Arabidopsis genome sequencing in late 2000 has been one of the major accomplishments of plant scientists. Nearly 70 percent of the 26,000 genes in this species were functionally classified, but the actual biological function of the majority of these genes is not known. One of major tasks before the Arabidopsis geneticists is to decipher the functions of all uncharacterized genes, using a systematic approach in a high-throughput manner.
The current methods such as use of gene knockouts, microarrays, etc. could not be successfully applied to this end because of functional redundancy between genes. In the December 2002 issue of Trends in Plant Science (vol. 7:531-534), Richard G.H. Immink and G. C. Angenent at Plant Research International, Wageningen, The Netherlands, dwell on this problem and suggest that mapping of protein–protein interactions has paved the way towards determination of protein functions. The authors suggest that functions may be attributed to many unknown proteins based on the protein reaction maps.
The article begins with a description of some known and new techniques for the identification of protein–protein interactions. Because of their importance as regulatory proteins, the authors focused their study on the MADS-box transcription factor family and used that body of information to understand functional genomics. When a part of DNA constituting a gene is duplicated, the newly formed segment is called a paralogue. This event that has taken place a number of times in Arabidopsis genome seems to be common throughout the plant kingdom. The presence of a large number of paralogues makes analysis of MADS-box transcription factors difficult because of functional redundancy i.e., their tendency to retain their original function.
While describing the usefulness of the yeast two-hybrid GAL4 system to identify protein–protein interactions, the authors point out that it has its limitation. In certain situations the reporter genes undergo autoactivation i.e. activation before the two reacting proteins come in close proximity.
The authors describe an alternative technique called yeast two-hybrid CytoTrap system to identify protein–protein interactions. It is a cytoplasmic system that relies on the Ras signal-transduction cascade for its operation. Ras genes encode proteins that play an important role in transmembrane signaling. The sequence of reactions takes place in the following manner: The bait and prey proteins are fused with the human SOS protein (hSOS) and a particular kind of signal peptide generated as a result of myristylation (= attachment of an myristic acid). In the next step the hSOS protein becomes attached to the membrane of the yeast cells. The Ras signal transduction cascade is activated as a result of anchoring of hSOS to the membrane, enabling the yeast cells to grow at 37°C.
Recent methods for screening protein-protein interactions:The authors then describe two other methods that are used to screen protein-protein interactions in living plant cells: (a) the protein complementation assay (PCA) and (b) fluorescence resonance energy transfer (FRET). As the name suggests, in the PCA method an enzyme called murine dihydrofolate reductase (mDHFR) is separated into two inactive complementary pieces. One separated piece is fused to the bait- and the other to the prey protein. Subsequent interaction between the two (bait and prey) leads to the restoration of the binding ability of mDHFR to tether itself to fluorescein-conjugated methotrexate (fMTX). By means of fluore- scence microscopy or fluorescence-activated cell sorting (FACS), mDHFR–fMTX complexes can be located inside the cell, while free fMTX particles move away from it.
The other technique termed FRET also identifies protein–protein interactions in plants. Fluorescence energy transfer takes place when emission of two fluorescent molecules overlaps with one another and their excitation spectra come in close contact (<100 Å). In this system, two fusion products are made: (a) in one, a bait protein is fused to cyan fluorescent protein (CFP) and (b) in the other; the prey protein is fused to yellow fluorescent protein (YFP). Interaction ensues as soon as CFP and YFP proteins are in close proximity, resulting in fluorescence energy transfer between the two chromophores, triggered by excitation of CFP. This reaction can be observed by various spectroscopy methods.
Predicting protein functions based on interaction patterns: Protein functions are predicted based on the results of interaction that takes place between a specific, well-characterized protein and the unknown dimer partner. The authors give an example to illustrate this point. In petunia, MADS-box protein FBP2 acts in conjunction with B- and C-class proteins to establish floral-organ identity. The same MADS-box protein also interacts with the D class MADS-box protein FBP11, which is known to control ovule development, thereby suggesting that FBP2 has a dual role, participating in both floral-organ identity and ovule formation.
Identification of functionally redundant genes:The two-hybrid system is also used to identify functionally redundant genes. For instance, it has been shown that unlike Arabidopsis, in petunia there are two B-type MADS-box proteins (FBP1 and pMADS2) in the PISTILLATA family. The function of these two proteins is indistinguishable probably because the two genes, encoding the above two proteins, originated recently through gene duplication (paralogues) and as such they are functionally redundant. The authors give a few more examples from Arabidopsis such as APETALA1 (AP1) and CAULIFLOWER (CAL) involved in the control of inflorescence architecture.
Identifying functional homologues: The authors point out that sequence alignment procedure can be used to compare genes from different species to determine their expression pattern but is not always applicable to identify functional homologues. In view of this difficulty, the authors suggest that the analysis of conserved interactions (‘interlogues’) be undertaken in order to discover functionally homologous proteins for gene annotation and comparative genomics. By means of two well-documented examples, the authors illustrate this point. In one, the interaction between the rice MADS-box protein OsMADS18 and the histone-folding protein NF-YB1 in mouse has been shown, demonstrating the conservation of the NF-Y proteins in both animals and plants. The other example, cited by the authors, is the homology between the Norway spruce MADS-box protein DAL13 and the Arabidopsis class-B protein PISTILLATA.
On the basis of results reported earlier and those obtained by the authors, the latter have proposed that heterologous two-hybrid screenings is a powerful tool to identify functional homologues among plant species. In support of their hypothesis, they present interaction maps between the Arabidopsis SEPALLATA3 (SEP3) protein and 23 known petunia MADS-box proteins. The authors advocate this method also for species for which data on mutants are voluminous and are not amenable to easy analysis.
Prospects: In conclusion, the authors suggest that this technology will prove advantageous to screen cDNA expression libraries of crop plants with well-characterized proteins from model species such asArabidopsis. Such a comparison will allow identification of homologues between model plants and their counterparts in crops, making it possible to transfer genomic data in an attempt to improve crop plants. The possibility of such a technology is enormous in the near future, in view of rapid developments in high-throughput interaction screens and the emerging protein-array technology.
Click here for details