Transcriptome Sequencing and Expression Analysis of Genes Related to Anthocyanin Biosynthesis in Leaves of Malus ‘Profusion’ Infected by Japanese Apple Rust

Abstract

Anthocyanins play many roles in plants, including providing protection from biotic and abiotic stresses. Japanese apple rust (Gymnosporangium yamadae Miyabe ex G. Yamada) causes serious diseases in plants of the genus Malus and results in reduced fruit production and quality. However, few studies have been done to unravel the molecular mechanisms of anthocyanin formation in rust-infected apple leaves. To identify new regulatory genes in apple leaves that may be involved in regulating rust-induced anthocyanin biosynthesis, we measured anthocyanin content and sequenced the transcriptomes of rust-infected and uninfected tissues of Malus ‘Profusion’ leaves. Significant color changes and anthocyanin enrichment (especially cyanidin-3-galactoside chloride) occurred in infected tissues, whereas no significant color change and a low anthocyanin level were observed in uninfected tissue. We identified 10,045 differentially expressed genes (DEGs) in these two tissue types, including 6021 genes that were upregulated in the infected tissue and 4024 genes that were downregulated. We also identified five structural genes that are putative regulators of anthocyanin biosynthesis. In addition, 56 MYB genes, 36 bHLH genes, and one WD40 gene were identified among the obtained DEGs. According to the phylogeny of the amino acid sequences of transcription factors known to be involved in anthocyanin biosynthesis, one MYB gene (MYB114-like) and two bHLH genes (bHLH33 and bHLHA-like) may relate to anthocyanin biosynthesis in rust-infected apple leaves. These data will provide insights into the molecular mechanisms underlying anthocyanin accumulation upon rust infection.

1. Introduction

Malus ‘Profusion’, an ornamental, deciduous crabapple tree or shrub in the Rosaceae family, is widely used in landscaping due to its beautiful colors and attractive shape. Like other members of the genus Malus, Malus ‘Profusion’ is vulnerable to Japanese apple rust [1]. The rust fungus, Gymnosporangium yamadae (Miyabe ex G. Yamada, 1904) infects apple trees and shrubs, causing serious diseases [2]. The G. yamadae fungal parasite afflicts coniferous trees, apple trees, and other species in the genus Malus [3]. In an earlier field experiment, we found obvious red patches on rust-infected tissues present on the leaves of Malus ‘Profusion’. In addition, our preliminary physiological analysis revealed that the major red pigments in the red patches of rust-infected leaves of Malus ‘Profusion’ were anthocyanins. It is worth noting that Lu et al. also found that the content of flavonoids, especially anthocyanins, was significantly increased in the infected symptomatic tissue of apple leaves, and it was speculated that they play an important role in rust resistance in Malus [4].

The anthocyanin biosynthetic pathway has been well studied in plants. There are many enzymes that can catalyze anthocyanin synthesis, including chalcone synthetase (CHS), dihydroflavonol 4-reductase (DFR), and anthocyanin synthase (ANS) [5]. However, genes encoding these key enzymes (structural genes) are usually regulated by transcription factors, such as MYB, bHLH, and WD40, which are upstream regulators of anthocyanin metabolism [6]. These three transcription factors form the MYB-bHLH-WD40 transcription complex and regulate the biosynthesis of anthocyanin in plants [7]. Related studies indicate that MdMYB1, MdMYB10, and MdMYBA in the MYB transcription factor family play a regulatory role in the anthocyanin biosynthetic pathway in apple fruit (Malus domestica Borkh.) [8]. Studies using transgenic plants also showed that the expression of these genes plays a key role in anthocyanin accumulation and fruit coloration, and they cooperate with the MdbHLH3 and WD40 proteins to interact with promoters of the anthocyanin biosynthetic gene such as MdANS [9,10,11].

Anthocyanins are a water-soluble natural pigment belonging to the flavonoid family of organic molecules. They are also the main red pigments found in the Rosaceae family. Moreover, anthocyanins are associated with pathogen resistance. Laima et al. found that anthocyanin extracts from European Viburnum (cranberry) species had strong inhibitory effects on many common pathogenic bacteria [12]. Zhao et al. suggested that the antibacterial activity of purple corn extract is related to its anthocyanin content [13]. Another study showed that the differential regulation of the anthocyanin biosynthesis pathway resulted in the enrichment of anthocyanins in apple leaves, thereby improving resistance to scab and fire blight [14,15]. These results point to anthocyanins playing a role in defending against fungal pathogens in infected tissues.

Transcriptomic analysis is commonly used to identify differentially expressed genes (DEGs) and to quantify their relative expression in different tissues, including infected and uninfected leaf tissues [5]. In this study, by comparing the contents of anthocyanin-related metabolites and the expression levels of anthocyanin structural genes and associated transcription factors in rust-infected and uninfected leaf tissues of Malus ‘Profusion’ plants, we identified putative anthocyanin biosynthetic genes that participated in the response to rust infection in Malus. These results will guide the improvement of rust control and facilitate the selection of new rust-resistant Malus cultivars.

2. Materials and Methods

2.1. Plant Materials

Eight-year-old Malus ‘Profusion’ plants were used as the study material. The plants were grown in the crabapple germplasm nursery of Northwest A&F University, Yangling, China, maintained by standard horticultural management and infected with Japanese apple rust (Gymnosporangium yamadae Miyabe ex G. Yamada) under natural conditions. On 8 June 2018, three trees (n = 3; considered as biological replicates) separated by a distance of three meters, were marked and from each tree, three visibly healthy and three rust-infected leaves were collected (total n of leaves per tree = 6; total nr of leaves harvested for analyses = 18). The healthy and infected leaves from each tree were respectively pooled to composite samples for the analyses (n = 3). The harvested leaf samples were washed three times with distilled water. Red infected tissues (RIT) were collected from the red area of infected leaves by sterile blades, and uninfected leaf tissues (UIT) were collected from healthy leaves (Figure 1A). The samples were then immediately frozen in liquid nitrogen and stored at −80 °C for RNA extraction and the analyses of total anthocyanin and monomer phenol content.

2.2. Total Anthocyanin Content

First, 0.1 g lyophilized leaf sample was weighed and soaked in 10 mL extraction solution (97:3 methanol: hydrochloric acid). The samples were then incubated at 4 °C for 48 h in the dark and extracted using ultrasound for 1 h. Next, we centrifuged the mixture at 6000 rpm for three minutes and collected the supernatant. Total anthocyanin content was determined by measuring absorbance on a UV-3802 spectrophotometer (Unico, Dayton, OH, USA) and applying the following formula: A530 − 0.25 × A657. Cyanidin chloride (>95% HPLC, Sigma, St. Louis, MO, USA) was used to generate the calibration curves. Finally, total anthocyanin content was normalized to the dry weight of each sample.

2.3. Anthocyanin Biosynthesis-Related Compounds Content

First, we weighed out 0.4 g of the lyophilized leaves, soaked them in 8 mL of methanol at 4 °C for 48 h in the dark, and stirred them using ultrasonic waves for 1 h. Next, we centrifuged the mixture at 6000 rpm for three minutes and extracted the supernatant, which was then set aside. The substances within the samples were separated using a Shimadzu LC-20AT high-performance liquid chromatograph (HPLC) and a Lachrom-C18 column (4.6 mm × 250 mm, 5 μm, Hitachi, Tokyo, Japan). The proposed anthocyanin biosynthesis-related compounds peak was isolated by several analytical HPLC runs. The structure of the unknown substance was determined using LC–MS by comparing it with a standard as described by Li et al. [16]. Mobile phase A contained 0.04% aqueous formic acid and phase B contained chromatographic grade acetonitrile. The gradient elution was as follows: 0–40 min, 95% A; 40–45 min, 60% A; 45–60 min, 0% A. The column temperature was maintained at 40 °C, the injection volume used was 10 μL, the flow velocity was 0.5 mL/min, and the detection wavelengths used were 280, 360, and 520 nm. All samples were filtered using a 0.22 μm filter before the analysis. Three technical replicates of each sample were analyzed. The type and concentration of the anthocyanin biosynthesis-related compounds in each extract were determined, and the content of each compound was calculated using the peak area normalization method.

2.4. RNA Extraction, Library Construction, and De Novo Transcriptome Assembly

Total RNA from both the RIT and UIT samples was extracted using the RNA prep polysaccharide polyphenol plant total RNA extraction kit (Tiangen Biotech, Beijing, China). The purity, density, and integrity of RNA samples was determined using a Nanodrop microvolume spectrophotometer, a Qubit 2.0 fluorometer, and an Agilent 2100 bioanalyzer to ensure that RNA samples were of sufficient quality for transcriptome sequencing. A library was assembled with high-quality RNA, and qualified libraries were sequenced using Sequencing by Synthesis (SBS) on an Illumina HiSeqX sequencing platform to obtain raw sequence data. Library construction and sequencing was performed by Beijing Baimax Technology Co., Ltd., Beijing, China.

The original sequencing data were cleaned by removing the adapter sequences and low-quality reads. The resulting clean reads were assembled into a transcriptome using the apple (Malus domestica) genome published by Nicolas et al. as the reference genome [17]. The reads were used for comparison and assembly using the HISAT2 [18] and StringTie [19] systems. The expression levels of each unigene were normalized to fragments per kilobase of transcript per million mapped reads (FPKM) [20]. The Pearson correlation coefficient r was used to evaluate correlations among replicates and among samples, and the rationality of sample selection was tested by examining the coefficient of determination (r²).

2.5. Gene Identification, Gene Annotation, and Analysis of DEGs

Analysis of the differentially expressed genes was performed as per the DeSeq method of Wei et al. [21]. DEGs were identified using the following threshold criteria: corrected p-value (equivalent of false discovery rate; FDR) <0.05 (using Benjamini–Hochberg correction methods) and a fold change >2 or <−2. Functional annotation and classification was conducted by BLASTing (BLAST: Basic Local Alignment Search Tool) the gene sequences against seven public gene and protein databases including the nonredundant protein (Nr) database, the Swiss-Prot protein (SwissProt) database, the Gene Ontology (GO) database, the Eukaryotic Orthologous Groups of proteins (KOG) database, the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, the Protein family (Pfam) database, and the Cluster of Orthologous Groups of proteins database (COG). GO hierarchical clustering was performed with the topGO R software package [22]. The enrichment of a specific GO term was considered significant when the D (distance) value ≤0.01 using the Kolmogorov–Smirnov test.

2.6. Phylogenetic Analysis and Sequence Alignment

The protein sequences of MYB, bHLH, and WD40 were obtained from the NCBI (National Center for Biotechnology Information) non-redundant protein (Nr) database. Phylogenetic analysis was conducted using MEGA version 6 using the neighbor-joining method with 1000 bootstrap replicates. Sequence alignment was performed in DNAMAN Version 6.

2.7. Quantitative Real Time PCR (qRT-PCR) Analysis

Total RNA was extracted from the RIT and UIT samples, respectively. We used RNase-free DNaseI (Tiangen, Beijing, China) as per the manufacturer’s protocol to remove genomic DNA. Next, we used a PrimeScript RT reagent kit (Takara, Otsu, Japan) to reverse-transcribe 1 μg of total RNA. Primers for qRT-PCR were designed and are listed in Table S1. 18S rRNA (DQ341382) was used as the internal control and UIT was used as the contrast group for normalization processing. In addition, the relative expression level of each gene of UIT was set to 1. The relative expression level of DEGs in the samples was determined using 2× SYBR real-time PCR mixture kit (BioTeKe, Beijing, China) on the StepOnePlus real-time PCR system (Applied Biosystems, Waltham, MA, USA). The reaction procedure was a three-step process as follows: The first step was 94 °C for 2 min, the second step was 40 cycles of 94 °C for 15 s and 60 °C for 15 s, and the third step was 95 °C for 15 s, 60 °C for 1 min, and 95 °C for 15 s. The 2^−∆∆Ct method was used to calculate the relative expression levels of each gene [23].

2.8. Data Analysis

All experimental data were analyzed using paired t-tests, and p < 0.05 was defined as the threshold of statistical significance. Statistical tests were performed using IBM SPSS Statistics version 22 (IBM SPSS, Chicago, IL, USA), while graphical analysis and plotting were performed using GraphPad Prism Version 7 (GraphPad Software, San Diego, CA, USA).

3. Results

3.1. Analysis of Total Anthocyanin Content and Composition

The rust-infected red leaf tissues (RIT) were isolated from rust-infected diseased tissues (RDT) and uninfected leaf tissue (UIT) (Figure 1A). UIT was used as the control for all anthocyanin analyses. The total anthocyanin content of RIT was 18.50 times greater than that in UIT (Figure 1B). Moreover, HPLC data revealed that the cyanidin-3-galactoside chloride content in RIT was 15.80-fold greater than that in UIT, which was similar to the magnitude of the difference in total anthocyanin between infected and uninfected tissues. The most abundant anthocyanin in Malus ‘Profusion’ leaves was cyanidin-3-galactoside chloride (Figure 1C). In addition, naringenin, dihydroquercetin, catechin, and epicatechin, which are the main byproducts of the anthocyanin biosynthesis pathway, were more abundant in RIT than in UIT (by a factor of 1.57, 2.38, 2.59, and 4.89 times, respectively;

Table 1. Type and content of anthocyanin biosynthesis-related compounds in UIT and RIT samples.

Compound	$\mathbf{UIT}/\mathbf{mg}\mathbf{·}{\mathbf{kg}}^{\mathbf{-1}}$	$\mathbf{RIT}/\mathbf{mg}\mathbf{·}{\mathbf{kg}}^{\mathbf{-1}}$
Catechin	72.63 $\pm$ 1.00 b	188.45 $\pm$ 0.45 a
Chlorogenic acid	513.59 $\pm$ 12.56 b	1947.46 $\pm$ 142.48 a
Coumaric acid	107.24 $\pm$ 2.94 b	171.26 $\pm$ 4.82 a
Cyanidin-3-galactoside chloride	44.71 ± 0.48 b	706.59 ± 7.52 a
Dihydromyricetin	281.39 $\pm$ 7.62 b	384.08 $\pm$ 16.89 a
Dihydroquercetin	1485.13 $\pm$ 72.57 b	3525.75 $\pm$ 61.70 a
Epicatechin	54.89 $\pm$ 2.45 b	268.37 $\pm$ 0.41 a
Eriodictyol	271.32 $\pm$ 7.08 a	278.33 $\pm$ 8.13 a
Naringenin	1222.09 $\pm$ 31.85 b	1915.27 $\pm$ 100.36 a
Phloretin	10,404.03 $\pm$ 116.34 b	28,553.27 $\pm$ 576.97 a
Proanthocyanin B1	92.67 $\pm$ 2.47 b	123.94 ± 3.19 a
Proanthocyanin B2	4.56 $\pm$ 0.15 b	38.50 $\pm$ 0.53 a

Data are the mean values ± standard deviation (SD) of three biological replicates. The different letters represent significant differences between UIT and RIT, as judged by the paired t-test (p < 0.05).

). The levels of these metabolites were significantly different between the UIT and RIT, as revealed by the paired t-test (p < 0.05).

3.2. Library Construction and Transcriptome Sequencing

We constructed six UIT and RIT transcriptome libraries based on three biological replicates for each tissue type. We obtained 49.14 Gb high-quality data after removing adapter sequences and low-quality reads. Bases with a Q score of 30 or higher (i.e., base call accuracy ≥ 99.9%) accounted for more than 93% of all the bases. The GC content (the ratio of Guanine and Cytosine content to the total nucleobase) ranged from 46%–50%, and the average contrast ratio was 70% (

Table 2. Descriptive statistics of sequencing data.

Sample Name	Clean Reads	Clean Bases	Q30	GC Content
RIT1	28,534,074	8,533,976,408	94.10%	50.04%
RIT2	22,858,851	6,835,393,636	93.71%	50.01%
RIT3	27,789,895	8,305,725,370	93.40%	49.84%
UIT1	32,429,472	9,701,511,274	93.58%	46.84%
UIT2	27,360,624	8,189,892,014	93.46%	47.18%
UIT3	25,312,509	7,573,777,970	93.97%	46.79%

). These results confirmed the suitability of the obtained transcriptomic data for further analyses.

Pearson correlation coefficients (Pearson’s r values) were used to determine the correlation in gene expression level (log10 (FPKM+1)) between tissue samples and among biological replicates for each tissue. The closer the r² values were to 1, the stronger the correlation between the two samples. As shown in Figure 2, the r² values between samples of the RIT and UIT tissues were less than 0.4, while the correlations between different biological replicates of the same tissue were over 0.99. These results suggested that the samples have good repeatability.

3.3. Annotation and Identification of DEGs

We considered genes that differed in expression between UIT and RIT samples at corrected p < 0.05 (using Benjamini–Hochberg correction methods) and had a minimum difference in expression of >(2)-fold or <(−2)-fold to be differentially expressed genes (DEGs). To show the overall distribution of the DEGs, we performed hierarchical cluster analysis and constructed volcano and minus average (MA) maps (Figure 3). After screening, we identified 10,045 DEGs. These included 6021 genes that were upregulated in RIT relative to UIT and 4024 genes that were downregulated. Next, a total of 9479 unigenes were annotated based on BLAST searches against seven public databases (

Table 3. Annotation of non-redundant unigenes using queries of public databases.

DEG Set	Total 1	COG	GO	KEGG	KOG	NR	Pfam	Swiss-Prot
RIT vs. UIT	9749	4242	3950	3609	4978	9628	8204	7818
Genes 2
Annotated (%)	100.0	43.5	40.5	37.0	51.1	98.8	84.2	80.2

¹ Total represents the total number of annotated unigenes. ² The percentages of annotated genes indicate that different types of data account for the total number of annotated unigenes.

). Only 296 of the 10,045 DEGs were not annotated, and therefore the total annotation rate was 97.05%.

3.4. Analysis of DEGs

3.4.1. GO Enrichment Analysis of the DEGs

We performed a GO enrichment analysis of all DEGs. GO classifications are divided into three major categories: biological processes, cellular components, and molecular functions. These categories are subdivided further. In our research, 8374 DEGs were assigned to the ‘biological processes’ category; 4341 DEGs were assigned to the ‘cellular components’ category; and 6141 DEGs were assigned to the ‘molecular functions’ category. These DEGs were further classified into 52 functional subcategories (Figure 4). Next, in order to identify significantly enriched GO terms for DEGs, we conducted a hierarchical clustering analysis on the DEGs and listed the ten most significantly enriched GO terms in

Table 4. The most prevalent GO enrichment results among the differentially expressed genes.

GO_ID	Term	Annotated	Significant 1	D Value 2
GO: 0015979	photosynthesis	242	152	3.60 × 10−20
GO: 0055114	oxidation-reduction process	1726	603	6.90 × 10−13
GO: 0009765	photosynthesis, light harvesting	35	26	7.50 × 10−11
GO: 0006098	pentose phosphate shunt	103	52	2.60 × 10−10
GO: 0010027	thylakoid membrane organization	97	51	6.40 × 10−10
GO: 0019252	starch biosynthetic process	77	41	1.90 × 10−8
GO: 0000023	maltose metabolic process	62	35	1.90 × 10−8
GO: 0010207	photosystem II assembly	65	34	2.10 × 10−7
GO: 0035304	regulation of protein dephosphorylation	35	23	7.00 × 10−7

¹ The number of significant DEGs under each GO term; ² D value represents the conspicuousness statistic of the enriched term. The smaller the D value, the more significant the enrichment.

. In total, 142 GO terms were significantly enriched with a D value ≤0.01 (Table S3). Photosynthesis (GO_ID: 0015979), oxidation-reduction processes (GO_ID: 0055114), and photosynthesis and light harvesting (GO_ID: 0009765) were the three most significant GO enrichment terms associated with the DEGs (Table 4). Each of these processes may be involved in anthocyanin biosynthesis.

3.4.2. KEGG Enrichment Analysis of the DEGs

We mapped the identified DEGs to 124 pathways in the KEGG database. The 20 pathways that showed the most significant enrichment are shown in Figure 5. Of these 20 terms, seven were related to metabolism, seven were related to biosynthesis, three were related to photosynthesis, and the remaining three were related to the glycolysis/gluconeogenesis, ribosome, and pentose phosphate pathway, respectively. Of the 20 most significantly enriched pathways, the three most enriched pathways were photosynthesis-antenna proteins, flavonoid biosynthesis, and photosynthesis. The flavonoid synthesis pathway includes many structural genes involved in anthocyanin biosynthesis. We then used KEGG annotation to analyze the function of these DEGs and identified seven structural genes that may participate in anthocyanin biosynthesis in response to rust infection (

Table 5. Putative structural genes involved in anthocyanin biosynthesis identified from transcriptome data.

GeneID	log2 Ratio (RIT/UIT)	FDR	Annotation
MD13G1285100	3.10	0	chalcone synthase [Malus domestica Borkh.]
MD04G1003300	4.16	0	chalcone synthase [Sorbus aucuparia Linn.]
MD01G1118100	2.10	2.91 × 10−42	chalcone isomerase [Pyrus pyrifolia (Burm. f.) Nakai.]
MD15G1353800	2.52	0	flavanone 3-hydroxylase [Prunus avium Linn.]
MD15G1024100	3.00	0	dihydroflavonol 4-reductase [Malus domestica Borkh.]
MD03G1001100	3.12	0	leucoanthocyanidin dioxygenase [Malus domestica Borkh.]
MD06G1071600	2.69	0	leucoanthocyanidin dioxygenase [Malus domestica Borkh.]

). Of these, MD04G1003300 and MD13G1285100 were identified as chalcone synthases (CHS); MD06G1071600 and MD03G1001100 were identified as leucoanthocyanidin dioxygenases (LDOX/ANS); and MD01G1118100, MD15G1353800, and MD15G1024100 were identified as chalcone isomerase (CHI), flavanone-3-hydroxylase (F3H), and dihydroflavonol 4-reductase (DFR), respectively. This gene set encodes most of the key enzymes involved in the anthocyanin biosynthesis pathway [24].

3.4.3. Analysis of Transcription Factors in Identified DEGs

Anthocyanin synthesis is known to involve many transcription factors (TFs), including those of the MYB, bHLH, and WD40 families. Some members of these families can form a ternary complex (MBW) that can regulate the expression of anthocyanin synthase [7]. In this study, we found 56 MYB genes, 36 bHLH genes, and one WD40 gene among the identified DEGs (Table S2).

To screen for MYB genes associated with anthocyanin biosynthesis, we conducted a phylogenetic analysis of the deduced amino acid sequences of the 56 MYB genes from the DEGs and other MYB genes that are known to participate in anthocyanin biosynthesis in plants. This analysis was performed according to the protocol published by Luo et al. [5]. One gene, MYB114-like (MD17G1261000), showed a relatively higher degree of homology with other MYBs involved in anthocyanin biosynthesis, including MCMYB10, MrMYB1, and NtAN2 (Figure 6), which have been proven to promote anthocyanin biosynthesis [25,26,27]. The amino acid sequence of MD17G1261000 was analyzed and compared to other MYB genes associated with anthocyanin biosynthesis. A highly conserved R2R3 sequence was found at the N-terminus of MD17G1261000 (Figure 7), which indicated that MD17G1261000 is a R2R3MYB transcription factor. To date, studies in diverse species have agreed that anthocyanin biosynthesis is regulated by a complex that includes R2R3MYB [28]. Meanwhile, another highly conserved [D/E]L$\times$₂[R/K]$\times$₃L$\times$₆L$\times$₃R sequence has been found in the R3 domain of MD17G1261000 (Figure 7). This conserved sequence was found to bind to an R-like bHLH protein [29]. In addition, a conserved “ANDV” sequence was also found in the R3 domain of MD17G1261000 (Figure 7). This is a signature MYB-associated motif known to promote anthocyanin biosynthesis [30]. Finally, we found a short conserved motif (motif 6) in the C-terminal of MD17G1261000 (Figure 7), which was previously identified as a marker of positive regulation of anthocyanin biosynthesis in vegetative cells and/or reproductive tissues [30]. Taken together, these data indicate that MD17G1261000 may be a R2R3MYB transcription factor that recruits the R-like bHLH proteins, thereby activating anthocyanin biosynthesis.

Next, we performed a similar phylogenetic analysis on the identified 36 bHLH genes (from the DEGs) and other bHLHs genes that are known to be involved in anthocyanin biosynthesis in plants. Two DEGs, bHLHA-like (MD03G126300) and bHLH33 (MD07G1137500), from this study were found to be in the same clade as the bHLH genes known to regulate anthocyanin biosynthesis (Figure 8). Genes clustered into the same clade as MD03G126300, including MdbHLH3, MdbHLH33, and NtAn1a, were proven to be bHLH genes involved in anthocyanin biosynthesis [31,32]. The deduced amino acid sequence alignment showed that MD03G126300 and MD07G1137500 contained box 11, 18, and 13 at the N-terminal domain (Figure 9), all of which are involved in the interaction between MYB and bHLH [33]. These data show that MD03G126300 and MD07G1137500 probably interact with the MYBs. Among all DEGs, we identified only one WD40 gene (MSI4-like, MD09G1193500) that may participate in anthocyanin biosynthesis; it was upregulated in RIT compared with UIT. However, the specific molecular mechanism needs to be further verified.

3.5. Quantitative Real-Time PCR (qRT-PCR)

To confirm the reliability of the transcriptome data, we examined the expression of eleven DEGs involved in anthocyanin biosynthesis in Malus ‘Profusion’ in the control and rust fungus-infected leaf samples. These DEGs included CHS (MD04G1003300), CHI (MD01G1118100), F3H (MD15G1353800), DFR (MD15G1024100), ANS (MD03G1001100), MYB114-like (MD17G1261000), bHLHA-like (MD03G126300), bHLH33 (MD07G1137500), and MSI4-like (MD09G1193500). RT-PCR data indicated that changes in the expression of 11 genes in UIT and RIT were consistent with our transcriptome sequencing results. In addition, MCMYB10, which plays an important role in the normal development of leaf pigments in Malus spp. [34], was also verified with RT-PCR (Figure 10).

4. Discussion

Plants are affected by many environmental factors during their life cycle and have evolved many adaptive mechanisms to cope with adverse environments and external stress. When exposed to pathogens, plants initiate active defense responses that involve adjusting their physiology and producing bacteriostatic and antifungal metabolites [35,36]. These metabolites are often enriched in areas near the site of pathogen infection, resulting in differences in coloration between normal and infected tissues [37,38,39]. In this study, red lesions were observed around the rust-infected tissues on the leaves of Malus ‘Profusion’. Such infections can be present as early as June. Pigment analysis found that the main chromogenic substance present in these red lesions was cyanidin-3-galactoside chloride.

Anthocyanins are important products of flavonoid metabolism. As colorful pigments, they not only attract pollinators and seed dispersers but also protect plants against pathogen infections [40]. Previous studies have shown that anthocyanins are involved in plant resistance to various fungal pathogens including Botrytis cinerea, bacterial blight, and gray mold [41,42,43]. Our results show that the content of cyanidin-3-galactoside chloride in RIT was significantly higher than that in UIT (Figure 1C). Moreover, in RIT, the contents of several flavonoids were also significantly higher than those in UIT (Table 1). Taken together, the evidence accumulated here suggests that the anthocyanin content of rust-infected leaves of Malus ‘Profusion’ may be an active defense mechanism adopted by plants to resist rust infection.

Anthocyanin biosynthesis is controlled by multiple structural enzymes, including CHS, CHI, F3H, DFR, and ANS. [44]. In this study, we identified seven anthocyanin biosynthetic genes as DEGs between the RIT and UIT samples. Among the structural genes verified by qRT-PCR, ANS is a key gene involved in anthocyanin biosynthesis. ANS can catalyze the conversion of the colorless leucoanthocyanidins to red cyanidin [45] and contribute to the redness of plant tissues. In this study, we found that the expression of ANS in the RIT samples was approximately 10.65-fold higher than that in UIT. Our results show that the higher expression of ANS in RIT is related to the redness of infected spots.

Structural genes associated with anthocyanin biosynthesis are regulated by transcription factors including those of the MYB, bHLH, and WD40 gene families [46]. Some of these transcription factors can form a MYB-bHLH-WD40 ternary complex and jointly regulate the synthesis of anthocyanins [7]. Of the TFs regulating the anthocyanin synthesis pathway, MYB TFs are the most well studied [5]. In this study, 56 MYB genes were found among the identified DEGs. Phylogenetic analysis of the MYB genes in M. ‘Profusion’ and other plants resulted in the identification of MYB114-like (MD17G1261000), a gene that had the highest homology to the MYB genes known to regulate anthocyanin biosynthesis in other plant species. Further comparative analysis revealed that the N-termini of MD17G1261000 had a highly conserved R2R3 domain. In addition, MD17G1261000 also contained a conserved [D/E]L$\times$₂[R/K]$\times$₃L$\times$₆L$\times$₃R structure, which was previously shown to be associated with bHLH binding [29]. We also found two motifs, an “ANDV” sequence and motif 6, in the R3 domain and the C-terminal of MD17G1261000, respectively. To date, motif 6 has only been found in MYB genes that promote anthocyanin biosynthesis in dicotyledons [30]. Another significant motif in MYB genes responsible for promoting anthocyanin biosynthesis is “ANDV”. Here, the ANDV motif was found at positions 100–103 of the R2R3 domain in MD17G1261000 [30]. Taken together, our data indicate that MD17G1261000 may be a R2R3-MYB gene that recruits the R-like bHLH proteins to activate anthocyanin biosynthesis in response to rust infection, although more studies are needed for validation. In addition, typical box 11, 18, and 13 motifs were found in the protein sequences of the two selected bHLH transcription factors: bHLHA-like (MD03G126300) and bHLH33 (MD07G1137500). These structures play an important role in the interaction with MYB transcription factors, leading to enhanced transcription of anthocyanin biosynthetic genes [33]. These findings suggest that the two bHLH genes in Malus may also participate in the regulation of anthocyanin biosynthesis, but additional studies are needed to confirm this. Only one WD40 gene, MSI4-like (MD09G1193500), was identified among the DEGs, and it was upregulated in the RIT samples compared with the UIT samples. Its involvement in anthocyanin biosynthesis remains to be elucidated.

In addition, we found that the level of anthocyanins in Malus ‘Profusion’ leaves, which is induced during normal development at a high carbon–nitrogen ratio, is regulated by MYB10 [34,46,47]. qRT-PCR analysis revealed that the expression of MYB10 in RIT samples was 2.08 times greater than in the UIT samples. Although MYB10 may not be the main regulator, it has a role in the formation of red lesions. In addition, the expression of MYB114-like (MD17G1261000) in the RIT samples was 29.51-fold greater than that in the UIT samples, consistent with that observed for ANS expression. This result suggests that MYB114-like (MD17G1261000) may be the key MYB TF that regulates anthocyanin biosynthesis in response to rust infection.

Based on our results, we propose the molecular mechanism responsible for anthocyanin biosynthesis at sites of rust infection on the leaves of Malus ‘Profusion’. The expression of MYB114-like (MD17G1261000) was activated in rust-infected areas. The resulting protein can combine with bHLH33 (MD07G1137500), bHLHA-like (MD03G126300), and MSI4-like (MD09G1193500) to form the MBW structure. The MBW complex promotes the expression of ANS and other genes in the infected area. This in turn makes the infected area appear red due to anthocyanin biosynthesis. The specific mechanism should be further characterized in future studies.

5. Conclusions

In summary, rust infection results in increased anthocyanin contents in infected apple leaves, suggesting that it plays a role in restricting the spread of the rust infection of these pigments in Malus leaves. Our results show that MYB114-like may be a key MYB TF that regulates anthocyanin biosynthesis in response to rust infection. bHLH33 and bHLHA-like may be key bHLH TFs that regulate anthocyanin biosynthesis in response to rust infection. These findings provide insights into the molecular mechanisms underlying anthocyanin accumulation upon rust infection. Japanese apple rust (Gymnosporangium yamadae Miyabe) is an alternate parasitic fungus. Its telial stage occurs on junipers. Spherical or hemispherical galls are formed on the infected junipers and harm the junipers’ growth and development. Furthermore, rusts are a widespread threat to forestry plants such as pine, poplar, willows, spruce, etc. Improved knowledge about these mechanisms may help to select tree lines with improved resistance to Pucciniales rusts, many of which cause severe problems in forests and tree plantations across the world.