The senescence of flower petals is a gradual process of programmed cell death (PCD) (Rogers, 2006). Based on the morphological changes, petal senescence can be categorized as wilting, withering, or abscission of turgid petals (Van Doorn, 2001). In petunia (Petunia × hybrida), the flower petals are fused to form a corolla, which wilts during senescence (Jones et al., 2009). Petal senescence can be induced by different factors, including aging, pollination, abiotic stresses, and pathogens, where different forms of PCD are involved (Van Doorn and Woltering, 2008). A sequence of events that includes nuclear fragmentation, protein degradation, nutrient mobilization, and membrane leakage accompanies petal senescence (Jones et al., 2009). Various senescence-related genes are upregulated during petal senescence, including genes involved in macromolecule degradation, nutrient remobilization, hormone biosynthesis and signaling, and cell death regulation (Broderick et al., 2014; Tsanakas et al., 2014; Wang et al., 2018; Yang et al., 2019).

Autophagy is an intracellular process that is responsible for the transportation and degradation of toxic or damaged cellular components in eukaryotic cells (Su et al., 2020). Autophagy assists in maintaining cell homeostasis and modulates developmental PCD in plants (Üstün et al., 2017). The most prominent and well-studied autophagy pathway is macroautophagy (hitherto referred to as “autophagy”), where the cellular components are enclosed in double-membrane vesicles (i.e., autophagosomes) to be transported to the vacuole for degradation (Feng et al., 2014). Autophagosomes are constructed collaboratively by proteins encoded from a variety of autophagy genes (ATGs) (Su et al., 2020). ATGs are involved in the regulation of PCD during petal senescence (Shibuya, 2012). The expression of ATGs, including ATG1, ATG4, ATG5, ATG6, ATG7, ATG8a, b, d-f, ATG13, and Phosphatidylinositol 3-Kinase (PI3K), is upregulated during petal senescence in petunia, Japanese morning glory (Ipomoea nil), and hibiscus (Hibiscus rosa-sinensis) (Shibuya et al., 2011; Broderick et al., 2014; Trivellini et al., 2016; Quijia Pillajo et al., 2018). In our previous study, silencing ATG6 or PI3K in petunia results in accelerated petal senescence, as well as reduced flower number and biomass (Lin and Jones, 2021).

ATG6/VPS30/Beclin1 is a central component of the autophagy pathway and is involved in the regulation of PCD (Cao and Klionsky, 2007; Menon and Dhamija, 2018). ATG6 is a single copy gene in Arabidopsis (Arabidopsis thaliana), barley (Hordeum vulgare), and grape (Vitis vinifera), as well as tobacco (Nicotiana benthamiana) and tomato (Solanum lycopersicum), which are in the same family (Solanaceae) as petunia (Tang and Bassham, 2018). Although the copy number of ATG6 in petunia has not been confirmed, only one ATG6-like protein was found when blasting the petunia genomes with Arabidopsis ATG6 sequence.¹ In addition to its function in petal senescence, ATG6 also plays an important role in the regulation of leaf developmental and stress-induced senescence. Barley leaves with silenced ATG6 are more susceptible to darkness, oxidative stress, nutrient deficiency, and salt stress (Zeng et al., 2017). Arabidopsis, tobacco, and wheat (Triticum aestivum) with suppressed ATG6 show early leaf senescence under normal conditions and enhanced pathogen-induced PCD in leaves when infected with viral, bacterial, or fungal pathogens (Liu et al., 2005; Patel and Dinesh-Kumar, 2008; Yue et al., 2015). Even though most of the studies of ATG6 in plants focus on leaf senescence, similarities have been identified between leaf and petal senescence, including the upregulation of autophagy genes (Price et al., 2008; Wagstaff et al., 2009). Understanding the regulatory mechanisms of leaf senescence can facilitate research on petal senescence, and vice versa.

Ethylene interacts with autophagy during petal senescence (Liao and Bassham, 2020). In ethylene-treated petunia flowers, the expression of ATG8a-d increases, while the application of an ethylene inhibitor suppresses ATG8a-d expression (Shibuya et al., 2013). Treating pollinated flowers with an ethylene inhibitor delays the induced expression of ATG8a-d (Shibuya et al., 2013). Ethylene-induced ATG expression in flowers is thought to be regulated by ethylene-responsive transcription factors (TF). Arabidopsis ethylene-responsive TFs EIL and AP2 interact with the promoters of ATG8a and ATG8h in yeast one-hybrid assays (Wang et al., 2020). In tomato, direct binding of ethylene-responsive TF ERF5 to the promoters of ATG8d and ATG18h has been shown in vitro (Zhu et al., 2018). Correlations are also found between autophagy and ethylene biosynthesis. The expression of a key ethylene biosynthesis gene, 1-aminocyclopropane-1-carboxylic acid synthase (ACC synthase, ACS), decreases in the senescing petals of ATG6- or PI3K-silenced petunias (Lin and Jones, 2021). In PI3K-overexpressing tobacco, the production of ethylene is enhanced as a result of increased expression of another ethylene biosynthesis gene ACC oxidase 1 (ACO1) despite decreases in the expression of ACO2 (Dek et al., 2017).

The remobilization of mineral nutrients is a critical process during petal senescence (Jones, 2013). Evolutionarily, the purpose of flower petals is to attract pollinators for sexual reproduction. Because it is energetically costly to maintain these elaborate structures, the flower petals senesce once the flower is pollinated, or when the stigma is no longer receptive to pollination (Jones et al., 2009). To preserve resources, nutrients are remobilized from the senescing petals to other tissues or organs for recycling (Borghi and Fernie, 2020). Nutrient remobilization during petal senescence is regulated by autophagy (Avila-Ospina et al., 2014). Nuclear fragmentation and DNA mass reduction are delayed in pollinated flowers treated with 3-methyladenine (3-MA), an inhibitor of autophagy protein PI3K (Yamada et al., 2009). The application of the autophagy inhibitor concanamycin A limits the growth of the ovary in pollinated flowers, suggesting that autophagy is a key regulator in pollination-mediated nutrient recycling (Shibuya et al., 2013). Phosphorus (P) and nitrogen (N) are two of the major nutrient elements remobilized during petal senescence (Jones, 2013), and an increase of N and P is found in the ovary as a sink organ after pollination-induced senescence (Hew et al., 1989). Ethylene plays an important role in autophagy-mediated nutrient remobilization during petal senescence (Jones, 2013). In both pollinated and unpollinated senescing flowers, the remobilization of N and P is reduced in ethylene-insensitive transgenic petunia compared to the ethylene-sensitive wild-type control (Jones, 2008).

The clustered regularly interspaced short palindromic repeats (CRISPR) and associated endonuclease 9 (Cas9) system is a powerful tool for reverse genetic analysis and horticulture crop improvement (Karkute et al., 2017). Mutagenesis of the ACO genes in petunia or a NAC transcription factor in Japanese morning glory using CRISPR/Cas9 successfully delays petal senescence in these plants (Shibuya et al., 2018; Xu et al., 2020, 2021). To better understand the function of PhATG6 in the regulation of corolla senescence and nutrient recycling in petunia and to characterize ethylene’s role in autophagy-mediated corolla senescence, we analyzed stable PhATG6-knockout (KO) mutants generated with CRISPR/Cas9-mediated gene editing.

Knocking out PhATG6 in petunia accelerated corolla senescence by 2–3 days (Figure 3A), indicating that autophagy is involved in delaying corolla senescence. This result supports our previous finding, where silencing PhATG6 using Virus-induced Gene Silencing (VIGS) results in decreased flower longevity by 3–4 days in petunia (Lin and Jones, 2021). Even though knocking out PhATG6 with CRISPR did not seem to reduce flower longevity as much as VIGS, this could be due to the difference in the plant growing environment and petunia cultivars. The PhATG6-KO lines (‘Mitchell Diploid’) were grown under greenhouse conditions, while the PhATG6-silenced petunias (‘Picobella Blue’) were grown in growth chambers. The proliferation of virus may also play a role by placing more stress on the VIGS plants, leading to a more severe reduction of flower longevity. In addition to ATG6, the genetic modifications of another autophagy gene, PI3K, also affect the longevity of flowers (Dek et al., 2017; Lin and Jones, 2021).

The morphological cell death processes in flower corollas vary in different species. In Japanese morning glory, corolla PCD starts from the edge of the petal limbs and gradually moves toward the base of the petals (Shibuya et al., 2014). Our results showed that petunia corolla PCD began in the corolla neck connecting the limb and tube and proceeded to the rest of the corolla, with the majority of petal cells maintaining intact plasma membranes until the last 1–2 days of the senescence process (as shown by Evans Blue staining; Figure 2). Similarly, in Alstroemeria (Alstroemeria peruviensis), the PCD in petals does not occur until the last stages of flower senescence, as demonstrated by the electrolyte leakage, DNA laddering, and protease activity in petals (Wagstaff et al., 2003). Cut daylily (Hemerocallis sp.) flowers also do not show reduced membrane stability until visible senescence symptoms appear (Chakrabarty et al., 2009). Our results provide further evidence that the loss of membrane integrity does not occur gradually but at the late stage of petal senescence.

PhATG6 is functionally involved in fruit (capsule) and seed production in petunia (Figure 4). Autophagy is the major regulator of nutrient recycling and therefore indirectly affects the nutrient supplies for seeds (Ren et al., 2014). In rice (Oryza sativa), the overexpression of ATG8a or ATG8b increases grain number and yield per plant, which are reduced in plants with RNAi-suppressed ATG8b compared to the wild-type plants (Yu et al., 2019; Fan et al., 2020). Rice atg7 mutants and ATG8b-RNAi plants also show reduced grain quality, demonstrated by the chalky appearance and the starch and protein composition of the grains (Sera et al., 2019; Fan et al., 2020). Arabidopsis mutants of ATG5 or ATG7 produce fewer seeds per silique, fewer siliques per plant, and lower seed yield than the wild-type plants (Guiboileau et al., 2012; Barros et al., 2017; Lornac et al., 2020). Arabidopsis plants with heterozygous PI3K (+/−) gene produce fewer numbers of seeds per silique (Lee et al., 2008). The number of siliques per plant is decreased in Arabidopsis heterozygous atg6 (+/−) mutants (Fujiki et al., 2007; Harrison-Lowe and Olsen, 2008). Similarly, we observed smaller capsules and fewer seeds per capsule in the PhATG6-KO petunias compared to the controls (Figures 4A,B,D). The average weight of individual seeds was higher in most of the PhATG6-KO lines than the control plants (Figure 4E), consistent with what is observed in Arabidopsis atg7 mutants (Barros et al., 2017). Although we did not find a difference in capsule maturation time between the PhATG6-KO plants and the controls (Figure 4C), a study in pepper (Capsicum annuum) shows an increase in autophagy gene expression, autophagy protein abundance, and autophagosome-like vesicles in mature fruit, suggesting that autophagy is involved in fruit ripening (López-Vidal et al., 2020).

Ethylene is involved in the regulation of accelerated corolla senescence in the PhATG6-KO lines. As a regulator for senescence in plants, ethylene plays a key role in autophagy-mediated petal senescence (Shibuya et al., 2013). The early elevation of ethylene levels was likely the reason for early corolla senescence in the PhATG6-KO lines. Our results suggested that autophagy could affect the timing of ethylene biosynthesis. However, in tobacco plants with overexpressed PI3K, even though the detached flower senescence is accelerated, and the ethylene production is higher in the entire detached flowers, no difference in the timing of ethylene production is observed (Dek et al., 2017). The difference in ethylene production timing may indicate different regulatory mechanisms of corolla senescence. The expression of ethylene biosynthesis genes PhACS and PhACO1 was also induced earlier in the PhATG6-KO corollas compared to the controls, but later than the elevated ethylene levels (Figures 5B,C). As shown in carnation flowers, ethylene treatment can induce the expression of ethylene biosynthesis genes (Jones, 2003). Therefore, the increase in ethylene biosynthesis gene expression in petunia corollas was likely induced by the earlier produced ethylene. Multiple ACS genes with different roles have been found in Arabidopsis, tobacco, carnation, orchid (Phalaenopsis sp.), tomato, and potato (Solanum tuberosum) (Yip et al., 1992; Destéfano-Beltrán et al., 1995; Liang et al., 1995; Bui and O’Neill, 1998; Jones and Woodson, 1999; Ge et al., 2000), while little is known about the function of additional ACS genes in petunia. It is possible that the ethylene produced earlier in the corollas of PhATG6-KO flowers was synthesized by another ACS member in petunia. Another possibility is that the ethylene could be synthesized using ACC translocated from other parts of the flowers, or the ethylene itself could be translocated, as shown in carnation and orchid (Reid et al., 1984; Woltering, 1990; O’Neill et al., 1993). Even though ACS expression increased at the end of corolla senescence, ethylene production decreased, indicating that ACS activity or stability could be affected by posttranscriptional regulation during corolla senescence.

The early upregulation of PCD-related gene expression supported early PCD in the senescing corollas of PhATG6-KO lines. ATG6 is an essential component of the PI3K protein complex in the autophagy pathway (McKnight and Yue, 2013). ATG6 serves as a binding hub for different interacting proteins to regulate the composition, subcellular localization, and/or function of the PI3K complex and consequently affect autophagy activity (He and Levine, 2010; Hill et al., 2019). Therefore, knocking out ATG6 could result in direct and/or indirect effects on autophagy and other PCD regulatory mechanisms. Our previous study shows that ATG6 expression is upregulated when PI3K is silenced, and ATG8d expression is downregulated when ATG6 is silenced (Lin and Jones, 2021). This study showed that PhATG6, PhPI3K, and PhATG8d expression was upregulated earlier in the PhATG6-KO corollas than the controls, correlated with the accelerated senescence (Figures 6A–C). The early induction of these genes was likely a result of the early increase of ethylene production, as ethylene treatment induces autophagy gene expression in petunia and Japanese morning glory (Shibuya et al., 2013; Quijia Pillajo et al., 2018). Apoptosis is another type of PCD modulated by caspases, and the interaction between autophagic PCD and apoptotic PCD is based on caspase-mediated cleavage of ATG6 (Djavaheri-Mergny et al., 2010; Wirawan et al., 2010). In plants, the apoptosis-like PCD is regulated by metacaspases, a group of cysteine proteases similar to mammalian caspases (Minina et al., 2013). RNAi-suppressed PhMC1 similarly causes accelerated corolla senescence in petunia (Chapin et al., 2017). The expression of PhMC1 was also upregulated earlier in PhATG6-KO lines (Figure 6D), suggesting accelerated PCD in the corollas and indicating a potential interaction between autophagy and metacaspases at the transcriptional level.

Knocking out PhATG6 in petunia negatively impacted nutrient remobilization during corolla senescence of unpollinated flowers. During leaf senescence, autophagy plays an important role in N remobilization as shown in the atg mutants of Arabidopsis, rice, and maize, where the remobilization efficiency of N is reduced (Guiboileau et al., 2012; Li et al., 2015; Wada et al., 2015; Yu et al., 2019; Fan et al., 2020). Even though the remobilization of N was not affected in the PhATG6-KO lines during corolla senescence, P remobilization was decreased in the KO lines (Figures 7A,B). The expression of two petal senescence-specific genes, a cysteine protease PhCP10 and a phosphate transporter PhPT1, was induced earlier in the PhATG6-KO lines, and PhCP10 showed a trend of downregulated expression in the KO plants (Figures 7C,D), indicating that autophagy affects the transcription of genes involved in nutrient degradation and transportation. Cysteine proteases play an important role in protein degradation during petal senescence (Jones et al., 1995; Wagstaff et al., 2002; Arora and Singh, 2004). PhCP10, a homolog to Arabidopsis Senescence-Associated Gene 12 (SAG12), is exclusively expressed in senescing petals and its expression is delayed in ethylene-insensitive petunias, corresponding to the delayed corolla senescence (Jones et al., 2005). PhPT1, a high-affinity phosphate transporter, belongs to the Phosphate Transporter1 family in petunia (Wegmüller et al., 2008; Chapin and Jones, 2009). Compared to the other petunia high-affinity phosphate transporters, PhPT1 is highly expressed in senescing petunia corollas (Chapin and Jones, 2009). Ethylene affects the expression of PhPT1, as the transcript levels of PhPT1 decrease in ethylene-insensitive plants during corolla senescence, and ethylene treatment induces PhPT1 expression in corollas (Chapin and Jones, 2009). The early induction of PhCP10 and PhPT1 expression in the PhATG6-KO lines was likely the result of the early increased ethylene production. Autophagy may affect nutrient remobilization via ethylene-mediated transcriptional regulation of genes involved in nutrient degradation and transportation.

Knocking out a central autophagy gene PhATG6 in petunia led to accelerated petal senescence, decreased P remobilization, and reduced seed yield, likely due to early expression of senescence-related genes and early ethylene production. Our study contributes to the current understanding of autophagy in senescence and nutrient remobilization and provides information for the improvement of flower longevity in ornamental plants.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

YL and MLJ conceptualized the project and designed the experiments. YL conducted the experiments and led the writing of the manuscript. MLJ acquired resources and funding for the study, supervised the project, and edited the manuscript. All authors contributed to the manuscript revision and approved the submitted version.

The funding for this study was sponsored by the American Floral Endowment, the OSU D.C. Kiplinger Floriculture Endowment, and the OSU Alumni Grants for Graduate Research and Scholarship. Salaries and research support were provided in part by State and Federal funds appropriated to the College of Food, Agricultural and Environmental Sciences, The Ohio State University. Department of Horticulture and Crop Science Manuscript #21-15.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.