The Plant Cell, Vol. 7, 573-588, May 1995 O 1995 American Society of Plant Physiologists Small Cysteine-Rich Antifungal Proteins from Radish: Their Role in Host Defense Franky R. G. Terras,' Kristel Eggermont,' Valentina Kovaleva,b Natasha V. Raikhel,b Rupert W. Osborn,' Anthea Kester,c Sarah 6. Rees,' Sophie Torrekemd Fred Van Leuven,d Jozef Vanderleyden,' Bruno P. A. Cammue,' and Willem F. Broekaert 'I' a F.A. Janssens Laboratory of Genetics, Katholieke Universiteit Leuven, W. De Croylaan 42, 6-3001 Heverlee, Belgium Plant Research Laboratory, Department of Energy, Michigan State University, East Lansing, Michigan 48824-1312 ZENECA Agrochemicals, Jealott's Hill Research Station, Bracknell, Berkshire RG12 6EY, United Kingdom d Center of Human Genetics, Katholieke Universiteit Leuven, Herestraat 49, 6-3000 Leuven, Belgiurn Radish seeds have previously been shown to contain two homologous, 5-kD cysteine-rich proteins designated Raphanus sativus-antifungal protein 1 (RsAFPl) and RsAFP2, both of which exhibit potent antifungal activity in vitro. We now demonstrate that these proteins are located in the cell wall and occur predominantly in the outer cell layers lining differ- ent seed organs. Moreover, RsAFPs are preferentially released during seed germination after disruption of the seed coat. The amount of released proteins is sufficient to create a microenvironment around the seed in which fungal growth is suppressed. Both the cDNAs and the intron-containing genomic regions encoding the Rs-AFP preproteins were cloned. Transcripts (0.55 kb) hybridizing with an RsAFP1 cDNA-derlved probe were present in near-mature and mature seeds. Such transcripts as well as the corresponding proteins were barely detectable in healthy uninfected leaves but accumu- lated systemically at high levels after localized fungal infection. The induced leaf proteins (designated RsdFP3 and RsdFP4) were purified and shown to be homologous to seed RsAFPs and to exert similar antifungal activity in vitro. A chimeric RsAFP2 gene under the control of the constitutive cauliflower mosaic virus 35s promoter conferred enhanced resis- tance to the foliar pathogen Alternaria longipes in transgenic tobacco. The term "plant defensins" is proposed to denote these defense-related proteins. INTRODUCTION In vegetative plant tissues, a series of dynamic defense mech- expressed in transgenic plants (Broglie et al., 1991; but see anisms can be triggered upon wounding or perception of Neuhaus et al., 1991; Alexander et al., 1993; Yoshikawa et al., microorganisms. Newly formed carbohydrate material can 1993; Zhu et al., 1994). Many stress signals are able to induce be deposited in the cell wall in response to penetration atternpts the expression of defense-relatedp roteins not only locally (e.g., by fungal hyphae (Aist, 1976), and preexisting cell wall proteins in the vicinity of the infection site) but also in distant, non- can be oxidatively cross-linked upon wounding and elicitor stressed leaves. This phenomenon is known as systemic treatment (Bradley et al., 1992). 60th responses result in an acquired resistance (Ross, 1961; Tuzun et al., 1989; Ward et induced fortification of the cell wall. Another strategy fol- al., 1991; Uknes et al., 1992). Candidate signal molecules in- lowed by plants to thwart invaders is based on the localized volved in the immunization of a plant against subsequent production of antimicrobial low molecular weight secondary infection are salicylic acid (Malamy et al., 1990; Mdtraux et metabolites known as phytoalexins (Van Etten et al., 1989; al., 1990; Gaffneyet al., 1993; Delaneyet al., 1994) and methyl Maher et al., 1994). Furthermore, the synthesis of many pre- jasmonate (Farmer and Ryan, 1990; Farmer et al., 1992; Xu sumed defense-related proteins is induced when plants are et al., 1994). Finally, some incompatible interactions between confronted with pathogens (Linthorst, 1991). Among these pro- pathogens and plants can trigger localized necrosis of cells teins are the pathogenesis-related (PR) proteins. Members of (Dixon and Lamb, 1990). This hypersensitive response can five different PR protein families have been shown to possess most probably be explained by "gene-for-gene" incompatibil- antifungal activity in vitro (Mauch et al., 1988; Woloshuk et al., ity (Ellingboe, 1981). A well-studied example of such a 1991; Hejgaard et al., 1992; Niderman et al., 1993; Sela- mechanism is the interaction between different tomato lines Buurlage et al., 1993; Ponstein et al., 1994), and some PR pro- and different Cladosporium fulvum races. In the plant, C. ful- teins confer enhanced resistance to fungal diseases when vum produces race-specific elicitors (3-kD cysteine-rich peptides) that trigger the hypersensitive response only in To whom correspondence should be addressed. tomato plants with the matching resistance gene (de Wit et 574 The Plant Cell al., 199 e2oh)Tc. curree nhhtc yfeop ersensitive responssei mediated by reactive oxygen species produced early in the plant-pathogen interaction (Levin teea l., 1994). Thus, plant defense mechanisms have been studied very intensiveln yvi egetative tissues. However, littles ki nown about the strategies used by seeds to survive and germinate in sub- strates densely populated with microorganisms. A particularly vulnerable stage occurs at early germination when the seed coat, which forn meafsf ective physical barrier ag-aimins t crobes, is disrupted and the young seedling becomes exposed to the soil. We have recently characterized a novel family of 5-kD cysteine-rich antifungal proteins (AFPs) from seef droas dish (Terras et al., 1992b) and four other crucifers, including Arabidopsis (Terras et al., 1993). Radish (Raphanus sativus) seeds contain almost equal amo ouiwsnto tfofso rms, Rs-AFP1 and Rs-AFP2, that exert antifungal activity against a broad spectrum of plant pathogenic filamentous fungi by causing hyperbranching and growth reduction of the hyphal tips. These proteins have little or no effect on bacteria and cultured hu- man cells (Terras et al., 1992b). Using radish seed as a model system, we studied the release of Rs-AFPs during germina- tion and the significance of this phenomenon with respect to Figu. r1eR eleasf Aoe ntifungal Compouny Gdbse rminating Radish seedling protection against fungal pathogens. In addition, the Seeds. expression of Rs-AFPs or Rs-AFP-like proteins was examined e mniOcrograf pomu rified Rs-A saFap ePwpph1o ltis etiadti ons indi- in nonstrd esstnrseaesd sed leao vvaeT elsipd .rhaettse umed cated be nhyt umbe. 1Rr adish seeds e aphtt osin tiaino dtanah sc2 t role of Rs-AFPs in host defense, Rs-AFP2 was constitutively seed coat, whern einaacsi s sedet depaa doshss eiet iod3n cs oat n transgeexnpiirce ts osbeadcco plants that were subsequently (along half of the seed periphery). The fungus P. tritici-repentis was analyzed for resistance to a foliar fungal pathogen. used in this assay. ) AA(ssay plates containing five cereal agar. (B) Assay plates containing five cereal agar supplemented with 50 RESULTS ng/mL Pron.Ease (C) Assay plates as given in (A). The Rs-AFP1 solution and the seeds were autoclaved. Release of Rs-AFPs from Germinating Seeds ) D(Assay plates containing five cereal agar supplemented with001 \iM abscisic acid. To study the release of antifungal compounds during germi- nationf o radish seeds,e wd evelopeda bioassayn i which seeds were allowedo t germinatea nom edium supporting growth seed coat, eithe yrb germinatio ybn rmo echanical incision. of a fungal colony. When the edges of the expanding colony e phreOtv fiously purified radish seed proteins with antifungal approachee ghdte rminating seeda g, rowth inhibition ha-ploa properties, including Rs-AFPS sa2, lbumins (Terra tesa l., peared around the seed, as shown in Figure 1A. The growth 1992b), and a nonspecific lipid transfer protein (Terras et al., inhibition effect coul edbm imickey dba pplying1 \ if goe ither 1992a), only Rs-AFPs could restrict growth of fungal colonies purified Rs-AFP1 or Rs-AFP2 to a well in the agar medium e ahintg ar diffusion bioassat aay mou0 n2nt<sg (datotna (shownr of Rs-AFP1n i Figure 1A). Additione ht feo ndoprotease shown). These results suggest that Rs-AFPs are the predomi- Pronasee ht ot E medium resultede ht ni abolitione ht fo inhi- nant proteinaceous antifungal compounds released from bition zoe ngeehsrm tc aiynyuabsbteind gs s swaeaee lld germinating radish seed. In the assay shown in Figure 1, the Rs-AFP1 (Figure 1B). Likewise, autoclaved seed orautoclaved fungus Pyrenophora tritici-repentis was used because this Rs-AFP1 lost their inhibitory capacities (Figure 1C). When seed fungus grows very evenly. Similar results (da tostnha own) were germs ipnaraetwvie oepy n nlhaatbdnetd td fihtoioonr mone obtained when using other fungi, for example, Fusarium cul- abscisic acid to the medium, no growth inhibition halo was morum and Pyricularia oryzae. observed aroun nidna tact seed. Under these conditions, how- To verify and quantify the presumed Rs-AFP release from ever, seeds with a mechanically applied incision in their seed radish seeds, seeds wa mithe chanically applied incnisiion coat were still capable of releasing their antifungal components their seed coats were imbibed in water, after which the imbibi- (Figure 1D). Thus, radish seed release a heat-sensitive pro- tion solution was analyzed by gel electrophoresis. Proteins teinaceous antifungal compound only after disruption of their released from a single imbibing seed were loaded on two Plant Defe5n7s5ins replica SDS-polyacrylamide gels. After a 30-min incubation, B a clear band correspondinge ht ot Rs-AFPss aw detectedno an immunoblot, as depicted in Figure 2B. After a 4-hr incuba- tione ha,t mounf otR s-AFPs releasee hitmd ni bibition medium exceeded 1 ng, the amount required to mimic the fungal growth inhibition halo formed aroe gunenord minating seed (see Fig- ure 1). As shown on the Coomassie blue-stained gel (Figure 2A), other proteins were released as well, but Rs-AFPs were the most abundant proteins released in the imbibition medium. Neither Rs-AFPs nor any other protein could be detected in the imbibition medium of intact (nongerminated and nonin- cised) seed even after a 12-hr incubation period. Comparison e ohtotf tal proted niinma munoblot pattee rhinmt fso bibition H medium from incised seed and of a crude seed extract further illustrated that Rs-AFPs were preferentially released. The amount of Rs-AFPs in a seed extract corresponding to one Figure 3. Tissue Print Immunolocalization of Seed Rs-AFPs. seed and in the imbibition medium of one incised seed (after a 4-hr incubation) was estimated to be 5 and 1.5 u,g, respec- )SAe(ction stained with amido br ltaoocftak l pr eovhtaesTinc u.lar tively. For total protein, these values are 800 and 4.8 u.g, bundle is visible in the center of the hypocotyl. respectively. Thus, ~30°/o of the total Rs-AFP content of a sin- (B) Immune section. gle incised see sradwe leased inte oihmt bibition medium after Bars = 1 mm. C, cotyledon; E, endosperm; H, hypocotyl; SC, seed coat. 4 hr. In contrast, the release of total protein from incised seed amountede tooho nt0tatl .yl6f os%o luble protein contfeon t a crude seed extract. Preferential releas foeR s-AFPs during seed imbibition sug- gestn sae xtracellular locatior notfh ese proteins. Tissue print immunolocalization revealed that Rs-AFPs are present at high levels in the outer cell layers of and in the spaces between the different seed organs, as illustrated in Figure 3. Closer ex- aminationy b immunofluorescence microscopyf o semithin seed sections confirmed that Rs-AFPs occur most abundantly in the outer cell wall layer lining the surface of cotyledons, hypocotyl, and endosperm, as depicted in Figures 4A to 4D. Rs-AFP eshtw ni allsf o inner cellsf o these organs were barely detectable by immunofluorescence microscopy. However, im- B munogold electron microscopic examination of ultrathin sections of mature radish seed revealed that Rs-AFPs specifi- cally ree mshitd indeid le lamellaf eco ell walls througheohutt different seed tissue;sF (4en ddons eapFee igrEsmu4r ,es cotyledond nhsa ypocotyl, dat toasn hown). Rs-Ae tFhrPauss extracellular seed proteins -poreps tean t sitions where the first contacts with invading fungi occur, namee lohytu, ter surfa echdtei ffosf erent seed orgadnnas e hctoatine ghit nfso tercellular spaces. This locations ic on- Figure 2. Electrophoretic Analysis of Proteins Released from Rad- sistent with a role for Rs-AFPs in the protection of germinating ish Seeds. seedlings against fungal infection. (A) Coomassie blue staining. (B) Immunostaining. f Ros- ALgFa1 Pncnoe1f Rn s;o 1tsnla0a-g iAnn 0eFsP51; l,a2ne s CDNA Structurf eo Rs-AFPs 3, imbibition medium coe rirnencsipsoeodn odsienteg d0 m a3ifnte; r lanes 4, imbibition medium corresponding to one incised seed after cDNA clonef oRs s-AFPs were obtainey sbd creenina gr ad- 4 hr; lanes 5. imbibition medium corresponding to one intact seed af- ter 4 hr; lanes 6, imbibition medium corresponding to one intact seed h sseeid cDNA lib2 ir nnaX2dir.eZy fpAOePnII d.ent positive a; frhtla e21nr e ,s7c rude seed extract correspondingo ot ne-twentieth plaques, four containedn a insertf o ^400, pbw herea-nis eht oa fs eed. Size ehmt sfo olecular mass markers (not showne ria)n di- sert length of the remaining 18 phages varied from 250 to 300 cated at right in kilodaltons. bp. Nucleotide sequences were determined for the four 400- 576 The Plant Cell HEj ()CEHE CW X OB Figure 4. Immunolocalization of Seed Rs-AFPs in Seed Tissue Sections. Tissue sections in (A) to (D) are immunofluorescence micrographs; sections in (E) to (G) are immunogold micrographs. (A) Immune section with a border between hypocotyl and cotyledon. (B) Preimmune section as shown in (A). )CIm( mune section wit ahb order between cotyledd enoann dosperm. (D) Preimmune section as shown in (C). )EIm( mune sectiof oen ndosperm tissue. ) FE(nlarg )e Eeiznhm(otd nneifcoien at ted wia thr ectangle. (G) Preimmune section of endosperm tissue. CE, cotyledon epiderm, cWiseC; ll wa, EelEl;n dosperm epiderm, EihsHy; pocotyl epiderm, LimsM; iddle laml ieob ,lBolaOd, ; BpyPr; otein body. Bars in (A) to (D) = 20 urn. Bars in (E) to (G) = 0.5 urn. Arrowheads in (E) and (F) show the presence of gold particles. Plant Defensins 577 bp inserts and for six of the other inserts. All of the 400-bp A cDNAs corresponded to the full-length Rs-AFP1 transcripts, which contained open reading frames of 240 bp. The nucleo- 1 GTTTATTAGTGATUTOGCTAAGTTTGCGTCCATCATCGCACTTCTTTTTGCTGCTCTTGTTCTTTTT -29 M A K F A S I I A L L F A A L V L F tide sequence and deduced amino acid sequence for one of V these clones, pFRG1, are depicted in Figure 5A. From a com- 69 GCTGCTTTCGAAGCACCMCAATGGTGGAAGCACA~GTTGTGC~GGCCAAGTGGGACATGGTCA parison of the experimentally determined N-terminals equence - 1 l A A F E A P T M V E A p K L C E R P S G T W S of RsAFP1 (Terras et al., 1992b) with the deduced amino acid 138 GGAGTCTGTGG~CAATAACGCATGCAAGAATCAGT~A~AACC~~G~G~CGACATGGATCT sequence of pFRGl, we concluded that the mature protein (51 t l 3 G V C G N N N A C K N Q C I N L E K A R H G S amino acids) is preceded by a 29-amino acid peptide show- 207 TGCAACTATGTCTTCCCAGCTCACAAGTGTATCTGCTACTTTCCTTGT~TTTATCGC-CTCTTTG ing all the characteristics of a signal peptide (von Heijne, 1986). + 3 6 C N Y V F P A H R C I C Y F P C * The occurrence of a signal peptide is in agreement with the 276 GTGAATAGTTTTTATGTAATTTA=A=~~G~CAGTGTCACTATCCAT~GTGATTTTAAGACATG extracellular location of the mature protein. The calculated mo- 315 TACCAGATATGTTATGTTGGTTCGGTTATAC-2GTTTTATTCACCP lecular mass of the predicted mature protein (5685 D) B corresponds well with the mass estimated by SDS-PAGE (5 kD; Terras et al., 1992b). Five of the smaller sequenced inserts 1 A~ACATA~AT?TAFA~~-c~AGG.~~TAGTAG~GA~F~G~~.~G~~TGcT~F~A~cA~~G=c were 5'truncated cDNAs corresponding to Rs-AFPl, whereas -29 M A K F A S I I V the remaining characterized positive clone contained a 5'trun- V cated cDNA corresponding to RsAFP2. The 5' end of the 69 C~~C~C.~TSGTTGFTF=~G=CG~=~=~GC.T.G~T~~CG~G~CC~C~~GG~G.-GCAc?G~G~=G - 2 O L L F V A L V V F A A F E E P T M V E A u RsAFP2 cDNA was rescued by ligation of an oligonucleotide 136 =?~.~.??Q?CC~?~???ACA=GG~C~GGAGTC~G~GG~TAATAACGCATGCAAGAATCAGTGCATT to the 3' end of single-stranded cDNAs followed by polymer- +4 C Q R P S G T W S G V C G N N N A C K N Q C I ase chain reaction (PCR) amplification. The nucleotide 207 C~CTTGA~~ACCU~TGGGTCTTG~CTATGTCTTCCCAGCTCA~GTGTATCTGTTATTTC sequence of the full-length RsAFP2 cDNA and the deduced + 2 l R L E K A R E G S C N Y V F P A H K C I C Y F amino acid sequence are given in Figure 58. Again, the ma- 216 CCTTGT~TTCCAT~CTCTTCGGTGGTTMTAOTOTGTGCGCATATTACATATAATT~GTTTGT +so P c * ture protein (51 amino acids) is preceded by an N-terminal prepeptide of 29 amino acids. PCR amplification of the genomic 315 GT~CTATTTATTAGTGACTTTATGACATGTGCCAGGTATGTTTATGTTGGGTTGGTTGTAATAT~ 414 AAGTTCACGGATAATAAGATGATAAGCTCACGTCGCCAAAAXAA fragments encompassing the Rs-AFPl and Rs-AFP2 coding region yielded bands larger in size than the products obtained C by amplification of the corresponding cDNA regions, indicat- ing the presence of introns. Therefore, the nucleotide Rs-AFPI QKLCERPSGTWSGVCGNNNACKNQCINLEKAWGSCNYVFPAllKCICYFPC sequences of both genomic fragments were determined. The Ra-AFPZ QKLCQRPSGTWSGVCGNNNACKNQCIRLEKAWGSCNYVFPAHKCICYFPC intron in the Rs-AFP1 coding sequence is 112 bp in length, Ra-ALFP3 KLXERSSGTWSGVXGNNNAXK?TQXIRLEGAQHGSXNYVFPAHFXIXYFPX Rs-AFP4 QRLCERSSGTWSGVCGNNNACKNQCINLEGARHGSCNYIFPYHRCICY whereas the Rs-AFP2 intron spans 98 bp. Both introns occur at the same position in the open reading frames, namely, be- Figure 5. Nucleotide Sequence of cDNAs and Genomic Regions and tween the codons for amino acids -9 and -8 of the N-terminal Amino Acid Sequences of Rs-AFPs. prepeptide. (A) Nucleotide sequence and deduced amino acid sequence of the From the nucleotide sequence data, the complete amino full-length Rs-AFPI cDNA insert of pBluescript SK- phagemid pFRGI. acid sequences of Rs-AFP1 and RsAFP2 could be deduced (E) Nucleotide sequence and deduced amino acid sequence of the (Figure 5C). It thus appears that RsAFPl and RsAFP2 are full-length Rs-AFP2 cDNA. The depicted nucleotide sequence was nearly identical peptides differing only at residues 5 and 27. obtained by combining a 5' truncated cDNA sequence isolated from The open reading frame encompassing the RsAFP2 prepro- the seed cDNA library and the 5end rescued by anchor PCR (under- tein is 91% identical to the Rs-AFP1 coding sequence. scored with dotted line) with an antisense primer specific to a part of the 3' noncoding region (double underlined). (C) Complete amino acid sequences of mature seed RsAFPl and Rs- AFP2 deduced from (A) and (6)a nd N-terminal sequence data of the Expression of RsAFPs during Seed Development and RsAFPs (Terras et al., 1992b). N-terminal amino acid sequences of in Biologically Stressed Leaves leaf RsAFP3 and RsAFP4 were determined in this study. Start and stop codons, experimentally determined N-terminal amino The tissue-specific expression patterns of the RsAFPs as well acid regions, and putative polyadenylation signals are underlined in as the temporal expression pattern of the RsAFPs during seed (A) and (6).T he cysteines of RsAFP3 were not derivatized before au- development were determined by RNA gel blot analysis using tomated Edman degradation; therefore, the putative cysteine residues an Rs-AFPl cDNA-derived RNA probe. In seeds, RsAFP of Rs-AFP3 are indicated by Xs in (C). The positions of the intron se- mRNAs start to accumulate at the onset of desiccation (45 days quences are indicated by arrowheads in (A) and (E). Asterisks in (A) and (6)i ndicate stop codons. Nucleotide sequences were submitted postanthesis) and are still present in mature, dry seed, as to GenBank and have accession numbers U18557 (RsAFPl) and shown in Figure 6. Therefore, Rs-AFPs are expressed only dur- U18556 (RSAFP2). ing the final stage of seed development when the seeds are prepared to detach from the mother plant. Cross-hybridizing transcripts of the same length (0.55 kb) were also present at 578 The Plant Cell the chromatographic profiles of both samples (depicted in Figures 8A and 8B, respectively), two peaks become promi- nent in A. brassicola-infected leaves that are virtually absent SEEDS ldpa) in healthy leaves. Both peaks coeluted with antifungal activity IS 20 25 35 45 Dry d conntaaD (inif nFeosikgde u tpr5eep 8tfiBdo)ee, swrhaich called Rs-AFP3d naR s-AFPe h4tp rof eptidee htes ni arliest and latest eluting peaks, respectively. Both Rs-AFP3 and Rs- bk[ 0.5 ? AFP4 cross-reacted with an antiserum raised against Rs-AFP1 when assessed by immunoblotting of SDS-PAGE gels (data t e scohohonnwcTenn) .traf tRiood Rsn-s snA--FAaePF3P r4 LEAVES quired to obtain 50% inhibition of fungal growth were similar CONTROL ALTERMARIA BOTRYTIS Hef 12 0 1+ 3+ 1- 3 ++3 t I 1-3 ++3 2+-2 to those of the seed Rs-AFPs (Table 1). Moreover, fungi treated with Rs-AFP3 or Rs-AFP4 displayed a hyperbranched mor- '"<** Hi " : 1.20kb phology similar to that of fungi treated with seed Rs-AFPs (data 0.55 kb not shown). The final proof that leaf Rs-AFPs are homologous to seed Rs-AFPs was provided by the amino acid sequences. Figure 6. RNA Gel Blot Analysis of Rs-AFP Expression in Radish. The complete amino acid sequence consisting of 51 residues e ahnTalyzed samples repres5 en A f1noetgNot x tRtara lcted from was determiner oRdf s-AFP3, where eaphstu tative last three eihntdicated tissuee dhsTe. signation hypocotyls ree fshetu rosct cu- amino acids were missing for Rs-AFP4. Both sequences are lent storage organs of the radish plant, dpa, days postanthesis; 0, total shown in Figure 5C and can be compared with the amino acid RNA isolated from leaves collected just after inoculation w0 i5t5h- nL sequencf esoes ed Rs-AFe oPhvsTe .rall amino acid sequence drops of H2O (CONTROL), 50 5-nL drops of A. brassicola spores (at homology of leaf Rs-AFPs with seed Rs-AFPs is ~90°/o. Rs- 5 x 105 spores per mL), 50 5-nL drops of Botrytis cinerea spores (at t oRe AdnbsFi g -tPoAuetb4Fs d, tPea3dh, we eihtnht zyme 5 x 10s spores per mL), or 50 5-nL drops of 0.2% (w/v) mercuric chlo- pyroglutamate aminopeptidaseo ot btain amino acid sequence ride; 1 +, 2+, and 3+ represent total RNA isolated from leaves collected signals. Similar to the seed Rs-AFP isoforms (Terras et al., ,rrhe2 s247p ,de4nc8at, ively, after inoc- 3u rdleanptaior e-n2s; ent I992b), Rs-Aa FcPy sc4l aizthheuds eghlutta smaine tA oiNstaRol lated from noninoculated leaves froe mshta me plants as given above a, rfh 2tr7 edena 8rs4 pectively. N-terminal amino acid. low len vroiee n holtlehoisyt d astspn ,vntoe eatcmus obsty ,l- Enhanced Disef aTsrea onTsogleernainc c Teobacco derived succulent storage organ (all harvested from 5-week- Expressing RS-AFP2 old plants), or flowers (harvested from adult radish plants). In leave ass ,econd hybrids oizabiwnseg br tvkrea 2dn..s1c rfiop t e hcTharacteriste ihcRt fsos -AFPs emerging from thdinsa This transcript was also apparent in stems and flowers after previous work (Terras et al., 1992b) prompted us to analyze overnight exposure (data not shown). This mRNA species has expression of these proteins in a heterologous tobacco sys- tebneyoe tn identified. to eamste sdeihssets ase resistanf scoue ch transgenic plants. Because Rs-AFP-like proteins seem to be at least weakly expressed in leaves, we were interested in whether their ex- pression was affected by stress. Therefore, different stress 12 3 4 56 factors were applied to leaves of 5-week-old plants, after which treated and untreated leaves from the same plants were col- lected separater olyfis olatiof ont otal l RegbN AloNAtR. analysis (Figure )6r evealed that induction abovw ocle ehot n- stitutive expression level of leaf Rs-AFPs was triggered by fungal in ytrfbee acdttnimoane nt with mercuric chloride. This induction was not restricted to stressed leaves because un- treated leaves showed an almost identical response, implying Figure 7. Induction of Rs-AFP-Like Proteins in Infected Radish Leaves. the transduc atsi otfrone ss signal throue pghlhta nt. Moreover, Four-microgram amounts of leaf protein after partial purification (see the systemic accumulation of the 0.55-kb transcript was ac- Methods) were electron pShaDo rSen-spoeod lydancrayl almeigde coe amhcptca uynmibeudl a fto5io-nk D proteins that were subjectedo ti mmunoblot analysis using anti-Rs-AFP1 antibodies. Lane recognized by antibodies raised against Rs-AFP1, as shown fo gR1n 00,1s -AFP1; laf o gnRn 00e2s ,2 -AFP1; lane ,3 uninfected eimhbtmy unoblot analyn sFiiis g.u7re leaves; lan, e4u ninfected leavef osr adish plants infected wit. Ahb ras- I nnea ffoo ritts ola etihent duced Rs-AFP-like proteins from sicola (see legend to Figure 6) collected 72 hr after inoculation; lane infected leaves, the total basic protein fractions from healthy 5, infee cstaehmd tele r pa; fs ,lullaloeav4o6asneaannee f sve td ses and Alternaria brass/co/a-infected leaves were separated on inoculated with HO c roahlflte e2cr t7eind oculation. Samples were 2 a reverse-phase chromatography column. When comparing prepareds ad escribedr oft obacco (see Methods). Plant Defensins 579 pollinated, and the progeny were screened for homozygous, hemizygous, and azygous genotypes. Expression of Rs-AFP2 ine hth omozygous T liness awa gain checkedy b immuno- 2 blot analysis of crude leaf protein fractions (Figure 9B). Based on immunoblot anao ilwnytsd fioes pendent leaf protein sam- ples per plant from at least two siblings per T line, we 2 concluded that Rs-AFP2s awe xpressedt a levels ranging from f Rs-oAFP2 2 ng 1 ot f to3taol leaf( pr o teign m rep 4 ug .2 ot 0.6 per g of fresh leaf tissue) in the five independent homozygous T lines, with line 6014 displaying the highest expression level 2 .)2(Ta ble Crude leaf protein fractions were also tested for in vitro an- tifungal activity against the tobacco foliar pathogen A. longipes usinga q uantitative microplate assay (Broekae treat l., 1990). The antifungal active ihtct yfor ude leaf protein fractions pre- pared from the transgenic homozygous T line expressing 2 Re sh-hAigtF htPea2s t levels (line 6s m0a1ow4r) e -t0h1an fold higher than that of extracts from untransformed or T azy- 2 gous (null line) plane thsTo. ther homozygous T lines 2 displayed only a two- to fourfold higher antifungal activity rel- ae cthiotv onet trols (Tab. l)Ue2 nde gehitrv en assay conditions, purified seed Rs-AFP2 inhibited growth of A. longipes by 50% ata concentrationf o 1.5 ng/mL. A le.hoT ngipes microcultures treated with crude leaf protein fractions from transgenic line 6014 clearly displayee hdht yperbranched morphology typi- cally causey Rbd s-AFPs (Terrat sae l., 1992b; this study, Figure 9C). Thus, Rs-AFP2 eb nfac unctionally expresseda ni heter- ologous tobacco system. The transgenic T lines expressing Rs-AFP2 were analyzed 2 for disease resistanc. Al eoot ngipee shaT. re ehlte afo sions formed on leaves of the homozygous T line 6014 was on av- 2 2O ELUTION TIME (minutes) 45 erage eight- and sevenfold lower than that of lesions formed n aazy gnoous T line derived efsrohammt e primary trans- 2 Figure 8. Purification ot Rs-AFP-Like Proteins from Infected Radish formant and on an untransformed plant, respectively, as Leaves. eho tlde eNsfpoioi ncnF. t0esii%g d 1wur4ee r6e v nisioble S eebpahastri caf tpioorno tein fracf trioaodn isa rhe vlenearovsees- infection spots of the homozygous line, versus 31 and 34% phase chromatography column (see Methods). e ahzd yuftognn rotaruas nsformed lines, respectively. Simi- (A) Basic protein fractiof o ng 0h2 fo ealthy leaves. lar results were obtained in three additional independent tests. (B) Basic protein fraction of 20 g of A. brassicola-infected leaves. The On the other hand, none of the other homozygous transgenic inset shows a gel blot of proteins contained in peak 1 (lane 1) and peak lines were more resistant than the control lines (data not 2 (lane) a2 fter SDS-PAGE. LanR ec ontaine mhst olecular mass mark- shown), indicating that Rs-AFP2 expression in transgenic ers: myoglobi7 n1k( D), myoglobI iI n(dn1as I4 .4 kD), myoglobi8n( I kD)d ,m 6 mnykyoD(aogI ) g( lI,o2IlIob.Ib5in i n kD). FSAU, full-scale absorption units. Tablen V1I i.tro Antifungal Activ foiRty adish aAF nPis Synthetic Low Ionic Strength Medium Tobacco leaf discs were transformy cboecd ultivation waith Fungus Rs-AFP1a Rs-AFP2a RS-AFP3 Rs-AFP4 disarmed Agrobacterium strain harboring the plant transfor- mation vector pFRGS (containing the chimeric gene for A. brass/cola 15 5 constitutive expression of Rs-AFP2; schematically shown in Botrytis cinerea 8 9 Figure 9A). Forty T kanamycin-resistant plants were selected F. culmorum 5 11 0 R aCnaPblya f sntoshiseo edi r e gpehrnet osrmeonffeo ce The given quantitative measun iv rroieft ro antifungal activehitt syi the introduced gene as well as on Rs-AFP2 expression as as- protein concentration (micrograms per milliliter) required to obtain 50% sessy eimbd munoblotting. After self-pollination, segregation growth inh sidbi eidtirnoiavne d from dose-response curves (Broekaert analyse ihpts for ogens payw erformea g dybe rmination assay et al., 1990). a In vitro antifungal activities as previously determined (Terras et al., on kanamycin-containing medium. Five T, lines segregating 1992b). 1 (re:3sistant/suscepto ikbatlen amycin) were rd esteanlifn-aed 580 The Plant Cell plants must reach a threshold level before displaying an in- creased disease tolerance phenotype. H DISCUSSION B D /VJ£ TRANSGENIC LINES RS-AFP2 e dhaTta presentn ethid is paper strongly sug-gneas te thhat t tifungal Rs-AFPs are an important component of the defense system of radish seeds. The extracellular localization of the Re sp-hrAetF fen dPsrinese eandt ial releaf tsohee se proteins in the imbibition medium during germination can contribute to the control of seed- and soil-borne fungal diseases, thus WATER RS-AFP2 enhane ccihhntag ncf esoes edling surd vrienvpaar loduction. Many seed- dna soil-borne fungal diseases causing important losses in yield and seedling stand are effectively controlled by treating seeds with chemical fungicides (Bateman, 1977; Koch and Leadbeater, 1992), which is a common practice in modern agriculture. Hence, Rs-AFPs may be considered a bi- ological equivalenf to seed-applied chemical fungicides. Althoug. (lh1a9 S9 2wt)e ehgale ve previously sho-wnn taha t tifungal chitine arrseaeles ased from barley seeds during imbibitiot onnd saw, eti monstrated thae thtr eleased amounts are physiologically meaningr opfufro lte eshcett eifoodn. Surprisingle yeht, xpressiof noR s-AFP-like proteinn sri ad- AZYG HOMOZYG t orens tsriic hto esstdei t euadblss o occun lreis af tissues. e RhTs-AFP counterparn tlsie ave seratsr ond sgnyalys temi- cally induced upon fungal infection and can therefore be classifiw eetn yad sap R Pep fo roteine W.i solate-sR dowt AFPs from. Ab rass/co/a-infected radish leaves: theye r~a 90% homologous to the previously purified seed isoforms (Terras et al., d 1e9xn9e2artb c )omparn avibtirleo antifungal activity. 4 These results also indicate that the pathogen-induced leaf Rs- 1* t oidnAe FenPrtaisc o aste led Rs-AFPs, which implies that •Wk Tab. l2eR s-AFP2 Expression Ln eiVv dientrlaso Antifungal Figure 9. Analysis of Transgenic Tobacco. Activity of Leaf Protein Extracts from Transgenic Tobacco Lines (A) Diagram of the T-DNA region of pFRG8. LB, left border; RB, right Rs-AFPn IV2 itro Anti- borde, HHr; indlll site; A+, polyadenylation se Cihgtan SfMao5 l3V Content3 fungalb promoter; Rs-AFP2, coding regf iRoons -AFP2; Enh, -eS3n5 hanced (lig/mL Total Activity CaMS V5p3 romoter (Kayt e al., 1987); KanR (nptll), chimeric gene Plant Line Leaf Protein) (Units/mL) conferring kanamycin resistance wia tnh opaline synthase promoter d tene rnamhetoi ndmantyoacr in phosphotransferasI cIe oding region. Untransformed <0.2 30 ) EBxp(ressf iRoons -AFn Ph2io mozygous T lines -vmisuia lyizbed Azygous line 6014 <0.2 30 2 munoblotting. UNTRANS, untransformed. Homozygous line 6014 2.4 320 nIv )iCtr(o antifungal activi ftoyw ater again. sAlto ngipes (nega- Homozygous line 7003 0.6 60 tive control), purified Rs-AF3 nP tga2 /mL (positive contrdonla), Homozygous line 7006 0.6 60 partially purified leaf extractsf o azygous (AZYGd n)a homozygous Homozygous line 7009 1.3 120 (HOMOZYG) T 6014 lines derived from the same primary transform- Homozygous line 8001 1.1 120 2 e adhnilTe ute. htxito tfrnoas e certqas uie vpahrltoe otnett in fraction a RS-AFP2 levels were determined by densitometric scanning of im- derived from 0.15 g of leaves (fresh weight) per mL of culture medium. munoblot luminograms. Total protein levels weree dehtetr myinbed Micrographs were taken a 4ff toh2ienr r cubatit oa2n2 °C. Coomassie blue dye binding method (Bradford, 1976). n viItro b antifungas lm aaecatwin svpuitao ryertd ially purified leaf extracts using A. longipes as a test fungus (see Methods for addi- tional experimental details). Plant Defensins 581 - 60 the enhanced tolerance of Rs-AFP2-expressing tobacco line (Y-E 6014 could be caused by enhanced expression of defense- E related proteins. However, the activity of chitinases, glucanases, o E 40 and peroxidases in line 6014 measured both before and after m infection with A. longipes was not significantly different from c .O- that in similarly treated untransformed plants (I. Penninckx and u) 3! 20 W.F. Broekaert, unpublished results). al m From the A. brassicola-infected radish leaves, we were able e to purify -2 pg of RsAFPs (RsAFP3 and RsAFP4 together) > m o per g of fresh leaf tissue. Taking into account the results ob- untrans azyg homozyg tained with the transgenic tobacco plants (increased disease tolerance at RsAFP2 levels of 12 pg per g of fresh leaf tis- Figure 10. Transgenic Lines Analyzed for Disease Resistance. sue), this would mean that the accumulated amounts of Shown is a quantitative analysis of the tolerance to A. longipes infec- RsAFPs in infected radish leaves are not sufficient to stop the tion of the RsAFPP-expressing tobacco line 6014 (homozyg) relative pathogen from growing within the plant. However, preliminary to azygous T2 plants derived from the same primary transformant results obtained from tissue printing diseased leaves showed (azyg) and untransformed control (untrans) plants. The average lesion that the amount of Rs-AFPs accumulating locally around in- area was determined from a total of 64 spot infections (four spots per fection sites is higher than that in the remaining part of the leaf, two leaves per plant, and eight plants per line) 7 days after inocu- leaf (F.R.G. Terras, unpublished results). lation with 5-WLd rops containing 1000 spores per mL in 50 mM glucose. Moreover, spray application of RsAFP2 on the surface of Lines on top of bars represent standard errors. sugar beet leaves controls disease development by Cercospofa beticola as efficiently as the chemical fungicide hexaconazole, they are products of different genes. This finding is consistent when compared on a molar base (De Bolle et al., 1993), and with the existence of eight to 10 genes that hybridize with an chemical compounds that, like Rs-AFPs, cause increased Rs-AFP1 cDNA probe, as shown by DNA gel blot analysis hyphal branching in vitro are used to control foliar diseases (F.R.G. Terras, unpublished results). Of interest in this context in crops (Robson et al., 1989; Wiebe et al., 1990). is the observation of Chiang and Hadwiger (1991) that pea pods A closer examination of the arrangement of the cysteine and challenged with E solaniaccumulate mRNAs transcribed from glycine residues in the RsAFP amino acid sequences reveals two genes encoding 5-kD proteins whose deduced amino acid the putative existence of a “cysteine-stabilized a-helix” motif sequences show 45 and 51% identity with RsAFP1. The on- characterized by the occurrence of the sequences CXXXC, set of the accumulation of these pea transcripts coincides with GXC, and CXC (where X stands for any amino acid). This mo- the initial suppression of the growth of the pathogen (Chiang tif was originally reported to occur in endothelin (Kobayashi and Hadwiger, 1991). Thus, Rs-AFP-like proteins could, to- et al., 1991; Tamaoki et al., 1991), a mammalian peptide with gether with other pathogenesis-induced antifungal proteins vasoconstricting activity (Yanagisawa et al., 1988), and later such as chitinases and glucanases, contribute to the control shown to dictate the three-dimensional structure of insect of funga1 diseases in vegetative tissues. Although we were un- defensin A (Bonmatin et al., 1992). lnsect defensin A is a able to demonstrate expression of RsAFP-like genes in 40-amino acid cysteine-rich peptide with antibacterial prop- flowers, expression of genes encoding RsAFP-like proteins erties. It is induced in the hemolymph of the fleshfly (Phormia has been reported to occur in tobacco flowers (Gu et al., 1992), terranovae) after challenge with bacteria (Lambert et al., 1989). potato flowers (Moreno et al., 1994), and Petunia inflata pistils The y-thionins isolated from cereal seeds (Colilla et al., 1990; (Karunanandaa et al., 1994). Mendez et al., 1990) have the same arrangement of their cys- Additional evidence for a role of Rs-AFPs in plant defense teine residues as the Rs-AFPs (and thus are more homologous came from the analysis of the disease tolerance of transgenic to the Rs-AFPs than to the a-o r p-thionins; Terras et al., 1992b; tobacco plants constitutively expressing RsAFP2. The leaves Bohlmann, 1994) and display a three-dimensional fold highly of such Rs-AFP2-expressing plants displayed a significantly similar to that of insect defensins (Bruix et al., 1993). The pre- decreased susceptibility to A. longipes infection relative to un- liminary nuclear magnetic resonance-based three-dimensional transformed control plants. The difference in disease tolerance structure of RsAFP1 (F. Fant, L. Santos, W. Vranken, K. Boulez, between T2 line 6014 and the other (lower expressing) trans- J.C. Martins, and F.A.M. Borremans, personal communication) genic lines indicates that RsAFP2 must be expressed above fully supports the structural homology of this protein with a certain threshold leve1 before a disease tolerance pheno- y-thionins and insect defensins. Recently, a cysteine-rich pep- type can be observed. This is consistent with the fact that tide with sequence homology to the Rs-AFPs (18 of 44 residues purified seed RsAFP2 also has a sharp dose-dependent identical after alignment to the RsAFP1 sequence, including growth inhibition effect on fungi in in vitro assays. Moreover, the cysteine-stabilized a-helical motif) was isolated from the we also observed disease resistance to A. longipes when as- hemolymph of bacterially infected fruit flies. Like the RsAFPs, saying high-expressing multilocus integrant TI lines in this peptide is not active against bacteria but displays strong preliminary tests (data not shown). We have also verified that antifungal activity. This 44-residue peptide starts to accumu- 582 The Plant Cell late in the insect hemolymph as early as 1 hr after bacterial mortar containing 2 mL of RTW. The suspension was then allowed challenge and is the major immune-induced peptide produced to stand for 3 hr at 4OC, followed by removal Of Particulate material by fruit flies (Fehlbaum et al., 1994). Hence, it appears that bY centrifugation (5 min at ll,ooo rpm). plants and animals have developed similar weapons to com- Estimation of radish antifungal proteins in the seed imbibition sob bat microbial infection. tion or seed extract was done by densitometry scanning of an x-ray film obtained after chemiluminescent detection of RsAFPs on an im- On the well-documenatendtif ungal Of the munoblot (see the following section) and standardized to a twofold RsAFPs (Terras et 1992b; this study), their presumedr ole dilution SerieS of purified RsAFPl, Determination of total protein was in host defense*a s wella s their structuraaln d functionahlo - performed by the Bradford protein assay (Bradford, 1976) using BSA mologies with pathogen-inducible insect defensins, we propose as a standard. the term "plant defensins" to describe the RsAFPs and their homologs in plants. Protein Gel Electrophoresis and lmmunoblottlng METHODS Samples were analyzed by SDS-PAGE (separating gei: 15% total acryi- amide, 0.5% bis-acrylamide; stacking gel: 5% total acrylamide, 2% bis-acrylamide)o n two replicate gels. One gel was stained with Coomas- Biological Material sie Brilliant Blue R 250; proteins separated on the other were electroblotted (semidry, 1.5 hr in 48 mM glycine, 39 mM Tris, 20% [vlv] Radish (Raphanus sativus cv Ronde Rode Kleine Witpunt) seeds were methanol at 1 mA/cm2) onto nitrocellulose (0.2-pm pore size; Hoefer obtained from AVEVE (Leuven, Belgium). Growth of fungi and har- Scientific Instruments, San Francisco, CA). lmmunoblots were pro- vesting and storage of fungal spores were done as described previously cessed as described (Boehringer Mannheim Colloquium No. 2/1992), (Broekaert et al., 1990). The following fungal strains were used: Alter- and detection was performed using the enhanced chemiluminescence nariabrassicola (MUCL 20297; Mycotheque Université Catholique de method (Amersham). Anti-RsAFP1 lgGs were obtained by treating Louvain, Louvain-Ia-Neuve,B elgium), A. longipes (CBS 620.83; Cen- antiserum from immunized rabbits with caprylic acid followed by traalbureau voor Schimmelcultures, Baarn, The Netherlands),F usarium (NH4)2S04p recipitation according to Harlow and Lane (1988). The IgG culmorum (IMI 180420; lnternational Mycological Institute, Kew, UK), precipitate was then redissolved in PBS (the initial serum volume) and and Pyrenophora tritici-repentis (MUCL 30217). dialyzed against PBS. The IgG fraction was passed over an affinity matrix consisting of cyanogen bromide-activated Sepharose 6MB (Pharmacia) to which total radish seed proteins were coupled (spe- Agar Diffusion Bioassay for Antifungal Activity cific anti-Rs-AFP antibodies do not bind to purified Rs-AFPl coupled to this matrix; F.R.G. Terras, unpublished results). The partially puri- Radish seeds were surface sterilized by the following procedure: 1 min fied primary antibodies were diluted 1:5000 in the immunoblotting in 70% (vhr) ethanol, 10 min in a sixfold diluted commercial hypochlorite experiments. solution containing 0.1% (w/v) SDS, and five washesof 5 min in sterile water. The fungus /? tritici-repentis was grown on five cereal agar (five cereal baby food instant flakes, 20 g/L, and agar, 8 glL; Broekaert et lmmunolocalization al., 1990). With a sterile cork borer, a mycelial plug was removed from a plate colony and placed mycelium-side down in the center of a 9-cm For tissue print immunolocalization, mature radish seed were cut in Petri dish containing five cereal agar. Surface-sterilized seeds were halves. With each half, a print was made on nitrocellulose (0.22-pm buried in the agar 3 cm from the center. Purified proteins were applied pore size) prewetted in Tris-buffered saline (100 mM Tris, 150 mM NaCI, in a total volume of 5 pL in a well (3-mm diameter and 3-mm depth) pH 7.5). Prints were stained for total protein with 0.1% (w/v) amido black punched in the medium 3 cm from the center. The seeds were posi- in 25% (vlv) isopropanol, 10% (vhr) acetic acid for 1 min and destained tioned in the agar simultaneously with the mycelial plug; purified by three 10-min washes in 25% (vk) isopropanol, 10% (v/v) acetic acid. proteins were applied 24 hr after inoculation of the plates with the Prints for immunolocalizationo f Rs-AFPsw ere processed as described mycelial plug. The assay plates were incubated for a total of 5 days. above, except that the secondary antibodies were conjugatedt o alkaline phosphatase and the detection was performed with 5-bromo-4-chloro- 3-indolyl phosphate and nitro blue tetrazolium. lmbibition of Radish Seed Whole radish seed were fixed and subsequently embedded in Lon- don Resin White resin (Polysciences, Warrington, PA) according to Radish seeds were soaked in running tap water (RTW) for 30 min. De Clercq et al. (1990). Ultrathin (60 nm) sections were made with an With a scalpel, an incision was made in the seed coat (along half of Ultracut S microtome (Reichert-Jung, Vienna, Austria) equipped with the seed periphery), and the incised seeds were incubated (in a 1.5- adiamond knife and collected on nickel grids precoated with Formvar mL Eppendorf tube; five seeds in 250 pL of RTW) for an additional (Polyscience, Niles, IL). Semithin sections (1 pm) were produced with 30 min or 4 hr at room temperature. lntact seeds were incubated (five a glass knife and collected on glass slides pretreated with 4% (w/v) seeds in 250 pL of RTW) for4 or 12 hr. A small hole was then punched gelatin and 0.4% (w/v) KCr(S04)2. Ultrathin sections were further in the bottom of the Eppendorf tube, and the imbibition medium was processed as follows: 30 min in 5% (whr) Na104;t hree times for 2 min separated from the seeds by centrifugation (5 min at 3000 rpm). The in H20; 10 min in 0.1 N HCI; three times for 2 min in H20;t wo times fluid was collected in a second Eppendorf tube. Particulate material for 30 min in PBS, 0.1% (vhr) Tween 20, 2% (wlv) BSA, pH 7.2; 1.5 hr in the imbibition medium was removed by centrifugation (5 min at 11,000 with partially purified anti-RsAFP1 antibodies or preimmune serum rpm). A crude seed extract was obtained by grinding 50 seeds in a diluted 50-fold in PBS, 0.1% (v/v) Tween 20, 2% (w/v) BSA, pH 7.2;
Description: