INTRODUCTION

Flower meristem identity genes control the developmental transition of plants from the vegetative to reproductive phase. Several flower meristem identity genes have been cloned from Arabidopsis; among them, LEAFY (LFY) is one of the earliest flower meristem identity expressed genes 1, 2. LFY encodes a plant-specific transcription regulator 3, 4 and plays a major role in flower initiation 1, 2. LFY transcripts are expressed at low levels during the vegetative phase, which are increased upon the transition from the vegetative to reproductive phase 5. Mutations in LFY lead to the conversion of the flower meristems into inflorescence meristems and result in the replacement of flowers by lateral shoots 6, 7, 8, 9. Constitutive expression of LFY promotes reproductive transition and results in early flowering 10, 11. In the complex flower development pathway, LFY is centrally located and acts as a link between the upstream flower timing genes and downstream floral homeotic genes 5. LFY integrates environmental and endogenous signals to control flower timing 12. It has also been shown that LFY activity is required for the activation of all the three classes of floral organ identity genes 4, 13, 14, 15.

Recent studies showed that the complex regulatory network of flower development is largely conserved across distantly related plants, especially among dicotyledonous plant species 16. LFY homologs have been cloned from several species of gymnosperms and angiosperms, including both monocots and dicots 9, 17, 18, 19, 20, 21, 22, 23, 24. Their functions are highly conserved with slight variations among different species. Constitutive expression of LFY promoted flower initiation not only in Arabidopsis 11, but also in heterologous species, such as tobacco 25, aspen 11, rice 21 and citrus 26. Ectopic expression of LFY homologous genes from other plant species in the Arabidopsis lfy mutant can partially or fully complement the lfy mutation 18, 27.

Papaya (Carica papaya L.) is a fruit crop grown widely in tropical and subtropical lowland regions. It is a dicotyledonous, diploid plant species and relatively close to the fully sequenced model plant Arabidopsis from the point of view of family level 28. Papaya is one of the few plant species that produces three sex forms, male, female and hermaphrodite, and provides a unique opportunity to examine the applicability of the floral development pathway established in hermaphrodite to dioecious plants. Our previous study demonstrated that papaya has a primitive Y chromosome controlling sex determination 29. Cloning and characterization of flower development genes from papaya will help us to understand the roles of floral meristem and organ identity genes in sex determination and differentiation in plants.

Here, we report the cloning and characterization of LFY ortholog in papaya, PFL. It shares a high similarity and identity with the LFY orthologs from other dicot species and its expression pattern is similar to that of other orthologous genes. Our results showed that PFL is a functional ortholog of the Arabidopsis LFY gene expressed in all three sex forms.

MATERIALS AND METHODS

Plant materials

Three different papaya cultivars, SunUp, Kapoho and AU9, were germinated and planted in Kunia station, Oahu. Young leaf tissues of SunUp, Kapoho and AU9 were collected for genomic DNA isolation. Hermaphrodite and female flower buds, young leaf tissues and root tissues were collected for total RNA isolation. Shoot apical meristem (SAM), floral meristem and leaf meristem tissues from SunUp and Kapoho were collected and fixed at different floral developmental stages for in situ hybridization.

cDNA library construction

Total RNA was isolated from hermaphrodite and female flower buds using the method described by Hall et al 30. PolyA+ RNA was isolated from the total RNA using PolyATtract mRNA Isolation System (Promega, WI, USA) according to manufacturer's instruction. About 5μg polyA+ RNA was used to synthesize double stranded cDNA using the ZAP-cDNA Synthesis kit (Stratagene, CA, USA). Size-selected cDNA (100 ng) was ligated with 1μg EcoRI/XhoI digested Uni-ZAP XR lambda vector and packaged with GigapackII packaging extract (Stratagene, CA, USA). Libraries of 8×106and 3.8×106 recombinants, respectively, were generated. A total of 18,432 (48×384) recombinant clones for each library were picked and archived in 384-well plates containing freezing medium. High-density (18,432 clones per 22.5 × 22.5 cm2) and low-density (4,608 clones per 22.5×22.5 cm2) membranes were made using a Genetix Q-BOT (Genetix, UK).

Southern analysis

Isolation of papaya genomic DNA from leaves was based on the procedure described by Tai and Tanksley 31 with minor modifications. Three restriction enzymes, HindIII, EcoRI and XbaI were used to digest papaya genomic DNA from the three cultivars, Kapoho, SunUp and AU9. Genomic DNA samples (10 μg) were incubated with 5 U of restriction enzyme/μg DNA in a total volume of 50 μl at 37°C for 4 h. Approximately 10 μg digested genomic DNA was electrophoresed through 0.8% agarose gel in 1×TBE buffer. After electrophoresis, the gel was blotted onto Hybond N+ membranes (Amersham, UK) using standard methods 32. The probes were labeled by random primer labeling system (RediPrimer II, Amersham, UK). The procedure for hybridization and autoradiography were as described by Chittenden et al 33.

In situ hybridization

Apices, flower meristems, and leaf meristems of papaya were harvested at different stages and fixed in FAA fixation solution (3.7% formaldehyde, 5% acetic acid, 50% ethanol). Samples were then dehydrated in increasing concentrations of ethanol and embedded in paraffin (Paraplast Plus, VWR, PA, USA). Sections (8 μm thick) were cut and mounted onto ProbeOn Plus slides (Fisher Scientific, PA, USA). The amplified PFL cDNA fragment was cloned into TOPOII vector (Invitrogen, CA, USA). Antisense and sense RNA probes were generated with opposing SP6 and T7 promoters from the PFL cDNA subclone and labeled with digoxigenin (DIG) (Roche, Germany). The methods for tissue pretreatment and in situ hybridization were as described by Jackson 34.

RT-PCR analysis

Total RNA was treated with RNase-free DNase I (Promega, WI, USA) and quantified using a spectrophotometer (Shimadzu, Japan). About 5 μg of total RNA were reverse transcribed using RETROscript Kit (Ambion, TX, USA). PCRs were performed on the cDNA using the primer pair PFL-Sub10-3 (5′-CCA GAA CAT AGC CAA GGA GC) and PFL 2403 (5′-AAT GGC AAG ACG AGG ATG TG).

Real-Time quantitative RT-PCR assays

One primer pair (A1: 5′-ATA GTG CAG GCG CCA GTC AG; A2: 5′-GGC TCA AGC TGT TCA TCA TC) had a Tm of >55°C and was designed using Clone Manager 6 (Sci Ed Central) to produce a PCR product of 223 bp. About 2 μg of total RNAs were treated with RNase-free DNase I (Promega, WI, USA) and reverse transcribed using TaqMan RT kit (Applied Biosystems, CA, USA). cDNA products were diluted 5.5 fold for use in real-time RT-PCR amplification. Platinum Quantitative PCR SuperMix-UDG (Invitrogen, CA, USA) was used for real-time RT-PCR amplification. Reactions were run on a DNA Engine OPTICON 2 Real-Time PCR detection system (MJ research, MA, USA). Thermocycler conditions were 2 min at 50°C followed by 10 s at 95°C and 40 cycles of 15 s at 95°C, 20 s at 58°C, and 30 s at 72°C. For each sample, an RNA sample without reverse transcriptase was included to control for genomic DNA contamination. All presented PCR data were generated from a minimum of three independent reactions for each biological replication. Actin gene was used as a standard to control for sample variation in the total amount of RNA in different reactions.

Phylogenetic analysis

Multiple alignments of conceptual amino acid sequences were generated by using Clustal W 1.81 (http://www.cmbi.kun.nl/bioinf/tools) with a gap open penalty of 10.00 and a gap extension penalty of 0.05. Phylogenetic analyses were conducted using the Neighbor-Joining (NJ) method as implemented by the PHYLIP program package. Pam matrix was used to generate distance. Bootstrap analysis (1,000 replicates) was performed to assess the support of individual branches.

RESULTS

Cloning of PFL

The Arabidopsis LFY cDNA was used as a probe to screen the papaya BAC library. Two positive BAC clones, 42B17 and 52E1, were identified and confirmed by southern hybridization. The sizes of these positive BAC clones were estimated as 60 kb and 175 kb, respectively. The smaller BAC clone, 42B17, was digested with HindIII and shotgun subcloned into pPCR-Script vector. The sub-clones were screened with the LFY probe and the positive sub-clone was sequenced from both ends. Direct sequencing of the BAC clone was carried out at the 3′ end of the sub-clone where the third exon of PFL was interrupted by a HindIII site.

A full-length sequence of PFL was obtained through directly sequencing of BAC clone 42B17 via a primer walking approach. Three exons were predicted based on sequence homology and conserved splicing sites. The sizes of three exons were 429 bp, 315 bp and 357 bp, respectively. Two primers were designed at the beginning of the first predicted exon and the 3′-end of the third predicted exon. The cDNA fragment of PFL was obtained by amplification of the cDNA from papaya hermaphrodite flower buds using these two primers. This cDNA fragment was cloned into pCRII-TOPO vector (Invitrogen, CA, USA) and sequenced. The sequence result matched exactly the predicted open reading frames. Comparison between PFL cDNA and genomic sequence confirmed that PFL has three exons and two introns and that the splicing sites are highly conserved among PFL, FLO, and LFY.

PFL cDNA was about 1.1 kb and predicted to encode a protein of 367 amino acids. The deduced amino acid sequence of PFL was aligned with other FLO/LFY-like proteins (Fig. 1). This alignment revealed two highly conserved regions and two short variable regions. A proline-rich region near the amino terminus and an acidic central region were found in the variable regions of most of angiosperm FLO/LFY-like proteins. The proline-rich region was not found in the papaya PFL. However, a highly acidic region was observed in PFL between glutamate 182 and glutamate 188 where it was preceded by a short basic region (between lysine 171 and lysine 178), which is similar to other FLO/LFY-like proteins. PFL also shares with other dicot FLO/LFY-like proteins a region containing leucines repeated every 7 or 8 amino acids (positions leucine 69 to leucine 99) and a region with alternating basic-acidic-basic stretches (positions arginine 115 to arginine 126).

Figure 1
figure 1

Sequence comparison of FLO/LFY-like proteins. The leucine repeats are marked with “*”. The basic- and acidic-regions are labeled as “+” and “-”, respectively.

The comparison of PFL proteins to other FLO/LFY-like proteins showed that PFL protein is more closely related to the other dicot FLO/LFY proteins than to their monocot counterparts or to gymnosperm FLO/LFY-like proteins. PFL protein shares 66% and 70% similarity with LFY and its snapdragon counterpart FLO. It shares a higher similarity with LFY homolgous proteins of two woody tree species, California sycamore (Platanus racemosa) and black cottonwood (Populus trichocarpa) at 75% and 76%, respectively (Tab. 1).

Table 1 Comparison of amino acid sequences between papaya and other FLO/LFY homologs

Southern hybridization analysis of papaya genomic DNA was performed using the subcloned PFL genomic fragment as a probe to determine whether PFL is a single copy gene in papaya. Only one hybridizing band was seen with Hind III digestion (Fig. 2). One strong band with some weak bands were seen with EcoRI digestion and XbaI digestion. Since the sequencing result showed there are EcoRI and XbaI restriction sites within the PFL gene and only one cDNA fragment was amplified from the mRNA of papaya hermaphrodite flower buds, it is likely that only one copy of PFL gene exists in the papaya genome. This same probe was also used to label the papaya BAC library (data not shown). Nine positive BAC clones were obtained. All nine positive BAC clones showed the same pattern after they were digested with HindIII and XhoI and labeled with the PFL probe. This result reinforced our conclusion that only one copy of PFL exists in the papaya genome.

Figure 2
figure 2

(A) Restriction map of genomic clone of PFL of selected sites. The probe used in the southern hybridization was indicated in the line underneath. (B) Southern hybridization result of Kapoho (K) and SunUp (S) papaya genomic DNA digested with three different enzyme, EcoRI, HindIII, and XbaI, and labeled with PFL probe. S: size control. The same fragment with the probe was used as the size control.

A phylogenetic tree was constructed using the predicted protein sequences to determine the evolutionary relationships between PFLand other FLO/LFY-like proteins (Tab. 2, Fig. 3). The phylogenetic tree shows three FLO/LFY-like proteins from two different gymnosperm species, NLY, PrFLL and GinLFY, clustered together and quite separate from the proteins of the angiosperms. Five FLO/LFY-like proteins from monocot species, RFL, LtLFY, ZFL1, ZFL2 and JunefLF cluster in a group separate from the dicot species. LFY, VcLFY and BOFH from three different species within Cruciferae formed a distinct cluster. Similar to this, NFL1, NFL2 and TOFL from two different species within the Solanaceae formed another sub-cluster. The general phylogenetic relationships found in this tree are consistent with the currently accepted species phylogenies. The PFL protein is closely related to other FLO/LFY-like proteins in dicots, but it is very different from its monocot or gymnosperm counterparts.

Table 2 List of FLO/LFY homologs cloned in different plant species
Figure 3
figure 3

Phylogenic relationship among FLO/LFY-like proteins. Bootstrap analysis (1,000 replicates) was performed to assess the support of each branch. Bootstrap values (percentages) are shown on branches when over 50. Branch lengths are proportional to the number of amino acid substitutions.

Expression of PFL during papaya flower development

The flowering times of different papaya genotypes are strictly related to the number of nodes produced before flowering 35. Three papaya genotypes were chosen to test whether the flowering time of papaya is related to the expression of PFL. Saipan Red, an early-flowering genotype, flowers at 17-20 nodes. Kapoho, a late-flowering genotype, flowers at 45-47 nodes. SunUp is intermediate and flowers at 32-34 nodes. The SAM tissues from these three genotypes were collected at 5 node intervals beginning at the 5-node stage. These tissues were used to determine the stages at which the PFL would be turned on using in situ hybridization. The in situ hybridization revealed that PFL mRNA was already detected in the SAM of young seedlings of all three genotypes at the 5-node stage (Fig. 4).

Figure 4
figure 4

In situ hybridization of PFL transcripts in vegetative tissues of papaya. (A) Longitudinal section of SAM and leaves hybridized by the antisense PFL RNA probe. Expression was detected in the SAM and in the adaxial face and margins of leaves (L). (B) A consecutive section of a was hybridized by the sense PFL RNA probe as control. (C) Longitudinal section of vegetative meristem hybridized by the antisense PFL RNA probe. Expression was observed in vegetative meristem (VM) and leaf primordia. (D) A consecutive section of c was hybridized by the sense PFL RNA probe as control.

In situ hybridization was also carried out to study PFL's spatial expression pattern (Fig. 5). PFL mRNA was detected at a very high level in flower primordia and at a relatively lower level in leaf primordial. Expression of PFL was observed in the adaxial face and margins of leaf and axillary buds. PFL was also expressed in developing floral organs but at lower levels than in the flower primordia. At later stages of flower development, PFL expression was restricted to the two inner whorls of the flower.

Figure 5
figure 5

In situ hybridization of PFL transcripts in floral tissue of papaya. (A) Longitudinal section of a floral branch meristem (FM) hybridized with the antisense PFL RNA probe. Hybridization signals were observed in floral meristems (FM) at different developmental stages. (B) A consecutive section of (A) was hybridized with the sense PFL RNA probe as control. (C) Longitudinal section of inflorescence hybridized with the antisense PFL RNA probe. Expression is detected in all floral organ primordia. (D) A consecutive section of c was hybridized with the sense PFL RNA probe as control.

RT-PCR results showed that PFL is expressed not only in reproductive stage, but also in vegetative stage (Fig. 6). PFL's expression was found in root and leaf tissue as well as flower tissue. Quantitative Real-time PCR was conducted to test PFL's expression levels in papaya flowers of three different sex types. The expression levels of PFL were the same between male and female flowers, but Marginal difference was detected between hermaphrodite and unisexual flowers (P< 0.05) with slightly elevated expression level in unisexual flowers (Fig. 7).

Figure 6
figure 6

RT-PCR result. Actin gene was used as positive control. R: root; l: leaf; yb: young flower buds; mb: mature flower buds; c: carpel; -: negative control.

Figure 7
figure 7

PFL expression analysis in papaya different sex-type flowers by quantitative Real-Time PCR. Error bars represent standard deviations calculated from three replications.

DISCUSSION

Papaya is a perennial semi-woody tropical fruit tree that flowers throughout the year once flowering is initiated. The pattern of flowering in papaya is different from seasonal flowering annual and perennial plant species. PFL expression was found in roots, SAM, leaf primordia, leaves and axillary buds throughout the vegetative phase. This is consistent with documented observations that all known FLO/LFY orthologs, except FLO from snapdragon, are expressed in both vegetative and floral meristems. However, during the reproductive phase development, PFL expression in the flower primordia is at a significantly elevated level, suggesting that PFL has a role in initiating flowers and regulating floral organ identity genes similar to other LFY/FLO homologues. This also suggests that there might be regulatory genes interacting with PFL to elevate PFL expression to a critical level in order to initiate flowering.

Papaya flowers are formed in the axils of leaves. After the completion of transition from vegetative phase to reproductive phase, its SAM continuously initiates new leaves and new flowers. In contrast, the SAM of Arabidopsis and Antirrhinum only produce flowers and bracts after the transition. It is not clear whether PFL functions in maintenance of an indeterminate phase so that the SAM can continuously produce leaf and floral primordia as proposed for the LFY orthlogs in pea, grapevine and tomato 19, 22, 36. Such a role would be in line with the result that PFL expressed in both vegetative and reproductive phases and the observation that a vegetative phase co-exists with a reproductive phase after flower initiation.

Hermaphrodite and female flowers of papaya are determinate while male flowers are partially determinate. Flowering in papaya has no effect on leaf morphology, although a novel function of FLO/LFY orthologs in regulating leaf morphogenesis has been reported previously 19, 22. However, initiation of male flowers results in production of long and branched peduncle with many clusters of male flowers. This unique trait in male trees is more likely associated with the sex determination gene and less likely affected by PFL since there is no difference in expression levels between male and female flowers.

With male, female, and hermaphrodite co-existing in the same species, papaya is an ideal system to examine the roles of floral meristem and organ identity genes in sex determination and differentiation and to test the applicability of ABCE model in dioecious species. Our preliminary analyses resulted in marginal difference between hermaphrodite and unisexual flowers. This difference might reflect the role of PFL in regulation of unisexual organ development, or simply due to the difference of developmental stages of the same sized hermaphrodite and unisexual flowers.

Homologous genes in species controlling the basic pathways have evolved with variation in their amino acid sequences and modified functions of their original roles. For flower meristem identity, the early events of flower development appear to be conserved mechanisms that are retained in a wide variety of plant species. Sequence analysis showed PFL was more closely related to with LFY homolgous proteins of two woody tree species, California sycamore (Platanus racemosa) and black cottonwood (Populus trichocarpa). Phylogenetic analyses of this target gene also revealed that PFL was not as closely related to LFY as suggested by previous systematic study on angiosperm families 28. This is likely due to different rate of divergence among genes under different selection pressures in the genome due to their roles in various developmental pathways. Comparison between PFL and other FLO/LFY homologs showed that the gene structure and splicing sites are highly conserved among the angiosperms. Moreover, PFL shares two highly conserved regions with other FLO/LFY homologs, indicating that these two-conserved regions might relate to the basic functions of FLO/LFY-like proteins. LEAFY protein regulates the expression of downstream floral homeotic genes by binding the sequences in their enhancers 4, 13, 37, 38. The two-conserved regions are involved in DNA binding 39 and the divergent functions between FLO/LFY homologs correspond to substitutions in these two-conserved regions 39.

Proline-rich motifs and acidic-rich motifs are typical for transcriptional activators and may be important for the function of FLO/LFY-like proteins. A proline-rich region near the amino terminus and an acidic central region are found in most angiosperm FLO/LFY-like proteins. Since both of these features are also found in certain types of transcription factors, these two domains are thought to be important for the function of FLO/LFY-like proteins. PFL lacks the proline-rich region and has a short acidic region at corresponding position. Since these two domains are located at variable regions, they might not be functionally significant. Similarly, the FLO/LFY-like proteins from pine, apple and Eucalyptus also lack the proline-rich region 18, 40, 41. Ectopic expression of these FLO/LFY orthologs in Arabidopsis causes the premature conversion of shoots into flowers, as does 35S:LFY18, 40, 41. These regions may be subject to evolutionary changes so that different species formed their own genetic regulation system to adapt to the specific environment.

Our results showed that PFL was expressed in flower buds of all three sex types, likely due to its role as a positive activator of AP3 and AG 3, 4. The marginal difference of PFL expression between hermaphrodite and unisexual flowers revealed by quantitative PCR was not sufficient to suggest a role of PFL as a regulatory element for sex differentiation in papaya. Further functional analyses of PFL and its interaction with other regulatory genes would clarify its role in flower development and sex differentiation in papaya.