Biochemical and molecular characterization of Alternaria alternata isolates highly resistant to procymidone from broccoli and cabbage

Alternaria alternata, a causal agent of leaf blights and spots on a wide range of hosts, has a high risk of developing resistance to fungicides. Procymidone, a dicarboximide fungicide (DCF), has been widely used in controlling Alternaria leaf blights in China for decades. However, the resistance of A. alternata against DCFs has rarely been reported from crucifer plants. A total of 198 A. alternata isolates were collected from commercial fields of broccoli and cabbage during 2018–2019, and their sensitivities to procymidone were determined. Biochemical and molecular characteristics were subsequently compared between the high-level procymidone-resistant (ProHR) and procymidone-sensitive (ProS) isolates, and also between ProHR isolates from broccoli and cabbage. Compared with ProS isolates, the mycelial growth rate, sporulation capacity and virulence of most ProHR isolates were reduced; ProHR isolates displayed an increased sensitivity to osmotic stresses and a reduced sensitivity to sodium dodecyl sulfate (SDS); all ProHR isolates showed a reduced sensitivity to hydrogen peroxide (H2O2) except for the isolate B102. Correlation analysis revealed a positive cross-resistance between procymidone and iprodione, or fludioxonil. When treated with 10 μg/mL of procymidone, both mycelial intracellular glycerol accumulations (MIGAs) and relative expression of AaHK1 in ProS isolates were higher than those in ProHR isolates. Sequence alignment of AaHK1 from ten ProHR isolates demonstrated that five of them possessed a single-point mutation (P94A, V612L, E708K or Q924STOP), and four isolates had an insertion or a deletion in their coding regions. No significant difference in biochemical characteristics was observed among ProHR isolates from two different hosts, though mutations in AaHK1 of the cabbage-originated ProHR isolates were distinct from those of the broccoli-originated ProHR isolates.

The DCFs, including procymidone and iprodione, have been registered to control ALSB of cabbage and broccoli for decades in China (Ma & Michailides 2004). The biochemical mechanism of DCFs against phytopathogenic fungi is documented to regulate high osmolarity glycerol (HOG) MAP kinases, interfering with the phosphorylation of transduction-associated proteins and glycerol biosynthesis (Cui et al. 2002;Lin & Chung 2010). DCFs and phenylpyrroles are reported to target the same enzyme, a fungal two-component histidine kinase (HK), which belongs to group III hybrid histidine kinases involved in an osmotic-regulatory signal transduction cascade of mitogen-activated protein kinase (MAPK) signaling pathway, i.e., HOG-MAPK (Avenot et al. 2005;Lin & Chung 2010;Ren et al. 2016). This cascade not only perceives extracellular stimuli, but also maintains intracellular osmotic homeostasis. Positive crossresistance between DCFs and phenylpyrroles is frequently observed in phytopathogenic fungi highly resistant to fludioxonil, including Botrytis cinerea (Sang et al. 2018), Bipolaris maydis (Han et al. 2017), Stemphylium solani (Wu et al. 2015) and Sclerotinia sclerotiorum (Duan et al. 2013). The typical structure of fungal group III HKs from major phytopathogenic fungi is composed of three domains: Histidine kinases-Adenylate cyclases-Methyl accepting proteins and Phosphatases (HAMPs, functionally named as "sensor domain") at a highly variable N-terminus, a central transmitter domain composed of histidine kinase A (HisKA) and cognate histidine kinase-like ATPase catalytic subdomains (HATPase_c), and a C-terminus receiver domain harboring a three amino-acid signature (DDK) (Herivaux et al. 2016).
Due to the extensive application of DCFs, the resistance against this type of fungicides has been reported in several phytopathogenic fungi, including Alternaria spp. (Dry et al. 2004;Avenot et al. 2005;Luo et al. 2008Luo et al. , 2012, Sclerotinia spp. (Ma et al. 2009;Kuang et al. 2011;Duan et al. 2013), B. cinerea (Fraile et al. 1986;Oshima et al. 2002;Cui et al. 2004;Grabke et al. 2014), Magnaporthe grisea (Motoyama et al. 2005), Stemphylium vesicarium (Alberoni et al. 2010), Neurospora crassa (Fujimura et al. 2000;Miller et al. 2002;Ochiai et al. 2010) and Cochliobolus heterostrophus (Yoshimi et al. 2003(Yoshimi et al. , 2004. The molecular mechanisms of DCFresistance include amino acid substitution in group III HK proteins, deletion or insertion mutations in their coding regions (Cui et al. 2004;Dry et al. 2004;Luo et al. 2008). Mutations in AaHK1 (insertion or deletion) of A. alternata lead to premature termination of the AaHK protein (Dry et al. 2004). Similarly, diverse mutations are found in AbNIK1 of Alternaria brassicicola, including nonsense mutations, a frameshift and a single amino acid substitution (E753K) (Avenot et al. 2005), while deletions and single amino acid substitutions are detected in AlHK1 of Alternaria longipes (Luo et al. 2008(Luo et al. , 2012. In brief, it is intriguing to observe a high degree of diversity among mutations in HKs of the DCF-resistant isolates from Alternaria spp.. The mutations of HK1s are involved in the development of resistance to DCFs in fungal pathogens under certain selective pressure (Dry et al. 2004). Recently, growers has raised concerns about the decreased efficacy of DCFs against ALSB on cabbage and broccoli in Jiangsu Province of China. The objectives of this study were to (i) investigate whether procymidoneresistant (Pro R ) populations of A. alternata has developed in commercial fields of cabbage and broccoli that were repetitively exposed to procymidone; (ii) compare biochemical characteristics between Pro S and Pro HR isolates, and also between Pro HR isolates from broccoli and cabbage; (iii) reveal possible molecular resistant mechanisms of A. alternata against procymidone.

Results
Frequencies of resistance against procymidone in natural populations of A. alternata from broccoli and cabbage The 198 isolates of A. alternata were obtained during 2018-2019, including 114 isolates from broccoli and 84 isolates from cabbage in Yancheng City, Jiangsu Province, China. Of the 114 isolates from broccoli, 7 highlevel procymidone-resistant (Pro HR ) isolates and 60 lowlevel procymidone-resistant (Pro LR ) isolates were detected with resistance frequencies of 6.14 and 52.63%, respectively (Table 1). Of the 84 isolates from cabbage, in contrast, 3 Pro HR isolates and 45 Pro LR isolates were found with resistance frequencies of 3.57 and 53.57%, respectively (Table 1). The data indicate that DCFresistant populations of A. alternata has developed in commercial fields of broccoli and cabbage.
The EC 50 values of all Pro HR isolates from broccoli and cabbage were more than 100 μg/mL, while the values for Pro S isolates were less than 1 μg/mL (Table 2).

Cross-resistance
All the Pro HR isolates from broccoli and cabbage were also highly resistant to fludioxonil and iprodione based on their EC 50 values ( Table 2). The Spearman's rank correlation coefficient between the sensitivity against procymidone and fludioxonil, or iprodione was 0.9870 or 0.9824 (P < 0.001), respectively, suggesting a positive correlation (a positive cross-resistance) between procymidone and fludioxonil or iprodione was present in these tested A. alternata isolates.
Mycelial growth, sporulation and virulence in Pro HR and Pro S isolates of A. alternata The mycelial growth rates of the Pro HR isolates on fungicide-free potato dextrose agar medium (PDA) were significantly reduced compared with those of the Pro S isolates (P = 0.05) (Table 3). No significant difference in mycelial growth was observed among the Pro HR isolates from cabbage and broccoli (Table 3). Compared with Pro S isolates, the sporulation capacity of the Pro HR isolates (except for isolates B46 and B88) was significantly reduced (P = 0.05) (Table 3). Furthermore, the virulence of the Pro HR isolates (except for isolates B35 and B53) on broccoli detached leaves was also significantly weakened compared with that of the Pro S isolates (P = 0.05) ( Fig. 2 and Table 3). Intriguingly, all Pro HR isolates from cabbage and broccoli fluctuated markedly in sporulation capacity and virulence.
Sensitivities of Pro HR and Pro S isolates of A. alternata to stress agents The group III HK is involved in osmotic-regulatory signal transduction cascade of MAPK signaling pathway (Lin & Chung 2010). To determine whether mutations in AaHK1 affect the sensitivity of A. alternata to osmotic and oxidative stresses, cell membrane permeability or cell wall integrity, the mycelial growth inhibition rates of the Pro HR and Pro S isolates on PDA supplemented with an osmotic stress agent (1.2 M KCl or 1.2 M NaCl), an oxidative stress agent (10 mM hydrogen peroxide, H 2 O 2 ), a cell membrane stress agent (0.05% Sodium dodecyl sulfate, SDS) or a cell wall damage agent (0.05% Congo red, CR) were determined.
The data showed that all Pro HR isolates were strongly inhibited by 1.2 M KCl or 1.2 M NaCl, with mycelial growth inhibition rates ranging from 84 to 95% or from 85 to 95%, indicating that they were more sensitive to osmotic stresses than Pro S isolates (51-78% and 64-78%  Means in a column followed by the same letter are not different according to Fisher's least significant difference (P = 0.05). Pro S , procymidone-sensitive isolates; Pro HR , high-level procymidone-resistant isolates; EC 50 , effective concentration for 50% inhibition of mycelial growth mycelial growth inhibition by 1.2 M KCl and 1.2 M NaCl, respectively) (P = 0.05) (Fig. 3a, b). However, the mycelial growth rates of the Pro HR isolates were higher than those of the Pro S isolates when treated with 0.05% SDS (P = 0.05) (Fig. 3c). Under 0.05% CR treatment condition, the radial growth rates of the Pro HR isolates displayed different degrees of inhibition relative to those of the Pro S isolates (Fig. 3d). Additionally, except for the Pro HR isolate B102 (having no mutation in AaHK1), the mycelial growth rates of the Pro HR isolates were higher than those of the Pro S isolates when treated with 10 mM H 2 O 2 (Fig. 3e). Furthermore, the sensitivity to stress agents varied greatly among the Pro HR isolates from broccoli or cabbage.

Mycelial intracellular glycerol accumulations (MIGAs) in Pro HR and Pro S isolates of A. alternata
DCFs are reported to affect glycerol biosynthesis and phosphorylation in HOG-MAPK pathway (Cui et al. 2002). Under procymidone-free condition, no significant difference was observed in MIGAs between the Pro S and Pro HR isolates (P > 0.05). After treated with 10 μg/mL procymidone for 5 h, MIGAs in Pro S isolates increased significantly and were 2.70-3.04 folds of those under procymidone-free condition, higher than that of 1.02-1.27 folds by the ten Pro HR isolates under the same conditions (P = 0.05) (Fig. 4). Furthermore, no significant difference was observed in MIGAs among the Pro HR isolates from broccoli and cabbage (P > 0.05) (Fig. 4). Sequence alignments of AaHK1 from Pro S and Pro HR isolates of A. alternata To characterize mutations in AaHK1 proteins, the complete nucleotide sequences of AaHK1 in all Pro HR and 20 arbitrarily selected Pro S isolates from broccoli and cabbage were sequenced and aligned. The results showed that the deduced AaHK1 protein is 1330 amino acids (aa) in length, possessing six HAMP repeat domains at Nterminus, followed by a His Kinase A (phospho-acceptor) domain (HisKA), an HK-like ATPase domain (HATPase_ c), and a response regulator domain at C-terminus (Fig. 5).
The twenty Pro S isolates possessed identical AaHK1 sequences with no genetic diversity, as illustrated by the four representative isolates (C10, C28, B16 and B21). No mutation was observed in AaHK1 of all Pro LR isolates from broccoli and cabbage (data not show). All mutations in the coding regions of AaHK1 from Pro HR isolates are presented in Table 4. Except for the Radial mycelial growth of A. alternata isolate was measured after 7 days of incubation on potato dextrose agar (PDA) plates at 25°C. Spores were harvested from the surface of 7-day-old PDA colony with sterile water (0.1% Tween 20). Lesion areas were determined on broccoli leaves at 7 days post-inoculation (dpi). The mean value ± SD of each isolate was calculated from three independent experiments. Means in a column followed by the same letter are not different according to Fisher's least significant difference (P = 0.05). Pro S , procymidone-sensitive isolates; Pro HR , high-level procymidone-resistant isolates from those from broccoli. Nucleotide sequence accession numbers corresponding to AaHK1 from the tested Pro HR and Pro S isolates are presented in Table 5.
The relative expression level of AaHK1 in Pro S and Pro HR isolates of A. alternata The relative expression level of group III HK gene in B. cinerea is correlated with its resistance against histidine kinase inhibitors (Ren et al. 2016). Here, we investigated changes of AaHK1 expression in Pro S and Pro HR isolates against procymidone. In the absence of procymidone, the expression of AaHK1 was relatively low in the Pro S and Pro HR isolates (Fig. 6). When treated with procymidone at 10 μg/mL for 12 h, the relative expression level of AaHK1 in the Pro S isolates was 3.36-3.65 folds of that of untreated Pro S isolates, much higher than that of 0.89-1.22 folds in the ten Pro HR isolates under the same condition (Fig. 6). Furthermore, no significant difference in the relative expression of AaHK1 was observed among Pro HR isolates from two different hosts (broccoli and cabbage).

Discussion
Procymidone, a DCF fungicide for ALSB control, has been extensively used for decades in China, however, there is a growing concern about its decreased efficacy nowadays. Our data in this study indicate that DCFresistance has developed in natural populations of A. alternata from commercial fields of broccoli and cabbage in Yancheng, Jiangsu Province, China. Pro LR populations developed faster than Pro HR populations in these fields during 2018-2019, and Pro HR populations from broccoli were larger than those from cabbage, which may be attributed to different selective pressures exerted by the use of a fungicide in different ways in the field. Long-term monitoring of population change should be performed to detect the dynamic of Pro HR population and formulate corresponding management strategy in time. To our knowledge, there has been no report so far on DCF-resistance in A. alternata from commercial fields of broccoli and cabbage. DCF-resistance has been reported in several Alternaria species, including A. alternata (Dry et al. 2004), A. longipes (Luo et al. 2008), A. brassicicola (Avenot et al. 2005) Fig. 4 Comparison of mycelial intracellular glycerol accumulations (MIGAs) among A. alternata Pro S and Pro HR isolates. The mycelia of each isolate were treated with procymidone at 0 or 10 μg/mL after growth in PDB for 5 h. Bars denote the standard errors of three repeated experiments. The asterisk indicates significant difference between procymidone-treated sample and control. The Pro S isolates include C10, C28, B16 and B21. The Pro HR isolates include C5, C13, C40, B30, B35, B46, B53, B88, B102 and B108 and A. arborescens (Ma & Michailides 2004). However, the resistance mechanism of these fungal pathogens to DCFs and phenylpyrroles has not yet been determined. To elucidate the possible molecular mechanism of DCF-resistance, the nucleotide sequences of AaHK1 of A. alternata Pro HR isolates from broccoli and cabbage were sequenced and aligned in this study. The results demonstrate that a high level of nucleotide sequence homology exists between AaHK1 and os-1 from N. crassa (Ochiai et al. 2010), B. cinerea (Ren et al. 2016), A. brassicicola (Avenot et al. 2005) and A. longipes (Luo et al. 2008).
Abundant mutation types were detected in AaHK1 of the Pro HR isolates from cabbage and broccoli, including deletions at HAMPs or between HATPase_c and Rec domains, an insertion at the Rec domain, amino acid substitutions at HAMPs or HATPase_c domain. The observed premature termination mutations in AaHK1 were generated by a 2-bp deletion, an 8-bp insertion or C to T transition. Intriguingly, no mutation was detected in AaHK1 of B102 (a Pro HR isolate from broccoli), with the same situation occurring in DCF-resistant isolates of A. alternata (from pistachio), A. arborescens (Ma & Michailides 2004) and A. brassicicola (Avenot et al. 2005). Additionally, in DCF-resistant A. alternata isolates from passionfruit, frameshift mutations caused by a 4-bp and an 11-bp insertion at HAMP5 and HAMP1 of AaHK1, respectively, confer resistance to DCFs (Dry et al. 2004). The diverse mutations are also found in homologous HK1s of other Alternaria species. The iprodioneresistant A. brassicicola isolates possess an amino acid substitution (E to K) at the HisKA domain, a frameshift caused by a 2-bp deletion at HAMP4, or nonsense  In yeast and B. cinerea, cellular responses to osmotic stress are controlled by the HOG-MAPK signaling pathway through which a range of osmotic responses including glycerol synthesis are regulated (Brewster et al. 1993;Cui et al. 2002). In this study, when treated with 10 μg/ mL procymidone, MIGAs and the expression level of AaHK1 were significantly increased in the Pro S isolates compared with those under procymidone-free conditions, while this increase in the ten Pro HR isolates was at a relatively low level, the same as that observed in fludioxonil-resistant B. cinerea isolates (Ren et al. 2016).
The results indicate that AaHK1 plays a role in the HOG-MAPK pathway, and a negative correlation exists between the resistance level of A. alternata to procymidone and the increase of MIGAs under procymidonestressed condition. It was reported that the high level iprodione-or fludioxonil-resistant isolates of A. alternata from passionfruit exhibited moderate sensitivity to osmotic stress, identical to that of A. brassicicola (Dry et al. 2004;Avenot et al. 2005). Nevertheless, in this study, all the Pro HR isolates from broccoli and cabbage were highly sensitive to osmotic stresses. The difference in sensitivity to osmotic stress agents in Alternaria spp. suggest that mutations in HK1s might affect signal transduction of the HOG-MAPK signaling pathway, consequently resulting in distinct responses to stress agents.
For other regulatory functions of HK1s in the development of DCF-resistance, Steel (Steel 1996) reported that the resistance to DCFs is positively correlated with the activity of an anti-oxidant enzyme, catalase. Compared with Pro S isolates, the sensitivity of Pro HR isolates (with mutations in AaHK1) to oxidative stress (H 2 O 2 ) was reduced significantly, whereas, no significant difference in sensitivity to H 2 O 2 was observed in B102, a Pro HR isolate having no mutations in AaHK1. This suggests that mutations leading to inactivation of HK1 may result in constitutive induction of catalase, as observed in DCFresistant B. cinerea and A. alternata isolates from passionfruit (Steel & Nair 1995;Steel 1996). In contrast, the sensitivity to oxidative stress in HK mutants of A. longipes from tobacco was not affected (Luo et al. 2012). These results indicate that HKs from A. alternata and A. longipes have a different regulatory function in signal transduction during oxidative stress. In addition, all Pro HR isolates were more resistant to cell membrane agent (0.05% SDS) than those of Pro S isolates, indicating that the group III HK of A. alternata is probably associated with membrane permeability. The enhanced tolerance to H 2 O 2 and SDS in Pro HR isolates is suggested to protect their cells from the deleterious effects of DCFs. Fig. 6 The expression of AaHK1 gene in A. alternata Pro S and Pro HR isolates. The mycelia of each isolate were treated with procymidone at 0 or 10 μg/mL after growth in PDB for 12 h. Bars denote the standard errors of three repeated experiments. The asterisk indicates significant difference between procymidone-treated sample and control. The Pro S isolates include C10, C28, B16 and B21. The Pro HR isolates include C5, C13, C40, B30, B35, B46, B53, B88, B102 and B108 Fitness is an extremely essential parameter for evaluation of the potential risk of fungicide-resistant populations. Many DCF-resistant isolates display a fitness penalty relative to DCF-sensitive isolates, as those found in S. sclerotiorum (Kuang et al. 2011), B. cinerea (Ren et al. 2016) and Penicillium expansum (Li & Xiao 2008). Although most Pro HR isolates of A. alternata in current study was also reduced in mycelial growth rate, sporulation capacity and virulence, the Pro HR isolates B46, B35 and B53 were increased significantly in sporulation capacity or pathogenicity compared with Pro S isolates. This suggests that Pro HR populations of A. alternata from commercial fields of broccoli and cabbage as a whole possessed no superiority over Pro S populations (wild type), and was in a stage of slow development in these fields during 2018-2019. However, the Pro HR isolates with high fitness may have the potential to develop into a large group if no reasonable control methods are adopted. Therefore, the growers should be encouraged to use mixed fungicides with different modes of action (e.g., QoIs, DMIs or SDHIs) to delay the development of DCF-resistant A. alternata populations in these fields.

Conclusions
The resistance to procymidone has developed in natural populations of A. alternata infecting broccoli and cabbage crops in Jiangsu Province of China. Most Pro HR isolates displayed a fitness penalty in mycelial growth rate, sporulation and virulence compared with Pro S isolates. The Pro HR isolates were highly sensitive to osmotic stress. In addition, a positive cross-resistance was observed between procymidone and fludioxonil or iprodione. Mutations in AaHK1 were involved in the resistance of A. alternata isolates to DCFs. No significant differences in biochemical characteristics was observed between Pro HR isolates from broccoli and cabbage.

Sample collection, causal agent isolation and identification
All the 198 isolates of A. alternata were isolated from symptomatic leaves during 2018-2019. Diseased leaves were collected from different commercial fields of broccoli and cabbage at Yancheng City, Jiangsu Province, China, where procymidone (a member of DCFs), alone or mixed with other fungicides with different mode of action, had been extensively used to control ALSB twice or three times in a growing stage since 1980s. Any two neighboring samples were separated from each other at least 50 m, with each sample placed in an individual envelope. To obtain A. alternata isolates, small tissue pieces cut from lesion edges were sterilized in 1% NaClO for 3 min, washed three times with sterile distilled water, and transferred onto Petri dishes containing PDA medium supplemented with 100 μg/mL streptomycin sulfate (Solarbio Science & Technology Co., Ltd.) (Ren et al. 2016). After incubation at 25°C for 6 days, a mycelial plug was cut from the margin of an actively growing colony, and transferred to the center of a fresh PDA plate. Conidia were harvested using sterile distilled water after 6 days of incubation at 25°C, and the single-spore isolate was obtained by picking a single conidium under a microscope (Zhang et al. 2017). In total, 198 singlespore isolates were obtained. The representative isolates were further identified by polymerase chain reactions (PCR) with two primer pairs ITS1/ITS4 and KOG1058F/ KOG1058R (Table 6) (Woudenberg et al. 2015). The ITS and KOG1058 sequences of each isolate were blasted and congeneric sequences were obtained from the National Center for Biotechnology Information (NCBI) GenBank database. Phylogram analysis was processed by MEGA7 software based on ITS and KOG1058 sequences to further identify the isolates.

Fungicides
All fungicides used are technical grade. Fludioxonil and iprodione (98% active ingredient; generously provided by Yangzhou Younuo Chemicals Co., Ltd., Yangzhou, China) were dissolved in methanol to obtain a stock solution of 2 × 10 4 mg/L. Procymidone (99.2% active ingredient; provided by Jiangsu Xinyi Chemical Co., Ltd., Xuzhou, China) was dissolved in acetone to obtain a stock solution of 2 × 10 4 mg/L. All stock solutions were kept at 4°C before use.

Determination of resistance frequencies of A. alternata isolates to procymidone
Sensitivities of 198 single-conidium isolates of A. alternata to procymidone were determined by discriminatory doses of mycelial growth inhibition method. In brief, procymidone at 0, 10, 50 and 100 μg/mL were used to determine sensitivities of all isolates for low-resistance (LR), moderate-resistance (MR) and high-resistance (HR). The isolate was considered procymidone-sensitive (Pro S ) when the MIC (minimum inhibitory concentration) value was less than 10 μg/mL. Similarly, 10 < MIC< 50 μg/mL for low-level procymidone-resistant (Pro LR ) isolates, 50 < MIC< 100 μg/mL for moderate-level procymidone-resistant (Pro MR ) isolates, and MIC> 100 μg/mL for high-level procymidone-resistant (Pro HR ) isolates. Mycelial plugs (5 mm in diameter) were cut from the margin of a 6-day-old colony, and transferred to the center of Petri dishes containing PDA medium (one plug per dish) amended with procymidone at the above-mentioned concentrations. At 6 days of incubation, the sensitivity of each isolate to procymidone was determined based on the mycelial colony growth inhibited by procymidone. The resistance frequency was determined using the following formula: resistance frequency (%) = (the number of resistant isolates/the number of the total isolates tested) × 100%. The experiment was performed three times with three replications for each isolate.

Determination of sensitivities of A. alternata isolates to procymidone
To determine the sensitivity of A. alternata isolates to procymidone, the fungicide was prepared in a serial concentrations of 0, 25, 50, 100, 200, 400 and 800 μg/mL for Pro HR isolates, and another serial concentrations of 0, 0.03, 0.1, 0.3, 1, 3 and 9 μg/mL for Pro S isolates. The 6day-old mycelial plugs (5 mm in diameter) were separately transferred to the center of PDA plates with abovementioned concentrations of procymidone. Percentage of inhibition was evaluated and EC 50 (effective concentration for 50% mycelial inhibitive growth) was calculated for each tested isolate using Data Processing System (DPS 7.05). The experiment was performed three times with three replications for each isolate.

Cross-resistance test among different fungicides
The sensitivity of A. alternata isolates to fludioxonil and iprodione were determined by measuring mycelial growth rate. One serial concentrations (0,25,50,100,200,400 and 800 μg/mL) were set up to test the sensitivity of resistant isolates to fludioxonil or iprodione, and another serial concentrations (0, 0.025, 0.05, 0.1, 0.2, 0.4, 0.8 and 1.6 μg/mL) for sensitive isolates. EC 50 values were determined with the same method as abovementioned. Spearman's rank correlation coefficient for the log 10 -transformed EC 50 values was used to analyze cross-resistance pattern between procymidone and fludioxonil or iprodione. The experiment was performed three times with four replicates per concentration.
Determination of radial growth rates and sporulation capacities of A. alternata isolates Six-day-old mycelial plugs (5 mm in diameter) of A. alternata isolates were transferred to the center of Petri dishes containing PDA medium (one plug per dish), then placed in an incubator at 25°C in darkness. At 6 days of incubation, the average colony diameter for each isolate was measured in two perpendicular directions. The experiment was performed three times with three replications for each isolate.
For sporulation test, conidia were harvested from the surface of six-day-old PDA colonies washed with 10 mL sterile water (containing 0.1% Tween 20), filtered through four layers of sterile lens papers, and centrifuged at 5000 rpm for 1 min. The pellets were re-suspended in 1 mL sterile distilled water. The concentrations of conidial suspensions were measured with a hemacytometer under microscope. Six replicates were performed for each isolate, and the experiment was repeated three times.

Virulence assay
Broccoli (var. Hanxiu) leaves were used to determine the pathogenicity of A. alternata isolates. Leaves were surface-disinfected with 75% alcohol for 10 s, and inoculated with six-day-old hyphal plugs (5 mm in diameter) of each isolate. Before inoculation, a sterilized needle was used to produce wound. All the inoculated leaves were transferred onto three layers of moistened filter

Determination of intracellular glycerol accumulations in A. alternata isolates
A commercial assay kit (Applygen Technologies Inc., Beijing) was used to determine intracellular glycerol accumulations in mycelia of A. alternata isolates in accordance with the manufacturer's instruction. In brief, all isolates were incubated in potato dextrose broth (PDB) in a shaker (175 rpm) at 25°C for 2 days. At 5 h after treatment with 10 μg/mL procymidone, mycelia of each isolate were collected with sterile distilled water, lyophilized and grounded in mortars amended with liquid nitrogen (Ren et al. 2016). Mycelial powder (0.1 g) was then transferred into a 2 mL centrifuge tube, and mixed in 1 mL glycerol extraction buffer (Applygen Technologies Inc., Beijing, China). The mixture was vortexed for 5 min and centrifuged at 5000 rpm for 20 min, and the supernatants were used to analyze mycelial intracellular glycerol concentrations (Duan et al. 2013). This experiment was repeated three times with four replicates for each isolate.

DNA extraction and sequence analysis of AaHK1
Genomic DNAs were extracted from mycelia of A. alternata isolates with a commonly used cetyltrimethylammonium bromide (CTAB) method. Based on the sequence of AaHK1 (GenBank accession No. GQ414508.1) deposited in NCBI GenBank database, five primer pairs (Table 6) were designed to amplify the complete nucleotide sequence of AaHK1. Polymerase chain reactions (PCR) were conducted in a 25 μL volume, containing 0.25 μL of high-fidelity LA polymerase (5 U/mL) (TaKaRa, Japan), 1 μL of each primer (10 mM), 2.5 μL of 10× LA polymerase buffer (Mg 2+ free), 50 ng of genomic DNA, 2.5 μL of MgCl 2 (25 mM), 4 μL of dNTP mixture (2.5 mM of each dNTP) and 12.75 μL of ddH 2 O. The PCR program consisted of an initial denaturation at 94°C for 5 min; followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 56°C for 30 s, and extension at 72°C for 50s; and a terminal extension at 72°C for 10 min. Each PCR product was purified using the QIAquick PCR Purification Kit (Qiagen, MD, USA), and sequenced by the Sangon Biotech Co., Ltd. (Shanghai, China). The amino acid sequences were analyzed by DNAMAN software (version 6.0; Lynnon Biosoft Bioinformatic Solutions). PCR amplifications were performed three times for each isolate to avoid sequence mismatches during PCR amplification and sequencing.

Determination of the gene expression level of AaHK1
To analyze the relative gene expression level of AaHK1 in A. alternata isolates, mycelial plugs were incubated in flasks containing PDB, with six replicate flasks for each isolate. After 2 days of incubation at 25°C in a shaker, three flasks were treated with 10 μg/mL of procymidone, and the other three were added with the same volume of sterile water and used as controls. After 12 h, the mycelia of each isolate were harvested for RNA extraction using Total RNA Isolation kit (Promega, USA). Further, cDNA was synthesized using PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Japan). Primers used for quantitative PCR (qPCR) are shown in Table 6. qPCR was performed in an ABI 7500 Real-Time Detection System (Applied Biosystems) using SYBR Green I fluorescent dye detection. Amplification was performed in a 20-μL volume containing 10 μL of iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories), 1 μL of each of the forward and reverse primers (10 μM), 0.4 μL of 50 × ROX Reference Dye 2, 1 μL of reverse transcription product, and 6.6 μL of ddH 2 O. The expression of AaHK1 was normalized to that of β-tubulin gene. The transcription level of AaHK1 in each isolate relative to that of β-tubulin gene was calculated using the 2 -ΔΔCt method (Wei et al. 2019). The experiment was repeated three times.

Statistical analysis
The SIGMA-STAT Statistical Software Package (SPSS Science, version 11) was used to analyze the data. Fisher's protected least significant difference (P = 0.05) test was calculated to evaluate statistical significance.