Assembling plant RuBisCo in E. coli

Optimising photosynthesis is a significant goal for those researching ways to secure the human food supply on a planet that will continue to see the emergence of threats to agriculture from climate change. We’ve written a number of articles in this vein, predominantly on the effort to be able to engineer C4 photosynthesis into major crops that use the less efficient C3 photosynthesis (see here, here, here and here for a few examples).

In the research article “Plant RuBisCo assmbly in E. coli with five chloroplast chaperones including BSD2″ published in Science, researchers from the Max Planck Institute of Biochemistry sought to install a functional RuBisCo enzyme from our favourite model organism, Arabidopsis thaliana, into E. coli, the model bacteria with a short generation time.

The researchers explain that the purpose of attempting to install plant RuBisCo into the bacteria is to assist the effort seeking to improve the catalytic efficiency of the enzyme, allowing it to be studied closely in conjunction with genetic engineering to hasten the results of this line of research.

The genes that transcribe the subunits of RuBisCo have been known for some time and the arrangement of eight large subunits (RbcL) and eight small subunits (RbcS) is well characterised. But as with the ongoing quest to understand protein folding patterns, the process by which the holoenzyme comes to be in its final conformation, which supplementary proteins are required to assist the assembly of the enzyme and therefore which supplementary genes need to be expressed for the functional formation of the enzyme, are not well understood.

By attempting to build the enzyme from scratch, we learn more about the necessary genetic components required to build it, as these researchers found.

The Study

The researchers first attempted to express A. thaliana RuBisCo in E. coli by transferring the RbcL and RbcS genes along with two additional factors, RbcX and Raf1 (the functions of which were further elucidated later in the paper), but this attempt was unsuccessful.

Therefore, they attempted the feat again using three separate plasmids, one containing the RbcL and RcbS components, another containing chloroplast chaperonin proteins (C60αβ/C20), and a final plasmid containing four assembly chaperones (Raf1, Raf2, RbcX and BSD2). Using separate plasmids allowed the researchers to test which components were required to form functional enzymes. Expressing all genes resulted in the formation of proteins that migrated on polyacrylamide gel electrophoresis at the same position as AtRuBisCo and which had a similar affinity and carboxylation rate of CO2 as native RuBisCo. Plasmids that did not express either sets of chaperones failed to form proteins that migrated to the same position.

Accordingly, the RbcL and RbcS genes plus the genes encoding the various chaperones transferred into the E. coli appears to have resulted in the expression of functional RuBisCo.

Chaperone Requirements

Being able to functionally express RuBisCo from a higher order plant in E. coli is a significant step forward, but understanding the purpose and need for each of the chaperones remains outstanding. To probe a bit deeper into the requirements for functional RuBisCo enzymes, the researchers varied and substituted a number of the genes and monitored the outcome.

The first reported variation was to substitute the C60αβ/C20 chaperones with GroEL/GroES, endogenous chaperones from E. coli. The researchers found that even overexpressing the endogenous chaperones couldn’t match the level of functional expression of RuBisCo compared to when A. thaliana chaperones were used. Cpn60α and Cpn60β were both required for efficient expression of the enzyme, while expression of a plasmid without C20 still resulted in the formation of RuBisCo, albeit with a lower rate of expression. Overexpressing GroES in the absence of C20 resulted in the production of RuBisCo with similar efficiency as when C20 is coexpressed.

RuBisCo in E Coli1

Figure 2 from article. Chloroplast charperonin requirements to form functional RuBisCo. Only when the RbcL, RbcS and folding chaperones were expressed with C60αβ/C20 (lane 2) or with C60αβ/GroES (lane 9) did RuBisCo successfully form.

In relation to the Raf1, Raf2, RbcX and BSD2 factors, the researchers deleted each of these in turn to test the effect on the final outcome of the enzyme to gain further insight into the function of each.

Deleting Raf1, Raf2 or BSD2 resulted in no RuBisCo production. Previous research has demonstrated that Raf1 works to assemble the RbcL subunits into RbcL8, whilst the function of Raf2 remains to be diagnosed.

RbcX deletion didn’t completely halt RuBisCo production but reduced assembly rates by 50% to 60%.

BSD2, what does it do?

The research delved much deeper into the function of BSD2, a gene that isn’t contained in all photosynthetic organisms and appears to have evolved subsequent to the endosymbiosis of the organism that led to chloroplasts, unlike the other three chaperones.

It was theorised in the paper that, given the possible evolutionary origins of the protein, that it may be related to the importation of the RbcS subunits into the chloroplast. When RbcS wasn’t expressed, a RbcL8/BSD2 complex formed. When RbcS was reintroduced, the RbcL8/BSD2 complex no longer formed; instead the RuBisCo RbcL8S8 enzyme formed. Omission of RbcS and Raf1 resulted in the lack of formation of the RbcL8/BSD2 protein, suggesting that Raf1 plays a role in the formation of this intermediary phase of the holoenzyme.

Given that it appears BSD2 plays a role late in the construction of the RbcL8S8 enzyme, the researchers tested whether altering the BSD2 and RbclS expression would have similar affects on the cyanobacterium Synechococcus elongatus. Without the expression of SeRbcS, SeRbcL and BSD2 formed a SeRbcL/BSD2 complex which, when SeRbcS was reinstroduced, resulted in fully formed SeRuBisCo, similar to what happened with the Arabidopsis based genes when expressed in E. coli. Therefore, this experiment helps to confirm the hypothses that BSD2 play a late but significant role in the final formation of RuBisCo.

RuBisCo in E Coli2

Figure 6D from article. Model of protein folding sequence and required chaperone involvement to form functional RuBisCo.

Taking our knowledge of this process a step further, the structure of the BSD2 protein and how it may interact with RbcL to form a RbcL8/BSD2 complex were analysed.

BSD2 imaging showed that it forms a hairpin structure that centres around two zinc atoms. Imaged as part of the RbcL8/BSD2 complex, the crystallised structure shows that two BSD2 molecules are bound within antiparallel RbcL2 units, appearing to join the two RbcL subunits together. One of the zinc-containing regions of the BSD2 interacts with the C-terminal end of one RbcL unit while the other zinc-containing region interacts with the N-terminal of the adjacent RbcL subunit. It appears that the conformation of the RbcL subunits when interacting with BSD2 allows RbcS subunits to access and bind to the required sites on RbcL. Mutating the BSD2 at contact sites resulted in a loss of function of the protein.


This piece of research will hopefully assist to increase efficiency in RuBisCo research, research which aims to increase photosynthesis efficiency with resultant improvements in crop yields and food security. The research is somewhat Feynmann-esque, referring to the note scribbled on his chalkboard when he passed that:

“What I cannot create, I do not understand”. Richard Feynmann, February 1988

Although this research is useful in its application for further research, the attempt to recreate the enzyme from the scratch also helps us to further understand the function of those genes and moving us closer to the depth of knowledge Feynmann aspired to.

Great research resulting in a method that we expect to see in use and built upon in future.


Host-delivered RNAi to control Wheat Leaf Rust

Wheat is a major staple food worldwide. Accordingly, threats to wheat production pose a significant risk to global food production and security. One major threat to wheat production is Wheat Leaf Rust (“WLR”), a disease caused by the fungus Puccinia triticina which has the ability to cause severe yield losses. Of further concern, the application of fungicides to control the disease can lead to the development of reduced sensitivity or resistance to the fungicide, let alone the economic and environmental impacts that fungicide use entails.

Given the serious threat that WLR entails and the problems with using fungacides, genetic studies have been used to identify a number of resistance genes which have been used to breed cultivars with greater resistance to the fungus. However, the use of resistant cultivars in monoculture applies significant pressure on the fungus to evolve and develop new methods of virulence.

The authors of the paper being presented today outline these important problems noting that, when combined with the effects of a changing climate, disease outbreaks are at risk of increasing and further threatening food security.

A Potential Solution?

After discussing the life cycle and infectivity of the fungus, the researchers behind the paper published in the Plant Biotechnology Journal theorised that by using the increasing genetic knowledge about the pathogen, particularly the genes implicated in their virulence factors that enable it to infect wheat plants, it may be possible to target and repress those genes using RNA-interference (RNAi).

We have written about RNAi previously in the context of RNAi sprays as a possible alternative to traditional pesticides, potentially being used in Bioclay-based pesticide and in transgenic plant pathogen resistance.

The researchers outline the possibility of using host-delivered RNAi (“HD-RNAi”), inserting genetic material into the plant that will translate into a double-stranded piece of RNA complementary to one of the virulence genes of the pathogen. When the dsRNA piece is cleaved into small interfering RNA (siRNA) duplexes in the plant and delivered to the pathogen, the siRNAs bind to the transcripts of the complementary virulence gene in the parasite and interfere with their further translation.

Using recent studies that characterised virulence genes of rust fungi, the researchers chose to target the Pt Map-kinase (PtMAPK1) and Pt cyclophilin (PtCYC1) genes, attempting to silence their translation and discover whether effective silencing of these genes will afford any protection to the plants against the fungus.


HD-RNAi constructs targeting each gene were transformed (separately) into immature embryos of a WLR susceptible wheat cultivar via projectile bombardment, creating 14 transgenic To lines harbouring the hp-PtMAPK1RNAi construct and 20 To lines aimed at silencing the PtCYC1 gene. These lines were selfed to produce T1 seeds. Polymerase chain reaction (PCR) was used to verify successful insertion of the constructs and reverse-transcription PCR (RT-PCR) was used to verify transcription of the construct. Furthermore, there was no effect on phenotype seen in the transformed lines.

The transformed lines and a set of controls were then challenged with the Pt urediniospores when they were two weeks old. 12 days after being challenged, the plants were scored on a 0 to 4 scale depending on the size and abundance of uredinia, with 0 being little to no infection and 4 being highly susceptible.  A total of 13 of the transformed lines (6 of the PtMAPK1, 7 of the PtCYC1) showed varying resistance to the fungus compared with the controls. Those lines with the greatest levels of resistance showed symptoms of infection between 1 and 3 days later than the control lines did.

LWR Fig 1

Figure 1 from article showing WLR infection levels in select transgenic lines compared to control (on the left).

To further investigate whether the HD-RNAi were the cause of the difference in susceptibility to the fungus, the researchers looked at the rate of fungal growth within the leaves of the plants by comparing the ratio of fungal DNA to plant DNA between the transgenic lines and the control lines. Analysis showed that the amount of fungal biomass was reduced by nearly 80% in each line compared to the controls. Accordingly, it appears that the HD-RNAi constructs have managed to interfere with the fungus’ ability infect and grow in the plant.

Using the quantification of fungal biomass to group the plants with differing levels of WLR resistance, the connection between the level of resistance and the severity of disease was tested. It was found that there was a close relationship between the two criteria, with the lowest disease severity correlating with the lowest ratio of fungal biomass to plant biomass.

LWR Fig 5

Figure 5 from article. (a) Comparison of fungal and plant biomass within infect control and transgenic wheat lines. (b) Transcript levels of genes targeted by the RNAi construction compared to control. (c) Confocal microscopy of infected cells.

To test whether the HD-RNAi constructs were actually interfering with the targeted fungus genes, quantification of transcript levels of the fungus genes was performed 5 days post infection of both transgenic and control lines. Transcript levels of both target genes was reduced by between 41% and 65% in transgenic lines compared to controls, and the reduction of transcript levels of each particular gene was specific to the transgenic line expressing the HD-RNAi construct targeting that specific gene. Further, the measured transcript levels correlated with the level of disease severity in the plant, demonstrating a link between the reduction in target transcript, level of fungal biomass and disease severity.

Finally, three transgenic lines from each HD-RNAi transformation showing increased disease resistance were selfed to obtain T2 generations of each. PCR and RT-PCR confirmed the presence and translation of the gene construct in that generation and the plants were challenged with Pt urediniospores when they were two weeks old. Similar to the T1 generation, infection scores were lower in the transgenic lines compared to control lines, with a similar delay in onset of symptoms previous described.

The same analysis of fungal biomass and transcript abundance was measured, again with similar findings compared to the T1 generation.


The results of the research are promising, and the discussion section of the paper explored the possibility of targeting more than one virulence gene in any one transformation event or targeting different virulence genes as we learn more about the genetics of the fungus.

Further discussion centres on how the HD-RNAi constructs are delivered to the pathogen, a process not fully understood. It is hypothesised that the HD-RNAi transcripts may enter the pathogen via vesicle-mediated transport across the haustorial interface. However, the delivery method requires further research to confirm this and to confirm how inter-kingdom RNA is able to interfere with one another.

Why there was a variation in resistance levels amongst the different transgenic lines was also discussed, with hypothesis such as incomplete knockdown of the target gene, variability of gene expression levels and target gene transcript turnover rates.


The study builds upon our growing knowledge of plant and pathogen genetics, pairing the two with our understanding of RNAi to explore the possibility of controlling a significant and increasing threat to wheat production and avoiding the fiscal and environmental problems that come with increasing fungicide use. The results of the study are promising and should direct further research. This further research may consider whether different pathogen genes can be targeted, whether targeting different combinations of genes will have greater effect, and testing whether the problem of forced evolution of the pathogen to overcome our advances will mean that our crop defence strategies will still require us to employ agricultural practices to reduce the evolutionary pressure our genetic strategies apply.

A great study with possible applicability to a range of pests in other important crops.

CRISPR-S: Quick and Cheap Selection of Transgenic Rice Plants

The CRISPR/Cas9 (Clustered Regularly Interspaced Short Palindromic Repeats) genome-editing technology is, by itself, a significant recent addition to the biologists’ toolkit. The tool, which gives researchers the ability to direct gene edits (whether that be single point mutations, deletions of a chunk of the genome or the addition of new genetic codes) to specific sections of the genome, has seen tremendous uptake in use both in systems biology research and genetic manipulation studies.

Not only does the tool have the impact of a sledgehammer with the precision of a laser, but manipulations of the tool have resulted in a swathe of CRISPR-derived tools that either improve the tool in some specific way or alter the effect that the nuclease has on the genome being edited.

A letter published in Wiley’s Plant Biotechnology Journal in August this year described a modification of the CRIPSR tool to allow easy identification of successfully edited rice plants and simple identification of second generation plants which, while having the desired genetic modification, no longer contained the piece of transfered CRISPR/Cas9 genetic material within the genome.

The CRISPR Modification

The paper starts with the researchers noting that in the course of editing genomes in plants there are two steps that need to be successfully completed.

To edit a plant genome, the CRISPR/Cas9 genetic material needs to be transferred and inserted into the genome to be edited. This is usually performed with Agrobacterium tumefaciens which randomly inserts the CRISPR/Cas9 sequences, promoter etc into the plant genome. When the sequence is subsequently expressed, the guide RNA and the Cas9 nuclease are expressed which in turn go on to attach and edit the desired part of the genome.

For this first step to be successful, the researchers not only need to ensure that the required modification is going to be made, but must also identify the regenerated plants that have been successfully modified. Although the rate of successful editing can be quite high it is not usually 100% effective and the process of confirming a successful edit within the first generation of the modified plant can be time consuming and expensive in itself.

Once the successfully modified plants have been identified, the CRISPR/Cas9 gene insert must be removed from subsequent generations of the plant so that only the modification to the gene of interest remains. Again, this can require time consuming and expensive genetic screening processes to identify the successfully transformed, but CRISPR/Cas9 gene-free, plants.


The researchers who penned the paper theorised that it may be possible to create a gene insert that self-reported whether it was successfully introduced by causing a change in phenotype (aside from the change being introduced by the CRISPR element itself). To do this they proposed that, by introducing a gene segment into the plant that contained the CRISPR/Cas9 sequence plus a sequence that transcribed a piece of hairpin RNA complementary to some other gene of known function, the result would be the required genetic modification plus RNA interference of a marker gene that could be used for identification of successful transformation.


Figure 1 (e) from the article demonstrating the successful transformation of the gene in bentazon-susceptible plants (top left), the transgene-silenced and therefore resistant plants (top right), the second generation plant with the edited cadmium transport gene but lacking the CRISPR transgene, therefore regaining bentazon resistance (bottom left), and the edited second generation plant with the transgene remaining in situ (bottom right).

To demonstrate the theory they sequenced a piece of DNA to insert into rice plants. The DNA being inserted contained a CRISPR/Cas9 segment which directed an edit to a gene which encoded a cadmium transporter, plus a segment of DNA encoding a piece of hairpin RNA that, if successfully inserted into the genome of the plant, would result in the plant being susceptible to the herbicide bentazon. Bentazon susceptibility would then be the marker for successful transformation of the first generation of plants (which would assume successful editing of the gene for cadmium transportation) and also serve as a marker for those second generation plants that had lost the CRISPR gene segment given that the lack of presence of that gene would then restore the plant’s resistance to bentazon.

Testing the theory

The researchers created the sequence and inserted it into japonica rice genotype Xidao #1 via Agrobacterium tumefaciens. The tillers of the first generation of plants were divided and grown as two subplants, one of which was sprayed with bentazon, the other which wasn’t (meaning that if one subplant was susceptible to the herbicide and died, thus confirming successful transmission, the other subplant also with the successful transformation remained). Of the 96 independent transgenic events, 29 subplants were susceptible to the herbicide with 67 subplants unaffected.


Figure 1 (b) from the article demonstrating the bentazon resistant (and therefore unedited) first generation plant (left) and susceptible (and therefore edited) first generation plant (right).

The 29 subplants with the RNAi affecting bentazon resistance were then sequenced at the site of the cadmium transporter gene to test whether the gene segment had not only created the hairpin RNA but had also edited this particular gene. The researchers found that all of the plants displaying the susceptibility marker also contained the desired gene edits at the cadmium transporter gene.

Thirdly, the researchers used quantitative real-time PCR to test the abundance of the Cas9 and the bentazon susceptibility genes to check whether it was their gene construct that was causing the changes being noted. They found that in the susceptible offspring the transcripts of the Cas9 gene was significantly higher and transcripts of the gene that normally confers resistance to bentazon was significantly lower (causing the susceptibility to the herbicide) than those found in the resistant offspring. Thus, it would seem that the gene construct had successfully resulted in transcription of the Cas9 nuclease and the RNAi construct, the latter reducing the transcript abundance of the bentazon resistance gene.

To test whether their marker construct could also be used to distinguish transgene-free second generation rice plants they created 16 lines (72 seedling each) from the successfully edited first generation plants. They again sprayed the plants with bentazon with the hypothesis that those generations which had lost the transfer DNA containing the CRISPR/Cas9 and RNAi gene segment would regain their resistance to the herbicide. Of the plants that died, subsequent analysis proved that the transgene DNA remained in the genome, while the resistant plants had lost the transgene segment in the recombination of genes during the creation of the second generation.


It appears that these researchers have created a method, combining CRISPR/Cas9 and RNAi, to enable easy identification of successfully transformed plants and to then confirm that the transgene component has been removed in subsequent generations. This is particularly useful in those countries where current legislation does not consider plants that have been edited to remove a section of DNA (as opposed to being edited to add a piece of DNA) as “genetically modified”, therefore avoiding regulatory control. Being able to easily confirm that gene deletion has occurred and then confirm that the transgene has been removed will reduce the time and cost associated with confirming the modifications using gene sequencing.

It will be interesting to see if the process can be used in other plants. The time and expense spared could see the speed of research projects increase and open up the technology for projects that lack funding to allow for repeated primer creation and PCR-based sequencing to confirm genetic alterations.

A Study on Studies that Report Adverse Effects of Transgenic Food & Feed

Our policy at the Legume Laboratory is to report on the latest science surrounding the way we feed ourselves and possible methods of feeding a growing population in a world of limited resources and escalating environmental pressures.

Although we have a keen interest in the public discourse surrounding genetically modified crops, delving into popular arguments against their use falls foul of the line we have drawn in the sand for what we will report on.

However, an article in the Plant Biotechnology Journal reviewed a number of studies that are usually cited by proponents of the anti-GMO movement as evidence of the adverse health effects of these crops. Although this is usually one tenet of a larger argument against the use of genetically modified food, the use of published scientific papers to support part of the argument opens it up the rigors of scientific critique. This review gave us the ability to write about the subject.

Which papers were reviewed?

Before deciding which papers were to be reviewed, the authors of the review point out the noticeable foible whenever reading a commentary on genetically modified crops – that the only commonality between the variety of “GMO” foods is the process of creating them. Although transgenic crops may be created by the same process, the type of modification made or the trait inserted into the crop are individually known as ‘events’. The paper reports that while the FDA has reviewed 153 individual events, the majority of studies that report negative effects and are the focus of negative commentary relate to relatively few events. Although the studies reporting adverse effects are highly publicised, studies reporting no safety concerns garner little public attention.

In selecting articles for review, the researchers took the citations listed in four review papers which concluded that there are negative effects related to transgenic agriculture (Domingo and Bordonaba (2011); Dona and Arvanitoyannis (2009); Magana-Gomez and de la Barca (2009) and Seralini (2011)) plus those cited by the organisations ‘GM Free USA’, ‘Coalition for a GM Free India’ and ‘GM Watch’. Of these, studies not published in scientific journals or which didn’t evaluate whole food/crops but only pure proteins were excluded.

The result was 35 studies which met the criteria. The authors make a number of observations about the selected studies, including that:

  • they make up fewer than 5% of the total number of studies assessing the safety of transgenic crops;
  • nearly half (43%) of the studies looked at a herbicide resistant soy bean the product of one particular transgenic event;
  • 23% of the studies focused on another single event, an insect-resistant maize;
  • 9% studied crops not commercially available;
  • 20% of the studies did not indicate which transgenic event was being evaluated;
  • 31% of papers were authored by the Malatesta group at two particular Italian universities;
  • Only one article was published in a high ranking journal – (Ewen & Pusztai (1999)). The article was published with an analysis written by the editors which elucidated several significant criticisms of the study. The editors also commented that by publishing the article they did endorse the study or its conclusion but instead thought that there was a public interest in publishing it so that it could be evaluated by other scientists and to avoid cries of suppression from scientists claiming the existence of these negative effects.

Conflicts of interest

We start this section by first pointing out that the authors of this paper declare their own conflict of interest, noting that the lead author is from a lab which obtains funding from GM developing companies, while the second author has performed publicly-funded research into transgenic crops.

All of the 35 papers they studied however contained no declarations of conflict. Further analysis by these researchers revealed that half of the papers did not provide any information on where funding for the research was obtained while 40% of the papers showed no conflict of interest due either to funding sources or personal affiliations of the authors to groups known to have particular positions on transgenic agriculture.

What about the quality of the studies?

In order to assess the quality of the studies the researchers critically analysed the studies using a standard used by a number of authorities and outlined in a number of papers to assess the safety of the transgenic food. The criteria included:

  1. Control and experimental varieties should be from matching or very similar genotypes, be grown in the same fields, under the same conditions and in the same seasons to avoid confounding factors affecting the results;
  2. Statistical tests should be chosen before the study is performed, avoiding tests being chosen or changed throughout the experiment to obtain a desired result; and
  3. If results show some difference between the transgenic and control crops, the study should be contrasted with similar previous studies that obtained different results and a hypothesis put forward as why there is a discrepancy between results.

In finding that all of the papers failed to meet one or more the criteria the researchers point out problems with each of the studies, some of the bigger issues being:

  • Three of the papers did not perform any experiment and instead analysed preexisting data with different statistical methods to find significant differences between the study and control groups where the original paper did not. Two of these papers managed to find significant effects at low doses but at higher doses the effect was not seen, contrary to the dose-response reaction normally seen in toxicology studies;
  • Ayyadurai and Deonikar (2015) predicted increased formaldehyde and decreased glutathione in GM soya bean crops based on reviewing data from 6,000 studies without specifying the data used and without validating the prediction experimentally;
  • A series of studies by the Malatesta group contained multiple flaws, particularly that levels of isoflavones, which themselves can have a physiological effect on mammals, weren’t measured between transgenic and control crops, adding an uncontrolled confounding variable into the studies;
  • Of four studies emanating from Professor Infascelli’s group, two  were investigated, retracted and the authors reprimanded for digitally manipulating images used in the papers which found GM feed detectable in goats;
  • Finamore et al. (2008) studied the ingestion of maize in mice, finding that it caused an immune response. However, they dismissed the effects of mycotoxins on the immune system stating that they were at acceptable levels when in reality they were twice as much as allowed;
  • Trabalza-Marinucci et al. (2008) don’t advise the variety of maize used in the study nor mention mycotoxin effects while Magana-Gomex et al. (2008) have a similar problem, not specifying whether genetically similar soya beans of similar origins were used;
  • Seralini et al. (2014), a republished article of one previously retracted in 2012, reported elevated tumour and mortality rates in rats. The paper was heavily criticised for such reasons as:
    • the reported tumour rate is within the normal range for the species of rat experimented with;
    • the data was statistically analysed in multiple ways and, when corrected for multiple comparisons by scientists reviewing the study, the results obtained by the authors disappear;
    • the 10 rats used was too small a sample size for the type of study;
    • a dose-response relationship wasn’t observed;
    • the results section contained pictures of treated rats with huge tumours but no pictures were included of the control rates;
    • the results conflict with previous studies.
  • Fares and El-Sayed (1998) concluded that potatoes with added Bt protein – added as a crude extract without analysis of the crudeness of the extract – caused adverse effects.
  • Some examples cited do not show any adverse effect, contrary to how they are reported in the reviews and websites that they are cited in.

In contrast to the small set of error-prone reports which conclude that there is a link between genetically modified food and adverse health effects, there are somewhere around 473 papers that do not have the same quality problems. These papers cover a significantly greater number of transgenic events. Throughout all of these studies, there have been no unfavourable or concerning findings made about adverse health effects.


The weight of rigorously obtained evidence is firmly supportive of the safety of transgenic crops. Although safety concerns should always be investigated, given the weight of evidence in support of the safety of transgenic

Does C4 photosynthesis occur in wheat seeds?

In our humble opinion, the debate about the conclusions drawn in a recent paper regarding the possibility of C4 photosynthesis being active in wheat seeds is one of the more interesting scientific debates raging at present.

The debate started with a paper published by Rangan et al last year which described findings they had made about differential expression of genes implicated in C4 photosynthesis within wheat seeds. These expression levels differed to the gene expression levels in other parts of the plant. We wrote an article describing the research here.

The response to that article (and our post) was that the evidence provided, although novel and interesting, wasn’t enough to justify the conclusion that C4 photosynthesis was in fact occurring within the wheat seed. Noting the criticisms, we wrote an updated piece covering two published responses to the Rangan article.

But the debate hasn’t stopped there…

Back and forth we go

The journal Plant Physiology published a series of letters between the researchers who published the initial research and two researchers with the differing opinion. The series of letters outlines the main points of difference between the propositions.

In defence of the C4 pathway conclusion

The series starts with a defence of the conclusions drawn from the research. Citing the earlier work of Bort et al (1995), a paper which described labeled carbon assimilation differences between leaves and seeds of wheat and barley and which found no significant difference between the location of the assimilated labeled carbon, the defenders argued that the research did not distinguish between carbon assimilated in the glumes covering the seed (the Bort et al paper) and carbon assimilated in the pericarp (the Rangan et al research).


Dissection of Wheat Glume showing glume and grain.


Figure from article showing cross-section of wheat seed, particularly the pericarp.

The distinction between the two asserted is that the Rangan et al research shows that pericarp, as opposed to the glume, demonstrates elevated C4 gene transcription.

Further, it is suggested that the carbon source supplying the pericarp is not from the capture of external CO2 but instead comes from the endosperm capturing respired carbon, the carbon being derived from bicarbonate in the developing seed tissue, and moving outwards to the pericarp for use in the asserted C4 pathway. For this reason it is suggested that the conclusions drawn by Bort et al, that there was no evidence for C4 photosynthesis in the ears of C3 cereals from their labeled-carbon pulse experiment, is a false basis to deny their conclusion as it failed to account for this inside-out carbon delivery.

To the contrary

The first point raised in contradiction to the Rangan et al research repeats the earlier criticism – that although the gene expression profiles reported in the disputed paper are novel, interesting and worthy of further research, they alone are not enough to justify the conclusions made. To be able to make such a conclusion, evidence demonstrating flux of the metabolites through the C4 pathway. Increased expression of the relevant genes is not enough to conclude that the pathway is operating in the seed particularly as all the genes involved in C4 photosynthesis are also expressed in C3 plants.

Although the evidence for the existence of the pathway in wheat seeds is scarce and contradictory, the defenders of the rebuttal suggest that the Bort et al article is evidence that the PEP carboxylase activity assists in intermediate reactions in metabolic pathways other than C4 photosynthesis.

Further, the evidence in the Rangan et al paper suggested that the expression of genes encoding the Rubsico enzyme was minimal. The contention is that, if the increased transcription of C4 genes is relied upon as evidence of increased activity within the pathway, reduced transcription of this vital photosynthetic enzyme must lead to the conclusion that the increase in PEP carboylase activity must be in aid of some pathway other than a photosynthesis pathway. Coupled with this, the low concentration of CO2 in the pericarp, where the Rubisco is located, is contrary to the high concentration of CO2 around Rubsico common in C4 tissues. Accordingly, expending energy to run a C4 pathway when the carbon would flow to Rubsico within the pericarp without the need for the pathway is contrary to expected energy conservation measures.

…and back again

In response to the criticism that evidence of flux through the pathway is lacking, the defenders of the Rangan et al paper suggest that such evidence has already been reported in a 1976 paper in the C3 intermediate barley and the relevant proteins isolated in research reported in 1986. It is argued that their research has showed that C4 specific versions of the genes were expressed in the seeds compared to the C3 versions expressed in the leaves.

In relation to the Rubisco levels, the Rubsico levels reported in the paper were for the whole seed, not just the pericarp and earlier work has shown that Rubsico is specifically expressed in the pericarp with very limited levels in the endosperm. As such, it is argued that the although the levels were rather limited in the seed as a whole, they are in fact present in high concentrations in the pericarp, the site where the C4 pathway is said to concentration CO2 levels.

Returning to the problem of the Bort et al paper, the researchers again suggest that the experiment conducted in that research could not have tested for the possibility of a C4 pathway supplying CO2 to the pericarp from within the seed, after which it would be utilised by Rubsico concentrated in the pericarp. Accordingly, they argue that the conclusions of that study fail to invalidate their conclusions.

As for the possibility that CO2 could more efficiently diffuse to the pericarp, it is contended that if that was the case it would not be concentrated enough at the site of the Rubisco and would therefore all too easily escape the seed. Their contention is that the presence of a C4 pathway in the seed is necessary for adequate CO2 concentration to the pericarp-located Rubisco.

…and one last response

The final of the letters concedes that the labeling of CO2 inside the endosperm may yield different results compared to the labeling study in the Bort et al paper, but cites the 1976 research as using isolated pericarps in their experiment and therefore doesn’t provide appropriate evidence support the C4 pathway inside the endosperm.

In relation to the diffusion versus C4 pathway dispute, it is suggested that any CO2 which is not passed from the endosperm to the pericarp through intermediates in the pathway would diffuse outwards following Fick’s law whether or not it moves through the tissue as malate. If it moved as malate, again the cost of such a pathway would be more than if it moved by simple diffusion.


The last paragraph of the final letter sums up the research as it stands and the debate about the possible conclusions that can be drawn. The research in the Rangan paper provides “an exiting new piece of the puzzle” in the quest to understand how wheat seeds increase carbon gain. It would seem that to put the debate to rest, the flux of the carbon through a C4 pathway must be demonstrated unequivocally along with measurement the location and activity of the required enzymes.

A passionate and reasoned debate over plant science which is wonderful to see.

Speed Breeding

Edit: We’ve added a link to the published version of the paper reported on here. The final version was published online in Nature Plants on 1 January 2018.

The generation time of an organism varies from organism to organism. One of the defining features of model species is their short generation times, allowing fast analysis of, for example, gene mutations on phenotype (think of mice, zebrafish and the fruit fly). The shorter the generation time, the greater the number of experiments and analysis which can be performed within a set period of time.

In the world of plant science, Arabidopsis thaliana is generally considered the model organism and many papers on genetics and gene technologies use A. thaliana as the subject plant from which study results are drawn.

But conclusions drawn from a model organism doesn’t always translate to similar results in another organism – think of promising medical studies in mice that haven’t had the same effect in humans. Accordingly, model organisms may not always be appropriate for use to draw conclusions about other organisms.

Such is the case with the translation of studies in A. thaliana in some important crops such as wheat and barley. The problem with studying genetic traits in such crops is that the generation time of the plant may be long, significantly impeding important research.

Introducing Speed Breeding

Given the problem with generation times, a small army of researchers led by Lee Hickov of the Queensland Alliance for Agriculture and Food Innovation (and recent recipient of the Queensland Young Tall Poppy Scientist of the Year – congratulations Lee) developed a new protocol for speed breeding and tested a number of important crop plants for their response to the new protocol (here is the published article in Nature Plants).

The Protocol

The protocol is pretty simple. Instead of using a standard photoperiod used when growing plants under study, the researchers increased the day length to 22 hours and reduced the length of time the plant is in darkness to 2 hours.

In fact, the paper specifies three protocols that all use this photoperiod, each of the protocols having a slight variation in the other growth conditions including the third protocol which is meant to be a less expensive set up able to generate similar results to that found in laboratory conditions.

Testing the Protocol

The first testing of the protocol took place in Norwich, United Kingdom, and compared the growth rates of bread wheat, durum wheat, barley and a model grass Brachypodium distachyon when grown using the speed breeding protocol or grown under usual greenhouse conditions with no supplemental lighting or heating.

The crops grown in the modified setting consistently flowered earlier than the control plants, taking approximately half the time to begin flowering with B. distachyon taking an average of 26 days and wheat taking on average 37 to 39 days. A significant change in the amount of time required to reproduce a target number of generations.

To ensure that the faster flowering plants didn’t suffer from loss of seed viability, the researchers tested whether the speed-bred plants were able to generate the following generation of plants. The researchers quantified effects on three phentoypic variables related to viability, including:

  • wheat seed counts per spike – a decrease compared to control;
  • spikes per plant – healthy number in the speed-bred plants; and
  • viable seed production – no difference.

Moving back from Norwich to Queensland, the researchers tested a second but very similar protocol with spring wheat, barley, canola and chickpea cultivars. This protocol used high pressure sodium lamps to produce the 22-hour day for these plants and compared the effect of this protocol to the same plants grown in glass house with a 12 hour day/night cycle. This time, both treatment and control groups were subject to the same temperatures for the day and night portions.

Similar to the first test, the time to anthesis was significantly less under the speed breeding conditions, but this time the wheat plants produced more spikes, grain number didn’t show the same reduced effect as it did the first protocol and the flowering times of the different crop species were more uniform than in the control. The researchers point out that this last effect, crops flowering at the same time, is advantageous when genotype crossing is to be performed. Seed viability was retained even when harvested 14 days post-anthesis and subsequently subjected to cold treatment for 4 days, indicating that further time savings can be made in the production of generation times.

In canola and chickpea there was no significant difference in seed production between the treatment and control groups.

Speed Breeding

Figure 1 from article demonstrating the photoperiod and temperature ranges under the speed breeding (left) and control (right) conditions and the number of generations which can be produced in wheat, barley, chickpea and canola under each condition.

Overall, the number of generations that can be produced under speed breeding conditions in a year were:

  • wheat – 5.7 generations;
  • barley – 5.4 generations;
  • canola – 3.8 generations;
  • chickpea – 4.5 generations.

The possibility of using the speed breeding technique in conjunction with single seed descent breeding and research programs was the basis of a seed viability analysis in wheat and barley. Using 100-cell trays, seed viability was confirmed for both genotypes at 80% two weeks after anthesis and 100% at four week post-anthesis. Accordingly, single seed descent programs can also be used in conjunction with speed breeding.

A low-cost speed-breeding protocol was also tested. In this set-up, LEDs were used exclusively in conjunction with a conventional split-system air conditioner to create speed breeding conditions and garnered similar results.

What about effects other than generation times?

Although faster generation times are extremely useful, if the effect of the speed breeding is to alter the crops’ phenotype, research programs would be plagued by unwanted confounding factors and breeding programs would be rendered ineffective.

The research team therefore mutated the awn supressor B1 locus and the Reduced height (Rhd) genes in wheat and obtained the expected phenotypes under the faster generated crops.

Testing any variation in disease responses was performed by inoculating wheat spikes with the bacteria which causes fusarium head blight, one set of spikes belonging to a susceptible cultivar and the second belonging to a head blight-resistant cultivar. The speed-bred cultivars showed the expected resistant and susceptible phenotypes.

Finally, the use of plant breeding in genetic engineering programs was addressed by looking at transformation efficiency of barley seeds grown under the protocol and, again, found no difference when compared to conventionally grown transformed plants.


Hickey’s team has successfully tested a protocol that will greatly assist research efforts into these studied plants and whose results will likely be replicable in many other plants. They cite sunflower, pepper and radish as plants that have shown similar responses to being grown in extended daylight conditions.

The ability quickly test, for example, the effect of knocking out or over-expressing a particular gene will certainly lead to more efficient test results and, hopefully will be another stepping stone to overcoming some of the limits to efficient research we currently endure.


C4 Photosynthesis – Testing Candidate Maize Genes in Rice

It would be no surprise to regular readers of the Legume Laboratory that we hold significant interest in the possibility of engineering C4 photosynthesis into C3 crops. Projects such as the C4 Rice Project are dedicating their efforts to this goal and its promise of more efficient crop growth and improved food security in the face of a changing climate.

Scientific Reports recently published an article funded through the C4 Rice Project which was looking to take the next step in procuring this biological engineering feat.

Maize is an interesting plant as it contains the two forms of photosynthesis, utilising C4 photosynthesis in foliar leaves while husk leaves use C3 photosynthesis. Previous studies have compared the transcriptomes of the two types of maize leaves to identify differences in gene expression. Given both leaves contain the same genome, hidden within the 283 differences in gene expression should be those differences which convert a leaf from using C3 photosynthesis to using C4 photosynthesis.

Working from this base, the researchers selected 60 of these genes to analyse further. The choice to look at 60 genes instead of the 283 differently regulated genes identified was born of the lack of high throughput technology needed to test gene functions efficiently enough for all to be analysed. The 60 genes tested were selected as they were predominantly transcription factors or leucine rich repeat receptor like kinases. A maize ubiquitin promoter was ligated upstream of the each individual gene resulting in 60 gene constructs that were individually transformed into rice plants.

In order to identify whether the genes made any change to transgenic rice plants, the researchers compared the leaf anatomy of the transformed leaves to the wild-type leaves. Regression analysis of vein number and leaf width in rice plants found a linear relationship between the two (a wider leaf will have a greater number of veins proportional to the change in width when compared to a narrower leaf). Given the difference in cells surrounding veins in C4 photosynthetic plants compared to C3 plants, there is a difference in the space between veins in the two plant types. The hope was to identify changes in the ratio of vein number to leaf width in transformed plants, an indication of a possible change in the method of photosynthesis.


The constitutive expression of 47 of the 60 genes made no observable difference to the ratio of veins to width of leaf normally seen in rice plants and the plants exhibited similar phenotypes to the wild type plants.

Three genes resulted in the transformed plantlets being unable to regenerate. The authors explain that rice tissue culture requires an excess of cytokinin to auxin in order for shoot growth to be promoted. Research into the three genes involved (ZmIDD16, ZmbHLH106 and ZmHCA2) revealed homologous genes in Arabidopsis that show evidence of effects on auxin biosynthesis. As a result, it was hypothesised that each of the three genes have resulted in an increase in auxin biosynthesis, throwing out the ratio of cytokinin to auxin needed for regeneration to occur, although further research was noted as necessary to confirm this hypothesis.

A different set of three genes disrupted shoot and root development, again believed to be due to alterations in the cytokinin/auxin ratio. The YUCCA gene seemingly caused the biosynthesis of excessive amounts of auxin, resulting in characteristic curled leaflets and hairy roots. A gene encoding a SMALL AUXIN UPREGULATED RNA (SAUR) disrupted the formation of lateral roots, with homologous genes demonstrating a disruption in normal auxin transport, reducing auxin-related root elongation and retarding initiation of lateral roots. The third gene (ZmSACL3) affected cytokinin, inhibiting its synthesis and resulting in similar phenotypes to that caused by the YUCCA gene due to the altered auxin/cytokinin ratio.

The overexpression of transcription factor R2R3 MYB caused intermediate veins to lose sclerenchyma cell connections to the abaxial epidermis. The loss of strengthening cells resulted in curled leaves. A different and novel gene resulted in the opposite effect on sclerenchyma formation, causing an increase in the proliferation of this cell type and resultant thicker walls than in wild type plants. Further, a couple of larger than usual bundle sheath cells around each vascular bundle formed. The result of the overexpression of this unnamed gene were infertile plants with short shoots and roots.

Figre 4 from C4 OX study

Figure 4 from article showing the effects of over expression of an as-yet unidentified gene. Figure b shows the reduced root growth while figure c shows the reduced shoot formation. Panel g shows normal sclerenchyma in wild type rice plants while panels h and l show the increased cell wall thickness described above.

ZmWRKY12, a gene which inhibited callus formation and therefore had to be inserted behind an inducible promoter to allow phenotypic analysis, caused dwarfing and reduced root growth in transgenic plants. The vein number to leaf width ratio was unaffected by the transgene but the normally lobed mesophyll cells in rice plants were lost in the transgenic lines. The orthologous gene in Arabidopsis is known to be involved in the formation of secondary cell walls, but these researchers have discovered that it also seems to play a role in cell lobing, a phenomenon for which the genetic basis was previously unknown.

Figure 5 from C4 OX study

Figure 5 from article showing the effects of ZmWRKY12 expressed in rice plants. The most interesting phenotype change was the change in mesophyll cell lobing from wild-type (panels d, f, and h) to the lack of lobing in the transgenic lines (panels e, g, and i).

Finally, three closely related Zinc finger nuclease genes showed normal vein spacing but caused spindly plants with drooping stems and leaves. It was theorised that the cause of the phenotype was due to alterations in the production gibberllic acid in the plants.


This paper is a great example of the incremental advancement of knowledge that builds upon earlier increments and is built upon by later increments. Those increments fill gaps in knowledge, or rule out particular cause/effect relationships, which help to direct and propel research towards its particular end (although, of course, research and enquiry need not always have a particular end to be directed at).

In the case of this research, the experiment resulted in some interesting and novel results, even though it didn’t yield any specific advance towards understanding the difference in gene regulation between the two forms of photosynthesis. That being said, we know that constitutive expression of any of 60 genes isn’t enough to trigger C4 photosynthesis. That leaves the remaining 223 genes to test, or perhaps a temporally regulated expression of one or more of the candidate genes, or some other combination of gene expression not yet contemplated.

The possible causes of C4 photosynthesis are numerous. This study also highlighted the deficiency in technology related to the efficient transformation, validation and phenotyping of transgenic plants. At present we lack the ability to quickly test over-expression of individual genes, and this is a relatively simple analysis to perform. The possibility that differential expression of a number of the genes being required to establish C4 photosynthesis requires significant advances in both wet lab and computation biology if we are to quickly make significant leaps in this type of research.

Great work by a great group of committed scientists.

Synbio in Space!

The challenges of surviving in space (with or without hostile enemies intent on dominating the universe) has made and will continue for a long time to make for a good sci-fi story. The resourcefulness required to survive in the vast emptiness between (and on) planets also forces us think deeply about better ways of propelling, feeding, housing and medicating ourselves. The cost of launching extra weight into space plus the limited lifespan of necessities such as food and medicine leaves us with complex questions about how to sustain ourselves as we look to explore more of our solar system and, hopefully one day, beyond it.

A 2015 review in the Journal of the Royal Society Interface considered a number of space missions and analysed how synthetic biology may assist those missions in the four resource areas mentioned above; propellant, food, habitat and medication. Of most interest, given the significance of the challenges presented, is a six-person, 916 day return voyage to Mars which would include 420 days of traveling there and back plus a 496 day stay on the red planet.

What are we going to have to do to be real-life Martians?


Although one of the shorter sections of the paper, how we will advance food production being the focus of this blog means that we will put the section of food production in space up-front and centre.

Based on astronauts living on the International Space Station, living in space requires 1.83kg of food per crew member per day. As a result, the six person Mars mission requires around 10 tonnes of food that needs to be shipped in order to sustain a mission if all of it was shipped from Earth although more accurate assessments of food amounts required to be produced whilst on mission have a lower estimate of 4.5 to 5 tonnes.

So the ability to produce food using available elements and compounds whilst away from the Earth is a significant problem. And the most accessible food production option is to use photosynthesizing bacteria and plants. Producing food using bacteria has the most significant weight savings, with Spirulina (a nutritional biomass made from either Arthrospira platensis or Arthrospira maximum) being promoted in earlier papers as a suitable option to producing food in space compared to attempting to grow and harvest plants. According to those previous papers, it is possible to produce 5.271kg of Spirulina per day using three 2,000 litre bioreactors. This rate of production would sustain the crew for the entirety of their stay on Mars and for the return to Earth and save a significant amount of launch mass.


Spirulina powder.

So how can synthetic biology help if Spirulina is already producible in adequate amounts to sustain the majority of a Mars trip? If you can imagine eating the same substance every meal for 706 days straight, one problem is easy to identify. It is suggested in the paper that synbio efforts could be directed to diversifying the textures and the tastes produced by the bacteria so that a variety of ‘foods’ can be consumed by those on the Mars mission.

Increasing the rate that the bacteria is able to reproduce will also aid the weight reductions at lift off and increase the redundancy built into the rations of food for the trip. There are some reported increases in production rates reported in the paper, but, for example, whether increased photosynthetic efficiency can be engineered into the bacteria to increase production is a question common to current research efforts in earth-bound food production.

Finally, being able to increase the nutrient content in the food produced to aid general health and space-specific health problems will increase the chances of a successful mission.


The shelf-life of pharmaceuticals are shortened in space travel and at present the issues caused by this are solved by  sending more supply missions. Given the time and distance involved in a trip to Mars, such a solution has little practical use.

Synthetic biology has the ability to play a role in readily producing pharmaceutical substances during a mission to Mars. For example, acetaminophen (paracetamol) can be produced using a modified chromisate pathway in E. coli. Unfortunately, space travel doesn’t suit E.coli as the compounds it requires to survive are not readily available. But if an autotroph can be manipulated to contain the same pathway and produce the same substance, we may have a new way to produce the pharmaceutical with the restricted inputs. The cyanobacterium Synechocystis sp. PCC 6803 has been identified as a possible candidate that will survive on the inputs available in space and could host the appropriate pathway to produce acetaminophen.


Figure 8 from article demonstrating the proposed pathway to create acetaminophen from Synechocystis sp. PCC 6803.

It can be imagined that with enough research we should be able to produce a range of pharmaceuticals or at least their components from the limited inputs available in space.


The further we go and the more weight we carry, the greater the fuel required. Complicating the matter further, the more fuel we carry into space, the more fuel we require to lift the extra fuel off the ground. The problem of needing more fuel to lift even more fuel to be able to propel more weight into space is one that permeates through the problems of supplying enough food and equipment to astronauts traveling for long periods of time.

The possibility of producing fuel in space is limited, according to this paper, to using methanogens to create methane. The type of methanogen used in bioreactors effects the quality of methane gas created and the rate of production. Further, a number of methanogens are able to use the compounds, particularly CO2, created by the crew of a space mission.

Synthetic biology uses to address the problem of fuel production in space will likely focus on the production of methane by methanogens. Particularly, the ability to produce methane using the least amount of inputs and with the greatest nutrient recycling efficiencies will be key. Similarly, modification of cyanobacterial production of ethylene, an alternative propellant, for greater production efficiency is another possible use of synthetic biology to aid space flight.

Habitat Construction

Lastly, the ability of erect a habitat on Mars will require either the supply of material from Earth and/or the use of material available on Mars with equipment supplied from Earth. The proposed use of 3D printing to construct habitat allows the use of finer materials in an additive manufacturing process. But the possibility of biologically produced materials (biopolymers) provides a method of producing the base materials using inputs available away from the Earth and significantly reducing the launch mass of materials dedicated to habitat construction.


Figure 7 from article demonstrating the shipping mass saved for habitat construction in missions to Mars (left) and the Moon (right).

Synthetic biology efforts focused on increasing the synthesis of biopolymer material per unit of bacteria will result in less equipment in the form of bioreactors being required to travel with the crew with the time for construction of the habitat being reduced as well.

Synbio in Space

At its most basic level, the quest to man a mission to Mars creates problems that must solved using the most efficient means possible. As a result, the use of synthetic biology to solve problems in the four areas listed above looks little further than what inputs are available in space and how they can be used to solve the problems spaceflight presents.

However, advances in modifying organisms to meet these challenges will likely lead to flow-on effects in the same areas here on a better-resourced planet. For example, should it be possible to create different textures and tastes in Spirulina in the confines of space then one would expect that it will be possible to use different inputs and different genetic circuits to create a variety of foods. Further, food produced should be more efficient, sustainable and economical than traditionally generated food.

Genetic Underpinnings of Heat Resistant Potatoes

Identifying the genetic basis of a particular trait is the building block on which molecular and transgenic breeding strategies build upon. How we connect the dots between a genotype and trait is the basis of a past post “How do we find genes related to traits? A review of bulked sample analysis”.

Despite only being behind rice and wheat as the most important crop worldwide, quantitative trait loci (QTLs) for heat tolerance in potatoes have not, until recently, been identified.

Potatoes are particularly sensitive to heat, with temperatures above the ideal maximum of 22°C resulting in significant decreases in growth and yield. The higher temperatures have a number of effects:

  1. Strongly suppresses tuberisation;
  2. Reduces the proportion of carbon assimilated for tuber starch;
  3. Reduces chlorophyll levels and CO2 fixation rates;
  4. Halts tuber dormancy resulting in early sprouting and secondary growths.

A recent study in the Plant Biotechnology Journal sought to identify QTLs for heat tolerance by comparing differences in gene expression and tuber yield between progeny of a biparental diploid potato (06H1) when grown at 22°C and 28°C.


Temperature response analysis and gene identification

Single node cuttings were used as a method to quickly induce tuberisation of subject potato plants, allowing a faster assessment of tuber yield (see this article discussing its use – paywalled unfortunately). 170 genotypes from the 06H1 population were analysed using 6 nodal cuttings from each to assess tuber yield under 22°C and another 6 for each genotype assessed at 28°C. On average, tuber yield at the lower temperature (1.017 grams) was greater than for nodes grown at the higher temperature (0.285 grams). The difference in performance was used in combination with single nucleotide polymorphism (SNP) maps of the 06H1 genome to identify genetic markers which may be connected to the genes which alter expression rates at the two temperatures.

fresh weight tuber yield nodal cutting assay

Figure 1 from article. Comparison of tuber fresh weight yield when grown at 22°C and 28°C.

Statistical analysis of the differences in expression levels at each marker found 3 markers which may be connected to the difference in phenotype. Using the Potato Genome Browser the researchers browsed the list of genes connected to the markers for suitable gene candidates for further research. Of the three markers, one marker contained a possible gene candidate, that encoding Heat Shock Cognate 70, a gene that has been suggested as involved in the heat tolerance in cabbage and potato.

With a candidate gene identified, the gene was isolated at amplified for analysis of allelic differences within the gene, of which 4 were identified and named A1, A2, A3 and A4. Nodes of potatoes containing each the four alleles were assessed for tuber yield, with allele A2 having the greatest fresh weight at both 22°C and 28°C and combined A2A3 genotypes having the highest yield. Gene transcript levels were assessed in plants grown at both temperatures which showed that the HSc70 gene was upregulated in A2A3 plants but not in the A1A4 plants, while the HSc70 allele up-regulated in plants grown at the warmer temperature were all from the A2 allele.

Analysis of the amino acids of each of the alleles themselves found that the A2 and A3 proteins contained an additional C-terminal sequence of KIEEVD (thought to be involved in interactions with co-chaperones), while no other variations were unique in a single allele.

What was unique to the A2 allele was a difference in the promoter sequence where a run of ten TA repeats was found as opposed to a run of four TA repeats in the promoters of the other repeats. The researchers later hypothesize that the greater number of TA repeats may result in enhanced expression levels, a hypothesis based on previous research in yeast, although the specific effect of the repeat TA sequence is not known.

Confirming the effects of the HSc70 A2 allele

To test the A2 allele and its promoter sequence, the gene and upstream component was transferred into Nicotiana benthamiana plants and subjected to 45°C heat in a growth chamber. After 24 hours the expression of the HSc70 gene was 7 times greater than in the control plants and cell membrane injury in the transgenic plants were significantly lower.

The unique promoter was then targeted for further analysis. The A2 allele was put downstream of a series of promoters with TA repeats varying between four and ten times. After being exposed to the higher temperatures the level of HSc70 transcription was measured. Alleles with eight to ten TA repeats in the promoter showed statistically significant elevated transcript levels compared to alleles with four to six TA repeats in the promoter.

HSc70 expression levels

Figure 6 from article. a) HSc70 expression levels for promoters containing 4, 6, 8 and 10 TA repeats and b) cell membrane injury percentage for plants containing the same numbers of TA repeats.

A2 allele-containing plants with significantly increased expression under 28°C heat were selected for analysis under 40°C conditions for 24 hours in comparison to wild-type plants. Extracts of plants after 4 hours at this temperature showed HSc70 expression being enhanced by up to 50 times more than in control plants. However, after 24 hours the transcript levels fell back to that seen prior to being subjected to the extreme heat. The A2 allele-containing lines has significantly less cell membrane damage compared to the wild-type.

Finally, the ability of the A2 allele to protect tuber yield at elevated temperatures was tested. Using the same nodal cutting tuber yield assay as before, A2 allele containing plants were compared with wild Desiree potato plants when grown at 20°C showed little difference in tuber yield. However, at 28°C the Desiree plants reduced yield by 75% compared to yield at 22°C. By comparison, transgenic lines had significantly lower yield reductions, with some lines have twice the fresh and dry weight yield as the wild type.


The identification of the heat shock cognate protein 70 within the potato genome and the quantification of the effects of the specific A2 allele demonstrates the significance of quality genome research, genome annotation and the ability of powerful molecular tools to seek out expression variances that allow us to pinpoint the basis of variances in phenotype. Coupled with well prepared and executed experiments, the development of these tools will continue to enable us to extract information that will assist us to assist crops and food production.

Whether it is through transgenic technology or molecular breeding, the identification of genes such as the one discovered in this study will assist food production meet the challenges of a changing climate.

Identifying the specific allele of the HSc70 protein that contributes to its resistance to the negative effects of elevated temperature is surely a big step into future-proofing the security of this important crop.

C4 Photosynthesis: Discovery, Understanding and Future Development

“The C4 pathway is one of the most complicated biochemical pathways likely to be placed in front of a university undergraduate, and attempts at C4 engineering have only served to reveal additional uncertainties about how the pathway operates, how it is regulated at the level of gene expression and post-translationally, and some of the physical properties of the anatomical specialization of Kranz C4 leaves.”

Robert Furbank (2016) “Walking the C4 pathway: past, present, and future”, Journal of Experimental Botany, vol. 67, no. 14,  pp. 4057 – 4066

Whilst writing our post titled “C4 Photosynthesis Evolution: Why some intermediates may not assist our understanding” we came across a 2016 paper written by Professor Furbank covering the discovery of C4 photosynthesis, the research and stories behind our understanding of this important photosynthetic trait, the gaps in our knowledge that remain to be filled and where the research may lead moving forward.

Readers of this blog may recall that Professor Furbank generously gave us the time to explain the problems in a research article proposing that C4 photosynthesis already existed in wheat which greatly assisted our updated post on the issue. He has also recently been a guest on the Australian Broadcasting Corporation’s “The Science Show” talking about C4 photosynthesis and the positive effects it could have on our food security (listen here).

Discovery of an important trait

Prior to the biochemical description of C4 photosynthesis detailed by Hal Hatch and Roger Slack in 1966, the special properties of particular plants that were later understood to use C4 photosynthesis had been documented since the turn of the 20th century. Professor Furbank’s review lists the important research articles which led the way to our current understanding, such as the discovery of Kranz anatomy, plants increased water use efficiency and increased photosynthetic rates compared other plants.

The separation between the collection and concentration of CO2 in mesophyll cells before pumping the CO2 to bundle sheath cells, where it is reduced and decarboxylated, is the most important and distinguishing trait used to differentiate C4 photosynthesis from other forms of photosynthesis. The concentration of CO2 allows Rubsico to operate at greater efficiency, reducing the competition for the active site of the enzyme that normally plays out between CO2 and O2 in non-C4 species, benefiting the growth rate and resource efficiency of the plant.

Furbank - C4 diagram

Figure from article demonstrating the Kranz anatomy (ring of purple bundle sheath (BS) cells and mesophyll cells (M) within the plant tissue. The numbers on the right of the figure denote; 1. carbonic anhydrase, 2. PEP carboxylase, 3. malate dehydrogenase, 4. NADP-malic enzyme, 5. pyruvate orthophosphate dikinase.

Unpublished data, criticisms of legitimate research findings and unnoticed publications of important research in obscure or foreign-language journals hampered early progress in discovering C4 photosynthesis until the seminal work of Hatch and Slack (who we are proud to learn are Australian scientists) in 1966.

Hatch and Slack’s 1966 research paper used 14C-labelled CO2 to study the phenomenon, applying a pulse of labeled CO2 and then illuminating the subject leaf to follow the progression of the labeled CO2 as it made its way through the photosynthetic cycle. This method is commonly known as a ‘pulse-chase’ analysis. The idea for the experiment came from similar earlier experiments conducted previously by other researchers but which had provided results that lacked certainty in the identification of the labeled compound. Whilst having chatting over a beer during the Australian Biochemical Society conference held in Hobart in 1965, Hatch and Slack decided to try their hand at replicating these earlier experiments in a manner that would reduce or remove the uncertainty contained within the earlier results and shed further light on the strange data that was being produced in these earlier works.

In something reminiscent of the oversight of Gregor Mendel’s early research on genetics in pea plants, it was discovered in the late 1960’s that a Russian researcher named Yuri Karpilov had reported similar findings from his research as that obtained by Hatch and Slack in an article he published in 1960 in a little known journal.

What wasn’t grasped at the time was that this newly discovered pathway resulted in a pool of CO2 concentrated in some cells while Rubsico, an enzyme difficult to extract, was present and operating in the separate bundle sheath cells. This discovery was slowly uncovered in a series of papers in the late 1960s and early 1970s, as was the importance of the concentration of CO2 in reducing photorespiration and its link to improved plant performance.

As our methods of isolating and probing at different cells have improved, so has our knowledge of the variety of biochemical pathways that exist within the C4 pathway. However, we are still grasping at which of these pathways provides the greatest flux of metabolites. When we acquire this understanding it will focus our attention to those pathways likely to result in the greatest improvements in plant performance and harness our attempts to engineer C4 photosynthesis in important crops.

Recent research

Since the significant increase in our knowledge of this important photosynthetic pathway around the time of Hatch and Slack’s 1966 paper, genetic transformation of plants and the increase in our ability to extract and subject plant data to bioinformatics techniques has continued to push research forward. The use of green foxtail millet as a C4 model grass by the biofuels community has allowed for the development of transformation protocols to help probe deeper into the genetics, transriptomics and protein products that form the basis of this system. Further, this information is crucial to our engineering attempts aimed at improving crop productivity.

Similarly, next-generation sequencing (NGS) has revolutionized research probing the evolution of C4 photosynthesis and how they may have evolved from their C3 ancestors. Probing the differences between the genetics and gene transcripts between the differentiated cells is steadily helping us understand and pinpoint which genes are absolutely necessary for a functioning C4 pathway. Analysis of C4 meristem tissue as it grows and the adjustment of gene transcription through this growth is also assisting in deciphering what gene regulation is required if we are to successfully install C4 photosynthesis into important C3 crops.

Where to now?

There is still important research to be done and discoveries to be made in C4 photosynthesis research.

Quantifying the amount of fixed CO2 that is leaked from mesophyll cells, under what conditions greater or lesser amounts of CO2 is leaked and what physical properties of the bundle sheath cells may change the rate of flux are all questions that remain to be answered if we are to successfully engineer the system into C3 crops. As with many areas of research, new tools to extract the information and avoid the uncertainty that plagues the more indirect methods of investigation we are currently limited to will greatly assist the research.

Further information about the positioning of important organelles within the cells and the ratios of bundle sheath cells to mesophyll cells are yet to be understood and which could prove to hamper engineering efforts.

Finally, despite the significant leap in the genetic and transciptomic knowledge we have seen of recent times, there is still only a developing understanding of gene regulation in the growing plant tissue to ensure the separation of the two parts of the carbon fixation and decarboxylation process within the tissue. Important transcription elements specific to this demarcation of activity between the cells are still being investigated and isolated.

The future

Significant research effort is being funded to enable the discovery of these important missing pieces of information and into attempting to transfer C4 photosynthesis into important C3 crops. The C4 Rice Project, funded by the Bill and Melinda Gates Foundation, is one such effort that looks very promising.

So, can we supercharge a VW to perform like a Porsche? The progress being made and the dedication to the problem must lead us to a sense of optimism that it will happen within our lifetimes.