g25810),XTH24 (At4g30270), XTH25 (At5g57550), XTH26(At4g28850), XTH27 (At2g01850), XTH28 (At1g14720),XTH29 (At4g18990), XTH30 (At1g32170), XTH31(At3g44990), XTH32 (At2g36870), XTH33 (At1g10550).Arabidopsis XTHs were compared with other plant XTHgenes or ESTs by means of BLAST similarity searches(Altschul et al. 1997) at The Arabidopsis InformationResource [The Arabidopsis Information Resource (TAIR):http://www.arabidopsis.org/cgi-bin/blast)]. The degree ofidentity and similarity between pairs of sequences was cal-culated using the GAP programme of the GCG package,with the default parameters. Sequences were aligned byPileup, from the GCG package. Phylogenetic reconstructionwas carried out by the neighbour-joining method (Saitu &Nei 1987), based on distances corrected for a high rate ofsubstitution, using ClustalW (Thompson, Higgins & Gibson1994) (version 1.81) from GCG. The phylogenetic tree wasgenerated using default parameters, except for correctionfor the substitution rate based on Kimura distances.Bootstrapping was performed with 1000 replicates. The

Celery XTH1 and Arabidopsis XTH33 in plant response to aphid 189


http://www.ncbi.nlm.nih.gov/BLAST

http://www.ncbi.nlm.nih.gov/BLAST

http://www.arabidopsis.org/cgi-bin/blast

phylogenetic tree was displayed with TreeView software(Page 1996).

Other angiosperm EST sequences related to celeryXTH1 and Arabidopsis XTH33 were found using BLASTsimilarity searches in Asparagus officinalis (CV707639.1),Antirrhinum majus (AJ794821.1), Glycine max(CV288396.1, CX710147.1), Solanum tuberosum(DR034932.1, CV302294.1), Citrus sinensis (CX047376.1,CX047377.1) and Populus balsamifera (CV227458.1,CN519887.1) with significant identity scores and BLAST Evalues.

Cloning of the Arabidopsis XTH33 promoterand coding region and plant transformation

The XTH33 promoter was cloned with Gateway cloningtechnology (Invitrogen) using a binary vector derived frompBI101-GUS. This binary vector was constructed by trans-ferring the Gateway R1-R2 recombination cassette intopBI101-GUS (Jefferson, Kavanagh & Bevan 1987). AHindIII-XbaI fragment containing the R1-R2 recombina-tion cassette (including ccdB and the CmR genes) wasligated to the coding sequence of the UidA gene present inpBI101-GUS. This destination vector, pBI101-R1R2-GUS,was verified by the subsequent cloning of a 35S promoter(unpublished). The XTH33 promoter region (1245 bpupstream from the initial ATG) was amplified by PCR,using sense 5′AAAAAAGCAGGCTTAAAATCAAGAA

TAAAACAG3′ and antisense 5′AAGAACGATGGGTCGTTGTGTTTTATTCATCT3′ primers containingXTH33-specific sequences (underlined) and recombinationsite-specific sequences (1/2 attB1 or 1/2 attB2 for sense andantisense primers, respectively). A second PCR was per-formed with primers reconstituting intact attB recombina-tion sites on the 5′ and 3′ PCR products. The resultingattB-pXTH33 product was introduced into pDONR 207(Invitrogen) by BP recombination, according to the manu-facturer’s recommendations, followed by LR recombina-tion in the pBI101-R1R2-GUS destination vector. Theresulting plasmid, pXTH33-GUS, consisting of pXTH33-GUS-tNOS cassette in a pBI101 derivative, was checked bysequencing. In this construct, the XTH33 promoter drivesexpression of the UidA gene (b-glucuronidase).

For ectopic XTH33 expression, the full-length codingregion was amplified by RT-PCR, using the followingprimers: 5′ATGAAGATTATGTGGGAAACAGC3′ and5′TCAGTTGCACTCAGCAGGCATG3′. It was intro-duced into an intermediate pCR2TOPO vector between a1600 bp promoter region of AtPP2-A1 (At4g19840) and theNOS terminator. The XhoI-BamHI fragment, correspond-ing to the chimeric gene pPP2-A1-XTH33-tNOS, was intro-duced in BamHI-SalI sites of the multiple cloning siteeither of the pBIN19 binary vector (Bevan 1984), or thepCAMBIA1300 binary vector (http://www.cambia.org/),creating the pPP2-XTH33 vectors. pBIN19 and pCAMBIAderivatives confer kanamycin or hygromycin resistance,respectively, for selection in planta.

Table 1. Characteristics of six XTH genes expressed in the phloem of celery

Gene Clone

GenBank accessionnumber

XTHgroupa

cDNAlibraryb

Transcriptional response toaphid infestationc

Tissue specificityd(EST) 3 dpi 7 dpi

XTH1 BIP0760 CN254566 3 ai† 1.24 2.49** n.d.XTH2 BIP1068 CN254817 3 ai† n.s. n.s. n.d.XTH3 BIP0938

BIP0635CN254704CN254464

1 ai† n.s. 1.58* n.d.

XTH4 BIP0934BIP0841

CN254701CN254625

1 ai† 2* 1.31* n.d.

XTH5 PA0495 BU693307 2 stg‡ n.d. n.d. All tissuesXTH6 BIP1089 CN254829 1 ai† 2* n.s. n.d.

*P < 10-2, **P < 10-3.†cDNA library generated from the phloem of plants infested by aphid and enriched in induced sequences.‡cDNA library generated from the phloem of plants grown in standard growth conditions.aCelery XTH genes were classified into the XTH groups defined by Campbell & Braam (1999), after multiple alignments of the sequences,using ClustalW (Thompson, Higgins & Gibson 1994).bAll XTH cDNAs were obtained by suppressive subtractive hybridization (SSH), using the phloem of infested and control plants (Divol et al.2005), with the exception of the XTH5 cDNA, which was obtained from a phloem cDNA library produced from phloem extracted from aplant grown in standard conditions (Vilaine et al. 2003).cThe response to aphids was determined using a set of cDNA macroarrays for differential hybridization with complex probes correspondingto the mRNAs from the phloem of the petioles of infested and uninfested plants (Divol et al. 2005). The infested/uninfested ratio is indicatedwhen statistically significant.dTissue specificity was determined using a set of cDNA macroarrays for differential hybridization with complex probes corresponding tomRNAs from the phloem, xylem or storage parenchyma of the petioles of plants grown in stg conditions (Vilaine et al. 2003).ai, aphid-induced library; stg, standard growth; n.s., non-significant; n.d., not determined; EST, expressed sequence tag.

190 F. Divol et al.


http://www.cambia.org

The pXTH33-GUS and pPP2-XTH33 plasmids wereintroduced into Agrobacterium tumefaciens GV3101(Koncz & Schell 1986) by electroporation. The resultingAgrobacterium strains were used to transform ArabidopsisCol-0 by floral dip treatment (Clough & Bent 1998). Trans-formants were selected on the proper antibiotics. ResistantpXTH33-GUS plants in the progeny (T2) were analysed forb-glucuronidase activity, and T2 pPP2-XTH33 plants wereanalysed for RT-PCR. For practical purposes, homozygouskanamycin-resistant T3 plants were also obtained fromlines presenting a single locus insertion.

RNA isolation and Northern blot analysis

Total RNA was isolated from frozen tissue, as described byVerwoerd, Dekker & Hoekema (1989), for Northern blotexperiments. Total RNA (10 mg) was subjected to electro-phoresis in denaturing 1% agarose gels containing formal-dehyde. The RNA was transferred onto nylon membranefilters (Genescreen, Life Science, Boston, MA, USA) bycapillary action, overnight, in 10 ¥ SSC (1.5 M NaCl, 0.15 Msodium citrate). The membrane was then washed in 2 ¥SSC, and the RNA was cross linked to the membrane byUV irradiation, using a UV Stratalinker 2400 apparatus(Stratagene). Probes were generated by random primingusing the Prime-a-Gene labelling kit (Promega, Madison,WI, USA), in the presence of a-32P dCTP, to label Spin-X-purified (Costar, Cambridge, MA, USA) cDNA fragments.The probes used to detect Arabidopsis XTH33 and PR1transcripts on Northern blots were obtained by insertingpartial cDNAs amplified by RT-PCR with gene-specificprimers (PR1, 5′GCAGAACAACTAAGAGGCAAC3′and 5′TCACTGACTTTCTCCAAACAAC3′; XTH33,5′AGCCTTTCACGACTATACCC3′ and 5′ATGACTTTGTATCTTGGCTT3′) into pCR2TOPO (Invitrogen). Pre-hybridization (4 h) and hybridization (overnight) werecarried out at 50 and 65 °C, respectively, in the bufferdescribed by Church & Gilbert (1984). The membrane wasthen washed at medium stringency [2 ¥ SSC, 0.1% sodiumdodecyl sulphate (SDS) at 37 °C for 30 min and 0.2 ¥ SSC,0.1% SDS at 55 °C for 30 min]. Blots were then placedagainst X-ray film (Kodak).

RT-PCR

Reverse transcription (RT) was performed with 1 mg totalRNA, after DNAse treatment, using the Superscript IIenzyme (Invitrogen), according to the standard protocol.RT-PCR was tested and normalized with EF-1a A4 primers.Each PCR was carried out with an equivalent amount ofcDNA, as determined on the basis of EF-1a A4 amplifica-tion product intensity. PCR was carried out with XTH33sense 5′ACCCGTCTAAACCGATGTCTCTATACG3′and antisense 5′ATGACTTTGTATCTTGGCTTATCACTG3′ primers, and EF-1aA4 sense 5′ATGCCCCAGGA-CATCGTGATTTCAT3′ and antisense 5′TTGGCGGCACCCTTAGCTGGATCA3′ primers. PCR programmeswere designed for each gene such that PCR products were

recovered during the exponential phase (40 and 20 cycles, at45 and 50 °C, respectively, for XTH33 and Ef-1a PCR prod-ucts). To control that the signal obtained from RT-PCR wasnot due to genomic DNA, control reactions were per-formed omitting the addition of reverse transcriptase to theinitial step.

In situ hybridization

The sixth petiole of celery plants was used for in situ hybrid-ization.The plants were fixed by two incubations, for 30 mineach, in 4% (v/v) formaldehyde in phosphate-bufferedsaline under vacuum, and were left in fixation buffer over-night at 4C. The fixed tissues were washed, dehydrated andembedded in paraffin, essentially as described by Jackson(1991). Paraffin sections (12 mM) were cut and attached topercolated glass slides (Dako, Buckinghamshire, UK).Anti-sense oligonucleotide probes were synthesized from celeryXTH1 oligonucleotide templates, using digoxigenin (DIG)(DIG oligonucleotide tailing kit; Roche Diagnostics, India-napolis, IN, USA), according to the manufacturer’s instruc-tions. The antisense oligonucleotide used to detect XTH1was 5′GCACTGAAAAATCCGTAATAGTA3′. A controlhybridization was carried out, using a sense oligonucleotide:5′TATCCTCTCGTGGTCCGATTTAT3′. In situ hybrid-izations with oligonucleotide probes were performedaccording to the manufacturer’s instructions, at a tempera-ture of 60 °C. DIG-labelled probes were immunodetectedusing an alkaline phosphatase-conjugated anti-DIG Fabfragment visualized with nitroblue tetrazolium and5-bromo-4-chloro-3-indolyl phosphate substrates.

Histochemical b-glucuronidase staining

b-glucuronidase (GUS) activity was assayed histochemi-cally as previously described (Jefferson et al. 1987). Sampletissues were fixed in 80% ice-cold acetone for 20 min,washed three times in water and placed under vacuum for20 min at room temperature before incubation at 37 °Covernight in the reaction buffer [1 mg mL-1 5-bromo4-chloro 3-indolyl glucuronide (X-Gluc), 100 mM potas-sium phosphate pH 7, 0.1% (v/v) Triton X 100] and in thepresence of 0.5 mM potassium ferrocyanide and potassiumferricyanide. Sample tissues were washed once in waterthen cleared in successive bathes of 70 and 96% ethanolbefore transfer to a glass slide for microscopy. Stained sec-tions were visualized with a binocular microscope. Imageswere captured with a video camera (ProgRes C10plus) andProgRes CaptureBasic version 1.2.01 software (Jenoptik,Clara Vision, Massy, France).

Cytochemical xyloglucan oligosaccharides(XGO) assay

Xyloglucan endotransglycosylase (XET) activity in celerypetioles was localized using the in situ assay based on the



incorporation of sulforhodamine-labelled xyloglucan oli-gosaccharides (XGO-SRs) described by Vissenberg et al.(2000). The XGO-SRs were kindly provided by Stefan Fry(Fry 1997). XGO-SRs were dissolved in 25 mM methane-sulphonic acid (MES) buffer, pH 5.5 (Vissenberg et al. 2000)before use. Thin hand-cut sections of celery petiolessampled from infested and non-treated plants were incu-bated in this solution in the dark for 1 h and were thenwashed for 10 min in ethanol/formic acid/water [15:1:4(v/v/v)] to remove any remaining unreacted XGO-SRs.Incubation in 5% formic acid removed apoplastic, non-wall-bound XGO-SR. Samples were then inspected under aLeica DMRB epifluorescence microscope (Leica Microsys-tems, Rueil Malmaison, France), using green (534–542 nm)excitation light. Images were taken with a Sony XC77CEcharge-coupled device (CCD) camera and analysed withOptimas 6 software (Bioscan, Edmonds, WA, USA).

Statistical analysis

All statistical analysis were performed using S-plus soft-ware for Unix, version 3.2. (MathSoft, Inc., PTC, Needham,MA, USA). For the bioassay analysis, data corresponding toadult preference were subjected to the non-parametric Wil-coxon signed-rank test. Wilcoxon signed-rank test was pre-viously used to analyse differences in colonizationpreference (Zavala et al. 2004). Data corresponding to thenymph distribution, reported to adults number, were sub-jected to the Mann–Whitney test.

RESULTS

Up-regulation of XTH1 in the phloem of celeryis specific to aphid infestation

In celery (Divol et al. 2005), infestation with M. persicae ledto a systemic phloem-specific increase in XTH1 transcriptaccumulation (Table 1). Compared to other XTHs deregu-lated by aphids (XTH3 and XTH4, Table 1), XTH1 showeda significant effect persisting 1 week after infestation (Pval-Cond 6.8 e-04, mean ratio 2.49, Table 1). Levels of mRNA forthis gene increased strongly after infestation, as shown byNorthern blots with RNA extracted from plants 3 and 7 dafter infestation (Fig. 1a). Because this induction may resultfrom early wounding due to aphid probing, the accumula-tion of XTH1 mRNA was followed in both wounded anduntreated plants to check the response of XTH1 to wound-ing. No modification on the accumulation of XTH1 tran-script was found following wounding (Fig. 1b). By contrast,MT5, which encodes in celery a metallothionein specificallyexpressed in the phloem (Vilaine et al. 2003), displayed anincrease in mRNA accumulation after wounding, similar tothat observed 3 and 7 d after aphid infestation, in thephloem of celery petioles, consistent with observations thatsome metallothionein genes respond to wounding(Snowden, Richards & Gardner 1995; Choi et al. 1996; Buttet al. 1998).

Celery XTH1 is expressed in specific cell layersin the phloem

In situ hybridization experiments were carried out onthe petioles of infested or untreated celery plants. ADIG-labelled oligonucleotide probe complementary toXTH1 mRNA was used for these experiments. In petioles,XTH1 transcripts were detected only in the phloem fromplants infested with M. persicae especially in the zone of theconducting phloem, which covered sieve elements and com-panion cells (Fig. 2a–d). XTH1 transcript accumulation wasnot detected in the petioles of uninfested plants or in thexylem and parenchyma of petioles isolated from infestedplants. This result is consistent with those for Northern blotanalysis and differential hybridization experiments (Fig. 1,Divol et al. 2005).

The up-regulation of celery XTH1 is notassociated with an increase in XET activity

The possible relationship between this increase in XTH1transcription and an increase in XET activity was

7 dpi

7 dpi3 dpi 3 dpi

XTH1

U I U I

(a)

rRNAs

3 dpi

(b)

XTH1

U I U W W U W W

MT5

rRNAs

Figure 1. Systemic accumulation of celery XTH1 transcripts inaphid infested, wounded and control plants. (a) Northern blotanalysis of 10 mg of total RNA from the phloem of infested (I) oruninfested (U) celery plants 3 and 7 d post infestation (dpi) withMyzus persicae aphids. The filter was hybridized with a probe forXTH1 derived from the cDNA clone BIP0760 (accession numberCN254566). The gel was stained with ethidium bromide andrRNA levels compared to control for equal loading. (b) Northernblot analysis of 10 mg samples of total RNA from the phloem ofinfested (I) celery plants at 3 dpi, wounded (W) celery plants, 3and 7 d after wounding, and control (U) celery plants. Each lanerepresents an independent plant. The filter was hybridized with aprobe for XTH1 derived from the cDNA clone BIP0760(accession number CN254566) and with a probe for MT5 derivedfrom the cDNA clone BIP1217 (Genbank accession numberCN254935). Below each blot, a control for equal loading, basedon the comparison of rRNAs levels on the ethidiumbromide-stained gel, is shown.

192 F. Divol et al.


investigated using the in situ assay developed by Vissenberget al. (2000). This assay is based on the incorporation byactive XET of sulforhodamine-labelled XGO-SRs intofresh tissue sections. Hand-cut thin sections of petioles frominfested and uninfested plants were assayed for XGO-SRincorporation. As previously observed in celery (Vissen-berg et al. 2000), the phloem tissue was a major site for XETactivity, with bright staining of companion cells, sieve ele-ments and phloem parenchyma cells (SupplementaryFig. S1a). However, XET activity was not significantlyhigher in the petioles of infested plants than in the petiolesof uninfested plants (Supplementary Fig. S1b). However, asseveral XTH genes are expressed in the phloem of celerypetioles, it was not possible to distinguish the individualcontributions of each of these genes to overall XET activity.

Celery XTH1 is related to the ArabidopsisXHT33 gene and to other XTH33-like genesin angiosperms

The full-length XTH coding sequence was determined byprimer extension, using a celery phloem cDNA library.XTH1 (GenBank accession number DQ204724) encodes a307-amino acid XTH with a predicted molecular weight of34.6 kDa. XTH1 contains the conserved domain thought to

function as the catalytic site for both hydrolase and trans-ferase activities (Fig. 3a) (Okazawa et al. 1993; Campbell &Braam 1999). Multiple alignments of the protein sequencesof XTH1 and the 33 XTH proteins of Arabidopsis(Yokoyama & Nishitani 2001) revealed significant aminoacid sequence identity, ranging from 33.2 to 61.3%. Aneighbour-joining tree was constructed, using the alignedpredicted sequences of Arabidopsis proteins and the pre-dicted protein sequence of XTH1. The tree was rooted ontwo additional XTHs from monocots, belonging to a dis-tinct XTH group (Campbell & Braam 1999; Rose et al.2002). The resulting cladogram indicated that XTH1belonged to the third group (Fig. 3b) and was related to theArabidopsis XTH33 (At1g10550), sharing 61.3% aminoacid sequence identity (E value: 1E-107). Little is knownabout the expression pattern of XTH33; Yokoyama &Nishitani (2001) reported that transcript levels for this genewere close to the limit of detection.

A BLAST search was carried out in plant databases, andsignificant similarity was found with the EST of theangiosperms A. officinalis, A. majus, G. max, S. tuberosum,C. sinensis and P. balsamifera (56.0–77.8% amino acid iden-tity with XTH1 and 1E-84–1E-105 E value).These ESTs wereobtained from cDNA libraries representing organs sampledin many cases in stress conditions. The frequency of theseESTs was low, suggesting that these genes were weaklyexpressed in the organs or the tissues that were tested(roots, stolons, inflorescences, developing seeds or vasculartissues). These sequences clustered with celery XTH1 andArabidopsis XTH33 (data not shown), and share commonmotifs within the glycosyl-hydrolase domain (Fig. 3a). Incontrast to some other XTHs of group 3 (Campbell &Braam 1999), they have no C-terminal extension.

Transcriptional regulation of Arabidopsis XH33by aphid infestation

The transcriptional regulation of XHT33 was evaluated byNorthern blot analysis with total RNA extracted from therosette or the stem of uninfested or infested plants, 24, 48and 72 h after infestation. XHT33 transcripts were notdetected in any of these conditions, suggesting that this genewas expressed very weakly, if at all (data not shown). Forcomparison, the transcriptional regulation of PR1(At2g14610), a gene induced by aphids within 24 h of infes-tation in Arabidopsis (Moran & Thompson 2001), wasassessed by Northern blotting on the same plant samples.PR1 was induced only in the leaves of the rosette, as soon as24 h after infestation (Fig. 4a). There was no obviousincrease in PR1 mRNA accumulation in distal organs, suchas the stem of infested plants, confirming that this gene wasnot induced systemically in response to aphid infestation, aspreviously described (Moran & Thompson 2001).

XTH33 expression was analysed by semi-quantitativeRT-PCR analysis, using gene-specific primers. Again thecDNA derived from XTH33 transcripts was not detectablein the rosette leaves or stem, neither from healthy norinfested plants (Fig. 4b). Transcripts were detected in the

Xyl

Xyl

Xyl

Xyl

Phl

PhlPhl

Phl

(a)

(b)

(c)

(d)

Con

trol

+ A

phid

Figure 2. Phloem localization of celery XTH1 mRNA. (a–d) Insitu hybridization of transverse sections of petioles fromuninfested leaves (a & c) and infested leaves (b & d) 7 d afterinfestation with Myzus persicae. (a & b) Antisense riboprobe(10 ¥). (c & d) Antisense riboprobe (40 ¥). Bars: 0.1 mm (a & b),20 mm (c & d). Phl, phloem; xyl, xylem.



stems of infested plants 72 h after infestation in someexperiments (Fig. 4b), but this response was not consis-tently detected, suggesting that the level of mRNA accumu-lation in these conditions was close to the limit of detection.Because the level of XTH33 mRNA accumulation wasapparently not sufficient for detection using standard pro-cedures, an alternative approach was developed to assessXTH33 expression.

Pattern of transcription of Arabidopsis XTH33

The regulation of XHT33 expression was analysed usingtransgenic lines expressing a transcriptional fusion in whichthe XHT33 promoter drove expression of the UidAreporter gene. Several independent transgenic lines pre-senting consistent b-glucuronidase (GUS) activity patternswere obtained. GUS activity was analysed further on theprogeny of two homozygous lines. GUS activity was weak inall tissues and organs of these plants (Fig. 5). Most expres-sion was observed in the vegetative rosette on plants up to3 week old.

Expression appeared to be tightly controlled duringdevelopment. Early expression was observed after germi-nation, in the cotyledons and epicotyl (Fig. 5a), withoutrestriction to the vascular tissues. In the first true leaves,

GUS activity was observed for 2–3 weeks, decreasingrapidly thereafter (Fig. 5b). In older plantlets (4–5 weeksold), GUS activity was no longer detected in these organs(Fig. 5c), but was still detected in expanding leaves. Weakactivity was observed in the stem and stem leaves, andlater on in flowers, including petals, stamens and carpel.Interestingly, GUS activity was not observed in emergingorgans, such as newly formed leaves, but was insteaddetected later, suggesting that GUS activity is associatedwith elongation rather than cell division. This pattern isillustrated by the GUS activity observed in a series of leavesof the rosette, from the expansion stage to the fully expandedstage (Fig. 5d–f), with activity progressively regressing fromthe middle of the leaf to the leaf margin. The strong devel-opmental control of XTH33 expression was also revealed bythe pattern of GUS activity in flowers of different ages(Fig. 5i–k). Early activity was observed in the anthers andcarpels. This activity gradually disappeared, to be replacedby strong activity in the elongating filaments of stamens.Interestingly, no activity was observed in the roots, in plantsgrown in vitro or in the greenhouse, with the exception ofsome secondary roots, in which GUS activity was occasion-ally observed in the root cap (Fig. 5l).This pattern of activitywas not significantly affected by stress responses, such aswounding or virus infection (data not shown).

0.1

XTH11

XTH3

XTH1

XTH2

XTH31

XTH32

AgXTH1

XTH33XTH27

XTH28

XTH29

XTH30

HvPM2

HvPM5

XTH26

XTH14XTH12XTH13

XTH15XTH16

XTH25XTH21XTH20XTH19

XTH17XTH18 XTH24

XTH22XTH23

XTH10XTH6 XTH7

XTH8

XTH4XTH5XTH9

Group 1

Group 3

Group 2

Group 4

(b)(a)

Figure 3. Cladogram and coding sequence indicating structural similarities between celery XTH1 from celery and other plantxyloglucan endotransglycosylase/hydrolase (XTH) enzymes. (a) Multiple alignment of celery XTH1 with the most similar plant XTH. Thebox underlined in green represents the presumed catalytic site for hydrolase and transferase activities. The motifs underlined in red areshared by XTH33 and celery XTH1, and these specific sequences and are not found in other Arabidopsis XTHs. Two motifs within theglycosyl-hydrolase domain, – +146-TGREEKFY(L/P)WFDPT +159 and +175-VFLVDN(I/V)P(V/I)R-+184 – were not found in otherArabidopsis XTH. (b) Cladogram of predicted sequences of Arabidopsis proteins and the predicted protein sequence of celery XTH1.The three main groups of XTH are indicated. Two barley sequences, PM5 (X93173) and PM2 (X91660), representing a fourth subgrouppresent only in monocots, were used to root the cladogram. At, Arabidopsis thaliana; Ao, Asparagus officinalis; Cs, Citrus sinensis; Gm,Glycine max.

194 F. Divol et al.


Vascular pattern and aphid response ondetached leaves

The expression pattern was also analysed on plants infestedwith aphids.The overall pattern of expression was not modi-fied. Nevertheless, a complementary assay was carried outwith detached leaves stored in tubes for 48 h. In contrastwith what was observed for whole plants, GUS activityappeared to be stronger in the vascular tissues in detachedleaves (Fig. 5g–h). This effect was even more marked indetached leaves infested with aphids for 48 h. This suggeststhat pXTH33 responds to leaf cutting primarily in the vas-cular tissues, and that this effect is stronger in the presenceof aphids.

Increased preference of aphids to xth33mutant line

An xth33 mutant was obtained from the Salk Institute col-lection.This mutant had a T-DNA insertion in the third exon(xth33-1) of the gene (Supplementary Fig. S2). Plants

homozygous for the insertion in the XTH33 gene wereobtained for further characterization. No visual phenotypicalteration was observed on the plant architecture or organmorphology in the xth33 mutant line, either in vitro or in thegreenhouse. Fertility was normal, and the growth and devel-opment of the plants were identical to those of controlplants.

The involvement of XTH33 in plant–aphid interactionswas analysed by studying the response of an xth33 mutantline to aphid infestation. Because XTH33 is expressed invegetative rosette of plants aged up to 3 weeks old,3-week-old plants were used for bioassays with M. persicaebased on the choice test described by Cabrera y Poch et al.(1998). More adults and more nymphs were found onxth33-1 (Fig. 6a) mutant plants than on wild-type Col-0plants, indicating greater preference of aphids for themutant (63% of the aphids settled preferentially on thexth33-1; xth33-1 versus wild type: Wilcoxon signed-ranktest for adults, P < 0.0001; Mann–Whitney test for nymphs,P � 0.02). A similar bioassay was carried out with a SalkInstitute Collection mutant carrying a T-DNA insertion inthe PP2-A1 gene, to rule out the possibility of this effectbeing linked to other features of the T-DNA, such asexpression of the resistance marker. No significant effectwas observed in this line (Fig. 6b), indicating that the pref-erence of aphids for the xth33 mutants was linked tochanges in the expression of XTH33 rather than to thepresence of the T-DNA in these lines (pp2-A1 versus wildtype: Wilcoxon signed-rank test, P � 0.72; Mann–Whitneytest for nymphs, P � 0.74).

Ectopic expression of Arabidopsis XTH33 incompanion cells does not confer significantprotection against aphids

Transgenic pPP2-XTH33 lines displaying ectopic expres-sion of XTH33 were produced. In these lines, the XTH33coding region was fused to the phloem-specific AtPP2-A1promoter, which drives expression in the sieve element/companion cell complex (SE/CCC). These lines differed intheir degree of XTH33 up-regulation (SupplementaryTable S1). They were used for the bioassay describedearlier. Aphids, as deduced from the number of adults andnymphs found on each plant, did not prefer wild-type plantsover pPP2-XTH33 plants (Fig. 6c), regardless of theamount of XTH33 transcript accumulation (pp2-xth versuswild type: Wilcoxon signed-rank test for adults, P � 0.5;Mann–Whitney test for nymphs, P � 0.76).

DISCUSSION

Phloem sap-feeding insects, such as aphids and whiteflies,establish feeding sites in the sieve elements. The small pro-portion of phloem in the organs (Sjölund 1997) makes itdifficult to investigate events taking place in the phloem.Such studies require the development of specific tools, suchas the use of phloem-enriched fractions or phloem sap. The

U I U I U I24 h

Local

PR1

PR1

rRNAs

rRNAs

System

ic

XTH33

EF1a

System

ic

48 hU I U I

(a)

(b)

72 h48 h

72 h

Figure 4. Local and systemic accumulation of ArabidopsisXTH33 and PR1 transcripts in response to aphid infestation. (a)Northern blot analysis of 10 mg of total RNA from rosette leaves(local) or cauline leaves (systemic) of Arabidopsis Columbia-0(Col-0) infested (I) or uninfested (U). The filter was hybridizedwith a PR1 probe for positive control of local induction afterinfestation. XTH33 expression was not detected in Northern blotexperiments (not shown). The gel was stained with ethidiumbromide, and rRNA levels were compared to control for equalloading. (b) RT-PCR analysis of XTH33 mRNA levels inArabidopsis stems 48 and 72 h after infestation with Myzuspersicae aphids (I) and in control uninfested plants (U),respectively. XTH33 expression could be detected only after 72 hin the stems of infested plant. The mRNA concentration of allsamples was equal, as confirmed by RT-PCR with EF1a primers.



response to an infestation with the generalist phloem-feeding aphid M. persicae was therefore analysed inphloem-enriched tissues from A. graveolens, to investigatethe response in the primary target of aphids, the phloem(Divol et al. 2005). Infestation triggered the systemicup-regulation of several genes, including several associatedwith cell wall remodelling. Genes encoding XTHs con-stituted the main class of genes up-regulated in theseconditions.

Pattern of celery XTH1 expression in responseto aphid feeding

Four of the six XTH genes expressed in celery phloem weresystemically induced by aphids, 3 or 7 d after infestation, orat both time points (Table 1).The largest increase in mRNAaccumulation was observed for XTH1, and this effect per-sisted for at least 1 week after infestation. It was not corre-lated with an increase in XET activity. Nevertheless, an

(a) (b) (c)

(d) (e) (f)

(i) (j) (k) (l)

(g) (h)

Figure 5. Histochemical localization of b-glucuronidase activity in various organs from transgenic Arabidopsis plants expressing thechimeric pXTH33-GUS gene. Plants were grown in different conditions (in vitro for younger stages or in greenhouse for older plants) andb-glucuronidase activity was assayed in whole plants or detached organs. (a) Fifteen-day-old plantlets grown in axenic culture, (b)30-day-old plantlets grown in axenic culture, (c) 45-day-old plantlets grown in axenic culture. (d–f) Details of leaves observed on 3- to6-week-old plantlets grown in axenic culture. (g–h) Detached leaves observed 48 h after cutting with (h) or without (g) aphids. During theassay, individual detached leaves were maintained for 48 h in tubes, with the petiole kept in water. Plants or leaves were infested witheight adult aphids per leaf. (i–k) Details of flowers observed on 10- to 12-week-old plants grown in the greenhouse. (l) Details of a lateralroot from an 8-week-old plant. Bars: (a–i) 2 mm, (j–k) 0.5 mm and (l) 0.2 mm.

196 F. Divol et al.


increase in XET activity resulting from XTH1 expressioncould have been masked by the activity of the other XTHs,because at least five other XTH genes are expressed in thistissue, leading to the high XET activity in the phloemobserved in the absence of stress, in this work, and inanother report (Vissenberg et al. 2000). Alternatively,XTH1 can be involved in the hydrolysis of xyloglucansrather than in their rearrangement. Such an activity canplay a role in signalling, via the release of xyloglucans intosieve elements.

The specificity of celery XTH1 expression was also inves-tigated. As aphids puncture several cell layers during earlyprobing, triggering some wound responses, the response ofXTH1 to wounding was assessed. Wounding did not modifythe expression of this gene, in contrast to a gene expressedin the phloem and a non-specific marker of stress, encodingthe metallothionein MT5 (Vilaine et al. 2003). The expres-sion of XTH1, specific to the phloem tissue, was restrictedto phloem transport cells, as determined by in situ hybrid-ization. Both observations suggest that XTH1 participatesto the plant response to aphids, although the lack ofincrease in XET activity does not support the hypothesisthat XTH1 induces significant cell wall modifications.

Transcriptional regulation ofArabidopsis XTH33

Based on sequence similarity, the closest of celery XTH1,XTH33 was analysed. RT-PCR analysis and Northern blotanalysis suggested a low level of expression for XTH33.These results were consistent with a survey of the transcrip-tional regulation of the 33 Arabidopsis XTH genes based onreal-time RT-PCR analysis (Yokoyama & Nishitani 2001)and data available on Genevestigator (Zimmermann et al.2004), both indicating a low steady-state level of transcrip-tion for this gene. It was necessary to use transgenic linesharbouring a transcriptional fusion between the promoterregion of XTH33 and the UidA gene (b-glucuronidase) tostudy XTH33 transcriptional regulation.

xth33-1

WT

Adults: 60%

Nymphs: 65%

0

5

10

15

20

25

30

0 5 10 15 20 25 30

(a)

pPP2-XTH

WT

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Adults: 50%

Nymphs: 50%

(c)

pp2-a1-1

WT

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Adults: 49%Nymphs: 52%

(b)

Figure 6. In vitro assay of aphid preference on an xth33 mutantline, ectopic pPP2-XTH33 lines and the wild-type (WT) line[Columbus-0 (Col-0)]. For each assay (i.e. one individual Petridish with three wild-type and three mutant plants, and onaverage, 20 apterous adults), each dot represents the totalnumber of aphids obtained on mutant plants (on the x-axis) andon wild-type plants (on the y-axis). The numbers at the top andon the right represent the percentage of aphids (adults ornymphs) on the mutant line versus the total number of aphids(on mutant and wild-type lines). Dots below the median linecorresponds to experiments in which mutant line plants weremore attractive than wild-type plants, and dots above the medianline correspond to experiments in which wild-type plants weremore attractive than mutant plants. The trend observed forxth33-1 (most dots below the median) was confirmed by thehigher mean numbers of adults or nymphs on mutants than onthe wild type. (a) xth33-1 mutant line versus wild type; (b) pp2-a1mutant line versus wild type; (c) pPP2-XTH33 transgenic linesversus wild type. �, adults; , nymphs.



These studies indicate that XTH33 expression was tightlyregulated during development and was mostly associated tovarious organs, such as leaves, petals, stamens or carpels,undergoing expansion, despite an atypical pattern in theleaf margins. XTH33 expression was not affected by wound-ing or aphid infestation. These observations support a rolein cell wall remodelling during cell expansion rather than inoligosaccharide signalling. Unlike celery XTH1, Arabidop-sis XTH33 was not expressed specifically in the vasculartissues, although some preferential expression was observedin the veins of detached rosette leaves, 48 h after detach-ment. Interestingly, neither celery XTH1 nor ArabidopsisXTH33 responded to wounding, whereas only celery XTH1responded unambiguously to aphids.

Transcription regulation of XTH1, XTH33 andother XTHs in response to aphid infestation

The eudicot celery and Arabidopsis belong to distant taxa,Asterids and Eurosids II. XTH1 and XTH33 share signifi-cant identity (61.3%) and both belong, with otherangiosperms sequences, to an homogeneous subgroup ofXTH.We observed that the transcription of these genes wasdifferent, at the tissue level, and in response to woundingand aphid infestation. This clearly indicates a divergenceeither in the regulation of the pathway or in the pathwaysthemselves involved in the response in celery and Arabi-dopsis to M. persicae. Because some processes, such as anorganism’s response to pathogens and environmental stress,are prone to divergence between organisms (Lespinet et al.2002), the role of celery XTH1 and Arabidopsis XTH33 inthe mechanisms of aphid response may also have diverged,despite sequence similarity. Besides XTH1 and XTH33,very few XTHs were shown previously deregulated byaphids: in Arabidopsis, XTH22 (former TCH4) in infectedleaves in response to M. persicae (Moran et al. 2002),whereas XTH22 is also induced by wounding (Cheong et al.2002); in Beta vulgaris, BvXTH1 and BvXTH2 (group 2)up-regulated by wounding and in response to M. persicae, atthe feeding site (Dimmer et al. 2004); in celery, two otherXTHs, from group1, showed a systemic response to aphids(Divol et al. 2005). These observations do not predictwhether this up-regulation benefits the plant defencemechanisms or the aphid settlement. Other tools, such asreverse genetics, were therefore necessary to define thecontribution of XTH, and subsequently of cell wall modifi-cations, in plant–aphid interaction.

Arabidopsis XTH33 and plant responseto aphids

A mutant line with disrupted XTH33 gene was analysed. Itsphenotype was indistinguishable from that of wild-typeCol-0 plants, and the preference of the aphid M. persicae forxth33-1 or wild-type genotype was investigated. The choicetest was performed on 3-week-old plants, at a stage corre-sponding to constitutive expression in the leaves on which

the aphids settled. A significant increased preference ofaphids was observed compared with wild-type plants, sug-gesting a role for XTH33 in the protection of the plantagainst aphids rather than in the weakening of natural bar-riers against insects for the aphid benefit. The tightly devel-opmentally regulated pattern of expression of XTH33supports the hypothesis of a role in cell wall modifications;this provides an indirect evidence that a modification of cellwall properties affect plant–aphid interactions. This obser-vation is consistent with the fact that pectinase and b-1,3-glucanase were found in aphid saliva, an observation thatled the authors to speculate that the behaviour of aphids isassociated with the composition of the cell wall (Campbell& Dreyer 1985; Dreyer & Campbell 1987). More recentobservations (Caillaud & Niemeyer 1996) describe thathost acceptance or rejection is associated with the phloem-sealing system of the plant and establish the role of phloem-specific barriers involved in plant–aphid interaction.

In celery, a few XTHs are deregulated following aphidinfestation, and possibly, these enzymes are activated invarious combinations depending on tissues, local or long-distance response or time schedule of the response. Theymay have additive effects in plant defence machinery. Insuch a case, it is predictable that a more dramatic biologicaleffect on aphid preference can be achieved by affectingthem simultaneously, given that manipulation of multipleXTHs, key cell wall remodelling enzymes, does not affectplant development. More physiological, biochemical andgenetic studies should contribute in the future to clarify theexact contribution of cell wall remodelling enzymes inplant–aphid interaction.

Studies of various mutants with impaired defence path-ways or senescence have demonstrated the complexinvolvement of salicylic acid-, jasmonic acid-, ethylene-dependent signalling pathways (Moran & Thompson 2001;Ellis, Karafyllidis & Turner 2002; Mewis et al. 2005; Pegada-raju et al. 2005), but little attention has been paid to otherclasses of mutants. As defence pathways are often alsoinvolved in developmental processes, it is uncertainwhether they can be manipulated to improve crop protec-tion against the generalist M. persicae. The results obtainedwith a mutant affected in a gene possibly involved in cellwall remodelling suggest that alternative strategies can beexplored, such as exploitation of collections of mutantsaffected in cell wall composition (Mouille et al. 2003).

Ectopic expression of Arabidopsis XTH33 inthe phloem is not sufficient to conferprotection against aphids

The increase in aphid preference for mutant plants withaltered XTH33 expression may result from a direct effect ofa lack of XTH activity, resulting in modification of the prop-erties of the cell wall, in particular in the phloem, or from anas yet unknown indirect effect. Because XTH33 isexpressed in all cell types, it was not possible to concludewhether the preference resulted from a defect in the super-ficial cell layers where aphids settle, or from a defect in the

198 F. Divol et al.


phloem sieve tubes where aphids feed on. Transgenic linesexpressing XTH33 ectopically in the SE/CCC of thephloem were generated and tested in the aphid choiceassay. No significant effect on aphid behaviour wasobserved in 3-week-old plants grown in vitro. Thus, alteringXTH33 expression in SE/CCC appears not sufficient toincrease protection against aphids. This may be because theaccumulation of XTH33 is not a limiting factor in this tissue,in terms of plant–aphid interactions. Alternatively, theprimary effect of XTH33 may not be in the SE/CCC, but inother cell layers. Finally, we cannot exclude the possibilitythat total XTH activity in the phloem or in other tissues istightly controlled and that the ectopic expression of a singlegene is not sufficient to affect XTH activity. These findingsconfirm the difficulty of manipulating plant–aphid interac-tions, given the complexity of the physiological and devel-opmental modifications occurring in plants during aphidinfestation (Girousse et al. 2003, 2005). In addition, it hasbeen shown that some protection mechanisms againstaphids are triggered after the induction of a phloemresponse peaking after 48 h (Sauge et al. 2006). In the com-patible interaction between Arabidopsis and M. persicae, itwould be of interest to determine whether the plantresponse involves delayed enhanced susceptibility or pro-tection, as described in various genotypes of peach (Saugeet al. 2006). Such studies would probably provide a betterframework for assessing the behaviour of the generalistaphid M. persicae and a more conclusive analysis of theprotection mechanisms operating in this system.

ACKNOWLEDGMENTS

We thank Dr. S. Fry for generously providing XGO andXGO-SR reagents, Dr. R. Delorme for providing the M.persicae clone and M. Castelain for providing seeds of ahomozygous pp2a1 mutant line.We also thank Dr.V. Braultand J.-P. Lacroze for carrying out additional experiments onxth33 mutant. We are grateful to H. Morin and the labora-tory of cytology of Versailles for helpful discussions for insitu hybridization experiments. We thank J.-P. Meunier, H.Ferry and K. Goffron for greenhouse facilities, Dr. A. Ver-meulen and A. de Courcel (Vilmorin Clause & Cie) fortheir support in this project and Dr. J.-C. Palauqui for dis-cussions. The project was supported by grants from Géno-plante (CI2000 005) and Vilmorin Clause & Cie.

REFERENCES

Alonso J.M., Stepanova A.N., Leisse T.J., et al. (2003) Genome-wide insertional mutagenesis of Arabidopsis thaliana. Science301, 653–657.

Altschul S.F., Madden T.L., Schäffer A.A., Zhang J., Zhang Z.,Miller W. & Lipman D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.Nucleic Acids Research 25, 3389–3402.

Asano T., Masumura T., Kusano H., Kikuchi S., Kurita A., ShimadaH. & Kadowaki K. (2002) Construction of a specialized cDNAlibrary from plant cells isolated by laser capture microdissection:

toward comprehensive analysis of the genes expressed in the ricephloem. Plant Journal 32, 401–408.

van Bel A.J.E. (2003) The phloem, a miracle of ingenuity. Plant,Cell & Environment 26, 125–149.

van Bel A.J.E. & Gaupels F. (2004) Pathogen-induced resistanceand alarm signals in the phloem. Molecular Plant Pathology 5,495–504.

Bevan M. (1984) Binary Agrobacterium vectors for plant transfor-mation. Nucleic Acids Research 12, 8711–8721.

Burstin J. (2000) Differential expression of two barley XET-relatedgenes during coleoptile growth. Journal of Experimental Botany51, 847–852.

Butt A., Mousley C., Morris K., Beynon J., Can C., Holub E.,Greenberg J.T. & Buchanan-Wollaston V. (1998) Differentialexpression of a senescence-enhanced metallothionein gene inArabidopsis in response to isolates of Peronospora parasiticaand Pseudomonas syringae. Plant Journal 16, 209–221.

Cabrera y Poch H.L., Ponz F. & Fereres A. (1998) Searching forresistance in Arabidopsis thaliana to the green peach aphidMyzus persicae. Plant Science 138, 209–216.

Caillaud C.M. & Niemeyer H.M. (1996) Possible involvement ofthe phloem sealing system in the acceptance of a plant as host byan aphid. Experientia 52, 927–931.

Campbell B. & Dreyer D.L. (1985) Host-plant resistance ofsorghum: differential hydrolysis of sorghum pectic substances bypolysaccharases of greenbug biotypes (Schizaphis graminum,Homoptera Aphididae). Archives of Insect Biochemistry andPhysiology 2, 203–215.

Campbell P. & Braam J. (1999) Xyloglucan endotransglycosylases:diversity of genes, enzymes and potential wall-modifying func-tions. Trends in Plant Science 4, 361–366.

Cheong Y.H., Chang H.S., Gupta R., Wang X., Zhu T. & Luan S.(2002) Transcriptional profiling reveals novel interactionsbetween wounding, pathogen, abiotic stress, and hormonalresponses in Arabidopsis. Plant Physiology 129, 661–677.

Choi D., Kim H.M., Yun H.K., Park J.A., Kim W.T. & Bok S.H.(1996) Molecular cloning of a metallothionein-like gene fromNicotiana glutinosa L. and its induction by wounding andtobacco mosaic virus infection. Plant Physiology 112, 353–359.

Church G.M. & Gilbert W. (1984) Genomic sequencing. Proceed-ings of the National Academy of Sciences of the USA 81, 1991–1995.

Clough S.J. & Bent A.F. (1998) Floral dip: a simplified methodfor Agrobacterium-mediated transformation of Arabidopsisthaliana. Plant Journal 16, 735–743.

Daie J. (1987) Sucrose uptake in isolated phloem of celery is asingle saturable transport system. Planta 177, 474–482.

Dimmer E., Roden L., Cai D., Kingsnorth C. & Mutasa-GöttgensE. (2004) Transgenic analysis of sugar beet xyloglucan endo-transglucosylase/hydrolase Bv-XTH1 and Bv-XTH2 promotersreveals overlapping tissue-specific and wound-inducible expres-sion profiles. Plant Biotechnology Journal 2, 127–139.

Divol F., Vilaine F., Thibivilliers S., Amselem J., Palauqui J.C.,Kusiak C. & Dinant S. (2005) Systemic response to aphid infes-tation by Myzus persicae in the phloem of Apium graveolens.Plant Molecular Biology 57, 517–540.

Douglas A.E. (2006) Phloem-sap feeding by animals: problems andsolutions. Journal of Experimental Botany 57, 747–754.

Dreyer D.L. & Campbell B.C. (1987) Chemical basis of host-plantresistance to aphids. Plant, Cell & Environment 10, 353–361.

Ellis C., Karafyllidis I. & Turner J.G. (2002) Constitutive activationof jasmonate signaling in an Arabidopsis mutant correlates withenhanced resistance to Erysiphe cichoracearum, Pseudomonassyringae, and Myzus Persicae. Molecular Plant-Microbe Interac-tions 15, 1025–1030.

Esquerre-Tugaye M.T., Boudart G. & Dumas B. (2000) Cell wall



degrading enzymes, inhibitory proteins, and oligosaccharidesparticipate in the molecular dialogue between plants and patho-gens. Plant Physiology and Biochemistry 38, 157–163.

Fanutti C., Gidley M.J. & Reid J.S.G. (1993) Action of a purexyloglucan endo-transglycosylase [formerly called xyloglucan-specific endo-(1–4)-ß-d-glucanase] from the cotyledons of ger-minated nasturtium seeds. Plant Journal 3, 691–700.

Fry S.C. (1989) Cellulases, hemicelluloses and auxin-stimulatedgrowth: a possible relationship. Physiologia Plantarum 75, 532–536.

Fry S.C. (1992) Xyloglucan endotransglycosylase, a new wall-loosening enzyme activity from plants. Biochemical Journal 282,821–828.

Fry S.C. (1997) Novel ‘dot-blot’ assays for glycosyltransferases andglycosylhydrolases: optimization for xyloglucan endotransglyco-sylase (XET) activity. Plant Journal 11, 1141–1150.

Gatehouse A.M.R. (2002) Plant resistance towards insect herbi-vores: a dynamic interaction. New Phytologist 156, 145–169.

Giavalisco P., Kapitza K., Kolasa A., Buhtz A. & Kehr J. (2006)Towards the proteome of Brassica napus phloem sap. Proteomics6, 896–909.

Girousse C., Faucher M., Kleinpeter C. & Bonnemain J.L. (2003)Dissection of the effects of the pea aphid Acyrthosiphon pisumfeeding on assimilate partitioning in Medicago sativa. New Phy-tologist 157, 83–92.

Girousse C., Moulia B., Silk W. & Bonnemain J.L. (2005) Aphidinfestation causes different changes in carbon and nitrogen allo-cation in alfalfa stems as well as different inhibitions of longitu-dinal and radial expansion. Plant Physiology 137, 1474–1484.

Heidel A.J. & Baldwin I.T. (2004) Microarray analysis of salicylicacid and jasmonic acid-signalling in responses of Nicotianaattenuata to attack by insects from multiple feeding guilds. Plant,Cell & Environment 27, 1362–1373.

Hyodo H., Yamakawa S., Takeda Y., Tsuduki M., Yokota A., Nishi-tani K. & Kohci T. (2003) Active gene expression of a xyloglucanendotransglucosylase/hydrolase gene, XTH9, in inflorescenceapices is related to cell elongation in Arabidopsis thaliana. PlantMolecular Biology 52, 473–482.

Jackson D.P. (1991) In situ hybridization in plants. In MolecularPlant Pathology:A Practical Approach (eds D.J. Bowles, S.J. Gurr& M. McPhereson), pp. 163–174. Oxford University Press,Oxford, UK.

Jefferson R.A., Kavanagh T.A. & Bevan M.W. (1987) GUS fusions:beta-glucuronidase as a sensitive and versatile gene fusionmarker in higher plants. EMBO Journal 6, 3901–3907.

Kaloshian I. & Walling L.L. (2005) Hemipterans as plant patho-gens. Annual Review of Phytopathology 43, 491–521.

Koncz C. & Schell J. (1986) The promoter of TL-DNA gene 5controls the tissue specific expression of chimeric genes carriedby a novel type of Agrobacterium binary vector. MolecularGeneral Genetics 204, 383–396.

Lespinet O., Wolf Y.I., Koonin E.V. & Aravind L. (2002) The roleof lineage-specific gene family expansion in the evolution ofeukaryotes. Genome Research 12, 1048–1059.

Macedo T.B., Bastos C.S., Higley L.G., Ostlie K.R. & Madhavan S.(2003) Photosynthetic responses of soybean to soybean aphid(Homoptera: Aphididae) injury. Journal of Economic Entomol-ogy 96, 188–193.

Matsui A., Yokoyama R., Seki M., Ito T., Shinozaki K., TakahashiT., Komeda Y. & Nishitani K. (2005) AtXTH29 plays an essentialrole in cell wall modification during the development of trache-ary elements. Plant Journal 42, 525–534.

Mewis I., Appel H.M., Hom A., Raina R. & Schultz J.C. (2005)Major signalling pathways modulate Arabidopsis glucosinolateaccumulation and response to both phloem-feeding and chewinginsects. Plant Physiology 138, 1149–1162.

Miles P.W. (1999) Aphid saliva. Biological Reviews 74, 41–85.Moran P.J. & Thompson G.A. (2001) Molecular responses to aphid

feeding in Arabidopsis in relation to plant defense pathways.Plant Physiology 125, 1074–1085.

Moran P.J., Cheng Y., Cassell J.L. & Thompson G.A. (2002) Geneexpression profiling of Arabidopsis thaliana in compatible plant–aphid interactions. Archives of Insect Biochemistry and Physiol-ogy 51, 182–203.

Mouille G., Robin S., Lecomte M., Pagant S. & Hofte H. (2003)Classification and identification of Arabidopsis cell wall mutantsusing Fourier-transform infrared (FT-IR) microspectroscopy.Plant Journal 35, 393–404.

Nakazono M., Qiu F., Borsuk L.A. & Schnable P.S. (2003) Laser-capture microdissection, a tool for the global analysis of geneexpression in specific plant cell types: identification of genesexpressed differentially in epidermal cells or vascular tissues ofmaize. Plant Cell 15, 583–596.

Nishitani K. & Tominaga R. (1992) Endoxyloglucan transferase, anovel class of glycosyltransferase that catalyzes transfer of asegment of xyloglucan molecule to another xyloglucan molecule.Journal of Biological Chemistry 267, 21058–21064.

Okazawa K., Sato Y., Nakagawa T., Asada K., Kato I., Tomita E. &Nishitani K. (1993) Molecular cloning and cDNA sequencing ofendoxyloglucan transferase, a novel class of glycosyltransferasethat mediates molecular grafting between matrix polysaccha-rides in plant cell walls. Journal of Biological Chemistry 268,25364–25368.

Oparka K.J. & Turgeon R. (1999) Sieve elements and companioncells – traffic control centers of the phloem. Plant Cell 11, 739–750.

Page R.D.M. (1996) TreeView: an application to display phyloge-netic tree on personal computers. Computer applications. Bio-sciences 12, 357–358.

Park S.J., Huang Y. & Ayoubi P. (2006) Identification of expressionprofiles of sorghum genes in response to greenbug phloem-feeding using cDNA subtraction and microarray analysis. Planta223, 932–947.

Pautot V., Holzer F.M. & Walling L.L. (1991) Differential expres-sion of tomato proteinase inhibitor I and II gene22s duringbacterial pathogen invasion and wounding. Molecular Plant-Microbe Interaction 4, 284–292.

Pegadaraju V., Knepper C., Reesse J. & Shah J. (2005) Prematureleaf senescence modulated by Arabidopsis phytoalexin deficient4gene is associated with defense against the phloem-feedinggreen peach aphid. Plant Physiology 139, 1927–1934.

Qubbaj T., Reineke A. & Zebitz C.P.W. (2005) Molecular interac-tions between rosy apple aphids, Dysaphis plantaginea, and resis-tant and susceptible cultivars of its primary host Malusdomestica. Entomologia Experimentalis et Applicata 115, 145–152.

Romo S., Jimenez T., Labrador E. & Dopico B. (2005) The gene fora xylogluan endotransglucosylase/hydrolase from Cicer arieti-num is strongly expressed in elongating tissues. Plant Physiologyand Biochemistry 43, 169–172.

Rose J.K.C., Braam J., Fry S.C. & Nishitani K. (2002) The XHTfamily of enzymes involved in xyloglucan endotransglucosyla-tion and endohydrolysis: current perspectives and a new unifyingnomenclature. Plant and Cell Physiology 43, 1421–1435.

Ruiz-Medrano R., Xoconostle-Cazares, B. & Lucas W.J. (2001) Thephloem as a conduit for inter-organ communication. CurrentOpinion in Plant Biology 4, 202–209.

Saitu N. & Nei M. (1987) The neighbor-joining method: a newmethod for reconstructing phylogenetic trees. Molecular BiologyEvolution 4, 406–425.

Sauge M.H., Mus F., Lacroze J.P., Pascal T., Kervella J. & PoësselJ.M. (2006) Genotypic variation in induced resistance and

200 F. Divol et al.


induced susceptibility in the peach-Myzus persicae aphid system.OIKOS 113, 305–313.

Shin Y.K., Yum H., Kim E.S., Cho H., Gothandam K.M., Hyun J. &Chung Y.Y. (2005) BcXTH1, a Brassica campestris homologue ofArabidopsis XTH9, is associated with cell expansion. Planta 2,1–10.

de Silva J., Jarman C.D., Arrowsmith D.A., Stronach M.S., Chen-gappa S., Sidebottom C. & Reid J.S.G. (1993) Molecular charac-terization of a xyloglucan-specific endo-(1→4)-ß-d-glucanase(xyloglucan endo-transglycosylase) from nasturtium seeds. PlantJournal 3, 701–711.

Sjölund R.D. (1997) The phloem sieve element: a river runsthrough it. Plant Cell 9, 1137–1146.

Snowden K.C., Richards K.D. & Gardner T.C. (1995) Aluminum-induced genes – induction by toxic metals, low calcium andwounding and pattern of expression in root tips. Plant Physiol-ogy 107, 341–348.

Tabuchi A., Mori H., Kamisaka S. & Hoson T. (2001) A new type ofendo-xyloglucan transferase devoted to xyloglucan hydrolysis inthe cell wall of azuki bean epicotyls. Plant and Cell Physiology42, 154–161.

Thompson G.A. & Goggin F.L. (2006) Transcriptomics and func-tional genomics of plant defence induction by phloem-feedinginsects. Journal of Experimental Botany 57, 755–766.

Thompson J.D., Higgins D.G. & Gibson T.J. (1994) CLUSTAL W:improving the sensitivity of progressive multiple sequence align-ment through sequence weighting, position-specific gap penal-ties and weight matrix choice. Nucleic Acids Research 22,4673–4680.

Tjallingii W.F. (2006) Salivary secretions by aphids interacting withproteins of phloem wound responses. Journal of ExperimentalBotany 57, 739–745.

Verwoerd T.C., Dekker M.M. & Hoekema A. (1989) A small-scaleprocedure for the rapid isolation of plant RNAs. Nucleic AcidsResearch 17, 2362.

Vilaine F., Palauqui J.C., Amselem J., Lemoine Kusiak C.R., &Dinant S. (2003) Toward phloem deciphering: a transcriptomeanalysis of the phloem of Apium graveolens. Plant Journal 36,67–81.

Vissenberg K., Martinez-Vilchez I.M., Verbelen J.P., Miller J.G. &Fry S.C. (2000) In vivo colocalization of xyloglucan endotransg-lycosylase activity and its donor substrate in the elongation zoneof Arabidopsis roots. Plant Cell 12, 1229–1237.

Vissenberg K., van Sandt V., Fry S.C. & Verbelen J.P. (2003) Xylo-glucan endotransglucosylase action is high in the root elongationzone and in the trichoblasts of all vascular plants fromSelaginella to Zea mays. Journal of Experimental Botany 54,335–344.

Vissenberg K., Fry S.C., Pauly M., Hofte H. & Verbelen J.P. (2005)XTH acts at the microfibril-matrix interface during cell elonga-tion. Journal of Experimental Botany 56, 673–683.

Voelckel C., Weisser W.W. & Baldwin I.T. (2004) An analysis ofplant–aphid interactions by different microarray hybridizationstrategies. Molecular Ecology 13, 3187–3195.

Vorwerk S., Somerville S. & Somerville C. (2004) The role of plantcell wall polysaccharide composition in disease resistance.Trends in Plant Science 9, 203–209.

de Vos M., Van Oosten V.R., Van Poecke R.M., et al. (2005) Signalsignature and transcriptome changes of Arabidopsis duringpathogen and insect attack. Molecular Plant-Microbe Interac-tions 18, 923–927.

Walz C., Giavalisco P., Schad M., Juenger M., Klose J. & Kehr J.(2004) Proteomics of curcurbit phloem exudate reveals anetwork of defence proteins. Phytochemistry 65, 1795–1804.

Will T. & van Bel A.J.E. (2006) Physical and chemical interactionsbetween aphids and plants. Journal of Experimental Botany 57,729–737.

Wu Y., Jeong B.R., Fry S.C. & Boyer J.S. (2005) Change in XETactivities, cell wall extensibility and hypocotyl elongationof soybean seedlings at low water potential. Planta 220, 593–601.

Yokoyama R. & Nishitani K. (2001) A comprehensive expressionof all members of a gene family encoding cell-wall enzymesallowed us to predict cis-regulatory regions involved in cell-wallconstruction in specific organs of Arabidopsis. Plant and CellPhysiology 42, 1025–1033.

Zavala J.A., Patankar A.G., Gase K., Hui D. & Baldwin I.T.(2004) Manipulation of endogenous trypsin proteinaseinhibitor production in Nicotiana attenuata demonstrates theirfunction as antiherbivore defenses. Plant Physiology 134, 1181–1190.

Zhu-Salzman K., Salzman R., Ahn J.E. & Koiwa H. (2004)Transcriptional regulation of sorghum defense determinants against a phloem-feeding aphid. Plant Physiology 134,420–431.

Zimmermann P., Hirsch-Hoffmann M., Hennig L. & GruissemW.K. (2004) Genevestigator. Arabidopsis microarray databaseand analysis toolbox. Plant Physiology 136, 2621–2632.

Received 28 March 2006; received in revised form 25 July 2006;accepted for publication 28 July 2006

SUPPLEMENTARY MATERIAL

The following supplementary materials are available forthis article:

Figure S1. Xyloglucan endotransglycosylase (XET) activityin celery petiole.Figure S2. Characterization of xth33-1 mutant line.Table S1. Characteristics of pPP2-XTH33 lines.

These materials are available as part of the online articlefrom http://www.blackwell-synergy.com/doi/abs/10.1111/j.1365-3040.2006.01618x(This link will take you to the article abstract.)

Please note: Blackwell Publishing is not responsible for thecontent or functionality of any supplementary materialssupplied by the authors. Any queries (other than missingmaterial) should be directed to the corresponding authorfor the article.



http://www.blackwell-synergy.com/doi/abs/10.1111

Documents

Involvement of the xyloglucan endotransglycosylase/hydrolases encoded by celery XTH1 and Arabidopsis XTH33 in the phloem response to aphids