Work

Functional Genomics in Plants

Year: 1998
Type: review
Abstract: The sequencing of the first genome of a higher plant, Arabidopsis, is progressing at a rapid pace. As we are writing these words, almost one-third of the estimated 100 Mb is available in public databases, and with appropriate funding the complete genome sequence is predicted to be finished by the year 2001. Sequencing programs for other plant genomes such as rice are planned, and with new generations of more efficient sequencing machines (Marshall and Pennisi, 1998), the sequence of several other plant genomes may become available in the coming decade. In the wake of these sequencing efforts, plant research enters an exciting period in which genome-wide approaches are becoming an integral part of plant biology, with potentially highly rewarding but as yet unpredictable biotechnological applications. This is reflected by the current frenzy with which new agricultural biotechnical companies are being founded and the rate at which existing companies are investing in the development of tools to exploit and further expand this wealth of information. The term functional genomics can be referred to as the “development and application of global (genome-wide or system-wide) experimental approaches to assess gene function by making use of the information and reagents provided by structural genomics” (Hieter and Boguski, 1997). With these approaches the focus of the analysis is shifted from individual components to biological systems. It involves the use of high-throughput methods for the study of large numbers of genes (ideally the entire set) in parallel. Gene “function” can be considered from several points of view: it can mean biochemical function (e.g. protein kinase), cellular function (e.g. a role in a signal transduction pathway), developmental function (e.g. a role in pattern formation), or adaptive function (the contribution of the gene product to the fitness of the organism). Having identified a new sequence, the comparison with sequence databases is the simplest way to obtain (essentially biochemical) functional information. Currently, about 50% of newly identified genes show sequence similarity to previously described genes. However, computerized analyses are generally not sufficient to define gene function with a high level of confidence, and experimental confirmation is needed in most cases. Indirect information on cellular or developmental function can be obtained from spatial and temporal expression patterns; for example, the presence of mRNA and/or protein in different cell types, during development, during pathogen infection, or in different environments. The subcellular localization and posttranslational modifications of proteins can be informative as well. Knocking out or overexpressing the gene permits the gene sequence to be linked to a phenotype from which a cellular role or a role in development may be deduced. Finally, the fitness of plants carrying mutations or natural variants for the gene can be compared in different environments with wild-type plants to study the adaptive function. In this Update we provide a review of the ever-growing toolbox for the global study of gene function in plants, indicating the potential and the limitations of the different techniques. Where appropriate, we will also draw a parallel with the more advanced technologies in bacterial, yeast, and animal systems. A rapid way to establish an inventory of expressed genes is by determining partial sequences of cDNA called ESTs. In this approach, single-pass sequences of 300 to 500 bp are determined from one or both ends of randomly chosen cDNA clones. The sequences are sloppy and have a relatively high error rate, but they are sufficiently accurate to unambiguously identify the corresponding gene in most cases. Thousands of sequences can thus be determined with a limited investment. EST information is present in public databases for a variety of species, including a number of plants (Höfte et al., 1993; Newman et al., 1994; Cooke et al., 1996; Yamamoto and Sasaki, 1997). A recent release (March 23, 1998) of the EST database of the National Center for Biotechnology Information (TableI) lists more than 39,000 Arabidopsis ESTs. Powerful programs for rapid sequence comparisons with databases are accessible via the Internet (Table I). The Institute for Genome Research (Table I) very conveniently has ordered overlapping ESTs into tentative contigs. It is difficult to know exactly how many different genes are represented by these ESTs, because cDNA clones frequently are truncated at the 5′ end, and sequencing of 5′ ends of cDNA clones often yields nonoverlapping sequences corresponding to the same mRNA. However, the fact that 56% of the predicted genes exactly matched an EST within 1.9 Mb of contiguous genomic sequence (Bevan et al., 1998) suggests that more than one-half of the total set of Arabidopsis genes is already represented by an EST. It should be noted that several companies possess large private EST databases for various crop plants (e.g. maize and soybean), the access to which can be negotiated on a case-by-case basis. Useful websites for plant genomics Useful websites for plant genomics While we await the complete genome sequence, EST databases have proven to be a tremendous resource for finding genes and for interspecies sequence comparison, and have provided markers for genetic and physical mapping and clones for expression analyses. The relative abundance of ESTs in libraries prepared from different organs and plants in different physiological conditions also provides preliminary information on expression patterns for the more abundant transcripts. EST sequencing programs in Arabidopsis and rice have been extremely successful in the discovery of new genes. However, rare transcripts and transcripts of genes that are induced under specific conditions (e.g. biotic and abiotic stresses) are not represented in EST databases. The only sure way to gain access to the entire set of genes is to determine the complete genomic sequence. This sequence also provides information on the global structure of the genome, including the relative order of genes on the chromosomes, which is extremely valuable for positional cloning strategies. At present the genomes of rice and Arabidopsis, both model plant species with small, information-rich genomes, are being sequenced on the basis of coordinated multinational efforts. For the Arabidopsis Genome Initiative, specialized sequencing laboratories from the United States, Europe, and Japan are continuously releasing sequence data in public databases (Table I) annotated with the results from database searches and gene-prediction algorithms (Rounsley et al., 1998). The first results from the large-scale sequencing of a 1.9-Mb region (Bevan et al., 1998) confirmed that the Arabidopsis genome is extremely gene rich (one gene every 5 kb, on average) and poor in repeated elements such as retrotransposons. The total gene complement of Arabidopsis is estimated to be around 21,000 protein-coding genes. About one-half of the predicted genes on the 1.9-Mb segment can be assigned to a functional category based on similarity with known proteins. A major problem with genomic sequences is how to distinguish coding regions from noncoding intergenic sequences and introns. Comparisons with EST and cDNA sequences and sequence similarity to known coding sequences can be used to assign intron positions for many genes. However, for the genes that do not match sequences in the databases, the coding sequences need to be predicted from the genomic sequence. For the human genome, programs such as GRAIL (Table I) have been developed to predict the beginnings and ends of genes and intron positions with high reliability (Uberbacher and Mural, 1991). In plants splicing signals are still poorly defined, and in Arabidopsis introns are often very small; the average size is around 200 bp and on average six or seven introns are found per gene. A gene-prediction algorithm based on neuronal networks was developed for plant sequences. The program, NetPlantGene, which is freely available (Table I), allows for a reliable prediction of introns (Hebsgaard et al., 1996; Tolstrup et al., 1997). A better comprehension of splicing mechanisms in plants is needed to improve gene-prediction algorithms in the future. Knowing when and where a gene product (RNA and/or protein) is expressed can provide important clues to its biological function. The facility with which a high-throughput approach can be used for the study of gene expression depends largely on the level of regulation that is being addressed. Technologies have been developed for the parallel study of mRNA and protein levels in yeast, animals, and plants. Additionally, procedures for the systematic study of protein-protein interactions using two-hybrid screens (Fromont-Racine et al., 1997) are being developed for yeast and may be adopted for plant systems in the future. SAGE is a logical extension of EST sequencing. An inventory of transcripts is established based on very short cDNA sequence tags (9–11 bp) from the 3′ part of mRNA molecules, concatenated, and cloned before sequencing (Velculescu et al., 1995). These short tags have proven sufficiently long to unambiguously identify corresponding genes in databases (Velculescu et al., 1997). Expression patterns for the different genes are reflected by the relative abundance of individual tags. SAGE patterns have been studied in human and in yeast (Polyak et al., 1997; Velculescu et al., 1997), but to our knowledge, the technique has not yet been applied to plants. A prerequisite for the identification of the tags is the availability of large sequence databases for the species under study. The technique is powerful but not very convenient for the comparison of many different samples and for the study of the rarer transcripts. High-density DNA arrays and the monitoring of mRNA levels for large numbers of genes in parallel. A, cDNAs or open reading frames (ORFs) identified in genomic sequences were PCR amplified and printed onto a solid support (microscope slides or nylon filters) using a gridding robot. Single-stranded cDNA probes were synthesized from total mRNA populations using reverse transcriptase in the presence of labeled nucleotides and hybridized to the DNA arrays. Hybridization signals were detected using a two-dimensional radioactivity detector such as a Phosphorimager or, for fluorescent probes, a modified confocal laser scanning microscope. The signals were quantified and processed using specialized software. The intensity of the hybridization signal is proportional to the abundance of the corresponding mRNA in the pool used to synthesize the probe. B, Sector of a yeast genome microarray. The total array was 18 × 18 mm and contained 6400 distinct, PCR-amplified ORFs. Two probes labeled with different fluorochromes were mixed and hybridized, and the two colors were quantified simultaneously. (Adapted from DeRisi et al., 1997.) Different systems have been developed depending on the source of target DNA and the nature of the solid support and the detection system. The simplest system, which uses nylon filters in combination with radioactive probes, can be put to work with a minor investment. It requires a commercial, all-purpose gridding robot and a radioactivity detection system such as a Phosphorimager. Software for the detection and quantification of signals is commercially available. The system was tested out on some 800 partially sequenced cDNA clones in Arabidopsis, with reproducible results (Desprez et al., 1998). Filters with a density of up to 6144 cDNA on a 12- × 8-cm surface have been made (H. Höfte, unpublished data). The technique can be used to detect low-abundance mRNAs (down to 1:10,000 of the total mRNA population). A more sophisticated system uses DNA microarrays printed at a high density on pretreated glass slides (Schena et al., 1995; DeRisi et al., 1997). This system allows the use of fluorescent probes, and hybridization signals are detected using an adapted confocal laser microscope (Fig. 1B). The use of two different fluorophores allows the simultaneous detection of hybridization signals of two probes in a single hybridization experiment, resulting in highly reproducible data. The system has a high sensitivity: mRNAs of an abundance down to 1:100,000 can be reliably detected, which in yeast corresponds to approximately 0.15 mRNA molecules per cell. The major drawback of this technique is its cost and the requirement for a specialized arraying robot and scanner. Also, the arrays cannot be reused, which further increases cost. An alternative technology is referred to as DNA chips. It is based on a method to synthesize large amounts of different oligonucleotides in situ on a glass support using light-directed, solid-phase, combinatorial chemistry developed by Affymetrix (Santa Clara, CA). Oligonucleotides corresponding to the genes of interest are hybridized with fluorescent probes, and signals are detected using techniques similar to the ones used for the microarrays. A hallucinating experiment was described recently (Wodicka et al., 1997) in which 260,000 25-mers, corresponding to a nearly complete set of yeast genes, were synthesized in situ on four 1.28-cm2 grids. Each open reading frame predicted from the genomic sequence was represented by about 20 oligonucleotides to cancel out differences in hybridization behavior of different oligonucleotides and cross-hybridization with related sequences. In addition, the hybridization value for each oligonucleotide was corrected for the hybridization value for a second, negative oligonucleotide with an sequence for a single The method a very high and reliably detected a present at one in to 20 The is to a human with elements genes for expression DNA for the complete set of Arabidopsis genes may become in the provided appropriate funding is available for the For the the technology to plant in public In the genome-wide analysis of gene expression will become for plants. public laboratories are up systems for Arabidopsis (Desprez et al., and a companies are already filters or including hybridization and data analysis with cDNA microarrays. can be with these they can be used to in a systematic way mRNA levels in different in the of development or in different information can be in databases that may be sequence databases. of such databases for yeast and Arabidopsis (Table I). global expression can be used to genes based on spatial and expression For example, during the of the of yeast to of genes be based on or et al., 1997). elements may be identified by the sequences of genes of the same the of in gene with expression of for known genes, should the study of expression For example, in yeast, in the expression of genes with known function were of the during the these same expression will be in the in developmental Software tools to such comparisons need to be techniques such as in situ hybridization on various organs are to obtain more information on as or can also provide information on gene Information on mRNA levels is not sufficient to obtain a complete of the way gene expression is within the cell. expression data are more but are more difficult to obtain in a parallel The term was to to the total protein complement expressed by the is for the the method of to study the abundance and posttranslational modifications of several proteins in parallel and 1997). the and of the system was and algorithms for quantification have been protein on are more identified with methods for and more with the use of techniques and 1997). The are based on the to determine very the of of amounts of protein from with predicted from sequences in protein databases allows for the of a sequence to the protein provided that a protein database is available. databases for yeast, and In plants more limited databases for and A database for proteins in Arabidopsis is under development (Table I), which not only provides information on abundance and posttranslational but also on the localization of a of proteins. This approach can be to other and protein provided that efficient procedures are available. the various the approach has been used in yeast to study gene function the of or and for the analysis of in protein on two-dimensional In maize comparison of of nearly for the gene the identification of new for the and 1998). In Arabidopsis were used to developmental and the of the of in one of the which was confirmed et al., its the technique for the monitoring of only a abundant proteins. based on of of protein may be developed in the for the study of total protein on the study of found in natural populations or induced by The analysis of the of this in mapping populations allows to determine the number of genetic for the and relative on the This can be as is generally the for or as with most of such as and plant which are by large numbers of genes. approaches have been extremely successful in recent for the study of the genetic and for in plant procedures to relatively a high of a in every gene in the to the genes is obtained using positional cloning cloning the gene based on its on the genetic This is in model species such as Arabidopsis, for which genetic with many and genetic markers and for which an almost complete physical of a of overlapping DNA cloned in yeast or chromosomes, has been et al., 1997; et al., 1996; et al., 1998). The total genomic sequence will provide the physical With of these with the genomic sequence in positional cloning will be the being the and for the mapping and the mapping of the mapping based on will this as well. of a DNA within a gene gene at the the or can be to use a (e.g. or by the sequence or The presence of a with a allows the expression of the gene at the of to be are by by populations can be by PCR using specific for the gene and for the The can be by cloning procedures and by PCR using The as in are the level of the of at one in which depends on the number of in the the of of the and the number of per For the Arabidopsis genome, we that a of around is needed to provide a high level of of gene at populations of plants, with maize such as or or the of have been in several plant species and The level of depends on the size and of the In addition, the can be with or gene that will on the expression of the gene at the of 1998). For example, a to the of the can when to a that will the expression of the gene. populations can be for the gene in specific cell or in specific with expression patterns and can be from such can also be used to cell for developmental et al., 1995). which is based on the use of an carrying a or into the region the the of mutations in which of a gene a phenotype et al., mutations are generally or The of approaches is relative facility in large populations of and to use the of many elements to to linked which it to the for in the of the the other it is more difficult to with but it results in per are to as as we do not show and 1997). In Arabidopsis highly efficient in techniques have been in large populations of the of linked to in and et al., 1993; and 1997). A way to obtain information on the function of a gene identified by sequencing is to a and study the phenotype of the resulting In many model (e.g. and can be used to target mutations into specific genes by the wild-type gene with a In plants has proven extremely difficult because of the of and some has been and 1995; et al., 1997), gene by is not considered on a large and other have to be to the functional analysis of large numbers of genes. in Arabidopsis using large populations of plants with an such as the of or a populations of plants were to an In Arabidopsis an average is about kb, which that the genome can be into around have a of at one in one about × DNA is from of plants and into can be detected by PCR in as large as several individual on the nature of the and can be into or to the of single carrying the PCR are on DNA using oligonucleotide from the gene of interest and from the PCR is extremely to interactions to the of the it is often to several PCR are on an to a and hybridized to probes from the gene of interest and from the This is to PCR and only to both probes are studied the has been confirmed by PCR and sequence the can be into and into individual The of the can be more or depending on the that has been adopted for the of the The first of mutations in Arabidopsis et al., et al., 1998) show that analysis of is often to detect the of a In our more than were obtained from PCR of a et al., and most of the not show when in conditions unpublished data). it may be to identify the conditions for the expression of the An is the for which a in is only at et al., 1998). mutations may have very that are not in experimental and have to be to differences and et al., 1998). Finally, it from large-scale sequencing programs that genomes are with many genes in (Bevan et al., 1998). The functional analysis of such gene may often the of carrying mutations in different A further extension of this involves the use of an approach to sequence the regions of it is to and sequence a large number of from or individual The can be by the sequences with the genomic sequence. of such sequences will be established and can be for in gene. of the carrying the sequenced will become available from the public and private laboratories are in such and these tools should become available in the future. It is that the interest from private companies will with the public release of these with the of the first sequence of a plant genome, a is becoming available that will improve the of research in of plant This large of in model for systematic gene function searches and methods for the study of the regulation of entire of genes at the mRNA or protein A in the will be to databases information on such as sequence, mRNA and protein and Also, tools will be for the efficient of this wealth of information and to the interactions that the EST sequencing and mapping or genomic sequencing of a crop species may to the of information from model species to the of crop plants, to be used for crop in the future. is for information on cDNA for unpublished and for elements for expressed sequence analysis of gene expression
Language: English
Subfield: Plant Science
Open Access status: green
OpenAlex ID: W1986221860
FWCI: 2.335
Cites: 48
Cited by: 198
Related to: 9