How do the downregulation and overexpression of the PA

On the basis of my results, it seems that genetic modifications in one gene of the PA pathway caused changes throughout the phenolic pathway. Previously, it has been found that GM plants may show unintended changes in other genes (e.g. Hjältén et al.

2007). These changes have been reported to induce non-predictable traits, such as growth changes (Hjältén et al. 2007, Axelsson et al. 2011). Genetic modification may also cause pleiotropic effects, if the locations and the numbers of the transgene are not completely controlled (Novak & Haslberger 2000).

Nevertheless, ANRi birches provided new and surprising changes in the concentration of phenolics and also in responses to nitrogen levels in the stems (II). Further, the ANRi birches showed a strong decrease in growth, and biosynthesis in the phenolic pathway turned towards phenolic acids and flavonols (Figure 3, II). Overexpression of the MYB134 gene resulted in the accumulation of PAs in hybrid aspen (Mellway et al. 2009, III) and in decreased content of PGs, which affected leaf consumption by leaf beetles (Figure 3, III).

The competition for biosynthetic precursors within the phenolic pathway may due to trade-off in P. tremuloides (Donaldson et al. 2006, Osier & Lindroth 2006). Especially environment with limited resources, such as low fertilization or weak light availability have caused trade off between PAs and PGs in several genotypes in P. tremuloides. (Donaldson et al. 2006, Osier & Lindroth 2006).

3.2.1 Downregulation of ANR caused alterations of chemotype and phenotype

Gene inhibition causes silencing or damage of a known gene. In other words, mRNA level is reduced, but not completely eliminated (Mocellin & Provenzano 2004). Inhibition of ANR gene caused dramatic changes in the phenolic pathway in B.

pendula. Concentrations of PAs decreased clearly in response to inhibition of the ANR gene. Similarly, antisense downregulation of the ANR gene in Populus trichocarpa has been found to cause a decrease of PA concentration and an increase of anthocyanidins in the leaves of Populus (Wang et al. 2013b). By contrast, concentration of flavonols accumulated in the studied tissues and the concentrations of several other compounds of the phenolic pathway changed (II) (Figure 3). As expected, most of the upstream phenolics of the phenolic pathway, such as flavonol glycosides, accumulated as much as ten times more in ANRi birches compared to WT due to PA inhibition (Figure 3 red and blue arrows). However, only kaempferols decreased due to ANR inhibition (Figure 3). There were also several phenolic compounds in ANRi birches that were not detected in WT birches (II). ANR inhibition caused alterations in the chemotype of the birches and several steps of the phenyl propanoid route changed.

Thus the available carbon was allocated differently in different parts of the pathway than normal (Figure 3). Generally, ANR genes play a crucial role in PA synthesis in all plant organs in different plant species (Kovinich et al. 2012, Wang et al. 2013b).

Leaves and stems responded differently to downregulation of ANR (II), both in chemistry (Figure 3 red and blue arrows, II) and

41 in gene expression. This result supports the existence of organ-specific regulation of phenolic synthesis. It is proposed that plant tissues have consistent synthesis of PAs, which relies on a balance between the activity of the activator and of the repressor MYBs transcription factor (Paolocci et al. 2010, Yoshida et al. 2015). This creates a feedback system between PAs and other phenolics via MYB factor regulation (Yoshida et al. 2015). In my study (II), the regulation of feedback may also be due to the transcription factor of MYBs. Because the MYB factors are important regulators in phenolic pathway the feedbacks between different phenolics be caused by MYB regulation. For example, some low-molecular weight phenolics, such as salidroside in leaves, and cinnamic acid derivatives (possible precursors of other phenolics) and phenolic aglycones in stems increased in the ANRi birches (Figure 3 red and blue arrows). This indicates positive feedback regulation in the upstream phenolics of ANRi birches. The downregulation of cinnamic acid 4-hydroxylase (C4H) in tobacco (Nicotiana tabacum L. cv Xanthi) caused negative feedback that reduced PAL activity in the plant (Blount et al. 2000). The flow of precursors into the phenylpropanoid synthesis is controlled, at least in part, via feedback regulation between PAL and production of cinnamic acid (e.g. Blount et al. 2000) (Figure 3).

Figure 3. The major biosynthetic routes to the various classes of phenylpropanoid compounds, and the effect of ANRi and overexpression of MYB on their concentrations (Phenylalanine ammonia-lyase = PAL, cinnamic acid 4-hydroxylase = C4H, 4-coumarate: CoA ligase = 4CL,

43 chalcone synthase = CHS, chalcone isomerase = CHI, flavanone 3-hydroxylase = F3H, cinnamoyl-CoA reductase = CCR, stilbene synthase

= STS, dihydroflavonol reductase = DFR, flavonol synthase = FLS, anthocyanidin synthase = ANS, leucoanthocyanidin reductase LAR, anthocyanidin reductase = ANR). Red arrows represent increased (arrow up) or decreased (down) content in the leaves of the ANRi birches (II). Blue arrows represent increased (up) or decreased (down) content in the stems of ANRi birches (II). Green arrows represent increased (up) or decreased (down) content in MYB134 aspens’ leaves (III). Arrows indicate the statistical difference between WT and (at least two) GM lines, ns means non-significant results. The pathway is a combination of following references: Crozier et al. 2000, Xie et al. 2003, Bowsher et al. 2008, Mellway et al. 2009, Babst et al. 2010, Singh et al. 2010, Ali et al. 2011, Dixon et al. 2013, Yoshida et al. 2015.

Not only was the chemotype of ANRi birches different compared with that of WT birches, but the changes in corresponding phenotype were also considerable (II). The color of the stems and leaves was more reddish or brownish in the ANRi birches than in the WT plants, indicating increased anthocyanin content (Figure 4). Moreover carotenoids cause the red dyes in B. pendula and the level of carotenoids is sensitive for air pollution (Yamaji et al.

2003, Sillanpää et al. 2008). The color might prove the pathway was turned to anthocyanins or carotenoids (II). Similarly, ANR inhibition caused an increased concentration of anthocyanidins in the leaves of Populus trichocarpa (Wang et al. 2013b).

The dwarf-like growth types of Taxus and Tsuga correlated with less nuclear flavanols and increasing flavonols (Feucht et al.

2014). Many flavonol glycosides and anthocyanin have been shown to accumulate in higher amounts in the dwarf-like plants (Besseau 2007). Accordingly, in the ANRi birches that showed a high concentration of flavonols, their phenotypes appeared to be dwarf- and bushy-like (II). Discrepancy between results of studies in conifers and in deciduous trees might be due to fundamental differences between conifer and deciduous trees.

Flavonoid accumulation has also been found to cause/induce decreased plant growth in Arabidopsis (Brown et al. 2001, Besseau et al. 2007). These Arabidopsis studies indicated changes in

transport of auxin that strongly influences plant growth and development (Woodward & Bartel 2005). Disrubtions of auxin transport might be one reason for the phenolic changes in ANRi birches.

3.2.2 Overexpression of the MYB134 gene implied trade-offs Overexpression of a transgene can cause an energetic cost to a plant, and lead to reallocation of energy and resources between growth, defense and reproduction. The overexpression of MYB134 caused a substantial increase of PA concentration in the two studied MYB lines (III). These lines also had a reduced total salicylate concentration, while the phenolic acids accumulated markedly (Figure 3 green arrows). Total concentration of phenolics was at a lower level in the MYB lines, whereas the PAs accumulated. These results implied a fairly strong trade-off between PAs and low-molecular weight phenolics, especially salicylates, and is in accordance with previous studies (Julkunen-Tiitto 1989, Veteli et al. 2002, Payyavula et al. 2009, Mellway et al.

2009, Boeckler et al. 2014). In addition, all lines, except for the WT, contained the flavonol, quercetin 3-galactoside. Thus, MYB overexpression caused some unexpected phenolic accumulation in aspen (Figure 3 green arrows, III).