Control of seed size and yield by vernalisation

A major problem with UK winter oilseed rape (WOSR) cultivation is the large inter-annual fluctuation in on-farm yields, varying from around 3.0 to 3.9 tonnes per hectare. Weather and climate are known to be critical factors in this yield instability, and climate change may be a factor in the current 'yield plateau' in UK agriculture. For this reason breeding and selection of new varieties suited to changing environments is widely appreciated to be a key long-term goal for UK agriculture. Targeting this effort requires an understanding of the precise biological mechanisms through which weather variation influences yield on farms. It is here that this proposal breaks new ground.
In the UK and Northern Europe high yielding so-called 'winter' crops are distinguished from spring crops by their requirement for a period of prolonged winter cold before flowering, a process known as vernalisation. While the effect of vernalisation on flowering has been known for many years, here we show for the first time that vernalisation has a further major role in the control of seed size. We show that seeds from more strongly vernalised plants are around 30% larger than seeds from more weakly vernalised plants, and this size increase comes without any penalty in fruit or seed number. These effects are large enough to account for a substantial fraction of the 30% inter-annual variation in yield observed on farms in the UK. In addition, we provide new evidence that winter cold during a very precise 23-day period is highly correlated to UK rapeseed yields, suggesting that on a landscape scale vernalisation may be important for WOSR crop yield performance.
The molecular mechanism underlying the vernalisation process is understood in detail. A key transcription factor gene, FLOWERING LOCUS C (FLC), is highly expressed before winter and is gradually silenced during the experience of cold weather by the accumulation of repressive chromatin marks. FLC expression in leaves and the shoot apical meristem controls flowering, but FLC is also strongly expressed in seeds where it has previously been implicated in seed vigour. Seed size on the other hand is controlled primarily by the expansion of the seed coat that occurs after fertilisation, creating the space into which the embryo and endosperm grow. Our preliminary study shows that in rapeseed both FLC and a second gene AUXIN RESPONSE FACTOR 2 (ARF2), a known seed size regulator in Arabidopsis, can both be genetically linked to the control of seed size by vernalisation. Furthermore, online genomic scale datasets show that ARF2 is likely a direct target of FLC.
In this proposal we will elucidate the mechanism by which vernalisation controls seed size and examine the importance of this process in on-farm yield variation in WOSR. Our preliminary analysis has revealed a very strong candidate for this mechanism and our experimental programme is designed to investigate this in detail. Using the JIC experimental farm we will test whether winter cold affects WOSR yield by affecting seed size, and how the key molecular events of vernalisation and seed development unfold in a real field scenario. By taking lines carrying genetic variation at FLC and ARF2 loci through to maturity we will directly test whether temperature in early December affects rapeseed yield parameters through a mechanism involving these two genes.
One goal is to identify new combinations of of FLC and ARF2 that maintain large seed size despite only weak vernalisation. Markers for these alleles can then be used by breeders to select new varieties with a more robust yield performance, ameliorating an important negative effect of warmer winters on UK agricultural performance.

A key goal of breeders is to develop varieties with high yields both in National List trials and on-farms which perform dependably despite variation in seasonal weather patterns. The data presented here show that high UK WOSR yield is driven by increased exposure to winter cold, and demonstrate for the first time that in addition to the well-known regulation of flowering, vernalisation also has a major impact on seed size. We observed that this effect is large, with vernalisation producing more than a 50% gain in TGW in some lines. This conclusion that vernalisation also influences seed size is further supported by a QTL study which implicates FLC and ARF2 in the mechanism, two transcription factors known to be central to vernalisation and seed size control respectively.
The aim of this proposal is to determine the molecular mechanism by which vernalisation affects seed size and consequently WOSR yield, and to test the significance of this biological process for UK on-farm WOSR yields. We will achieve this by artificial manipulation of temperature in field trial plots to examine effects on yield, and by analysing the molecular events leading to vernalisation and seed size variation in the field. Allelic variation at FLC and ARF2 in rapeseed and Arabidopsis will be used to test genetically for a role of these genes in seed size control by vernalisation. Finally ChIP and transient gene expression approaches will be used in rapeseed to determine the effects of BnFLC manipulation on ARF2 gene expression. The results of this work will have implications for winter crop yields during an era when climate change will lead to both warmer winters, and to more inter-annual variation in winter temperatures. Understanding the biological mechanisms underlying year-on-year variation in yield could lead to early prediction of yield facilitate breeding of new varieties that are better suited to the challenge of climate variability in the 21st century.

The UK rapeseed harvest is worth an average (last 5 years) of £815 million per year. With yield varying by up to 30% each year, this corresponds to a annual value of up to £260 million which could be subject mainly to varation in weather and climate.
The impacts from this work can be divided primarily into two distinct areas:
The first is impact on breeding: this project will yield haplotypes of WOSR which via allelic variation at FLC and ARF2 loci are likely to have enhanced Thousand Grain Weight under a wider range of vernalisation intensities. We will design markers for these haplotypes, and the high copy number of FLC and ARF2 in WOSR suggests that there is likely value in stacking haplotypes to increase effect size. Our aim will be to gather positive alleles together in a single background to expedite use in breeding programmes, i.e. to produce optimised pre-breeding germplasm (see pathways to impact). Using the ecotilling data we will be able to assign each variety a haplotype set for FLC and ARF2 loci. We will make this data accessible via the JIC Brassica webpages so breeders can see which useful haplotypes may already be present in their populations and which they are missing.
The second is yield prediction: with a simple linear model in hand we were able to make some yield predictions for the 2015/16 season by Christmas 2015. December 2015 temperatures averaged 8C, 4C higher than normal, and the model predicts that 2016 harvest yields should be 0.5-0.6 t ha-1 lower than average (which is 3.6tha-1; i.e. on farm yields should be3.0-3.1 tha-1). The final value according to AHDB was 3.0-3.2 tha-1 which is in line with our simple estimates. The work in Objective 5 will improve our predictive abilities still further. This ability to predict yields could be useful for farmers for planning, or to know when to sell the rapeseed crop to the futures market. It could also help planning in supply chains. To aid with yield prediction, we will produce a website widget which will predict yields and advertise this prominently (see pathways to impact).