Brassica rapa Has Three Genes That Encode Proteins Associated with Different Neutral Lipids in Plastids of Specific Tissues

Abstract

Plastid lipid-associated protein (PAP), a predominant structural protein associated with carotenoids and other non-green neutral lipids in plastids, was shown to be encoded by a single nuclear gene in several species. Here we report three PAP genes in the diploid Brassica rapa; the three PAPs are associated with different lipids in specific tissues. Pap1 and Pap2 are more similar to each other (84% amino acid sequence identity) than to Pap3 (46% and 44%, respectively) in the encoded mature proteins. Pap1 transcript was most abundant in the maturing anthers (tapetum) and in lesser amounts in leaves, fruit coats, seeds, and sepals; Pap2 transcript was abundant only in the petals; and Pap3 transcript had a wide distribution, but at minimal levels in numerous organs. Immunoblotting after sodium dodecyl sulfate-polyacrylamide gel electrophoresis indicated that most organs had several nanograms of PAP1 or PAP2 per milligram of total protein, the highest amounts being in the anthers (10.9 μg mg⁻¹ PAP1) and petals (6.6 μg mg⁻¹ PAP2), and that they had much less PAP3 (<0.02 μg mg⁻¹). In these organs PAP was localized in isolated plastid fractions. Plants were subjected to abiotic stresses; drought and ozone reduced the levels of the three Pap transcripts, whereas mechanical wounding and altering the light intensity enhanced their levels. We conclude that the PAP gene family consists of several members whose proteins are associated with different lipids and whose expressions are controlled by distinct mechanisms. Earlier reports of the expression of one Pap gene in various organs in a species need to be re-examined.

Special neutral lipids are present in chloroplasts and non-green plastids in leaves and other organs in plants during vegetative and reproductive growth and under stresses (Deruere et al., 1994; Rabbani et al., 1998; Ting et al., 1998; Vishnevetsky et al., 1999). These lipids, including carotenoids, triacylglycerols, and steryl esters, may accumulate in high amounts. In the chloroplasts the carotenoids function as light receptors in photosynthesis and quenchers of excess energy in photoprotection (Bartley and Scolnik, 1995; Havaux, 1998), and the triacylglycerols may be a temporary storage of energy or harmful free fatty acids from damaged lipids. In the chromoplasts in petals and fruits, the pigments function as attractants for animals in pollination and fruit dispersion. In the elaioplasts in the tapetum cells of the anthers, the steryl esters are to be deposited extracellularly onto the pollen surface for waterproofing and other purposes.

The above neutral lipids are located in the thylakoid membranes or in special structures such as fibrils, tubules, crystalloids, and globuli, etc. Within these structures the neutral lipids are present in a matrix covered with a layer of amphipathic lipids (phospholipids and glycolipids) and unique proteins, termed plastid-lipid associated proteins (PAPs), fibrillin, or chromoplast-specific protein (CHR). The deduced amino acid sequences of PAPs from the cloned genes of several diverse species are highly similar regardless of the types of neutral lipids that the PAPs cover (Pozueta-Romero et al., 1998; Vishnevetsky et al., 1999). They all have a putative plastid-targeting N-terminal peptide and a mature polypeptide of about 30 kD. The tertiary structure of the protein is unknown, and the amino acid sequences do not have a long hydrophobic stretch.

PAP genes are expressed in leaves and other organs. They are especially active in specific organs during certain phases of development. These phases include the formation of chromoplasts during colorization of the fruits (Knoth et al., 1986; Deruere et al., 1994) and petals (Vishnevetsky et al., 1997), and the production of elaioplasts in the tapetum during the maturation of the anthers (Ting et al., 1998). The expression of PAP genes can be induced by externally applied hormones (Vishnevetsky et al., 1997) and stresses (Oren-Shamir et al., 1993; Pruvot et al., 1996; Chen et al., 1998; Gillet et al., 1998), when there are increases in the plastid neutral lipids. The hormones such as ethylene and abscisic acid likely represent secondary messengers produced by the plants as a response to developmental changes or environmental stresses. The environmental stresses could include drought, wounding, and oxidation; the enhanced PAP could coordinate with the increased carotenoids to exert photoprotection.

In all of the above research on the PAP genes and their encoded proteins in diverse species, only one gene from each species has been studied. The expression of the gene during development or under environmental stresses was characterized by RNA-blot hybridization. Southern-blot hybridization has suggested the presence of a single PAP gene in bell pepper (Pozueta-Romero et al., 1998) and potato (Oren-Shamir et al., 1993). Two polypeptides in bell pepper samples, one of which could represent a breakdown product, were recognized by immunoblotting after SDS-PAGE (Chen et al., 1998). In citrus (Moriguchi et al., 1998) and a hexaploid potato variety (Oren-Shamir et al., 1993), Southern-blot hybridization has revealed more than one copy of the PAP gene.

We have characterized three PAP genes in the diploid Brassica rapa. Three equivalent genes in Arabidopsis are present in the GenBank, but have not been studied. The three B. rapa genes encode PAPs of highly similar amino acid sequences. However, their developmental and tissue-specific expressions are strikingly distinct, and their induction or repression in the leaves by applied abiotic stresses are coordinated. Here we report on the characteristics of these three members of the PAP gene family and their expression and we analyze the structures of the proteins. Our findings show that the earlier reports of treating the PAP transcripts in leaves, petals, and fruits in a species as being from the same PAP gene need to be re-examined.

RESULTS

Three to Four Pap Genes Were Found in the Diploid B. rapa

We used an incomplete-length cDNA clone of a Pap gene (termed Pap1), which encodes a PAP located in the elaioplasts in the tapetum of B. rapa (Ting et al., 1998) to screen a cDNA library of florets and obtained full-length cDNA clones of Pap1 and Pap2, and a gDNA library and obtained a Pap1 gDNA. We were unsuccessful in obtaining a Pap2-gDNA clone. We obtained a Pap2 gDNA fragment by PCR with gDNA and primers corresponding to the terminal sequences of the Pap2-cDNA.

The Arabidopsis genome contains three Pap genes. T01472 (termed At-Pap1) and T04905 (termed At-Pap2) bear close similarities (80% and 81% identity in the sequence encoding the open reading frame [ORF]) to B. rapa Pap1 and Pap2, respectively. A third Arabidopsis sequence, AAC3672 (termed At-Pap3), is distantly related. We used primers corresponding to the ORF of At-Pap3 to perform PCR with floret cDNA and gDNA and obtained Br Pap3-cDNA and Br Pap3-gDNA, respectively. A summary of the three genes and their transcripts is shown in Figure 1.

Figure 1.

Major features of the three B. rapa Pap genes and their mRNAs. The flanking sequence of Pap1, but not those of Pap2 or Pap3, has been characterized. The complete sequences of Pap1 and Pap2 mRNA are known, whereas the sequence of Pap3 is known only at the coding region. The white box on the mRNA represents the sequence encoding the mature protein, and the black box depicts the sequence encoding the putative plastid targeting peptide. The lengths of the three mRNAs observed by RNA-blot hybridization are indicated. Sequences representing the gene-specific (probe A) and nonspecific (probe B) probes for each Pap are indicated. The numbering of the nucleotides starts at 1 for A in the first ATG. Restriction sites are E (EcoRI), X (XbaI), H (HindIII), Xh (XhoI), and B (BamHI). Pap1 (AF290566), Pap2 (AF290567), and Pap3 (AF290568), Pap1-cDNA (AF290563), Pap2-cDNA (AF290564), and Pap3-cDNA (AF290565) have been recorded in GenBank.

Pap1 and Pap2 are more similar to each other (79% identity) than to Pap3 (43% and 44%, respectively) in the sequences encoding the mature proteins, whereas the sequences of the three genes encoding the 5′-untranslated region (UTR) and the putative plastid-targeting peptide (to be described) are less similar to one another. We used a gene-specific probe and a relatively nonspecific probe for each Pap gene (Fig. 1) for Southern-blot hybridization (Fig. 2). The Pap1 gene-specific probe detected one fragment from each of the EcoRI, HindIII, and XbaI reaction products; the lengths of these fragments are consistent with the gene sequence. The Pap1 nonspecific probe detected EcoRI (3.8 kb), HindIII (5.5 kb), and XbaI (2.3 kb) fragments, as expected (Figs. 1 and 2), but also detected, presumably from a new gene or allele, a 8.0-kb EcoRI fragment, a 6.6-kb HindIII fragment, and 7.0- and 6.0-kb XbaI fragments. The Pap2 gene-specific probe detected two EcoRI fragments, one HindIII fragment, and one XbaI fragment, consistent with the gene sequence. The Pap2 nonspecific probe detected EcoRI (3.8 kb), HindIII (5.5 kb), and XbaI (2.3 kb) fragments (corresponding to those fragments of Pap1), and also a 6.6-kb HindIII fragment and 7.0- and 6.0-kb XbaI fragments of the presumed new gene or allele. Either of the two Pap3 probes recognized the same single fragment produced from each of the three restriction enzymes.

Figure 2.

Southern-blot hybridization of the three Pap genes. Gene-specific (probe A) and nonspecific (probe B) probes for each gene (see Fig. 1) were used. Restriction enzymes are E (EcoRI), H (HindIII), and X (XbaI). The sizes of the DNA fragments are indicated on the left.

The results suggests that Pap1, Pap2, and Pap3 each have only one copy in the genome. The presumed new gene would be more similar to Pap1 than Pap2 and relatively dissimilar to Pap3.

Figure 3 is a pileup of the amino acid sequences of PAP available in the GenBank. A phylogenetic tree constructed on the basis of these sequences is shown in Figure 4. Three sequences from the gDNA of the cyanobacteria Synechocystis encode putative proteins (GenBank accession nos. BAA17246, AAD38023, and BAA14161; one of which was mentioned in Vishnevetsky et al., 1999) that have low residue similarities with all the PAP (31% or less, data not shown). These putative bacterial proteins probably are unrelated to the PAP.

Figure 3.

Alignments of the amino acid sequences of all reported PAPs. The alignments were made with the MegAlign program from DNAstar using the Clustal method. Residues identical in all PAPs are shaded. The residue immediately after the putative cleavage site, which has been predicted in cucumber and bell pepper and determined in B. rapa PAP1 and PAP2 by N-terminal sequencing of the mature protein in the current study, is underlined. Horizontal bars denote residue gaps. Arabidopsis Pap1 (T01472), PAP2 (T04905), and PAP3 (AAC3672) are from GenBank and the names are temporarily assigned in reference to their similarities to the respective B. rapa PAP. Other PAPs are from potato (T07825), wild potato (CAA10372), bell pepper (S56633), tobacco (T03635), citrus (BAA34702), cucumber (T10179), tomato (T07125), and pea (AAD02288).

Figure 4.

A phylogenetic tree of PAPs of different species. The tree was constructed with the aligned sequences shown in Figure 3. Only the portions of the sequences encoding the mature proteins were used. The scale represents branch distance as the number of residue changes between neighbors.

The Pap Genes in B. rapa, Arabidopsis, and Other Species Have Introns at Similar Locations

The Pap genes in B. rapa and other diverse species have an intron at the same location midway in the ORF (Fig. 5). Most of them have another intron at an identical location downstream from the first intron. The exceptions include B. rapa Pap2, which does not have a second intron, and tobacco Pap, which has its second intron further downstream.

Figure 5.

Locations and lengths of introns in Pap of diverse species. The length (in basepairs) of each intron (represented by a short horizontal bar) is indicated. One or two introns are present in each gene. The first introns and the second introns (except tobacco) in all genes have identical or very similar locations. The complete gDNA sequences of Pap in tomato and tobacco are not available and the introns at the available sequences are indicated. Only bell pepper Pap has a Ts (a long interspersed repetitive element) inserted at the 5′ terminus of the first intron. See legend of Figure 3 for sources of the Pap genes.

Pozueta-Romero et al. reported (1998) that the first intron in bell pepper Pap has a Ts of the long interspersed repetitive element family inserted at the 5′ border (Fig. 5); this Ts insertion results in the failure of the intron to be removed from a percentage of the Pap transcripts during in vivo RNA splicing. There is no such Ts inserted into any of the intron in any of the B. rapa or Arabidopsis Pap genes. Also, we did not find a transcript of any of the three B. rapa Pap genes that was longer than the expected length after in vivo RNA splicing (data not shown).

The Transcripts of the Three Pap Were Present at Very Different Levels in Various Organs during Development

RNA-blot hybridization was performed using gene-specific probes of the three Pap genes (Fig. 1). Pap1 transcript had a wide distribution in green organs, including anthers, sepals, seeds, fruit coats, and leaves, of different developmental stages (Fig. 6). It was most abundant in the anthers at stage 2 when the tapetum cells started to accumulate PAP1 and steryl-ester globules in the elaioplasts (Wu et al., 1997; Ting et al., 1998). In leaves, the level of Pap1, as well as Pap2 and Pap3, transcripts was highest in mature leaves, medial in young leaves, and least in senescing leaves. Pap1 transcript was at very low levels in petals and pistils, and not detectable in roots.

Figure 6.

RNA-blot hybridization of various B. rapa organs of different developmental stages. Gene-specific probes for the three Pap genes were used. Left and right panels are different blots, and samples of stage 3 anthers and stage 3 petals were used as cross references in the blots on the right panels. The radioactive blots were exposed to x-ray films for a short (for Pap1 and Pap2) or long (about 10×) time (for Pap2 and Pap3) to reveal the relative levels of transcripts in different samples. Ethidium bromide-stained 25S and 18S rRNA in the gel reveal that equal amounts of RNA in each sample were used.

Pap2 transcript was present almost exclusively in the petals (Fig. 6). Nevertheless, it could be detected in most organs when the radioactive RNA blot was extensively exposed to the x-ray film. In these other organs the weak signals detected by the Pap2 gene-specific probe represented authentic Pap2 transcript rather than cross-hybridized Pap1 transcript because stage 1 and stage 2 anthers exhibited the most abundant Pap1 transcript, but had little Pap2 transcript.

Pap3 transcript was present at very low levels compared with Pap1 transcript in various organs. It could be detected only after the radioactive blot was extensively exposed to the x-ray film. It was present in essentially all the organs of various developmental stages at levels that were not drastically different among the samples (Fig. 6).

Overall, Pap1 transcript was present in diverse organs, but most abundantly in the anther tapetum. Pap2 transcript is present mainly in the petals. Pap3 transcript was ubiquitous among various organs, but at very low levels. The prevalence of the Pap transcripts in various organs was comparable with the levels of their respective PAP proteins (next section).

The Levels of PAP1 and PAP2 Were Much Higher Than That of PAP3 in Various Organs and Especially High in the Anthers and Petals, Respectively

PAP1 and PAP2 are very similar in amino acid sequences, whereas they are quite dissimilar to that of PAP3 (Figs. 3 and 4). We prepared two types of chicken antibodies, one against PAP1 (for the detection of PAP1 and PAP2) and another against PAP3. Antibodies against PAP1 recognized two proteins of about 34 kD in immunoblots after SDS-PAGE of the total proteins of the various organs (Fig. 7). They specifically recognized a protein of 34 kD in the extracts from the anthers, sepals, fruit coats, and leaves, and a protein of 36 kD in the extracts from the sepals and petals. The proteins of 34 and 36 kD were identified as PAP1 and PAP2, respectively, by N-terminal sequencing. The 34-kD protein from the anthers had VAEK(Q) VAEEAIESA, and the 36-kD protein from the petals had VIDAEDELDPE. The sizes of PAP1 (34 kD) and PAP2 (36 kD) relative to each other are consistent with those of the two proteins (26,496, and 28,063 daltons, respectively) calculated from the deduced amino acid sequences. The sizes of the PAP as determined by SDS-PAGE are higher than those deduced from amino acid sequences; this discrepancy has been reported on the PAPs from other species (e.g. Deruere et al., 1994). Antibodies against PAP3 did not recognize any protein in the extracts from the various organs, although they were active toward the recombinant antigen (data not shown).

Figure 7.

SDS-PAGE and immunoblotting of proteins of the total extracts and the plastid fractions of different B. papa organs. The gel was stained with Coomassie Blue (top panel) or treated for immunoblotting using antibodies against PAP1. The two short bars on the right indicate the 36-kD PAP2 and the 34-kD PAP1, respectively. Positions of M_r markers are shown on the far right.

The amounts of the three PAPs in the total extracts of the various organs were determined semi-quantitatively by immunoblotting. PAP1 was present at a level of 10.9 μg per 1 mg of protein in the extract of anthers and less than 1 μg per 1 mg of protein in the extracts of other organs (Table I). PAP2 was present in 6.6 μg per 1 mg of protein in the extract of petals and less than 0.5 μg per 1 mg of protein in the extract of other organs. The abundance of PAP1 in the anthers and PAP2 in the petals was consistent with the prevalence of Pap1 and Pap2 mRNAs in these two respective organs (Fig. 6). The high amount of PAP1 in the anthers (1.1% of the total proteins) apparently reflects the abundance of the protein, which covers the steryl-esters globules of 0.5 μm in diameter in the predominant elaioplasts in the tapetum (Wu et al., 1997; Platt et al., 1998). It is analogous to the 8% oleosin in B. rapa seeds in which the oil bodies of 0.6 μm are covered with the protein (Huang, 1992). PAP1, rather than PAP2 and PAP3, was also present in measurable amounts in all organs possessing chlorophyll. The amount of PAP2 in the petals correlated with the content of carotenoids in the chromoplasts (Table I). Pap2, as well as Pap1 and Pap3, transcripts in leaves decreased during senescence (Fig. 6), which is indicative of PAP2, as well as PAP1 and PAP3, playing a role only in chloroplasts of healthy leaves. PAP3 in the extracts of all the organs was not detected and determined to be present in less than 0.02 μg per 1 mg of protein (Table I).

Table I.

Contents of various constituents in the total extracts of Brassica organs

Constituentsa	PAP1	PAP2	PAP3b	Chlorophylls	Steryl Esters	Carotenoids
	μg per mg−1 proteins
Leaves	0.92	<0.05	<0.02c	49	NDd	4.1
Petals	<0.05	6.6	<0.02	0	ND	292
Anthers	10.9	<0.05	<0.02	1.0	247	0.03
Sepals	0.50	0.41	<0.02	16	ND	0.6
Fruit coats	0.18	<0.05	<0.02	26	ND	0.36
Seeds	0.73	<0.05	<0.02	18	ND	0.08

^a Mature leaves, stage 5 petals, stage 3 anthers, stage 4 sepals, stage 5 fruit coats, and stage 4 maturing seeds were used. 

^b PAP amounts were determined semiquantitatively by immunoblotting after SDS-PAGE. 

^c When PAP in an extract was not detected, the value of <0.05 or <0.02 was assigned on the basis of results using the recombinant antigen of decreasing amounts. 

^d ND, Not determined. 

PAP1 and PAP2 Were Localized in Isolated Plastids of Various Organs

The lone PAP detected in selected organs in a few plant species had been localized in the plastids (Deruere et al., 1994; Vishnevetsky et al., 1996; Pozueta-Romero et al., 1997; Kessler et al., 1999). In the three B. rapa PAP amino acid sequences deduced from the DNA sequences, a high percentage of the N-terminal 40 to 50 residues are hydroxylated and basic (Fig. 3). Such a high percentage is a characteristic of an N-terminal plastid-targeting peptide.

We tested whether PAP1 and PAP2 in various B. rapa organs were located in the plastids. The total extract and the plastid fraction of each organ were subjected to SDS-PAGE and then immunoblotting with antibodies against PAP1 (and also PAP2; Fig. 7). In each organ many more different proteins in the total extract than in the plastid fraction were resolved by SDS-PAGE. PAP1, of 34 kD, in the extracts of anthers, sepals, fruit coats, and leaves were recovered in the plastid fractions. In a similar manner, PAP2, of 36 kD, in the extracts of sepals and petals was recovered in the plastid fractions.

Only a small percentage of PAP1 of 34 kD in the anther extract was recovered in the plastid fraction at the same location in the immunoblot (visible, but not on the photo after photography). This protein was subjected to N-terminal sequencing, and the result indicated that it was PAP1. In the plastid fraction an immunodetected protein of a higher apparent M_r (Fig. 7) was observed; presumably, this protein was modified from PAP1 during the plastid isolation procedure.

The finding that PAP1 and PAP2 were located in the plastid fractions is consistent with the presence of putative plastid-targeting peptides in the nascent proteins. We should mention that the above plastid fractions obtained after gradient centrifugation probably contained contaminated organelles.

The Levels of Pap Transcripts in the Leaves Changed after the Plants Were Stressed

B. rapa plants grown in pots inside a growth chamber at a light intensity of about 600 μE m⁻² s⁻¹ at 16°C were subjected to various stress treatments. The levels of the three Pap transcripts in the leaves were monitored by RNA-blot hybridization with the use of gene-specific probes. The PAP proteins were not monitored because the immunoblotting barely detected PAP1 and could not detect PAP2 and PAP3 (Fig. 7).

Withholding water to the plants greatly reduced all three Pap transcripts during a 5-d period (Fig. 8). During this period the Relative Water Contents of the leaves in successive days were 90% (at the start of the treatment), 90%, 90%, 90%, 65%, and 40%. Re-watering the plants changed the Relative Water Content of the leaves to 80% after 1 d and led to a substantial increase of all three Pap transcripts.

Figure 8.

RNA-blot hybridization of leaf samples after the plants had been subjected to various stress treatments or light intensity for the designated days. Gene-specific probes for the three Pap genes were used. The radioactive blots were exposed to x-ray films for a short (for Pap1) or long (about 10×) time (for Pap2 and Pap3) to reveal the relative levels of transcripts in different samples. 25S rRNA in the gel reveal that equal amounts of RNA in each sample were used. Top left (drought), After 5 d of withholding watering, the plants were re-watered. Relative Water Contents (RWC) in the leaves are shown at the bottom. Top right (ozone), Plants were grown in ozone at 0.075 μL L⁻¹; ozone at a higher concentration led to irreparable damage to the leaves. Bottom left (wounding), The leaves were wounded mechanically. Bottom right (light intensity), Light intensity received by the plants in E (μE m⁻² s⁻¹) are indicated.

The ozone effect on plants is related to oxidative stresses (Chen et al., 1998). In B. rapa plants in a growth chamber containing 0.075 μL L⁻¹ ozone, the levels all three Pap transcripts remained unchanged for the first 1 to 2 d and then declined gradually (Fig. 8). During the 4 d of treatment, the leaves remained green and appeared healthy. Subjecting the plants to twice the concentration of ozone (0.15 μL L⁻¹) generated necrotic spots on the leaves. These leaves were not studied because the treatment was drastic and the results would have little physiological relevance.

Mechanical wounding of the leaves caused a gradual increase in the levels of all three Pap transcripts (Fig. 8). The levels of the Pap transcripts peaked at d 1 through 3 and then declined.

In all the above stress treatments, the light intensity was maintained at about 600 μE m⁻² s⁻¹. The levels of Pap in some plant species increased at high light intensity (Chen et al., 1998; Gillet et al., 1998). B. rapa plants are temperate crops and grow well at lower light intensities. We tested whether the levels of the Pap transcripts in B. rapa leaves would change under lower or higher light intensity (Fig. 8). Plants grown in pots in a greenhouse (about 26°C/18°C of 14-h/10-h day/night cycle) at a light intensity of about 1,000 μE m⁻² s⁻¹ were transferred to a growth chamber at 16°C of continuous light of 100 μE m⁻² s⁻¹ for 7 d (leaves losing some green color), then 300 μE m⁻² s⁻¹ for 14 d (leaves regaining the green color), and finally back to the greenhouse. When the plants were transferred to a different light and temperature regime, the levels of all three Pap transcripts increased, although the increase was not very drastic or persistent.

In the above studies of Pap transcripts in stressed plants, Pap1 transcript and PAP1 were much more abundant than the other two Pap transcripts and PAPs. This assessment was made on the basis of the times required to expose the different RNA blots to x-ray films for visualization and the quantification of PAPs in the control leaf samples (Fig. 6).

DISCUSSION

The existence of three Pap genes in diploid B. rapa and Arabidopsis indicates that the Pap gene family is small, but has distinct members. Earlier studies of Pap expression and PAP dealt with a lone gene in a particular species, and the findings on the same or different species were not always consistent. Pap transcript was found in high amounts in organs containing chromoplasts such as the petals and fruit coats; it was present (Oren-Shamir et al., 1993; Vishnevetsky et al., 1996; Pozueta-Romero et al., 1997) or absent (Deruere et al., 1994) in the leaves. Also, the amount of Pap transcript was found to be related to (Deruere et al., 1994) or independent of (Moriguchi et al., 1998) the accumulation of carotenoids in the fruit coats. These discrepancies can be easily explained by our current findings of the existence of different genes whose sequences are very similar, but whose expressions are drastically different at the temporal and spatial levels. The use of a gene nonspecific probe will detect all the three Pap transcripts. In B. rapa, Pap2 is highly expressed only in the yellow petals, whereas Pap1 is highly expressed in the anthers and leaves. The use of a Pap2 nonspecific probe would have detected Pap2 in the petals and, erroneously, Pap1 in the leaves. It is clear that the control of expression of the Pap genes in other species needs to be re-examined by first characterizing the Pap gene family in that particular species.

It has been suggested that PAPs in the plastids perform two functions (Deruere et al., 1994; Monte et al., 1999; Vishnevetsky et al., 1999). First, PAP on the thylakoid membranes modulates the action of the carotenoids involved in the light reaction of photosynthesis and in photoprotection. Second, PAP stabilizes carotenoids and other neutral lipids (e.g. steryl esters) accumulated in subplastid structures such as fibrils and globules. Our current findings raise the possibility that the above two functions, especially the second function of lipid stabilization, are exerted by different PAPs. In B. rapa, PAP2 is located in the chromoplasts in the yellow petals and presumably stabilizes the accumulated carotenoid structures. Pap2 transcript and PAP2 are present in minimal amounts in other organs and decreases in senescing leaves. PAP1, rather than PAP2, is associated with steryl esters in the tapetum elaioplasts. The three Pap transcripts are ubiquitous in most organs, but Pap1 transcript and PAP1 are much more prevalent. PAP1 may be the dominant constituent of the internal membranes in the chloroplasts where it modulates the action of the carotenoids. Although there are clear-cut demonstrations of PAP being associated with carotenoid-containing structures in chromoplasts, there is no experimental evidence or theoretical implications that PAP is in direct contact with the carotenoids on the thylakoid in chloroplasts. The three PAP might have distinct subplastid locations and associate with different lipids.

Earlier reports have described some characteristics of PAP on the basis of its amino acid sequence (Deruere et al., 1994; Ting et al., 1998; Vishnevetsky et al., 1999). These characteristics include an acidic pI, the absence of Cys, several consecutive aspartic/glutamic residues (nos. 58, 73, 91, 98, and 180 in PAP1), a cell adhesion motif (RGD, no. 308 in PAP1), a sequence of several tandemly situated glutamic/aspartic residues each followed by two non-charged residues (no. 238 in PAP1). Here we describe some significant features of the protein structure. PAP from different species can be divided into two groups on the basis of sequence similarities (the first 10 and the remaining three listed in Figs. 3 and 4). Group 1 members have a long hydrophobic stretch consisting of 16 residues flanked on either side by a basic residue (KWILVYTSFVGLFPLLSR, no. 154 in BrPAP1) and predicted to be a β-structure (Chou and Fasman, 1978). If this stretch were to form a hairpin inserted into the lipid structure (fibril, globule, etc.), its length would reach a depth no longer than the length of the acyl moiety of the surface phospholipid/glycolipid. In addition, there is no Pro residue at the center of the stretch that would promote a turn to form a hairpin. In contrast, oleosin, the structural protein on seed oil bodies, has a hydrophobic stretch of 72 residues and several Pro residues in the middle of it; these features would enable the stretch to form a hairpin penetrating into the lipid core of the oil body (Huang, 1992). Therefore, the hydrophobic stretch of PAP should not exist as a hairpin structure penetrating into the neutral lipid core as proposed (Vishnevetsky et al., 1999). Rather, it would only interact with the surface of the lipid structure, presumably with the phospholipids and not glycolipids because the flanking basic residues could interact with the phosphate groups of the phospholipids. Group 2 members do not have this hydrophobic stretch of 16 residues (Fig. 3); their corresponding stretch has a Glu residue near the center and does not have the flanking basic residues. Group 1 and Group 2 members contain a putative amphipathic α-helical structure of 22 residues (STNAK—, no. 209 in PAP1). The prediction is based on the alternation of hydrophobic and hydrophilic residues, the biphasic location of the hydrophobic and hydrophilic residues on a helical wheel (drawing not shown), and the length of 22 residues being sufficient to form six turns of an α-helix. This amphipathic helix would run parallel to the surface of the lipid structure (globule. fibrillin, etc.). The hydrophobic phase would interact with the acyl moieties of surface amphipathic lipids, whereas the hydrophilic phase would have its basic residues interacting with the phosphate groups of the phospholipids and more importantly with adjacent acidic PAP. Whether the stretch forms an α-helix remains to be seen, and a Pro residue, a potential helical breaker, is near the center of the stretch. Overall, the mode of PAP interacting with the lipid structures in the plastids is less similar to that of oleosin, which penetrates into the neutral lipid core of the seed oil bodies. It resembles more that of apolipoprotein, which runs along the surface of the lipoproteins in mammals (Small, 1992).

The expression of the Pap genes in various species can be altered by the external applications of hormones or stresses. The hormonal effects on Pap expression (Deruere et al., 1994; Vishnevetsky et al., 1999) are comparable with those in normal physiological processes (e.g. fruit ripening and floral induction) or related to the possible role of the hormone (e.g. abscisic acid) being an intermediate in signal transduction of the stress. Earlier studies have shown that mechanical wounding, drought, and oxidative stress increase the expression of Pap in the leaves of some species (Chen et al., 1998; Gillet et al., 1998; Rey et al., 2000). The idea is that when stressed, the leaves synthesize more PAP to modulate the action of carotenoids for the extra load of light reaction and photoprotection or to stabilize the accumulated neutral lipids. In our studies mechanical wounding and light intensity alteration increase, whereas drought and ozone applications decrease the expression of the three Pap genes. The negative effects of stresses on Pap expressions that we have observed are not necessarily contradictory to the earlier findings because the growth conditions, the levels of stresses applied, and the species were different. These earlier studies focused largely on finding induction of the Pap and other related genes, and presumably the conditions for the suppression of Pap expression were not explored.

MATERIALS AND METHODS

Plant Materials

Brassica rapa var. R500 seed was obtained from Calgene (Davis, CA) and was used to produce flowering plants in 5-gallon pots in a greenhouse maintained at 26°C/18°C with a 14-h/10-h day/night cycle.

For developmental studies of the floret parts, the florets were divided into six developmental stages. From stages 1 through 6, the lengths of the sepals were 2, 3, 4, 5, 6, and 6 mm, and the color was green at stage 1, gradually turning greenish yellow at stage 6. The lengths of the petals were 2, 3, 4, 5, 7, and 8 mm, and the color was greenish yellow at stages 1 and 2 and rapidly changed to yellow at stages 3 through 6. The length of the anthers was 2, 2, 2.5, 3, 3, and 3 mm, and the color gradually changed from light green to yellow. The lengths of the pistils were 2, 3, 4, 5, 6, and 9 mm, and the color remained green except for the stigma, which was greenish white. At stages 1 through 4, the sepals covered the floret. At stage 5, the sepals had split open, exposing the yellow petals. At stage 6, the floret was completely open and the pollen had matured.

For developmental studies of the fruit parts, the fruits were divided into six developmental stages. For stages 1 through 6, the lengths of the fruit (silique) not including the stalk were 30, 50, 60, 70, 70, and 70 mm, and the color was green at stages 1 through 4 and became yellow and brown at stages 5 and 6, respectively. The diameters of the round seed were 1, 1, 1.5, 2.2, 2.5, and 2.5 mm, and the seed was green and soft at stages 1 through 3 and turned brown and hard at stages 5 and 6. The fruit coat and the maturing seed of the various stages were obtained.

Leaves were collected from the greenhouse-grown plants. Young leaves were collected from one-half-expanded green leaves. Mature leaves were those that had just fully expanded. Senescing leaves were mature leaves that had started to become yellow.

Roots were collected from 6-d-old seedlings grown in 1:1 (v/v) perlite:vermiculite in the greenhouse.

Stress Treatments of Plants

Each plant was grown from seed in a 1-gallon pot of soil in a growth chamber maintained at 16°C and at a light intensity of 600 μE m⁻² s⁻¹ in a 12-h/12-h light/dark cycle. After 35 d of growth the plant reached a height of about 0.3 m. It was active in vegetative growth and produced no flower buds. The plant was then subjected to a stress treatment, and mature leaves were sampled 2 h after the start of the light period.

In drought treatment the potted plant was placed on a pan for a few days during which it was watered by placing water in the pan. Drought treatment was initiated by withholding watering. During a 5-d period, the daily Relative Water Content in the leaves was 90%, 90%, 90%, 65%, and 40%. After the 5th d, watering was resumed and after 1 d, the relative water content became 80%.

In ozone treatment, ozone was applied by passing a mixture of air and ozone (generated using a battery of UV lights) continuously through the chamber. The ozone concentration in the chamber was adjusted to 0.075 μL L⁻¹.

In mechanical wounding treatment, the mature leaves were gently pinched as described (Gu and Walling, 2000) with a pair of forceps to which two squares (1 × 1 cm) of stainless steel wire mesh (24-grid) were mounted. One leaf was pinched at 10 different locations, and the pinched area was about 5% of the leaf area.

In experiments to alter the light intensity received by the plants the plants were grown in pots in a greenhouse (about 26°C/18°C of 14-h/10-h day/night cycle) at a light intensity of about 1,000 μE m⁻² s⁻¹. They were transferred to a growth chamber at 16°C of continuous light of 100 μE m⁻² s⁻¹ for 14 d, then 300 μE m⁻² s⁻¹ for 7 d, and finally placed back at the greenhouse.

Isolation of Plastids from Various Tissues/Organs

Elaioplasts from stage 3 florets were isolated according to the method described earlier (Wu et al., 1997). The florets were chopped to a fine mince with a razor blade in a Petri dish containing a grinding medium of 0.05 m HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid]-NaOH, pH 7.5 and 0.8 m Suc (125 florets of 3.1 g per 8 mL). The homogenate was filtered through a layer of Nitex cloth (20 × 20-μm pore size), which removed the microspores. The filtrate (4 mL) was placed in a 17-mL centrifuge tube. Successive layers of 4 mL each of 0.4, 0.2, and 0 m Suc solutions (all containing 0.05 m HEPES-NaOH, pH 7.5) were placed on top of the filtrate. The tube was centrifuged at 9,000 rpm in a rotor (SW 28.1, Beckman Instruments, Fullerton, CA) for 2 h. The elaioplast fraction banding at the interface between 0.2 and 0 m Suc solutions was collected.

Plastids from other tissues/organs (mature leaves, stage 4 sepals, stage 4 maturing seeds, stage 5 fruit coats, and stage 5 petals) were isolated using the method described in the preceding paragraph with the following modifications. The filtered homogenate was placed in a 17-mL centrifuge tube on top of successive layers of 2 mL of 2.2, 1.5, and 1.15 m Suc solutions. On top of the homogenate, successive layers of 1.5 mL of 0.6, 0.4, 0.2, and 0 m Suc solutions were applied. The green plastids of leaves, sepals, seeds, and fruit coats banded at the interface of 1.15 and 1.5 m Suc solutions, whereas the yellow plastids of petals banded at the interface of 0.2 and 0 m Suc solutions.

SDS-PAGE, Immunoblotting, and Microsequencing

All procedures followed those described earlier (Wu et al., 1997). Proteins were separated by 12.5% (w/v) SDS-PAGE for 2.5 h at 100 V and were stained with Coomassie Blue. For protein-blot analyses, proteins on the gels were transferred to a nitrocellulose filter membrane. The filter was subjected to immunodetection with the use of chicken antibodies against bacteria-synthesized PAP1 and PAP3 (to be described). A semi-quantitative determination of each of the three PAPs in a sample was carried out as follows. SDS-PAGE and immunoblotting were performed with the sample and serial dilutions (5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%) of known weights of the recombinant PAP1, PAP2, or PAP3 (to be described) and 6-his-ALL (a recombinant maize protein [K. Hsieh and A. Huang, unpublished data] as control for subtracting the immunosignal resulted from the antibodies recognizing the epitope of the N-terminal six His residues in 6-his-PAP1). Visual comparison of these samples allowed us to estimate the relative amounts of the three PAPs in a particular sample. For sequence analyses, proteins on the gels were transferred to a polyvinylidenedifluoride membrane.

Quantification of Proteins, Chlorophylls, and Carotenoids

Total homogenates of the samples were prepared as described before. The protein content was determined by the Bradford method (Smith et al., 1985). Lipids in the homogenate was extracted with 80% (v/v) acetone and were separated by thin-layer chromatography (silica gel 60A, Whatman, Clifton, NJ) using 70:20:10 (v/v) petroleum ether:1,4-dioxane:acetone. Silica containing the visible chlorophyll and carotenoid spots were scrapped, and the pigments were eluted with 80% (v/v) acetone. The quantities of the chlorophyll (Lichtenthaler, 1987) and carotenoids (Rabbani et al., 1998) were determined spectrophotometrically.

Isolation of Three Pap Genes

Stage 2 floret and stage 4 anther cDNA libraries of B. rapa (Kim and Chung, 1997) were screened with ³²P-labeled Bcp32 (=Pap1, containing an incomplete-length coding region) cDNA (Ting et al., 1998). Out of 150,000 plaques, 21 positive plaques were obtained. Restriction enzyme mapping showed that the cDNA inserts in these plaques consisted of Pap1 cDNA or Pap2 cDNA. One cDNA, the longest, from each group was studied further; they are termed BrPap1 and BrPap2, respectively. A genomic library of B. rapa (Kim and Chung, 1997) was screened with ³²P-labeled Bcp32 cDNA. Out of 280,000 plaques, five positive plaques were obtained. Restriction enzyme mapping showed that all the five plaques were gDNA of Pap1. A 3.8-kb EcoRI fragment of Pap1 gDNA was subcloned into pBluescript SK (Stratagene, La Jolla, CA).

Pap3 cDNA was cloned by reverse transcriptase-PCR. First-strand cDNAs were synthesized from 1.5 μg of total RNAs from stage 3 florets. The total RNAs were incubated in 100 μL of 1× Moloney murine leukemia virus reverse transcriptase buffer (Stratagene) containing 50 μm each of dATP, dCTP, dGTP, and dTTP, 40 units of RNase Block (Stratagene), and 15 μmol of Pap3C primer (5′-TCAGAGCTCAAGCAGAGAGCT-3′), which was based on the 3′ terminal region of the ORF of Arabidopsis sequence of AtT32F12 (accession no. AAC36172) encoding a putative PAP, at 70°C for 5 min and then at room temperature for 15 min. One microliter of Moloney murine leukemia virus reverse transcriptase (50 units μL⁻¹, Stratagene) was added, and the mixture was incubated at 42°C for 2 h. Second-strand cDNA synthesis and amplification were performed by PCR with the use of 5 μL of the first-stranded reaction mixture as a template, 0.5 μm of each of Pap3 n primer (5′-ATGGCTACGCTCTTCACCGTC-3′), which was based on the 5′ terminal region of the ORF of Arabidopsis sequence of AtT32F12, and Pap3C primer.

For the cloning of Pap2 and Pap3 genomic DNA, 1 μg of gDNA was used as a template, and Pap2 n (5′-TAAAAAAACAAAACAATGGCGACGGTT-3′, corresponding to the 5′-flanking region of the sequence encoding the N terminus of PAP2) and Pap2C (5′-GCATTAAAGAGTTCAAGGGTTCAAGAG-3′, corresponding to the 3′-flanking region of the sequence encoding the C terminus of PAP2), Pap3 n and Pap3C were used as the primers, respectively. The conditions for amplification were 35 cycles of 94°C for 1 min, 50°C for 1 min, 72°C for 1 min, and 72°C for 10 min. The PCR products were gel-purified using QIAGENE MAX (Qiagen, Valencia, CA), and cloned into pGemT (Promega, Madison, WI) using a cloning kit.

Preparation of Gene Nonspecific and Specific Probes for Pap1, Pap2, and Pap3

The gene nonspecific probes were cDNA sequences representing a major portion of the sequence encoding the mature protein (to be shown in Fig. 1). The probes for Pap1 and Pap2 were 0.52- and 0.51-kb fragments cut with XhoI and XbaI from the respective cDNA. The nonspecific probe for Pap3 was a 0.61-kb fragment cut with BamHI and HindIII from Pap3 cDNA.

The gene-specific probes were cDNA sequences representing a portion of the 5′-UTR and the sequence encoding the putative plastid-targeting peptide (to be shown in Fig. 1). The probe for Pap1 was a 0.46-kb fragment obtained by PCR with the use of Pap1 cDNA and the primers 5′-AGACCATGTCGATTCCC-3′ (corresponding to the 5′ terminus of the mRNA) and 5′-CACTTTCTCAGCCACCG-3′ (corresponding to the 3′ terminus of the mRNA region encoding the putative plastid targeting peptide). The probe for Pap2 was a 0.26-kb fragment obtained by PCR using Pap2 cDNA and the primers 5′-GAATAATTAAAAAAAC-3′ (corresponding to the border of the 5′-UTR and the sequence encoding the putative plastid targeting peptide) and 5′-CATCAATAGAGCCGATC-3′ (corresponding to the 3′ terminus of the mRNA region encoding the putative plastid-targeting peptide). The conditions of the PCR reactions were those described in the preceding section. The probe for Pap3 was a 0.37-kb cDNA fragment cut with EcoRI and BamHI from the cDNA.

The levels of 25S rRNA in various samples were used to standardize the signals observed in northern-blot hybridization. A 25S rRNA cDNA fragment (0.48 kb) was amplified by PCR using the primers 5′-GAATTCACCAAGTGTTG-3′ and 5′-TCGAATCTTAGCGACAA-3′. These two primers were selected on the basis of the cauliflower 25S rRNA gene sequences (accession no. X60324).

The above DNA fragments were separated by gel electrophoresis, eluted from the gel, and ³²P-labeled using Multiprimer DNA Labeling Kit (Amersham Pharmacia, Piscataway, NJ). The labeled probes were purified using ProbeQuantTM Micro Columns (Amersham Pharmacia) according to the manufacturer's protocol.

Genomic DNA-Blot and RNA-Blot Hybridization

DNA was isolated from young leaves by a procedure described earlier (Kim and Chung, 1997). Ten micrograms of DNA was digested with EcoRI, HindIII, or XbaI. The fragments were electrophoresed on a 1% (w/v) agarose gel and blotted onto a Hybond N membrane (Sambrook et al., 1989). The membrane was pre-hybridized at 65°C in 5× SSPE, 5× Denhardt's solution, 0.5% (w/v) SDS, and 100 μg mL⁻¹ salmon sperm DNA for 1 h and sequentially hybridized with ³²P-labeled gene probe for Pap1, Pap2, or Pap3 (preceding section). Afterward, the membrane was subjected to high-stringency washes in 2× SSC/0.1% (w/v) SDS for 20 min, 1× SSC/0.1% (w/v) SDS for 30 min, and 0.1× SSC/0.1% (w/v) SDS for 30 min, all at 65°C.

Total RNAs were isolated from the various tissues/organs as described (Verwoerd et al., 1989). An aliquot of 30 μg of total RNA of each sample was fractionated on a 1.3% (w/v) formaldehyde gel and transferred to a Hybond N membrane (Amersham Pharmacia). The membrane was pre-hybridized at 65°C in 0.5 m phosphate buffer, pH 7.2, 7% (w/v) SDS, 1% (w/v) bovine serum albumin, and 0.1 m EDTA, pH 8.0, for 2 h, hybridized with the ³²P-labeled gene probe, and then washed, as described in the preceding paragraph.

DNA Sequencing and Sequence Analyses

Sequencing was carried out with an ABI373 automated fluorescent sequencer (Applied Biosystems, Foster City, CA).

Sequencing of Pap1, Pap2, and Pap3 cDNA clones in pBluescript SK was performed with gene-specific primers and M13 forward and reverse primers of the universal priming sites in the vector.

A 5.4-kb fragment of Pap1 gDNA (consisting of 2.2-kb 5′-flanking region, 1.4-kb coding region, and 1.8-kb 3′-flanking region) was cut into five smaller fragments by restriction enzymes, which were subcloned into pBluscript SK. Sequencing was performed with gene-specific primers and M13 forward and reverse primers of the universal priming sites in the vector. The PCR products of genomic DNA of gPap2 and gPap3 were subcloned into pGemT (Promega) cloning vector for sequencing.

Database searches were performed using the BLAST Network Service. Amino acid alignments and phylogenetic analysis were performed according to the Clustal method using the DNASTAR software (DNAstar, Madison, WI).

The assignments of the first ATG codon in the Pap transcripts were made partly on the proper alignments of their deduced amino acid sequences with those of Pap in other species. They were also made on the basis of the length of the mRNA determined by RNA-blot hybridization (to be described), the occurrence of ATG codons, and the consensus sequences at the translation initiation sites in plant genes (Lutcke et al., 1987).

Expression of Pap1 and Pap3 in Escherichia coli for the Production of PAP Mature Proteins and Preparation of Antibodies

The sequence encoding the mature protein of PAP1 was amplified by PCR using Pap1 cDNA as a template. The 5′ primer (5′-GGGCTAGCGAGGAAGCCATCGAGTCTGCGGAG-3′) corresponded to the amino acid sequence EEAIESAE near the N terminus and contained an NheI site (underlined). The reverse primer (5′-AAAAGCTTTTAAGGGTTTAAGAGAGAGCTTCC-3′) corresponded to the amino acid sequence GSSLLNP near the 3′ terminus and contained a HindIII site (underlined). The PCR product was digested with NheI and HindIII. The fragment was ligated in-frame into an NheI/HindIII-digested pRSETB (Invitrogen, San Diego) to generate pRSET-6His-PAP1.

The sequence encoding the mature protein region of PAP3 was obtained by digestion of Pap3 cDNA with BamHI and HindIII (to be shown in Fig. 1). The fragment was cloned in-frame into a BamHI/HindIII-digested pRSETB to generate pRSET-6His-PAP3.

The plasmids pRSET-6His-PAP1 and pRSET-6His-PAP3 were used to transform E. coli BL21(DE3) pLysE (Novagen, Madison, WI) according to the manufacturer's instructions. The procedure for inducing the synthesis of recombinant proteins followed the protocol supplied by Invitrogen with the use of isopropyl-β-d-thiogalactopyranoside. The optimal concentration of isopropyl-β-d-thiogalactopyranoside (1 mm), temperature (37°C), and time (3 h) to induce the transformed E. coli to produce large amounts of the recombinant protein was determined empirically. After induction, the cells were harvested by centrifugation at 4,000g for 15 min at 4°C. The cells were resuspended in 10 mL of lysis buffer (50 mm NaH₂PO₄, pH 8.0, 10 mm Tris-HCl, pH 8.0, 8 m urea, and 100 mm NaCl). Afterward, the 6His-recombinant proteins were subjected to TALON metal-affinity resin chromatography (CLONTECH Laboratories, Palo Alto, CA), according to the procedure provided by the company and further purified by SDS-PAGE. The gels corresponding to the recombined PAP1 and PAP3 of 27 and 31 kD, respectively, were cut and used to produce antibodies in chickens (Wu et al., 1997).