Identical homologs of the Galanthus nivalis agglutinin in Zea mays and Fusarium verticillioides

Plant Physiology and Biochemistry 49 (2011) 46e54

Contents lists available at ScienceDirect

Plant Physiology and Biochemistry journal homepage: www.elsevier.com/locate/plaphy

Research article

Identical homologs of the Galanthus nivalis agglutinin in Zea mays and Fusarium verticillioides Elke Fouquaert a, Willy J. Peumans a, Godelieve Gheysen b, Els J.M. Van Damme a, * a b

Department of Molecular Biotechnology, Lab. Biochemistry and Glycobiology, Ghent University, Coupure links 653, 9000 Gent, Belgium Department of Molecular Biotechnology, Lab. Applied Molecular Genetics, Ghent University, Coupure links 653, 9000 Gent, Belgium

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 July 2010 Accepted 27 September 2010 Available online 8 October 2010

The structural domain corresponding to the Galanthus nivalis agglutinin (GNA) is a mannose-binding motif that was originally discovered in plants but according to recent data also occurs in other eukaryotes and prokaryotes. Transcriptome analyses revealed that Fusarium verticillioides expresses a protein (FvGLLc1) identical to a recently identified cytoplasmic/nuclear GNA-like lectin from maize (ZmGLLc). The FvGLLc1 and ZmGLLc gene sequences are nearly identical in the coding region as well as in the intron and the 5 and 3 prime untranslated regions. However, whereas the Fusarium genome contains only a single gene with an intron, both an intronless and an intron containing lectin gene can be amplified from maize DNA. Southern blot analysis confirmed the presence of this cytoplasmic GNA-like gene in the maize and rice genome. A comparative analysis of the products amplified by different PCRs using genomic DNA from Fusarium species and maize DNA samples from sterile as well as contaminated plant material strongly indicated that the GNA-like sequence found in maize grown under sterile conditions is not derived from a contaminating Fusarium species. Furthermore, using a PCR-based approach it could be demonstrated that this particular type of lectin occurs also in other plants from distant taxa and is markedly conserved. Ó 2010 Elsevier Masson SAS. All rights reserved.

Keywords: Agglutinin Evolution Lectin Fusarium Zea mays

1. Introduction During the past two decades numerous proteins have been identified in plants that are structurally and evolutionary related to the Galanthus nivalis agglutinin (GNA) [1]. GNA itself is an (oligo)mannoside binding lectin that was originally isolated from snowdrop bulbs and is now considered the prototype of a large family of carbohydrate-binding proteins found in organisms of different taxa [2,3]. Biochemical analyses combined with molecular cloning and subcellular localization studies demonstrated that GNA is synthesized as a preproprotein with an N-terminal signal peptide and a Cterminal propeptide, and follows the secretory pathway to its final destination in the vacuolar compartment [4,5]. Since all other related plant proteins isolated thus far are synthesized as preproproteins it seemed evident that the GNA-like plant lectins

Abbreviations: GNA, Galanthus nivalis agglutinin; FvGLLc, cytoplasmic GNA-like lectin from Fusarium verticillioides; GLLc, cytoplasmic GNA-like lectin; ZmGLLc, cytoplasmic GNA-like lectin from Zea mays. * Corresponding author. Tel.: þ32 92646086; fax: þ32 92646219. E-mail addresses: [email protected] (E. Fouquaert), Willy.Peumans@ telenet.be (W.J. Peumans), [email protected] (G. Gheysen), ElsJM. [email protected] (E.J.M. Van Damme). 0981-9428/$ e see front matter Ó 2010 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.plaphy.2010.09.018

represent a family of typical vacuolar lectins. However, screening of transcriptome databases revealed that some plants express proteins that closely resemble the vacuolar GNA-related lectins but lack the signal peptide and C-terminal propeptide [6] and, as was confirmed by confocal microscopy of a green fluorescent protein (EGFP) tagged protein, reside in the nucleocytoplasmic compartment [4,7]. Moreover, using the same approach it could be demonstrated that removal of the signal peptide and C-terminal propeptide converts the normally vacuolar GNA into a cytoplasmic/ nuclear protein [4]. GNA-related proteins and/or corresponding genes have been described in fishes (e.g. Fugu rubripes), fungi (e.g. Aspergillus oryzae), slime molds (e.g. Dictyostelium discoideum) and in the freshwater sponge Lubomirskia baicalensis [8,9,10,11,12]. All these (putative) proteins are synthesized without a signal peptide and Cterminal propeptide. This implies that a cytoplasmic/nuclear location is the rule rather than the exception within the family of eukaryotic GNA-related proteins. The development of the concept of cytoplasmic/nuclear plant lectins was based primarily on the identification of a small set of ESTs encoding strikingly conserved cytoplasmic GNA-like lectins (GLLc) in Zea mays, Triticum aestivum and Medicago truncatula [6]. To check whether similar lectins or corresponding genes are

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

present in other plants the publicly accessible databases were screened. These in silico analyses yielded - apart from a few Sorghum bicolor ESTs - no additional plant sequences but revealed that a sequence virtually identical to the Z. mays and T. aestivum ESTs is present in and expressed by the plant parasitic fungus Fusarium verticillioides (syn. Gibberella moniliformis). Here we report detailed analyses of GLLc sequences in Z. mays and F. verticillioides to check whether the maize sequence can be ascribed to contamination with F. verticillioides or another fungus. In addition, we provide evidence that sequences (nearly) identical to the maize GLLc are present in DNA samples from other grasses, several dicots, the gymnosperm Picea abies and the liverwort Marchantia polymorpha. 2. Results 2.1. Identification of identical cytoplasmic/nuclear GNA-related lectins in the transcriptome of Zea mays and the phytopathogenic fungus Fusarium verticillioides Screening of the plant transcriptome databases in 2004 provided evidence for the occurrence in plants of cytoplasmic/ nuclear orthologs of the classical vacuolar GNA-like lectins [6]. Though only a few corresponding ESTs could be retrieved, their sequences revealed a marked conservation of the primary structure of the cytoplasmic GNA-like lectins [13]. To further investigate this issue, a comprehensive screening of the publicly accessible databases was undertaken. Surprisingly a set of 41 ESTs with a nucleotide sequence virtually identical to that of the Zea mays GLLc (ZmGLLc) was detected in the transcriptome database of the plant parasitic fungus Fusarium verticillioides (syn. Gibberella moniliformis) (Fig. 1A). In addition, a set of 30 ESTs encoding a protein differing at 16 positions from ZmGLLc was retrieved in the transcriptome database of Gibberella zeae (syn. Fusarium graminearum). It appears, therefore, that both F. verticillioides and G. zeae express a protein (further referred to as FvGLLc1 and GzGLLc1, respectively) identical or very similar to ZmGLLc (Fig. 1A). Moreover, a detailed comparison revealed that the maize and F. verticillioides ESTs are apart from a single nucleotide - identical within the open reading frame (357 nucleotides) and also share a marked sequence identity in both the 50 and 30 untranslated region (UTR) (the sequence identity being 41/42 and 78/86, respectively) (Fig. 1B). Blast searches of the completed F. verticillioides genome indicated that FvGLLc1 is encoded by Locus: FVEG_12398.3 (located on Chromosome 4; Supercontig 19: complement 567878e568518) with the following annotations: gene name ¼ hypothetical protein similar to mannose-binding lectin; gene product name ¼ hypothetical protein similar to mannose-binding lectin. The sequence of FVEG_12398.3 perfectly matches that of the F. verticillioides EST except that it contains a 63-basepair intron inserted between the 2nd and 3rd nucleotide of a TAT codon (Y69). 2.2. DNA from axenically grown maize plants contains a sequence corresponding to the FvGLLc1 and ZmGLLc sequences The sequence encoding ZmGLLc was successfully amplified from DNA isolated from shoots of axenically grown maize plants. A nested PCR performed on this DNA using primers complementary to the 50 and 30 end of the EST yielded two fragments of 321 and 384 bp, respectively (Fig. 2A). Cloning and sequencing revealed that the 321 bp fragment was identical to the Z. mays (and F. verticillioides) ESTs whereas the 384 bp fragment corresponded to the genomic sequence of FvGLLc1 (Fig. 1B). The introns in ZmGLLc and FvGLLc1 have the same length and are located at exactly the same position in the sequence. Moreover the intron sequences of both

47

lectins differ by only 1 nucleotide (Fig. 1B). These findings imply that DNA extracted from shoots of axenically grown maize plants contains (i) a sequence that matches the ZmGLLc/FvGLLc ESTs and (ii) a sequence that is virtually identical to the genomic sequence of the FvGLLc1 gene (including the intron). 2.3. Southern blot analysis confirms the presence of ZmGLLc1 in the maize genome Genomic DNA from F. verticillioides and axenically grown maize and rice was digested with BamHI and HindIII, two restriction enzymes that will not cleave the FvGGLc1 sequence, and hybridized using the full-length ORF of FvGLLc1 as a probe. As expected two fragments of 3526 bp and 3767 bp were detected in the genome of F. verticillioides when genomic DNA was digested with BamHI and HindIII, respectively (Fig. 3, lane 1 and 2). Fragments larger than 10000 bp probably correspond to non-digested DNA. Hybridization of the FvGLLc1 probe to genomic DNA from maize and rice also yielded several bands. Digestion of genomic DNA from maize with BamHI yielded fragments of approximately 4000 and 5200 bp (Fig. 3 lane 3) whereas after cleavage with HindIII a fragment of approximately 1600 bp was observed (Fig. 3 lane 4). In the case of rice restriction fragments were estimated 4300 and 4800 bp after digestion with BamHI. Digestion with HindIII yielded one band of approximately 10000 bp. 2.4. The ZmGLLc sequences in Z. mays are not derived from a Fusarium contaminant In view of the high sequence identity between the lectin-like sequences from maize and Fusarium a PCR-based approach was used to investigate whether other Fusarium sequences could be found in the genomic DNA from maize. Though Fusarium verticillioides can be considered as a ’prime suspect’ (because of the sequence identity between the ZmGLLc and FvGLLc1 ESTs and the fact that F. verticillioides is the most commonly reported fungal parasite on maize) it is evident that other fungi cannot be excluded. Therefore a comprehensive overview of fungal EST/genomic sequences encoding orthologs of FvGLLc1 was made. As shown in Table 1, (multiple) GLLc orthologs were identified in several Fusarium/Gibberella/Nectria and Aspergillus/Neosartoya species. Sequence alignments revealed that the genome of Fusarium oxysporum (which is very closely related to F. verticillioides) contains a gene (FOXG_13130.2 in database http://www.broad.mit.edu/ annotation/genome/fusarium_group/, further referred to as FoGLLc1) sharing 87% sequence identity at the nucleotide level with FvGLLc1 (94% at the amino acid level; 8 amino acids difference with FvGLLc1) (Fig. 1A). Of all other identified fungal orthologs GzGLLc1 from Gibberella zeae shares the highest sequence identity with FvGLLc1. However, GzGLLc1 differs at 16 positions (out of 118) from FvGLLc1 (Fig. 1A; Suppl. Fig. 1). In F. verticillioides a second protein referred to as FvGLLc2 is expressed that definitely belongs to the family of GNA-like lectins but shares only 43% identity at the amino acid level. Similarly, in the genome of F. oxysporum a second GNAlike lectin was found; FoGLLc2 shows 95% sequence identity to FvGLLc2. The sequence divergence (at the nucleotide level) between the corresponding genes is even higher especially within the intron (Suppl. Fig. 1). Based on this analysis, one can reasonably conclude that only F. verticillioides or a very closely related species -assuming that the maize material is contaminated- can give rise to the ZmGLLc1 EST/genomic sequences. Hence the search for a possible fungal contaminant was focused on F. verticillioides/F. oxysporum. A series of PCR experiments was set up using primers designed to amplify specific fragments of the F. verticillioides/F. oxysporum

48

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

A

B

Fig. 1. (A) Sequence alignment of cytoplasmic GNA-like lectins from Zea mays (ZmGLLc), Fusarium verticillioides (FvGLLc1), Fusarium oxysporum (FoGLLc1) and Gibberella zeae (GzGLLc1). Note that for the sake of clarity the sequences of other fungal orthologs are not included. Identical residues are indicated by asterisks and similar residues by dashes or colons. The amino acids that constitute the carbohydrate binding sites are boxed grey. (B) ClustalW alignment of the nucleotide sequences of fragments covering the 50 and 30 untranslated region, the open reading frame and intron of the genes encoding ZmGLLc and FvGLLc1.

genome. In each experiment, the following DNAs were used as a template: (i, ii) genomic DNA of F. verticillioides and F. oxysporum (positive control), (iii) genomic DNA from axenically grown maize shoots and (iv) genomic DNA isolated from shoots of maize seedlings grown from non-sterilized seeds. PCR with primers designed to amplify the gene encoding the second GLLc-lectin of F. verticillioides and F. oxysporum (FvGLLc2 and FoGLLc2, respectively) yielded a fragment of 371 bp in the reaction mixtures containing F. verticillioides and F. oxysporum DNA (Fig. 2C). However, no fragment could be amplified in the reactions driven by the DNA samples from maize. Analysis of the reaction products for a second PCR experiment aiming at the amplification of part of the gene encoding the

F. verticillioides hydrophobin, a typical fungal protein absent from plants revealed that only the F. verticillioides DNA yielded a fragment of 436 bp after the first PCR. Upon repeating the PCR using the products of the first reaction as a template, the F. oxysporum DNA also yielded an amplified fragment of the same size but no amplified fragment was obtained in the reaction mixtures containing the maize DNAs (Fig. 2B). Finally, a third PCR experiment was carried out using primers developed to detect contamination of maize with (a broad range of) mycotoxigenic Fusarium species in an early stage [14]. Although a 200 bp fragment was readily amplified in the reaction mixtures containing F. verticillioides and F. oxysporum DNA, no fragment was detected after the first PCR reaction when maize DNAs were used as

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

49

maize yielded a weak PCR fragment of expected size, indicating that this batch might be contaminated (Fig. 2D). 2.5. The flanking regions of the F. verticillioides FvGLLc1 gene are absent from the maize DNA

Fig. 2. PCR analyses performed to check a possible contamination of maize with Fusarium. Amplification was done on genomic DNA isolated from non-sterilized maize germinated on wet paper (lane 1), from maize seedlings grown in vitro (lane 2), from mycelium of Fusarium verticillioides (lane 3) and Fusarium oxysporum (lane 4). M shows the GeneRulerÔ 100 bp DNA Ladder Plus from Fermentas A: amplification of the GLLcgene; B: amplification of a F. verticillioides hydrophobin gene fragment; C: Amplification of FvGLLc2; D: amplification of intergenic spacer of rDNA of Fusarium species.

a template. Dilution of the product of the first reaction and reamplification in a second round of PCR still yielded negative results for the DNA from the axenically grown maize. In contrast, under the same conditions the sample with the DNA from the non-sterile

To investigate whether the maize sequence is confined to the coding sequence (plus 50 and 30 UTR) or comprises also (part of) the flanking regions of the FvGLLc1 gene a series of PCR experiments were set up using primers designed to amplify fragments of the F. verticillioides genome at different positions upstream and downstream of the FvGLLc1 gene. A region spanning approximately 13 kb was chosen (Fig. 4A). PCR using primers designed to amplify fragments of the gene downstream (probably a kinase, 814 bp) and upstream (ortholog of a predicted protein of Gibberella zeae PH-1 (XP383866), 1011 bp) of FvGLLc1 yielded amplified fragments of the expected length in the assays with F. verticillioides DNA but not in assays with maize DNA (Fig. 4B). Since these results indicated that the maize DNA lacks the sequences corresponding to the F. verticillioides genes in the direct vicinity of the FvGLLc1 gene, the search for matching sequences was concentrated on a region spanning the coding sequence plus 1.3 kb upstream and 1.7 kb downstream. As shown in Fig. 4B, all primers combinations designed to amplify the region upstream (covering up to 1290 bp) and downstream (covering up to 1711 bp) of the coding region of the FvGLLc1 gene yielded fragments of the expected length in the PCR with F. verticillioides DNA, but no fragment was amplified when DNA from axenically grown maize was used as a template. Similarly a nested PCR to amplify (part of) the promoter sequence was successful with F. verticillioides DNA but not with DNA from axenically grown maize. 2.6. Sequences (nearly) identical to ZmGLLc/FvGLLc1 can be amplified from DNA of taxonomically far distant plants To check whether ZmGLLc sequences are more common in plants genomic DNA was purified from four monocots (Z. mays, Table 1 List of cytoplasmic GNA-like lectins from plants and fungi. Sequences are shown in Supplemental data. Species

Code

Accession number/Locus

Galanthus nivalis Cytoplasmic plant homologs Helianthus tuberosus Marchantia polymorpha Medicago truncatula Oryza sativa Picea abies Triticum aestivum Zea mays

GNA

AAA33346

HtGLLc MpGLLc MtGLLc OsGLLc PaGLLc TaGLLc ZmGLLc

DQ122761 EU327114 DQ122760 DQ122758 DQ122762 DQ122759 DQ122756

Aspfl1 Aspfl2 Aspfl3 Aspor1 Aspor2 Aspor3 FoGLLc1 FoGLLc2 NhGLLc

AFL2G_00504.2 AFL2G_11048.2 AFL2G_11049.2 AO 090005000502 AO 090020000246 AO 090020000245 FOXG_13130.2 (chr12) FOXG_16844.2 (chr13) (>jgi|Necha2|102073| estExt_fgenesh1_pg.C_ sca_40_chr12_1_00047) FVEG_12398.3 (F. verticillioides supercont_3.21 Nt: 94660e95061 þ) FGSG_03689.3 FGSG_11368.3 FGSG_03828.3

Cytoplasmic fungal homologs Aspergillus flavus

Aspergillus oryzae

Fusarium oxysporum Fusarium solani ¼ Nectria haematococca

Fig. 3. Southern blot analysis of genomic DNA. The blot was hybridized with the PCRamplified FvGLLc1 gene as probe. Genomic DNA loaded in the odd lanes was digested with BamH1. Genomic DNA loaded in the even lanes was digested with HindIII. Lane 1 and 2: genomic DNA of Fusarium verticillioides; Lane 3 and 4: genomic DNA of Zea mays; Lane 5 and 6: genomic DNA of Oryza sativa. < ¼ 3526 bp; = ¼ 3767 bp;  ¼ approximately 4000 bp; / ¼ approximately 5200 bp; + ¼ approximately 1600 bp.

Fusarium verticillioides

FvGGLc1 FvGGLc2

Gibberella zeae

GzGLLc1 GzGLLc2 GzGLLc3

50

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

A Ogz1 Ogz3

Us1 Us2

Us3 Pro1 Pro2 Zm1 Zm2

K1 K3

5’

3’ Zm4 Zm3 Ds1 Ds2 Ds3

Ogz4 Ogz2

K4

K2

12987 bp

Forward primers Gene upstream of FvGGLc1 gene

B

FvGGLc1 gene

Gene downstream of FvGGLc1 gene

Reverse primers

Amplification product

Size

F. verticillioides

Z. mays

Zm1-Zm3* / Zm2-Zm4

420 bp / 384 bp

+

+

Ogz1-Ogz2 / Ogz3-Ogz4

800 bp / 687 bp

+

-

K1-K2 / K3-K4

769 bp / 713 bp

+

-

Pro1-Zm3 / Pro2-Zm4

738 bp / 679 bp

+

-

Us1-Zm3 / Us1-Zm4

1709 bp / 1691 bp

+

-

Us2-Zm3 / Us2-Zm4

1369 bp / 1351 bp

+

-

Us3-Zm3 / Us3-Zm4

1150 bp / 1132 bp

+

-

Zm1-Ds1 / Zm2-Ds1

1111 bp / 1093 bp

+

-

Zm1-Ds2 / Zm2-Ds2

1551 bp / 1533 bp

+

-

Zm1-Ds3 / Zm2-Ds3

2131 bp / 2113 bp

+

-

* ZmGLLc was also amplified with the primers Zm1-Zm3 after 35 cycles. Fig. 4. (A) Schematic representation of the FvGLLc1 and neighboring genes together with the primers used to verify which part of Fusarium can be retrieved in maize. (B) Overview of PCR amplification products amplified from F. verticillioides and Z. mays (þ) amplification; () no amplification.

T. aestivum, Hordeum vulgare and Oryza sativa), two dicots (M. truncatula and Helianthus tuberosus), a gymnosperm (Picea abies) and a liverwort (Marchantia polymorpha), and used as a template in a PCR amplification experiment with primers corresponding to the 50 and 30 ends of the coding sequence of a maize cDNA encoding ZmGLLc. Analysis of the reaction mixtures revealed that all DNA samples yielded fragments of the expected size. Only in maize two fragments were amplified. Cloning and sequencing revealed high sequence conservation between the GLLc orthologs from different species (Fig. 5; Suppl. Fig. 2). For example, the proteins from flowering plants differ by only 2 residues from that of a gymnosperm (P. abies) (Fig. 5). In addition, the GLLc sequence of the liverwort M. polymorpha revealed only 8 different residues (Fig. 5). The PCR amplification experiments yielded also additional information about the occurrence of intron sequences. Maize contains lectin sequences with and without an intron. In the other species only a single intron-containing (barley, rice and M. polymorpha) or an intronless (P. abies, H. tuberosus, M. truncatula and wheat) sequence was identified. The intron is invariably located between the 2nd and 3rd nucleotide of the TAT codon corresponding to Y69 in ZmGLLc (Suppl. Fig. 2). To highlight the marked sequence identity between FvGLLc1 and its plant orthologs a phylogenetic analysis was made of all currently known (putative) fungal and plant GLLcs (and GNA itself as the prototype of the vacuolar plant homologs). As shown in Fig. 6, all plant GLLcs form a single cluster together with FvGLLc1, confirming that FvGLLc1 is more closely related to the plant

sequences than to its orthologs from closely related Fusarium species. Apart from this obvious phylogenetic anomaly, the rest of the fungal sequences exhibit a more or less ’normal’ phyologeny. The same applies to GNA itself that - as could be expected on the basis of the low residual sequence identity between the vacuolar and cytoplasmic plant GNA-like proteins - was placed in a separate branch at the other end of the dendrogram. 3. Discussion Transcriptome analyses revealed that F. verticillioides, which is a widespread polyphagous fungal pathogen and infects several important crop plants like maize, rice, wheat and barley, expresses a protein identical to a recently identified cytoplasmic/nuclear GNAlike lectin from maize [13]. Though the maize genome is still not completed, the apparent absence of a matching genomic sequence raises the question of the origin of the ZmGLLc encoding ESTs found in the Z. mays transcriptome. Taking into account that (i) these ESTs are extremely rare (only 4 out of a total of >700,000 maize ESTs) and, in addition, (ii) are virtually identical to the far more abundant ESTs found in the transcriptome of a major fungal pathogen of maize, there is certainly a possibility that the maize ESTs are derived from F. verticillioides or another fungal contaminant present in the tissues from which the cDNA libraries were prepared. To investigate whether the ZmGLLc/FvGLLc1 sequences present in the maize DNA are possibly derived from F. verticilloides the GLLc sequence was amplified from genomic DNA from axenically grown

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

51

Fig. 5. Multiple sequence alignment of the amino acid sequences of cytoplasmic GNA-like lectins from Zea mays (ZmGLLc), Triticum aestivum (TaGLLc), Oryza sativa (OsGLLc), Medicago truncatula (MtGLLc), Hordeum vulgare (HvGLLc), Helianthus tuberosus (HtGLLc), Picea abies (PaGLLc) and Marchantia polymorpha (MpGLLc1). Sequences were deduced from the nucleotide sequences obtained by PCR with primers complementary to the 50 and 30 end of the EST encoding ZmGLLc. Identical residues are indicated by asterisks. Amino acids different from the amino acids in ZmGLLc are boxed. The amino acids that constitute the carbohydrate binding sites are boxed grey.

Fig. 6. Phylogenetic tree showing the relationships between the amino acid sequences of GNA-like lectins found in plants and fungi. Lectin abbreviations and accession numbers or loci can be found in Table 1. All sequences used for the construction of the dendrogram are listed in supplementary data. Bootstrap values higher than 50% are indicated (n ¼ 1000).

maize. PCR yielded two fragments with almost identical sequences to the Fusarium sequence. Surprisingly the smallest fragment contained an intronless sequence which is clearly absent in all Fusarium sequences. The combination of intron containing as well as intronless sequences in the DNA of axenically grown maize plants argues for a plant origin of the ZmGLLc. Since the genomic sequences of all identified fungal GLLc genes contain a single intron at a strictly conserved position (inserted between the 2nd and 3rd nucleotide of a TAT or TAC codon corresponding to a conserved Y residue), it is very unlikely, indeed, that the intronless amplification products are derived from a fungal template. Though indicative, the results of these PCR experiments provide no definitive proof for a plant origin of the amplified fragments. First, it cannot be guaranteed that the axenically grown maize seedlings are completely free of fungi. Second, it cannot be excluded that some close relative of F. verticillioides contains an intronless ortholog of the FvGLLc1 gene. Therefore, additional experiments were set up. Southern blot analysis using the Fusarium sequence as a probe revealed clear hybridization signals not only with genomic DNA of Fusarium but also with genomic DNA of axenically grown maize and rice. Since the hybridization patterns obtained for maize and rice are clearly different from that of Fusarium these results also argue against the hypothesis that the maize sequences result from a fungal contamination. Additional arguments against a contamination of the axenically grown maize plants used in our experiments by F. verticilloides or a closely related species was obtained from a series of PCR experiments which revealed that the sequences encoding fungal proteins such as a second lectin from Fusarium and hydrophobin could not be amplified from DNA purified from axenically grown maize. Similarly, a PCR assay to amplify the sequence corresponding to the intergenic spacer of rDNA of Fusarium [14] yielded negative results. Furthermore PCR amplification of the genomic region around the GLLc sequences in Fusarium and maize indicated that the sequence(s) in the maize DNA which comprise(s) the ZmGLLc sequence cover the coding sequence but not the promoter region of the F. verticillioides FvGLLc1 gene. In addition, the failure to amplify fragments covering the regions upstream and downstream of the coding sequence confirmed that the ZmGLLc sequences in maize are not due to contamination with F. verticillioides. Using a PCR-based approach it could also be demonstrated that this cytoplasmic GNA-like lectin occurs in plants from distant taxa and is markedly conserved. In contrast to plants, identical or nearly

52

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

identical lectins are not common in fungi. Several other related lectins could be identified in fungal species such as G. zeae, F. oxysporum and A. oryzae but these proteins share only a limited sequence identity with the plant GLLc. Though the occurrence of FvGLLc1 orthologs in other fungi cannot be excluded on the basis of the available sequence information, the present data argue against a widespread occurrence of GLLc orthologs in fungi. PAUP analysis of the available fungal and plant sequences revealed two major phylogenetic anomalies. First, the occurrence of nearly identical GLLc genes in plants from far distant taxa (angiosperms, gymnosperms and liverworts) markedly contrasts the obvious intra- and inter-species sequence heterogeneity of the vacuolar GNA homologs [3]. Second, the presence of a virtually identical gene in F. verticillioides and several plants is even more enigmatic. A horizontal transfer of a lectin gene from a plant to F. verticillioides would nicely explain the clustering of FvGLLc1 and the plant orthologs but cannot be reconciled with the position of the other fungal sequences in the dendrogram. Unfortunately an analysis of the codon usage pattern between FvGLLc and ZmGLLc was not conclusive because the difference in the codon usage between maize and Fusarium is too low. Screening of the publicly accessible plant genomic databases did not yield any sequence identical or similar to ZmGLLc. The apparent absence of these sequences from the genome databases of wheat, maize, rice and M. truncatula is difficult to explain if one assumes that the genes encoding the ZmGLLc orthologs are incorporated in the nuclear or organellar genome. Since it is rather unlikely that the same gene(s) were ’lost’ in multiple independent genome sequencing programs, there might be a need for an alternative explanation. As is discussed above, fungal contamination can be excluded as a source of the ZmGLLc sequences in the DNA of maize and M. polymorpha, and is very unlikely for the other species. Moreover, taking into account that the M. polymorpha DNA was prepared from an axenic culture (which has been maintained over a period of many years) any parasitic microbial contamination can be excluded in this particular case. Even if one assumes that the axenic culture was infected with an endosymbiotic prokaryote, it is unlikely that it is the source of the intron containing ZmGLLc sequence. 4. Methods 4.1. Plant material Maize (Zea mays) seeds were surface sterilized for 4 min with 70% ethanol and extensively rinsed with sterile water. To ensure sterility, the seeds were subsequently immersed in 5% (v/v) sodium hypochlorite for 20 min. After several rinses in sterile water, the seeds were transferred into glass jars filled with Murashige and Skoog (MS) medium containing 4.3 g/l MS micro and macro nutrients with vitamins (Duchefa, Haarlem, The Netherlands), 30 g/l sucrose, adjusted to pH 5.7 with 1 M NaOH and 8 g/l agar (Duchefa). Seedlings were kept in a growth chamber at 25  C, 70% relative humidity and a 16 h photoperiod for 3 weeks. Untreated seeds were germinated in 150 mm diameter Petri dishes on filter paper soaked in water. An axenic culture of Marchantia polymorpha was maintained by transferring calli onto fresh MS medium every two weeks. Buds of Picea abies and Helianthus tuberosus tubers were collected locally. Seeds of Triticum aestivum, Hordeum vulgare, Oryza sativa and Medicago truncatula were sterilized as described above for the maize seeds. 4.2. Retrieval of sequences Sequences encoding GLLc-like proteins were retrieved by BLAST searches using both the amino acid and nucleotide

sequences of the Zea mays GLLc (ZmGLLc) as a query. In a later phase all newly identified fungal sequences were also used as queries. All retrieved EST sequences were analysed individually. In the absence of complete ESTs, contigs were reconstructed from ESTs showing overlaps of at least 200 identical nucleotides. Searches were completed on April 30, 2009. The following databases were screened for the presence of EST and/or genomic sequences encoding fungal proteins sharing sequence similarity with FvGLLc1: 1. National Center for Biotechnology Information (NCBI: http:// www.ncbi.nlm.nih.gov/): 38 fungal genomes and all fungal ESTs 2. Consortium for the Functional Genomics of Microbial Eukaryotes (COGEME: http://www.cogeme.man.ac.uk/): Phytopathogenic Fungi and Oomycete EST Database 3. The Institute for Genomic Research (TIGR: http://tigrblast.tigr. org/tgi/): TIGR Gene Indices 4. Saccharomyces Genome Database (SGDÔ: http://www. yeastgenome.org/): EST database 5. Aspergillus oryzae EST Database (www.nrib.go.jp/ken/EST/db/) 6. Aspergillus database (http://www.broad.mit.edu/annotation/ genome/aspergillus_group/) 7. Fusarium database (http://www.broad.mit.edu/annotation/ genome/fusarium_group/) 8. DOE Joint Genome Institute (http://www.jgi.doe.gov)

4.3. PCR amplification of genomic DNA fragments DNA was extracted from young shoots of axenically grown Zea mays, Triticum aestivum, Hordeum vulgare, Medicago truncatula, Oryza sativa, buds from healthy Helianthus tuberosus tubers, young buds of Picea abies and callus of Marchantia polymorpha using the protocol described by Stewart and Via [15]. Fusarium verticillioides (strain FGSC 7600) and Fusarium oxysporum were grown on potato dextrose agar medium for 5 days and the DNA was extracted using the FastDNA Spin kit in an automatic homogenizer (FastPrep Instrument, MP Biomedicals and Qbiogene, Irvine, CA, USA) following the manufacturer’s recommendations. Genomic sequences encoding GLLc orthologs were amplified by a nested PCR using primers derived from the N - (Zm1: 50 ATGGGTTACGGCACTCTCGA 30 and Zm2: 50 GACAACGGCGACTGGCTCAT 30 ) and C-terminal (Zm3: 50 TTATTTGTTGCTCTTGGAAG 30 and Zm4: 50 AGACCAGATGGGAGTAGCTC 30 ) sequences of the Zea mays EST MEST340-H06.T3 ISUM5-RN (gi|18176447|gb|BM351398.1| BM351398). The reaction mixture for amplification of genomic DNA sequences contained 10 DNA polymerase buffer, 1.5 mM MgCl2, 0.4 mM of each dNTP, 2.5 units of Taq polymerase (Invitrogen, Carlsbad, CA), 200 ng of genomic DNA and 2 mL of the appropriate primer mixtures (5 mM), in a 50 mL reaction volume. After denaturation of the DNA for 2 min at 94  C amplification was performed for 25 cycles through a regime of 15 s template denaturation at 94  C followed by 30 s primer annealing at 47  C and 1 min primer extension at 72  C using a Perkin Elmer DNA Thermal Cycler 9600. The cycling conditions of the nested PCR were the same as in the first PCR, except that the annealing temperature was 55  C. The PCR fragments were purified using Qiaquick PCR Purification kit (Qiagen, Hilden, Germany) and cloned in TOPO pCR2.1-TOPO cloning vector using the TOPO cloning kit from Invitrogen (Carlsbad, CA). Plasmids were isolated from purified single colonies on a miniprep scale using the alkaline lysis method [16] and sequenced by the dideoxy method [17]. Introns were identified by alignment of the nucleotide sequences of the corresponding ESTs and the PCR product amplified from the genomic DNA.

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

4.4. Southern blot analysis Genomic DNA was isolated from Fusarium verticillioides mycelium, and axenic maize and rice shoots. Plant and fungal materials were frozen and ground in liquid nitrogen using a mortar and pestle. The finely powdered material was transferred into 50 ml falcon tubes containing 5 ml g1 tissue extraction buffer (100 mM TriseHCl, pH 8 containing 20 mM EDTA, 1.4 M NaCl, 2% cetyl trimethyl ammonium bromide, and 2% polyvinylpyrrolidone 40). The powder was thoroughly mixed with extraction buffer, incubated at 60  C for 1 h, and the DNA isolated using a phenol/chloroform extraction protocol [15]. Purified genomic DNA (10 mg) was completely digested with BamHI or HindIII in a single digest at 37  C for 7 h. Separation of the digest in a 0.8% TBE agarose gel was followed by an overnight blotting onto Hybond-Nþ nylon membrane (Amersham Biosciences) using a standard protocol [18]. The FvGLLc1 gene amplified with the primers Zm1 and Zm3 (i.e. a sequence covering most of the ORF and the intron) was used as a probe and labelled with (a-32P)-dCTP using the DecaLabelÔ DNA Labelling kit (Fermentas St. Leon-Rot, Germany) according to the manufacturer’s instructions. After prehybridization for 3 h the membrane was hybridized overnight at 60  C following standard protocols [18]. The hybridized membrane was washed twice in 2  saline sodium citrate (SSC) for 15 min, 2  SSC þ 0.1% SDS for 30 min, 0.5  SSC for 15 min and 0.1  SSC for 10 min, each at 60  C. The membrane was exposed to a radioactive sensitive screen and scanned with the FujiFilm Fluorescent Image Analyzer FLA-5100 (FUJI, Dusseldorf, Germany). 4.5. Control experiments performed to check for possible contamination of maize with Fusarium Genomic DNA of Fusarium verticillioides, Fusarium oxysporum, in vitro grown maize and non-sterilized maize shoots was isolated using the FastSpin DNA kit (FastPrep Instrument, MP Biomedicals and Qbiogene, Irvine, CA, USA) following the manufacturer’s recommendations. All PCR reactions were performed in a 50 ml reaction volume containing 200 ng template, 10 DNA polymerase buffer, 1.5 mM MgCl2, 10 mM dNTP’s, 5 mM of each primer and 1.25 U Taq DNA Polymerase (Invitrogen) and run in an AmplitronIIR Thermolyne apparatus (Dubuque, Iowa, USA). Amplification of a second lectin FvGLLc2 was achieved using a nested PCR with specific forward primers FvL1 (50 ATGGGCAGACT CGACAACGA30 ) and FvL2 (50 CAACTGGCTGTACCCAGGTG30 ) and reverse primers FvL3 (50 TTAAATCTGGTTGGTGCTAG 30 ) and FvL4 (50 TCGCCCAAACGGCCTCAGCC 30 ). The PCR programs consisted of 25 repetitive cycles with a denaturation step at 94  C for 15 s, an annealing step at 50  C for 30 s and an elongation step at 72  C for 1 min. The cycles were preceded by an extra denaturation step of 2 min at 94  C and were ended by an extra elongation step of 5 min at 72  C. Amplification of a hydrophobin from the plant parasitic fungus F. verticillioides was performed with primers complementary to the 50 (FvH1) and 30 (FvH2) end of an EST encoding this hydrophobin (DR663129) (FvH1 ¼ 50 ATGCAGTACATGACCATCGTCG 30 and FvH2 ¼ 50 TCAGATAAGGCTGCTGAGAGCA 30 ). Thermal cycling conditions were: 2 min denaturation at 94  C followed by 25 cycles (15 s 94  C, 30 s 50  C, 1 min 72  C); the final elongation reaction was performed at 72  C for 5 min. This first PCR product was diluted 1:50 and used as template for a second PCR. The specific conditions and the primers for this second PCR were as described for the primary PCR. A PCR for detection of contamination with Fusarium species was adapted from Jurado et al. [14]. This PCR assay is based on IGS

53

sequence (intergenic spacer of rDNA), a multi-copy region in the eukaryotic genome that permits to enhance the sensitivity of the assay in comparison with PCR assays based on single-copy sequences. The forward primer Fus1 (50 CGCACGTATAGATGGACAAG 30 ) was combined with the reverse primer Fus2 (50 GGCGAAGGACGGCTTAC 30 ) in the following amplification protocol: 2 min 94  C, 25 cycles (35 s 94  C, 30 s 67  C, 30 s 72  C), 5 min 72  C. The PCR was performed twice; the template of the second PCR was a 1:50 dilution of the PCR product of the first PCR. 4.6. PCR amplification of genomic region around FvGLLc1 gene in Fusarium and maize Nested PCRs were performed on genomic DNA isolated from Fusarium verticillioides and Zea mays to amplify the genomic sequence upstream and downstream of the FvGLLc1 gene. The gene downstream of FvGLLc1 gene, annotated as a kinase was amplified using the primers K1 (50 CATTTAACCATAGCTGCCCCGTTC 30 ), K2 (50 AGTTGCTGGAGCTCCTTGATCAAGG 30 ) and K3 (50 GTCATCCTGGTT CTTCCGTAACTC 30 ), K4 (50 CCATTGATATCAAGCATGGC GGCTG 30 ). The gene upstream of FvGLLc1 is annotated as an expressed ortholog of a predicted protein of Gibberella zeae PH-1 (XP383866) and was amplified with the primers Ogz1 (50 GCGATATCT GGTGAATAGCG 30 ), Ogz2 (50 ACCGAACTTTCTTACTGTTG 30 ) and Ogz3 (50 CTGACCAGAAACGACTTCAT 30 ), Ogz4 (50 TTCGTCAACAAACGGCTTTG 30 ). In addition, three forward primers were designed in the region upstream from the FvGLLc1 gene: Us1 (50 CCAGGTTTATGTCTTGTCAG 30 ), Us2 (50 AGTCACAACGCTTGTCGGAA 30 ) and Us3 (50 CAACCTGTCACAAATTCACA 30 ) and three reverse primers downstream the FvGLLc1 gene: Ds1 (50 ACAGGCACTCACTTGCGTCA 30 ), Ds2 (50 GCCAGATATA AATACCGAGA 30 ) and Ds3 (50 CGGAACCTCCATGACATGAT 30 ). A schematic overview of the different primers is shown in Fig. 4A, and the different primer combinations tested are listed in Fig. 4B. All PCR programs consisted of 25 repetitive cycles (15 s 94  C, 30 s 50  C, 1 min 72  C), preceded by a denaturation step of 2 min at 94  C and ended by an elongation step of 5 min at 72  C. The entire FvGLLc1 gene together with part of its promoter was amplified using forward primers which were developed in the conserved promoter region of FvGLLc1 and GzGLLc1. PCR was performed using the following two primer pairs: forward primer Pro1, 50 -CGGAGAGGAGAAGATCATGT-30 and reverse primer Zm3; forward primer Pro2, 50 -TGATACGTGCAACTACATCATG-30 and reverse primer Zm4. Cycling parameters were as follows: 2 min 94  C, 25 cycles (15 s 94  C, 30 s 50  C, 1 min 72  C), 5 min 72  C. 4.7. Phylogenetic analysis Protein sequences were aligned using the ClustalW program [19]. Parsimony analyses on the alignments were conducted with PAUP* (phylogenic analysis using parsimony) version 4.0b10 [20]. Nonparametric bootstrap support was obtained by resampling the data 1000 times using parsimony. Heuristic searching used 100 random taxon addition replicates, holding 100 trees at each step, tree bisection-reconnection branch swapping, MulTrees, Collapse, and Steep Descent options, and no upper limit for trees held in memory. The phylogenetic tree is visualized using Treeillustrator [21]. Acknowledgments This research was supported by the Fund for Scientific ResearchFlanders (FWO grants G.0201.04 and G.0022.08). We thank Prof. K. McCluskey (University of Missouri, Kansas City, MO USA) for the Fusarium verticillioides strain FGSC 7600, Prof. M. Höfte (University

54

E. Fouquaert et al. / Plant Physiology and Biochemistry 49 (2011) 46e54

of Ghent, Belgium) for providing Fusarium oxysporum, Prof. H. Hamada (Okayama University, Okayama, Japan) for giving the axenic culture of Marchantia polymorpha and Prof. B. Hause (Leibniz Institute für Pflanzenbiochemie, Halle, Germany) for the Medicago truncatula seeds. We are grateful to Prof. P.S. Schnable (Iowa State University, USA) for providing the EST clone encoding ZmGGLc (Genbank Accession No. BM351398). Appendix. Supplementary material Supplementary material associated with this paper can be found, in the online version, at doi:10.1016/j.plaphy.2010.09.018. References [1] E.J.M. Van Damme, N. Lannoo, W.J. Peumans, Plant lectins, Adv. Bot. Res. 48 (2008) 107e209. [2] E.J.M. Van Damme, A.K. Allen, W.J. Peumans, Isolation and characterization of a lectin with exclusive specificity towards mannose from snowdrop (Galanthus nivalis) bulbs, FEBS. Lett. 215 (1987) 140e144. [3] E.J.M. Van Damme, P. Rougé, W.J. Peumans, Carbohydrate-protein interactions: plant lectins. in: J.P. Kamerling, G.J. Boons, Y.C. Lee, A. Suzuki, N. Taniguchi, A.J.G. Voragen (Eds.), Comprehensive Glycoscience-From Chemistry to Systems Biology. Elsevier, New York, 2007, pp. 563e599. [4] E. Fouquaert, S.L. Hanton, F. Brandizzi, W.J. Peumans, E.J.M. Van Damme, Localization and topogenesis studies of cytoplasmic and vacuolar homologs of the Galanthus nivalis agglutinin, Plant. Cell. Physiol. 48 (2007) 1010e1021. [5] E.J.M. Van Damme, H. Kaku, F. Perini, I.J. Goldstein, B. Peeters, F. Yagi, B. Decock, W.J. Peumans, Biosynthesis, primary structure and molecular cloning of snowdrop (Galanthus nivalis L.) lectin, Eur. J. Biochem. 202 (1991) 23e30. [6] E.J.M. Van Damme, A. Barre, P. Rougé, W.J. Peumans, Cytoplasmic/nuclear plant lectins: a new story, Trends. Plant. Sci. 9 (2004) 484e489. [7] E. Fouquaert, W.J. Peumans, E.J.M. Van Damme, Confocal microscopy confirms the presumed cytoplasmic/nuclear location of plant, fish and fungal orthologs of the Galanthus nivalis agglutinin, Commun. Agric. Appl. Biol. Sci. 71 (2006) 141e144.

[8] E. Jung, P. Fucini, M. Stewart, A.A. Noegel, M. Schleicher, Linking microfilaments to intracellular membranes: the actin-binding and vesicle-associated protein comitin exhibits a mannose-specific lectin activity, EMBO. J. 15 (1996) 1238e1246. [9] M. Machida, K. Asai, M. Sano, et al., Genome sequencing and analysis of Aspergillus oryzae, Nature 22 (2005) 1157e1161 (64 co-authors). [10] S. Tsutsui, S. Tasumi, H. Suetake, Y. Suzuki, Lectins homologous to those of monocotyledonous plants in the skin mucus and intestine of pufferfish, Fugu rubripes, J. Biol. Chem. 278 (2003) 20882e20889. [11] S. Tsutsui, M. Okamoto, S. Tasumi, H. Suetake, K. Kikuchi, Y. Suzuki, Novel mannose-specific lectins found in torafugu, Takifugu rubripes, Comp. Biochem. Physiol. 1D (2006) 122e127. [12] M. Wiens, S.I. Belikov, O.V. Kaluzhnaya, A. Krasko, H.C. Schröder, S. PerovicOttstadt, W.E. Müller, Molecular control of serial module formation along the apical-basal axis in the sponge Lubomirskia baicalensis: silicateins, mannosebinding lectin and mago nashi, Dev. Genes. Evol. 216 (2006) 229e242. [13] E. Fouquaert, D.F. Smith, W.J. Peumans, P. Proost, J. Balzarini, S.N. Savvides, E.J.M. Van Damme, Related lectins from snowdrop and maize differ in their carbohydrate-binding specificity, Biochem. Biophys. Res. Commun. 380 (2009) 260e265. [14] M. Jurado, C. Vázquez, S. Marín, V. Sanchis, M.T. González-Jaén, PCR-based strategy to detect contamination with mycotoxigenic Fusarium species in maize, Syst. Appl. Microbiol. 29 (2006) 681e689. [15] C.N. Stewart Jr., L.E. Via, A rapid CTAB DNA isolation technique useful for RAPD fingerprinting and other PCR applications, Biotechniques 14 (1993) 748e750. [16] R.C. Mierendorf, D. Pfeffer, Direct sequencing of denatured plasmid DNA, Meth. Enzymol. 152 (1987) 556e562. [17] F. Sanger, S. Nicklen, A.R. Coulson, DNA sequencing with chain-terminating inhibitors, Proc. Natl. Acad. Sci. U S A 74 (1977) 5463e5467. [18] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning, a Laboratory Manual, second ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 1989. [19] R. Chenna, H. Sugawara, T. Koike, R. Lopez, T.J. Gibson, D.G. Higgins, J.D. Thompson, Multiple sequence alignment with the Clustal series of programs, Nucleic. Acids. Res. 31 (2003) 3497e3500. [20] D.L. Swofford, PAUP*: Phylogenetic Analysis Using Parsimony (* and other methods) version 4. Sinauer Associates, Sunderland, MA, 2002. [21] G. Trooskens, D. De Beule, F. Decouttere, W. Van Criekinge, Phylogenetic trees: visualizing, customizing and detecting incongruence, Bioinformatics 21 (2005) 3801e3802.