Preventing unwanted breakups

Abstract

Cell separation is an important biological process in plants that is precisely regulated both spatially and temporally. Key separation events include abscission of organs such as leaves and fruit and dehiscence events such as pod shatter in canola and other Brassicas. Polygalacturonases (PGs) are enzymes essential for the degradation of pectin, an important component of the adhesive material between cells. Although there are around 70 PG genes with overlapping expression domains, recent analysis has revealed the function of several PGs in specific aspects of Arabidopsis reproductive development. Upstream regulators that control the expression domain of some of these PGs during reproductive development have also been identified. This information provides new strategies to control unwanted cell separation events in various crops.

Many different biological events that occur during the plant lifecycle involve a process known as cell separation. Obvious examples include leaf and fruit abscission in which individual plant organs detach from the main plant body. However, several less visible separation events also occur, such as microspore separation during pollen development, the emergence of lateral roots through the root surface, and the formation of air spaces within leaves.^1,2 Fruit softening can also be regarded as a form of cell separation as similar biochemical processes are involved (see below).^3,4

As a controlled biological process, cell separation requires the coordination of several biochemical pathways, particularly those related to breaking the adhesion between adjacent cells at the point of separation. Cell adhesion is maintained in part by the presence of homogalacturon-rich pectin which is commonly found in the middle lamella region of cell walls between adjacent cells. Pectin degradation consequently plays an important role in cell separation, and one of the key types of enzyme involved in this process is known as endopolygalacturonase (PG). PGs catalyze random hydrolysis of α-1,4-glycosidic linkages in galacturonic acid, a polymer that forms the main chain of the homogalacturan region of pectin.^5,6

Many genes encoding putative PGs have been identified and characterised to varying degrees, in species such as Arabidopsis, tomato and rice.^3,7 These genes exist in reasonably large gene families with Arabidopsis and rice, for example, predicted to possess 69 and 59 PGs, respectively.^8,9 Sequence divergence within the PG gene family occurred prior to the divergence of major angiosperm families, and there are at least 21 PGs shared between Arabidopsis and rice that are thought to have been present in their most recent common ancestor.^3,8 Other cell wall proteins such as expansin and xyloglucan endotransglycosylase (XET) also consist of multiple gene families in plants.^10,11 In the absence of comprehensive biochemical analyses, it is generally not known whether members of these gene families, which often display different expression patterns, possess different substrate specificities.^10,11 However, for PGs, it is now clear that individual genes in different families and related genes within families can possess discrete and essential biological functions in specific cell separation events.

The best characterized physiological roles for PGs are fruit softening and leaf abscission in tomato, and several cell separation events associated with reproductive development in Arabidopsis.^12–17 In tomato, antisense plants in which a fruit-expressed PG was downregulated exhibit delayed fruit softening, demonstrating that PGs are required for this process to occur normally. Similarly, Virus-induced gene silencing (VIGS) of another tomato PG delays, but does not prevent, leaf abscission, presumably because of redundancy between different PGs.¹² In Arabidopsis, the best studied PGs are QUARTET3 (QRT3) and a group of three related proteins encoded by QRT2, ARABIDOPSIS DEHISCENCE ZONE POLYGALACTURONASE1 (ADPG1) and ADPG2. These genes exhibit various levels of redundancy depending on which separation event is examined. This is likely to reflect each gene's individual expression domain, which are only partly overlapping, and partial redundancy with other as yet unidentified PGs. During microspore separation, which ocurrs as part of pollen formation, both QRT2 and QRT3 (Fig. 1) are essential with no obvious requirement for ADPG1/2.^16–18 By contrast, QRT2 and ADPG1/2 all contribute to anther dehiscence (to release pollen) while QRT2 and ADPG2 promote floral organ abscission and ADPG1/2 are required for pod shatter, the separation of the valves from the siliques to allow seed dispersal.¹⁷ Although each of these genes is also expressed in other tissues where cell separation occurs (see below), the biological significance of this is not yet clear.

Figure 1.

PGs are required for cell separation events during reproductive development in Arabidopsis. (A) SEM image of a qrt2 mutant, which is defective in one of three closely-related PGs, allowed to self-pollinate. Pollen grains remain functional but are arranged in tetrads (arrowhead) due to the failure of the microspores to separate during pollen development. (B) Regulation of DZ formation in Arabidopsis siliques involves the interaction of several putative transcription factors and the eventual localized expression of specific PGs. Proteins shown in grey are not expressed in the indicated tissue. Recent work demonstrates that IND is a promising target for reducing pod shatter in canola. (C) Pollinated siliques (+) and unpollinated pistils (−) express an ADPG1:GUS reporter at the junction between the valves and the replum, despite the fact that unpollinated pistils often do not shatter. (D and E) Transverse sections of WT (D) and ind (E) siliques in the region indicated by the dotted box in (B). Triangles indicate where the DZ forms in WT but not ind mutants. (F) Transverse section of a WT silique as pod shatter begins. GUS staining (blue) represents ADPG1 expression. (G) ADPG1, ADPG2 and QRT 2 are three closely-related PGs required for various cell separation events during reproductive development. ADPG1 and ADPG2 have likely orthologs in canola (PGAZBRAN and RDPG1) that are possible targets for manipulating pod shatter.

In regard to pod shatter, ADPG1 appears to play a more important role than ADPG2. Compared with WT plants, adpg1 mutants exhibit reduced pod shatter unless external mechanical pressure is applied to the siliques. By contrast adpg2 mutants appear to shatter normally while adpg1 adpg2 double mutants do not shatter even with mechanical pressure.¹⁷ Despite the important role of ADPG1, expression of this gene is not in itself sufficient for cell separation. For example, ADPG1 is expressed in the funiculus of unfertilised ovules, which do not abscise.¹⁷ Similarly, ADPG1 is expressed (Fig. 1) in the presumptive DZ of unfertilized pistils which, although remaining green for at least 1 week after anthesis, do not set seeds and, at least under some conditions, do not shatter.^19,20

The biological importance of cell separation combined with increasing understanding of the functional roles of individual PGs provides new opportunities to manipulate crop performance. In this addendum we discuss possible approaches for crop improvement by either increasing or decreasing PG activity to modify specific cell separation events.

Decreasing PG Activity to Reduce Seed Loss Prior to Harvest

Although there are several cell separation events that could potentially be altered in crops, including pollen release (anther dehiscence), cut flower life (floral organ shedding), and fruit harvest (fruit abscission), reducing pod shatter in crops such as canola (also called oil seed rape) is a particularly attractive target. While pod shatter aids seed dispersal in the wild, it can cause economically significant decreases in seed yield in cultivated crops. For example, for canola, in which the seeds are harvested for oil, up to 50% of the potential yield can be lost via premature pod dehiscence and seed abscission.²¹ Similarly, pod shatter prior to harvest can also cause yield losses in crops such as soybean.²²

Although reducing premature pod shatter in canola due to wind or physical impact between adjacent plants is highly desired, it is important that pod shatter should still occur at harvest to allow efficient release and harvest of seeds. At present, the cutting of canola stems so that the plants can be laid on the ground a few weeks before harvest is often used to reduce the incidence of unwanted pod shatter. This is a relatively expensive procedure known as windrowing, and alternative genetic solutions have been sought over several decades. Because existing genetic variation for this trait is relatively limited and difficult to quantify, little progress has been made using conventional breeding approaches.^23–27

Dehiscence of the Arabidopsis silique is very similar to the process in canola, and has been used as an effective model in which to study pod shatter.²⁸ During ovary and fruit development cell fate specification must occur to form the dehiscence zone (DZ), a layer in which cell separation occurs to allow the silique to open. In Arabidopsis, the DZ consists of a few cell layers separating the replum from the edges of the two fused carpels.²⁸ Genetic approaches have revealed that several genes encoding transcription factors are required for DZ differentiation.^29–33 Key proteins include FRUITFUL (FUL), SHATERPROOF1/2 (SHP1/2) and INDEHISCENT (IND), which act together to define the position of the pod DZ (Fig. 1), in part by regulating local levels of the plant hormone auxin.^{29–31,34,35} Once the pod DZ has been established, and following seed set and fruit development, PGs are expressed in the DZ where they degrade pectins and allow separation to occur. Additional enzymes involved in degrading other cell wall components are also required.¹ In canola, and possibly to a modest extent in Arabidopsis, physical pressures set up as the silique dries cause mechanical stresses that promote separation.

Based on these results, one strategy is to prevent pod shatter in Brassica fruit by interfering with DZ formation.^36,37 While this approach has been at least partially successful, one hurdle for crop application is the need to retain sufficient DZ formation to enable consistent pod shatter and seed release at harvest. It has recently been demonstrated that canola contains two IND orthologs, which are excellent targets for reducing pod shattering in canola.³⁸

An alternative approach to reduce seed loss in the field prior to harvest is to apply knowledge of the role of the PG genes expressed in the pod DZ.^39–40 For example, driving expression of the RNase encoded by Barnase using the ADPG1 promoter has been shown to successfully inhibit pod shatter in canola.⁴¹ However, now that a role for ADPG1 and ADPG2 in Arabidopsis pod shatter has been established, targeting the equivalent canola genes represents a promising alternative strategy. There are two ADPG1/ADPG2-related genes (RDPG1 and PGAZBRAN) in canola (Fig. 1) that have been identified previously as PGs potentially involved in pod shatter and floral organ abscission, respectively.^42,43 Amino acid sequence similarity and expression studies indicate that RDPG1 is the ortholog of ADPG1 and PGAZBRAN corresponds to ADPG2.^42,43 Depending on the degree of functional conservation between the species, this result suggests that reduced RDPG1 and/or PGAZBRAN expression may lead to siliques that do not shatter as readily as normal siliques, but do shatter at harvest during threshing. For example, mimicking the Arabidopsis adpg1 phenotype in canola, possibly by silencing RDPG1 expression, may produce siliques that will only shatter during the mechanical forces experienced during threshing at harvest.

The above discussion focuses on manipulation of pod shatter as a way to reduce seed losses in canola. However, a potential alternative solution is to prevent seed abscission directly regardless of whether the silique is intact. In Arabidopsis, an abscission zone (AZ) also forms between each seed and the maternal plant. Thus, either preventing the formation of the seed AZ or preventing PG expression in the canola seed AZ could be used to reduce seed loss even if pod shatter occurs and the silique valves detach. An advantage of this approach is that, because mature seeds are relatively large compared to the funiculus, mechanical forces during harvest and threshing are likely to break the funiculus and allow seed separation independent of a normally functioning AZ.

Similar to the pod DZ, the seed AZ is likely to require a cascade of transcription factors to regulate its formation followed by expression of PGs and other cell wall degrading enzymes at seed maturity. In fact, two putative transcription factors required for seed AZ formation have been identified in Arabidopsis. SEEDSTICK (STK) is a MADS box protein related to FUL and SHP1/2, that regulates funiculus development including seed AZ formation.⁴⁴ HECATE3 (HEC3), is a putative basic Helix-Loop-Helix (bHLH) transcription factor closely related to IND that is required for seed AZ formation.^17,45 Further extending the parallels between the pod DZ and seed AZ, ADPG1 is also expressed in the seed AZ. However, presumably because of redundancy with other PGs, a defect in seed abscission was not detected in the adpg1 mutant.¹⁷ Thus, reduced expression of the canola orthologs of these and/or related genes could be used as an alternative approach to reducing seed losses associated with unwanted pod shatter (Fig. 2).

Figure 2.

HEC3 is required for Arabidopsis seed abscission. During the final stages of silique development two distinct cell separation events occur in WT Arabidopsis plants: pod shatter and seed abscission. IND and HEC3 are closely-related bHLH transcription factors required for normal formation of the silique dehiscence zone and seed abscission zone, respectively. The hec3 ind silique on the right has been opened manually to reveal the seeds. Preventing seed abscission is a potentially novel mechanism to reduce seed losses in crops such as canola.

In conclusion, exciting progress has been made over the last few years in identifying genes that regulate the formation of abscission and dehiscence zones, combined with a more detailed understanding of the physiological roles of individual PGs in the subsequent cell separation events. Using this knowledge it should now be possible to design effective approaches to regulate specific aspects of cell separation in important agricultural and horticultural crops.

Addendum to: Ogawa M, Kay P, Wilson S, Swain SM. ARABIDOPSIS DEHISCENCE ZONE POLYGALACTURONASE1 (ADPG1), ADPG2 and QUARTET2 are polygalacturonases required for cell separation during reproductive development in Arabidopsis. Plant Cell. 2009;21:216–233. doi: 10.1105/tpc.108.063768.