Plant Physiol. (1991) 97, 139-1460032-0889/91/97/01 39/08/$01 .00/0

Received for publication February 11, 1991Accepted May 6, 1991

Acid Phosphatase-1 from Nematode Resistant Tomatol

Isolation and Characterization of its Gene

Valerie M. Williamson* and Greggory ColwellDepartment of Nematology, University of California, Davis, California 95616

ABSTRACT

Aps-1 encodes acid phosphatase-1, one of the many acidphosphatases present in tomato (Lycopersicon esculentum Mill.).Aps-1 is closely linked to Mi, a gene conferring resistance againstnematodes. Thus, a clone of Aps-1 would provide access to theregion of the genome containing Mi. Acid phosphatase-1 waspurified from tomato suspension culture cells. Fragmentary aminoacid sequences were derived from the purified protein and fromits proteolytic and chemical digestion products. One of theseamino acid sequences was used to design an oligodeoxyribonu-cleotide probe expected to hybridize to acid phosphatase-1cDNA. This probe identified, in a cDNA library, a clone encodingthe carboxyl-terminal sequence of a protein that is very similar,but not identical, to acid phosphatase-1. Using this clone, wediscovered a second cDNA clone that corresponds in its carboxylterminal sequence to acid phosphatase-1 but, surprisingly, re-tains sequences of an Aps-1 intron. The second cDNA clone wasused to detect both a cDNA clone and a genomic clone corre-sponding to Aps-1. The identity of these clones was confirmedby sequence analysis and by the correlation of a restrictionfragment length polymorphism with two Aps-1 alleles in a segre-gating tomato population. The deduced amino acid sequence ofthe Aps-1 open reading frame predicts a hydrophobic animo-terminal signal sequence and a mature protein with a molecularweight of 25,000. The amino acid sequence of this protein has astrong similarity in size and sequence to a vegetative storageprotein of soybean.

Apases2 belong to a broad group of enzymes that catalyzethe hydrolysis of inorganic phosphate from phosphomonoes-ters at low pH. Apases are ubiquitous in nature and activityhas been reported in a wide range of plants (1, 8, 12, 16).Plant Apases vary substantially in size, tissue and subcellularlocalization and regulation of expression ( 1, 8, 12, 13, 16, 2 1,

23). Many roles have been postulated for these enzymes inplants, including a role in release of inorganic phosphate fromorganic phosphate in the environment under conditions ofphosphate limitation (13). Plants examined thus far containseveral different Apases.There are several Apase isozymes in tomato that are distin-

'This work was supported by U.S. Department of Agriculturegrant No. 88-37234-3538 to V.W.

- Abbreviations: Apase, acid phosphatase; RFLP, restriction frag-ment length polymorphism; kb, kilobase

139

guished by their electrophoretic mobility and/or other prop-erties (23, 24). A 57 kD Apase is induced by phosphatedeprivation and secreted by suspension culture cells into themedium (13). Two other acid phosphatase isozymes, Apase-1 and Apase-2, have been mapped to genetic loci, Aps-l andAps-2, respectively (24). Two alleles ofAps-l, designated Aps-I' and Aps-l', are found in domestic tomato. The encodedproteins, Apase- 1+ and Apase- 1' are separated by starch gelor cellulose acetate electrophoresis (6). Apase- 1 is the electro-phoretic form commonly found in the cultivated tomato,Lycopersicon esculentum. Apase- ' is a variant found in thewild species Lycopersicon peruvianum and in many tomatolines that carry Mi, a gene conferring resistance to root-knotnematodes (24). The corresponding allele, Aps-l', was intro-duced into cultivated tomato with Mi when this resistancegene was introduced from the wild species L. peruvianum(26). Aps-] and Mi are tightly linked (less than 0.89 recom-bination units apart) on chromosome 6 (21). The tight linkageof Aps-l to Mi has made Aps-] an obvious starting point forchromosome walking to the nematode resistance gene, Mi.

Studies on the localization and regulation of phosphatasesare hampered by the inability to distinguish all species. How-ever, because of its distinctive electrophoretic mobility, suchanalyses of tomato Apase-1 expression have been possible.Activity of Apase-1 is developmentally regulated (21), i.e.activity is relatively high in leaflets, hypocotyl tissue, andapical meristems of young plants but is absent in dry orgerminating seed, cotyledons, leaf petioles, petals, anthers,and mature fruits (21). Activity increases in root tips of bothresistant and susceptible tomato plants after root-knot ne-matode infection (K. Lambert and V.M. Williamson, unpub-lished results). Apase-1 activity is not induced by phosphatedeprivation and is not secreted into the medium by suspensionculture cells (V.M. Williamson, unpublished observation).The Apase- 1 protein, a dimer of about 51,000 D, has beenpurified from tomato cell suspension culture and some of itsproperties were determined (23). The enzyme has a very lowpH optimum for activity, pH 3.5 to 4.0, compared withApases isolated from tomato fruit or other plant species. Italso has a distinct substrate specificity, preferring sugar phos-phates to ATP as substrate.We report here the cloning and characterization of cDNA

and genomic clones of Aps-l'. Isolation of the clone relied onthe use of a long oligodeoxyribonucleotide probe designedfrom fragmentary amino acid sequence data obtained fromanalyses of the purified protein. The identity of clones was

WILLIAMSON AND COLWELL

verified by the presence of an RFLP that correlated with theAps-l+ and Aps-l' alleles of tomato plants. Although RFLPsare rare within cultivated tomato lines, they are commonbetween the DNA of tomato and its wild relatives (4, 15).Because the Aps-J' allele originated from the wild species, L.peruvianum, the DNA corresponding to this gene was likelyto display an RFLP.

MATERIALS AND METHODS

Plant Materials

VFNT cherry tomato (Lycopersicon esculentum Mill. cvVFNT; genotype: Aps-l', Mi) seed was obtained from C.Rick, University of California, Davis; Roma (Aps-lJ, mi) andRossol (Aps-l+, Mi) are from Petoseed, Woodland, CA.;Castlerock II (Aps-l+, mi) and Sun 6082 (Aps-l', Mi) arefrom G. Campbell, Sunseeds Genetics, Inc., Hollister, CA.The cultivar Sun 6082 is a result of five backcrosses, screeningfor the Aps-l' allele, to Castlerock II.

DNA Libraries

Two tomato cDNA libraries and one genomic library werescreened in this work. The first, an amplified cDNA libraryof VFNT cherry tomato, constructed from polyadenylatedRNA of 15-d-old seedlings in Xgtl0, was obtained from W.Gruissem, University of California, Berkeley. The secondcDNA library, constructed in plasmid pARC7 using polyad-enylated RNA from hypocotyl tissue of a nematode-resistantline, was obtained from S. O'Neill, University of California,Davis. A genomic library, derived from a partial Sau3Adigestion of VFNT cherry tomato DNA cloned into Charon35, was obtained from R. Fischer, University of California,Berkeley.

Protein Purification

Throughout the purification, fractions containing Apase-1were identified and distinguished from other Apase isozymesby using cellulose acetate electrophoresis and a specific stainfor Apase activity (23).Tomato cell suspension cultures were propagated as previ-

ously described (23) at 27°C in 2800 mL Fernbach flasks (1L medium/flask). Six liters of culture were harvested whentheir settled volume was 50% of the total volume after allow-ing cells to settle for 15 min at room temperature. Cells werepoured through a 150 ,m pore sieve then rinsed with ice-cold50 mM Tris-HCl, pH 7.6. Protein was extracted from cells bythe addition of 200 mL of ice-cold extraction buffer (100 mMsodium acetate, pH 4.7, 0.25% Triton X-100, 5 mm DTT)per L of cell suspension. Apase-1 and other proteins werereleased from cells by shaking at room temperature for 10min at 100 rpm. The cells were separated from the extract bypouring over one layer of nylon membrane (44 gm pore,Small Parts, Inc.). The pH of the extract was adjusted to 7.6using 2 M Tris-HCl, pH 8.3. Particulates were removed bycentrifugation at 5000g for 10 min. The supernatant wasloaded onto a 100 mL Q-Sepharose (Fast-Flow, Pharmacia)radial flow column (Sepragen) equilibrated with 50 mM Tris-HCI, pH 7.6, at 4°C. The column was washed with 125 mM

NaCl, 50 mm Tris-HCl, pH 7.6, and Apase- 1 was eluted with225 mm NaCl, 50 mm Tris-HCl, pH 7.6. Fractions with Apase-1 activity were collected and adjusted to a final concentrationof 1 M (NH4)2SO4 by adding 3 M (NH4)2SO4. The solution wasloaded onto a 1.5 x 40 cm phenyl-Sepharose CL-4B (Phar-macia) column equilibrated with 1 M (NH4)2SO4. The columnwas washed with one column volume of 600 mM (NH4)2SO4,50 mM Tris-HCl, pH 7.6, followed by one column volume of450 mM (NH4)2SO4, 50 mM Tris-HCl, pH 7.6. Apase-1 waseluted by applying 350 mM (NH4)2SO4 in 50 mm Tris-HCl,pH 7.6. Fractions with activity were pooled and dialyzedagainst 25 mM Tris-HCl, pH 7.6, containing 0.1% Triton X-100. Ampholytes Bio-Lyte 3/5 and Bio-Lyte 4/6 (both ob-tained from Bio-Rad) were added to a final concentration of1% (w/v) each in a 40 mL volume. Proteins were separatedfor 3.5 h at 12 W in a Rotophor preparative isoelectricfocusing apparatus (Bio-Rad). Fractions with Apase- 1 activitywere pooled and dialyzed against 50 mM Tris-HCl, pH 7.6,and applied to a 3 mL Bio-Gel HTP (Bio-Rad) columnequilibrated in 50 mM Tris-HCl, pH 7.6. The column waswashed with two column volumes of 20 mm sodium phos-phate, 50 mM Tris-HCl, pH 7.6, and eluted with 30 mMsodium phosphate, 50 mM Tris-HCl, pH 7.6. Fractions withactivity were pooled and dialyzed against 50 mm Tris-HCl,pH 7.6. The sample was applied to a 12 mL Sepharose-Q ionexchange column and eluted in a 0 to 300 mm NaCl gradientin 50 mM Tris-HCl, pH 7.6. Fractions containing activitywere pooled, then dialyzed against 25 mM ammonium bicar-bonate, and lyophilized.

Amino Acid Sequence Analysis

Apase- 1 was cleaved with endoproteinase Lys-C (Boehrin-ger-Mannheim) or cyanogen bromide (Sigma) to obtain pep-tides for sequence analysis. For the endoproteinase digestion,7.5 to 10 mg of Apase- I was resuspended in 60 mL of 25 mMTris-HCl, pH 8.5, 1 mm EDTA, 0.2% SDS, then denaturedby boiling 10 min. Lys-C (330 ng in 70 mL of 25 mm Tris-HCI, pH 8.5, 1 mm EDTA) was added and the mixture wasincubated 4 h at 25°C. The digested Apase- 1 was lyophilizedand resuspended in 50 mL Laemmli sample buffer (17) withthe SDS reduced to 1.75% (w/v) and 2-mercaptoethanolfreshly added to a 0.25% (v/v) final concentration. For thecyanogen bromide cleavage, approximately 20 ,g of purifiedprotein was treated as described by Gross (14).

Peptides were denatured by boiling for 7 min and separatedon a 1 mm thick, 15% SDS-polyacrylamide minigel (Bio-Rad) prepared using the solutions of Laemmli (17) at 40 mA.Electrophoretically resolved peptides were electroblotted ontoImmobilon P PVDF membrane (Millipore) and stained withCoomassie brilliant blue according to Matsudaira ( 19) exceptthat the gel was presoaked in transfer buffer for 30 min.Coomassie brilliant blue-stained peptide bands were cut outand stored at -20°C. Peptides were sequenced directly fromthe membrane using an Applied Biosystems model 470 se-quenator operated by the Protein Structure Laboratory, Uni-versity of California, Davis.

140 Plant Physiol. Vol. 97, 1991

TOMATO ACID PHOSPHATASE-1 GENE

Screening DNA Libraries

The cDNA library in Xgt 1O was plated on 150 mm Petridishes of LBM (Luria broth with 8 mM MgSO4) using Esche-richia coli C600 as host. Screening was carried out using thealkaline denaturation technique of Benton and Davis (2).Bacteriophage DNA was fixed onto Nytran membranes(Schleicher & Schuell, Inc.). The membranes were prehybrid-ized at 37°C in a solution of 5x SSPE (lx SSPE = 0.18 M

NaCl; 10 mM NaPO4, pH 7.7; 1 mM EDTA), 5x Denhardt'sreagent (10), 0.1% SDS, 10% dextran sulfate, 0.05% pyro-

phosphate, and 100 ,g/mL sheared, denatured herring testesDNA as carrier. The 53-nucleotide probe (Fig. IC) was end-labeled (20) to a specific activity of 1.2 x 108 cpm/,ug andadded to the prehybridization solution to a final concentrationof 2 x 106 cpm/mL. Membranes were incubated for 48 h at42°C and then washed (final wash, 0.5x SSPE, 0.1I% SDS, at41 C). The filters were exposed to Kodak XAR-2 film withintensifying screen at -80°C for 16 h. Putative positiveplaques were picked and subjected to two additional roundsof screening using the oligodeoxyribonucleotide probe de-scribed above.The genomic DNA library in X phage Charon 35 was plated

using E. coli KH802 as host on 150 mm Petri dishes of LBM.After alkaline denaturation, neutralization, and incubationfor 2 h at 80°C under vacuum, the filters were prehybridizedin 5x SSPE, 5x Denhardt's reagent (10), 0.1% SDS, 50%formamide, and 100 ,ug/mL sheared, denatured herring testesDNA for 4 h at 42°C. The cDNA used as a probe was excisedfrom Seaplaque low melt agarose gel (FMC BioProducts),melted, and labeled using random hexamers (1 1). The probe,with specific activity of 1 x 109 cpm/gg insert, was added tothe prehybridization solution at 1 x 106 cpm/mL. After 16 hof hybridization at 42°C, the blots were washed (final wash,0.5x SSPE, 0.1% SDS at 58°C). The filters were exposed toKodak XAR-2 film with intensifying screen at -80°C for 14

A 1 2 3kd

° mi.vW4~~~~~~~~~~~~~~~~ 3.

25. 7_ _ , ..

18 . 4 _b

14 3.4-peptide G

_ .. 4-peptide H

B N-terminus DELKXTTWXFVVETNNL

peptide BB EIXXVSXEAGEXA

peptide CC NAGFHDXXKLILXG

peptide G SVDLGXXG

peptide H TATTYKSERRNAMVEEGFRIVG

A T T Y K S E R R N A M V E5-ACT GCT ACT ACT TAT AAA TCT GAA AGA AGA AAT GCT ATG GTT GAN (

G G

h. Putative positives were purified by repeating the aboveprocedure for two addition rounds using the same probe.The tomato cDNA library constructed in plasmid pARC7

and propagated in E. coli DH5a was spread onto 150 mmPetri dishes of LBM with 100 jig/mL ampicillin. Plates wereincubated at 370C for 5 h, then colonies were lifted ontoNytran membranes, which were placed colony-side up ontoLBM + ampicillin plates and incubated for 15 h at 370C. Asecond membrane was wetted in 2x SSPE, blotted dry, andplaced onto the first membrane. Both membranes were placedbetween two pieces ofWhatman 3MM filter paper and pressedbetween two glass plates. The filter paper was sprinkled withwater and, together with the membranes, placed in foil andautoclaved for 5 min, baked 30 min, and washed twice in 2xSSPE, 1% SDS. Colonies were screened using duplicate mem-branes to reduce false positives. The membranes were prehy-bridized, hybridized, and washed to the same final stringencyas discussed for the genomic library screen.

Subcloning and Sequencing

Enzymes for subcloning and sequencing were purchasedfrom Promega unless otherwise indicated. Nucleotides andrandom hexamer primers were purchased from Pharmacia.Radioisotopes were purchased from DuPont-NEN ([y32p]dATP and [a35S]dATP) and ICN ([a32P]dCTP). X recombi-nant DNA was isolated by the procedure ofVerma (28) exceptthat a liquid lysate was used in place of a plate lysate. Frag-ments to be subcloned were electroeluted from agarose gelsusing a Elutrap (Schleicher & Schuell, Inc.) and precipitatedwith ethanol. The fragments were inserted into the bacterio-phage M13 vectors mpl8 and mpl9 that had been digestedwith the appropriate restriction enzyme and dephosphoryl-ated (30). Templates for sequencing were prepared as de-scribed by Carlson and Messing (7). Universal primers andoligodeoxyribonucleotide primers, ranging between 17 and

Figure 1. Apase-1 protein and peptides derivedfrom it. A, An 18% polyacrylamide SDS gel wasstained for protein using a silver stain (Sigma).Mol wt standards (Bethesda Research Labora-tories), purified Apase-1, and purified Apase-1digested with endoproteinase Lys-C were ap-plied to lanes 1, 2, and 3, respectively. The Lys-C digestion products, peptides G and H, areindicated on the right. B, Amino-terminal aminoacid sequence of purified Apase-1 (N-terminus)and of peptides obtained after cleavage withcyanogen bromide (peptides BB and CC) orendoproteinase Lys-C (peptides G and H) arepresented. Single letters designate amino acidsidentified, except that X designates amino acidsthat could not be assigned unambiguously. C,Sequence of oligonucleotide probe AP53 derivedfrom peptide H sequence. This probe is eightfolddegenerate due to the incorporation of mixturesof two nucleotides at the three positionsindicated.

E G FGAA GGT TT-3'

A

141

WILLIAMSON AND COLWELL

20 nucleotides in length, were generated by K. Shaw (Depart-ment of Medical Pathology, University of California, Davis).The dideoxy chain termination sequencing method (25) wasemployed using [a 35S]dATP and a Sequenase Kit (U.S. Bio-chemical). Termination products were resolved on a buffergradient gel (5).DNA sequences were analyzed using the software package

from the Genetics Computer Group (GCG) of Madison, WI,on a VAX/VMS mainframe.

Preparation of Genomic DNA and Southern Blotting

For RFLP analysis, tomato DNA was extracted from leavesas described by Bernatzky and Tanksley (4). DNA was di-gested to completion with restriction enzymes as recom-mended by the manufacturer (Promega Corp., Madison, WI).About 15 ,gg ofdigested DNA was separated at 35 V overnighton a 0.8% agarose gel. Southern transfers used Nytran mem-branes as recommended by the supplier (Schleicher & Schuell,Inc.). Hybridization conditions were as described in Bernatzky(3). cDNA inserts were labeled by the random hexamer label-ing method (1 1). Approximate sizes of hybridizing fragmentswere determined by comparison to X DNA digested withEcoRI and/or HindIll.

RESULTS

Protein Analysis

Acid phosphatase- 1 was purified from cell suspension cul-ture of VFNT tomato cells as described in "Materials andMethods." VFNT tomato contains the Aps-] allele of Aps-].Cellulose acetate electrophoresis was used to assay fractionsfor Apase-1 activity and to distinguish this isozyme fromothers present in tomato (23). Our analysis revealed thatApase- 1 is a minor constituent of the total acid phosphataseactivity present in tomato. We found that Apase-I activitywas released from suspension culture cells after incubation inan extraction buffer containing sodium acetate and Triton X-100, eliminating the need for cell disruption. In this and otherregards, the purification procedure reported here differs sub-stantially from that previously reported (23) and includes,after the chromatographic separations, a preparative isoelec-tric focusing step. Apase- activity was recovered at pH 4.5to 4.8, indicating that the protein was acidic. The purifiedprotein was present as a single, major band after SDS-PAGE(Fig. IA). The procedure resulted in 300 ,g of purified Apase-1, corresponding to a yield of 50 ,g/L of suspension culture.Attempts to determine the amino-terminal sequence of

purified Apase- resulted in the peptide sequence shown inFigure B. The amino acid yields from the sequence analysiswere approximately 5% of the levels expected, suggesting thatthe amino-terminus ofmost molecules ofApase-l is modified.Coomassie brilliant blue-stained bands representing the Lys-C digestion products, peptides G and H (Fig. IA), and cyan-ogen bromide cleavage products, peptides BB and CC, gavethe other amino acid sequences shown in Figure 1 B.

Isolation of cDNA and Genomic DNA Clones

Long, synthetic oligodeoxyribonucleotide probes based on

preferred codons for each amino acid in the target peptide

sequence have been used successfully to clone a number ofgenes from complex genomes (29). We used this strategy toscreen available cDNA libraries for a clone corresponding toAps-]. The peptide H sequence was used to derive the eight-fold degenerate 53-deoxyribonucleotide probe (AP53) shownin Figure IC. The sequence of this probe was designed byselecting codons found to be most abundant in two tomatogenes, chalcone synthase (22) and polygalacturonase (9). Thisprobe was used to screen a cDNA library of VFNT cherrytomato DNA in XgtIO. From a screen of approximately 3.5x 105 plaques, 21 clones were obtained that hybridized to the

R BEr .oAA:"'

._g,.... Hi

£_ ..J , w.

..;4 q -S R ......... A.

q..>

Boss

Z._._Wr' +:(- ..

X

> I:

iJ} 4

O f ..

*ADs aile'e?

Figure 2. RFLP in tomato genomic DNA correlates with the Aps-1allele using probe pAA2. Genomic DNA isolated from tomato wasdigested with EcoRI and analyzed by Southern blot hybridizationusing 32P-labeled pAA2 insert as a probe. Lane 1 contains DNA fromVFNT cherry tomato (Aps-1'); lane 2 contains DNA from Rossol (Aps-1+); lane 3, from Castlerock (Aps-1+); and lane 4 from Sun6082 (Aps-11). Sizes of mol wt markers are indicated on the left. Polymorphicbands in each lane are indicated by arrows and estimated sizes onthe right. The Apase-1 isozyme present in each plant is indicated atthe bottom of the figure. Hybridization conditions were moderatelystringent (final wash at 580C in 0.5 x SSPE).

142 Plant Physiol. Vol. 97, 1991

2? 3.

.I.-

*a',: Im "I1.40

TOMATO ACID PHOSPHATASE-1 GENE

probe. These clones were divided into six groups after restric-tion digestion and cross-hybridization analysis.

Because Aps-l' is derived from the wild species, L. peru-viantum, it was likely that the DNA fragment containing Aps-1 would display an RFLP when compared with that contain-ing Aps-J+, the common tomato allele ofAps-1. DNA from arepresentative of each of the six groups of candidate cloneswas used to probe genomic blots of tomato DNA from fourtomato varieties that differ in the Aps-] allele present. Of thesix clones, only one, pAA2, containing an 0.5 kb insert,revealed an RFLP in these varieties. That RFLP correlatedwith the Aps-] allele (Fig. 2). Four major bands were seen ineach lane of the EcoRI-digested DNA shown in Figure 2 andone of these bands was dimorphic (a 3.0 kb band was seen invarieties with the Aps-l' allele and a 5.0 kb band was seen invarieties with the Aps-l+ allele). The most intense band onthe Southern blot (approximately 3.5 kb in size) is not poly-morphic, suggesting that pAA2 does not correspond to Aps-1, but has a sequence similar to that ofAps-]. This suggestionwas confirmed by sequence analysis (see below).pAA2 was used to probe a genomic DNA library ofVFNT

cherry tomato in X phage Charon 35. From a screen ofapproximately 4.2 x 105 plaques, five clones that hybridizedto pAA2 were obtained. DNA from two of these clones, whendigested with EcoRI, displayed a restriction fragment of 3.0kb which hybridized to pAA2. This fragment correspondedin size to the Aps-l'-specific polymorphic band in the South-ern blot of VFNT cherry tomato DNA. The restriction frag-ment map of the 15.5 kb insert in one of these clones, pGAP,is shown in Figure 3. The other clone was similar in size buthad slightly different end points (not shown).pAA2 was also used to probe a plasmid cDNA library in

pARC7 constructed from hypocotyl mRNA. Of approxi-mately 3 x 105 total plaques screened, 23 hybridized to pAA2.Six colonies were isolated, and one, containing a plasmiddesignated pAp la, hybridized intensely to a Sau3A fragmentofpGAP under stringent conditions. The cDNA insert in thisclone was estimated to be 1.2 kb, long enough to encode aprotein the size of the mature Aps-]. Another cDNA clone,designated pAplb, was obtained by rescreening the pARC7

pAA2

library with the pAp 1 a insert. pAp lb contained an insert ofapproximately the same length as that in pAp 1 a but lackedthe BamHI restriction site of the pApla insert (see Fig. 3).

DNA Sequence Analysis

The nucleotide sequences of the inserts in pAA2, pAp 1 a,and pAp lb were determined. The entire sequence from bothstrands was determined, except for the 40 nucleotides preced-ing the polyadenylate tail of pAp 1 a and pAp lb where se-quence was obtained on one strand only. The entire nucleo-tide sequence of pAp lb (Fig. 4), excluding the polyadenylatetail, was 1072 nucleotides long and contained an open readingframe of 255 amino acids extending from the first ATG. Thederived amino acid sequence from this open reading frameincludes a sequence identical to that of peptide H, as well assequences of the other peptides in Figure 1B, strongly sup-porting the identity of pAp lb as Aps- ' cDNA.The sequence of the insert in pAA2 was 469 nucleotides

long and the open reading frame showed a high degree ofsimilarity to that of Aps-1. However, the derived amino acidsequence was only 68% identical to peptide H, indicating thatthe pAA2 insert did not encode Apase- 1, but rather a relatedprotein. The short length of the pAA2 insert and the lack ofan obvious initiation site near the 5' end of the cDNAindicated that pAA2 was not a full-length cDNA clone.The cDNA clone pAp 1 a encoded 124 nucleotides ofDNA

excluding the polyadenylate tail. The 555 nucleotides of thissequence preceding the polyadenylate tail were identical tothose ofpAp lb. Upstream from this region, the two sequencesdiverged completely. Analysis of the pAp 1 a sequence showedno open reading frame extending through the region upstreamofthis point ofdivergence in any ofthe three reading registers.The lack of an open reading frame suggested that this cDNAclone might be a cloning artifact, or, alternatively, mightcontain an unspliced intron. To clarify its origin, a 3.2 kbSacl fragment of the genomic clone pGAP containing ho-mology to pAp 1 a (see Fig. 3) was subcloned into M13 andpartially sequenced using primers that had been synthesizedfor the cDNA sequencing. The genomic fragment contained

AAAA

0.1 kbI4

pGAP

0.1 kbi

Figure 3. Inserts from Aps- 1 genomic and cDNAclones. The three cDNA clones analyzed here,pAA2, pApl a, and pApl b, are represented asthick horizontal bars with the position of thepolyadenylate tail at the 3' end of the transcriptindicated to show orientation. The insert fromthe genomic DNA clone pGAP is shown with ahorizontal arrow indicating the position of theAps-1 transcript. Approximate positions of in-trons and 11 are indicated by hatched bars.Dotted lines connect matching sequences on thecDNA and genomic clones. In the cDNA pApl a,intron 11 has been spliced correctly, but intronis still present. pApl b represents a full-length, ornearly full-length, cDNA clone of Aps-1. Restric-tion enzyme cleavage sites are indicated as fol-lows: B, BamHI; E, EcoRI; H, HindIlIl; S, Sacd. PLindicates the position of the polylinker of Charon35.

143

WILLIAMSON AND COLWELL

AGAATTAAGTAATCCA

CCTACTCTCAACACAATTTCCTCTCTTATACAACTATATACTATACAATACGTATTCTTT

ATCTCTGGTATTTCATTGACTCCATTACTTCCCTATTATCATTCTTGAAGAATTTACTAC

ATGAGGATTTTCGTGTTCTTGGTGTTGTTGACTGTTGCAATCGGAACTGAAAATCTCAATM R I F V F L V L L T V A I G T E N L N

TCTCATGTGTTTCCAAGGCCATTGATTATTGAGTATCCTGAAAAACAATTGAGGGATGAGS H V F P R P L I I E Y P E K Q L R ID E

TTGAAGTGTACGACTTGGAGGTTTGTTGTTGAAACGAATAATTTAAGTCCATGGAAGACGL K C T T W R F V V E T N N L S P W K T

ATTCCAGAGGAATGTGCTGATTATGTCAAGGAATATATGGTGGGTCCAGGTTATAAGATGI P E E C A D Y V K E Y M V G P G Y K M

GAGATTGATAGGGTTTCGGATGAGGCAGGAGAATATGCCAAAAGTGTTGATTTGGGAGATE I D R V S D E A G E Y A K S V D L G D

GATGGAAGAGATGTGTGGATTTTTGATGTTGACGAAACTTTGCTTTCTAATCTTCCTTATD G R D V W I F D V D E T L L S N L P Y

TATTCTGATCATCGTTATGYATTGGAGGTATTTGATGATGTGGAATTTGATAAATGGGTTY S D H R Y G L E V F D D V E F D K W V

GAGAATGGAACGGCGCCAGCCTTGGGGTCCAGCTTGAAGCTTTATCAAGAAGTTCTGAAAE N G I A P A L G S S L K L Y Q E V L K

CTGGGATTCAAAGTTTTCTTGCTGACTGGGCGCAGTGAAAGACACAGAAGTGTTACTGTGL G F K V F L L T G R S E R H R S V T V

IIGAGAATTTGATGAATGCTGGATTCCACGATTGGCACAAGCTCATTCTGAGAGGCTCGGACE N L M N A G F H D W H K L I L R G S D

GACCATGGCAAAACAGCAACAACCTATAAATCAGAGAGACGAAATGCGATGGTAGAAGAAD H G K T A T T Y K S E R R N A M V E E

GGTTTCCGCATAGTGGGCAACTCAGGAGACCAGTGGAGTGATCTGCTAGGCTCCTCTATGG F R I V G N S G D O W S D L L G S S M

TCTTATCGCTCATTCAAGCTTCCAAACCCGATGTATTACATTCTTTAAAGTAACTAATAGS Y R S F K L P N P M Y Y I L

GTTTGGTAGTCCATGTTGATGCAACATGCCAATGATTATTTCTTTTCACTGTCTAATGGC

16 digested with EcoRI. In plants homozygous for the Aps-l'76 allele, a 3.0 kb band was present. This is the expected size for

136 an EcoRI fragment containing the 5' end of Aps-] (see mapof pGAP in Fig. 3). From the pGAP map, we predict that a

196 second band of over 5 kb containing the 3' end of Aps-l'should also be present in lane 1. This is consistent with the

256 16 kb fragment observed. A 5.0 kb band was present only inplants with the Aps-l+ allele and presumably contained the316 5' end of Aps-l+. The 16 kb fragment containing the 3' end

was not polymorphic in this population. The segregationpattern shows that the probe pAp 1 a hybridizes only to a single

436 locus corresponding to Aps-J under high stringency condi-tions. Twenty additional plants (for a total of 32) were ana-

496 lyzed by Southern blot analysis and, again, the presence ofthe 3.0 kb band, the 5.0 kb band, or both, corresponded for

556 every plant tested to the Aps-] 'lAps-i', Aps-J+/Aps-l+, or theheterozygous genotypes, respectively (not shown).

616

676

DISCUSSION

Evidence for the Isolation of Aps-1

736 We have obtained and analyzed the nucleotide sequence ofthe cDNA clones pApla, pAp1b, and pAA2. Our results

796 indicate that pApla, pAplb, and a genomic clone, pGAP,correspond to the Aps-J locus. The pAA2 cDNA insert ap-

856 pears to encode a similar, but distinct protein. Two lines ofevidence support the identity of pAplb as an intact, full-

916

97 6 ..

TACCTTATAAACAATGGAACTGTAATTCTTGTAATCCCTTAAACTGGATCATTGATGATT 1036

CAATTAATGTATGCCGTATCAGCAACAATTGTGCAT 1072

Figure 4. Nucleotide sequence of cDNA insert pAp1 b, encoding Aps-1. The entire sequence of the sense strand of pAp1 b and the aminoacid sequence of the open reading frame corresponding to Apase-1are presented. Positions of the Apase-1 peptide sequences obtainedin this work are indicated by horizontal bars under the derivedsequence. The position of the amino-terminus of the purified proteinis indicated by a vertical, cross-hatched bar. Positions of intronsand 11 are indicated by labeled arrows.

4V

40-

.4*

a contiguous region matching exactly in sequence the 5'region ofpAp I a that diverged from the pAp lb sequence. Thisresult supports the hypothesis that pAp la contains an un-spliced intron. The 3' splice consensus dinucleotide AG oc-curs at the appropriate position within the intron (not shown).DNA sequence analysis of the SacI genomic fragment alsoserved to localize the position of another intron in Aps-l asshown in Figures 3 and 4. We have not eliminated thepossibility that another intron exists 5' to the SacI fragment.

Apase-I Isoforms Correlate with a pApla-Detected RFLP

We tested whether our clones corresponded to the Aps-]locus by using pAp la to probe a Southern blot ofDNA of 12different tomato plants from an F2 population segregating forthe Aps-J alleles Aps-J' and Aps-J+ (Fig. 5). Two hybridizingbands were seen in each lane when the genomic DNA was

Figure 5. Co-segregation of RFLPs with Aps-1 allele using probepApia. DNA was isolated from each of 12 tomato plants from apopulation segregating for the alleles Aps-11 and Aps-1+. The Aps-1allele of each plant was determined by cellulose acetate electropho-resis and is indicated for each plant by 1/1 for plants homozygousfor Aps- 11 +/+ for those homozygous for Aps- 1+, and 1/+ forheterozygous plants. DNA was digested with EcoRI and probed withthe insert of pAp1a. The 3.0 kb band containing the 5' end and the16 kb fragment containing the 3' end of Aps-1' are indicated on theright, as is the 5.0 kb band presumably containing the 5' end of Aps-1+. Hybridization is specific to Aps-1 under the high stringencyconditions used for this blot.

Plant Physiol. Vol. 97, 1991144

TOMATO ACID PHOSPHATASE-1 GENE

length clone ofAps-1. There is complete identity ofthe aminoacid sequences determined from the Apase- 1 peptides withthe amino acid sequence deduced from clone pAplb. Inaddition, the RFLP pattern obtained using pApla as probecorrelates exactly with the Aps-l allele present in 32 differentplants, providing strong genetic evidence that pApla wasderived from a transcript of Aps-].The 1072 nucleotide sequence of the insert of pAplb

(shown in Fig. 4) appears to represent most or all of the Aps-1 transcript. The first ATG, at nucleotide 137, is preceded bya stop codon in the same reading frame, implying that theentire coding region is represented. The sequence correspond-ing to the amino-terminus of the purified protein was 39amino acids downstream from the first methionine, indicatingthat Apase- 1 is processed. The processed amino-terminus,particularly the first 15 amino acid residues, is rich in hydro-phobic residues as expected for a signal sequence. The calcu-lated mol wt of the mature protein is 24,981.

Homology and Function

A computer search of amino acid sequence banks revealedno similarity of the Apase- 1 sequence to any other Apase inthe GCG protein data bases searched, even though thesedatabases include acid phosphatase sequences from eight or-ganisms, including animals, bacteria, and fungi. No otherplant Apase sequences are present in the GCG database.Surprisingly, the sequence most similar to that encoded byAps-l was a 28 kD soybean protein (18). Alignment of thetwo sequences reveals a 44% identity in amino acid residuesbetween the mature forms of the two proteins. The proteinsare nearly identical in size of precursor and mature forms.The S28K protein and related proteins in soybean are abun-dant in soluble extracts of soybean seedling stems and accu-mulate to high levels (5% of soluble protein) in leaves (18,27). In contrast, based on their low representation in cDNAlibraries, neither the Aps-] nor the AA2 sequence is highlyexpressed in tomato. The S28K protein and Apase-1 differconsiderably in observed and deduced ionic charge: S28K isa basic protein (isoelectric point 8.6 [18]); Apase-1 is acidic(calculated isoelectric point is 4.7 for the mature protein).Although S28K and Apase- 1 are related in sequence, it is notknown whether they have any common function. It would beof interest to test the soybean protein for phosphatase activity.The presence of a signal peptide and our ability to extract

Apase- 1 activity from suspension cells by bnref treatment withdetergent and salt suggest that this enzyme may be associatedwith the cell wall, as is a portion of the S28K protein (18).The Aps-] clone will expedite further studies on tissue andsubcellular localization of this protein. The cDNA clonepApla appears to encode a partially processed Aps-] RNA.This clone may represent a rare, anomolous event, or, moreinterestingly, may reflect the involvement of RNA processingin regulation of expression of Aps- 1.The Southern blots show simple segregation of a polymor-

phic band with the corresponding Aps-] alleles, indicatingthat Aps-J is a single copy gene. When pAA2, apparentlyencoding the partial sequence of a gene similar to, but distinctfrom, Aps-1, was used to probe tomato DNA at moderatestringency, two to four bands were evident, depending on the

restriction enzyme used (Fig. 2), indicating that there are twoor three related sequences in tomato. The pAA2 and Aps-]cDNAs are 83% identical in nucleotide sequence within thecoding region but divergent in the 3' untranslated region. Thehigh similarity (85% identity of predicted amino acid se-quence) between Apase- 1 and the partial sequence of AA2suggests that AA2 may encode another acid phosphatase.

Walking to Mi

A major incentive for cloning Aps-l was its proximity tothe nematode resistance gene Mi. One cloning strategy, chro-mosome walking, requires DNA markers located very closeto the gene of interest. The 15 kb insert in the genomic clonepGAP that contains Aps-] provides a starting point for chro-mosome walking to Mi.

ACKNOWLEDGMENTS

We thank W. Gruissem, R. Fischer, and S. O'Neill for generouslyproviding tomato libraries, J.Y. Ho and H.M. Ma for providingtomato DNA, and J. Gardner for help with protein sequence inter-pretation. We are grateful to J. Harada, F. Ho, J. Frazer, H. Ma. andG. Bruening for helpful comments on the manuscript.

LITERATURE CITED

1. Baker JE, Takeo T (1973) Acid phosphatase in plant tissue 1.Changes in activity and multiple forms in tea leaves and tomatofruit during maturation and senescence. Plant Cell Physiol 14:459-471

2. Benton W'D, Davis RW (1977) Screening lambda-gt recombinantclones by hybridization to single plaques in situ. Science 196:180-182

3. Bernatzky R (1988) Restriction fragment length polymorphism.In SB Gelvin, RA Shilperoort, DPS Verma, eds, Plant Molec-ular Biology Manual Cl. Kluwer Academic Publishers, Dor-drecht. The Netherlands, pp 1- 17

4. Bernatzky R, Tanksley SD (1986) Toward a saturated linkagemap in tomato based on isozymes and random cDNA se-quences. Genetics 112: 887-898

5. Biggin NID, Gibson TJ, Hong GF (1983) Buffer gradient gels and-3'S label as an aid to rapid DNA sequence determination. ProcNatl Acad Sci USA 80: 3963-3965

6. Bolkan HA, Williamson VM, W'aters CM (1983) Use of celluloseacetate electrophoresis as an alternative to starch gel electro-phoresis for detecting root-knot nematode resistance in tomato.Plant Dis 71: 100 1-1003

7. Carlson J, Messing J (1984) Efficiency in cloning and sequencingusing the single-stranded bacteriophage M 13. J Biotechnol 1:253-264

8. Ching TM, Lin T-P, Metzinger RJ (1987) Purification andproperties of acid phosphatase from plump and shriveled seedsof triticale. Plant Physiol 84: 789-795

9. DelaPenna D, Alexander DC, Bennett AB (1986) Molecularcloning of tomato fruit polvgalacturonase: analysis of polyga-lacturonase mRNA levels during ripening. Proc Natl Acad SciUSA 83: 6420-6424

10. Denhardt DT (1966) A membrane-filter technique for the detec-tion of complementary DNA. Biochem Biophvs Res Commun23: 641-646

11. Feinberg AP, Vogelstein B (1983) A technique for radiolabelingDNA restriction fragments to high specific activity. AnalBiochem 132: 6-13

12. Giordani R, Nari J, Noat G, Sauve P (1986) Purification andproperties of an acid phosphatase from .wsleptals cCrassbail-tlatex. Plant Sci 43: 207-212

13. Goldstein AH, Baertlein DA, McDaniel RG (1988) Phosphate

145

WILLIAMSON AND COLWELL

starvation inducible metabolism in Lveopersicon escilentnun.I. Excretion of acid phosphatase by tomato plants and suspen-

sion-cultured cells. Plant Physiol 87: 711-71514. Gross E (1967) The cyanogen bromide reaction. In CHW Hirs.

ed, Methods in Enzymology, Vol XI. Academic Press, NewYork, pp 238-255

15. Helentjaris T, Slocum M, Wright S, Schaefer A, Nienhuis J(1986) Construction of genetic linkage maps in maize andtomato using restriction fragment length polymorphisms.Theor Appl Genet 72: 761-769

16. Kruzel M, Morawiecka B (1972) Acid phosphatase of potatotubers (Sol/nutm tuberostum L.). Purification, properties, sugarand amino acid composition. Acta Biochim Pol 29: 321-329

17. Laemmli UK (1970) Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 227: 680-685

18. Mason HS, Guerrero FD, Boyer JS, Mullet JE (1988) Proteinshomologous to leaf glycoproteins are abundant in stems ofdark-grown soybean seedlings. Analysis of proteins andcDNAs. Plant Mol Biol 11: 845-856

19. Matsudaira P (1987) Sequence from picomole quantities ofproteins electroblotted onto polyvinylidene difluoride mem-

branes. J Biol Chem 262: 10035-1003820. Maxam A, Gilbert W (1977) A new method for sequencing

DNA. Proc Natl Acad Sci USA 74: 56021. Medina-Filho HP (1980) An electrophoretic variant as a tool for

breeding tomatoes for nematode resistance. PhD thesis. Uni-versity of California, Davis

22. O'Neill SD, Tong Y, Spoerlein B, Forkmann G, Yoder JI (1990)Molecular genetic analysis of chalcone synthase in Lvcopersi-con escldentum and an anthocyanin deficient mutant. MolGen Genet 224: 279-288

23. Paul EM, Williamson VM (1987) Purification and properties ofacid phosphatase- 1 from a nematode resistant tomato cultivar.Plant Physiol 84: 399-403

24. Rick CM, Fobes J (1974) Association of an allozyme withnematode resistance. Tomato Gen Cooperative Rep 24: 25

25. Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing withchain-terminating inhibitors. Proc Natl Acad Sci USA 74:5463-5467

26. Smith PG (1944) Embryo culture of a tomato species hybrid.Proc Am Soc Hort Sci 44: 413-416

27. Staswick PE (1988) Soybean vegetative storage protein structureand gene expression. Plant Physiol 87: 250-254

28. Verma M (1989) Use of ammonium sulfate precipitation andguanidine isothiocyanate lysis to isolate lambda DNA. Bio-techniques 7: 230-232

29. Wood WI (1987) Gene cloning based on long oligonucleotideprobes. In SL Berger, AR Kimmel, eds, Methods in Enzymol-ogy, Vol 152. Academic Press, New York, pp 443-447

30. Yanisch-Perron C, Vieira J, Messing J (1985) Improved phagecloning vectors and host strains: nucleotide sequences of theM 13mp 18 and pUC 19 vectors. Gene 33: 103-119

Plant Physiol. Vol. 97, 1991146