Source: SciDev.Net

Genetic engineering techniques are frequently proposed as ways to increase crop yields, especially in areas of the developing world where the people suffer from malnutrition and agricultural productivity is low. However, despite 40 years of biochemical and physiological research, there have been very few cases that led directly to improved cultivars with better yield.

This research has greatly improved our understanding of molecular traits and the factors associated with crop yields. However, the fact that there are so few examples of this research leading directly to better crop-yield suggests that we should be cautious how much molecular biology can improve crop yields in the near term.

The challenge of boosting yield potential

There are two basic approaches to increasing yield potential:

• Increasing plants' overall physiological capacity to produce harvestable yield under various environmental conditions (yield potential).
• Reducing the effects of biological stresses such as diseases, insects, and weeds that prevent plants from reaching their theoretical yield potential.

The first approach depends on the straightforward logic of identifying a crop plant's specific function or functions that could be improved to increase yield potential. Scientists have used this approach to target particular metabolic 'control points' so that plant genes thought to limit a crop's basic yielding capacity could be targeted for improvement.

The concept of modifying a crucial biochemical or physiological step to achieve yield increase is not new, however.  The advent of 'scientific agriculture' after World War II resulted in previously unimagined increases in crop yield. Relatively cheap fertilisers became available, and the addition of quantities of specific nutrients to the soil, especially nitrogen, was an important factor in increasing yields.  So also was the development of plant varieties suitable for high-fertility conditions. National averages in various countries reached six to eight tons per hectare.

But progress was slow in research aimed at improving specific biochemical or physiological traits.  This was certainly not due to lack of effort.  Scientists identified superior cultivars and established genetic heritability for some important traits that they thought were associated with yield.  However, few improved varieties with enhanced yield were developed using this approach.

There are several explanations for these disappointments.  First, the anticipated benefits were sometimes simply not transferable to the field situation.  The activity of a metabolic pathway can be enhanced or diminished, but it may be irrelevant in increasing crop yield.

Second, molecular modification often had a greatly diminished effect on crop growth as its influence moved up through a crop's organisational hierarchy, from the molecular level to the organelle, cellular, organ, plant and crop levels. Biochemical compensation by other pathways may also moderate the effect of the original modification

Finally, the challenge that appears to have often proved to be a critical limitation was a research structure in which it was difficult to carry a concept of genetic variation from the process level through to the development of a viable commercial variety.  The rare successes in the past in developing crop varieties using this approach illustrate that this research requires integrated, multidisciplinary teams with career and financial commitments to sustain long-term research efforts lasting 12 to 15 years.

Possible target traits for genetic transformation

Photosynthesis

Photosynthesis is the process by which plants use sunlight to assimilate carbon.  Considerable early excitement was generated as genetic lines with superior leaf photosynthetic activity were identified and the photosynthetic capacity was successfully bred into progeny lines of several crops, including maize, wheat, and soybean.  Disappointment followed when these 'improved' lines failed to produce yield increases.   

The reason for this disappointment is that the benefits of improved photosynthesis at the cellular and leaf level do not translate directly into larger grain yield.  In particular, improved photosynthesis — which boosts carbohydrate production — may lead to larger vegetative growth, which will increase the plant's demand for nitrogen.  Unless the plant can take up more nitrogen from the soil, grain yield may actually decrease.  This is because the limited supply of nitrogen may end up producing bigger plants rather than more grain.  The plant's needs for resources such as carbohydrates, nitrogen and sulphur need to be addressed together.

Nitrogen assimilation

In the past, a plant's nitrogen accumulation has been a crucial feature of yield increases.  Usually, this has resulted from making more nitrogen available to plants (for example, by applying nitrogen fertilisers to the soil), and by breeding new plant varieties that can take up and store more nitrogen.  A key factor in these successes has been improving the amount of accumulated nitrogen that gets stored in the grain, rather than being 'locked up' in the plant's vegetative tissues.

A target for genetic engineers has been to increase 'nitrogen-use efficiency.'  However, efforts to improve a plant's nitrogen metabolism by changing its genome are unlikely to succeed because plant biochemistry is already extremely 'efficient' in nitrogen uptake and use.  It may be difficult to improve whole plant traits because it seems unlikely that engineering a single or even a few genes can easily manipulate these traits.

Seed growth

A large amount of past research has focused on increasing seed growth rates and overcoming the problem of seed-embryo abortion.  However, researchers have found that plants are well endowed with redundancies and backups to optimise grain production in a range of environments.  Plants tend to compensate for an increase in one factor by decreasing others.  Research has shown that in a community of plants, if one seed is growing more quickly, the plant compensates by changing the number of seeds or the duration of seed growth, resulting in little or no increase in overall yield.

In a few isolated cases, researchers have succeeded in improving seed growth rates by genetically engineering plants to be less sensitive to phosphorus feedback inhibition in the grain.  However, they found that seed growth was accompanied by increases in the growth of individual plants.  As a result, there was no change in the overall harvest index (the ratio between the harvested grain and the total accumulated crop mass) for rice and only a small increase in wheat.

Drought stress

Scarce water is a critical limitation on crop yield in many places. How efficiently a crop uses water directly influences yield potential, but it is not very flexible because of the physical and physiological characteristics of gas exchange in leaves (transpiration).  Although there has been some success in improving water-use efficiency in wheat, any major increases in yield still depend on more water being available. This means that the crop must access more soil water in dry-land conditions.

Much biochemical and physiological research has focused on drought tolerance, to enable plants to survive long periods of drought.  However, for most annual grain crops, a drought severe enough to threaten the plant's survival will inevitably result in such a low yield that survival is a moot point.  Therefore, there is probably no point in trying to use inserted genes from so-called 'resurrection plants' to sustain food production during droughts.  The original premise for the trait is irrelevant to the cropping situation.

Successful research to achieve yield increase

Although this wealth of research has considerably improved our understanding of plant growth and crop yield in grain crops, a targeted approach of increasing specific physiological traits has resulted in very few improved cultivars with increased yield potential.  The lack of success illustrates the difficulties involved in translating insights and breakthroughs at the micro level into real improvements in crop quality at the macro level.

These failures offer important lessons for molecular genetics research, which is even further removed from grain yield.  The lesson is that genetic engineering research will probably confront many of the same obstacles that have limited the impact of previous biochemical and physiological research.

Nevertheless, there are a few successful examples in which physiological research and genotypic selection has played an integral role in developing useful new cultivars that produce better yields.  Three such cases are described in boxes 1, 2 and 3.

The rare successes of increasing crop yield potential by altering a targeted physiological trait may offer important lessons for achieving success using GM techniques.  It is clear that the existence of genetic diversity for a particular trait — whether from natural diversity or transgenic techniques — is only a small, first step in achieving measurable yield improvements for farmers.  The challenges have been to understand the altered trait's effect on crop performance in the field and to exploit the trait in a crop-breeding programme.  The examples described in the boxes had characteristics in common, that were crucial in achieving success using the physiological approach. 

(1) Early assessment of the potential beneficial trait.

Early in each programme, the researchers gave considerable attention to understanding and documenting the trait of interest under field conditions, rather than relying on extrapolations from laboratory study.  Under what conditions would the trait be beneficial?  What are the consequences of trait expression on crop performance?

Integrating genomics, mapping and physiology might enable scientists to develop molecular hypotheses that begin at the top of the trait hierarchy rather than in the laboratory.  Input from whole-plant physiologists and agronomists will be needed to make an early assessment of how proposed genetic modifications might improve the plant.  It may also be advisable to use systems analysis technology to make an early assessment of the proposed GM trait or crop's economic or commercial viability.

(2) Effective phenotyping of genetic modifications

A crucial challenge in using a transgenic approach to improve yield is forecasting what will happen when a transformed trait is expressed. Trait expression depends on both the physical environment in which the plants grow and the genetic environment into which the trait has been inserted.

The past successful studies gave considerable attention to characterising trait expression under a range of field environments.  It was not sufficient simply to know that the genetic advantage existed in the plant.  Rather, it was necessary to document the trait's level of expression under a range of conditions in which successful varieties will be commercially grown.

Phenotypic expression requires extensive testing, including evaluating plant performance and yield in a cropping situation.  This is likely to increase the demand for rapid and inexpensive methods for phenotyping plants, especially for traits that do not have a readily visible expression.

(3) Multi-disciplinary effort

The successes documented in boxes 1 to 3 involved contributions from different disciplines throughout the research effort, including crop physiologists, agronomists and breeders.  These three disciplines are still important and necessary in developing GM plants offering new genetic variability.  The molecular genetic approach merely adds a further layer to this team.  Indeed, system and environmental analysts may need to be involved as well, to assess where and when the expression of a GM trait might be commercially beneficial.

The GM approach will likely need early involvement of all disciplines simply to move the new plants forward for field assessment.  In past successes, researchers used early field screening to identify candidate lines with both the desired trait and a reasonable capacity for growth under field conditions.  GM plants may require early attention from all participants, to move the trait into more viable plant material before starting trait assessments.

This interactive, team-based approach will require the team to be well integrated and coordinated at all stages of the research programme.  This will probably present new challenges, especially for public research organisations.

(4) Long-term commitment

The examples show that it took 12 to15 years to move from the initial studies of a physiological trait to the release of a commercial variety with improved yield.  Although the molecular genetics approach can speed up some of the steps involved, additional efforts to fully document the consequences of introducing a new trait could well offset this gain.  It seems likely that a successful programme to generate improved cultivars will require a team of scientists to work together for more than a decade.  This may be difficult because the traditional time horizons among the various disciplines tend to differ widely.  Further, commitment to a team effort may be difficult when the probability of individual recognition may not be high.

The greatest limitation, however, may be the financial commitment required for long-term team research.  Public funding based on a series of two- to three-year grants does not encourage initiating team research to undertake a high-risk, multi-year project.  Private companies may be in no better position to fund long-term research, because of the large uncertainty about eventually generating a commercially viable product. 

Reprinted from Trends in Plant Science, 9, Sinclair et al, Crop transformation and the challenge to increase yield potential, pp 70-75, Copyright (2004), with permission from Elsevier