Developing pineapple fruit has a small transcriptome dominated by ...

RESEARCH PAPER

Developing pineapple fruit has a small transcriptomedominated by metallothionein

Richard Moyle1, David J. Fairbairn1, Jonni Ripi1, Mark Crowe2 and Jose R. Botella1,*

1 Plant Genetic Engineering Laboratory, Department of Botany, University of Queensland, QLD 4072, Australia2 Institute of Molecular Biosciences, University of Queensland, Australia

Received 5 April 2004; Accepted 20 August 2004

Abstract

In a first step toward understanding the molecular basis

of pineapple fruit development, a sequencing project

was initiated to survey a range of expressed sequences

from green unripe and yellow ripe fruit tissue. A highly

abundant metallothionein transcript was identified dur-

ing library construction, and was estimated to account

for up to 50% of all EST library clones. Library clones

with metallothionein subtracted were sequenced, and

408 unripe green and 1140 ripe yellow edited EST clone

sequences were retrieved. Clone redundancy was high,

with the combined 1548 clone sequences clustering

into just 634 contigs comprising 191 consensus se-

quences and 443 singletons. Half of the EST clone

sequences clustered within 13.5% and 9.3% of contigs

from green unripe and yellow ripe libraries, respect-

ively, indicating that a small subset of genes dominate

the majority of the transcriptome. Furthermore, se-

quence cluster analysis, northern analysis, and func-

tional classification revealed major differences

between genes expressed in the unripe green and ripe

yellow fruit tissues. Abundant genes identified from the

green fruit include a fruit bromelain and a bromelain

inhibitor. Abundant genes identified in the yellow fruit

library include a MADS box gene, and several genes

normally associated with protein synthesis, including

homologues of ribosomal L10 and the translation

factors SUI1 and eIF5A. Both the green unripe and

yellow ripe libraries contained high proportions of

clones associated with oxidative stress responses

and the detoxification of free radicals.

Key words: Ananas comosus, EST, metallothionein, non-

climacteric fruit, pineapple, fruit ripening.

Introduction

Pineapple (Ananas comosus L. Merrill) fruit cropping andprocessing are important horticultural industries in manycountries with tropical climates. In terms of worldwideproduction, pineapple is currently the third most importanttropical fruit after bananas and mangoes. Pineapple iscultivated mainly for fresh or canned fruit and juice, but isalso the only source of bromelain, a complex proteolyticenzyme used in the pharmaceutical market and as a meat-tenderizing agent.

Pineapples are diploid (2n=50) perennial monocotyle-donous plants with a terminal inflorescence consistingof 50–200 individual hermaphrodite flowers that clustertogether. As individual fruits develop from the flowers, theyjoin together forming a cone shaped, compound, juicy,fleshy multiple fruit of approximately 30 cm or more inheight. Pineapple maturity is evaluated by the fruit eyeflatness, the extent of skin yellowing and by aroma. It takes3–5 months for pineapple fruit to reach maturity andpineapple fruit quality is highest if the fruit matures onthe plant. Pineapples harvested prematurely do not continueto ripen or sweeten, as there are no starch reserves in thefruit to be converted to sugar.

In general, fleshy fruits are divided in two large groups,climacteric and non-climacteric, based upon the presence orabsence of an autocatalytic ethylene burst during ripening.Pineapple is an example of a non-climacteric fruit. It doesnot have an autocatalytic ethylene burst during ripening,and exogenous application of ethylene does not rapidlyaccelerate fruit ripening. In climacteric fruits, ethyleneplays a crucial role in the control of the ripening process, byregulating the transcription of a large number of genes. Theethylene biosynthetic pathway is well established (Yang andHoffman, 1984) and the expression patterns of themain ethyl-ene biosynthetic genes, 1-aminocyclopropane-1-carboxylic

* To whom correspondence should be addressed. Fax: +61 7 33651699. E-mail: [email protected]

Journal of Experimental Botany, Vol. 56, No. 409, ª Society for Experimental Biology 2004; all rights reserved

Journal of Experimental Botany, Vol. 56, No. 409, pp. 101–112, January 2005

doi:10.1093/jxb/eri015 Advance Access publication 1 November, 2004

Downloaded from https://academic.oup.com/jxb/article-abstract/56/409/101/467467by gueston 04 April 2018

acid (ACC) synthase (ACS) and ACC oxidase (ACO),have been well characterized (Giovannoni, 2001). Con-versely, almost nothing is known of ripening processesin non-climacteric fruit, although ripening-related genesfrom non-climacteric fruits have been reported with grape,Citrus unshiu, and strawberry among those most studied.Davies and Robinson (2000) undertook the isolation of 17ripening-associated cDNAs by differential screening ofa shiraz grape library. An EST sequencing project on grapebuds has also been reported, but the library clones wereisolated to investigate gene expression leading up to budrelease from dormancy (Pacey-Miller et al., 2003). A small-scale sequencing project from Citrus unshiu resulted in theisolation of 297 EST sequences (Moriguchi et al., 1998). Inaddition, an EST sequencing project has been reported fromstrawberry, resulting in the isolation of 1100 clone sequen-ces (Aharoni et al., 2002; Aharoni and O’Connell, 2002).Microarray analysis of the strawberry clones identified112 genes that were ripening-regulated (Aharoni andO’Connell, 2002). A number of previously identified genesinvolved in cell wall softening and pigmentation weredifferentially expressed during strawberry fruit ripening.Oxidative stress-related genes were also observed as up-regulated in ripening strawberry fruits, and it has beensuggested that strawberry fruits may contain a transcrip-tional programme responsive to oxidative stress inducedduring ripening (Aharoni and O’Connell, 2002). Despitethese efforts to isolate ripening-related clone sequences, therelatively small number of genes available for study hashampered the identification of genes controlling the ripen-ing processes in non-climacteric fruits.The development and automation of high-throughput

sequencing technology provides an opportunity to studygenomes at both the DNA and RNA level. Genome-sequencing projects can lead to the identification of thecomplete catalogue of genes within an organism, independ-ent of gene expression levels. Alternatively, large-scale end-sequencing of cDNA library clones generates expressedsequence tags (ESTs) that can be used to obtain a snap-shotof gene expression in a particular tissue and stage ofdevelopment at a fraction of the cost of genome sequencing(Adams et al., 1993). EST sequencing has proved a popularand cost-effective method of isolating vast collections ofcoding sequences from a plethora of species. As at February2004, the GenBank dbEST (database of expressed sequencetags) contained 20 039 613 EST sequences from 611organisms (Boguski et al., 1993). Substantial numbersof EST sequences have been generated from a numberof plants, with plant species comprising 12 of the 30 mostabundant EST collections. In addition, comparing ESTclones generated from different tissues at different stagesof development provides a method for comparing relativegene expression and identifying subsets of genes andcorresponding functions that probably play vital roles withinspecific tissue types and stages of development.

Despite the economic importance of pineapple as a crop,surprisingly little research has been undertaken to under-stand the molecular basis of pineapple fruit development.A search of the GenBank nucleotide sequence database(Benson et al., 2003) yielded only 51 sequence entries forAnanas comosus (as of February 2004), the majority beingisolated from root and vegetative tissues. Almost half ofthese sequence entries resulted from a study reporting 24genes isolated from pineapple roots by differential display(Neuteboom et al., 2002). This study reports the first ESTsequencing project from pineapple, resulting in the isolationof 1548 fruit EST clone sequences. As far as is known, this isthe first report of EST sequencing on any pineapple tissue orindeed on any bromeliad. The isolation of a range of pine-apple gene sequences fromgreen unripe and yellow ripe fruittissue is a first step toward building a valuable resource thatwill help in understanding the molecular basis of fruit devel-opment and will aid pineapple improvement programmes.

Materials and methods

Plant material

Commercial field-grown pineapples (A. comosus L. cv. SmoothCayenne) were harvested early in the morning and graded accordingto ripening appearance. Selected pineapple fruits were separated intotop, middle, and bottom thirds, with the skin and pith tissues re-moved. The remaining fruit flesh was snap frozen in liquid nitrogen,pulverized, and stored at �80 8C.

cDNA library construction and sequencing

PolyA RNA was isolated from the middle third of pineapple 4 (Fig.1), a mature size but unripe green pineapple, and pineapple 9 (Fig. 1),a mature ripe yellow pineapple, using the oligo dT dynabead kit(Dynal). PolyA RNA was transcribed into cDNA and subsequentlyamplified using the SMART cDNA synthesis kit (Clontech Labora-tories, Palo Alto, CA). SMART cDNA synthesis utilizes a mechanismto enhance the proportion of cDNA containing full-length codingsequences, allows for PCR amplification of the synthesized cDNA,and introduces SfiI restriction enzyme sites for directional cloning.The resulting cDNA was size fractionated by electrophoresis through0.8% agarose gels. cDNA fragments running higher than approxi-mately 700 bp were recovered and purified using the GeneClean II kit(Bio101). The resulting cDNA pools were digested with SfiI re-striction enzyme and directionally cloned into pBS-SfiI. pBS-SfiI wasengineered by inserting a linker containing SfiI directional cloningsites between the EcoRI and XhoI restriction enzyme sites ofpBLUESCRIPT (Stratagene). Transformed colonies were picked at

Fig. 1. Commercial field-grown pineapples harvested and sorted 1–9according to visible ripening characteristics.

102 Moyle et al.


random and placed, using a grid pattern, on duplicate 25 cm2 LBampicillin plates, one of which was subsequently screened at highstringency (0.13 SSC/0.1%SDS at 65 8C) with a radiolabelled metal-lothionein clone Ac180 using a colony lift procedure (Sambrooket al., 1989). Colonies not hybridizing to the metallothionein Ac180probe were picked from the duplicate plate for growth in 96 formatdeep well plates. A robotic platform (Biomek 9600) was used toreplicate the 96 well clone plates into multiple glycerol stock bankcollections. Cultures were grown from glycerol stock plates and 59end sequencing was conducted on purified plasmid preparations.

Bioinformatics

Raw sequences were manually edited for sequence quality, trimmedof plasmid contaminant and polyA tail in the sequence viewer pro-gram Chromas v2.13 (Technelysium). Clones yielding less than 150bp of insert sequence were eliminated from the data set. All EST se-quences were submitted into the GenBank dbEST. Edited sequenceswere compiled into contigs or singletons using SeqMan sequenceassembly software (DNASTAR Inc. Madison, USA), and key par-ameters of minimum 90% match over at least 45 bp overlap. Com-parison analysis of the contig consensus sequences was conductedwith the advanced basic alignment search tool, BLASTX server(Altschul et al., 1990) and the National Center for BiotechnologicalInformation (www.ncbi.nih.gov) non-redundant protein database andresulting information parsed from nearest neighbour hits. Contigconsensus sequences returning hits better than the 10E�20 E-valuecutoff were BLASTX searched against MATDB (mips.gsf.de/proj/thal/db/index.html) to retrieve putative functional classifications.Functional classes and subclasses retrieved were subjected to manualcuration.

Northern analysis

Total RNA was extracted from meristem, young leaves (shorter than3 cm in length), mature-sized green leaves, the middle third ofpineapples 1–5 and 7–9 (Fig. 1) using a pineapple RNA extractionprotocol (Cazzonelli et al., 1998). Total RNA was isolated from roottissue using TRIzol (Invitrogen). Five microgram aliquots of totalRNA from pineapple 1–5, 7–9, meristem, young leaf, old leaf, androots were size fractionated through two 1% agarose gels, andtransferred to HyBond N+ membranes (Amersham) by capillarytransfer (Sambrook et al., 1989). The duplicate membranes werecross-linked by UV and prehybridized in Church and Gilbert buffer(Church and Gilbert, 1984) at 50 8C. Radiolabelled probes weresynthesized using the Strip EZ DNA kit (Ambion) and hybridized tomembranes at 55–65 8C. Membranes were progressively washed toa high stringency of 0.1% SSC/0.1% SDS at 60 8C prior to exposurein phosphoimager cassettes. Membranes were stripped and reprobedaccording to the Strip EZ DNA kit instructions (Ambion).

Real-time PCR

For real-time quantitative PCR experiments (RT-qPCR), 1 lg of totalRNA was reverse transcribed using Superscript III (Invitrogen)following the manufacturer’s instructions. The resulting cDNA wassubsequently taken up in a volume of 250 ll and the fruit andmeristem samples further diluted 100-fold. Pineapple metallothioneinAc180 and b-actin target genes were amplified using gene-specificprimers designed from the coding sequence, and in the case of Ac180over an RNA splice junction, using the Primer Express 1.5 software(Applied Biosystems, Foster City, CA). The primer pairs for Ac180were 59-TTCTGCCTCATTTATTGGCTCAT-39 and 59-AGCCGT-TTCCCTTCTTCACA-39 and for b-actin 59-ATGGAAGCTGCG-GGTATTCA-39 and 59-CCACCACTGAGCACGATGTT-39. Bothprimers pairs amplified 101 bp amplicons with similar melting tem-peratures of 81 8C. RT-qPCR was carried out using the ABI PRISM7700 sequence detector and SYBR Green Master mix (Applied

Biosystems) using primers at a final concentration of 0.1 lM each and2.5 ll (the equivalent of 10 or 0.1 ng total RNA) of cDNA astemplate. PCR-cycling conditions comprised an initial polymeraseactivation step at 95 8C for 10 min, followed by 45 cycles at 95 8Cfor 15 s and 60 8C for 1 min. Real-time DNA amplification wasmonitored and analysed using the Sequence Detector 1.7 program(Applied Biosystems). Threshold cycle data were exported toEXCEL and the quantification of Ac180 gene expression relative tothe b-actin reference gene was calculated using the Pfaffl equation(Pfaffl, 2001). This equation takes into account differing primerefficiencies. Cycle threshold values were corrected for genomic DNAcontamination in the RNA sample. The young leaf sample was usedas a calibrator to allow comparisons across samples to be made.RT-PCR experiments were repeated on three separate occasions induplicate.

Results and discussion

cDNA library construction

PCR amplification of both unripe green and ripe yellowfruit cDNA and subsequent fractionation through an aga-rose gel revealed an intense band within the green andyellow fruit cDNA pools. Subsequent sequence analysis ona test plate of 96 randomly picked library colonies resultedin the identification of 40 clones as a member of the metal-lothionein gene family. Subsequent colonies were pre-screened with radiolabelled Ac180 to reduce redundancydue to the abundant metallothionein clone. One problemtraditionally associated with EST library construction is theoverabundance of short and truncated EST fragments due toincomplete reverse-transcription and the ligation bias to-ward small inserts. This can result in a disproportionatelyhigh number of false undiscovered ‘novel’ sequences dueto insufficient coding sequence to establish identity accur-ately by homology to existing coding sequences. To minim-ize the overabundance of very short inserts and enhancethe proportion of full-length and large EST fragments in thelibraries, the amplified cDNA were size-fractionated andthose running at less than approximately 700 bp werediscarded. Although the resulting cDNA pools still containfragments of less the 700 bp, it is possible that this size-fractionation strategy may subsequently result in the under-representation of very short genes in the cDNA libraries.Restriction mapping of 24 randomly picked clones revealedan estimated average EST insert size of approximately1.2 kb for each library. A total of 480 unripe green pine-apple and 1536 ripe yellow pineapple clones were sub-sequently picked for sequencing.

Sequence analysis

Raw sequence quality was consistently high, with over 80%of clones returning an average Phred20 score of more than700 bp. After editing raw sequences for quality anddiscarding insert sequences of less than 150 bp, a total of408 green unripe and 1140 yellow ripe pineapple EST se-quences were retrieved at an average read length of 785 bp.

Gene expression in pineapple fruits 103


http://www.ncbi.nih.gov

All edited EST sequences have been submitted to theGenBank dbEST.The large read lengths obtained compare favourably with

other recently reported EST sequence projects. For example,Bhalerao et al. (2003) report average read lengths of 349 bpand 355 bp from autumn leaf and young leaf libraries, whileRonning et al. (2003) report an overall average read lengthof 525 bp from various potato EST libraries. Maximum readlengths are desirable, as short truncated EST clones con-taining 39 untranslated sequence with little or no codingsequence cannot be accurately identified by homology toprotein coding sequences in public databases. The lengthof the sequences generated also impacts on EST clusteranalysis, as the longer the EST sequences are, the morelikely that multiple EST clones encoded from a single genewill overlap during contig assembly.

Cluster analysis reveals high redundancy withinthe pineapple fruit EST libraries

All edited EST sequences were clustered using SeqMansequence assembly software (DNASTAR). The 1548 se-quences clustered into 634 contigs. The unripe green andripe yellow sequences clustered with 62% and 54% re-dundancy, respectively. Further analysis of all contigscontaining two or more green and/or yellow fruit clonesrevealed that 1105 ESTs (71% of all sequences) clusterwithin the 191 contigs, whereas just 443 ESTs (29%)represent singletons. To investigate the high level ofredundancy, contigs were listed by clone abundance—fromcontigs containing the largest number of EST clones tothose containing a single EST clone from either unripegreen or ripe yellow libraries. The resulting data wereconverted to percentages to enable a direct comparisonbetween the green unripe and yellow ripe cumulativeabundance of EST clones (Fig. 2). The results reveal 50%of EST clone sequences cluster within just 13.5% and 9.3%

of contigs from green unripe and yellow ripe libraries,respectively. Thus a small subset of genes appears todominate the majority of the transcriptome in both unripegreen and ripe yellow fruit. However, as library construc-tion involved an amplification step, the possibility thatamplification bias may have contributed to the highredundancies cannot be discounted.

Bioinformatics

Contig consensus sequences were exported in FASTAformat and batch BLASTX searches performed againstthe GenBank non-redundant database. A stringent BlastXE-value cut-off of 10E�20 was chosen to ensure that theannotations were based only on genes with a high degree ofsimilarity to this study’s cDNA clones (Hu et al., 2003;Whitfield et al., 2003). A semi-automated process ofparsing BLAST hits and manually curating the putativeannotations resulted in a spreadsheet of information in-cluding cloneID, contig number, number of clones in eachcontig, nearest BLASTX neighbour, accession number ofmatch, length of match, percentage similarity, and putativeannotation (supplementary data at JXB online).

Approximately 17% of all sequences did not havesignificant homology to coding sequences in the GenBanknon-redundant database and therefore could not be anno-tated by similarity. This ‘undiscovered’ sequence subsetwould probably contain some novel coding sequences thathave not previously been discovered. However, it is un-realistic to suggest that all of the undiscovered EST subsetencode for novel proteins with no significant homology tothose in public sequence databases. Instead, many clonesfalling under the undiscovered sequence subset are likely tocontain insufficient coding sequence to accurately assign anannotation based on homology. It is also possible that someof these unidentified sequences may not encode proteins,but instead may function as RNA molecules. Furthermore,some of the undiscovered sequence ESTs may contain shortstretches of protein coding sequence homologous to pro-teins in the non-redundant database but, due to a small openreading frame or short length of overlapping codingsequence, do not produce an expected value better thenthis cutoff limit of 10E�20.

The majority (83%) of EST clones were putativelyannotated by similarity to coding sequences in the Gen-Bank non-redundant database with over 55% of all ESTclones sharing 80% or higher similarity with existingsequences. Approximately 17% of clones had similarityof between 60–80%, and less than 5% shared lower than60% similarity to known coding sequences. Contig con-sensus sequences returning a valid BLASTX hit fromGenBank were individually searched against the MIPSMATDB database (Schoof et al., 2002). Manual curationof MIPS MATDB functional classification resulted inthe assignment of major functional classes and subclasses

Fig. 2. The cumulative abundance of EST clone sequences from unripegreen and ripe yellow fruit libraries was calculated by sorting contigscontaining the most abundant clones to singletons. The cumulativenumbers of clones and contigs are displayed as a percentage of the unripegreen or ripe yellow clone and contig totals.

104 Moyle et al.


to 1184 EST clones generated from the green unripe andyellow ripe pineapple libraries.

Metallothionein clone Ac180 is very highlyexpressed in pineapple fruits

Fractionating an aliquot of the unripe green and ripe yellowpineapple cDNA through an agarose gel revealed thepresence of a very highly abundant transcript of approxi-mately 500–600 bp in length. Despite gel purification toenrich for cDNA greater than 700 bp, sequence analysis ofan initial test plate of 96 clones picked resulted in theidentification of 40 clones as a member of the metal-lothionein gene family, subsequently named Ac180. Ex-cluding the polyA tail, the metallothionein EST clones wereapproximately 500 bp in length, matching the estimatedsize of the abundant transcript in the amplified fruit cDNApools. When subsequent library colonies were screenedusing radiolabelled Ac180 as a probe, up to 50% of the fruitlibrary colonies screened hybridized strongly. Reports ofhighly abundant clones isolated in plant EST sequencingprojects include 14% of EST sequences from a poplaryoung leaf library encoding the small subunit of Rubisco(Bhalerao et al., 2003) and 20% of EST sequences froma poplar cambium transition to dormancy library encodinga major storage protein (Schrader et al., 2004). Further-more, a small-scale EST sequencing project from a Citrusunshiu fruit library identified 20% of clone sequences asencoding for metallothionein (Moriguchi et al., 1998).However, as far as is known, there have been no reports ofa single gene product dominating a transcriptome to theextent that the metallothionein clone Ac180 appears todominate the unripe green and ripe yellow pineapple fruittissue. However, as the libraries were constructed fromPCR amplified cDNA, the possibility that the abundance ofthe metallothionein clone Ac180 is exaggerated due toamplification bias cannot be discounted.

The metallothionein Ac180 expression was further ana-lysed by quantitative real-time PCR using first strandcDNA synthesized from the middle third of the ninepineapples harvested (Fig. 1). b-actin (Ac42) was used asa normalization control (Fig. 3). The expression of themetallothionein Ac180 gene increased during fruit devel-opment and was highest in mature fruit. Expression washighest in fruit tissue, but high levels relative to b-actinwere also observed in root and leaf tissue, particularly in theolder leaf tissue. Metallothionein Ac180 expression wascalculated to be more than 50-fold higher than b-actin(Ac42) in the ripening fruit samples 7–9, confirmingmetallothionein transcripts are highly abundant in ripepineapple fruit.

A second metallothionein gene, Ac181, was also dis-covered in the yellow fruit library with 17 copies isolated.No copies of Ac181 were isolated from the green fruitlibrary. The two metallothionein sequences share approxi-mately 80% identity at the amino acid and nucleotide level.

Although Ac180 appears to be considerably more abundantthan Ac181, Northern analysis revealed both genes exhibita similar expression pattern across the range of ripeningpineapple fruits sampled, and are similarly up-regulated inolder leaves (Fig. 4).

Metallothioneins are small cysteine-rich proteins re-quired for heavy metal tolerance in animals and fungi. Inplants, metallothionein genes are up-regulated in responseto heavy metal stress (van Hoof et al., 2001), natural andinduced leaf senescence (Buchanan-Wollaston, 1994; Chenet al., 2003), ethylene-induced abscission (Coupe et al.,1995), drought and salt stress (Oztur et al., 2002), lightstress (Dunaeva and Adamska, 2001), nematode infection(Potenza et al., 2001), pathogen-induced necrosis (Buttet al., 1998), abscisic acid (Chatthai et al., 1997; Reynoldsand Crawford, 1996), wounding (Choi et al., 1996),sucrose starvation and heat shock (Hsieh et al., 1996),and cold stress (Reid and Ross, 1997). Metallothioneingenes up-regulated during fruit development have beenisolated from climacteric fruits such as banana (Clendennenand May, 1997), apple (Reid and Ross, 1997), and kiwifruit(Ledger and Gardner, 1994) and non-climacteric fruit suchas grape (Davies and Robinson, 2000), Citrus unshiu(Moriguchi et al., 1998), and strawberry (Nam et al., 1999).Two kiwifruit metallothionein-like clones were differ-entially expressed during fruit development. pKIWI4 wasup-regulated early while pKIWI3 was up-regulated late infruit development. Similarly, two banana metallothionein-like clones were found to be differentially expressed duringfruit ripening, pBAN3-23 expression decreased during fruitripening while pBAN3-6 expression peaked during a laterstage of fruit ripening. For Citrus unshiu, 62 metall-othionein clones were isolated from the 297 clones analysedfrom a mature fruit library. A comparison of the expression

Fig. 3. Quantitative real-time PCR analysis of metallothionein cloneAc180 gene expression in various pineapple tissues.



of the mature fruit abundant metallothionein clone,citMT45, and a metallothionein clone previously isolatedfrom developing fruit, citMT36, again revealed a differentialexpression pattern. CitMT36 was detected in all organstested whereas citMT45 was strongly up-regulated in fruitfrom 70 d after fertilization.It is not yet clear why metallothioneins are up-regulated

during fruit development, or how they function in plants.One potential link between the reported up-regulation ofmetallothionein during various stress responses and theabundance during pineapple fruit ripening and in old leavescould be oxidative stress. Reactive oxygen species accu-mulate during stress conditions such as wounding, abscis-sion, necrosis, pathogen infection, drought and salt stress,and multiple genes involved in oxidative stress responsesand the detoxification of free radicals have been isolatedfrom these fruit libraries. As northern analysis revealed thata number of other abundant genes share a similar pattern ofexpression to the Ac180 and Ac181 metallothionein genes(Fig. 4), it is therefore possible that a suite of genes are up-regulated during fruit ripening and in ageing leaves bycommon causal and regulatory mechanisms.

Unripe green and ripe yellow pineapple fruits expressa high proportion of oxidative stress-related genes

Functional classifications were assigned to contig con-sensus sequences that had an E-value better than 10E�20,using a manually curated system based on MIPS MATDBfunctional classification. Results obtained from contig con-sensus sequences were subsequently assigned to the indi-vidual EST sequences clustering within each contig. The

highly abundant metallothionein gene Ac180 was excludedfrom the functional classification analysis.

Functional class distribution indicated that 10% of unripegreen and 6% of ripe yellow clones encode proteins fallingwithin the cell rescue, defence, and virulence major class(Fig. 5). This represents an unusually high proportion ofclones, although similar percentages of cell rescue, defence,and virulence clones have been found in senescing leaves(Bhalerao et al., 2003). The majority of the clones clusteredwithin the subclasses representing detoxification and stressresponses, particularly oxidative stress and the detoxifica-tion of reactive oxygen species and free radicals. Indices ofoxidative processes have been measured in tomato fruits(Jimenez et al., 2002). Hydrogen peroxide content, lipidperoxidation, and protein oxidation were all found to in-crease at the breaker stage (fully developed green fruit justprior to turning red). Six EST copies of a (S)-2-hydroxy-acidgene (Ac3388) were identified from the green pineapplefruit. (S)-2-hydroxy-acid oxidase (glycolate oxidase) catal-yses a reaction resulting in the production of H2O2. It isa known source of peroxide produced during photorespira-tion and is enhanced during abiotic stress responses (Mittler,2002). As (S)-2-hydroxy-acid oxidase appears to be abun-dant in the green unripe fruit, it is speculated that it isa significant source of peroxide, contributing to a state ofoxidative stress within the fruit. Peroxiredoxin and glutathi-one transferase are two genes involved in the detoxificationof H2O2. Peroxiredoxin (Ac21) and glutathione transferase(Ac135) were among the most abundant gene sequencesfrom the green fruit library, with six EST copies isolatedeach (Table 1). Three copies of peroxiredoxin (Ac21) werealso isolated from the yellow fruit library. Clones encodingadditional glutathione transferase genes were isolated fromthe green fruit library (three copies of Ac134 and two copiesof Ac54), and the yellow fruit library (three copies ofAc55, three copies of Ac2038, and two copies of Ac2248).Glutathione transferases are up-regulated in many plants inresponse to a range of stress conditions (Conklin and Last,1995;Marrs andWalbot, 1997;Mittova et al., 2003;Moons,2003; Ruzsa et al., 1999; Skipsey et al., 1997; Thom et al.,2002) and have been reported as up-regulated during fruitripening in strawberry (Aharoni et al., 2002; Aharoni andO’Connell, 2002). In plants, the functions of glutathionetransferase include detoxification of xenobiotics and thestabilization of flavanoids (Dixon et al., 2002). In addition,stress-inducible glutathione transferases have been shown toprotect plants from oxidative injury by functioning asglutathione peroxidases (Cummins et al., 1999; Roxaset al., 1997). It is probable that some glutathione transferasegenes are up-regulated by the general oxidative stress causedduring a range of stress conditions (Dixon et al., 2002), andthat oxidative stress also causes up-regulation of glutathionetransferases in fruits. Similarly, plant peroxiredoxins are up-regulated during stress. They function as antioxidant pro-teins, detoxifying various peroxide substrates (Dietz, 2003;

Fig. 4. Northern hybridization analysis of total RNA with abundantpineapple fruit clone probes. RNA was extracted from the middle third ofpineapples 1–9 (Fig. 1), fruit meristem (M), young leaf (YL), old leaf(OL), and root tissues. Duplicate membranes were prepared and probedwith EST clones from the contig clusters Ac124 (fruit bromelain), Ac122(bromelain inhibitor), Ac180 (metallothionein), Ac181 (metallothionein),Ac146 (MADS box protein), Ac3085 (S-like ribonuclease RNS2), Ac183(mitochondrial ADP/ATP carrier protein), Ac75 (protein translationfactor SUI1), and Ac8 (60S ribosomal protein L10). The northern wasalso probed with pineapple actin as a control for observing loadingdifferences.

106 Moyle et al.


Hofmann et al., 2002). In addition, two ascorbate peroxidasegenes, Ac50 (two copies) and Ac3096 (one copy), onemonodehydroascorbate reductase gene (Ac3333), and a glu-taredoxin gene (Ac37) were isolated from the yellow fruitlibrary. Ascorbate peroxidase and monodehydroascorbatereductase also function to scavenge H2O2 (Mittler, 2002)while glutaredoxin is described as an electron donor toperoxiredoxin (Rouhier et al., 2003). Three superoxidedismutase genes, Ac119 (one copy), Ac3200 (one copy)and Ac3219 (three copies) were isolated from the yellowfruit library. One copy of Ac3219 was also isolated fromthe green fruit library. Superoxide dismutase functions as

a scavenger of O�2 (Alscher et al., 2002). The relative

abundance of antioxidant proteins and enzymes of the ROSscavenging pathways in pineapple fruits supports the viewthat ripening is anoxidative phenomenon requiring a balancebetween the production of ROS and their removal byantioxidant systems.

The green unripe and yellow ripe pineapple fruitexpress different subsets of abundant genes

Unripe green and ripe yellow fruit clones were sorted bycontig abundance for direct comparison. The fruit librariesshared few abundant genes in common, with most abundant

Fig. 5. The distribution of EST clones by modified MIPS based functional classification for ripe yellow (A) and unripe green (B) fruit.



gene species being preferentially isolated from either theunripe green or ripe yellow fruit (Tables 1, 2). Abundantclones preferentially isolated from green fruit included abromelain inhibitor (Ac124) and a fruit bromelain (Ac122).The bromelain inhibitor was the most abundant clone typefrom the green fruit library with 52 clones isolated. Northernanalysis confirmed these two genes are down-regulatedduring fruit ripening (Fig. 4). Why bromelain and a bro-melain inhibitor are simultaneously expressed in develop-ing fruits is unclear.Cinnamyl alcohol dehydrogenase (CAD) (Ac176) clones

were isolated from both the green and yellow fruit libraries,but were more abundant in the green than the yellow fruitlibrary A CAD gene was reported to be up-regulated duringfruit ripening in strawberry (Aharoni et al., 2002; Blanco-Portales et al., 2002). CAD up-regulation in strawberryfruits has been suggested to play a role in the interconversionof aldehydes and alcohols implicated in flavour as well asthe lignification of vascular elements (Aharoni et al., 2002).Eight copies of a putative PHD transcription factor

(Ac3400) were isolated from the green fruit library, whilenone were identified among the yellow fruit library clones.It is possible that this putative transcription factor directschanges in gene expression during the course of fruitripening. Conversely, a MADS box transcription factorhomologue (Ac146) was the most abundant clone typeisolated from the subtracted yellow fruit library, but was notisolated from the green fruit library. MADS-box genesencode for a large family of transcription factor proteinsinvolved in regulating various aspects of plant develop-ment, particularly flowering and organ or meristem identity.Analysis of the Arabidopsis genome revealed 107 genesencode MADS box proteins, although the function of mostremain unknown (Parenicova et al., 2003). Northern

analysis of developing pineapple fruit confirmed thatAc146 expression increases during ripening and is mostabundant in yellow ripe fruit tissue (Fig. 4). No expressionwas detected in meristem, leaf or root tissue. It is somewhatunusual for a transcription factor to be among the mosthighly expressed genes, and, therefore, it is hypothesizedthat the differential abundance of this MADS box geneprobably plays a significant role in regulating differentialgene expression during fruit ripening.

Interestingly, a protein translation factor SUI1, a eukary-otic translation initiation factor 5A isoform (eIF 5A), a 60Sribosomal protein L10, and a putative nascent polypeptide-associated complex alpha chainwere all abundant among theyellow fruit library clones yet absent in green fruit libraryclones. Why these translation-related genes are strongly up-regulated in the ripe yellow fruit is not immediately obviousas one would expect protein synthesis genes to be relativelyevenly expressed throughout the plant. However, recentresearch in non-plant systems has determined that L10,SUI1, and eRF5A perform alternative functions in additionto protein synthesis. The pineapple 60S ribosomal L10 geneis a homologue of the humanQM ribosomal protein (Dowdyet al., 1991) and the Jun interaction factor Jif-1 from chicken(Monteclaro and Vogt, 1993). Both QM ribosomal proteinand Jif-1 function as tumour suppressor genes (Dowdy et al.,1991; Monteclaro and Vogt, 1993), and Jif-1 has been dem-onstrated specifically to bind to c-Jun, causing the inhibitionof a certain AP-1 transcription factor function. SUI1 transla-tion initiation factors have been implicated in the nonsense-mediated decay pathway (NMD) in addition to their role inrecognition of the AUG codon during translation initiation(Cui et al., 1999). The NMD pathway is an mRNA surveil-lance pathway, responsible for recognizing aberrant mRNAtranscriptswith premature stop codons and accelerating their

Table 1. The 20 most abundant EST clones isolated from green unripe fruit

Putative annotation Pineapple clonecluster ID

ESTs from unripegreen fruit

ESTs from ripeyellow fruit

Fold enrichment ingreen unripe fruit

Bromelain inhibitor Ac124 52 3 –Mitochondrial ADP/ATP carrier protein Ac183 23 28 2Unknown protein Ac101 15 1 42S-like ribonuclease RNS2 Ac3085 15 39 1Undiscovered sequence Ac1885 9 27 1Similar to transcription factor IIB Ac115 8 22 1Fruit bromelain Ac122 8 1 22PHD finger transcription factor, putative Ac3400 7 0 –GAST-like gene product Ac35 7 0 –Glutathione transferase Ac135 6 0 –Putative cinnamyl-alcohol dehydrogenase Ac176 6 4 4Nuclear transport factor 2, putative Ac187 6 0 –Peroxiredoxin Ac21 6 3 6Putative dehydration-responsive proteinRD22 precursor

Ac3144 6 1 17

Putative (S)-2-hydroxy-acid oxidase Ac3388 6 0 –Ubiquitin conjugating enzyme Ac153 5 3 5Unknown protein Ac187 5 0 –Unknown protein Ac105 4 0 –Glyceraldehyde 3-phosphate dehydrogenase Ac136 4 1 11Putative fructose-bisphosphate aldolase Ac3389 4 0 –

108 Moyle et al.


subsequent decay (Hillman et al., 2004;Wilusz et al., 2001).Hypusine modified eIF5A is the active form of the proteinand is thought to act as a nucleo-cytoplasmic transporter ofmRNA from the nucleus to the cytoplasm, in addition totranslation initiation factor activity (Bevec and Hauber,1997; Rosorius et al., 1999). The eIF5A mediated trans-location of mRNAmay occur for only certain RNA species,as hypusine-modified eIF5A was found to bind RNA ina sequence-specific manner (Xu and Chen, 2001). Yeastcells, in which DHS has been inactivated are incapable ofdividing, prompted the proposal that hypusine-containingeIF-5A facilitates translation of a subset of mRNAs requiredfor cell division. In Arabidopsis, eIF5A expression peaksin rosette leaves at days 14 and 35 after emergence, coinci-dent with the onset of bolting and the later stages of leafsenescence (Wang et al., 2003). In tomato, eIF-5A mRNAshow a parallel increase in abundance in senescing flowers,senescing fruit, and osmotically stressed tomato leavesexhibiting programmed cell death (Wang et al., 2001).Northern analysis of the pineapple eIF5A shows an up-regulation of the eIF5A during fruit ripening and in olderleaves, consistent with the expression data reported forArabidopsis and tomato eIF5A (Wang et al., 2001, 2003).

Two highly abundant undiscovered sequences werepresent in the ripe yellow fruit library, but absent in greenfruit library. The undiscovered sequence Ac81 representedby 44 ripe yellow fruit clones appeared to have a very shortopen reading frame encoding 64 amino acids. Althoughannotated as an undiscovered sequence, there are twohypothetical genes from rice that share approximately64% similarity. The short open reading frame of only 64amino acids resulted in an expected value falling below the10E�20 threshold, leading to the annotation as an undis-covered sequence. The undiscovered sequence Ac157represented by 14 ripe yellow fruit clones is just 283 bp

in length and contains no significant open reading frame,the longest starting at the fifth ATG and being just 16 aminoacids in length. As 13 of the 14 ESTs for this gene begin atthe same base and there were nine variations in the polyAaddition base, it is believed that the full-length clonesequence for this transcript has been isolated rather thanpartial sequence clones containing only a 39 untranslatedsequence. Together with the lack of a reasonable open-reading frame, it is hypothesized that Ac157 may functionas an RNA molecule rather than encode for a protein.

Abundant genes found in common and in relativelyproportionate numbers between the unripe green and ripeyellow libraries included an S-like ribonuclease, mitochon-drial ADP/ATP carrier protein, and a putative transcriptionfactor IIB. However, northern analysis revealed that theS-like ribonuclease and mitochondrial ADP/ATP carrierprotein genes are up-regulated during ripening, with highestexpression in the ripe yellow fruit. S-like ribonucleaseshave been isolated from many plant species and charac-terized as being up-regulated during processes includingsenescence, phosphate starvation, wounding, and droughtstress (Liang et al., 2002; Ma and Oliveira, 2000; Salekdehet al., 2002; Taylor et al., 1993). It has been suggested thatS-like ribonucleases are up-regulated during senescenceand in response to phosphate starvation to play a role in theremobilization of phosphate.

Comparing clone distribution by functional classreveals major differences between the unripegreen and ripe yellow fruit libraries

A number of differences between the unripe green and ripeyellow fruit libraries were observed after clustering clonesby assigned functional classification (Fig. 5). The largestdifference was in the protein synthesis category, where10% of ripe yellow fruit clones and 2% of unripe green fruit

Table 2. The 20 most abundant EST clones isolated from yellow ripened fruit

Putative Annotation Pineapple clonecluster ID

ESTs in ripeyellow fruit

ESTs in unripegreen fruit

Fold enrichment inyellow ripe fruit

MADS box protein Ac146 45 0 –Undiscovered sequence Ac81 44 1 16S-like ribonuclease RNS2 Ac3085 39 15 1Mitochondrial ADP/ATP carrier protein Ac183 28 23 0.4Undiscovered sequence Ac1885 27 9 1Similar to transcription factor IIB Ac115 22 8 1Protein translation factor SUI1 Ac75 20 0 –Metallothionein Ac181 17 0 –60S ribosomal protein L10 Ac8 17 0 –Undiscovered sequence Ac157 14 0 –Glycine decarboxylase complex H-protein Ac138 12 1 4Eukaryotic translation initiation factor 5A isoform 1 Ac32 11 0 –Putative uricase Ac109 9 1 3Putative carboxymethylenebutenolidase Ac175 8 1 3Submergence induced protein 2A Ac117 8 0 –Putative auxin-repressed protein Ac172 8 0 –Putative nascent polypeptide-associated complex alpha chain Ac185 8 0 –Similar to protein disulphide isomerase Ac23 7 4 0.6Calmodulin 4 Ac127 7 1 3Cytochrome b5 Ac131 7 0 –



clones clustered. The difference is largely attributable tothree abundant gene products, a 60S ribosomal protein L10,an abundant SUI1 translation factor, and isoforms ofa eukaryotic translation initiation factor 5A.Nine per cent of unripe green fruit and 4% of ripe yellow

fruit clones fell within the cellular transport functionalclass. A single abundant gene encoding a mitochondrialADP/ATP carrier protein was largely responsible for thedifference, being twice as abundant in green fruit than inyellow fruit. Conversely, the transport facilitation classrepresented 2% of ripe yellow fruit clones and 1% of unripegreen fruit clones. The difference is largely due to the eightclones isolated from ripe yellow fruit encoding varioussubunits of H+ transporting vacuolar ATPase, comparedwith just one clone isolated from unripe green fruit.The transcription class contained 9% of clones from ripe

yellow fruit and 5% of unripe green fruit clones. Thepercentage difference can be accounted for by the presenceof the abundant MADS box gene AC146 from the ripeyellow library that was not found in the green fruit library.Interestingly, a closer inspection revealed no transcriptionfactor genes were isolated in common from both the unripegreen and ripe yellow libraries. The yellow fruit library con-tained three differentMADS box genes, a scarecrow-like pro-tein clone, a bHLH transcription factor, and two different zincfinger proteins. The green unripe library contained sevencopies of a putative PHDfinger transcription factor protein, anAPETALA2 homologue, and a G-box binding factor.The cellular communication and signal transduction

functional class contained approximately 3–4% of clonesfrom both the unripe green and ripe yellow libraries.However, closer inspection revealed major differences inthe composition of the subclasses. The green unripe librarycontained many clones that fell within the plant hormonalregulation subclass whereas the yellow unripe fruit pre-dominantly contained clones involved in Ca2+-mediatedand G-protein-mediated signal transduction.Approximately 3% of clones from each library fell

within the energy functional class. However, the composi-tion of energy-related clones differs between each library.The ripe yellow fruit library produced 18 clones related toelectron transport and membrane-associated energy con-servation that were not present in the unripe green fruitlibrary clones. These clones encode for photosystem IIprotein K (seven clones), cytochrome b5 (seven clones),and cytochrome b5 reductase (four clones). The greenunripe fruit produced six clones encoding a putative (S)-2-hydroxy-acid oxidase assigned to the respiration subclassthat were absent from the yellow fruit library clones.Differences were also observed between various metab-

olism subclasses. The ripe yellow library contained manyclones associated with cysteine biosynthesis that wereabsent from the unripe green library clones. In addition,12 clones of the subclass glycine degradation encoding forglycine decarboxylase were isolated from the ripe yellow

fruit library, whereas just one clone was isolated from thegreen fruit library. The nitrogen and sulphur utilization andmetabolism subclasses contained 11 clones from the yellowfruit library and only one from the green fruit library.

Approximately 37–40% of the unripe green and ripeyellow fruit sequences analysed shared homology to un-classified MATDB sequences, and as such were assigned asof unknown or not clear-cut function.

Summary

A pineapple EST sequencing project was initiated to studygene expression during fruit ripening and to develop a valu-able molecular resource for pineapple biotechnology.Significant differences have been identified in the mostabundant genes expressed in the green unripe tissue com-pared with the yellow ripe fruit tissue. The differential ex-pression patterns of the most abundant EST species isolatedwere validated by northern analysis of fruit tissue spanninga range of developmental and ripening stages, as well asapical meristem, leaf, and root tissues (Fig. 4). Northernanalysis confirmed the fruit bromelain and bromelain in-hibitor genes found to be abundant in the green fruit libraryare down-regulated during fruit ripening. Two metallothio-nein genes (Ac180 andAc181), aMADS box gene (Ac146),a protein translation factor SUI1 (Ac75), and a 60S ribo-somal protein L10 (Ac8) were all abundant in the yellowfruit library relative to the green fruit library. All were shownto be strongly up-regulated during fruit ripening by northernanalysis. With the exception of MADS box protein Ac146,the abundant clones isolated from the yellow fruit libraryexhibit a similar expression pattern of up-regulation duringfruit ripening and up-regulation in older leaves. These resultssuggest that a suite of genes up-regulated during ripening infruits is similarly up-regulated during maturation or senes-cence in leaves, perhaps by common regulatory processes.

The metallothionein clone Ac180 was extremely abun-dant in the green unripe and yellow ripened libraries. Over40% of all library colonies hybridized strongly to a radio-labelled Ac180 probe. The metallothionein Ac180 expres-sion level was further analysed by quantitative real-timePCR. Expression of metallothionein Ac180 was calculatedat over 50-fold higher than the b-actin control in ripeningfruit tissues. Why metallothionein is so strongly expressedin ripening pineapple fruit tissue remains unclear.

The present study provides a valuable resource for thefuture study of fruit ripening in pineapple, and in non-climacteric fruits in general. A number of genes likely toplay vital roles in fruit ripening and senescence have beenidentified, providing many candidate genes for futureanalysis. In addition, the isolation of the EST clone setwill allow microarrays to be constructed for the large-scaleexamination of gene expression over a range of fruit devel-opment stages. It is hoped to investigate the similarities anddifferences in genes expressed in this non-climacteric fruit

110 Moyle et al.


model compared with climacteric fruit systems. It is alsopossible that key gene sequences identified could prove use-ful for future molecular breeding programmes, as markersto aid selection of desirable traits in commercial pineapplelines. The identification of pineapple fruit-specific geneswill also facilitate the isolation of promoter sequencesconferring tight gene control during fruit ripening.

Supplementary data

For supplementary data, please refer to JXB online.

Acknowledgements

This work was supported by the Australian Research Council (ARC)and Golden Circle Limited.

References

Adams MD, Soares MB, Kerlavage AR, Fields C, Venter JC.1993. Rapid cDNA sequencing (expressed sequence tags) froma directionally cloned human infant brain cDNA library. NaturalGenetics 4, 373–80.

Aharoni A, Keizer LC, Van Den Broeck HC, Blanco-Portales R,Munoz-Blanco J, Bois G, Smit P, De Vos RC, O’Connell AP.2002. Novel insight into vascular, stress, and auxin-dependent and-independent gene expression programs in strawberry, a non-climacteric fruit. Plant Physiology 129, 1019–1031.

Aharoni A, O’Connell AP. 2002. Gene expression analysis ofstrawberry achene and receptacle maturation using DNA micro-arrays. Journal of Experimental Botany 53, 2073–2087.

Alscher RG, Erturk N, Heath LS. 2002. Role of superoxidedismutases (SODs) in controlling oxidative stress in plants.Journal of Experimental Botany 53, 1331–1341.

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990.Basic local alignment search tool. Journal of Molecular Biology215, 403–410.

Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, WheelerDL. 2003. GenBank. Nucleic Acids Research 31, 23–27.

Bevec D, Hauber J. 1997. Eukaryotic initiation factor 5A activityand HIV-1 Rev function. Biological Signals 6, 124–133.

Bhalerao R, Keskitalo J, Sterky F, et al. 2003. Gene expression inautumn leaves. Plant Physiology 131, 430–442.

Blanco-Portales R, Medina-Escobar N, Lopez-Raez JA,Gonzalez-Reyes JA, Villalba JM, Moyano E, Caballero JL,Munoz-Blanco J. 2002. Cloning, expression and immunolocali-zation pattern of a cinnamyl alcohol dehydrogenase gene fromstrawberry (Fragaria 3 ananassa cv. Chandler). Journal of Ex-perimental Botany 53, 1723–1734.

Boguski MS, Lowe TM, Tolstoshev CM. 1993. dbEST–databasefor ‘expressed sequence tags’. Nature Genetics 4, 332–333.

Buchanan-Wollaston V. 1994. Isolation of cDNA clones for genesthat are expressed during leaf senescence in Brassica napus. Identi-fication of a gene encoding a senescence-specific metallothionein-like protein. Plant Physiology 105, 839–846.

Butt A, Mousley C, Morris K, Beynon J, Can C, Holub E,Greenberg JT, Buchanan-Wollaston V. 1998. Differential ex-pression of a senescence-enhanced metallothionein gene in Arabi-dopsis in response to isolates of Peronospora parasitica andPseudomonas syringae. The Plant Journal 16, 209–221.

Cazzonelli CI, Cavallaro AS, Botella JR. 1998. Cloning andcharacterization of ripening-induced ethylene biosynthetic genes

from non-climacteric pineapple (Ananas comosus) fruits. Austra-lian Journal of Plant Physiology 25, 513–518.

Chatthai M, Kaukinen KH, Tranbarger TJ, Gupta PK, Misra S.1997. The isolation of a novel metallothionein-related cDNAexpressed in somatic and zygotic embryos of Douglas-fir: regula-tion by ABA, osmoticum, and metal ions. Plant Molecular Biology34, 243–254.

Chen HJ, Hou WC, Yang CY, Huang DJ, Liu JS, Lin YH. 2003.Molecular cloning of two metallothionein-like protein genes withdifferential expression patterns from sweet potato (Ipomoeabatatas) leaves. Journal of Plant Physiology 160, 547–555.

Choi D, Kim HM, Yun HK, Park JA, Kim WT, Bok SH. 1996.Molecular cloning of a metallothionein-like gene from Nicotianaglutinosa L. and its induction by wounding and tobacco mosaicvirus infection. Plant Physiology 112, 353–359.

Church GM, Gilbert W. 1984. Genomic sequencing. Proceedingsof the National Academy of Sciences, USA 81, 1991–1995.

Clendennen SK, May GD. 1997. Differential gene expression inripening banana fruit. Plant Physiology 115, 463–469.

Conklin PL, Last RL. 1995. Differential accumulation of antioxi-dant mRNAs in Arabidopsis thaliana exposed to ozone. PlantPhysiology 109, 203–212.

Coupe SA, Taylor JE, Roberts JA. 1995. Characterisation of anmRNA encoding a metallothionein-like protein that accumulatesduring ethylene-promoted abscission of Sambucus nigra L. leaf-lets. Planta 197, 442–447.

Cui Y, Gonzalez CI, Kinzy TG, Dinman JD, Peltz SW. 1999.Mutations in the MOF2/SUI1 gene affect both translation andnonsense-mediated mRNA decay. RNA 5, 794–804.

Cummins I, Cole DJ, Edwards R. 1999. A role for glutathionetransferases functioning as glutathione peroxidases in resistancetomultiple herbicides in black-grass.ThePlant Journal 18,285–292.

Davies C, Robinson SP. 2000. Differential screening indicatesa dramatic change in mRNA profiles during grape berry ripening.Cloning and characterization of cDNAs encoding putative cell walland stress response proteins. Plant Physiology 122, 803–812.

Dietz KJ. 2003. Plant peroxiredoxins. Annual Review of PlantBiology 54, 93–107.

Dixon DP, Lapthorn A, Edwards R. 2002. Plant glutathionetransferases. Genome Biology 3, REVIEWS3004.

Dowdy SF, Lai KM, Weissman BE, Matsui Y, Hogan BL, Stan-bridge EJ. 1991. The isolation and characterization of a novel cDNAdemonstrating an altered mRNA level in nontumorigenic Wilms’microcell hybrid cells. Nucleic Acids Research 19, 5763–5769.

Dunaeva M, Adamska I. 2001. Identification of genes expressed inresponse to light stress in leaves of Arabidopsis thaliana usingRNA differential display. European Journal of Biochemistry 268,5521–5529.

Giovannoni J. 2001. Molecular biology of fruit maturation andripening. Annual Review of Plant Physiology and Plant MolecularBiology 52, 725–749.

Hillman RT, Green RE, Brenner SE. 2004. An unappreciated rolefor RNA surveillance. Genome Biology 5, R8.

Hofmann B, Hecht HJ, Flohe L. 2002. Peroxiredoxins. BiologicalChemistry 383, 347–364.

Hsieh HM, Liu WK, Chang A, Huang PC. 1996. RNA expressionpatterns of a type 2 metallothionein-like gene from rice. PlantMolecular Biology 32, 525–529.

Hu W, Yan Q, Shen DK, et al. 2003. Evolutionary and biomedicalimplications of a Schistosoma japonicum complementary DNAresource. Natural Genetics 35, 139–147.

Jimenez A, Creissen G, Kular B, Firmin J, Robinson S,Verhoeyen M, Mullineaux P. 2002. Changes in oxidativeprocesses and components of the antioxidant system during tomatofruit ripening. Planta 214, 751–758.



Ledger SE, Gardner RC. 1994. Cloning and characterization of fivecDNAs for genes differentially expressed during fruit developmentof kiwifruit (Actinidia deliciosa var. deliciosa). Plant MolecularBiology 25, 877–886.

Liang L, Lai Z, Ma W, Zhang Y, Xue Y. 2002. AhSL28,a senescence- and phosphate starvation-induced S-like RNasegene in Antirrhinum. Biochimica et Biophysica Acta 1579, 64–71.

Ma RC, Oliveira MM. 2000. The RNase PD2 gene of almond(Prunus dulcis) represents an evolutionarily distinct class of S-likeRNase genes. Molecular and General Genetics 263, 925–933.

Marrs KA, Walbot V. 1997. Expression and RNA splicing of themaize glutathione S-transferase Bronze2 gene is regulated bycadmium and other stresses. Plant Physiology 113, 93–102.

Mittler R. 2002. Oxidative stress, antioxidants and stress tolerance.Trends in Plant Science 7, 405–410.

Mittova V, Theodoulou FL, Kiddle G, Gomez L, Volokita M, TalM, Foyer CH, Guy M. 2003. Coordinate induction of glutathionebiosynthesis and glutathione-metabolizing enzymes is correlatedwith salt tolerance in tomato. FEBS Letters 554, 417–421.

Monteclaro FS, Vogt PK. 1993. A Jun-binding protein related toa putative tumor suppressor. Proceedings of the National Academyof Sciences, USA 90, 6726–6730.

Moons A. 2003. Osgstu3 and osgtu4, encoding tau class glutathioneS-transferases, are heavy metal- and hypoxic stress-induced anddifferentially salt stress-responsive in rice roots. FEBS Letters 553,427–432.

Moriguchi T, Kita M, Hisada S, Endo-Inagaki T, Omura M.1998. Characterization of gene repertoires at mature stage of citrusfruits through random sequencing and analysis of redundantmetallothionein-like genes expressed during fruit development.Gene 211, 221–227.

Nam YW, Tichit L, Leperlier M, Cuerq B, Marty I, Lelievre JM.1999. Isolation and characterization of mRNAs differentiallyexpressed during ripening of wild strawberry (Fragaria vesca L.)fruits. Plant Molecular Biology 39, 629–636.

Neuteboom LW, Kunimitsu WY, Webb D, Christopher DA. 2002.Characterization and tissue-regulated expression of genes involvedin pineapple (Ananas comosus L.) root development. PlantScience 163, 1021–1035.

Oztur ZN, Talame V, Deyholos M, Michalowski CB, GalbraithDW, Gozukirmizi N, Tuberosa R, Bohnert HJ. 2002. Monitor-ing large-scale changes in transcript abundance in drought- andsalt-stressed barley. Plant Molecular Biology 48, 551–573.

Pacey-Miller T, Scott K, Ablett E, Tingey S, Ching A, Henry R.2003. Genes associated with the end of dormancy in grapes.Functional and Integral Genomics 3, 144–152.

Parenicova L, de Folter S, Kieffer M, et al. 2003. Molecular andphylogenetic analyses of the complete MADS-box transcriptionfactor family in Arabidopsis: new openings to the MADS world.The Plant Cell 15, 1538–1551.

Pfaffl MW. 2001. A new mathematical model for relative quantifica-tion in real-time RT-PCR. Nucleic Acids Research 29, 2002–2007.

Potenza C, Thomas SH, Sengupta-Gopalan C. 2001. Genes inducedduring early response toMeloidogyne incognita in roots of resistantand susceptible alfalfa cultivars. Plant Sciences 161, 289–299.

Reid SJ, Ross GS. 1997. Up-regulation of two cDNA clonesencoding metallothionein-like proteins in apple fruit during coolstorage. Physiologia Plantarum 100, 183–189.

Reynolds TL, Crawford RL. 1996. Changes in abundance of anabscisic acid-responsive, early cysteine-labelled metallothioneintranscript during pollen embryogenesis in bread wheat (Triticumaestivum). Plant Molecular Biology 32, 823–829.

Ronning CM, Stegalkina SS, Ascenzi RA, et al. 2003. Comparativeanalyses of potato expressed sequence tag libraries. Plant Phys-iology 131, 419–429.

Rosorius O, Reichart B, Kratzer F, Heger P, Dabauvalle MC,Hauber J. 1999. Nuclear pore localization and nucleocytoplasmictransport of eIF-5A: evidence for direct interaction with the exportreceptor CRM1. Journal of Cell Science 112, 2369–2380.

Rouhier N, Vlamis-Gardikas A, Lillig CH, Berndt C, SchwennJD, Holmgren A, Jacquot JP. 2003. Characterization of the redoxproperties of poplar glutaredoxin. Antioxididants and RedoxSignalling 5, 15–22.

Roxas VP, Smith Jr RK, Allen ER, Allen RD. 1997. Over-expression of glutathione S-transferase/glutathione peroxidaseenhances the growth of transgenic tobacco seedlings during stress.Nature Biotechnology 15, 988–991.

Ruzsa SM, Mylona P, Scandalios JG. 1999. Differential responseof antioxidant genes in maize leaves exposed to ozone. RedoxReports 4, 95–103.

Salekdeh GH, Siopongco J, Wade LJ, Ghareyazie B, Bennett J.2002. Proteomic analysis of rice leaves during drought stress andrecovery. Proteomics 2, 1131–1145.

Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: alaboratorymanual. NewYork: Cold SpringHarbor Laboratory Press.

Schoof H, Zaccaria P, Gundlach H, Lemcke K, Rudd S, KolesovG, Arnold R, Mewes HW, Mayer KF. 2002. MIPS Arabidopsisthaliana Database (MAtDB): an integrated biological knowledgeresource based on the first complete plant genome. Nucleic AcidsResearch 30, 91–93.

Schrader J, Moyle R, Bhalerao R, Hertzberg M, Lundeberg J,Nilsson P, Bhalerao RP. 2004. Cambial meristem dormancy intrees involves extensive remodelling of the transcriptome. ThePlant Journal 40, 173–187.

Skipsey M, Andrews CJ, Townson JK, Jepson I, Edwards R.1997. Substrate and thiol specificity of a stress-inducible glutathi-one transferase from soybean. FEBS Letters 409, 370–374.

Taylor CB, Bariola PA, delCardayre SB, Raines RT, Green PJ.1993. RNS2: a senescence-associated RNase of Arabidopsis thatdiverged from the S-RNases before speciation. Proceedings of theNational Academy of Sciences, USA 90, 5118–5122.

Thom R, Cummins I, Dixon DP, Edwards R, Cole DJ, Lapthorn AJ.2002. Structure of a tau class glutathione S-transferase from wheatactive in herbicide detoxification. Biochemistry 41, 7008–7020.

van Hoof NA, Hassinen VH, Hakvoort HW, Ballintijn KF,Schat H, Verkleij JA, Ernst WH, Karenlampi SO, TervahautaAI. 2001. Enhanced copper tolerance in Silene vulgaris (Moench)Garcke populations from copper mines is associated with increasedtranscript levels of a 2b-type metallothionein gene. Plant Physi-ology 126, 1519–1526.

Wang TW, Lu L, Wang D, Thompson JE. 2001. Isolation andcharacterization of senescence-induced cDNAs encoding deoxy-hypusine synthase and eucaryotic translation initiation factor 5Afrom tomato. Journal of Biological Chemistry 276, 17541–17549.

Wang TW, Lu L, Zhang CG, Taylor C, Thompson JE. 2003. Pleio-tropic effects of suppressing deoxyhypusine synthase expression inArabidopsis thaliana. Plant Molecular Biology 52, 1223–1235.

Whitfield CW, Cziko AM, Robinson GE. 2003. Gene expressionprofiles in the brain predict behavior in individual honey bees.Science 302, 296–299.

Wilusz CJ, Wang W, Peltz SW. 2001. Curbing the nonsense: theactivation and regulation of mRNA surveillance. Genes andDevelopment 15, 2781–2785.

Xu A, Chen KY. 2001. Hypusine is required for a sequence-specificinteraction of eukaryotic initiation factor 5A with postsystematicevolution of ligands by exponential enrichment RNA. Journal ofBiological Chemistry 276, 2555–2561.

Yang SF, Hoffman NE. 1984. Ethylene biosynthesis and itsregulation in higher plants. Annual Review of Plant Physiology35, 155–189.

112 Moyle et al.


Developing pineapple fruit has a small transcriptome dominated by ...

Documents

Transcript of Developing pineapple fruit has a small transcriptome dominated by ...