INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Leucyl aminopeptidases (LAPs; EC 3.4.11.1) are members of the M17 family of peptidases (Barrett et al., 1998). LAPs are ubiquitous being found in animals, plants, and prokaryotic cells. These hexameric metallopeptidases catalyze the release of the N-terminal residues from protein, peptide, fluorometric, and chromogenic substrates. The best characterized LAPs are from Bos taurus, Escherichia coli and tomato (Lycopersicon esculentum). X-ray crystal structures of the bovine and E. coli LAPs have provided insight into the LAP catalytic mechanism (Kim and Lipscomb, 1994; Sträter and Lipscomb, 1995; Sträter et al., 1999a). The roles of selected residues of the E. coli and tomato LAPs in chromogenic or peptide substrate catalysis, respectively, have been tested by site-directed mutagenesis (Sträter et al., 1999b; Gu and Walling, 2002).

In most plants, three classes of LAP-related polypeptides are detected using a tomato LAP antiserum, including the 66- and 77-kD LAP-like proteins and the 55-kD neutral LAP (LAP-N; Chao et al., 2000). Only in a subset of the Solanaceae is a second 55-kD LAP species (LAP-A) detected (Hildmann et al., 1992; Gu et al., 1996b; Chao et al., 2000). In tomato, LAP-A protomers have an acidic pI and are encoded by two genes (LapA1 and LapA2), which are expressed during floral and fruit development. The LapA genes are not expressed in foliage from healthy plants (Chao et al., 1999). However, LapA RNAs, proteins, and activities accumulate locally and systemically in leaves after wounding, Pseudomonas syringae pv. tomato and Phytophthora parasitica infection, and caterpillar feeding (Pautot et al., 1993, 2001; Gu et al., 1996b; Chao et al., 1999; Jwa and Walling, 2001). The activation of LapA gene expression by jasmonic acid (JA), abscisic acid, the phytotoxin coronatine (a JA mimic), and suppression of LapA by salicylic acid is consistent with the regulation of the tomato LapA genes by the wound octadecanoid pathway (Chao et al., 1999). LapA genes also respond to signals generated during water deficit and salinity stress (Chao et al., 1999). The potato (Solanum tuberosum) Lap RNAs also accumulate after wounding and exogenous abscisic acid and JA, but increases were not observed after water-deficit stress (Hildmann et al., 1992).

Like the eukaryotic and prokaryotic LAPs, the wound-induced LAP-A of tomato is a homo-hexamer (Gu et al., 1996b; Gu and Walling, 2000). The tomato LAP-A enzyme preferentially hydrolyzes substrates with N-terminal Leu, Arg, and Met and does not cleave substrates with N-terminal Asp, Glu, or Gly residues (Gu et al., 1999; Gu and Walling, 2000). Although LAP-A is similar to the homo-hexameric porcine LAP and E. coli PepA (LAP), differences in substrate specificity have been noted.

In contrast to the wound-induced LAP-A, there is a limited knowledge about the 66- and 77-kD LAP-like and 55-kD LAP-N proteins of tomato. The levels of these proteins are not modulated by defense signals in leaves (Gu et al., 1996b; Chao et al., 1999). Furthermore, these proteins are detected in both dicots and monocots (Chao et al., 2000). For example, in Arabidopsis, LAP proteins accumulate to similar levels in each vegetative and reproductive organ and do not increase in response to phytohormone or stress treatments (Bartling and Nosek, 1994). There is also biochemical evidence for multimeric LAP enzymatic activities in germinated barley (Hordeum vulgare) seeds (green malt) and in resting kidney bean (Phaseolus vulgaris) cotyledons (Kolehmainen and Mikola, 1971; Sopanen and Mikola, 1975; Mikkonen, 1992).

The biological roles for the plant, animal, and prokaryotic LAPs are not completely understood and they may be complex and species specific. For example, the E. coli LAP, also known as XerB, PepA and CarP, appears to be multifunctional. The E. coli LAP serves as an aminopeptidase (Vogt, 1970) and a DNA-binding protein that mediates both site-specific recombination at the cer site of ColE1 plasmids (Stirling et al., 1989) and transcriptional activation of the carAB operon (Charlier et al., 2000). The DNA-binding capabilities of the E. coli LAP are independent of aminopeptidase function (McCulloch et al., 1994; Charlier et al., 2000).

Substantially less is known about the role of the eukaryotic LAPs. Increases in LAP protein levels were detected during meiosis in meiocytes and their surrounding cells using an immunohistochemical assay in the basidiomycete Coprinus cinereus (Ishizaki et al., 2002); however, the exact role LAP plays in meiosis is not understood. Given the decreases in LAP activity that accompany lens aging, a role for LAP in cataract development has been proposed (Taylor, 1985; Sharma et al., 1996). In addition, the human LAP is induced by -interferon (Harris et al., 1992) and has been implicated in the processing of peptides released from the proteasome; these peptides are subsequently used for antigen presentation in the MHC I complex (Beninga et al., 1998).

The roles of plant LAP-N and LAP-A may be different given the differences in their distribution in the plant kingdom and responses to stress. To begin to understand the importance of the tomato LAP-N in plant growth and development, it was critical to isolate and characterize the tomato LapN gene product. Using a series of LAP-A domain-specific antisera, LAP-N and LAP-A were shown to be distinct protein species. The accumulation of LAP-N and LAP-A proteins in vegetative and reproductive organs was determined. LapN was a single-copy gene encoding a rare transcript, which contained two potential translational initiation codons that could give rise to a 60-kD preprotein or a 55-kD protein lacking targeting signals. Comparison of plant LAPs indicated that two classes of plant LAPs represented by LAP-N and LAP-A can be discerned by sequence identity and the presence of signature residues discriminating the LAP-N and LAP-A proteins. Evaluation of an overexpressed His₆-LAP-N enzyme showed that the His₆-LAP-N formed a multimeric complex with a biochemical properties and a substrate specificity distinct from the wound-induced LAP-A.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The LAP-N and LAP-A Proteins Are Diverged in their N-Terminal Domains

The LAP-A polyclonal antiserum detects four classes of proteins immunologically related to LAP-A (Gu et al., 1996b). The 55-kD LAP proteins with acidic pIs (LAP-As) accumulate in wounded leaves, whereas the 55-kD LAP proteins with neutral pIs (LAP-Ns) and the 66- and 77-kD LAP-like proteins are detected in both healthy and wounded tomato leaves. Comparisons of plant, animal, and microbe LAPs have shown that the COOH domain is highly conserved and harbors the two Zn²⁺-binding sites and catalytic domain of LAP, whereas N-terminal domains are diverged (Gu and Walling, 2002). Therefore, it was possible that antibodies recognizing the N-terminal domain of the LAP-A would discriminate the four classes of LAP-related polypeptides.

To test this hypothesis, antibodies recognizing different regions of the LAP-A protein were purified (Fig. 1A). Affinity-purified polyclonal antibodies to LAP-A domains A (residues 123-194), B (residues 195-233), D (residues 290-424), and F (residues 533-571) were incubated with 2D-PAGE immunoblots of proteins from wounded tomato leaves. For comparison, a 2D-PAGE immunoblot incubated with the LAP-A polyclonal antiserum is displayed (Fig. 1B). As previously observed, the polyclonal LAP-A antiserum recognized the 77- and 66-kD LAP-like proteins and the LAP-A and LAP-N proteins (Gu et al., 1996b). The affinity-purified LAP antibodies selected by the LAP-A domains B, D and F detected both the 55-kD LAP-A and LAP-N proteins (Fig. 1B). Only the domain A-specific antibodies were able to discriminate between the LAP-A and LAP-N proteins (Fig. 1B). The specificity of the domain A antibodies provided evidence that LAP-A presented antigens that distinguished the two different 55-kD LAP species.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Specificity of LAP-A domain-specific antibodies. A, Four GST-LapN fusion genes containing domains A, B, D, and F were constructed and overexpressed in E. coli. Black bars, LapA1-coding regions in the 1.9-kb full-length LapA1 cDNA clone pBLapA1 (Gu et al., 1996a), the partial LapA1 cDNA clone pDR57 corresponding to LapA1 nucleotides 325 to 1,843 (Pautot et al., 1993), and the GST-LapA fusion constructs. The glutathione-S-transferase (GST)-LAP-A fusion proteins contained the following amino acid residues: A (residues 123-194), B (residues 195-233), D (residues 290-424), and F (residues 534-571). White bars, LapA1 5'- and 3'-untranslated regions (UTRs). Gray bar, a small segment of the pGEX-3X polylinker was present in pDR57 and the GST-LAPA-F clone. Regions corresponding to each LapA1 subclone are described in detail in "Materials and Methods." B, Total proteins (100 µg) from wounded tomato leaves were fractionated by two-dimensional (2D)-PAGE, electroblotted, and incubated with a 1:500 (w/v) dilution of the LAP-A polyclonal antiserum or a 1:20 (w/v) dilution of affinity-purified LAP-A antibodies selected using the GST-LAPA-A, -B, -D, or -F fusion proteins (see "Materials and Methods"). LAP-A, LAP-N, and LAP-like proteins molecular masses (in kilodaltons) were determined by protein markers. The pH range (pH 8-4) of the isoelectric focusing (IEF) gels is indicated. LAP-A (), LAP-N (), and 66-kD () and 77-kD () LAP-like proteins are indicated.

LAP-N and LAP-A Protein Accumulation in Vegetative and Reproductive Organs

Although there are numerous papers monitoring changes in aminopeptidase activity during development (Walling and Gu, 1996), there are few studies examining aminopeptidase protein levels in plant development (Bartling and Nosek, 1994; Herbers et al., 1994; Hauser et al., 2001; Murphy et al., 2002). Therefore, it was of interest to more thoroughly examine the accumulation of the LAP-N, LAP-A and LAP-like proteins in vegetative and reproductive organs of tomato plants. Total proteins were extracted from leaves, stems, roots, and senescent leaves. These proteins were fractionated by 2D-PAGE, and protein blots were incubated with the LAP-A polyclonal antiserum (Fig. 2). The 66-kD LAP-like proteins and the 55-kD LAP-N polypeptides were detected at similar levels in stems, roots and leaves. In contrast, LAP-A and the 77-kD LAP-like proteins were not detected in stems or roots and were present at low levels in healthy and senescent leaves.

View larger version (60K): [in this window] [in a new window]   Figure 2.   Accumulation of LAP-A, LAP-N and LAP-like proteins in vegetative organs. Total proteins were isolated from tomato stems, roots, healthy leaves, and senescent leaves (see "Materials and Methods"). Total proteins (80 µg) were fractionated by 2D-PAGE, electroblotted, and incubated with a 1:500 (w/v) dilution of polyclonal LAP-A antiserum. LAP-A, LAP-N, and LAP-like proteins masses (in kilodaltons) are shown. The pH range (pH 8-4) of the IEF gels is indicated. LAP-A (), LAP-N (), and 66-kD () and 77-kD () LAP-like proteins are indicated.

Aminopeptidase activities in dormant seeds and cotyledons after germination vary within the plant kingdom (for review, see Walling and Gu, 1996). Aminopeptidase activities increase, decline or remain constant in cotyledons during the period of maximum storage protein mobilization. Hexameric LAP activities have been detected in barley cotyledons after germination and resting kidney bean seeds. If the tomato LAP or LAP-like proteins correspond to these hexameric enzymatic activities, LAP and LAP-like proteins should be detected in cotyledons after germination.

The levels of LAP-A, LAP-N and LAP-like proteins in cotyledons at four times after imbibition were monitored in 2D-PAGE immunoblots (Fig. 3). Small amounts of the LAP-N and 66-kD LAP-like proteins were detected in cotyledons 24 h after imbibition (Stage 1). Increases in both proteins were readily detected in Stage 2 (immediately after emergence from seed coats), Stage 3 (2 d after emergence), and Stage 4 (4 d after emergence) cotyledons. In contrast, the LAP-A polypeptides were below the level of detection in Stage 1 and Stage 2 cotyledons and were present at very low levels in Stage 3 and Stage 4 cotyledons (Fig. 3). The 77-kD LAP-like proteins were detected at low levels in Stage 2 through 4 cotyledons.

View larger version (70K): [in this window] [in a new window]   Figure 3.   Accumulation of LAP-A, LAP-N and LAP-like proteins in cotyledons. Total proteins were isolated from cotyledons at four times after germination (Stages 1-4; see "Materials and Methods"). Total proteins (80 µg) were fractionated by 2D-PAGE, electroblotted, and incubated with a 1:500 (w/v) dilution of polyclonal LAP-A antiserum. LAP-A, LAP-N, and LAP-like proteins masses (in kilodaltons) are shown. The pH range (pH 8-4) of the IEF gels is indicated. LAP-A (), LAP-N (), and 66-kD () and 77-kD () LAP-like proteins are indicated. These immunoblots required longer development times to visualize LAP and LAP-like proteins; therefore, the backgrounds in these blots were higher than in other figures.

To determine if the levels of the LAP and LAP-like proteins varied in the tomato reproductive organs, total proteins from four different stages of floral bud development and stamens, pistils, petals, and sepals from open flowers were analyzed in 2D-PAGE immunoblots (Fig. 4). LAP-N and LAP-like proteins accumulated to similar levels in developing floral buds and in all floral organs. In contrast, LAP-A proteins were abundant in 0.3-cm buds and increased markedly throughout bud maturation. LAP-A was present at extremely high levels in stamens, pistils, sepals, and petals of open flowers.

View larger version (36K): [in this window] [in a new window]   Figure 4.   Accumulation of LAP-A, LAP-N and LAP-like proteins in reproductive organs. Total proteins were isolated from tomato stamens, pistils, petals, and sepals from freshly open flowers and at four stages of floral bud development (0.3, 0.5, 0.7, and 1.0 cm in length). Total proteins (80 µg) were fractionated by 2D-PAGE, electroblotted, and incubated with a 1:500 (w/v) dilution of LAP polyclonal antiserum. LAP-A, LAP-N, and LAP-like polypeptide masses (in kilodaltons) and pH range of the IEF gels (pH 8-4) are shown. LAP-A (), LAP-N (), and 66-kD () and 77-kD () LAP-like proteins are indicated.

Isolation and Characterization of LapN

To identify a full-length LapN cDNA, an expression cDNA library was screened using the polyclonal LAP-A antiserum. LapN cDNA clones were identified from the pool of immunopositive clones based on their hybridization to a LapA probe spanning a LapA/LapN conserved domain (domain D) and lack of hybridization signal to a LapA probe derived from a more diverged region (domain A; Fig. 1A). The 1.9-kb LapN3 cDNA contained a single large open reading frame with 85% nucleotide sequence identity to the wound-induced LAP-A of tomato.

To estimate the number of the genes that encode LAP-N, genomic DNA blots were hybridized with a ³²P-labeled LapN cDNA probe (Fig. 5A). The LapN cDNA probe detected three EcoRI fragments in tomato genomic DNA. The 10- and 3-kb fragments represented the LapA1 and LapA2 genes, respectively (Gu et al., 1996a; Chao et al., 2000). The LapN gene was located on a 7-kb EcoRI fragment, and single-copy reconstructions indicated that LapN was present once in the haploid tomato genome.

View larger version (53K): [in this window] [in a new window]   Figure 5.   Genomic DNA-blot analysis and RNase protection studies. A, Genomic DNAs (10 µg) from the tomato cultivars Peto 238R (238R) and VFNT were digested with EcoRI, separated on a 0.7% (w/v) agarose gel and transferred to a nitrocellulose filter. The DNA blot was hybridized with a 32P-labeled pBLapN probe. The control (C) lane is a single-copy reconstruction with 68 pg of EcoRI-digested pBLapN. The sizes of DNA fragments were determined using the 1-kb DNA ladder (BRL) run in a parallel lane. B, Poly(A+) mRNAs were isolated from healthy and wounded tomato leaves. A 32P-labeled LapA1 3'-UTR antisense RNA probe was hybridized with 2 µg of poly(A+) mRNAs from healthy (lane 1) or wounded leaves (lane 2). A 32P-labeled LapN 3'-UTR antisense RNA probe was hybridized with 2, 5 and 10 µg (lanes 3-5, respectively) of poly(A+) mRNAs from healthy leaves. An M13mp18 DNA sequencing ladder (lanes A, G, C, and T) served as size markers. The sizes (nucleotides) of the major LapA1 and LapN protected fragments are indicated.

LapA and LapN transcripts were present at extremely low levels in healthy tomato leaves. LapN RNAs were not detected in RNA blots with 10 µg of total RNA or 2 µg of poly(A⁺) RNA (data not shown). Therefore, RNase protection studies using a ³²P-labeled LapN 3'-UTR antisense RNA probe and 2, 5 and 10 µg of poly(A⁺) RNA from healthy tomato leaves were performed (Fig. 5B). LapN transcripts were barely detected in 2 µg of poly(A⁺) mRNAs. Two clusters of the protected signals (154-157 and 165-166 nucleotides) were detected using 5 and 10 µg of poly(A⁺) mRNA (Fig. 5B, lanes 3-5). These data show that the most abundant LapN transcripts had 3'-UTRs that were approximately 145 and 156 nucleotides in length, which was 37 and 26 nucleotides shorter than the LapN cDNA clone that was sequenced. The relative abundance of each of the protected fragments in the clusters was similar. For comparison, RNase protection using a ³²P-labeled LapA1 3'-UTR RNA probe and 2 µg of poly(A⁺) RNAs from healthy or wounded leaves are displayed (Fig. 5B, lanes 1 and 2). The LapA riboprobe detected protected fragments of 140 and 146 nucleotides in the wounded leaf RNA samples, whereas LapA RNAs were not detected in healthy leaf RNA samples. These results are consistent with the previous report by Chao et al. (2000).

Structural Features That Distinguish LapN and LapA RNAs and Proteins

The sequence of the LapN cDNA predicted two potential translational initiation codons, which could generate a mature 55-kD LAP-N protein by different mechanisms. The translational initiation codon at nucleotide 31 was imbedded in a nucleotide context (uAuCAAUGGCc) with strong identity to the dicot translation codon consensus of aaA(A/C)aAUGGCu (Joshi et al., 1997). This deduced LapN translation product was 60 kD and had a pI of 7.8 (Fig. 6). This protein was similar in size (577 residues) and had 77% amino acid identity (87% similarity) with the wound-induced LAP-A1 and LAP-A2 preproteins of tomato (Gu et al., 1996a). Because the mature LAP-A and LAP-N proteins detected in 2D-PAGE immunoblots were 55 kD, this 5-kD N-terminal extension must be efficiently processed in vivo. Similar to the tomato LAP-A1, LAP-A2 and potato LAP presequences (Herbers et al., 1994; Gu et al., 1996a), the tomato LAP-N presequence contained several features similar to plastid transit peptides including a high percentage of Ser, Thr and positively charged residues and absence of acidic residues (for review, see De Boer and Weisbeek, 1991).

View larger version (133K): [in this window] [in a new window]   Figure 6.   Alignment of the deduced amino acid sequences of plant LAPs. The deduced amino acid sequence of the tomato LapN was aligned with LeLapA1, LeLapA2, and LeTPP24 from tomato (Milligan and Gasser, 1995; Gu et al., 1996a), StLAP from potato (Hildmann et al., 1992), AtLAP1 (Bartling and Weiler, 1992), AtLAP2 and AtLAP3 from Arabidopsis, OsLAP from rice (Oryza sativa), a partial PsLAP from parsley (Petroselinum crispum), and a partial BpLAP from white birch (Betula pendula; Valjakka et al., 1999). Amino acids identical to the tomato LAP-N polypeptide are shaded in gray. LAP-A signature residues are highlighted in yellow. Amino acid residues that are invariant in all 10 plant LAPs (except for one residue in AtLAP3 and two residues in LeTTP24), and the bovine LAP, E. coli PepA, Rickettsia prowazekii. PepA, human (Homo sapiens) LAP, and Salmonella typhimurium LT2. PepA are shaded in black with white letters. Amino acid residue numbers are indicated. Dots indicate gaps introduced to allow for the optimal alignment of the peptide sequences. Solid line, Pfam conserved domain Pfam00883; hyphenated line, Pfam conserved domain Pfam02789.

The transit peptide consensus motif of (Val/Ile)-X-(Ala/Cys)Ala that was described by Gavel and von Heijne (1990) and detected in the LAP-A proteins (Gu et al., 1996a) was not detected in the N-terminal region of LAP-N. Richter and Lamppa (1998) showed that the stromal endopeptidase that processes plant transit peptides primarily cleaves between Arg/Lys and Ala residues; this motif was not detected in LAP-A or LAP-N, although there was an abundance of basic residues in these proteins between residues 45 and 49 (Fig. 6). The more recently developed neural network program ChloroP predicts transit peptide cleavage sites and defined a consensus cleavage motif of Val-ArgAla-Ala-Ala-Val in plant plastid polypeptides (Emanuelsson et al., 1999). Although this motif was not detected in the LAP-N or LAP-A proteins, the ChloroP program predicted a cleavage site for LAP-N (cleavage site (CS) score = 5.581) and LAP-A (CS score = 5.365) at residues 42 (Pro-LeuCys-Ser-Arg-Arg) and 43 (Pro-LeuCys-Ser-Lys-Arg), respectively. This contrasts to the known N-terminal residues for LAP-A, which are Ile-54 or Gly-56 (Gu et al., 1996a). For comparison, the tomato Rubisco cleavage site had a CS score of 14.3 and its cleavage site (Val-ArgCys-Met-Gln-Val) was more consistent with the consensus. Finally, although HSP70-binding sites are detected in a majority of plastid transit peptides, these motifs were not evidenced in the tomato LAPs (Rial et al., 2000). Collectively, these data suggest that LAP proteins may have transit peptide-like characteristics, but motifs used for processing these putative transit peptides remained hard to define. Interestingly, cell fractionation studies indicate that the majority of LAP proteins are soluble and cytosolic (Gu et al., 1996a, 1996b).

A second in-frame initiation codon at nucleotide 151 (AAAgAAUGGCU), with an even stronger match to the dicot consensus, was also identified. If this initiation site was utilized, a 55-kD LAP-N protein with a pI of 6.29 would be synthesized. This protein was similar in size to the mature LAP-N proteins isolated from tomato leaves (Figs. 2-6; Gu et al., 1996b). It is possible that this initiation codon is utilized in vivo at some frequency. In vitro translations of LapN transcripts provide evidence that both translational initiation codons can be utilized in a cell-free system (C.J. Tu and L.L. Walling, unpublished data). It is possible that differential use of the two initiation codons in the LapN transcript and processing of the LAP-N presequences would determine the localization of LAP-N within the tomato cell. This is being tested using LAP-A- and LAP-N-specific antisera and immunolocalization.

Determination of the N terminus of the mature LAP-N proteins might resolve if one or both translational initiation codons were utilized in vivo. To this end, LAP-N polypeptides were purified from etiolated tomato seedlings using a four-step enrichment procedure. Enriched proteins were fractionated by 2D-PAGE, and three LAP-N polypeptides with small variations in mass and charge were identified by immunoblot analysis (data not shown). LAP-N proteins were excised and their N-terminal sequence determined by Edman degradation. Unfortunately, the N termini on all three LAP-N polypeptides were blocked.

There were numerous expressed sequence tag clones from both monocot and dicot species attesting to the ubiquity of the LAP enzymes in the plant kingdom (data not shown). Eleven plant LAPs were aligned to elucidate their relatedness to the tomato LAP-N (Fig. 6; Bartling and Weiler, 1992; Hildmann et al., 1992; Pautot et al., 1993; Milligan and Gasser, 1995; Gu et al., 1996a; Valjakka et al., 1999). The Conserved Domain Database (National Center for Biotechnology Information) identified two conserved Pfam domains (Fig. 6). Pfam00883 spanned the highly conserved catalytic domain (LAP-N residues 259-573). Pfam02789 was more variable in plant, animal, and microbe LAPs and spanned an N-terminal region corresponding to LAP-N residues 104 to 226.

The tomato LAP-N showed the highest degree of identity (97%) with the tomato TTP24 cDNA clone (Milligan and Gasser, 1995), suggesting that TPP24 and LapN were alleles. This cultivar variation was localized in four regions. First, the LAP-N presequence was six residues longer than the TPP24 polypeptide sequence, and there was no identity between the LAP-N and the TPP24 presequences until residue 12. Second, there was a nonconservative substitution from Ala-313 in LAP-N to an Arg in TPP24. Third, the LAP-N Asn-320 was deleted in the TPP24 polypeptide. All other plant LAPs retained Asn-320 (Fig. 6). Finally, residues 437 to 443 of the LAP-N sequence and the corresponding regions of TPP24 were distinct. The 8-residue sequence (ADALVYAC) found in LAP-N was substituted with a six-residue sequence (SVGISC) in TPP24. This region was imbedded in the highly conserved LAP catalytic domain (Fig. 6). Collectively, these data suggested that the TPP24 allele was a nonfunctional allele of LAP-N.

The wound-induced LAP-A preproteins of tomato and potato were only 77% identical with LAP-N. A previously characterized Arabidopsis LAP (LAP1; Bartling and Weiler, 1992) and two additional Arabidopsis LAPs (LAP2 and LAP3) were identified in the complete Arabidopsis genome sequence (The Arabidopsis Initiative, 2000). Although LAP2 and LAP3 were predicted to contain transit peptides similar to LAP-N (Fig. 6), LAP1 did not have this N-terminal extension (Bartling and Weiler, 1992). Comparisons of the overlapping 527-residue regions of Arabidopsis LAP1, LAP2, and LAP3 and rice LAP with the tomato LAP-N demonstrated identities ranging from 73% to 77%.

When the 11 plant LAP sequences were carefully inspected, 28 signature residues were identified. Signature residues were invariant in LAP-A or LAP-N and enabled classification of the plant polypeptides as LAP-N or LAP-A like (Fig. 6). The Arabidopsis, rice, white birch and parsley LAPs shared the tomato LAP-N signature residues. Only the potato LAP shared signature residues with tomato LAP-A. These signature residues included seven substitutions that changed residue charge, three substitutions that altered Pro residue locations and one three-residue insertion/deletion (residues 147-149; Fig. 6).

Similar to LAP-A (Gu and Walling, 2002), the LAP-N had strong identity with microbial and animal LAPs in the 251-residue COOH domain, ranging from 42% (bovine LAP) to 48% (E. coli PepA) identity (data not shown). More recently, LAP genes were identified in two cyanobacteria, Nostoc sp. PCC 7120 and Synechocystis sp. PCC 6803. The cyanobacterial LAPs lacked the first 70 residues of the LAP-N preprotein (corresponding to the transit peptide) but shared over 50% identity with LAP-N and 65% identity with each other (data not shown).

Characterization of a His₆-LAP-N Enzyme in E. coli

The strict conservation of primary sequence in the COOH domain, including the residues implicated in Zn ion coordination and substrate binding, suggests that LAP-N might have biochemical properties, a substrate specificity and catalytic mechanism similar to the tomato LAP-A (Gu et al., 1999; Gu and Walling, 2000, 2002). Both the FPLC-purified LAP-A and an affinity column-purified His₆-LAP-A assembled into hexamers in E. coli and had similar K_ms and substrate specificities (Gu et al., 1999; Gu and Walling, 2000). Therefore, to begin characterization of LAP-N, the LAP-N protein was overexpressed in E. coli as a His₆-LAP-N fusion protein. Total proteins were extracted in non-denaturing conditions and His₆-LAP-N was affinity purified. For comparison, His₆-LAP-A enzyme was purified in parallel. SDS-PAGE fractionation showed that the nickel affinity columns efficiently purified the 55-kD His₆-LAP-A and His₆-LAP-N proteins and no significant degradation products were detected by Coomassie Blue staining (Fig. 7A).

View larger version (30K): [in this window] [in a new window]   Figure 7.   Characterization of the His6-LAP-N enzyme expressed in E. coli. The His6-LAP-N and His6-LAP-A enzymes were overexpressed in E. coli and purified. A, Purified His6-LAP-N (2 µg) and His6-LAP-A (2 µg) were fractionated by 10% (w/v) SDS-PAGE and stained with Coomassie Blue. The molecular masses of protein markers run in a parallel lane are indicated in kilodaltons. B, Purified His6-LAP-N (3 µg) and His6-LAP-A complexes (3 µg) were fractionated by 7.5% (w/v) native PAGE. The complexes were visualized by silver staining. The molecular mass of the His6-LAP-A was previously determined and is shown in kilodaltons (Gu and Walling, 2000). C, Native His6-LAP-N (10 µg) was fractionated on native gels with six different acrylamide concentrations ranging from 4.5% to 9.0% (w/v). The purified His6-LAP-A (hexamer; 357 kD), porcine LAP (hexamer; 350 kD), and molecular mass markers including: bovine serum albumin (BSA) monomer (66 kD), BSA dimer (132 kD), urease trimer (272 kD), and urease hexamer (545 kD) were loaded in parallel lanes. Gels were stained with Coomassie Blue. The mass of each complex was determined by the relative mobility (RF) and the retardation coefficient (KR; Bryan, 1977). The marker proteins are indicated as black squares. The His6-LAP-A and pig LAP complexes are shown as white squares. The estimated molecular mass of the His6-LAP-N complex is shown as a white triangle.

To determine if the His₆-LAP-N protomer assembled into a multimeric form similar to His₆-LAP-A, the purified His₆-LAP-N and His₆-LAP-A enzymes were fractionated by native PAGE. There was a marked difference in the mobility of the His₆-LAP-A complex and His₆-LAP-N complex (Fig. 7B). The His₆-LAP-N complex ran more slowly than the His₆-LAP-A complex in native PAGE gels (Gu et al., 1996a) and exhibited diffuse silver staining (Fig. 7B). A second distinction was the fact that His₆-LAP-N was labile and could only be assayed in freshly prepared extracts, whereas His₆-LAP-A activity was stable after repeated freeze-thaw cycles.

Because both charge and molecular mass influence protein mobility in native gels, the molecular mass of His₆-LAP-N was calculated by measuring the relative mobility of the His₆-LAP-N using a set of gels with variable polyacrylamide content (4.5%-9%; Bryan, 1977). For comparison, the tomato His₆-LAP-A and the porcine LAP enzymes were included in these studies (Fig. 7C). The His₆-LAP-N had a mass of 365 kD and His₆-LAP-A was 350 kD, which was consistent with the previous determinations of 357 kD (Gu et al., 1999).

The substrate specificity of the affinity-purified His₆-LAP-N and His₆-LAP-A enzymes were determined using a set of nine amino acyl--naphthylamide (-NAP) substrates under standard LAP assay conditions. The ability of the LAP enzymes to hydrolyze the amino acyl bond was measured spectrophotometrically. The relative activity on each substrate was calculated by comparing with activities on Leu--NAP substrate (taken as 100%). His₆-LAP-A preferred substrates with N-terminal Leu, Met and Arg, as was previously shown (Gu et al., 1999) with a maximum hydrolysis rate of 24.5 µmol min¹ mg¹ (Leu--NAP; Table I).

View this table: [in this window] [in a new window]   Table I.   LAP-N and LAP-A1 enzyme activity on amino acyl- -NAP substrates

All nine of the amino acyl--NAP substrates were hydrolyzed by His₆-LAP-N at rates less than 1 µmol min¹ mg¹ (Table I). The His₆-LAP-N had a maximal activity on the Met--NAP substrate. However, it was 18-fold less active than His₆-LAP-A. Furthermore, the His₆-LAP-N hydrolyzed Leu--NAP at 0.59 µmol min¹ mg¹, which was 42-fold slower than His₆-LAP-A. Only Phe--NAP was cleaved more rapidly (3-fold) by His₆-LAP-N than His₆-LAP-A.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The LapA and LapN genes of tomato have distinct structures and expression programs. There are two tightly linked genes (LapA1 and LapA2) that encode the wound-induced LAP-A proteins (Gu et al., 1996a; Chao et al., 2000), whereas there was a single gene encoding LAP-N. The tomato LAP-N and LAP-A and all other plant and cyanobacterial LAPs had a highly conserved COOH domain that contained the residues important in catalysis and zinc ion coordination.

Several lines of evidence supported the idea that LAP-N and LAP-A were related, but distinct protein species and allowed other plant LAPs to be classified as LAP-N or LAP-A like. First, comparison of the deduced sequences of the tomato LAP-A1/LAP-A2 and LAP-N preproteins indicated that they were over 23% diverged. Second, 2D-PAGE immunoblots with affinity-purified antibodies showed that whereas epitopes in the central and COOH portions of these LAPs were shared, the most N-terminal domain of LAP-A and LAP-N (domain A; LAP-A residues 123-194) displayed distinct epitopes. Third, careful inspection of the LAP-N and LAP-A peptide sequences identified 28 signature residues that unambiguously classified plant LAPs as LAP-N- or LAP-A-like proteins. With the sole exception of the potato LAP, other plant LAPs characterized to date are LAP-N like. Because the LAP-N proteins were detected in all plant species examined (Bartling and Nosek, 1994; Chao et al., 2000), it is believed that the LapN-like genes were derived from an ancient ancestral gene. Furthermore, because the wound-induced LapA is not widely distributed in the plant kingdom (Hildmann et al., 1992; Chao et al., 2000), the tomato LapA1 and LapA2 genes may have arisen via duplication of LapN in a subset of the Solanaceae.

Correlated with their sequence divergence, the expression programs of LAP-A and LAP-N were distinct. The tomato LapN gene encoded a rare-class transcript. LAP-N proteins were detected at similar levels in all vegetative and reproductive organs. These results suggested that the expression pattern of LAP-N was similar to the Arabidopsis LAP (Bartling and Nosek, 1994). However, given the recent identification of the Arabidopsis LAP2 and LAP3 genes, the conclusions made by Bartling and colleagues may need to be reevaluated. It is possible that the Arabidopsis studies measured accumulation of all three LAP protein species, rather than LAP1, as initially thought (Bartling and Nosek, 1994).

Although the role of LAP-N is not yet understood, it is likely that LapN has a role in protein turnover required for cell maintenance in vegetative and reproductive organs. LAP-N may act on general proteins or peptides or may facilitate the turnover of specific polypeptides. Because the N-terminal residue of a protein can influence a protein's half-life (Varshavsky et al., 1997; Callis and Vierstra, 2000), it is possible that LAP-N may play a role in the regulation of ubiquitin-dependent protein degradation by exposing penultimate residues that influence protein half-life or by processing peptides released by the proteasome.

LAP-N may have a role in protein mobilization from cotyledons after germination. This conclusion was supported by the facts that LAP-N was present at low levels in imbibed tomato seeds and LAP-N levels increased after cotyledon emergence from seed coats and were retained for 4 d. The pattern of LAP-N protein accumulation was distinct from changes in LAP activities observed in barley and kidney bean cotyledons. Mikola and Kolehmainen (1972) showed that LAP activity remains constant in barley seeds before and after germination. In contrast, LAP activity declines before the mobilization of the majority of proteins from kidney bean cotyledons (Mikkonen, 1986). It has been postulated that the levels of the kidney bean LAP may still be sufficient for continued mobilization of protein reserves (Kolehmainen and Mikola, 1971; Mikkonen, 1992). To date, a correlation of LAP-N protein and activity levels in tomato is lacking because LAP-N activities in tissue extracts have not been detected using in situ gel assays. This contrasts to the easy detection of LAP-A activity after wounding (Gu et al., 1996b). The availability of transgenic tomato plants that have suppressed levels of the LAP-N protein should allow us to correlate changes in LAP-N protein accumulation and rate of storage protein hydrolysis (Pautot et al., 2001).

LAP-A accumulation patterns were distinct from LAP-N. LAP-A proteins were not detected in stems or roots and only low levels of LAP-A proteins were noted in cotyledons and healthy and senescent leaves. These data suggested that the LAP-A enzymes may have a limited role in modulating protein turnover in non-stressed vegetative organs. A similar pattern of expression was noted for the potato LAP (Hildmann et al., 1992; Dammann et al., 1997), which shared the LAP-A signature residues. Because LAP-A proteins are detected in leaves after wounding and infection (Pautot et al., 1993; Gu et al., 1996b) and in reproductive organs, LAP-A may have a more specialized role than LAP-N. It is possible that the LAP-A proteins have a role in protecting male and female reproduction organs from insect attack or pathogen infection (Milligan and Gasser, 1995; Pautot et al., 2001). Alternatively, LAP-A may have a specific or generalized role in protein turnover in response to stress.

In addition to distinct expression programs, the LAP-N and LAP-A enzymes may be biochemically distinct. This is supported by several observations. First, the multimeric His₆-LAP-N activity purified from E. coli had a diffuse silver staining pattern when compared with the tomato LAP-A or porcine LAP. Second, the purified LAP-N activity was unstable relative to the tomato LAP-A, requiring activity determinations to be performed in freshly prepared protein extracts. Finally, the His₆-LAP-N hydrolyzed most chromogenic substrates slowly and the substrate preferences of the purified His₆-LAP-N were distinct from His₆-LAP-A. Collectively, these data indicate that if LAP-N is a homohexameric enzyme, it is a very distinct enzyme from the tomato LAP-A. Alternatively, the tomato LAP-N expressed in E. coli may not have been able to assemble into its native quaternary structure due to the absence of its in vivo partner in E. coli. There is evidence in kidney beans and humans that heterohexameric LAPs are also present in eukaryotic cells (Sanderink et al., 1988; Mikkonen, 1992).

The biochemical characteristics of the tomato LAP-N contrasts with the Arabidopsis LAP1, which is classified as an LAP-N-like protein by virtue of its signature residues and percentage identity with the tomato LAP-N (Bartling and Weiler, 1992). The overexpression of an Arabidopsis LAP1-fusion protein in E. coli assembled into a stable homohexameric enzyme, although a high mass complex was also identified. AtLAP1's ability to hydrolyze Leu--NAP has not been tested; however, using the Leu-p-nitroanilide substrate (AtLAP1) had a specific activity similar to that reported for the tomato LAP-A1 and potato LAP (Bartling and Nosek, 1994; Herbers et al., 1994; Gu et al., 1999; Gu and Walling, 2000).

Based on the deduced sequences of the plant LAPs, LAPs may reside in two cellular compartments. The Arabidopsis LAP1 had no N-terminal targeting sequence, suggesting a cytosolic location (Bartling and Weiler, 1992). In contrast, all other plant LAPs had putative transit peptides suggesting localization within the plastid stroma (Herbers et al., 1994; Gu et al., 1996a). LAP localization may be more complex in tomato. Both LapA and LapN RNAs were predicted to encode a 60-kD precursor protein with a putative plastid transit peptide; however, cell fractionation studies suggest that despite the presence of plastid targeting signals, the majority of the 55-kD tomato LAPs that accumulate in tomato cells are soluble and cytosolic (Gu et al., 1996a, 1996b). In addition, the LapN transcript had a second in-frame translational initiation codon that could give rise to a 55-kD LAP-N form. The localization of LAP-N may be controlled by balancing translational initiation site use and LAP-N preprotein processing. Unfortunately, the N termini of the purified LAP-N polypeptides were blocked; therefore, data to support this hypothesis must be solely derived by immunolocalization of the tomato LAP-A and LAP-N proteins within tomato cells. Affinity-purified antisera that discriminate between LAP-A and LAP-N are being used to determine the cellular compartments of the tomato LAPs.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Materials

One-month-old tomato (Lycopersicon esculentum Peto 238R and VFNT cherry) plants were grown in a growth chamber with a 16-h-light (30°C)/8-h-dark (20°C) cycle. Leaf wounding and tissue harvest have been described by Pautot et al. (1991). Peto 238R seeds were imbibed in water and cotyledons were collected after 1 d (Stage 1), the day cotyledons emerged from seed coats (Stage 2), and 2 d (Stage 3) and 4 d after cotyledon emergence (Stage 4). Stamens, pistils, petals, and sepals were collected from mature plants. Stage 1 to 4 floral buds (0.3, 0.5, 0.7, and 1.0 cm in length, respectively) were collected. Roots, stems and mature and senescent leaves were harvested from 1.5-month-old tomato plants. Tissues were frozen in liquid N₂ and stored at 80°C until use.

Isolation of LAP-A Domain-Specific Antibodies

The sequence of the full-length LapA1 cDNA clone (pBLapA1) was previously described (GenBank accession no. U50151; Gu et al., 1996a). pDR57 is a partial LapA1 cDNA clone corresponding to nucleotides 325 to 1,843 (Pautot et al., 1993). pDR57 was digested with PstI and DNA fragments were end filled with T4 DNA polymerase and subcloned into SmaI-digested pGEX-3X to give rise to pGLAPA-A (nucleotides 390-600; domain A) and pGLAPA-B₉ (nucleotides 601-954). pGLAPA-B (nucleotides 601-718; domain B) was a spontaneous deletion mutant of pGLAPA-B₉ (Fig. 1A). The 886-bp PstI DNA fragment (nucleotides 954-1,843 plus the polylinker PstI site) was excised from a gel, digested with HincII and fragments were subcloned into SmaI-digested pGEX-3X to generate pGLAPA-F (nucleotides 1,617-1,843 plus polylinker sequences; domain F; Fig. 1A) and pGLAPA-E (not shown). To generate the pGLAPA-D subclone (nucleotides 891-1,287; domain D), HincII-digested pDR57 DNA was subcloned into SmaI-digested pGEX-3X (Fig. 1A). The in-frame fusions with GST were confirmed by DNA sequence analysis.

Escherichia coli cultures (250 mL) were grown at 37°C to OD₆₀₀ = 0.6 and isopropyl--D-thiogalactopyranoside (IPGT) was added to 0.1 mM. After 3 h of growth, cells were pelleted and resuspended in 5 mL of phosphate-buffered saline (PBS; 170 mM NaCl, 6.2 mM KCl, 12.6 mM Na₂HPO₄, and 2.2 mM KH₂PO₄ [pH 7.4]) containing 0.1% (w/v) Triton X-100 and 0.03% (w/v) SDS, sonicated for 1 min, and cooled on ice. The 1-min sonication/cooling cycle was repeated five times. The cell debris was pelleted by centrifugation at 4,350g for 20 min at 4°C. The supernatant was directly applied to a glutathione-agarose bead column according to the manufacturer's instructions (Sigma, St. Louis). After extensive washing with PBS, fusion proteins were eluted with 15 mM reduced glutathione. Accumulation of GST-LAP proteins in extracts was determined by Coomassie Blue staining and SDS-PAGE immunoblots.

The LAP-A1 polyclonal antiserum was previously described by Gu et al., (1996b). To isolate domain-specific antibodies from this serum, E. coli total protein extracts (80 µg) containing the GST-LAPA-A, -B, -D, or -F fusion proteins were fractionated by preparative SDS-PAGE and transferred to nitrocellulose filters. The regions of the blot containing the fusion proteins were identified by excising a strip of the blot and incubating the strip with a 1:1,000 (w/v) dilution of the LAP-A polyclonal antiserum. Region of the blot with a fusion protein was excised, incubated for 1 h in Tween-phosphate buffered saline (TPBS; PBS and 0.05% [w/v] Tween 20 [pH 7.4]) with 5% (w/v) dry milk, and incubated with undiluted LAP-A polyclonal antiserum with gentle shaking for 16 h at 4°C. After eight washes with TPBS (for a total of 72 h), the affinity-purified antibodies were eluted with a low-pH buffer (0.1 M Gly [pH 2.7]). The eluate was immediately neutralized by addition of 1 M sodium phosphate buffer (pH 7.7) to a final concentration of 50 mM (Gu et al., 1996b) and BSA was added to a final concentration of 50 mM. LAP affinity-purified antibodies were diluted 1:20 (w/v) before use in immunoblot analyses. The affinity-purified antibodies were stored at 4^o for no longer than 2 weeks.

Total Leaf Protein Isolation, SDS-PAGE, and 2D-PAGE Immunoblots

Total proteins from tomato leaves were isolated and fractionated by SDS-PAGE or 2D-PAGE as described by Wang et al. (1992). Native proteins from E. coli extracts were extracted as described previously (Gu et al., 1996a). Protein concentrations were measured with a modified Bradford method using BSA as a standard (Sedmak and Grossberg, 1977). Protein gels were stained with Coomassie Brilliant Blue R250 and transferred to nitrocellulose filters (Wang et al., 1992). Electro-transfer and immunoblot procedures were carried out according to Gu et al. (1996b).

LAP-N Protein Purification and Sequencing

Tomato seeds were germinated on moistened filter paper in petri dishes in the dark in a temperature controlled chamber at 25°C. One-week-old etiolated seedlings (25 g) were ground in liquid nitrogen. The powder was homogenized in 50 mL of PBS (pH 7.2) with 1 mM phenylmethylsulfonyl fluoride. The homogenate was centrifuged for 20 min at 12,000g at 4°C. The proteins were precipitated by 50% to 100% (w/v) (NH₄)₂SO₄ and were recovered by centrifugation. The protein pellet was resuspended in loading buffer (10 mL of PBS [pH 7.2], 1 mM phenylmethylsulfonyl fluoride, and 1 mM EDTA) and a Centricon-plus filtration system (30,000 M_r cutoff [MWCO]) was used to remove lower molecular mass polypeptides. Proteins were loaded onto a 5-mL DEAE-Sephadex column and unbound proteins were recovered and significantly enriched for LAP-N. Proteins were concentrated using a Centricon-plus filter (10,000 MWCO) and loading buffer was exchanged with lysis buffer (9.5 M urea, 1% [w/v] Nonidet P-40, 4% [w/v] ampholines [pH 5-7], 1% [w/v] ampholines [pH 3-10], and 5% [w/v] -mercaptoethanol). Proteins were separated by 2D-PAGE. The locations of LAP-N proteins were determined by 2D-PAGE immunoblots. LAP-N polypeptides were excised from six Coomassie Blue-stained gels, proteins were electroeluted and concentrated using a Centricon filter (10,000 MWCO). Purified proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. Membrane segments with LAP-N were sequenced using a sequencer (PE-Applied Biosystems, Foster City, CA) at the University of California Riverside Genomics Institute Core Facility.

Isolation and Characterization of LapN cDNAs

Total and poly(A⁺) RNAs were isolated from leaves of 1-month-old tomato cv Peto 238R plants according to Pautot et al. (1991). cDNAs were synthesized and packaged into gt11 SfiI-Not I cDNA arms according to the manufacturer's instructions (Promega, Madison, WI). The primary library contained 5.4 × 10⁶ recombinants. Approximately 2.3 × 10⁶ phage from the unamplified cDNA library was screened using a 1:500 (w/v) dilution of LAP-A polyclonal antibodies. To remove antibodies that cross-reacted with bacterial polypeptides, LAP-A antibodies were pre-incubated with E. coli Y1090 total proteins immobilized on nitrocellulose filters (Sambrook et al., 1989). Expression library screening was according to Sambrook et al. (1989) with the following modifications. All filters were pre-incubated with TPBS in 5% (w/v) dry milk for 1 h. Every wash step was at least 15 min. LAP-A antiserum-reactive clones were plaque purified by secondary and tertiary screenings. A total of 29 immunopositive -clones were identified.

Putative LapN clones were eluted from individual plaques in 500 µL of phage dilution buffer (0.01% [w/v] gelatin, 50 mM Tris-HCl [pH 7.5], 100 mM NaCl, and 8 mM MgSO₄). The cDNA inserts were amplified with PCR using the left and right gt11 primers as described by Gu et al. (1996a). cDNA inserts were separated on a 1% (w/v) agarose gel, and DNA blots were hybridized with a ³²P-labeled pDR57 probe (a general Lap probe) according to Walling et al. (1988). The six pDR57-positive clones were subsequently hybridized with³²P-labeled LapA domain A (pGLAPA-A, an LapA-specific probe) and domain D (pGLAPA-D, a general Lap probe) probes. LapA probes were labeled with [-³²P]-dCTP by nick translation according to the manufacturer's instructions (Gibco-BRL, Cleveland). LapN clones were identified as pGLAPA-A negative and pGLAPA-D positive. DNAs from six LapN clones were isolated (Sambrook et al., 1989), digested with Sfi I and NotI, and cloned into pGem11+ (Promega). EcoRI, Bam HI and HindIII restriction maps were determined. LapN3 contained the cDNA with longest 5'-UTR.

DNA and Protein Sequence Analyses

For DNA sequencing, the LapN cDNA insert in pGLapN was excised with NotI and SfiI, blunt ended with T4 DNA polymerase and subcloned into SmaI-digested pBluescript SK (pBLapN). A series of nested deletions were generated using exonuclease III (Henikoff, 1987). The nucleic acid sequences of both strands were determined by the dideoxy chain-termination method using Sequenase (United States Biochemical, Cleveland). The 1,943-bp cDNA contained 30 bp of the 5'-UTR followed by a 1,750-bp coding region, a 136-bp 3'-UTR and a 29-bp poly(A⁺) tail (GenBank accession no. AF510743). The deduced amino acid sequence of LapN was compared with several LAP proteins using the BLAST-2 program. The GenBank accession nos. are: tomato LapA1 (U50151), LapA2 (U50152), and TPP24 (U20594); potato (Solanum tuberosum) LAP (X77015); Arabidopsis LAP1 (X63444; At2g24200), LAP2 (AF424634; At4g30920), and LAP3 (AY090346; At4g30910); parsley (Petroselinum crispum) LAP (X99825); white birch (Betula pendula) LAP (Y14777); rice (Oryza sativa) LAP (The Institute for Genomic Research 2502.t0002); Nostoc sp. PCC 7120 LAP (NP_484281); and Synechocystis sp. PCC 6803 LAP (NP_441359). The tomato TTP6 cDNA (U20593) was not included in the LAP comparison because it is a partial LapA1 cDNA clone from line VF36 corresponding to LapA1 nucleotides 163 to 1,874. It has a four-nucleotide substitutions relative to the Peto238R LapA1 cDNA at position 829 (C  A), 830 (C  A), 1,090 (C  G), and 1,617 (C  T). Unlike AtLAP2 and AtLAP3, the AtLAP1 does not have a transit peptide. Over 2 kb of genomic sequences 5' to the predicted AtLAP1 translational start codon have been carefully inspected, and unlike AtLAP2 and AtLAP3, an exon encoding a transit peptide was not identified. This conclusion is also supported by two full-length AtLAP1 cDNA clones (AY062105 and AY035006). Percentage sequence identity was determined by BLAST2 comparisons and LAPs were aligned using the PileUp program (Genetics Computer Group, Madison, WI). Transit peptide predictions were made using the first 100 residues of LAP-A1, LAP-N and Rubisco (P08706) in the ChloroP program (Emanuelsson et al., 1999).

Genomic DNA Blots, RNA Blots, and RNase Protection Studies

Genomic DNA isolation, DNA-blot hybridization and washing conditions were performed as described by Walling et al. (1988). Peto238R and VFNT genomic DNAs (10 µg) were digested with EcoRI (Fig. 5A), EcoRV (data not shown) or Xba I (data not shown). DNAs were fractionated on a 0.7% (w/v) agarose gel. A single-copy equivalent of EcoRI-digested pBLapN (6.8 × 10² ng) was loaded in a parallel lane. DNAs were transferred to nitrocellulose filters and hybridized with a ³²P-labeled pBLapN probe. pBLapN was labeled with [-³²P]-dCTP by nick translation. The blot was exposed to Hyper-film-MP (Amersham, Piscataway, NJ) at 80°C with an intensifying screen for 48 h (DuPont, Wilmington, DE).

RNase protection assays have been described (Chao et al., 2000). ³²P-labeled antisense RNA probes were synthesized from SfiI-digested pBLapA1-3UTR (LapA1 RNA nucleotides 1,735-1,903) or SacI-digested pBLapN3-13 (LapN RNA nucleotides 1,754-1,945) using SP6 or T7 RNA polymerases (New England BioLabs, Beverly, MA), respectively, and [-³²P]-GTP. The antisense RNAs were incubated with RNase-free DNase I for 30 min, purified by fractionation over a Sephadex G-50 spin column, two phenol/sevag extractions, and ethanol precipitation before use. pBLapA1-3UTR was described previously (Gu et al., 1996a) and pBLapN3-13 was an LapN exonuclease III clone. pBLapA1-3UTR and pBLapN3-13 antisense RNAs (2.5 × 10⁶ cpm) were hybridized in a 30-µL reaction with 2 to 10 µg of healthy or wounded leaf poly(A⁺) RNA at 45°C and 51°C, respectively. Optimal hybridization temperatures were determined empirically using annealing temperatures in 3°C intervals from 42°C to 51°C. After hybridization and processing, RNAs were dissolved in 6 µL of double-distilled water and 4 µL of stop solution (95% [w/v] formamide, 20 mM EDTA, 0.05% [w/v] bromphenol blue, and 0.05% [w/v] xylene cyanol FF). RNA samples were denatured and fractionated on 6% (w/v) polyacrylamide gels. Sequencing ladders (M13mp18) were loaded in parallel lanes to determine the size of protected fragments. Gels were dried and exposed to Hyper-film MP for 16 h at 80°C with an intensifying screen.

Construction of His₆-LapN Fusion Protein Genes

The LapN primer 1 (5'-ATCGGATCCATGATTGCTCGTGATACTCTTGGTC-3') contained a Bam HI site (underlined) and an ATG translational start codon (italicized) followed by the 21 nucleotides of the mature LapN-coding region (nucleotides 192-203), which was inferred by similarity to LAP-A N terminus that was determined empirically (Gu et al., 1996a). The LapN primer 1 and the M13 reverse primer (5'-AGCGGATAACAATTTCACACAGGA-3') were used in a PCR reaction with pGLapN plasmid as a template. The 1.9-kb PCR product was gel purified (Qiagen USA, Valencia, CA) and digested with Bam HI and EcoRI. The LapN cDNA was assembled in pGem-11Zf(+) in two steps. First, the 148-bp Bam HI/EcoRI DNA fragment (LapN nucleotides 163-310) was ligated into Bam HI/EcoRI-digested pGem-11Zf(+) to generate pGB-E310, which was confirmed by DNA sequencing. Second, pGLapN was digested with EcoRI and the 1.6-kb EcoRI fragment (nucleotides 311-1916) was ligated with EcoRI-digested pGB-E310 plasmid to generate pGB-E1916.

The pGB-E1916 plasmid contained two Bam HI sites at LapN nucleotide 960 and adjacent to the initiation codon (see above). pGB-E1916 was digested with Bam HI and SacI to generate the 0.62-kb Bam HI (nucleotides 343-960) and the 0.95-kb Bam HI/SacI (nucleotides 961-1,916) fragments. Both fragments were subcloned into Bam HI/SacI-digested pQE30 to generate the plasmid pQLapN, which expresses the His₆-LAP-N from the tac promoter. To overexpress the His₆-LAP-A protein in E. coli, the pQLapA-M plasmid was used (Gu and Walling, 2000).

Overexpression and Purification of His₆-LAP-N and His₆-LAP-A Fusion Proteins

His₆-LAP-N and His₆-LAP-A1 overexpression and purification procedures were performed as described by Gu and Walling (2000) with minor modifications. The 500-mL cultures were grown at 25°C, and IPTG induction was for 4 h. The bacteria were resuspended in sonication buffer (50 mM NaPO₄ [pH 8.0] and 300 mM NaCl) at 5 volumes buffer per gram of cell pellet. Cells were frozen in a dry ice/ethanol bath and thawed in ice-cold water. Cells were lysed using five 1-min sonication pulses followed by cooling on ice for 1 min. The lysate was cleared at 10,000g for 20 min at 4°C. The His₆-LAP-N enzyme was bound to Ni-nitrilotriacetic acid resin (Qiagen USA) and eluted with a 30-mL gradient of 0 to 0.5 mM imidizole as previously described (Gu and Walling, 2000). Fractions with His₆-LAP-N were identified after SDS-PAGE or native PAGE by staining with Coomassie Brilliant Blue R-250 or immunoblot analyses using the LAP-A polyclonal antiserum.

His₆-LAP-N Mass Determination and Activity Assays

The purified His₆-LAP-N complex (10 µg) was fractionated on a set of six native polyacrylamide gels ranging from 4.5% to 9.0% (w/v). The protein complexes used as molecular mass standards (Sigma) were loaded in parallel lanes and included: BSA monomer (66 kD), BSA dimer (132 kD), urease trimer (272 kD), urease hexamer (545 kD), porcine LAP hexamer (330 kD; Sigma), and tomato His₆-LAP-A hexamer (357 kD). The gels were stained with Coomassie Brilliant Blue R-250 for 16 h and destained. The relative mobility of each protein was determined (Bryan, 1977).

Aminopeptidase activity assays using nine amino-acyl--NAP subtrates (Sigma) were performed according to the methods described by Gu et al. (1999). Two micrograms of purified His₆-LAP-N and His₆-LAP-A enzymes was assayed per reaction. Each reaction was performed in triplicate; the assays were repeated twice. Due to the instability of the His₆-LAP-N after freezing at 80°C, only freshly prepared His₆-LAP-A and His₆-LAP-N enzymatic were used in these assays.