gene family of Asteraceae: Evolution with substrate change ... - NCBI

Proc. Natl. Acad. Sci. USA Vol. 93, pp. 9033-9038, August 1996 Evolution

Duplication and functional divergence in the chalcone synthase gene family of Asteraceae: Evolution with substrate change and catalytic simplification (anthocyanin/flavonoid genetics/gene phylogeny/secondary metabolism/stilbene synthase)

YRJO HELARIUTA*t*, MiKA KOTILAINEN*, PAULA ELOMAA*, NISSE KALKKINEN*, KARE BREMER§, TEEMU H. TEERI*, AND VICTOR A. ALBERTt *Institute of Biotechnology, University of Helsinki, P.O. Box 45, FIN-00014 Helsinki, Finland; *The New York Botanical Garden, Bronx, NY 10458-5126; and §Department of Systematic Botany, Uppsala University, Villavagen 6, S-752 36 Uppsala, Sweden

Communicated by Michael T Clegg, University of California, Riverside, CA, May 20, 1996 (received for review November 29, 1995)

SS genes at the level of deduced amino acid sequence. Its expression pattern at both organ and cellular levels is not correlated with anthocyanin pigmentation, for which CHS provides the first committed biosynthetic step. Furthermore, the catalytic properties of the corresponding enzyme differ from CHS and SS, although the GCHS2 catalytic reaction and its role in vivo are not yet completely understood. In this study we show that the GCHS2-like genes in Asteraceae constitute a gene family, whose corresponding amino acid sequences share some consensus residues. Phylogenetic parsimony analysis of (i) the GCHS2 nucleotide sequence, (ii) further GCHS2-like genes screened from a Gerbera cDNA library, (iii) gene fragments amplified from various Asteraceae using GCHS2 family-specific primers, and (iv) CHS superfamily genes isolated from other angiosperms of subclass Asteridae indicates that GCHS2 probably evolved from CHS via a single gene duplication event that occurred before the diversification of Asteraceae. Structural and functional variation observed among the members of the GCHS2 gene family suggests that subsequent diversification has also taken place. A comparison of the catalytic properties of GCHS2 to parsley CHS shows that both substrate specificity and progressivity of catalytic reaction steps have changed during GCHS2 evolution.

Plant-specific polyketide synthase genes conABSTRACT stitute a gene superfamily, including universal chalcone synthase [CHS; malonyl-CoA:4-coumaroyl-CoA malonyltransferase (cyclizing) (EC 2.3.1.74)] genes, sporadically distributed stilbene synthase (SS) genes, and atypical, as-yetuncharacterized CHS-like genes. We have recently isolated from Gerbera hybrida (Asteraceae) an unusual CHS-like gene, GCHS2, which codes for an enzyme with structural and enzymatic properties as well as ontogenetic distribution distinct from both CHS and SS. Here, we show that the GCHS2like function is encoded in the Gerbera genome by a family of at least three transcriptionally active genes. Conservation within the GCHS2 family was exploited with selective PCR to study the occurrence of GCHS2-like genes in other Asteraceae. Parsimony analysis of the amplified sequences together with CHS-like genes isolated from other taxa of angiosperm subclass Asteridae suggests that GCHS2 has evolved from CHS via a gene duplication event that occurred before the diversification of the Asteraceae. Enzyme activity analysis of proteins produced in vito indicates that the GCHS2 reaction is a non-SS variant of the CHS reaction, with both different substrate specificity (to benzoyl-CoA) and a truncated catalytic profile. Together with the recent results of Durbin et al. [Durbin, M. L., Learn, G. H., Jr., Huttley, G. A. & Clegg, M. T. (1995) Proc. Natl. Acad. Sci. USA 92, 3338-3342], our study confirms a gene duplication-based model that explains how various related functions have arisen from CHS during plant evolution.

MATERIALS AND METHODS Plant Material. G. hybrida is a hybrid of two species (G. jamessonii and G. viridifolia) belonging to the tribe Mutisieae (Asteraceae subfamily Cichorioideae; ref. 16). We chose for analysis Leibnitzia (a closely related genus) and Onoseris (a more distantly related genus) from Mutisieae, Taraxacum (tribe Lactuceae) from Cichorioideae, and Dahlia (tribe Heliantheae) from subfamily Asteroideae (16). Mature plants of G. hybrida var. Regina (obtained from Terra Nigra, De Kwakel, The Netherlands) and seedlings of Leibnitzia anandria and Onoseris sagittatis were grown under standard greenhouse conditions. Leaf material of Dahlia sp. and Taraxacum sp. (collected from gardens in Helsinki) were also used as sources of DNA.

Plant-specific polyketide synthase genes constitute a gene superfamily. Genes encoding chalcone synthase [CHS; malonyl-CoA:4-coumaroyl-CoA malonyltransferase (cyclizing) (EC 2.3.1.74)] and flavonoids, their corresponding reaction products, seem to be universally distributed in plants (1-3). CHS genes have been isolated from a wide taxonomic spectrum from nonflowering seed plants to dicots and monocots (4, 5). Stilbene synthase (SS) genes coding for enzymes with a related activity to CHS have been isolated from species that accumulate stilbene phytoalexin (5-8). In addition to SS genes, other structurally unusual CHS-like genes or gene products have been reported (9-12). Recently, Tropf et al. (13) have provided evidence that SS genes have evolved from CHS genes via independent gene duplication events several times during seed plant evolution. Durbin et al. (14) have demonstrated an analogous mechanism leading to the evolution of structurally unusual genes in the genus Ipomoea. We have recently isolated a structurally unusual CHS-like gene, GCHS2, from Gerbera hybrida (Asteraceae; ref. 15). GCHS2 is '70% identical to typical CHS genes and the related

Isolation of GCHS17 and GCHS26 from a Genomic A Library. Nuclear DNA from Gerbera leaves was prepared by the method of Jofuku and Goldberg (17). A genomic library was constructed with LambdaGEM-11 vector (Promega) and was screened using GCHS1-3 cDNA clones as probes (15, 18). Short fragments (181 bp) from the clones were amplified using primers designed from the conserved region of the CHS genes Abbreviations: CHS, chalcone synthase; SS, stilbene synthase; 2-ME,

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in I accordance with 18 U.S.C. §1734 solely to indicate this fact.

9033

2-mercaptoethanol. Data deposition: The sequences reported in this paper have been deposited in the GenBank data base (accession nos. X91339-X91345). tTo whom reprint requests should be sent at present address: Department of Biology, New York University, New York, NY 10003.

9034

Evolution: Helariutta et al.

(15), and their sequences were used for classification and designation of the subcloning strategy. Both strands of the GCHS17 and GCHS26 clones were determined using the deletion strategy both manually and with an automated sequence determination system (ALF; Pharmacia; ref. 18). Sequence alignments were done using the CLUSTAL program of the PC/gene package (IntelliGenetics). Amplification of GCHS2-Like Sequences from Dahlia, Leibnitzia, Onoseris, and Taraxacum. Partially degenerate (inosine containing) primers including restriction enzyme sites were used to amplify fragments from various asteraceous species.

5'-TGCACTCGAGTGA(A/C/G/T)AA(A/G)ACAGC(A/ C/G/T)ATAAA(A/G)AA-3' and 5'-ACTGGGATCCACC(A/C/G/T)GGGTG(A/C/G/T)ACCATCCA(A/G)AA-3' corresponding to peptides C(D/E)KTAIKK and FWMVHPGG were used for specific amplification of GCHS2-like genes. For amplification, plant DNA was extracted by the method of Dellaporta et al. (19) and purified in isopycnic CsCl gradients (18). PCR was performed using a "touch-down" strategy: 10 times (940 75 s; 500 5 min adding -1° per cycle, slope +220, 10 per 10 s; 720 5 min) followed by 31 times (940 75 s; 530 2 min; 720 5 min). To verify that the amplification products did not contain any chimeric artifacts due to recombination among related gene family members (20), a second, independent amplification was performed. After PCR, fragments were separated from primers by gel electrophoresis, purified from agarose and digested with the corresponding restriction enzyme pair for cloning into plasmid pOK12 (21) for sequence analysis of both strands. Phylogenetic Analysis of CHS-Like Gene Sequences. Nucleotide sequences representing 20 sequences (corresponding to the amino acid sequence from Rll to S389 in GCHS1) were aligned (using CLUSTAL) based on their deduced amino acid sequences. The aligned sequences (available on request) were subjected to cladistic parsimony analysis using the program PAUP (22). Heuristic search options were random addition of sequences with 100 replicates and subsequent tree-bisectionreconnection branch swapping to generate multiple equally parsimonious trees. Branch support was estimated using parsimony jackknifing with 10,000 replicates (23). The trees were oriented by treating CHS superfamily sequences from taxa of the Asterid I and Asterid II clades of Chase et al. (24) as monophyletic, respectively. Expression of GCHS26 in Escherichia coli and Analysis of Corresponding Enzymatic Activity. Plasmids pHTT402, pHTT406 (15), and pHTT409 express GCHS2, parsley CHS, and GCHS26 genes from the vector pKKtac (25) in E. coli. In pHTT409, the intron in the genomic clone of GCHS26 was removed with the help of an oligonucleotide spanning the putative joint site of the exons. In these expression constructs the initiation of translation takes place at the plant genes' ATG. The vector without an insert served as the control. For enzyme production, E. coli DH5a cells (26) harboring the expression constructs were grown to OD600 = 0.8-1.0, induced for 1.5 h with 1 mM isopropyl 3-D-thiogalactoside at 280C, pelleted, and stored at -70°C. The enzymatic reaction was analyzed exactly as described in ref. 15. In certain experiments, 2-mercaptoethanol (2-ME) was added in the reaction to inhibit its progressivity. Production, Purification, and UV Spectrum Analysis of Benzoyl-CoA-Derived Products. The ethyl acetate extracts from a large-scale enzymatic reaction [in 6 x 1 ml of 50 mM Hepes-KOH (pH 7)/1 mM EDTA/100 ,uM benzoyl-CoA/32 gLM malonyl-CoA (2 x 1 ml: 800 nM [14C]malonyl-CoA)/200 gg of dialyzed E. coli protein extract] were evaporated in a vacuum centrifuge and redissolved in 10% methanol in water. This material was subjected to reverse-phase HPLC on a 0.4 X 10 cm LiChrospher 100 RP-18 (5 ,um) column (Merck) connected to a Beckman 126 system gold gradient HPLC pump. Chromatography was performed with a flow of 1 ml/min using

Proc. Natl. Acad. Sci. USA 93 (1996) a linear gradient of methanol (0-60% in 40 min) in H20. The eluent was monitored at 220 nm using a Waters 990 + diode array detector. For spectral information, data from 200 to 400 nm was collected at a resolution of 1.4 nm. Furthermore, the radioactivity of each fraction was measured by scintillation counter, and the peak fractions were analyzed by TLC.

RESULTS Isolation and Characterization of GCHS17 and GCHS26 Clones. Among 3.4 million plaque-forming units screened, eight A clones hybridizing to GCHS1-3 cDNA probes were isolated. According to the classification based on the amplification of a 181-bp fragment from a conserved region (15), the clones were deduced to represent four different sequences. Two novel genes, GCHS17 and GCHS26, having a continuous reading frame (except for an intron) showed similarity to the GCHS2 gene. The exon/intron boundaries as well as the start and stop codons of the reading frames were deduced based on the general similarity of CHS enzymes at the amino acid sequence level and by comparison to the GCHS2 cDNA sequence. GCHS17 is a truncated clone missing approximately the first 50 codons. It has a 28-bp first exon, a 1638-bp intron, and a 1016-bp second exon. GCHS26 harbors the entire reading frame: a 196-bp first exon, a 483-bp intron, and a 1016-bp second exon. We compared the deduced amino acid sequences of GCHS17 and GCHS26 to each other and to the other asteraceous CHS superfamily sequences. The sequences form two subgroups: usual CHS-like sequences (with 88-93% intragroup identity) and GCHS2-like sequences (83-84% intragroup identity). The identity between the two groups is 73-77%. Next, we compared these sequences to a CHS consensus sequence that contains the 260 residues identical in nine functionally verified sequences of a wide evolutionary spectrum (15). GCHS2 deviates at 49, GCHS26 at 47, and the truncated GCHS17 at 42 positions. At 18 comparable sites, all GCHS2-like sequences deviate in an identical way from the CHS consensus. At an additional six sites, they deviate in concert from the consensus of GCHS1, CHS from Dendranthema grandiflora, and GCHS3. For comparison, GCHS1 and the CHS from D. grandiflora deviate from the CHS consensus at five positions, whereas GCHS3 deviates at 10 sites. The sequence analysis suggests that, in the Gerbera genome, GCHS2-like genes form a family, at least in the sense that they share some diagnostic characteristics in the primary structure of the corresponding enzymes. Isolation and Characterization of Gene Fragments from Other Asteraceae Species Using Primers Designed for the GCHS2 Gene Family. To study the distribution of GCHS2-like genes in the Asteraceae, we designed specific primers for PCR amplification based on the GCHS2 diagnostic sites (Fig. 1). Amplification products of expected size were obtained from Leibnitzia, Onoseris, Taraxacum, Dahlia, and Gerbera (for control). In Fig. 1, deduced amino acid sequences corresponding to the amplification products of the four species are shown. Each fragment has a reading frame without stop codons, and the degree of deviation from the CHS consensus is of the same order of magnitude as that of GCHS2 family of Gerbera, suggesting that the fragments analyzed represent coding regions of CHS superfamily genes. Leibnitzia LACHS1 and Onoseris OSCHS1 sequences share the clear majority of features common to the GCHS2 family. LACHS1 follows the GCHS2 consensus in 13 of 16 comparable sites and OSCHS1 follows this consensus in 11 of 16 comparable sites. The Taraxacum and Dahlia sequences also share some residues diagnostic for GCHS2-like sequences. TXCHS1 shares four of seven comparable positions with the GCHS2 consensus, DHCHS1 shares three of seven comparable positions, and DHCHS2 shares three of nine comparable positions. Furthermore, the Taraxacum and Dahlia sequences often

Proc. Natl. Acad. Sci. USA 93


IC

GCHS1

KS I R M TEE L NP C Y APSLDI CDKSMIRKRYMHITEEYLKQN NMCAYMAPSLD 95

MUMCHS GCHS3 GCHS2 GCHS26 GCHS17

CDKSMIRKRYMHLTEEYLKEN NLCEYMAPSLD 95 CDKSMIRKRYMHITEEFLKEN SMCK MAPSLD 98 CEK AIKKRYHALTEIYLQEN TMCE MAPSLD 97 CDK AIKKRY ALTEEYLKQN SMCE MAPSL 98 CEK AIKKRY VLTEIYLEK NMCE MAPS 42

LACHS1

--------RYMVLTEEYLKEN NMCEIMAPSL --------RYMVLTEEYLKEN NMCEgMAPSL NMCGYN PSL --------RYMFLTEIFLKDN NMCEYA PSLI

cons.

+

OSCHS1 TXCHS 1 DCHS1 cons.

GCHS1

MUMCHS GCHS3

GCHS2 GCHS26 GCHS17

*

*

AKD AENN GARV

RYIHHTEEFLKEN

9035

VVCSEITAVTI

FAGGTVLRLAKDLAENNKGARVLVVCSEITAVT 200 FAGGTVLRLAKDLAENNKIARVLVVCSEITAVT 200 FAGGTVLRLAKDLAENNKGARVLVVCSEITAVT 203 AGGTVLRLAKDLAENNKGERVLLVCSEITAE 202 203 AGG VLRLAKDLAENNKG RVLVVCIEITA 147 AGG VLRLAKDLAENNK RVLVICSEIT *

*

DHCS

25 25

25 25

*

AGGTVLRIAKDLAENNKG RVLVVClEIl A AGGt LRLAKDLAENNK RVLVVCSEIT GG LRLAKDVAENNK RVLVVCEE

130 130 130

RVLVVCIElIA RVVV1i

130

G ~Vf~ LA D AENN

AGG VLRLAKDIAENNKG

PDS GAIDGHLREVGL FHLLKDVPG SKNI L AF AQT ISD N FW AHPGG MVSAAQTILPDSEGAIDGHLREVGLTFHLLKDVPGLISKNIEKALTTAFSPLGIDWNS IFWIAHPGG 309 MISAAQTILPDSEGAIDGHLREVGLTFHLLKDVPGLISKNIEKALTQAFSPLGISDWNSIFWIAHPGG 309 MVSAAQTILPDSEGAIDGHLIEVGLTFHLLKDVP LI KNIEKALIQAFSPLNI DWNSIFWIAHPGG 312 IVST IVST IVST +

LACHS1 OSCHS1 DHCHS2

+

IFAGGTVLR

(1996)

QTILPDIE

A XHLRE QTILPDSE A _HLRE A HLRE

QTILPDgE

+

IVSA QTILPISE IVS QTLLPDSE IVSA QTIIPDE

*

*

A A A

L

GLTF GLTF GLTF

MV KNIENAEKASPLGI DWNSVFWM HPGG 311 KNIEDV VKE SPLGISDWNSLFW HPGG 312 LEDVP MVSKNIEDA MKA SPLGISDYNSLFWM HPGG 256

*

HLRE HLRE

I3HLRE

DVP DVP

L

*

MI

*

+

GLTFLDVP MIS NIKDVDQAFSPLGISDWNSL GLTFjL DVP MIKNIEDVIVKA SPLGISDWNSL GLK MI HPLGflDWNSM -------------

-------

HLDVP

SNIENILMKA

193 193 33

FIG. 1. Alignment of deduced amino acid sequences for CHS superfamily genes of Asteraceae. Three selected regions are shown, corresponding to the sequence ranges of PCR-amplified genes (the forward and reverse primer sites are denoted with hyphens). Positions of final residues per block are marked for each polypeptide. cons., strict consensus of CHS sequences based on 9 sequences from a wide taxonomic spectrum (15). The amino acid residues deviating from this consensus are shown with a black background. Consensus within the GCHS2 family of Gerbera that represents deviation from CHS is indicated by * when compared with all CHS sequences and + against CHS genes of Asteraceae only. Note that in complete sequences (not shown) two asterisked sites lie between the upper two blocks and three others lie after the last block. MUMCHS, CHS of D. grandiflora (27). LACHS1, OSCHS1, TXCHS1, DHCHS1, and DHCHS2 are the amplified fragments of Leibnitzia, Onoseris, Taraxacum, and Dahlia, respectively. These partial sequences vary in size because of cloning aspects; in Taraxacum, a sequence prematurely ending at a BamHI site was isolated, whereas in Dahlia, two BamHI fragments were isolated separately.

deviate from the CHS consensus in the same positions as the GCHS2 family of Gerbera, even if the derived residue is not identical. In contrast, there are several sites that are shared by Taraxacum and/or Dahlia clones and the CHS consensus that deviate from the GCHS2 consensus. Phylogenetic Analysis of Asteraceae and Asterid CHS Superfamily Genes. To study the phylogenetic relationships of the GCHS2 gene family, we performed a parsimony analysis of 19 CHS-like sequences at the nucleotide level (Fig. 2). As recent publications (13, 14) have shown general overviews of CHS superfamily phylogeny, we included only selected CHSlike sequences available for angiosperm subclass Asteridae. These included CHS for Apiaceae, Asteraceae, Convolvulaceae, Scrophulariaceae, and Solanaceae. Based on prior phylogenetic results from the plastid gene rbcL (24), CHS-like sequences from the first two families were expected to appear as one monophyletic branch of Asteridae, whereas the latter three were expected to occur in another. Genes selected included two unusual UV-inducible CHS genes of Petunia (CHSB and CHSG; ref. 9), the CHSB-related Ipomoea sequences of (14), and the GCHS2-like genes of Gerbera and other Asteraceae. The parsimony analysis yielded 2 equally most-parsimonious trees of 2720 steps with a consistency index of 0.50 and a retention index of 0.52 (31). One of these trees is shown in Fig. 2. The alternative tree differs only in the placement of Leibnitzia and Onoseris GCHS2-like sequences relative to those of Gerbera (Fig. 2). In every case, GCHS2-like sequences are derived from within CHS as sister-group to CHS3. Jackknife branch support values (23) suggest uncontradicted support for the clade comprising all Asteraceae CHS-like sequences and for each of the two major clades containing GCHS2-like

sequences (Fig. 2). GCHS2-like genes appear to be monophyletic in all mostparsimonious trees, but support for this association is weak

(Fig. 2). Nevertheless, it is likely that GCHS2 has emerged from CHS as the result of a single gene duplication event, with subsequent differentiation during the evolution of the Asteraceae. This duplication event must have occurred prior to diversification of the Asteraceae because all GCHS2-like genes (of tribes Mutisieae, Lactuceae, and Heliantheae) would be derived from the CHS3 and CHS1 lineages (Fig. 2), both of which include Gerbera (Mutisieae). Phylogenetic analyses of morphological and other molecular data indicate that tribe Mutisieae is the basal-most lineage of cichorioid-asteroid Asteraceae (-23,000 species), with only subfamily Barnadesioideae (92 species) more primitive in the family (16). Comparison of the Enzymatic Activities of GCHS2 and GCHS26 with Parsley CHS. To investigate the catalytic properties of GCHS26 and compare them with those of the previously analyzed GCHS2 and parsley CHS (15, 32), we cloned the cDNA into the expression vector pKKtac and produced the enzymes in E. coli. A parsley CHS cDNA was used as a reference for CHS function. As a control for the E. coli background, the vector with no insert was also used. Fig. 3A shows the products of the three enzymes formed with 4-coumaroyl-CoA as a substrate. The initial product of a typical CHS reaction is chalcone, but in vitro most of it is converted nonenzymatically to naringenin in the course of the reactions, and this is the main product observed with parsley CHS in the chromatogram. In contrast, GCHS2 and GCHS26 reveal no formation of naringenin, but produce a faint (and blurred) signal (P0.08; ref. 15) near the start. The radioactivity at the front detected with low-pH extractions probably represents malonic acid liberated from malonyl-CoA by thioesterases in the extracts (15). Since in our previous study (15) we found that GCHS2 is able to use benzoyl-CoA as a substrate (leading to the accumulation of a product, P0.51) we tested the activity of GCHS26 similarly. Fig. 3B shows that GCHS26 produces two signals

9036


Proc. Natl. Acad. Sci. USA 93 (1996)

CHS (Antirrhinum: Scrophulariaceae)

~~1

CHSB (Petunia: Solanaceae) CHSA (Ipomoea: Convolvulaceae)

86

100 CHSB (Ipomoea: Convolvulaceae) CHSG (Petunia: Solanaceae)

CHSJ (Petunia: Solanaceae) 1CHSA (Petunia: Solanaceae) l...... CHS1 (Lycopersicon: Solanaceae) -

CHS (Petroselinum: Apiaceae) ^^"sC4 UiL;Hlll (Gerbera: Asteraceae) MUMCHS (Dendranthema: Asteraceae) GCHS3 (Gerbera: Asteraceae) TXCHS1 (Taraxacum: Asteraceae) DHCHS1 (Dahlia: Asteraceae) .0 OSCHS1 (Onoseris: Asteraceae) -

I

81

75

LACHS1 (Leibnitzia: Asteraceae)

50

GCHS26 (Gerbera: Asteraceae) GCHS2 (Gerbera: Asteraceae) GCHS1 7 (Gerbera: Asteraceae)

FIG. 2. Phylogenetic relationships of GCHS2-like genes within the context of the CHS gene superfamily of angiosperm subclass Asteridae. The tree shown is one of two most-parsimonious topologies differing only in the reversed placements of OSCHS1 (from Onoseris) and LACHS1 (from Leibnitzia). The tree was oriented by treating the two major clades of Asteridae (24) as monophyletic groups. The same orientation was also indicated by midpoint rooting (28). Branch lengths are proportional to numbers of hypothesized nucleotide changes under the accelerated transformation character optimization (29, 30). The scale bar indicates 50 nucleotide changes. Numbers at nodes are parsimony jackknife support values; values close to or greater than 63% may indicate nodes set off by uncontradicted synapomorphies, whereas values between 50% and 63% indicate nodes with some robustness to extra steps (23). Support values under 50% are not shown. All Asteraceae CHS-like sequences form a group supported by 67% of jackknife replicates. GCHS2-like sequences are shown to derive monophyletically from CHS1 and CHS3, but support for this relationship is weak. It is therefore possible that all GCHS2-like genes trace to a single duplication event that took place before the diversification of Asteraceae; DHCHS1 from Dahlia (of the derived subfamily Asteroideae) is embedded within an asteraceous CHS-like clade otherwise dominated by Gerbera and other taxa of tribe Mutiseae, which occupies a primitive position in Asteraceae phylogeny (16).

(PO.51 and PO.9) specific for benzoyl-CoA. This indicates that both GCHS2 and GCHS26 differ from typical CHS by their substrate specificity; both enzymes are strikingly inactive with the natural substrate of CHS, 4-coumaroyl-CoA, but both are able to use benzoyl-CoA instead. Furthermore, neither enzyme is able to catalyze the formation of products that are soluble to organic phase in high pH, which contrasts with parsley CHS (and the two CHS enzymes of Gerbera with both 4-coumaroyl-CoA and benzoyl-CoA as substrates; ref. 15). To characterize further the product specificity of each enzyme with benzoyl-CoA we purified the major products by HPLC and analyzed their identity based on mobility in TLC (Fig. 3C) and UV absorption spectrum (Fig. 3D). Parsley CHS produces four products (PO.51a, PO.51b, PO.67, PO.82), one of which (PO.51a) is produced by both GCHS2 and GCHS26 and one of which (PO.82) is produced also by GCHS26, but not significantly by GCHS2. Based on this analysis, the two products soluble at high pH are PO.5lb and PO.67. To understand the relationship of the compound shared between GCHS2 and GCHS26 (PO.51a) to the other reaction products, we studied the effect of 2-ME, a known inhibitor of the progressivity of the CHS reaction (33). With parsley CHS, 2-ME decreases the formation of the normal end-product naringenin (probably by disturbing a cysteine residue at the enzyme's active site; ref. 34) and increases the formation of a byproduct, bis-noryangonin, which is a reaction intermediate

(33). As a control, we first tested the effect of 2-ME on the reaction with 4-coumaroyl-CoA. As expected, treatment with 100 mM concentration blocked naringenin formation and led to the production of a high mobility signal likely ascribable to liberated malonic acid (Fig. 3E). With benzoyl-CoA as substrate, increasing the amount of 2-ME in the reaction led to reduced formation of the high-pH soluble compounds PO.51b and PO.67, whereas the accumulation of PO.51a (not soluble to organic phase in high pH) remained high (Fig. 3F). As with 4-coumaroyl-CoA, the high mobility signal became stronger. This signal (and no others) increased even in a second control reaction where malonyl-CoA was the only substrate (Fig. 3F). In conclusion, the effect of 2-ME on the parsley CHS reaction with benzoyl-CoA strongly suggests that PO.51a represents an intermediate of the normal CHS reaction. Indeed, NMR structural analysis of the PO.51 compound indicates the absence of the second aromatic ring typical for CHS reaction, and therefore incomplete cyclization in the GCHS2 reaction (I. Kilpelainen, personal communication). This catalytic truncation with substrate change implies a functional deviation from CHS consistent with the virtually unaltered floral pigmentation of Gerbera plants transgenic for an antisense-GCHS2 Agrobacterium construct (35).

DISCUSSION The GCHS2 gene of G. hybrida has been shown to be a novel member of the CHS gene superfamily, encoding an enzyme



A

9037

B -F

S

-F ...

Awf

-....

!im:

qp

0 .;..

- Nar

-f

- 0.08 -

V mc hi

V Mc

to

P mc hi

- 0.51

P mc lo

26 mc hi

26 mc lo

SD

S

2 2 mc mc hi lo

V mb hi

C

V mb lo

P mb hi

P mb lo

0} - s

0 26 mb hi

26 mb lo

2 mb hi

2 mb lo

D -F

_NW.

0

:

fW

26/0.82 P/0.82

:.

P/0.51 26/0.51 2/0.51 P/0.67 P/0.51 b 2/ 2 P/ P/ P/ P/ P 26/ 26/ 26 0.51 mb 0.82 0.51 0.67 0.51b mb 0.82 0.51 mb lo lo lo

220 nm

E

350 nm

F -F

-F

S - 0.51

S_ ,K.?

a '

P P P P P P P P V V P P mb mb mb mb m m m m mb mb mc mc hi hi lo lo hi hi lo J lIolo lo lo +

+

+

+

- 0.51

- Nar

-S

-S

P P P P mb mb mb mb hi lo hi lo 0 0 12 12

P mb hi 60

P mb lo 60

P mb hi 300

P mb lo 300

FIG. 3. Chromatographic analysis of in vitro CHS reaction products. (A) TLC analysis with 4-coumaroyl-CoA and malonyl-CoA as substrates. 2, GCHS2; 26, GCHS26; P, parsley CHS expressed in E. coli; V, E. coli control (vector); c,-coumaroyl-CoA; m, malonyl-CoA substrates of the reactions. The pH at extraction is labeled lo (pH 4) and hi (pH 8.8). (B) TLC analysis with benzoyl-CoA and malonyl-CoA as substrates. b, benzoyl-CoA. (C) TLC analysis of the radioactive fractions purified by HPLC run adjacent to the nonpurified reactions. 2/0.51, 26/0.51, 0.82, P/0.51, 05lb, 0.67, 0.82, unknown products of GCHS2, GCHS26, and parsley CHS reactions, respectively. (D) UV spectrum of the fractions analyzed in C; (E) TLC analysis of the effect of 100 mM concentration of 2-ME to parsley CHS reaction. (-) absence/(+) presence of 2-ME. (F) TLC analysis of the effect of increasing concentration of 2-ME to parsley CHS reaction. 0, 12, 60, and 300 mM concentrations of 2-ME. S, F: start and front of the chromatograms. Nar, position of naringenin; 0.08, 0.51 position of unknown products.

with structural and enzymological properties as well as ontogenetic distribution distinct from CHS and SS (15). Here, we describe the evolutionary and functional relationships of GCHS2 to other genes of the superfamily. The Gerbera genome

harbors a family of GCHS2-like genes with at least three members, and amplified fragments of GCHS2-like genes have been obtained from other species of Asteraceae. Based on the phylogenetic hypothesis presented (Fig. 2), two strongly-supported

9038


lineages of GCHS2-like genes have emerged from CHS by gene duplication and functional divergence during the evolution of Asteraceae. Although the relationship is only weakly supported, maximum parsimony indicates a monophyletic origin for both of these lineages, suggesting that a single gene duplication gave rise to all GCHS2-like genes before Asteraceae diversified. Together with the recent results of Tropf et al. (13) and Durbin et al. (14), our study confirms a gene duplication based model that explains how various related functions have arisen from CHS during plant evolution. Based on some common motifs in primary structure and the similar sequential reaction mechanism, it has been suggested that CHS itself shares a common origin with fatty acid synthases of primary metabolism (36, 37). In the reaction of SS, which uses identical malonyl-CoA and 4-coumaroyl-CoA substrates, only the final cyclization step of the reaction is modified (37). In the GCHS2 reaction, as studied here, both the substrate specificity as well as the progressivity of the reaction have been changed. Altered substrate specificity in a CHS-like enzyme has been reported for acridone synthase, an enzyme in the alkaloid biosynthetic pathway using N-methylanthraniloyl-CoA as substrate (11, 12). Nevertheless, the truncation of the CHS reaction in GCHS2 catalysis is a novel feature. It suggests that the initially relatively complex CHS reaction has been simplified along the protein evolutionary process, and that this simplification must have been selectively advantageous to have been retained. If the catalytic steps of the CHS reaction (and those of related fatty acid synthases) evolved stepwise over time, then the truncated reaction of GCHS2 may be considered a reversal to a more primitive condition in enzymatic evolution. Truncation of the CHS reaction leading to the accumulation of novel metabolites of secondary metabolism has been hypothesized previously in the context of p-hydroxyphenylbutan-2-one biosynthesis in raspberry (39) and the biosynthetic origin of bis-noryangonins (33). Our present results with the GCHS2 reaction support these untested suppositions and imply that catalytic simplification (i.e., evolutionary reversal), like substrate change, may be a recurring theme in CHS superfamily evolution. The GCHS2-like sequences from Taraxacum (tribe Lactuceae, subfamily Cichorioideae) and Dahlia (tribe Heliantheae, subfamily Asteroideae) form a well-supported lineage apart from those of the three species of the tribe Mutisieae (subfamily Cichorioideae), which indicates a further divergence in the GCHS2 family along with the diversification of Asteraceae. Additionally, the three GCHS2-like genes of Gerbera clearly differ from one other. GCHS17 and GCHS26 are generally expressed at a lower level than GCHS2, and the strong expression in floral organs typical for GCHS2 is lacking (data not shown). Furthermore, in the comparison of the catalytic properties of GCHS2 and GCHS26, a slightly different product specificity was observed (Fig. 3B). In the future, in combination with biochemical investigation of the function of GCHS2, these studies on the molecular evolution of GCHS2 will lead to a greater understanding of the role, biological significance, and diversity of GCHS2 as a novel enzyme of secondary metabolism in the Asteraceae. We thank Hans V. Hansen for the Leibnitzia and Onoseris seed material; Neil Courtney-Gutterson for the CHS sequence of D. grandiflora; and James S. Farris, Jaakko Hyvonen, Ilkka Kukkonen, Barbara Meurer-Grimes, Joachim Schroder, Lena Struwe, and Risto Vainola for valuable discussions or assistance. Eija Holma, Marja Huovila, Paivi Laamanen, and Keijo Virta are acknowledged for their excellent technical assistance. This work was partially funded by the Academy of Finland, The Swedish Natural Science Research Council, and the Lewis B. and Dorothy Cullman Foundation. 1. Swain, T. (1986) in Plant Flavonoids in Biology and Medicine: Biochemical, Pharmacological and Structure-Activity Relation-


2. 3. 4. 5. 6.

7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21. 22.

23.

24. 25. 26. 27.

ships, eds. Cody, V., Middleton, E., Jr., & Harborne, J. B. (Liss, New York), pp. 1-14. Stafford, H. (1991) Plant PhysioL 96, 680-685. Koes, R. E., Quattrocchio, F. & Mol, J. N. M. (1994) BioEssays 16, 123-132. Niesbach-Klosgen, U., Barzen, E., Bernhardt, J., Rohde, W., Schwarz-Sommer, Z., Reif, H. J., Wienand, U. & Saedler, H. (1987) J. Moi. Evol. 26, 213-225. Fliegmann, J., Schroder, G., Schanz, S., Britsch, L. & Schroder, J. (1992) Plant Mol. Biol. 18, 489-503. Schroder, G., Brown, J. W. S. & Schroder, J. (1988) Eur. J. Biochem. 172, 161-169. Melchior, F. & Kindl, H. (1990) FEBS Lett. 268, 17-20. Sparvoli, F., Martin, C., Scienza, A., Gavazzi, G. & Tonelli, C. (1994) Plant Mol. Bio. 24, 743-755. Koes, R. E., Spelt, C. E., van den Elzen, P. J. M. & Mol, J. N. M. (1989) Gene 81, 245-257. Shen, J. B. & Hsu, F. (1992) Moi. Gen. Genet. 234, 379-389. Baumert, A., Maier, W., Groeger, D. & Deutzmann, R. (1994) Z. Naturforsch. 49, 26-32. Junghans, K. T., Kneusel, R. E., Baumert, A., Maier, W., Groger, D. & Matern, U. (1995) Plant Mol. Bio. 27, 681-692. Tropf, S., Lanz, T., Rensing, S. A., Schr6der, J. & Schr6der, G. (1994) J. Mol. Evol. 38, 610-618. Durbin, M. L., Learn, G. H., Jr., Huttley, G. A. & Clegg, M. T. (1995) Proc. Natl. Acad. Sci. USA 92, 3338-3342. Helariutta, Y., Elomaa, P., Kotilainen, M., Griesbach, R. J., Schroder, J. & Teeri, T. H. (1995) Plant Moi. Biol. 28, 47-60. Bremer, K. (1994) Asteraceae: Cladistics and Classification (Timber, Portland, OR). Jofuku, D. & Goldberg, R. B. (1988) in Plant Molecular Biology: A Practical Approach, ed. Shaw, C. H. (IRL, Oxford), pp. 37-66. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY), 2nd Ed. Dellaporta, S. L., Wood, J. & Hicks, J. B. (1983) Plant Mol. Biol. Rep. 1, 19-21. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science 239, 487-491. Vieira, J. & Messing, J. (1991) Gene 100, 189-194. Swofford, D. L. (1993) PAUP: Phylogenetic Analysis Using Parsimony (Illinois Nat. His. Survey, Champaign), Version 3.1. Farris, J. S., Albert, V. A., Kallersjo, M., Lipscomb, D. & Kluge, A. G. (1996) Cladistics, in press. Chase, M. W., Soltis, D. E., Olmstead, R. G., Morgan, D., Les, D. H., et al. (1993) Ann. Mo. Bot. Gard. 80, 528-580. Takkinen, K., Laukkanen, M.-L., Sizmann, D., Alfthan, K., Immonen, T., Vanne, L., Kaartinen, M., Knowles, J. K. C. & Teeri, T. T. (1991) Protein Eng. 4, 837-841. Hanahan, D. (1983) J. Mo. Biol. 166, 557-580. Courtney-Gutterson, N., Napoli, C., Lemieux, C., Morgan, A., Firoozabady, A. & Robinson, K. E. P. (1994) BiolTechnology 12,

268-271. 28. Farris, J. S. (1972) Am. Nat. 106, 645-668. 29. Farris, J. S. (1970) Syst. Zool. 19, 83-92. 30. Swofford, D. L. & Maddison, W. P. (1987) Math. Biosci. 87, 199-229. 31. Farris, J. S. (1989) Cladistics 5, 417-419. 32. Schuz, R., Heller, W. & Hahlbrock, K. (1983) J. Biol. Chem. 258, 6730-6734. 33. Kreuzaler, F. & Hahlbrock, K. (1975) Arch. Biochem. Biophys. 169, 84-90. 34. Lanz, T., Schroder, G. & Schroder, J. (1990) Planta 181, 169-175. 35. Elomaa, P., Helariutta, Y., Kotilainen, M. & Teeri, T. H. (1996) Mol. Breeding 2, 41-50. 36. Kauppinen, S., Siggaard-Andersen, M. & von WettsteinKnowles, P. (1988) Carlsberg Res. Commun. 53, 357-370. 37. Verwoert, I. I. G. S., Verbree, E. C., Van der Linden, K. H., Nijkamp, H. J. J. & Stuitje, A. R. (1992) J. Bacteriol. 174, 28512857. 38. Kindl, H. (1985) in Biosynthesis and Biodegradation of Wood Components, ed. Higuchi, T. (Academic, New York), pp. 349377. 39. Borejsza-Wysocki, W. & Hrazdina, G. (1994) Phytochemistry 35, 623-628.