Fixing photosynthesis

C&I Issue 4, 2015

 Some two and a half billion years ago, nature evolved the two-step photosynthesis mechanism that all plants use to produce the oxygen we breathe. By combining two photosynthetic molecular complexes, pioneering microbes similar to today’s cyanobacteria acquired the ability to break up water molecules and liberate oxygen; this would not have been possible with either of the photosystems used individually. Some of these bacteria ended up inside plant cells and became the ancestors of today’s chloroplasts.

Without this innovation, our planet would have very little atmospheric oxygen, no ozone shield, and no life on dry land. For all its merit and gigantic impact, this crucial change in the history of life on Earth came with a tiny flaw that didn’t seem to matter at the time.

A tiny flaw

The trouble with this type of photosynthesis is that the key enzyme, rubisco – ribulose-1,5-diphosphate carboxylase/oxygenase – which is in charge of converting atmospheric carbon dioxide and water into organic molecules, isn’t very good at distinguishing the carbon dioxide molecule from the oxygen molecule and will use either of them. Initially, this lack of specificity didn’t cause problems because there wasn’t much oxygen around that might interfere with the desired fixation of carbon dioxide.

However, as species using this reaction and releasing oxygen as a waste product spread around the world and raised the oxygen content of the atmosphere to today’s level of nearly 21%, plants and bacteria have had to evolve new strategies and mechanisms to exploit photosynthesis in the presence of its byproduct.

C3 plants, for example, a group that includes most crop plants and incorporates the carbon dioxide from the air into a molecule with three carbon atoms, namely 3-phosphoglycerate, slowed down the turnover rate of the rubisco enzyme, allowing it to improve its selectivity. This came at the cost of having to produce more copies of the enzyme for a given turnover rate, which is one of the reasons why rubisco is now probably the most abundant protein on our planet.

Rubisco forms a complex of 16 protein molecules (eight of each subunit type) with no obvious functional relevance and its structure is thus interpreted as a way of cramming as much of the enzyme into the chloroplast as physically possible. Rubisco is the entrance to the Calvin cycle, the circular metabolism pathway that incorporates carbon dioxide into organic molecules, thus its efficiency ultimately determines the productivity of plants including crop plants.

Cyanobacteria, the closest present-day relatives of the microbes that originally developed two-step photosynthesis, kept the fast and poorly selective enzyme but built a protein shell around it. The carboxysome shell enables carbon dioxide in the shape of the negatively charged bicarbonate ion to access the enzyme, while the neutral oxygen molecule is excluded.

Meanwhile, C4 plants, which include many tropical species like maize (corn) and sugar cane, added a separate carbon dioxide fixation step upstream of the flawed rubisco enzyme. These plants first incorporate carbon dioxide into malate, which has four carbon atoms, hence the term C4 photosynthesis. C4 plants then transport the malate to a separate location where it releases the CO2 to be channelled into the Calvin cycle via the rubisco reaction.

Of these three possibilities, the first one that operates in the bulk of our crop plants is the least efficient by far. Estimates suggest that a switch to one of the alternative mechanisms could improve the yields of crops like wheat or rice by up to 30% while reducing the amount of nitrogen fertiliser required.

Constructing carboxysomes

In order to transfer the benefits of the cyanobacterial system to C3 plants, for example, one would have to replace the slow rubisco enzyme with the fast one and introduce the proteins that form the carboxysomes, along with any additional factors they may need for this. The group of Maureen Hanson at Cornell University, US, has now demonstrated that each of these two changes is possible in principle.

The researchers managed to express the carboxysomal proteins in tobacco plants using an approach based on the plant pathogen Agrobacterium tumefaciens, an established system for the experimental introduction of genes into plants. The proteins duly formed microscopically small vesicles, resembling carboxysomes. In addition, the researchers also succeeded in labelling a separate protein with a signal sequence for carboxysome incorporation, and they could confirm that this protein tends to associate with the vesicles.1

Then the researchers tackled the enzyme itself. Previous attempts to express the fast rubisco enzyme from cyanobacteria in plant chloroplasts had failed. The Hanson group tried two approaches, each time combining the rubisco gene with a different bacterial gene that might be helpful. In one experiment, they incorporated a molecular chaperone – ie a protein that assists the structure formation of other proteins by blocking undesirable interactions – and in another, they used a protein known to be involved in the encapsulation of rubisco into carboxysomes.2

In both cases, the researchers used tobacco plants and inactivated their own rubisco enzyme. They succeeded in producing transgenic plants with active photosynthesis. As was to be expected, the plants with the cyanobacterial enzyme but without the protective carboxysomes wilted and died in air, but in an atmosphere enriched with carbon dioxide they thrived.

The remaining challenge now is to introduce the complete package, ie rubisco, carboxysomes and any helper proteins, in a way that is suitable for plant propagation. As Dean Price and Susan Howitt put it in a commentary accompanying the second paper: ‘The advance can be likened to having a new engine block in place in a high-performance car engine – now we just need the turbocharger fitted and tuned.’

Spicing up rice

Other research groups are investigating the possibility of converting C3 plants into C4 ones. This conversion is likely to be very difficult, as it not only involves protein expression but also changes to the anatomy of the plant leaf.

Nevertheless, researchers at the International Rice Research Institute (IRRI) at Los Baños in the Philippines are trying to get the rice plant to use the C4 mechanism. The C4 rice project is led by Paul Quick from the University of Sheffield in the UK and is funded by the Bill and Melinda Gates Foundation. As rice is the staple food for nearly half the world population, any improvement to its yield would have an enormous impact on global food security.

The required changes are complex. On the anatomical level they involve the distribution of labour between two crucial types of leaf cells, the mesophyll cells and the bundle sheath cells (BSCs). In C3 plants, photosynthesis mainly takes place in the chloroplasts of the mesophyll cells, which make up the bulk of the leaf cross-section. The BSCs in these plants have a range of other functions, including regulation of liquid flows and hydraulic pressure, and they are grouped around the leaf veins. In C4 plants, however, the BSCs take up an important part of photosynthesis, including the rubisco reaction, shielded from ambient oxygen.

To enable this split between carbon dioxide assimilation and its introduction into the Calvin cycle, the leaves of C4 plants need to provide more space for the BSCs, give them more chloroplasts, and allow them close contact with the mesophyll cells so they can pass on the C4 compound for further processing. Many plants achieve this by increasing the density of leaf veins, so that the leaf mainly consists of veins sheathed by BSCs, which are in turn wrapped in mesophyll cells, leaving no space for bulk phase mesophyll cells. Thus, both cell types occur in comparable number and are always in close contact, enabling them to work together in photosynthesis. This arrangement is known as the Kranz – German for wreath – anatomy.

But how does one go about changing the leaf anatomy from the C3 version to the C4 version? In the absence of a detailed mechanistic understanding of leaf development, there is only trial and error.

The group of Erik Murchie at the University of Nottingham, UK, together with IRRI researchers in Asia, has used random mutagenesis and high-throughput screening on rice cultivars and managed to identify a number of mutants that display the desired leaf anatomy, ie with high density of veins.3 The researchers found that this trait appears to be tightly linked to a narrower shape of the leaf, possibly via the use of a shared growth factor such as auxin. While this link can help with rapid screening it may also hinder application of the mutated variants where the reduced leaf width leads to disadvantages in growth or yield.

In a recent review, Shanta Karki and colleagues at IRRI listed the changes that have to be made to turn rice into a C4 plant.4 In addition to inserting the basic C4 biochemical pathway into rice and increasing the leaf vein density, researchers will also have to find ways of increasing the number and size of chloroplasts in the BSCs to enable these cells to fulfil their new role in photosynthesis.

A further measure – essentially fine-tuning the division of labour between the two cell types – will be to reduce the activity of rubisco in mesophyll cells while increasing it in BSCs. Thus, in the engineered plant leaf, the Calvin cycle will mainly be active in BSCs, as is normally the case for C4 plants, rather than in mesophyll cells, where it happens in normal rice and other C3 plants. This reorganisation will also keep rubisco away from oxygen and thus solves the problem of photorespiration, the competing reaction in which rubisco uses molecular oxygen instead of carbon dioxide as a reactant.

These changes are a challenge for plant science, and there are still debates over which may be the most promising path. However, experts base their hopes on the observation that evolution has come up with C4 mechanisms at least 66 times independently in separate species, so it is clearly an option in the natural repertoire of plants, and researchers just have to find the right buttons to press.

Moreover, as Karki points out, recent technological advances have been speeding up gene identification and testing to an unprecedented extent, thus enabling the production of basic C4 rice prototypes for testing. In December 2014, IRRI announced the successful introduction of some C4 photosynthesis genes into rice plants, which were found to be active, even though the plants continued to rely on their normal C3 photosynthesis.

That said, IRRI researchers estimate that it may still take another 10 years for a C4 cultivar to be ready for the market. In the meantime, other research into improvements of rice and other plants, using genetic traits found in wild relatives to make them more resistant to diseases and tolerant of flooding, for instance, is also progressing. As highlighted in a recent special issue of the American Journal of Botany,5 science has a wide range of opportunities to link basic plant science to applied agriculture that can help provide food security for future generations.


1 M. T. Lin et al, Plant J., 2014, 79, 1.

2 M. T. Lin et al, Nature, 2014, 513, 547.

3 A. B. Feldman et al, PloS One, 2014, 9, e94947.

4 S. Karki, G. Rizal, W. P. Quick, Rice, 2013, 6, 28.

5 B. L. Gross, E. A. Kellogg and A. J. Miller, Am. J. Bot. 2014, 101, 1597.

Michael Gross is a science writer based at Oxford, UK  

Become an SCI Member to receive benefits and discounts

Join SCI