Analysis of gene expression during the fruit set of tomato: A comparative approach

Abstract

The study of tomato fruit set and early fruit development is of upmost importance due to their economic implications for crop production. Moreover, they are good subjects in the attempt to understand plant development control. In this work, we isolated more than 400 genes that are differentially expressed during these processes in tomato by using two complementary approaches: suppressive subtractive hybridization and genomic comparative analysis. We made a tomato flower library enriched in genes differentially expressed at 3 DPA when compared against anthesis. The library subtraction showed the high variability of genes and pathways implicated in this process. Using the AtGenExpress data from Arabidopsis, we detected 1879 genes differentially expressed during fruit set and early development. The expression of their orthologue tomato genes was tested by quantitative PCR, and more than 75% of the genes were also differentially expressed in tomato, meaning there is a conservation between these two species in spite of the significant differences in fruit morphology and development. We detected known and unknown fruit development genes and pathways, and this data will be a good source of information for future experiments about fruit set in tomato and Arabidopsis. We also showed that this transcriptomic comparative approach is very useful in identifying target genes of conserved processes in species where microarray facilities are not available, making the work with model species more profitable.

Introduction

The tomato (Solanum lycopersicon L.), like other crops, is grown to obtain its fruit, and its production is dependent upon fruit set and development among other factors. Fruit development is also a process of interest for the study of plant development regulation, and the tomato has been used as a model plant for climacteric fruit development [1]. Fruit development in tomato is divided in four phases: fruit set, cell division stage, cell expansion stage and ripening. In tomato, the fruit set can be defined as the restarting of cell division in the ovary after pollination or fecundation. The development of tomato fruit is independent of embryo development, and the linkage of these processes can be broken. Parthenocarpy, the production of fruits without seeds, is common in this species and can be caused by natural mutations, environmental factors or hormone treatments. The existence of a hormonal control in the fruit set is obvious and has been demonstrated by various studies reviewed in Ozga et al. [2] and Srivastava and Handa [3]. Gibberellins and auxins play a crucial role in this process in tomato, although it appears that other plant regulators might be implicated, such as cytokinins, ethylene and polyamines. The implication of these hormones has been shown by the measurement of endogenous levels in pollinated ovaries, in the unpollinated ovaries of parthenocarpic lines and by exogenous application. Natural and artificial mutants have demonstrated the existence of a genetic control in fruit set, but there is little known about this control. In tomato, there are several described mutations that produce parthenocarpic growth, such as pat, pat-2 and pat3/pat4 [4], [5], [6], [7], [8], [9], but they have not been characterized at the molecular level. Several genes are also described as being implicated in fruit set; among others, Aux/IAA transcription factor IAA9 [10], TM29, a tomato SEPALLATA homologue [11] and the diageotropica gene [12]. The panorama in other crop species is similar, and not much information is available.

As of this work, no global analysis of gene expression during fruit set has been published for tomato. Microarray or proteomic analyses have been used to analyze global expression during fruit development in tomato, but these analyses have not included the fruit set stage and are focused on the late stages of fruit development, from 7 DPA to ripening [13], [14], [15]. Digital northern is another approach for analyzing global expression, but the available tomato EST libraries do not include this stage of fruit development separately [13]. Therefore, more work is necessary in order to understand the genetic and hormonal control as well as the metabolic pathways of the fruit set in tomato and other crops, especially for the identification of the key genes and pathways. A suppressive subtractive hybridization (SSH) or SAGE that can be quick and useful approaches to create a library enriched in genes differentially expressed in any tissue, when there are no available EST libraries to perform a digital northern [16].

Arabidopsis is an excellent model for plant development and many data have been extrapolated to other species. The conservation of developmental and metabolic pathways, gene function and expression data between this species and other unrelated plants has been repeatedly reported. Arabidopsis genomic and transcriptomic data has been used as a tool for functional genomics in tomato [17]. Notwithstanding, these two species, with different types of fruits, seem too different to expect conservation. However, we are interested in fruit set, which may not be affected by fruit morphology. In fact, the Solanaceae family presents a high fruit variability which includes capsules, drupes, pyrenes, berries and several sorts of dehiscent noncapsular fruits [18]. Moreover, the dehiscence of capsular and noncapsular Solanaceae fruits presents similarities with the Arabidopsis dehiscence process [19], [20], [21]. Good examples of this conservation are the genes implicated in carpel and ovule development or in ethylene signaling ([22], [1] and references therein). In fact, comparative analyses of transcriptomic data between tomato, Arabidopsis and other species have already been done in late tomato fruit development and ripening, which shows the validity of this approach [13], [23].

There are several microarray platforms for Arabidopsis that have been used extensively in different studies about development, stress, etc. These experiments have produced large amounts of data regarding gene expression that may be very useful in the study of developmental or metabolic pathways in this species. One of these projects, named AtGenExpress, consists in the hybridization of the ATH1 array from Affymetrix with a complete selection of developmental and stress samples, which includes the development of carpel and silique [24], and which has been used in several studies [25], [26].

In this work, we isolated genes that are differentially expressed during the tomato fruit set by SSH, and we detected a high degree of correlation between the results of a suppressive subtractive hybridization of two stages of tomato fruit set and the AtGenExpress data. In order to increase the number of differentially expressed genes, we used the AtGenExpress data to detect genes regulated during Arabidopsis carpel and silique development, and we tested these genes in tomato by quantitative real-time PCR; in the end, more than 400 candidate genes were detected as being implicated in the fruit set of tomato.