Báo cáo khoa học: Survey of the Botrytis cinerea chitin synthase multigenic family through the analysis of six euascomycetes genomes

doi:10.1111/j.1432-1033.2004.04135.x

Eur. J. Biochem. 271, 2153–2164 (2004) (cid:1) FEBS 2004

Survey of the Botrytiscinereachitin synthase multigenic family through the analysis of six euascomycetes genomes

Mathias Choquer1, Martine Boccara2, Isabelle R. Gonc¸ alves3, Marie-Christine Soulie´ 2 and Anne Vidal-Cros1 1UMR 7613 CNRS/Universite´ Paris VI; 2UMR 217 INRA/INAP-G/Universite´ Paris VI; 3Structure et Dynamique des Ge´nomes, Institut Jacques Monod, Universite´s Paris VI/Paris VII, Paris, France

highlight three different transmembrane topologies of the CHS membranous isoenzymes. We found that the N-ter- minal region of the BcCHSI isoenzyme, and its orthologues in other euascomycetes, probably contain folded peptide motifs with conserved tyrosine residues. Their putative role is discussed. The BcCHSVII isoenzyme appeared to belong to a new class of CHS orthologues that was demonstrated by phylogenetic study to branch apart from division 1 and 2 of CHS.

Keywords: filamentous fungus; glycosyltransferase; hydro- phobic cluster analysis; phylogeny; plant pathogen.

We describe a strategy for systematic amplification of chitin synthase genes (chs) in the filamentous ascomycetes plant- pathogen Botrytis cinerea using PCR with multiple degen- erate primers designed on specific and conserved sequence motifs. Eight distinct chs genes were isolated, named Bcchs I, II, IIIa, IIIb, IV, V, VI and VII. They probably constitute the entire chs multigenic family of this fungus, as revealed by careful analysis of six euascomycetes genomes. Bcchs I, IIIa, IIIb, IV and VI genes were subjected to DNA walking and their deduced amino acid sequences were compared by hydrophobic cluster analysis (HCA) to localize putative residues critical for CHS activity. HCA also enabled us to

The cell wall of filamentous fungi is a protective structure mainly composed of the interconnected polysaccharides chitin, b-glucan, mannan and galactofuranose [1]. Chitin, the b-1,4 homopolymer of N-acetylglucosamine (GlcNAc), represents only 10–20% of the cell wall components. However, several mutants altered in their chitin content were severely impaired in hyphal growth and conidiophore development (for a recent review see [2]) or virulence in the case of pathogenic fungi [3,4]. Chitin biosynthesis thus appears to be a possible target for antifungics with the remarkable characteristic of being absent in the plant and mammal hosts.

organization composed of the catalytic domain surrounded by an hydrophilic N-terminus region and a hydrophobic C-terminus region [5]. Classes IV, V and VI, which belong to division 2, display an original modular organization with the same catalytic domain preceded by a cytochrome b5-like domain, and a myosin head-like domain observed for classes V and VI [6–8]. Filamentous fungi appear to contain up to 10 CHS isoenzymes spread among all classes of the two divisions. In contrast, the yeasts Saccharomyces cere- visiae and Candida albicans contain, respectively, three and four isoenzymes distributed in classes I, II and IV only [9,10]. Although considerable efforts have been directed towards the characterization of the CHS system in filamen- tous fungi, most of our knowledge comes from yeasts. In S. cerevisiae, class I is a repair enzyme during budding, class II is involved in primary division septa and class IV generates chitin at bud scars and in the lateral wall [2].

Chitin biosynthesis is performed by a set of membranous isoenzymes called chitin synthases (CHS; EC 2.4.1.16). These enzymes catalyse the multiple transfer of GlcNAc residues from UDPGlcNAc onto the nonreducing end of the growing chitin chain. Chitin synthases have been classified into six classes divided into two divisions, accord- ing to protein sequence similarities. Classes I, II and III, which belong to division 1, share a common protein

Botrytis cinerea (teleomorph Botryotinia fuckeliana) is a filamentous ascomycetes fungus responsible for the grey mold disease which causes significant damage to many crops worldwide. It is considered as one of the 10 most aggressive fungi on agronomic plants, which compel researchers to develop specific strategies. Given that B. cinerea strains developed resistance to many of the actual chemicals [11], we are currently investigating chitin biosynthesis in this fungus as a parallel target for fungicides. B. cinerea belongs to the leotiomycetes family for which chitin biosynthesis has been poorly documented. Indeed, only two chs genes were studied in the leotiomycetes Blumeria graminis [12].

In this paper, we present the survey of the chitin syn- thase multigenic family of the plant pathogen B. cinerea. We took advantage of the recent availability of some euasco- mycetes genomes to rationally design degenerate primers

Correspondence to A. Vidal-Cros, UMR 7613 CNRS/Universite´ Paris VI, 4 place Jussieu 75005, Paris, France. Fax: + 33 1 4427 7150, Tel.: + 33 1 4427 5159, E-mail: vidalc@ccr.jussieu.fr Abbreviations: ARE, AbaA response element; CHS, chitin synthase; CON, conserved region among CHS proteins; GlcNAc, N-acetyl- glucosamine; HAS, hyaluronate synthase; HCA, hydrophobic cluster analysis; MR, membrane-associated region; PMR, putative mem- brane region; PTR, putative transmembrane region; SGC, SpsA GnTI core; TBS, tyrosine-based signal. Enzyme: chitin synthases (EC 2.4.1.16). (Received 8 January 2004, revised 8 March 2004, accepted 31 March 2004)

2154 M. Choquer et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

corresponding to highly specific motifs. A PCR approach allowed us to identify eight chitin synthase genes in B. cine- rea. Besides revealing remarkable motives, sequence analysis led us to update the classification of chitin synthases.

provided in the kit. Outer and nested gene-specific primers were designed to the 5¢ and 3¢ ends of a known chs sequence with the intention of amplifying the adjacent unknown sequences amplified on the library with adaptor primers. For amplification reactions, we processed as described by the manufacturer.

Materials and methods

RT-PCR

Genomic DNA preparation and Southern analysis

The culture of Botrytis cinerea (Bd90 strain) was performed as described previously [13]. Genomic DNA was isolated from 1 g of mycelium using the Plant DNA isolation Kit (Roche). For Southern analysis, genomic DNA was diges- ted with HindIII, EcoRI and BamHI restriction enzymes and treated as described previously [14].

We used the QIAGEN(cid:4) OneStep RT-PCR kit. A 5¢ cDNA portion of the BcchsI gene was amplified with the for- ward primer (5¢-TCGATACCATGTCCTATAATCG) and reverse primer (5¢-TGGGATTGAGGGACATGCG CTGATGG). Amplification was performed on total RNA and yielded a 294 bp product. For expression analysis, we performed RT-PCR experiments as described previously [13].

Primer design and PCR amplification for chs identification

Cloning of PCR products

PCR products were separated on 1.2% agarose gels, excised and purified using the Ultrafree(cid:4)-DA centrifugal filter device (Millipore). Fragments were ligated using T/A strategy into the pGEM-T Easy Vector System 1 (Promega). Escherichia coli DH5a was transformed with the resulting plasmids by the CaCl2 method. Plasmids were extracted by the Wizard(cid:4) Plus Minipreps DNA purification System (Promega).

DNA sequencing and sequence accession numbers

Six sets of degenerate primers were designed to match highly conserved motifs within chitin synthases from euascomyc- etes species (Table 1). Amplifications were carried out in a 50 lL total volume containing 1 lM of each primer, 200 lM of each dNTP, 1 unit of Taq DNA polymerase (Sigma) in the corresponding amplification buffer and from 400– 800 ng of genomic DNA. Amplification conditions consis- ted in an initial denaturation step of 3 min at 94 (cid:2)C followed by 30 cycles of 1 min at 94 (cid:2)C, 1 min at 45 (cid:2)C, 50 (cid:2)C, 58 (cid:2)C, 55 (cid:2)C, 50 (cid:2)C, 50 (cid:2)C, 50 (cid:2)C for BcchsI, II, IIIa, IIIb, IV, V, VI and VII, respectively, and 1 min 30 s at 72 (cid:2)C, which were followed by a final extension step at 72 (cid:2)C for 7 min. PCR amplifications were performed in a GeneAmp PCR System 2400 thermal cycler (PerkinElmer).

Recombinant plasmids were prepared using the Wizard(cid:4) Plus Midipreps DNA purification System (Promega) and were subjected to DNA sequencing by ESGS Cybergene (Evry, France) group and Millegen Biotechnologies (Labe` ge, France). The encoded amino acid sequences were TM software (CEA, Saclay, predicted using DNA STRIDER France). Nucleotide and amino acid sequences for BcchsI, BcchsII, BcchsIIIa, BcchsIIIb, BcchsIV, BcchsV, BcchsVI and BcchsVII were assigned GenBank accession numbers AY515145, AY515141, AF494188, AF529208, AF215732, AY515142, AY515144 and AY515143, respectively.

DNA walking We used the Universal GenomeWalkerTM Kit (Clontech). Four libraries of adaptor-ligated genomic fragments were constructed with 1 mgÆmL)1 of B. cinerea genomic DNA restricted by blunt-end DraI, EcoRV, PvuII or StuI endonucleases and ligated to the GenomeWalker adaptor

Table 1. Primers used for B. cinerea chs genes amplification. F, forward primer; R, reverse primer.

Class Primer Motif Use Reference 5¢)3¢

[5]

a New class of gene confirmed in this study.

I and III 1 2 3 TMYNED TMYNED QNFEY CTG AAG CTT ACN ATG TAY AAY GAR GAT CTG AAG CTT ACN ATG TAY AAY GAR GAC GTT CTC GAG YTT RTA YTC RAA RTT YTG F F R II This study 4 5 FKGAEKG ADVFTANMY TTY AAR GGN GCN GAR AAR GG TAC ATR TTN GCN GTR AAN ACR TCN GC F R IV [14] 6 7 QVFEY WKFDDF CTG AAG CTT CAR GTN TTY GAR TA GTT CTC GAG AAR TCR TCR AAY TTC CA F R V This study 10 11 EIAEICQ IANQFDSKCV GAR ATH GCN GAR ATH TGY CA ACR CAY TTN SYR TCR AAY TGR TTN GCD AT F R VI This study 8 9 KVYSGLYE MIQVYEYYI AAR GTN TAY WSN GGN YTN TAY GA ATR TAR TAY TCR TAN ACY TGD ATC AT F R This study VIIa 12 13 QNF(MIV)YDME ITNEVCM CAR AAY TTY RTN TAY GAY ATG GA CAT RCA NAC YTC RTT NGT DAT F R

The chitin synthases of Botrytis cinerea (Eur. J. Biochem. 271) 2155

(cid:1) FEBS 2004

Sequence analysis

(DNA STRIDER

cluster analysis

taxa, different between fungi belonging to different especially between ascomycetes and basidiomycetes fungi (M. Choquer, I. R. Gonc¸ alves, M. Boccara and A. Vidal- Cros, unpublished results). Ascomycetes are composed of around 90% filamentous fungi called euascomycetes, among them B. cinerea, and 10% yeasts, divided into hemiascomycetes and archiascomycetes. We focused our analysis on euascomycetes genomes, because yeasts pos- sess a smaller number of chs genes. We thus discarded from this analysis the genomes of the hemiascomycetes Saccharomyces cerevisiae and Candida albicans, and those of the archiascomycetes Schizosaccharomyces pombe and Pneumocystis carinii. The euascomycetes genomes freely available presently are from three eurotiomycetes, namely Aspergillus fumigatus, Aspergillus nidulans (teleomorph Emericilla nidulans) and Histoplasma capsulatum (teleo- morph Ajellomyces capsulatus), and three sordariomycetes, namely Magnaporthe grisea, Neurospora crassa and Fusarium graminearum (teleomorph Gibberrella zeae) (Table 2).

Euascomycetes genomes were analysed by similarity search of the deduced protein sequences indicated in bold in Table 2 using TBLASTN [15] through the Center for Genome Research at the Whitehead Institute for Aspergillus nidulans, Fusarium graminearum, Magnaporthe grisea and Neuros- pora crassa (http://www-genome.wi.mit.edu/annotation/ fungi), the National Center for Biotechnology Information for Aspergillus fumigatus (http://www.ncbi.nlm.nih.gov) and the Genome Sequencing Center at Washington University Medical School for Histoplasma capsulatum (http://www.genome.wustl.edu/projects/hcapsulatum/). Pre- diction of transmembrane segments in proteins was based on PRED-TMR2 statistical analysis (http://biophysics.biol.uoa. gr/), TOPPRED2 statistical analysis (http://bioweb.pasteur.fr), TM Kite–Doolittle hydropathy profiles software) and hydrophobic (HCA) (http://smi.snv.jussieu.fr/hca/hca-form.html). Alignments of N-terminal extremity were performed by CLUSTALW [16] and HCA on the class I protein sequences from B. cinerea, Arthroderma benhamiae, Coccidioides immitis, Histoplasma capsulatum, Aspergillus nidulans, Aspergillus oryzae, Asper- gillus fumigatus, Ampelomyces quisqualis, Exophiala derma- titidis, Fusarium graminearum, Mycosphaerella graminicola, Magnaporthe grisea and Neurospora crassa. Global align- ment of protein sequences without gap end penalty was performed using the Needleman–Wunsch–Sellers algorithm.

Phylogeny

TBLASTN analysis performed on the genomic sequences identified a group of putative chs in addition to the six expected chs classes. This group, which will be discussed below, is constituted of AfchsD-like genes that are probably orthologous to the already described gene AfchsD of A. fumigatus. The fully sequenced genomes contained at least one representative for each class including the AfchsD- like group, except for class VI, for which no representative could be found in two out of six species. Besides this, duplication of the class III genes was found for three out of six species.

In parallel, when looking at all the euascomycetes chs sequences available in GenBank/EMBL/DDJB, we observed a duplication of class II in the eurotiomycetes Penicillium chrysogenum [23]. To our knowledge, this represents the only example of a euacomycetes chs dupli- cation outside the class III, discussed above.

Isolation of Bcchsfragments from division 1 and 2, and DNA walking for classes I, III, IV and VI

Amino acid alignment for phylogeny was performed with a segment-to-segment comparison using DIALIGN2 [17]. We aligned portions of 63 protein sequences beginning at the DX(D/G) motif and ending at the QXXRW motif of polysaccharide synthases, which correspond to 134 sites after gap removal. Tree construction was then carried out using different algorithms and distances. The parsimony tree and the PAM-Dayhoff distance matrix were carried out with the PHYLIP package [18]. For comparison, maximum likelihood distances of the protein sequences alignment were computed with TREE-PUZZLE version 5.0 [19], taking into account the rate of heterogeneity between sites and using the JTT substitution matrix [20]. Tree constructions with the PAM-Dayhoff and JTT distance matrix were then carried out using the BIONJ algorithm [21]. Bootstrap values were computed using 1000 replicates, created with the program SEQBOOT from the PHYLIP package [18]. Lastly, the phylogeny by maximum likeli- hood using the Dayhoff matrix was constructed with PHYML [22].

Identifying the chs multigenic family of six euascomycetes genomes led us to expect five to seven chs genes from divisions 1 and 2 in B. cinerea. We used five different combinations of degenerate primers, enabling us to amplify by PCR, portions of seven chitin synthase genes from genomic DNA of B. cinerea. These products were cloned, and their deduced amino acid sequences aligned with the corresponding chs fragments of A. fumigatus for compar- ison (Table 3). For the sake of consistency, we named the B. cinerea chs genes with the number of their deduced CHS class, adding a or b as an index for differenciating the duplicated genes of class III.

Results and discussion

Dissecting the chsmultigenic families of six euascomycetes genomes

With the aim of assessing the expected number of chitin synthase genes in the ascomycetes filamentous fungus B. cinerea, we explored the chs multigenic families from sequenced fungal genomes (Materials and methods). We noticed that the composition of the chs family was very

We succeeded in amplifying four gene fragments from division 1 in B. cinerea. BcchsI, IIIa and IIIb fragments, of 603, 674 and 615 bp in length, respectively (Fig. 1), were amplified with the degenerate primers 1, 2 and 3 (Table 1) described previously by Bowen et al. [5]. Our results provide evidence for a duplication of class III in a leotiomycetes species, as was already observed in some eurotiomycetes and sordariomycetes fungi (Table 2). A BcchsII fragment of 491 bp in length (Fig. 1) was amplified with degenerate

2156 M. Choquer et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Table 2. Chs multigenic families in six euascomycetes genomes and comparison with the B. cinerea family. In bold, previously confirmed chs genes; underlined, contig accession number for which a new chs sequence was found; in parenthesis, contig accession number for which an already known chs sequence was found. Class VII refers to the new class confirmed in this study.

Division 1 Division 2

Classification I II III IV V VI VII Taxonomy

Aspergillus fumigatus chsB (4882) chsE (4829) chsD (4836) Eurotiomycetesa Eurotialesb Trichocomaceaec 4843

Aspergillus nidulansd csmA (1.107) chsF (4922) (5172) chsE (1.25) 1.107 1.16 Histoplasma capsulatume 0.2 Magnaporthe grisea 25.13 absent Eurotiomycetesa Eurotialesb Trichocomaceaec Eurotiomycetesa Onygenalesb Onygenaceaec Sordariomycetesa Magnaporthaceaec 25.11 csm1 (2.1828) 2.1131 Neurospora crassa chsC/chsG (4951)/(4925) (4243)/(5237) chsB (1.43)/1.75 chs2 (17.1) chs2 (2.344) chs1 (3.199) 18.8 chs4 (2.1900) chs4 (3.585) 3.298 3.225 Fusarium graminearumf 3.225 absent Sordariomycetesa Sordarialesb Sordariaceaec Sordariomycetesa Hypocrealesb 1.65 Botrytis cinereag chsA (4928) (5210) chsC (1.78) chs1 (17.1) chs3 (2.1822) chs3 (3.221) chs1 (1.425) chsI chsA (1.117) chs3 (53.2) chs1 (2.784) chs2 (3.295) chs2 (1.122) 1.157/1.419 chsII chsIIIa/chsIIIb chsIV 1.105 chsV chsVI 1.104 chsVII Leotiomycetesa Helotialesb Sclerotiniaceaec

The taxonomy according to NCBI is a Class b Order c Family. d Teleomorph Emericilla nidulans; e teleomorph Ajellomyces capsulatus; f teleomorph Gibberella zeae; g teleomorph Botryotinia fuckeliana, sequences derived from this study.

Table 3. Percentage of identity between peptide fragments of B. cinerea and A. fumigatus chitin synthases. Each BcCHS peptide sequence was predicted from the nucleotide PCR fragments obtained with degenerate primers (Fig. 1). These peptide fragments correspond to a part of the highly conserved catalytic domain among chitin synthases (CHS), except for BcCHSV which corresponds to a part of the myosin head-like domain. Percentage amino acid identity was determined using global alignment (Needleman–Wunsch–Sellers algorithm) without gap end penalty; amino acid fragment length used for B. cinerea fragments are shown in parenthesis; bold, % identity within genes of a probable same class; italics, % identity within genes of a same probable division. Class VII is a new class confirmed in this study. –, Identity not determined because AfCHS-A, -B, -G, -C, -F and -D are lacking a myosin head-like domain.

Division 1 Division 2

I III III IV V VI VIIa II Classification

AfCHSA (748) AfCHSB (799) AfCHSG (911) AfCHSC (889) AfCHSF (584) AfCHSE (1498) Af4843 (780) AfCHSD (789) Fragment

a Percentage amino acid identities are from fragments of myosin head-like domains and are not comparable to those of the catalytic domain fragments.

primers 4 and 5 (Table 1), designed on the motifs FKGAEKGI and ADVFTANMY, respectively.

motifs EIAEICQ and IANQFDSKCV of the myosin head- like domain, respectively. A BcchsVI fragment of 396 bp in length (Fig. 1) was amplified with degenerate primers 8 and 9 (Table 1), that we designed on the motifs KVYSGLYE and MIQVYEYYI, respectively. Our results give evidence for a class VI in a leotiomycetes species, already observed in some eurotiomycetes and sordariomycetes fungi (Table 2). Motifs of classes II, V and VI were found to be strictly conserved and class-specific by alignment of orthologous

We succeeded in amplifying in B. cinerea three gene fragments from division 2. A BcchsIV fragment of 733 bp in length (Fig. 1) was amplified with degenerate primers 6 and 7 (Table 1) that we previously used to amplify chs homologues in M. grisea [14]. A BcchsV fragment of 1271 bp in length (Fig. 1) was amplified with degenerate primers 10 and 11 (Table 1), that were designed on the

BcCHSI (201) BcCHSII (163) BcCHSIIIa (207) BcCHSIIIb (205) BcCHSIV (227) BcCHSV (423) BcCHSVI (132) BcCHSVII (157) 82 68 56 54 14 – 15 23 63 85 51 49 29 – 14 18 55 60 87 74 17 – 23 24 54 64 84 76 17 – 23 17 16 29 16 15 87 – 50 21 17 17 18 25 45 69a 59 26 18 16 17 17 54 17a 86 24 14 18 19 24 28 – 12 85

The chitin synthases of Botrytis cinerea (Eur. J. Biochem. 271) 2157

(cid:1) FEBS 2004

showed 75 and 65% amino acid identity with AfCHSG and AfCHSC, respectively. BcCHSIIIb showed 59 and 57% amino acid identity with AfCHSG and AfCHSC, respectively. BcCHSIV showed 65% amino acid identity with AfCHSF.

sequences from around 10 euascomycetes species in each case (data not shown). Our primers could represent an alternative for amplification of these classes in euascomyc- etes species, more specifically for the class II sequences displaying the variant TMYNEN motif that may be resistant to amplification with the previous primers designed on the TMYNED consensus [5].

cDNA amplification of BcchsIIIa revealed four intron sequences which were not strictly conserved when compared to the predicted intron sequences of BcchsIIIb and the class III orthologues in other fungi (data not shown). In BcchsIV, only one intron sequence was predicted and checked by RT-PCR amplification (data not shown).

Fragments amplified in each class were used to probe B. cinerea genomic DNA, restricted by HindIII, EcoRI or BamHI. In all cases, Southern analysis revealed a single band, thus demonstrating that chitin synthases BcchsI, II, IIIa, IIIb, IV, V and VI are each encoded by a single copy gene (Fig. 2).

Bcchsinvitro expression and 5¢ noncoding region analysis for BcchsI,IIIa,IIIb and IVgenes

BcchsI, IIIa, IIIb, IV and VI genomic sequences were determined using the Universal GenomeWalkerTM Kit based on a step-down PCR technique which led to the unknown genomic DNA sequences adjacent to the frag- ments previously amplified with degenerate primers. We obtained the 5¢ noncoding region of the BcchsI, IIIa, IIIb and IV genes, the complete coding sequence of Bcchs IIIa, IIIb and IV and a large portion of the BcchsI and VI coding sequences.

We recently reported that expression of BcchsI and BcchsIV was constitutive and stable during exponential-like vegetat- ive growth. In contrast, the BcchsIIIa transcript steadily increased whereas the BcchsIIIb transcript could not be detected [13]. In the present work, we showed that BcchsII, V, VI and VII were all expressed during vegetative growth (data not shown).

The percentage of amino acid identity was determined using global alignment (Needleman–Wunsch–Sellers algo- rithm) without gap end penalty for BcCHSIIIa (911 amino acids), BcCHSIIIb (904 amino acids) and BcCHSIV (1218 amino acids) complete sequences, with their orthologues in A. fumigatus, AfCHSC (889 amino acids), AfCHSG (911 amino acids) and AfCHSF (1163 amino acids). BcCHSIIIa

We tried to correlate these results with the presence, in the available 5¢ noncoding region, of sequences that could modulate the expression. Although transcriptional regula- tion of chs genes has been poorly documented in filamen- tous fungi, recent results indicated that it could be implied during yeast-mycelium transition for dimorphic fungi [24], fruit body development [25] or stress [26].

Fig. 1. Analysis of protein sequences predicted for the euascomycetes CHS family isoenzymes and comparison with Botrytis cinerea sequences. Amplified fragments, which are not strictly overlapping between the different BcCHS, are indicated by double headed arrows. BcCHS sequences, determined through gene walking in the present study, are coloured in grey. In parenthesis, the amino acid (aa) length for complete BcCHS sequences is indicated. The undetermined sequences of BcCHS are not coloured. Remarkable features of the sequences were predicted on the CHS sequences deduced from fungal genomes and BcCHS sequences when possible. Motifs a–h of the catalytic domains are indicated in Table 4. Motifs a–g were considered for HCA analysis. Motifs c–g were used for phylogeny.

2158 M. Choquer et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

but whose significance is unclear. Up until now, we could not verify transcriptional regulation of BcchsI, IIIa, IIIb and IV by exposure of the fungus to light, temperature and various metals (data not shown).

Identification of residues probably critical for the b-glycosyltransferase catalytic mechanism

We could not identify with confidence, in our sequences, motifs corresponding to CAAT boxes and TATA boxes. We observed however, the four-fold repetition of the CACATT motif immediately followed by the five-fold repetition of the TAA motif, situated 761 bp upstream of the putative start codon of BcchsIIIa. Whether those repetitions play a role in the up-regulation of BcchsIIIa has to be confirmed.

The AbaA response element (ARE) sequence, corres- ponding to the CATTC motif recognized by the transcrip- tional factor AbaA, is another motif potentially involved in Bcchs transcriptional regulation. Indeed, Park et al. [27] recently described the expression activation of the AnchsC gene (class I) of A. nidulans during conidiophore develop- ment by AbaA. We found at least four putative ARE in sense or antisense orientation upstream of the BcchsI, IIIa, IIIb and IV coding sequences.

We also noticed 800 bp upstream of the putative start codon of BcchsI, the presence of a region rich in GATA motifs of 330 bp in length (in both sense and antisense orientation). These GATA were arranged in 19 putative nitrogen response elements (NIT2), corresponding to two GATA separated by less than 40 bp [28]. Although GATA motifs are known to be implied in many transcriptional regulations, as in response to light or nitrogen deprivation [29], we could not find a BcchsI expression sequence tag in a cDNA library which was obtained from B. cinerea grown under nitrogen deprivation (http://www.genoscope.fr). The significance of repeated GATA sequences in BcchsI might then correspond to another type of regulation.

Finally, when looking at the 5¢ noncoding region of the BcchsIIIb gene, we found a region rich in AT tracks (80%) of 336 bp in length. This region is situated 765 bp upstream of the putative start codon and may account for the expression inhibition observed for this gene. Indeed, AT-rich regions in the 5¢ noncoding region were demon- stated to promote DNA methylation and affect gene expression in N. crassa [30].

All b-glycosyltransferases, even if presenting low sequence identity, share a catalytic domain of around 200–300 amino acids in length defined by a conserved folding [31]. This domain was named SGC for SpsA GnT I Core, from the two nonprocessive b-glycosyltransferases SpsA and GnT I, whose crystal structures were recently resolved [32]. The domain contains one tyrosine and two aspartates, probably implied in substrate binding (motifs a, b and c in Fig. 3), as well as a third aspartate constituting the catalytic base (motif f in Fig. 3). This last residue has been proved to be essential for ScCHS2 activity of S. cerevisiae [33]. In the aim of localizing the corresponding residues in BcCHSI, IIIa, IIIb, IV and VI proteins, and because of the very low identity in the first half sequence of the SGC domains, we used hydrophobic cluster analysis (HCA) [34] to perform alignment with SpsA and GnT I. This method enabled us to identify in our sequences the four residues critical for substrate binding or glycosyltransfer (Fig. 3 and Table 4 [35]). Motif c most probably corresponds to the so-called (cid:2)DXD signature(cid:3) of the GT-A fold superfamily [36]. The aspartate residue at the third position of this signature has been shown to establish an interaction with the a- and b-phosphates of the UDP-sugar donor through coordination to a metal ion. However, the motif is not strictly conserved among the superfamily, appearing for instance as DD in SpsA and GnT I (Fig. 3). Concerning the CHS sequences of B. cinerea, this motif differed slightly, depending on the division, appearing as DXG in division 1 and DXD in division 2 (Table 4). Whether this could be related to a difference in metal requirement between the isoenzymes of the two divisions has to be studied.

Altogether, the 5¢ noncoding region of the chs genes exhibit typical features that differ from one gene to another

Fig. 2. Southern blot hybridization of B. cine- rea genomic DNA. Lanes I, II, IIIa, IIIb, IV, V, VI and VII, DNA probed with 32P-labeled Bcchs fragments from class I–VII. Sub-lanes H, E and B, DNA digested with HindIII, EcoRI and BamHI, respectively. Faint signals are boxed to facilitate the reading.

The chitin synthases of Botrytis cinerea (Eur. J. Biochem. 271) 2159

(cid:1) FEBS 2004

Fig. 3. HCA plots of SGC domain in chitin synthases from B. cinerea (BcCHSI, IIIa, IIIb, IV and VI) and S. cerevisiae (ScCHSII) compared with the nonprocessive glycosyl transferases SpsA and GnTI. In these plots, the protein sequences are written on a duplicated a-helical net and the contour of clusters of hydrophobic residues are automatically drawn (boxed). The standard one-letter code for amino acids is used except for proline, glycine, threonine and serine which are represented by a star, a diamond, a square and a square with a dot, respectively. The HCA multisequence alignments involve the visual analysis of hydrophobic cluster shapes, lengths and distribution along with an analysis of the strictly conserved residues (bold circles). The vertical lines indicate the proposed correspondence between the sequences, and which correspond in CHS to motifs a–h described in Table 4. The divergence among glycosyl transferases in motif c, appearing as DD or DXG or DXD, is illustrated at the bottom.

2160 M. Choquer et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Table 4. Sequence motifs specific to fungal CHS found in division 1, 2 and AfCHSD-like isoenzymes. Motif b not shown as corresponds to only one conserved aspartate implied in substrate binding. Motif h corresponds to a part of the CON2 region [37]. In bold, the residues presumed to be critical for CHS activity [31–33]. Alignments were performed by HCA for motifs a–g (Fig. 4) and by CLUSTALW for motif h (Fig. 5). AfCHSD-like refers to Class VII described at the end of this paper.

Motif code Motif specific to fungal CHS Division 1 consensus Division 2 consensus AfCHSD-like consensus Presumed function

a Motifs described in [35].

a c d e f g h P(C/V)Y(K/R)E DSDC QD(M/L)EY LPG LGEDRWL QRRRW TWG Substrate binding Substrate binding No hypothesis No hypothesis Catalytic base Processivity Processivity (T/P)XYXE DX(G/D)(T/C) QXXEYa LP(G/A)a L(A/G)EDRXLa Q(R/G)RRWa (S/T)WG T(M/Y)YNEa DXGT QNFEYa LP(G/A)a LAEDRILa QRRRWa (S/T)WG (T/P)(A/C)Y(S/T)E DADTa QV(Y/F)EYa LPGa LGEDR(YFE)La Q(R/G)RRWa (S/T)WGa

beginning of the myosin head-like domains and upstream of the cytochrome b5-like domains.

In a second step, we tried to localize in our sequences the motif QXXRW and the conserved region 2 (CON2) among CHS proteins (motifs g and h in Table 4 [33,37]). These sequences are highly conserved downstream of the SGC domain in all polysaccharide synthases but absent from nonprocessive glycosyltransferases (Fig. 3). Given this specificity, they have been postulated to be related to the processivity of the polysaccharide synthases. Their involvement in catalysis has been validated by site-directed mutagenesis on the class II isoenzyme from S. cerevisiae [33,37]. The motif QXXRW and the CON2 region were located in the BcCHSI, IIIa, IIIb, IV and VI sequences using CLUSTALW (motifs g and h in Figs 3 and 4, respectively). The alignment of the CON2 region was checked manually and we particularly considered that the tryptophan residue was the sole strictly conserved posi- tion in the CON2 region in all processive b-glycosyl- transferases.

In contrast, prediction of PTR and PMR in the C-terminus of CHSs from division 1 and 2 was facilitated by the recent description of the topological organization of the hyaluronate synthase from Streptococcus pyogenes SpHASA [39]. In this case, two membrane-associated regions (MR) were characterized inside the catalytic domain. As revealed by Kyte–Doolittle hydropathy profiles, the first one could be shared by all chitin synthases, where it would be positioned between motifs d and f (Fig. 1). As proposed for the hyaluronate synthase, it could participate to the formation of a pore implied in the translocation of the growing chitin chain through the membrane. The second MR of SpHASA, which was initially presumed to be a transmembrane region, results in the localization of the conserved tryptophan residue (Fig. 4, motif h) on the same face as the SGC domain. By analogy, we presumed for CHS that the PTR at the C-terminal extremity situated upstream of motif h was in fact a MR.

Prediction of topological organization of CHS proteins

As we observed that the localization of PTR and PMR was always conserved between the orthologous CHS proteins, we proposed the existence of three different topological organizations that we named A, B and C (Fig. 1). Topology A was shared by all division 1 sequences and presented PTRs that were all grouped in the C-terminus region downstream of the SGC domain (Fig. 1). In contrast, topologies B and C presented PTRs distributed on both sides of the SGC domain. Considering that the glycosyl transfer probably takes place into the cytoplasm

CHS paralogues from B. cinerea and their CHS ortho- logues from other euascomycetes species were subjected to prediction of putative transmembrane regions (PTR) and membrane-associated regions (PMR) by statistical analysis [38], Kyte–Doolittle hydropathy profiles (http://bioweb. pasteur.fr) and HCA. Each method suggested a different number of PTR or PMR. Prediction was particularly difficult in the N-terminus of CHS from division 2, and we still have some doubt about the presence of PMRs at the

Fig. 4. Amino acid alignment of CON2 regions from CHS isoenzymes, NodC and hyaluronate polysaccharide synthases (HAS) using CLUSTALW and checked manually. A tryptophan was found to be the only invariant site among the sequences. Numbering corresponds to the amino acid position among the ScCHS2 protein sequence. Asterisks correspond to the amino acids essential for ScCHS2 activity [37].

The chitin synthases of Botrytis cinerea (Eur. J. Biochem. 271) 2161

(cid:1) FEBS 2004

the motif YQLXD which was replaced by YIHPD in the chaetothyriomycetes E. dermatitidis and the motif QPXYSV which was absent from four sequences of sordariomycetes species (data not shown).

where the substrate UDPGlcNAc is located, we oriented each PTRs topology with the SGC domain facing the cytoplasm. Myosin head-like domains in class V and VI are thus postulated to also be localized into cytoplasm. Myosin head-like domains linked to CHS could play a role in the vesicular transport of these proteins. Indeed, the unconventional myosin MYO2 was proved to act in the proper localization of CHS3 (class IV) of S. cerevisiae [40]. Cytochrome b5-like domains, if not processed during translocation, should be periplasmic. Cytochrome b5-like domains of chitin synthases have probably lost their ability to bind a heme and have been postulated to serve as evolutionary templates for the development of a lipid binding site [41].

Identification of class I specific and conserved peptide motifs

Many hypotheses can be made concerning the role of these motifs. We retained for discussion a possible involve- ment in sorting or function. The motif MXYNRL is reminiscent of a tyrosine-based signal (TBS) exhibiting the tyrosine-X-X-hydrophobe consensus [42]. TBSs bind adap- tor protein complexes that are implied in the targeting of transmembrane proteins to different compartments of the endocytic and late secretory pathways. The YNRL consen- sus could be implied in a routing pathway specific to class I isoenzymes. It is interesting to remember that adaptor protein complex 1 has recently been proposed to be involved in ScCHS3 (class IV) recycling from the early endosome to the trans-Golgi network [43]. However, we propose another model that would impart a role to the whole combination of tyrosine-centered motifs. The N-terminal region of the class I isoenzymes is relatively rich in proline, which could let us suppose that it is not highly folded. However, the conserved tyrosine-centered motifs are composed of strong or moderately hydrophobic amino acids (M, Y, L or V) which are known to be favoured within the internal faces of regular secondary structures (a-helices and b-strands) and disfavoured within the main irregular secondary structures (loops). Careful analysis by HCA [34], which takes into account local proximities, allowed us to propose that the conserved tyrosine-centered motifs formed clusters of hydrophobic amino acids which are representative of b-strands secondary structures (Fig. 5). We propose that this structural arrangement could have a direct or indirect role in the overall function of these enzymes. Protein–protein interactions are known to often be mediated by contacts between b-strands allowing stabil- ization by sheet extension. Whether this kind of interaction is important for the function of class I isoenzymes is an interesting possibility.

Identification of an Afchsd-like gene in B.cinerea and highlighting of a new CHS class independent of division 1 and 2

Sequences of class I isoenzymes deposited in databanks diverge greatly in their N-terminal regions. When looking more precisely, we noticed that the longest N-terminal extensions corresponded to nucleotide sequences including one or more introns, situated very near to the 5¢ extremity of the coding sequence. This observation prompted us to reanalyse the available DNA sequences in all three reading frames. Indeed, we could find evidence at the 5¢ extremity of each class I-encoding sequence for the probable pres- ence of an intron of variable length, ranging from 50 to in the case of BcchsI, by 500 bp. This was verified, amplifying a cDNA portion encompassing the candidate start methionine codon. We could thus localize two nearby introns of 373 and 69 bp and delineate the N-terminal residue of the BcCHSI isoenzyme. When the corrected reading frames were taken into account, all the class I protein sequences exhibited comparable N-terminal exten- sion sizes. Furthermore, conservation of motifs with at least one conserved tyrosine appeared to be evident in this region that hitherto was known as very variable. The first motif corresponded to the consensus MXYNRL. As in B. cinerea, which displayed the MSYNRL sequence, the methionine residue of the motif most probably corres- ponds in each case to the start of the protein. This motif was followed by three others, namely YQLXD, QPXYSV and PXXHHDXYYXXP. These consensus sequences were each separated by around 30 amino acids. They were observed in 13 euascomycetes species, except for

As mentioned above, we observed in all analysed euasco- mycetes genomes (Table 2) a single copy gene probably orthologous (around 80% amino acid identity) to the

Fig. 5. HCA plot of the N-terminus of the class I isoenzyme of B. cinerea. HCA was performed as described for Fig. 3. The vertical lines indicate the conserved motifs for which b-strand structure is proposed.

2162 M. Choquer et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

already described AfchsD gene of A. fumigatus. AfCHSD was reported to be divergent from usual CHS isoenzymes with around 20–30% amino acid identity and its function still remains uncertain. Nevertheless, an AfchsD mutant showed a decrease of 20% in chitin content [44] and alignment of the AfCHSD-like isoenzymes revealed con- servation of sequence consensus motifs known to be correlated with fungal CHS activities (Table 4). For these reasons, we considered that the AfchsD-like genes corres- pond to a new type of chitin synthase.

With the aim of amplifying the corresponding gene in B. cinerea, we aligned the AfCHSD-like sequences that we identified in the six euascomycetes genomes sequenced to date, adding another AfchsD-like gene recently isolated in Coccidioides immitis, CiChs6. We designed degenerate primers 12 and 13 (Table 1) to the QNF(MIV)YDME and ITNEVCM motifs, respectively, that were strictly conserved and specific to these genes. A BcchsVII fragment of 471 bp in length was successfully amplified in B. cinerea (Fig. 1), and presented 85% amino acid identity with the corresponding fragment of AfchsD (Table 3). Our primers are thus probably effective in amplifying AfchsD-like genes in other euascomycetes fungi. Southern analysis, processed as described above, showed that the BcchsVII chitin synthase was encoded by a single copy gene in B. cinerea (Fig. 2).

were congruent and validated our phylogenetic study (data not shown). Sequences from division 1 were clearly monophyletic, however, this was not the case for sequences from division 2 as revealed by a low boostrap value (77%). We observed that AfCHSD-like isoenzymes formed a new group of sequences that was clearly branched apart from the divisions 1 and 2 (boostrap value 100%) (Fig. 6). We concluded that AfCHSD-like isoenzymes define a new class of CHS orthologues that we call class VII. We suggest reclassification of AfCHSD and CiCHS6, isoenzymes described above, which were first attributed to class VI.

Fig. 6. Phylogenetic tree connecting amino acid sequences. Multiple alignment was performed between motifs c to g. The tree was con- structed with the distance method using the BIONJ program and the PAM-Dayhoff distance. The bootstrap values calculated using 1000 replicates are indicated above each branch.

Conclusion

Phylogenetic relationships of CHS from the euascomyc- etes species were analysed with the aim of positioning the AfCHSD-like isoenzymes among the chitin synthase clas- sification. Because chitin synthases display multiple pre- sumed deletions, insertions and highly variable regions, the alignment of their entire sequence is impossible by classical multiple alignment software. For this reason and as previously described in numerous studies, we based our phylogenetic study on a fixed amino acid alignment corresponding to a part of the CHS catalytic domain that presents few gaps or insertions. This region of around 160 amino acids in length, begins at the DX(G/D)(T/C) motif and ends at the QXXRW motif (motifs c and g in Table 4 and Fig. 1). We analysed the CHS sequences taken from five euascomycetes genomes (A. fumigatus, A. nidulans, M. grisea, N. crassa and F. graminearum) and included the BcchsI, IIIa, IIIb, IV and VI chitin synthase genes derived from this study. Sequences of BcCHSII, BcCHSV, and BcCHSVII were too short or did not overlap this conserved region for the phylogeny. We used the sequences of the bacterial b-glycosyltransferase NODC and hyaluro- nate synthase (HAS) from vertebrates and bacteria as outgroups to allow the rooting of the euascomycetes CHS tree. All sequences were aligned with a segment-to-segment comparison using DIALIGN2 (data not shown). We observed, as others did, that sites which correspond to amino acids that are probably implied in catalysis and present in motifs c–g (described above) were almost invariant while the other sites were greatly evolved. We thus obtained a strong rate of heterogeneity between sites as revealed by a low parameter alpha of 0.88 ± 0.11, estimated from the data set using a gamma distribution.

the first

This paper describes the first identification of the chitin synthase multigenic family in a leotiomycetes fungus. Duplication of class III and presence of a class VI were observed in B. cinerea, although we noticed that it was not a general rule in other euascomycetes genomes. We provide evidence for time that conserved sequence N-terminal motifs exist on the BcCHSI isoenzyme and its orthologues, and propose putative structures in the N-terminal extremity of CHS from division 1. Finally, we

The phylogenetic tree was constructed with the BIONJ algorithm, using the PAM-Dayhoff distance and with bootstrap statistical analysis (Fig. 6). Trees obtained with other algorithms and distances (Material and methods)

The chitin synthases of Botrytis cinerea (Eur. J. Biochem. 271) 2163

(cid:1) FEBS 2004

update the CHS classification of the euascomycetes fungi by adding a new class (class VII) of CHS formed by the AfCHSD-like isoenzymes, which clearly branch apart from divisions 1 and 2.

16. Higgins, D.G. & Sharp, P.M. (1988) CLUSTAL: a package for performing multiple sequence alignment on a microcomputer. Gene 73, 237–244. 17. Morgenstern, B. (1999) DIALIGN 2: improvement of

the segment-to-segment approach to multiple sequence alignment. Bioinformatics 15, 211–218.

Acknowledgements

18. Felsenstein, J. (1993) PHYLIP (Phylogeny Inference Package) version 3.5c. Distributed by the Author. Department of Genetics, University of Washington, Seattle.

19. Schmidt, H.A., Strimmer, K., Vingron, M. & von Haeseler, A. (2002) TREE-PUZZLE: maximum likelihood phylogenetic ana- lysis using quartets and parallel computing. Bioinformatics 18, 502–504.

The authors wish to express their thanks to Gilles Labesse and Anne Poupon for their kind help in interpreting the HCA profiles, Caroline Levis for providing them with the EST clones of B. cinerea and Isabelle Jupin for giving them access to her image documentation system for RT-PCR analysis. I. G. is a member of the Universite´ Pierre et Marie Curie (Paris) and this work was supported by grant no. 4672 from the Association pour la Recherche sur le Cancer. 20. Jones, D.T., Taylor, W.R. & Thornton, J.M. (1992) The rapid generation of mutation data matrices from protein sequences. Comput. Appl. Biosci. 8, 275–282.

References

1. Bernard, M. & Latge, J.P. (2001) Aspergillus fumigatus cell wall: 21. Gascuel, O. (1997) BIONJ: an improved version of the NJ algo- rithm based on a simple model of sequence data. Mol. Biol. Evol. 14, 685–695. composition and biosynthesis. Med. Mycol. 39, 9–17. 2. Roncero, C. (2002) The genetic complexity of chitin synthesis in fungi. Curr. Genet. 41, 367–378. 22. Guindon, S. & Gascuel, O. (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst. Biol. 52, 696–704. 3. Munro, C.A. & Gow, N.A. (2001) Chitin synthesis in human pathogenic fungi. Med. Mycol. 39, 41–53.

23. Namgung, J., Park, B.C., Lee, D.H., Bae, K.S. & Park, H.M. (1996) Cloning and characterization of chitin synthase gene frag- ments from Penicillium chrysogenum. FEMS Microbiol. Lett. 145, 71–76.

4. Soulie, M.C., Piffeteau, A., Choquer, M., Boccara, M. & Vidal- Cros, A. (2003) Disruption of Botrytis cinerea class I chitin syn- thase gene Bcchs1 results in cell wall weakening and reduced virulence. Fungal Genet. Biol. 40, 38–46.

24. Nino-Vega, G.A., Munro, C.A., San-Blas, G., Gooday, G.W. & Gow, N.A. (2000) Differential expression of chitin synthase genes in Para- during temperature-induced dimorphic transitions coccidioides brasiliensis. Med. Mycol. 38, 31–39. 5. Bowen, A.R., Chen-Wu, J.L., Momany, M., Young, R., Szanis- zlo, P.J. & Robbins, P.W. (1992) Classification of fungal chitin synthases. Proc. Natl Acad. Sci. USA 89, 519–523.

6. Din, A.B., Specht, C.A., Robbins, P.W. & Yarden, O. (1996) chs-4, a class IV chitin synthase gene from Neurospora crassa. Mol. Gen. Genet. 250, 214–222. 25. Balestrini, R., Mainieri, D., Soragni, E., Garnero, L., Rollino, S., Viotti, A., Ottonello, S. & Bonfante, P. (2000) Differential expression of chitin synthase III and IV mRNAs in ascomata of Tuber Borchii Vittad. Fungal Genet. Biol. 31, 219–232.

7. Fujiwara, M., Horiuchi, H., Ohta, A. & Takagi, M. (1997) A novel fungal gene encoding chitin synthase with a myosin motor-like domain. Biochem. Biophys. Res. Commun. 236, 75–78. 26. Wang, Q., Liu, H. & Szaniszlo, P.J. (2002) Compensatory expression of five chitin synthase genes, a response to stress sti- muli, in Wangiella (Exophiala) dermatitidis, a melanized fungal pathogen of humans. Microbiology 148, 2811–2817.

8. Chigira, Y., Abe, K., Gomi, K. & Nakajima, T. (2002) chsZ, a gene for a novel class of chitin synthase from Aspergillus oryzae. Curr. Genet. 41, 261–267. 27. Park, B.C., Park, Y.H. & Park, H.M. (2003) Activation of chsC transcription by AbaA during asexual development of Aspergillus nidulans. FEMS Microbiol. Lett. 220, 241–246.

9. Cabib, E. & Schmidt, M. (2003) Chitin synthase III activity, but not the chitin ring, is required for remedial septa formation in budding yeast. FEMS Microbiol. Lett. 224, 299–305. importance of

28. Chiang, T.Y. & Marzluf, G.A. (1994) DNA recognition by the NIT2 nitrogen regulatory protein: the number, spacing, and orientation of GATA core elements and their flanking sequences upon NIT2 binding. Biochemistry 33, 576–582. 29. Scazzocchio, C. (2000) The fungal GATA factors. Curr. Opin. 10. Munro, C.A., Whitton, R.K., Hughes, H.B., Rella, M., Selvaggini, S. & Gow, N.A. (2003) CHS8 – A fourth chitin syn- thase gene of Candida albicans contributes to in vitro chitin synthase activity, but is dispensable for growth. Fungal Genet. Biol. 40, 146–158. Microbiol. 3, 126–131.

30. Tamaru, H. & Selker, E.U. (2003) Synthesis of signals for de novo DNA methylation in Neurospora crassa. Mol. Cell. Biol. 23, 2379–2394. 11. Leroux, P., Fritz, R., Debieu, D., Albertini, C., Lanen, C., Bach, J., Gredt, M. & Chapeland, F. (2002) Mechanisms of resistance to fungicides in field strains of Botrytis cinerea. Pest Manag. 58, 876–888.

12. Zhang, Z.G., Hall, A., Perfect, E. & Gurr, S.J. (2000) Differential expression of two chitin synthase genes of Blumeria graminis. Mol. Plant Pathol. 1, 125–138. 31. Saxena, I.M., Brown, R.M. Jr, Fevre, M., Geremia, R.A. & Henrissat, B. (1995) Multidomain architecture of beta-glycosyl transferases: implications for mechanism of action. J. Bacteriol. 177, 1419–1424. 32. Unligil, U.M. & Rini, J.M. (2000) Glycosyltransferase structure and mechanism. Curr. Opin. Struct. Biol. 10, 510–517.

13. Choquer, M., Boccara, M. & Vidal-Cros, A. (2003) A semi- quantitative RT-PCR method to readily compare expression levels within Botrytis cinerea multigenic families in vitro and in planta. Curr. Genet. 43, 303–309.

14. Vidal-Cros, A. & Boccara, M. (1998) Identification of four chitin synthase genes in the rice blast disease agent Magnaporthe grisea. FEMS Microbiol. Lett. 165, 103–109. 33. Nagahashi, S., Sudoh, M., Ono, N., Sawada, R., Yamaguchi, E., Uchida, Y., Mio, T., Takagi, M., Arisawa, M. & Yamada-Okabe, H. (1995) Characterization of chitin synthase 2 of Saccharomyces cerevisiae. Implication of two highly conserved domains as pos- sible catalytic sites. J. Biol. Chem. 270, 13961–13967.

34. Callebaut, I., Labesse, G., Durand, P., Poupon, A., Canard, L., Chomilier, J., Henrissat, B. & Mornon, J.P. (1997) Deciphering protein sequence information through hydrophobic cluster ana- 15. Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Anang, Z., Miller, W. & Lipman, D.J. (1997) Gapped BLAST and PSI- BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402.

2164 M. Choquer et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

lysis (HCA): current status and perspectives. Cell Mol. Life Sci. 53, 621–645. 41. Mifsud, W. & Bateman, A. (2002) Membrane-bound progester- one receptors contain a cytochrome b5-like ligand-binding domain. Genome Biol. 3, research0068.1–0068.5.

42. Bonifacino, J.S. & Traub, L.M. (2003) Signals for sorting of transmembrane proteins to endosomes and lysosomes. Annu. Rev. Biochem. 72, 395–447.

35. Ruiz-Herrera, J., Gonzalez-Prieto, J.M. & Ruiz-Medrano, R. (2002) Evolution and phylogenetic relationships of chitin syn- thases from yeasts and fungi. FEMS Yeast Res. 1, 247–256. 36. Bourne, Y. & Henrissat, B. (2001) Glycoside hydrolases and gly- cosyltransferases: families and functional modules. Curr. Opin. Struct. Biol. 11, 593–600.

43. Valdivia, R.H., Baggott, D., Chuang, J.S. & Schekman, R.W. (2002) The yeast clathrin adaptor protein complex 1 is required for the efficient retention of a subset of late Golgi membrane proteins. Dev. Cell 2, 283–294.

37. Yabe, T., Yamada-Okabe, T., Nakajima, T., Sudoh, M., Arisawa, M. & Yamada-Okabe, H. (1998) Mutational analysis of chitin synthase 2 of Saccharomyces cerevisiae. Identification of addi- tional amino acid residues involved in its catalytic activity. Eur. J. Biochem. 258, 941–947. 44. Mellado, E., Specht, C.A., Robbins, P.W. & Holden, D.W. (1996) Cloning and characterization of chsD, a chitin synthase-like gene of Aspergillus fumigatus. FEMS Microbiol. Lett. 143, 69–76.

Supplementary material

38. Liakopoulos, T.D., Pasquier, C. & Hamodrakas, S.J. (2001) A novel tool for the prediction of transmembrane protein topology based on a statistical analysis of the SwissProt database: the OrienTM algorithm. Protein Eng. 14, 387–390.

The following material is available from http://blackwell publishing.com/products/journals/suppmat/EJB/EJB4135/ EJB4135sm.htm Fig. S1. Chitin synthase sequence alignment used to con- struct the phylogenetic tree shown in Fig. 6. Fig. S2. Phylogenetic trees, except the one shown in Fig. 6.

39. Heldermon, C., DeAngelis, P.L. & Weigel, P.H. (2001) Topolo- gical organization of the hyaluronan synthase from Streptococcus pyogenes. J. Biol. Chem. 276, 2037–2046.

40. Santos, B. & Snyder, M. (1997) Targeting of chitin synthase 3 to polarized growth sites in yeast requires Chs5p and Myo2p. J. Cell Biol. 136, 95–110.