5-Fluoroindole Resistance Identifies Tryptophan Synthase Beta Subunit

Copright 8 1995 by the Genetics Society of America

5-Fluoroindole Resistance Identifies Tryptophan SynthaseBeta Subunit Mutants in Arabidqpsis thalianu

Andrea J. Barczak, Jianmin Zhao, Kim D. Pruitt and Robert L. Last Plant Molecular Biology Program, Boyce Thompson Institute for Plant Research and Section of Genetics and Deveiopment, Cornell University, fthaca, New York 14853-1801 Manuscript received May 27, 1994 Accepted for publication January30, 1995

ABSTRACT A study of the biochemical genetics of the Ambidopsis thaliana tryptophan synthase beta subunit was initiated by characterization of mutants resistant to the inhibitor 5-fluoroindole. Thirteen recessive mutations were recovered that are allelicto trp2-1, a mutation in the more highly expressedof duplicate tryptophan synthase beta subunit genes (TSBI).Ten of these mutations (trp2-2 through trp2-11) cause a tryptophan requirement (auxotrophs),whereas three (trp2-100 through trp2-102) remain tryptophan prototrophs. The mutations cause a variety of changes in tryptophan synthase beta expression. For example, two mutations (trp2-5 and trp2-8) cause dramatically reduced accumulationof TSB mRNA and immunologically detectable protein, whereastrp2-10 is associated with increased mRNA and protein. A correlation exists between the quantity of mutant beta and wild-type alpha subunit levels in the trp2 mutant plants, suggesting that the synthesis of these proteins is coordinated or that the quantity or structure of the beta subunit influences the stabilityof the alpha protein. The levelof immunologically detectable anthranilate synthase alpha subunit proteinis increased in the trp2 mutants, suggesting the possibility of regulation of anthranilate synthase levels in response to tryptophan limitation.

AST studies of the biochemistry and genetic regula- (NIYOGIet al. 1993), phosphoribosylanthranilate trans- P tion of amino acid biosynthesis in plants were ham- ferase (step 2 in Figure 1) trpl mutants (LAST and FINK pered by the paucity of useful mutants (LAST 1993). 1988; ROSE et al. 1992), and the tryptophan synthase This is in contrast to microorganisms, where amino acid alpha subunit (TSA; step 5 in Figure 1) t7p3 mutants auxotrophs are among themost frequently isolated gen- (unpublished) and the beta subunit (TSB; step 6 in eral class of mutants. The phenylalanine, tyrosine and Figure 1) tqZ-1 mutant (LAST et al. 1991). Although tryptophan biosynthetic pathways of plants are particu- mutations are not described for the remaining three larly interesting because they produce precursors to a proteins, Arabidopsis genes were characterized for an- variety of secondary metabolites, in addition to amino thranilate synthase alpha subunit (MA, step 1 in Figure acids for proteinsynthesis. For example, the tryptophan 1) (NIYOGIand FINK1992), PR-anthranilate isomerase branch provides precursors to indolic secondary metab (step 3 in Figure 1) (LI et al. 1995b) and indole-3-glyc- olites, includingthe hormone indole-%aceticacid erol phosphate synthase (step 4 in Figure 1) (LI et al. (WRIGHTet al. 1991; NORMANLYet al. 1993), antimicro- 1995a). bial phytoalexins (TSUJIet al. 1993), glucosinolates The enzyme tryptophan synthase is an especially at- (HAUGHN et al. 1991) and alkaloids (CORDELL1974). tractive target for detailed biochemical genetic analysis In contrast to the pathways leading to other amino in plants because the structureand function of bacterial acids, tryptophan biosynthesis is proving to be especially tryptophan synthase are very well understood (MILES well suited to genetic analysis in Arabidopsis thaliana 1991). In bacteria, the enzyme existsas an a2& hetero- (LAST and FINK1988; LAST et al. 1991; ROSEet al. 1992; tetramer, which catalyzes the conversion of indole-3 NIYOGIet al. 1993; reviewed in ROSE and LAST 1994) glycerol phosphate and serine to tryptophan and glycer- and other plants (FANKHAUSERet al. 1990; WRIGHTet aldehyde-3-phosphate. However, each subunit can less al. 1992). As shown in Figure 1, structural gene muta- efficiently catalyze a half-reaction: tions in Arabidopsis are available that cause defects in a-subunit activity: four of the seven proteins that convert chorismate into tryptophan, and the structural genes affected in these Indole-3-glycerol phosphate e, Indole mutants are characterized. These are the anthranilate + Dglyceraldehyde-%phosphate synthase beta subunit (step 1 in Figure 1) trp4 mutants &subunit activity:

Corresponding authoc Robert L. Last, Boyce Thompson Institute for Indole + L-Serine + L-tryptophan + H20 Plant Research, Tower Rd., Ithaca, NY 148531801. E-mail: [email protected] Although the reactions carried out by fungal TS are

Genetics 140 303-313 (May, 1995) 304 A. Barczak et al.

Chorismate We identified Arabidopsis mutants by selection for TRP4 growth on 5-fluoroindole. Genetic, molecular and im- Anthranilate munological analyses are presented for the 13 mutants that containrecessive tq2alleles. These lines are shown TW1 to range from severely affected tryptophan-requiring Phosphoribosylanthmdate auxotrophs to leaky mutants that grow in the absence of exogenous amino acid. The TSB subunit mRNA and @I protein accumulate to varying extents, ranging from l-(O-Carboxyphenylano)-ld~xy~b~~5.P approximately 5% to 150% of the wild type. Thus, the selection for 5-fluoroindole resistance yielded an allelic series of Arabidopsis trp2 mutations. Indole-3-glycerolphosphate MATERIALSAND METHODS

Trpsynth.se a TRP3 Plant strains and growth conditions: The Columbia ecotype of Arabidopsis was used as the wild-type line throughout this work. For all biochemical assays, rosette leaves of Sweek-old I plants grown in soilless mix (Peat-lite Redi Earth from W. R. Indole Grace, or Cornel1 Mix) were harvested into liquid nitrogen 5-Fluoroindole and stored at -80" until extraction.Unless otherwise specified, plant growth was at 22" under continuous illumination. For Trpsyn*S B sterile culture, plants were grown in 100-mmdiameter Petri E5-Fluorotryptophan plates sealed with parafilm (American National Can, Green- wich, CT) on sterile plant nutrient medium with 0.5% sucrose Tryptophan (PNS) (HAUGHN and SOMERVILLE 1986) solidifiedwith 0.75% FIGURE 1.-The tryptophan biosynthetic pathway in Arabi- Bacto agar. Plants were grown in soilless mix under 80-100 dopsis. The enzymatic steps are in boxed numbers, as dis pmol/m*/sec photosynthetically active radiation (PAR) from cussed in the text, and the four genes for which mutations mixed cool-white and incandescent lights or from MVR400/ are available are indicated. The presumed TSB-mediated con- U multivapor metal halide bulbs (General Electric). version of 5fluoroindole to 5fluorotryptophan is also indi- Mutagenesis: Mutagenesis was performed on batches of cated. 50,000 seeds, and the self"ferti1ized progeny Mn seeds were harvested from separate pools of 5000 MI plants. EMS mutagenesis was performed as described, using 0.25-0.30% (v/ identical, the two subunits are fused into a single poly- v) EMS (Sigma) (ESTELLEand SOMERVILLE1987). Gamma peptide: in Neurospma crassa and Saccharomyces cerevisiae irradiation (30 krad) was performed in a I3'Cs source (Gam- the amino-terminal domain is homologous to the alpha macell1000; AECL Industrial, Ontario, Canada) on seeds subunit and the carboxyl-terminus is homologous to that had imbibed in water for 4 hr at room temperature. the beta subunit (BURNSand YANOFSKY1989; WKIN Nitrosomethylurea (NMU, Sigma) mutagenesis was done by a modification of HAGEMANN (1982). Seedswere imbibed and YANOFSKY1982). overnight at room temperature in water and then treated at Two lines of evidence indicate that the plantenzyme room temperature for 4 hr with 150-200 pM NMU in a pH 5.0 is composed of separate subunits, reminiscent of pro- buffer (0.0485 M citric acid and 0.103 M sodium phosphate karyotes, The Arabidopsis (BERLYNet al. 1989; LAST et dibasic), with gentle shaking. The seeds were washed with500 al. 1991) and Zea mays (maize) (WRIGHTet al. 1992) ml water 10 times,and the NMU contaminated solutions and solids were detoxified overnight with 5 M NaOH. Seeds that beta subunit genesdo notcontain alpha subunit coding were mutagenized with diepoxybutane (DEB;Sigma) were information. Similarly, the Arabidopsis alpha subunit imbibed overnight in water at room temperature and then gene is devoid of sequences homologous to the beta treated with 22 mM DEB in water for 5 hr at room tempera- subunit (E. R. RADWANSKI,J. Zmo and R. L. LAST, ture, with gentle shaking. The seeds were washed five times unpublisheddata). Biochemical studies also suggest with 1 liter of water each time, before planting. The DEB solution was detoxified overnight with 2% sulfuric acid, and that theactive enzymefrom Nicotiana tabacum (tobacco) the othercontaminated materials were treated with 1% HZSO, and Pisum sativum (pea) is composed of alpha and beta for 24-48 hr. subunit activities (DELMERand MILLS 1968; CHENand Mutant selections and genetic manipulations: In selection BOLL 1971; NAGAOand MOORE 1972). experiments, Mnseeds were surface sterilized(hT and FINK Analysis the Arabidopsis trp2-1 mutant, which is 1988) and suspended at 500 seeds/ml in 0.1 % agar. Aliquots of of 500 seeds were mixed with 4.0 ml of molten 50" agar and deficient in TSB subunit activity, revealed it to be rela- poured onto the surface of 100-mmdiameter Petri plates con- tively insensitive to the toxic effects of 5fluoroindole taining 25 ml PNS supplemented with 200 pM 5-fluoroindole (LAST et al. 1991). This observation suggests that 5-flu- plus 40 pM tryptophan (to permit growth of tryptophan auxo- oroindole toxicity results from conversion to 5-flUOro- trophic plants), and the plates were incubated at 100 pmol/ tryptophan by wild-type TSB protein. In this respect, m2/sec PAR. Putative resistant plantswere identified after 14- 21 days, transferred to PNS containing 50 pM tryptophan for Arabidopsis responds to 5-fluoroindole in the same way 5-7 days to recover, and then transplanted to soilless mix to as do the fungi S. cerevisiae (MIOZZARIet al. 1977) and self-pollinate.The 5-fluoroindole resistancephenotype of MB, Copinus n'nmeus (TILBY1978). F1 or F2 seeds was rescreened by planting 100 seeds per plate Arabidopsis Tryptophan Mutants 305 containing 100 pM 5fluoroindole plus 40 pM tryptophan, TABLE 1 and the phenotype was scored after 14-21 days growth. Tryp tophan auxotrophy was assayed by comparing growth at 180 Sequencing primers pmol/m2/sec PARof seeds on minimal PNS medium containing or lacking tryptophan (LAST and FINK 1988; LAST et Primer Sequence al. 1991; ROSEet al. 1992). Nucleic acid and protein isolation and detection: Plant ge- KP 30 nomic DNA (CONE1989) and total RNA (DEVRIESet al. 1988) KP 31 were isolatedby published methods. Genomic DNA (AUSUBEL KP 32 et al. 1994) and RNA gel electrophoresis and blot transfers KP 33 (PRUIITand HANSON 1991), and blot hybridizations(CHURCH KP 34 and GILBERT 1984)were performed as described. The gel- KP 35 purified TSBl cDNA and genomic probes were a 0.8-kb SacI- KP 36 XhoI and a 4.2-kb EcoRI fragment, respectively (BERLYNet al. KP 37 1989).The TSAl cDNA probe, which encodes the Arabidopsis KP 38 TSA subunit, was a 1.2-kb EcoRI fragment from pERR65 (E. R. KP 39 RADWANSKI, J. ZHAO and R. L. LAST, unpublished data). DNA KP 40 probes were labeled with “P usingthe Megaprime DNA Label- KP 41 ing Kit (Amersham, Arlington Heights, IL), and unincorpe rated nucleotides were removedwith Nuctrap Push Columns KP 42 (Stratagene, La Jolla, CA). The abundance of TSA and TSB AB1 mRNA was normalized to 26s rRNA as previously described AB2 (PRUITTand LAST 1993). Production of rabbit antisera, pre AB3 tein sample preparation, protein gel electrophoresis, immu- AB4 noblotting and Phosphoimagerradioactivity quantification AB5 were performed as described by ZHAO and LAST (1995). AB6 DNA sequence analysis of ng2 mutant alleles: The TSBl genes were cloned from EcoRIdigested tq2-5 and tq2-8 mutant total DNA, which was separated by preparative agarose trp2-1, a recessive Arabidopsis mutation shown to cause gel electrophoresis. DNA was isolated from the region span- ning an estimated size range of 3-6 kb and purified with reduced TSB subunit activity due to a mutation in TSBl, Glassmilk silica beads (Bio 101 Inc., La Jolla, CA). We con- the more highlyexpressed of duplicate TSB genes firmed that the gel slicewas not contaminated with TSB2 DNA (LAST et al. 1991). This mutantwas found tobe resistant by testing a sample using Southern blot hybridization with to 5-fluoroindole, indicating that direct selection for5- the TSBl genomic probe, which hybridizes efficiently withthe fluoroindole resistance would yield mutants altered in TSB2 locus as demonstrated in LAST et al. (1991). The size- selected DNA was then ligated into EcoWdigested and alka- TSB activity. We reasonedthat these mutants might line phosphatase-treated AZAPII, packaged withGigapack include a wide variety of trp2 structural gene alterations, Gold, and infected into Escherichia coli XL1-Blue MRF’, using as well as mutations in other genes that less directly the recommendations of the manufacturer (Strategene). Ni- trocellulose plaque hybridizations were performed with di- influence TSB activity. This hypothesis was tested by goxigenin-labeled TSBl genomic probe and detected with subjecting a total of 390,000M2 seeds (obtainedby self- alkaline phosphatase conjugated antidigoxigenin FAB (Ge- pollination of plants derived from mutagenesis of seeds nius System, Boehringer Mannheim Corp., Indianapolis, IN). with EMS, NMU, DEB or gamma irradiation) to selec- The recombinant pZAPIIplasmid was excised using M13 helper phage ExAssist (Stratagene) and purified with a Qla- tion on medium supplemented with both 5-fluoroin- gen-tip 20 column (Qlagen Incorporated, Chatsworth, CA). dole and tryptophan. A total of 132 putative resistant Double-stranded plasmid DNA was sequenced with TSBl oli- mutants that exhibitedgood root growth and true leaf gonucleotides using the DyeDeoxy Terminator Sequencing development were transferred tosoil. These plantswere Kit (Applied Biosystems Incorporated, Foster City, CA) and the Thermolyne TEMP-TRONIC DB66925 thermocycler. The allowed to self-pollinate, and the M3 progeny were re- DNA fragments were purified with CentriSep spin columns tested for inhibitor resistance and tryptophan require- (Princeton Separations, Adelphia, PA) and then loaded onto ment. an Applied Biosystems 373A Automated DNA Sequencer. Nineteen independently derived 5-fluoroindole-re- DNA sequence of tq2-lOwas obtained by PCR amplification of total DNA. The 5’ half of the TSBl gene was amplified sistant mutants were chosen for further analysis be- with primers AB3 and AB4 and the 3‘ half of the gene was cause they exhibited strong inhibitor resistance in the amplified with primers AB5 and AB6. Single stranded tem- Ms (see Table 2). Two additional 5-fluoroindole-resis- plate was obtained by the method of HICUCHIand OCHMAN tant lines, which were originally identified as resistant (1989), and DNA sequence was obtained by manual sequenc- to Gmethylanthranilate (trp2-11 and FIR 17H2) (J. LI ing methods using Sequenase as specified by the manufacturer (US. Biochemical). The MacVector software (IBI, New and R. LAST, unpublished results), were also included Haven, CT) was used for sequence analysis. The accuracy of in this collection. Of these 21 mutants, 10 are trypto- the mutant DNA sequence was verified by sequencing the phan-requiring auxotrophs under our standard high- altered region on both strands. The oligonucleotides used light growth conditions, whereas 11 lines are capable are listed in Table 1. of growth on sterile minimal medium lacking trypto- RESULTS phan under all conditions tested. As described below, Isolation of 5fluoroindole resistant mutants: Previ- 13 of the 5-fluoroindole-resistantlines (including all ous 5-methylanthranilate resistance selections yielded of the auxotrophs and three prototrophs)recessive are 306 A. Barczak et al.

TABLE 2 (tryptophanprototrophy and 5-fluoroindole sensitiv- Mutant selection ity). Onequarter of the selfcross F2 progeny were of mutant phenotype (tryptophan requiring or 5-fluoro- New allele 5FI" isolate Tryptophan indole resistant). F2 progeny were not obtained for the designation designation Mutagen phenotype tq2-9 cross to wild type because the F1 plants died be- trp2-2 11Al EMS - fore setting seed, apparently due to a problem with the trp2-3 13H2 EMS - batch of soil used to grow these individuals. One of the trp2-4 11D2 EMS - auxotrophic mutants (tq2-7) was very difficult to cross trp2-5 llEl EMS - and was not outcrossed to the wild type. Although the t92-6 11F1 EMS - F2 obtained from the crosses of t92-8 and t92-11 were trp2-7 EMS 13A9 - not tested for tryptophan auxotrophy, they were scored trp2-8 17H8 Gamma - for 5-fluoroindole resistance, shown in Table 4. tq2-9 lFl0a NMU - as trp2-10 10C4 EMS - Results obtained with the prototrophic mutantsindi- trp2-11 150102 NMU - cate that these mutations are also recessive, as the F1 trp2-loo 11B4 EMS + progeny were uniformly 5-fluoroindole sensitive (Table trp2-101 12D1 EMS + 5). Surprisingly, analysis of the tq2-100 F2 cross yielded trp2-I 02 14B3 EMS + an unexpectedly high number of 5-fluoroindole-resis- 13B1" EMS + tant progeny (33%). When calculated for an expected 13H4" EMS + segregation ratio, these data result in a chi-square 17G3" Gamma + 3:l 17H2" Gamma + value of 32.1, which does not fit the expectation at P 1.45 NMU + = 0.05. F2 analysis of the other two prototrophic alleles 1A5a DEB + (tq2-101 and tq2-102) yielded the expected 3:l ratio. 1D8 NMU + As this cross was only performed once, the significance 14E9 EMS + of this skewed segregation pattern is unclear. "The indicated isolates were shown to contain recessive Allelism tests wereperformed to ask whether theana- Mendelian traits that complement trp.2. logue resistance and tryptophan requirement were caused by auxotrophic tq2 mutations, with the results trP2 alleles. Phenotypic characterization of the re- shown in Table 6. In each case it was demonstrated that maining eight prototrophic mutations indicated that the heteroallelic F1 plants were resistant to 5-fluoroin- four are recessive Mendelian mutations that comple- dole (and unable to grow in the absence of tryptophan ment thetq2 tester alleles (A.J. BARCZAK,unpublished if the new allele was auxotrophic), demonstrating allel- results). Preliminary datafrom complementation ismwith the tester mutation. Consistent with the hy- crosses among these four recessive mutants revealed a pothesis that thenew mutations are in tq2, tight linkage minimum of two complementationgroups. One of between each pair of mutations was demonstrated by these alleles (isolate 13H4) failed to complement the the absence of recombinant wild-type F2 progeny. This TSA-deficient trp3-1 mutation and is likely to be defec- analysis was initially complicated by the observation that tive in the alpha subunit. Crossesof the remaining it was difficult to obtain successful crosses using the four mutants with wild type yielded F2 progeny that severely affected tq2-1 auxotrophicmutant as the could not be reliably scored for 5-fluoroindole resis- tester line. To facilitate complementation crosses, two tance, hampering their further characterization. of the new auxotrophic 5-fluoroindole-resistant mu- As was previously observed with the tq2-1 mutation, tants were used in this analysis. First,the tq2-2 and t92- the auxotrophic plants are smaller than the wild type 3 mutations were shown to be allelic to tq2-1, and then when grown on nonsterile soilless potting mix with or the tq2-6 mutant was crossed to t92-2, demonstrating without tryptophan supplementation. However, growth allelism (Table 6). The remaining mutants were then of these new mutants is generally more robust than that crossed with tq2-1, tq2-3or tq2-6 to complete the com- of the original tq2-l mutant (A. J. BARCZAK,unpub- plementation and linkage analyses. lished results). Growth of the three prototrophic mu- Tryptophan synthase gene product accumulation in tants in soil is less severely affected than that of the ag2 mutants: It was previously shown that tq2-1 is an auxotrophs, consistent with the hypothesis that these allele of TSBI, the more highly expressed of duplicate are leakier tq2 alleles. TSB subunit structural genes in Arabidopsis (LAST et al. Genetic characterization of the B-fluoroindole-resis- 1991). RNA blot hybridization and immunoblotanalysis tant mutants The results of crosses to wild-type Arabi- were used to ask whether any of the new t92alleles were dopsis indicate that 5-fluoroindole resistance and trypte altered in accumulation of TSA or TSB gene products. phan auxotrophy are conferred by recessive monogenic Analysis of the levels of these mRNAs and proteins in traits in the mutants tested. As shown in Tables 3 and the 5-fluoroindole-resistant mutants revealed a number 4, the F1 progeny derived from crosses with nine of the of interesting differences. Figure 2 shows that the t92- auxotrophic mutants displayed the wild-type phenotype 8 and t92-5mutants had themost dramaticallyreduced Arabidopsis Tryptophan Mutants 307

TABLE 3 Results from crosses with wild type: tryptophan auxotrophy

Cross“ TYPe Trp-Total Trp+ % Trp-

~~~~ ~ ~ TRP2/ TRP2X tq2-2/ tq2-2 F1 6 6 0 0 F2 42 1 337 84 20 TRP2/ TRP2TRP2/ X tq2-3/tq2-3 13 FI 13 0 0 426 319 F2 426 107 25 TRP 2/ TRP2 TRP2/tq2-4 X tq2-4/ FI 6 6 0 0 429 320 109 25.4 109 320 F2 429 TRP 2/ TRP2 TRP2/tq2-5 X tq2-5/ FI 12 12 0 0 382 271 F2 382 111 29.1 TRP2/TRP2 X trp2-6/ tq2-6 FI 6 6 0 0 F2 423 323 100 23.6 TRP 2/ TRP2TRP2/ X trp2-8/ tq2-8 FI 10 10 0 0 TRP2/TRP2 X trp2-9/trp2-9 F1 9 9 0 0 TW2/ TRP2 X tq2-1O/ tq2-I 0 FI 8 8 0 0 400 296 F2 400 104 26 TRP2/TRP2 X tq2-1tq2-11 I/ FI 7 7 0 0 “Pollen from the M3 5-fluoroindole resistant mutant was crossed with wild type to generate F1 seeds. The heterozygous F, plants were allowed to self-pollinate to yield F2 progeny. accumulation of TSB mRNA (7% and 34% of wild type, As shown in Figures 3, bottom, and 4B, quantitative respectively). In contrast, tq2-9 and tq2-10plants had immunoblot analysiswith polyclonal antibodies re- modest, but reproducible, enhancement of TSB RNA vealed that TSB protein levels in the trp2 mutants vary (28% and 40% higher than the wild type, respectively). from very low (13% of wild type for trp2-5 and 15% for In all cases, TSA mRNA abundance is the same as, or trp2-8) to significantly greater than that ofwild type slightly higher than, that of the wildtype (data not (50-60% higher for trp2-10, trp2-100and trp2-102). The shown). Although the results obtained for TSA mRNA observation that the trp2-5 and trp2-8 mutants produce quantitation were consistent throughoutthree inde- small amounts of TSB protein is consistent with the pendent experiments, we are cautious about drawing dramatic reductions of mRNA (Figure 2). None of the strong conclusions about the significanceof the en- mutants accumulate stable nonsense peptides of lower hanced mRNA levels because the TSA mRNA hybridiza- molecular weight or proteolytic fragments that are de- tion signal is very weak in total RNA. tectable with the polyclonal antisera.

TABLE 4 Results of crosses with wild type: 5-fluoroindole resistance

Cross“ Type Total FIS~ FIR %FIR

TRP2/ TRP2 X tq2-2/ tq2-2 14 14 0 0 263 201 62 23.6 TRP2/ TRP2 X tq2-3,’ tq2-3 18 18 0 0 283 218 65 23 TRP2/ TRP2X tq2-4/ tq2-4 12 12 0 0 276 21 1 65 23.6 TRP2/ TRP2 X tq2-5,’ tq2-5 19 19 0 0 263 208 55 23.6 TW2/ TRP2 X tq2-6/ tq2-6 13 13 0 0 441 32 1 120 27.2 TRP2/ TRP2 X tq2-8/tq2-8 8 8 0 0 896 684 212 23.7 TRP2/ TRP2X tq2-9/ tq2-9 6 6 0 0 TRP2/ TRP2 X tq2-1O/ tq2-10 16 16 0 0 625 460 165 26.4 TRP2/ TRP2X tq2-1 I/ tq2-11 14 14 0 0 594 434 160 26.9 “Pollen from the M, 5-fluoroindole resistant mutant was crossed with wild type to generate F, seeds. The heterozygous F1 plants were allowed to self-pollinate to yield F2 progeny. * 5-Fluoroindole sensitive. 5-Fluoroindole resistant. 308 A. Barczak et al.

TABLE 5 Results of crosses with wild type: prototrophic mutants

Cross" Type Total FIS~ FIR %FIR

TRPZ/ TRPZ X trp2-100/ trp2-I00 FI 18 18 0 0 F2 3601088 728 33.1 TRP2/TRP2 X trp2-101/trp2-l01 F, 10 10 0 0 F2 3291252 923 26.2 TRP2/TRPZ X trp2-102/trp2-102 F, 12 12 0 0 F2 3251223 898 26.6 Pollen from the M3 5-fluoroindole resistant mutant was crossed with wild type to generate F1 seeds. The heterozygous FI plants were allowed to self-pollinate to yield F2 progeny. 5-Fluoroindole sensitive. ' 5-Fluoroindole resistant.

Despite the fact that these mutants have lesions in a type TSA subunit, whereas tq2-10 and tq2-100 have TSB gene and are presumably wild type for TSA, the reproducibly higher concentrations of both the TSA steady-state TSA protein concentrations are altered in and TSB proteins. These results are consistent with the the majority of the tr-2 mutants (Figures 3, top, and idea that the wild-type TSB protein stabilizes the plant 4).Figure 5 illustrates a trend between the level of TSA subunit in vivo. This is a plausible hypothesis be- accumulation of mutant TSB subunits and steady-state cause it is known that interactionsbetween the bacterial concentration of the wild-type TSA protein. At one ex- subunits influence the conformation of the alpha sub treme, tr-2-5and tq2-7 accumulate -50% of the wild- unit (MILES1991). Data from this laboratory indicate

TABLE 6 Results of complementation crosses

Tryptophanauxotrophy 5-fluoroindole resistance

Cross" Type %Trp-Total Trp- Trp+ FIS* FIR %FIR trpZ-I/trpZ-l X trp2-2/trp2-2 20 0 6 100 0 14 100 300 0 200 100 0 100 100 trp2-l/trpZ-l X trp2-3/trp2-3 10 0 10 100 300 0 200 100 0 100 100 trp2-2/trp2-2 X trp2-6/trp2-6 15 0 2 100 0 13 100 959 0 266 0 693 100 trp2-6/ trp2-6X trp2-4/ tq2-4 8 0 8 100 752 0 752 100 trpZ-I/trpZ-l X trp2-5/trp2-5 5 100 0 3 100 203 0 203 100 trp2-3/trp2-3 X trp2-7/tq2-7 13 100 0 6 100 88 1 0 88 1 100 trp2-6/trp2-6 X trp2-8/trp2-8 19 100 0 12 100 1013 0 1013 100 trp2-6/trp2-6 X tq2-9/trp2-9 5 0 5 100 966 0 966 100 trp2-3/trp2-3 X trp2-10/trpZ-l0 15 0 8 I00 0 7 100 360 0 360 100 tqt12-3/trp2-3X trp2-1 I/ trp2-I 1 10 0 10 100 73 0 73 100 tl;b2-3/trp2-3X trp2-100/trp2-100 21 10 0 0 0 11 100 191 0 191 100 trp2-6/ trp2-6 X trp2-101/ trp2-I 01 14 4 0 0 0 10 100 81 0 81 100 0 0 0 20 100 0 1285 100

" Pollen from the MS 5-fluoroindole resistant mutant was crossed with a plant of the indicated genotype to generate F1 seeds. The heterozygous F, plants were allowed to self-pollinate to yield F2 progeny. 5-Fluoroindole sensitive. ' 5-Fluoroindole resistant. Arabidopsis Tryptophan Mutants 309

180

El T -r I I II I I

FI(:uRE 2.-TSB mRNA accumulation in hp2 mutants. All values are corrected for rRNA hybridization, and the mutant results are normalized to the wild type (“100”). Each value represents the mean of triplicate blots from a single experi- ment, and error bars are standard error of the mean.

Anti. TSa

Anti. TSP

Flcum 3.-Immunoblot analysis of tryptophan synthase alpha and beta subunit accumulation in trp2 mutants. Dupli- FIGURE4.-Quantitation of tryptophan svnthase subunit with cate protein blots were probed antiserum against TSA levels in tv12 mutants. A and B show data for TSA and TSB (top) orTSB subunit (bottom) followed by “%protein A and results, respectively.The data plotted represent the cross-reac- then subjected to autoradiography. The doublet above the tive material that migrated with the band ofwilcl-type mobility. alpha subunit signals is nonspecific: it is reduced in intensity The mutant values are normalized to the wild-type signal for when affinity purified antibody is used and is present in ex- each blot (“100”). Each datum represents the mean ofresults tracts from lq3-2 mutant, which accumulates undetectable from eight immunoblots of extracts derived from three indc- levels of TSA mRNA and protein (E. R. RADWANSKI, J. Z1-1AO pendent sets of plants. Standard errors of the mean are indi- and R. L. LMT, unpublished data). cated. that the Arabidopsis TSA and TSB subunits co-purify during immunoaffinitychromatography, consistent Increased accumulationof anthranilate synthase alpha with the existence of stable subunit interactions (E. R. subunit: Previousresults demonstrated -50% in- RADWANSKI, J. ZHAO and R. L. LAST, unpublished re- creased anthranilate synthase enzvme activity in both sults). It is interesting to note that there are exceptions the tql-100 (Table 5 in NIY~GIP/ 01. 1999) and tqf12-I to the correlation betweenTSA and TSB. For example, (Table 3 in L~STP/ (11. 1991) mutants. We extended tq2-I and tq2-I1 have nearly wild-type levels of TSA these results by examining whether the concentration protein despite decreased mutantTSB subunit accumu- of immunologically detectable ASA protein was simi- lation. larly affected in leaf tissue of the thirteen /7;h2 mutants 310 A. Barczak et al.

.-e a c, trp2-10 9 140- n trp2- 100 tj cn trp2-9 trp2-102 F 120- rc 0 trp2-6 e 0 .= 100 ------Q - trp2- 1 tP2- 7 7 -F I E' tp2-3 80- 0= I tfp2-8 I Y trp24 trp2-2 I a trp2-101 .? 60- c, trp2-7 -Q I 2 tP2-5 I I 40 I'I'I'I'I~I'I~I'I' 40 60 80 100 120 140 160 180 160 140 120 100 0 80 2060 40 21 Relative Accumulationof TSP Protein

FIGURE5.-Comparison of tryptophan synthase subunit levels in trp2 mutants. The data plotted are from Figure 4, and the coefficient of linear regression (R') = 0.75. tested. The results summarized in Figure 6 indicate that these mutants have reproducibly increased levelsof ASA antigen, ranging from two- to sixfold higher than the wild type. Although further experimentswill be necessary to assess the significance of the allele-specific differences and toelucidate the mechanism(s)responsible for this increase, it is clear that a defect in the last step in tryptophan biosynthesis alters the regulation of the commiting enzyme ASA protein. Molecularbiological analysis of the lrp2 mutations: We tested the hypothesis that tr-2 mutations might be associated with deletionsor other rearrangements at the TSBl locus that would be detected by genomic Southern blot hybridization. When genomic DNA of the mutants was digested with restriction enzyme EcoRI, tq2-10 was the only allele that displayed an aberrant pattern (A. J. BARCZAK,unpublished results). This mutantyielded EcoRI fragments of 1.3 and 2.9 kb, instead of the wild-type 4.2 kb fragment. This result is consistent with a mutation that creates a new FIGURE6.-Anthranilate synthase levels in tq2 mutants. EcolU recognition site within the TSBl gene. As shown The mutant values are normalized to the wild-type for each blot ("100"). Each datum represents the mean of data from in Figure 7, DNA sequence analysis of tq2-10 revealed triplicate independent extracts, each assayed on duplicate im- a transition mutation at position 2859 of the TSBl gene munoblots, for a total of six experiments. Standard errors of that creates an EcoRI site and converts a phylogeneti- the mean areindicated. Extracts from tq2-1I are not included cally conserved glycine (at amino acid 266 in Figure 2 in this set of data because multiple extractswere not available. of BERLW et al. 1989) to a glutamate residue. To understand the molecular defects that led to dra- transition mutation thatresults in conversion of a phylo- matically reduced protein accumulation, the TSBl ge- genetically conserved glycine to arginine at aminoacid nomic DNA sequence was also obtained for tq2-5 and 351 (BERLINet al. 1989). In contrast, the gamma ray- trp2-8 (Figure 7). The EMS-induced allele tq2-5 is a induced allele tq2-8 has amore complex structure: Arabidopsis Tryptophan Mutants 31 1

TRP2 K WG G K P AAA TWGGG AAG CCG G 1

FIGURE 7.-DNA sequence changes inthe tq2-5 and tq2-8 and tq2-10 alleles. The TSBl gene is represented schematically, with exons as filled rect- angles and introns as thinner lines. The 5' 3' wild-type and mutant sequences are in- TSB1 dicated with the altered bases in bold- face type.

AGFEGI GCG GOA TTC GAG CGAATC m2 I I t-2-10 GCG GAA TTC GAG AGA ATC t-2-5 AE F ERI there is a single base pair deleted from a cluster of four important plants. Because such mutants might have al- guanine residues (at positions 2742-2745 of the wild- tered indolic secondary metabolism, these would be type sequence) and a transversion mutation six nucleo- particularly useful to obtain in plants thatproduce tides downstream (at position 2751 of the wild-type se- pharmacologically important indolic alkaloids. quence). Translation of this mutantopen reading We analyzed the DNA sequence of the TSBl locus frame is expected to yield a truncated protein of 251 for three mutants, two of which were found to contain amino acids (the inferred wild-type protein is 470long), missense mutations. The t7p2-10 mutation creates an which would contain 23 abberant amino acids at the EcoRI polymorphism as a result of a G to A transition. new COOH terminus. This mutation is expected to produce an altered protein with a negatively charged amino acid (glutamate) DISCUSSION in place of an absolutely phylogeneticallyconserved glycine (BERLYNet al. 1989). The role of this specific site Selection for resistance to 5-fluoroindole was used to was not defined from mutational or structural analysis identify 13 new mutations at the Arabidopsis TW2 lo- of the S. typhimun'um enzyme. However,it is in a region cus, which encodes TSBl, the more highly expressed of that interacts with the alpha subunit and is involved in duplicate tryptophan synthase beta subunit genes. The producing the hydrophobic tunnel (HYDEet al. 1988). selection also produced several other classes ofmutants, Presumably, the change from a small neutral amino including four lines that contain recessive alleles that acid to a charged residue in this important region af- complement trp2 and define at least two other comple- fects the catalytic activity of the enzyme, leading to a mentation groups (A.J. BARCZAK,unpublished results). tryptophan auxotrophy. tq2-5 also contains a G to A Interestingly, one of these complementing recessive al- transition mutation that changes a phylogenetically leles (isolate 13H4) causes accumulation of a variant conserved glycine to a positively charged arginine resi- TSA subunit that migrates unusually slowly on SDS due. This amino acid falls in a random coil domain PAGE gels (data not shown). This result suggests that between helices 11 and 12 (HYDEet al. 1988). Surpris- changes in the TSA protein can affect the efficiency of ingly, these two missense mutations both affectTSB the beta subunit-mediated conversion of 5-fluoroindole mRNA accumulation but in opposite directions. Whereas to 5-fluorotryptopan. This observation is consistent with trp2-10 causes enhanced mRNA and protein levels, trp2- results in Salmonella typhimunum, demonstratingthat 5 mutant plants accumulate only 34% of wild-type changes in the conformation of either subunit can af- mRNA and an even lower amount of protein. It is un- fect the catalyticactivity of the other (MILES 1991). clear why these mutations alter TSB mFWA and protein Thus, the selection for 5-fluoroindole resistance in Ara- accumulation. Further biochemical analysisof these bidopsis is versatile because it notonly yieldstrp2 alleles missense mutant proteinsand molecular analysis of the but also identifies plants with other types of defects. It remaining alleles are necessary to understand the rea- is likely thatthe 5fluoroindole resistance selection sons for differences in accumulation of various mutant could also be sucessfully used to identify TSdeficient beta subunits. mutants of other organisms, including agronomically The tq2-8 mutant, which has the lowest concentra- 312 A. Barczak et al. tions of TSB mRNA and protein, has both a single nu- account for the apparent influence of beta subunit on cleotide deletion and a transversion clustered within a alpha protein accumulation is that the plant uses a ge- 10-nucleotide region of the third exon of TSBl (Figure netic regulatory mechanism to maintain the 1:l subunit 7). This complex mutation is expected to cause a trans- stoichiometry required for maximal catalytic activity.In lational frameshift and premature termination of trans- this scenario, the cell would respond to accumulation lation following amino acid 251. Despite the expecta- of unusually low or high amounts of TSB protein con- tion thatthe mutant tsbl gene would encode a centration by reducing or stimulating the synthesis of truncated protein, theTSB protein that accumulates in TSA. This model is complicated by mutants such as trp2- tq2-8 plants is the same size as that in the wild type 1 and trp2-11, which deviate from the general correla- (Figure 3, bottom). This strongly suggests that theresid- tion between beta and alpha protein accumulation in ual TSB protein that accumulates in the trp2-8 mutant the trp2 mutants. is produced by the duplicate TSB2 gene. The reduced The increased ASA protein in trp2 mutants (Figure accumulation of TSB mRNA is consistent with the hy- 6) and enzyme activity in trpl and trp2 mutants (LAST et pothesis that the t92-8 mutation causes decreased sta- al. 1991; NIYOCI et al. 1993) might also reflect a genetic bility of the mutant tsbl mRNA. Analogous frameshift regulatory mechanism. In onescenario, decreased tryp- mutations were demonstrated to cause dramatically re- tophan biosynthetic capacity would directly trigger an duced mRNA accumulation for both a Glycine rnax (soy- increase in synthesis or stability of the committing en- bean) Kunitz trypsin inhibitor (JOFUKU et al. 1989) and zyme anthranilate synthase. An alternate model is sug- Phaseolus vulgaris (pintobean) phytohemagglutinin gested by the previous observation that mRNAs for both (VOELKERet al. 1990). In bothcases, the reduced tran- subunits of this enzyme are induced in response to script accumulation was due to mRNA destabilization stress induced by mechnical wounding and bacterial (JOFUKU et al. 1989; SULLIVANand GREEN1993). pathogenesis (NIYOGIand FINK 1992; NIYOCI et al. The two clustered single nucleotide changes in the 1993). It is possible that synthesis of anthranilate syn- gamma ray-induced trp2-8 allele contrast with the rela- thase is induced by a wider variety of stressagents, and tively large changes found in other characterized ioniz- the observed increase in anthranilate synthase in trp2 ing radiation-induced Arabidopsis mutations. For ex- mutants is the result of a tryptophan limitation-pro- ample, deletions were observed by Southern blot voked stress response. hybridization for all three gamma ray-induced chl3 al- The increased activity ofthis regulated enzyme might leles in the major nitrate reductase gene MA2 (WILKIN- be expected to increase the flux of chorismate into SON and CRAWFORD1991) and for chll-5, one of three indolic biosynthesis. This may account for the observa- gamma-induced mutations in a nitrate transporter gene tion that the trp2-1 mutant accumulates unusually high (TSAYet al. 1993). Similarly, fast-neutron mutagenesis levels of indolic compounds such as the ester and amide induced an inversion of the chalcone isomerase TT5 conjugates of indole-3-acetic acidand freeindole-3-ace- geneand x-irradiation caused a complex tt3 dihy- tonitrile (NORMANLYet al. 1993). Further characteriza- droflavonol4reductase allele consisting of two adjacent tion of this regulatory response should provide new deletions with an inversion of the large intervening re- insights into the regulation of the Arabidopsis trypto- gion (SHIRLEYet al. 1992). Although our Southern blot phan pathway in general and into the committing en- hybridization data do not rule out the possibilityof zyme in particular. These studies may also suggest novel large rearrangements with breakpoints beyond the 4.2- approaches to manipulating indolic secondary metabo- kb EcoRI fragment thatwas analyzed, it seems likely that lism in plants. the trp2-8 reduction in TSB mRNA is due to the two We thank JIAVANG LI for providing the t92-11 and 5FIR 17H2 mu- point mutations that were identified. This is because it tants and helpful conversations, andJoHNWOOLFORD and KRIS NIY- was previously demonstrated that the 4.2-kb fragment OGI for critical comments on the manuscript. This work was sup was sufficient to complement the tq2-1 mutation and ported by National Institutesof Health grantGM43134 and National restore wild-typeTSB enzyme activity in transgenic Science FoundationPresidential YoungInvestigator Award DMB 9058134 to R.L.L. AutomatedDNA sequencing was performed atthe (LAST plants et al. 1991). Cornell Bloanalytic Laboratory. Although mutations in the TRP2 gene were shown to affect levels of three different tryptophan biosynthetic proteins (TSA, TSB and MA),it seems likely that these LITERATURE CITED altered protein levels are caused by different mecha- AUSUBEL,F. M., R. BRENT,R. E. KINGSTON, D.D. MOORE,J. G. SEIDMAN nisms. The differences in TSBprotein accumulation are et al., 1994 Cuwent Protocols in Molecular Biology. Wiley Intersci- presumably the direct result of the mutations causing ence, New York. BERLYN,M. B., R. L. LAST and G. R. FINK, 1989 A gene encoding altered synthesis or stability of the mutant protein. In the tryptophan synthase p subunit of Arabidqbsis thaliana. Proc. contrast, the positive correlation between TSA and TSB Natl. Acad. Sci. USA 86 4604-4608. subunit accumulation (Figure 5) suggests that the TSA BURNS, D. M., and C. YANOFSKY,1989 Nucleotide sequence of the Neurospora crassa t7p-3 gene encoding tryptophan synthetase and protein is less stable in the absence of a normal interac- comparison of the t~p-3polypeptide with itshomologs in Saccham tion with the TSB subunit. An alternative hypothesis to myces cerm'siae and Escherichia coli.J. Biol. Chem. 264: 3840-3848. Arabidopsis Tryptophan Mutants 313

CHEN,J., and W. G. BOLL,1971 Tryptophan synthase: a two-compc- tryptophan analogues in Succhumyces cereuisiae. Arch. Microbiol. nent enzyme from pea plants. Can. J. Bot. 49: 1155-1163. 115 307-316. CHURCH,G. M., and W. GILBERT,1984 Genomic sequencing. Proc. NAGAO,J. T., and T. C. MOORE, 1972 Partial purification and prop Natl. Acad. Sci. USA 81: 1991-1995. erties of tryptophan synthase of pea plants. Arch. Biochem.Bio- CONE,K., 1989 Yet another plant DNA prep. Maize Genet. Newsl. phys. 149 402-413. 63: 68. NIYOGI,K. K., and G. R. FINK, 1992 Two anthranilate synthase genes CORDELL,G. A., 1974 The biosynthesis of indole alkaloids. Lloydia in Arabidopsis: defense-related regulation of the uyptophan 37: 219-298. pathway. Plant Cell 4: 721-733. NIYOGI,K R. L. L”, G. R FINK and B. KEITH,1993 Suppressors DELMER,D. P., and S. E. MILLS,1968 Tryptophan synthase from IC, of trpl fluorescence identify a new gene, TRP4, encoding the Nicotiana tabacum. Biochim. Biophys. Acta 167: 431-443. anthranilate synthase beta subunit. Plant Cell 5 1011-1027. DEVRIES, H. HOGEand T. BISSELING,1988 Isolation of total and S., NORMANLY,J., J. D. COHENand G. R. FINK, 1993 Arabidopsis thaliana polysomal RNA from plant tissues, pp. 1-13 in Plant Molecular auxotrophs reveal a tryptophan-independent biosynthetic path- Biology Manual, edited by S. B. GELVIN,R. A. SCHILPEROORTand way for indole-%aceticacid. Proc. Natl. Acad. Sci.USA 90: 10355- D. P. S. VEW. Kluwer Academic Publishers, Dordrecht, The 10359. Netherlands. PRum, K. D., and M. R. HANSON, 1991 Transcription of the Petunia ESTELLE,M. A,, and C. R. SOMERVILLE,1987 Auxin-resistant mutants mitochondrial CMSassociated Pcflocus in male sterile and fertil- of Arubidopsis thalianu with an altered morphology. Mol. Gen. ity-restored lines. Mol. Gen. Genet. 227: 348-355. Genet. 206 200-206. PRuITT, K. D., and R. L. LAST, 1993 Expression patterns of duplicate FANKHAUSER,H., F. FWHOUD and P. J. KING,1990 A tryptophan tryptophan synthase p genes in Arabidopsis thaliana. Plant Physiol. auxotroph of Hyoscyamus muticus lacking tryptophan-synthase ac- 102: 1019-1026. tivity. Planta 180 297-302. ROSE,A. B., and R. L. LAST, 1994 Molecular genetics of amino acid, HAGEMANN,R., 1982 Induction of plastome mutations by nitroso- nucleotide and vitamin biosynthesis,pp. 835-879 in Arubidopsis, urea-compounds, pp. 119-127 in Methodr in ChhophtMolecular edited by E. MEYEROWITZand C. SOMERVILLE. ColdSpring Har- Biology, edited by M. EDELMAN,R. B. HALLICK and N. H. CHUA. bor Press, Cold Spring Harbor, W. Elsevier Biomedical Press, New York. ROSE,A. B., A. L. CASSELMAN and R. L. LAST, 1992 A phosphoribosylanthranilate transferase gene is defective in blue fluorescent HAUGHN,G. W., and C. R. SOMERVILLE,1986 Sulfonylurea resistant Arabidopsis thaliana tryptophan mutants. Plant Physiol. 100: 582- mutants of Arabidopsis thaliana. Mol. Gen. Genet. 204 430-434. 592. HAUGHN, G. W., L. DAVIN,M. GIBLINand E. W. UNDERHILL,1991 SHIRLEY,B. W., S. HANLEY and H. M. GOODMAN,1992 Effectsof Biochemical genetics of plant secondary metabolites in Arabi- ionizing radiation on a plant genome: analysis oftwo Arabidopsis dopsis thalianu: the glucosinolates. Plant Physiol. 97: 217-226. transparent testa mutations. Plant Cell 4: 333-347. HIGUCHI,R. G.and H. OCHMAN,1989 Production of single-stranded SULLIVAN,M. L., and P. J. GREEN,1993 Post-transcriptional regula- DNA templates by exonuclease digestion following the polymer- tion of nuclearencoded genes in higher plants: the roles of ase chain reaction. Nucleic Acids Res. 17: 5865. mRNA stability and translation. Plant Mol. Biol. 23 1091-1104. HYDE,C. C., S. A. AHMED,E. A. PADLAN,E. W. MILESand D. R. DAVIES, TILBY,M. J., 1978 Inhibition of Corpinus n’neracs by 5fluoroindole. 1988 Three-dimensional structure of the tryptophan synthase Arch. Microbiol. 118: 301-303. a& multienzyme complex from Salmonella typhimurium J. Biol. TSAY,Y. F., J. I. SCHROEDER,K. A. FELDMANN and N. M. CRAWFORD, Chem. 263: 17857-17871. 1993 The herbicide sensitivity gene CHLl of Arabidopsis en- JOFUKU, K. D., R. D. SCHIPPERand R.B. GOLDBERG, 1989 A codes a nitrate-inducible nitrate transporter. Cell 72 705-713. frameshift mutation prevents Kunitz txypsin inhibitor mRNA ac- TSUJI,J., M. ZOOK, R. HAMMERSCHMIDT,R. L. LAST and S. C. SOMER- cumulation in soybean embryos. Plant Cell 1: 427-435. VILLE,1993 Evidence that tryptophan is not a direct biosynthetic intermediate of camalexin in Arabidopsis thalianu. Physiol. LAST,R. L., 1993 The genetics of nitrogen assimilation and amino Mol. Plant Pathol. 43: 221-229. acidbiosynthesis in plants: progress and prospects, pp. 297- VOELKER,T. A,, J. MORENO and M. J. CHRISPEELS,1990 Expression 330 in International Review of Cytology, edited by K. W. JEON,M. analysis of a pseudogene in transgenic tobacco: a frameshift mu- FRIEDLANDERand J. JARVIK.Academic Press, San Diego, CA. tation prevents mRNA accumulation. Plant Cell 2: 255-261. LAST,R. L., and G. R. FINK,1988 Tryptophan-requiring mutants of WIWNSON,J. Q., and N. M. CRAWFORD,1991 Identification of the the plant Arabidopsis thuliana. Science 240 305-310. Arabidopsis CHL3 gene as the nitrate reductase-structural gene LAST,R. L., P. H. BISSINGER,D. J. MAHONEY,E. R. RADWANSKI and NIAZ. Plant Cell 3 461-471. G. R. FINK,1991 Tryptophan mutants in Arabidopsk the conse- WRIGHT,A. D., M. B. SAMPSON,M. G. NEUFFER,L. MICHALCZUK,J. P. quences of duplicated tryptophan synthase p genes. Plant Cell SLOVINet al., 1991 Indole-%aceticacid biosynthesis in the mu- 3: 345-358. tant maize orangepmicarp, a tryptophan auxotroph. Science 254: LI,J., S. CHEN,Z. LIHUANGand R. L. LAST, 1995a Isolation of DNAs 998-1000. encoding the tryptophan pathway enzyme indole-%glycerol phos- WRIGHT,A. D., C. A. MOEHLENKAMP,G. H. H. PERROT,M. G. NEUFFER phate synthase from Arabidopsis thalianu. Plant Physiol. (in press). and K. C. CONE,1992 The maize auxotrophic mutant orange LI,J., J. ZHAO, A. B. ROSE, R.SCHMIDT and R. L. LAST, 1995b Arabi- pencarp is defective in duplicate genes for tryptophan synthase dopsis phosphoribosylanthranilate isomerase: molecular genetic p. Plant Cell 4: 711-719. analysis of triplicate tryptophan pathway genes. Plant Cell (in ZALIUN,H., and C. YANOFSKY,1982 Yeast gene TRP5: structure, func- press). tion, regulation. J. Biol. Chem. 257: 1491-1500. ZHAO, J., and R. L. LAST, 1995 Immunological characterization and MILES, E. W., 1991 Structural basis for catalysis by tryptophan syn- chloroplast localization of the tryptophan biosynthetic enzymes thase, pp. 93-172 in Advances in Enzymology and Related Areas of of the flowering plant Arabidopsis thuliana. J. Biol. Chem. 270: Molecular Biology, edited by A. MEISTER.John Wiley and Sons, 6081-6087. New York. MIOZZARI,G., P. NIEDERBERGERand R. HVITER, 1977 Actionof Communicating editor: J. CHORY