The use of transgenic organisms for studying over-expression and knock-out phenotypes has revolutionized scientific experimentation over the past decade. There have been innumerable reports and review articles dealing with the use of antisense strategies to down-regulate specific gene expression as well as to overexpress of particular cellular products by introducing the sense copy of a gene. The generation of such transgenic organisms generally involves directed DNA delivery followed by a ‘prescreen’ which eliminates organisms that do not show the expected phenotype; such transgenics are discounted as defective and tend not to be reported in the literature. (A good example of this is an apparently defective transgenic plant in which the transgene causes the endogenous gene to be silenced instead of being overexpressed.) Thus most of the published work represents just a selected subset of a number of attempts to generate transgenic plants and animals. It is conceivable that these exceptions would have yielded important insights into cellular functioning and regulation. But they were ignored: interpreting them was difficult on account of the limited knowledge of what might be going on. As it turns out, investigation of transgene silencing has begun to open up a whole new field, that of RNA-based long-range signalling in plants. Indeed, the wealth of information generated by studying these so-called defective transgenics has led plant geneticists to treasure them.

Plants in which the introduction of exogenous transgenes leads to an over-expression of useful products are potentially of immense economic importance. However, the presence of multiple copies of homologous (trans) genes in a plant nucleus can lead to the exact opposite, a drastic reduction of both host and transgene steady state mRNA levels (see Raghunand 1998 for a brief review). First observed in petunia in 1990 (Napoli et al 1990), transgene silencing, also known as ‘mutual inactivation’ or ‘co-suppression’, has been reported in a large number of other plant systems like tobacco, Arabidopsis and tomato. Researchers in this field see gene silencing as the plant’s way of correcting for gene over-expression. Several models have been suggested for this interesting phenomenon but the mechanism by which co-suppression occurs is still disputed. Any explanation must take into account the following observations. Firstly, the inactivation process is extremely specific: suppression will affect only those endogenous genes which have the same DNA sequence as the transgene. Further, co-suppression is not cell autonomous: silencing can spread within the entire plant, leading to what has been called ‘systemic acquired silencing’ (Palauqui et al 1997; Voinnet et al 1998).

The identity of the silencing message – a gene specific, mobile signal molecule that could transmit the co-suppression state throughout the plant – eluded plant geneticists for a long time. Protein factors were thought to be unlikely candidates because the genetic load on the host cell imposed by the requirement to code for the vast repertoire of proteins needed to confer specificity would be enormous. On the other hand, a nucleic acid, probably an RNA molecule, would be an ideal candidate for recognition and elimination of specific transcripts (Jorgensen et al 1998). This does not lead to a genetic load for the following reason. If the silencing message were to be a protein, the cell would need to synthesize a protein for every gene. On the other hand, if you have an RNA molecule doing the same job, a common protein, e.g., an RNA-dependent RNA polymerase, would simply copy the over-expressed mRNA transcribed from the transgene and generate a specific silencing message as and when required. How does the protein distinguish between transcripts coming from the transgene and those from the pre-existing mRNA? It could do so by detecting overexpression of a particular mRNA species or by detecting certain abnormal features in these RNA molecules like lack of splicing or double-strandedness. Thus the cell needs to make use of just one general protein, the polymerase enzyme, as against a large number. (As it happens, an RNA-directed RNA polymerase has been identified in tomato leaves; see Schiebel et al 1993). The problem that remained, however, was that there was no precedent for nucleic acids being able to traverse long distances within plants. In a recent study Lucas and co-workers at the University of California, Davis, have put an end to speculation by demonstrating that RNA molecules can actually be transported through the plant phloem and act as carriers of important information (Xoconostle-Cazares et al 1999).

The phloem serves as an advanced long-distance transport system, as a conduit for nutrient and hormone delivery to various tissues and organs. It has been long acknowledged that small molecules are transported through the phloem; but whether macromolecules like nucleic acids could negotiate its narrow channels was unknown. Genetic and molecular approaches had established that plant viruses move large nucleic acids into the phloem by expressing certain viral movement (VM) proteins. Lucas and his colleagues guessed that these viruses were probably mimicking an inherent plant transport system. By using an antibody to a VM protein they were able to pull out its plant paralog, CmPP16, from pumpkin phloem sap. Purified CmPP16 turned out to be an RNA-binding protein which mediated cell to cell transport of both sense and antisense RNA of different sequences, but was unable to affect the movement of single or double stranded DNA. They detected the presence of this protein and its RNA in ‘sieve elements’ which themselves have no nuclei and thus cannot synthesize RNA and lack the machinery to make proteins, implying that both the CmPP16 RNA and protein had moved in from adjacent cells. More interestingly, in pumpkin-cucumber grafts, CmPP16 and its mRNA could indeed traverse large distances – going by earlier work, as much as 30 cm – in the plant: endogenous pumpkin CmPP16 was present in the phloem sap of both the pumpkin plant and the cucumber graft. Further, sequence homology searches identified homologues of CmPP16 in rice, maize and Arabidopsis indicating that what we are looking at might just be the tip of the iceberg and similar RNA-transporting proteins might be present in other plant systems as well.

Abhilasha Gulati

In fact, from a historical view point one can trace the evolution of molecular genetics by studying the different research programmes to which an organism like Drosophila has been subjected at different points of time, i.e., all the way from the time of the transmission genetics of T H Morgan to contemporary research in molecular genetics and developmental biology. While it is well known that Morgan’s work led to very important scientific results, it is interesting to note that his work had important philosophical consequences as well. The work that Morgan did enabled his conversion from being a skeptic regarding the material nature of the Mendelian genes to becoming a firm supporter of it. Prior to his own work with the fruit fly, Morgan along with Bateson tended to see material theories of inheritance as bordering on the ancient doctrine of preformation. Morgan also shared Bateson’s idealism in thinking that there was no material basis for the existence of chromosomes in cell structure. But as Allen (1975, p 59) notes, this change in Morgan’s attitude in the light of his own experience signalled ‘. . . the beginning of a far-reaching theory of the physical basis of inheritance’. As this example shows, "model" organisms contribute not only to our understanding of the biological complexity of life but also help in removing our metaphysical presuppositions concerning organic phenomena.

Lest it should be thought that too much is being said about one organism, recent historical research shows that other organisms have also played similar roles in biological research (de Chadarevian 1998; Bonner 1999). In a detailed study of the role played by the nematode worm C. elegans in our understanding of development, de Chadarevian shows how the organism was used by Sydney Brenner to move from molecular biology, which ‘had become inevitable’, to the study of problems ‘which are new, mysterious and exciting’ in the domain of development (p 82). As had happened in the case of Drosophila, C. elegans was chosen because it fulfilled all the requirements that Brenner had specified: short life cycle, easily cultivable, and small enough to be handled in large numbers. But as Bonner informs us, Brenner had earlier thought of using the slime mould Dictyostelium for his study but since it lacked a nervous system he chose the nematode instead.

De Chadarevian’s paper offers a number of historical and philosophical insights regarding the use of this organism in the area of development and, in particular, developmental genetics. One such insight relates to the manner in which the worm at once offered enormous scope for genetic analysis and indicated a set of boundary conditions for the study of specifically developmental problems. Brenner, who had started with the idea of ‘microbiologizing’ and ‘taming’ the organism so as to study development directly through a complete genetic analysis, had to concede, 20 years later, that this was not possible. He had to accept that the original expectation that there would be a logic of development encoded in a genetic programme had to be abandoned; and he warned that the notion of a programme had to be handled carefully. He also saw that the representation of genetic space onto organismic space would not be a direct and explicit one. What was required in this move from molecular genetics to development was the realization that the cell was the basic unit of development and this necessitated an understanding of ‘how genes get hold of the cell’. Notwithstanding the advantages offered by ‘model’ organisms, their role in providing answers to all the major questions in genetics and developmental biology should not be over estimated. As Okada (1997) and Bonner (1999) point out, the use of a single organism in a specific research programme cannot by itself lead to overarching generalizations. In other words, model organisms play a double role: one, model organisms qua models indicate similarities between organisms. But at the same time, they also point to dissimilarities or differences, which implies that there is a limit to what the models can do under given circumstances.

M G Narasimhan