Abstract

   Ilex × attenuata ‘Sunny Foster’ represents a yellow leaf mutant
   originating from I. × attenuata ‘Foster#2’, a popular ornamental woody
   cultivar. However, the molecular mechanisms underlying this leaf color
   mutation remain unclear. Using a comprehensive approach encompassing
   cytological, physiological, and transcriptomic methodologies, notable
   distinctions were discerned between the mutant specimen and its wild
   type. The mutant phenotype displayed aberrant chloroplast morphology,
   diminished chlorophyll content, heightened carotenoid/chlorophyll
   ratios, and a decelerated rate of plant development. Transcriptome
   analysis identified differentially expressed genes (DEGs) related to
   chlorophyll metabolism, carotenoid biosynthesis and photosynthesis. The
   up-regulation of CHLD and CHLI subunits leads to decreased magnesium
   chelatase activity, while the up-regulation of COX10 increases heme
   biosynthesis—both impair chlorophyll synthesis. Conversely, the
   down-regulation of HEMD hindered chlorophyll synthesis, and the
   up-regulation of SGR enhanced chlorophyll degradation, resulting in
   reduced chlorophyll content. Additionally, genes linked to carotenoid
   biosynthesis, flavonoid metabolism, and photosynthesis were
   significantly down-regulated. We also identified 311 putative
   differentially expressed transcription factors, including bHLHs and
   GLKs. These findings shed light on the molecular mechanisms underlying
   leaf color mutation in I. × attenuata ‘Sunny Foster’ and provide a
   substantial gene reservoir for enhancing leaf color through breeding
   techniques.

   Keywords: chlorophyll deficient, chlorophyll synthesis, chloroplast
   development, carotenoid biosynthesis, photosynthesis, transcriptome

1. Introduction

   Leaf hue is a significant trait in ornamental plants, drawing public
   interest and scholarly inquiry [[36]1]. Leaf color mutants present a
   prime opportunity to delve into chlorophyll (Chl) metabolism and
   photosynthesis while unraveling the intricacies of chloroplast
   structure, development, and functionality [[37]2,[38]3]. The diverse
   leaf pigments, notably Chl, carotenoids (Car), and anthocyanins,
   contribute to the vibrant hues observed in plant foliage [[39]4,[40]5].
   Chl, as the primary pigment responsible for leaf greenness, holds
   particular importance, with leaf color mutants often exhibiting Chl
   deficiency [[41]6]. These mutations significantly affect pigment
   synthesis and chloroplast development, leading to alterations in
   pigment quantity, ratio, and leaf color phenotype [[42]7,[43]8]. Albino
   rice leaves exhibit markedly reduced Chl concentrations compared to
   their green counterparts [[44]9]. Correspondingly, the yellow-green
   mutant Cmygp in melon manifested a Chl content of approximately 57.14%
   in comparison to the green phenotype [[45]10]. Conversely, the deep
   green mutant ZD131 in cotton demonstrates a substantially elevated Chl
   content, reaching approximately 115% of its wild type [[46]11].

   The biosynthesis of Chl entailed a sophisticated process involving over
   twenty enzymatic steps. It commenced with the conversion of glutamate
   to 5-aminolevulinic acid (ALA). Subsequently, ALA underwent a series of
   conversions, including porphobilinogen (PBG), uroporphyrinogen III
   (Urogen III), coproporphyrinogen III (Coprogen III), protoporphyrin IX
   (Proto IX), magnesium-protoporphyrin IX (Mg-Proto IX),
   protochlorophyllide (Pchlide), ultimately culminating in the synthesis
   of chlorophyll b and chlorophyll a [[47]12]. Due to the intricate
   enzymatic processes governing Chl biosynthesis, disruptions at any
   stage led to a deficiency of green pigmentation in plant leaves and
   various leaf color phenotypes, as evidenced by mutants like ygl1 in
   Brassica napus and YX-yl in Panicum miliaceum [[48]13,[49]14]. Analysis
   of these Chl precursors is crucial for detecting metabolic
   abnormalities in Chl biosynthesis [[50]15]. When a process in Chl
   synthesis is inhibited, precursor substances upstream of the blocked
   site tend to accumulate significantly, whereas those downstream of the
   obstruction exhibited a marked decrease [[51]16]. For example, the
   yellow mutant lrysl1 identified in Lilium regale impeded the synthesis
   of ALA, consequently leading to Chl deficiency [[52]17]. Similarly,
   disruptions in Chl synthesis occurred during the conversion of Proto IX
   to magnesium protoporphyrin in the rice yellowish-green leaf mutants
   Chlorina-1 and Chlorina-9 [[53]18]. Multiple genes, including HEMC,
   CHLH, HEMD, CAO and SGR, are directly linked to Chl synthesis, with
   mutations in these genes resulting in altered Chl levels
   [[54]19,[55]20]. A point mutation in CsCHLI, a subunit of magnesium
   chelatase I crucial for catalyzing the insertion of Mg^2+ into
   protoporphyrin IX within the Chl synthesis pathway, led to the
   emergence of a golden cucumber leaf mutant [[56]21].

   Moreover, Chl-deficiency mutants often show distinct variations in
   thylakoid structure, hampering photosynthesis due to abnormal
   chloroplast and membrane structures [[57]22,[58]23]. Yellow-green
   leaves and malformed chloroplasts with limited membranes, incomplete
   stacking, and decreased photosynthesis are observed in the Petunia
   mutant E5059, characterized by Chl deficiency [[59]24]. Similarly, in
   the Arabidopsis var2 mutant, regions displaying green sectors with
   typical chloroplasts coexisting alongside variegated sectors exhibited
   abnormal chloroplasts lacking internal membrane structures and hampered
   photosynthesis [[60]25].

   In recent years, researchers have employed various innovative
   techniques, such as transcriptomics, proteomics, metabolomics, and
   integrated approaches, to unravel the complex mechanisms behind leaf
   color mutation [[61]26,[62]27]. For instance, physiological and
   transcriptome analysis of I. × ‘Whoa Nellie’ unveiled alterations in
   chloroplast ultrastructure and reduced gene expression linked to Chl
   biosynthesis and photosynthesis, leading to Chl deficiency in the
   mutant leaves [[63]24]. Similarly, transcriptomic analysis and weighted
   gene co-expression network analysis (WGCNA) profiling in Ginkgo biloba
   identified differentially expressed genes (DEGs) involved in pigment
   metabolism and putative transcription factors (TFs) associated with its
   leaf coloration [[64]28]. Through an integrated transcriptomic and
   metabolomic investigation in Cucumis melo, enzyme genes and metabolites
   predominantly associated with central carbon metabolism,
   photosynthesis, flavonoid metabolism, and energy metabolism were
   identified [[65]29]. Investigations utilizing iTRAQ-based quantitative
   proteomics analysis identified 168 differentially accumulated proteins
   that are involved in photosynthesis, Chl, and Car metabolism in the
   leaf color formation of the hybrid paper mulberry [[66]30]. These
   studies collectively highlighted the intricate nature of leaf color
   mutation mechanisms observed from various plant species.

   Leaf color mutations have been intensively studied across a range of
   plant species like tea [[67]31], rice [[68]8], and wheat [[69]32]. The
   Chl synthesis of albino ‘Huangjinya’ tea and the formation of
   photosynthesis-related proteins were inhibited, resulting in
   photosynthetic efficiency decrease [[70]33]. Similarly, the virescent
   yellow-leaf mutant in rice exhibited compromised chloroplast
   development and decreased Chl levels [[71]34]. The albino cucumber
   cultivar ‘g32’ obtained an albino phenotype that contributed to the
   disruption of chloroplast development [[72]23]. Due to impaired
   photosynthesis, the yellow NAU31 mutant had impaired chloroplast
   structure and reduced Chl and Car content [[73]35]. Although, in recent
   years, ornamental plants have also witnessed the emergence of
   Chl-deficiency mutants, including species like G. biloba [[74]36],
   Quercus shumardii [[75]37], Ulmus pumila [[76]38], Lagerstroemia indica
   [[77]39], and Acer palmatum [[78]40]. However, there is a paucity of
   comprehensive investigations into the molecular mechanisms underlying
   ornamental leaf color mutants. Among these, hollies (genus Ilex) stand
   out for their highly valued traits in landscape design, including
   diverse fruit colors, leaf textures and colors [[79]41]. The vibrant
   color of holly leaves is often regarded as a defining characteristic by
   many gardeners [[80]42]. Among the extensive range of Ilex cultivars,
   more than 165 are colored-leaf varieties [[81]43,[82]44]. Nevertheless,
   research on the leaf mutation mechanism in hollies is scarce.

   In our study, we delve into an important yellow leaf holly variety
   known as I. × attenuata ‘Sunny Foster’, a natural bud mutant of the
   wild type I. × attenuata ‘Foster #2’. It is a well-known Foster Holly
   for its attractive accent and bright yellow flush in direct sunlight,
   and received the Holly Society of America “Gene Eisenbeiss Holly of the
   Year” award in 2006 [[83]45]. As an excellent ornamental plant, it has
   been popularized and planted in most areas of America for its color
   leaves, fruits, tree form, good vigour, cold hardiness, and easy
   culture [[84]46,[85]47]. Although the cytological, physiological, and
   transcriptomic properties of this mutant remain unexplored, our
   investigation sought to quantify the disparities in these aspects
   between the mutant and its wild-type counterpart. By employing a
   multidisciplinary approach encompassing advanced microscopy techniques,
   physiological measurements, and high-throughput RNA sequencing, we
   aimed to provide a thorough comprehension of the determinants
   influencing leaf coloration. The findings greatly enhanced our
   comprehension of the fundamental mechanisms governing leaf color
   mutation in ornamental plants and furnished valuable resources to
   inform and guide ornamental plant breeding efforts.

2. Results

2.1. Yellow Hue and Brighter Color in the Mutant

   Leaf color differences were quantitatively assessed using parameters
   including L*, a*, b*, a*/b*, H° and C. Yellow leaves exhibited
   significantly higher L* and C values compared to green leaves,
   indicating a brighter appearance. Additionally, both green and yellow
   leaves had positive b* values, situating them in the red-yellow region
   of the CIELab color map ([86]Figure 1C). Notably, yellow leaves
   displayed a lower H° value (87.32°) compared to green leaves (118.40°),
   suggesting a shift towards a yellow hue in the mutant. The negative
   a*/b* ratio in green leaves signified their green coloration, with a
   larger absolute value indicating a deeper green tone ([87]Figure 1A,B).

Figure 1.

   [88]Figure 1
   [89]Open in a new tab

   Leaf phenotypes, color indices, pigments, and Chl precursors content of
   YT and WT. (A) YT and WT plants. (B) Leaf phenotypes of YT and WT.
   Scale bar = 1 cm. (C) Leaf color indices of YT and WT. (D) SPAD value
   of YT and WT. (E) Pigments ratios of YT and WT. (F) Pigments content of
   YT and WT. (G) Flavonoid content of YT and WT. (H) Chl precursors
   content of YT and WT. The values are presented as mean ± SD.
   Statistical significance was determined as follows: ns (not
   significant), * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001,
   using the t-test. ALA—5-aminolevulinic acid; Car—carotenoids;
   Chl—chlorophyll; Coprogen III—coproporphyrinogen; FW—fresh weight;
   Mg-Proto IX—Mg-protoporphyrin IX; PBG—porphobilinogen;
   Pchlide—protochlorophyllide; Proto IX—protoporphyrin IX; SD—standard
   deviation; SPAD—Soil and Plant Analyzer Development; Urogen
   III—uroporphyrinogen III; WT—wild-type I. × attenuata ‘Foster #2’;
   YT—mutant I. × attenuata ‘Sunny Foster’.

2.2. Reduced Chl Synthesis in the Mutant

   In yellow leaves, there were significant reductions in SPAD value, Chl
   a, Chl b, Chl a + b, Car, and flavonoid contents compared to green
   leaves, showing decreases of 97.63%, 98.53%, 95.28%, 87.73%, and
   97.81%, respectively ([90]Figure 1D,F,G). However, no significant
   difference was observed in anthocyanin content between the two
   varieties ([91]Figure 1F). Notably, the ratios of Car/Chl a, Car/Chl b,
   and Car/Chl a + b exhibited a significant increase in yellow leaves,
   approximately 10.48, 3.24, and 6.98 times the levels observed in green
   leaves ([92]Figure 1E).

   The contents of all seven Chl precursors were diminished in yellow
   leaves in contrast to green leaves. While the differences in the levels
   of ALA and PBG between yellow and green leaves were not significant,
   the disparities in the levels of Urogen III, Coprogen III, Proto IX,
   Mg-Proto IX, and Pchlide were statistically significant ([93]Figure
   1H), mirroring the trends observed in Chl contents ([94]Figure 1F).

2.3. Varied Leaf Anatomical Structure and Impaired Chloroplast Ultrastructure
in the Mutant

   The epidermal micromorphological characteristics observed under
   scanning electron microscopy were comparable between the mutant and
   wild-type leaves. The upper epidermal cells in both leaf types
   exhibited inconspicuous margins, accompanied by fine lines or distinct
   ridges on the cuticle surface. Additionally, no epidermal hairs were
   observed ([95]Figure 2A,D). The stomatal apparatus was exclusively
   distributed in the lower epidermis and displayed an irregular shape
   ([96]Figure 2B,E). The stomata were oval-shaped and composed of two
   guard cells and two accessory guard cells ([97]Figure 2C,F). Moreover,
   the inner margin of the outer arch cover of the stomata appeared to be
   nearly smooth ([98]Figure 2C,F). While the distribution pattern and
   form of stomata in yellow leaves were similar to those in green leaves,
   significant micromorphological changes were observed between the two
   leaf types. Although the stomatal length and width did not show
   significant differences between I. × attenuata ‘Foster #2’ (WT) and I.
   × attenuata ‘Sunny Foster’ (YT), yellow leaves exhibited a markedly
   lower stomatal density compared to green leaves ([99]Table S1).
   Additionally, while the stomata of green leaves were mostly open
   ([100]Figure 2E,F, [101]Table S1), the majority of stomata in yellow
   leaves were almost fully closed ([102]Figure 2B,C, [103]Table S1).
   These observations implied a potential impairment in heat dissipation
   through the stomatal regulation in the YT mutant.

Figure 2.

   [104]Figure 2
   [105]Open in a new tab

   Epidermal features of YT and WT. (A–C) are YTs. (D–F) are WTs. (A) The
   upper epidermis of YT. (D) The upper epidermis of WT. (B,C) Lower
   epidermis of YT. (E,F) Lower epidermis of WT. Scale bars = 100 μm
   (B,E), 30 μm (A,C,D,F). AC—arch cover; GC—guard cell; R—ridge;
   SA—stomatal apparatus; S—stoma; WT—wild-type Ilex × attenuata ‘Foster
   #2’; YT—mutant I. × attenuata ‘Sunny Foster’.

   Both types of leaves featured an apparent upper epidermis, palisade
   tissue, spongy tissue, and lower epidermis. In the wild-type green
   leaves, the majority of chloroplasts were situated within mesophyll
   cells, comprising both palisade mesophyll cells and spongy mesophyll
   cells ([106]Figure 3E). Notably, the palisade tissue housed a higher
   concentration of chloroplasts than the spongy tissue. The internal
   structure of yellow leaves closely resembled green leaves ([107]Figure
   3A). However, in yellow leaves, the sponge tissue appeared more
   spacious compared to wild-type plants. The relatively lighter color of
   cross-sections from yellow leaves was attributed, at least in part, to
   the reduced size of chloroplasts within these leaves. Additionally, a
   diminished Chl content also contributed to the lighter coloration.

Figure 3.

   [108]Figure 3
   [109]Open in a new tab

   Anatomical structures and chloroplast ultrastructures of YT and WT.
   (A–D) are YTs. (E–H) are WTs. Scale bars = 50 μm (A,E), 10 μm (B,F), 2
   μm (C,G), 1 μm (D,H). CP—chloroplast; CW—cell wall; GL—grana lamella;
   LE—lower epidermis; P—Plastoglobuli; PT—palisade tissue; SG—starch
   grain; SL—stroma lamella; ST—spongy tissue; UP—upper epidermis;
   VE—vesicle; WT—wild-type Ilex × attenuata ‘Foster #2’; YT—mutant I. ×
   attenuata ‘Sunny Foster’.

   There were notable distinctions in chloroplast ultrastructure between
   the green and yellow leaves ([110]Figure 3B–D,F–H, [111]Table 1).
   Chloroplasts in WT plants displayed spindle or spindle-shaped
   morphology, characterized by a well-organized inner membrane system
   with regularly distributed granal thylakoids containing abundant granal
   lamellae, starch granules, and limited plastoglobuli presence
   ([112]Figure 3F–H). Green leaves typically contained an average of
   approximately 9.6 chloroplasts per cell ([113]Table 1). On the
   contrary, chloroplasts in yellow leaves displayed swelling, lacked
   distinct grana, had unclear or absent stromal lamellae, and were
   densely packed with numerous vesicles and plastoglobuli ([114]Figure
   3B–D, [115]Table 1). Yellow leaves exhibited an average of
   approximately 3.9 chloroplasts per cell ([116]Table 1), indicating a
   reduction in number and altered morphology. Analysis from [117]Table 1
   revealed that chloroplasts in green leaves had a length of
   approximately 5.98 μm and a width of about 2.31 μm, resulting in a
   length/width ratio of about 2.67. Conversely, chloroplasts in yellow
   leaves had a length of approximately 6.46 μm and a width of about 5.34
   μm, yielding a length/width ratio of about 1.25, indicating an
   increased chloroplast volume in yellow leaves. While green tissue
   contained approximately 2.1 starch granules per chloroplast, the
   absence of starch granules was notably observed in chloroplasts of
   yellow leaves ([118]Table 1). In summary, alterations in the number,
   size, and morphology of chloroplasts in yellow leaves are evident,
   alongside incomplete development of chloroplasts.

Table 1.

   Comparative analysis of chloroplast ultrastructure of YT and WT.
     Ultrastructural Parameters         YT             WT
   Chloroplast number (per cell)  3.90 ± 0.74 ^b 9.60 ± 0.84 ^a
      Chloroplast length (um)     6.46 ± 0.86 ^a 5.98 ± 0.80 ^a
       Chloroplast width (um)     5.34 ± 1.16 ^a 2.31 ± 0.48 ^b
      Chloroplast length/width    1.25 ± 0.29 ^b 2.67 ± 0.52 ^a
   Starch grain number (per cell) 0.00 ± 0.00 ^b 2.10 ± 0.57 ^a
   [119]Open in a new tab

   Note: The data are expressed as means ± SD (standard deviation) from
   three independent biological replicates. Statistical significance was
   determined at p < 0.05, denoted by different lowercase letters.
   WT—wild-type Ilex × attenuata ‘Foster #2’; YT—mutant I. × attenuata
   ‘Sunny Foster’.

2.4. Decreased Photosynthetic Capacity and Delayed Growth in the Mutant

   Compared to the wild-type plants, the YT plants displayed significant
   reductions of 98.10% in net photosynthetic rate (P[n]), 86.26% in
   stomatal conductance (G[s]), and 79.90% in transpiration rate (E).
   Yellow leaves exhibited a 35.82% higher intercellular CO[2]
   concentration (C[i]) compared to the WT ([120]Table 2). The decrease in
   P[n] and G[s] led to a reduced capacity for CO[2] absorption, thereby
   limiting photosynthesis. However, the C[i] significantly increased in
   mutant plants ([121]Table 2), indicating a decreased rate of CO[2]
   utilization in mutants. The mutant exhibited reductions of 29.60% in
   fluorescence origin (Fo), 75.81% in fluorescence maximum (Fm), 94.25%
   in variable fluorescence (Fv), 76.39% in maximal quantum yield of PSII
   photochemistry (Fv/Fm), 77.08% in actual photosynthetic efficiency of
   PSII (ΦPSII), 95.83% in non-photochemical quenching (NPQ), and 88.24%
   in regulatory energy dissipation quantum yield (Y(NPQ)) compared to its
   wild type plants ([122]Table 2). These reductions suggest that the
   mutant had a compromised self-protective capacity and reduced
   photochemical efficiency. Non-regulatory energy dissipation quantum
   yield (Y(NO)) in the mutant was 59.77% higher compared to that of WT,
   indicating increased photodamage in the mutant. In combination, these
   findings pointed towards a deficiency in the photosynthetic capacity of
   the yellow mutant plants.

Table 2.

   Photosynthetic parameters in the leaves of YT and WT.
   Parameters YT WT
   Photosynthesis P[n] [μmol(CO[2]) m^−2 s^−1] 0.37 ± 0.55 ^b 19.47 ± 2.14
   ^a
   E [mol(H[2]O) m^−2 s^−1] 0.87 ± 0.21 ^b 4.33 ± 0.38 ^a
   G[s] [mol(H[2]O) m^−2 s^−1] 40.67 ± 8.33 ^b 296.00 ± 43.35 ^a
   C[i] [μmol(CO[2]) mol^−1] 428.00 ± 26.00 ^a 274.67 ± 5.13 ^b
   Chl fluorescence Fo 153.00 ± 13.00 ^b 217.33 ± 24.58 ^a
   Fm 184.33 ± 13.80 ^b 762.00 ± 61.88 ^a
   Fv 31.33 ± 4.62 ^b 544.67 ± 37.63 ^a
   Fv/Fm 0.17 ± 0.03 ^b 0.72 ± 0.01 ^a
   ΦPSII 0.11 ± 0.05 ^b 0.48 ± 0.01 ^a
   NPQ 0.02 ± 0.02 ^b 0.48 ± 0.11 ^a
   Y (NO) 0.87 ± 0.04 ^a 0.35 ± 0.02 ^b
   Y (NPQ) 0.02 ± 0.02 ^b 0.17 ± 0.03 ^a
   [123]Open in a new tab

   Note: The data are expressed as means ± SD (standard deviation) from
   three independent biological replicates. Statistical significance was
   determined at p < 0.05, denoted by different lowercase letters.
   C[i]—intercellular CO[2] concentration; E—transpiration rate;
   Fo—fluorescence origin; Fm—fluorescence maximum; Fv—variable
   fluorescence; Fv/Fm—maximal quantum yield of PSII photochemistry;
   G[s]—stomatal conductance; NPQ—non-photochemical quenching; P[n]—net
   photosynthetic rate; WT—wild-type Ilex × attenuata ‘Foster #2’; Y
   (NO)—non-regulatory energy dissipation quantum yield; Y
   (NPQ)—regulatory energy dissipation quantum yield; YT—mutant I. ×
   attenuata ‘Sunny Foster’; ΦPSII—actual photosynthetic efficiency of
   PSII.

   Mutant 2-year-old plants exhibited significantly shorter heights
   compared to their wild-type counterparts, measuring only 75% of the
   wild-type plants ([124]Figure 4A). The fresh mass of mutant plants’
   leaves, stems, and roots decreased dramatically by 23.81%, 35.69%, and
   48.21%, respectively ([125]Figure 4B). The dry mass of mutant plant
   leaves, stems, and roots significantly decreased by 5.83%, 40.42%, and
   52.22%, respectively ([126]Figure 4C). The levels of soluble protein in
   mutant plant leaves significantly decreased by 4.21%, while the levels
   of soluble sugar significantly increased by 8.05% ([127]Figure 4D,E).
   These findings suggest that mutant plant growth was significantly
   deterred.

Figure 4.

   [128]Figure 4
   [129]Open in a new tab

   Plant height, fresh mass, dry mass, and soluble sugar and protein
   content of YT and WT. (A) Plant height of YT and WT. (B) Fresh mass of
   YT and WT. (C) Dry mass of YT and WT. (D) Soluble sugar content of YT
   and WT. (E) Soluble protein content of YT and WT. The values are
   presented as mean ± SD. Statistical significance was determined as
   follows: ** p < 0.01, *** p < 0.001, **** p < 0.0001, using the t-test.
   DW—dry weight; FW—fresh weight; SD—standard deviation; WT—wild-type
   Ilex × attenuata ‘Foster #2’; YT—mutant I. × attenuata ‘Sunny Foster’.

2.5. Transcriptomic Modifications in Pigment Metabolism and Photosynthesis in
the Mutant

   After rigorous quality control, we obtained a total of 260,393,656
   clean reads, with 127,328,916 reads derived from the WT and 133,064,740
   reads from the YT mutant. These clean data exhibited Q20 and Q30
   percentages surpassing 96.77% and 91.39% and had a GC content ranging
   from 44.08% to 44.52%, indicating high quality. The clean data
   exhibited mapping rates to the I. latifolia genome ranging from 67.44%
   to 68.32% ([130]Table S2). Replicate samples displayed high
   correlation, ensuring the uniformity of RNA-seq data ([131]Figure 5A).
   [132]Table S3 provides an overview of the RNA-Seq data.

Figure 5.

   [133]Figure 5
   [134]Open in a new tab

   GO and KEGG enrichment of DEGs between YT and WT. (A) Pearson
   correlation between different samples. (B) DEGs were visualized using a
   volcano plot. On the x-axis, log[2] (fold change) is represented, while
   the y-axis depicts −log[10]Padjj values. Up-regulated genes are denoted
   by red points, whereas down-regulated genes are indicated by green
   points. (C) Clustering of genes between different samples. Red denotes
   genes that are up-regulated, and green indicates genes that are
   down-regulated. (D) GO analyses of the DEGs. (E) The 30 most enriched
   enriched GO terms. (F) The 20 most enriched KEGG terms. The x-axis
   illustrates the enrichment score, while the size of the bubbles
   corresponds to the number of DEGs. Bubble color transitions from purple
   to red indicate decreasing Padjs (increasing significance).
   WT—wild-type Ilex × attenuata ‘Foster #2’; YT—mutant I. × attenuata
   ‘Sunny Foster’.

   YT and WT exhibited a total of 3024 DEGs, with 1212 genes showing
   up-regulation and 1812 genes displaying down-regulation ([135]Figure
   5B,C). For deeper functional understanding, we conducted Gene Ontology
   (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment
   analyses. In the biological process category of the GO analysis, the
   most significantly enriched terms were “cellular carbohydrate metabolic
   process” (GO:0044262), “polysaccharide metabolic process” (GO:0005976),
   and “cellular glucan metabolic process” (GO:0006073). Moving to the
   cellular component category, terms like “Cell wall” (GO:0005618),
   “external encapsulating structure” (GO:0030312), and “apoplast”
   (GO:0048046) took prominence. Within the molecular function category,
   “Zinc ion binding” (GO:0008270), “glucosyltransferase activity”
   (GO:0046527), and “hydrolase activity, acting on glycosyl bonds”
   (GO:0016798) stood out prominently ([136]Figure 5D,E). These results
   suggest alterations in carbohydrate metabolism, cell membrane
   composition, and catalytic activities in the YT mutant, possibly
   contributing to its phenotype.

   The KEGG pathway enrichment analysis revealed that the DEGs were
   distributed across 122 pathways. Notably, pathways such as “starch and
   sucrose metabolism” (ko00500), “phenylpropanoid biosynthesis”
   (ko00940), “porphyrin metabolism” (ko00860), “biosynthesis of
   co-factors” (ko01240) and “biosynthesis of various plant secondary
   metabolites” (ko00999) showed significant enrichment (Padj < 0.05).
   Additionally, “flavonoid biosynthesis” (ko00941),
   “photosynthesis-antenna proteins” (ko00196) and “photosynthesis”
   (ko00195) were included in the top 20 pathways ([137]Figure 5F). Given
   the decreased Chl, Car, flavonoid contents and impaired photosynthetic
   efficiency in the yellow mutant, the pathways associated with porphyrin
   metabolism, Car biosynthesis, flavonoid biosynthesis, and
   photosynthesis warranted particular attention. Specifically, we
   identified 18 DEGs involved in the “porphyrin metabolism” pathway,
   seven DEGs in “carotenoid biosynthesis”, 10 DEGs linked to “flavonoid
   biosynthesis”, 16 DEGs related to the “photosynthesis”, and seven DEGs
   associated with “photosynthesis-antenna proteins” ([138]Table S5).

2.6. Altered Chl Biosynthesis in the Mutant

   Through KEGG enrichment analysis, we identified 18 DEGs encoding
   enzymes closely associated with porphyrin and Chl metabolism
   ([139]Figure 6, [140]Table S5). Among these DEGs, 15 genes related to
   Chl synthesis (e.g., GSA, HEMC, CHLD, PORA) and 2 genes linked to Chl
   degradation (e.g., SGR) were up-regulated, while HEMD (encoding
   S-adenosyl-L-methionine-dependent uroporphyrinogen III
   methyltransferase) associated with Chl synthesis showed
   down-regulation. Notably, the concentrations of Chl precursors,
   spanning from ALA to PBG, exhibited no substantial disparities between
   WT and YT. However, Urogen III, Coprogen III, Proto IX, Mg-Proto IX,
   and Pchlide displayed significant decreases in YT, suggesting an
   impediment from PBG to Urogen III synthesis. These findings indicate
   that the inhibition of Chl synthesis in YT plants was attributed to the
   decreased expression of HEMD, while the increase in SGR expression
   likely contributed to the reduction in Chl levels. The increased
   expression of other genes involved in Chl synthesis served as a
   compensatory mechanism arising from the deficiency of Chl.

Figure 6.

   [141]Figure 6
   [142]Open in a new tab

   DEGs involved in the chlorophyll metabolism pathway between YT and WT.
   The mean log[2] (Fold Change) values for the DEGs were computed based
   on three biological replicates for YT and WT. Up-regulated genes are
   denoted in red, while down-regulated genes are shown in blue.
   WT—wild-type Ilex × attenuata ‘Foster #2’; YT—mutant I. × attenuata
   ‘Sunny Foster’.

   Gene Set Enrichment Analysis (GSEA) analysis revealed markedly enriched
   molecular pathways, including RNA polymerase (ath03020), ribosome
   (ath03010), homologous recombination (ath03440), porphyrin metabolism
   (ath00860), and pantothenate and CoA biosynthesis (ath00770) in YT
   (nominal p-value < 0.05, NES > 1.5, FDR q-value < 0.25) ([143]Table
   S6). Beyond the previously discussed up-regulated DEGs, the enriched
   porphyrin metabolism pathway encompassed 14 DEGs, including CHLG, CHLP,
   CLH2, COX10, HEMF, HCAR, HY2, NOL, PORA, PPOX1 ([144]Figure S1,
   [145]Table S7). This underscores the significance of Chl
   metabolism-related genes in influencing leaf coloration in YT. The
   pronounced enrichment COX10 (encoding heme synthase) (ES = 0.618)
   highlights possible disruptions in the Chl-heme metabolic pathway, thus
   emphasizing their pivotal role in determining leaf coloration in I. ×
   attenuata ‘Sunny Foster’.

2.7. Down-Regulated Carotenoid and Flavonoid Biosynthesis in the Mutant

   Six DEGs encoding enzymes crucial for Car biosynthesis were identified,
   with four being down-regulated and two up-regulated in the mutant.
   Down-regulated DEGs included CCS, NCED, and CYP707A1, while LCYE and
   CYP707A2 were significantly up-regulated in the mutant ([146]Figure 7,
   [147]Table S5). The observed decrease in Car content in yellow leaves
   suggested an association with the down-regulation of genes involved in
   Car metabolism, indicating a disruption in Car biosynthesis pathways in
   the mutant plants.

Figure 7.

   [148]Figure 7
   [149]Open in a new tab

   DEGs involved in the carotenoid metabolism pathway between YT and WT.
   The mean log[2] (Fold Change) values for the DEGs were computed based
   on three biological replicates for YT and WT. Up-regulated genes are
   denoted in red, while down-regulated genes are shown in blue.
   WT—wild-type Ilex × attenuata ‘Foster #2’; YT—mutant I. × attenuata
   ‘Sunny Foster’.

   Given YT’s significantly decreased flavonoid content compared with WT,
   we subsequently analyzed the flavonoid biosynthesis pathway. We
   pinpointed 10 DEGs linked to flavonoid biosynthesis, with all
   exhibiting down-regulation in the mutant, including CYP92C6, CAMT,
   CYP98A2, HCT, ANR, and CHI ([150]Figure S2, Table S5).

2.8. Down-Regulated DEGs in Photosynthesis and Chloroplast Development

   The aberrant chloroplast ultrastructure and photosynthesis in the YT
   mutant prompted further investigation into the molecular mechanisms
   underlying these phenotypic changes. Significant differential
   expression was observed in 41 genes related to photosynthesis and
   chloroplast development when comparing the mutant with normal green
   leaves, as depicted in the heatmap presented in [151]Figure 8 and
   [152]Table S5. Specifically, while two DEGs associated with the PSII
   reaction center subunit (psbC) demonstrated reduced expression, two
   DEGs involved in the PSI reaction center subunit (psaA and psaB) and
   three DEGs (psbQ, psbK, and psbE) exhibited higher expression levels in
   the mutant ([153]Figure 8B,C, [154]Table S5). One cytochrome b6/f
   complex gene (petD), one photosynthetic electron transport gene (petF),
   and four F-type ATPase genes (atpI, atpA, atpB, atpE) exhibited higher
   expression levels in the mutant. On the contrary, the mutant exhibited
   reduced expression levels in three photosynthetic electron transport
   genes (petF and petH) ([155]Figure 8D–F, [156]Table S5). Three genes
   involved in carbon fixation in photosynthesis had increased
   transcription, while six genes were significantly down-regulated in the
   YT mutant, affecting carbon assimilation. The downregulation of these
   genes weakened CO[2] fixation ([157]Figure 8G, [158]Table S5). All nine
   genes associated with nitrogen metabolism were down-regulated
   ([159]Figure 8H, [160]Table S5). The down-regulation of these genes
   suggested a potential impairment of photosynthetic activity in the
   mutant plants.

Figure 8.

   [161]Figure 8
   [162]Open in a new tab

   The expression profiles of DEGs involved in photosynthesis and nitrogen
   metabolism between YT and WT. (A) DEGs involved in
   photosynthesis-antenna protein. (B) DEGs involved in photosystem II.
   (C) DEGs involved in photosystem I. (D) DEGs involved in cytochrome
   b6/f complex. (E) DEGs involved in photosynthetic electron transport.
   (F) DEGs involved in F-type ATPase. (G) DEGs involved in carbon
   fixation in photosynthetic organisms. (H) DEGs involved in nitrogen
   metabolism. The mean log[2] (Fold Change) values for the DEGs were
   computed based on three biological replicates for YT and WT.
   Up-regulated genes are denoted in red, while down-regulated genes are
   shown in blue. WT—wild-type Ilex × attenuata ‘Foster #2’; YT—mutant I.
   × attenuata ‘Sunny Foster’. The light-harvesting chlorophyll protein
   complex schematic in the top-left corner and the photosynthesis pathway
   schematic in the top-right corner are sourced from the KEGG pathway
   website ([163]https://www.kegg.jp, accessed on 22 December 2023).

   Differential expression of the light-harvesting chlorophyll protein
   complex was also observed between YT and WT in the
   photosynthesis-antenna protein pathway. Specifically, four DEGs
   encoding light-harvesting complex II Chl a/b binding proteins (LHCB)
   and three DEGs encoding light-harvesting complex I Chl a/b binding
   proteins (LHCA) were significantly up-regulated in the mutant
   ([164]Figure 8A, [165]Table S5). The increased expression of these key
   genes indicated a compensatory regulatory mechanism to efficiently
   capture and transfer light energy for photosynthesis in the mutant.

2.9. Differentially Expressed Transcription Factors in the Mutant

   A comprehensive analysis revealed 311 DEGs encoding putative
   transcription factors (TFs), spanning 43 TF families ([166]Figure 9,
   [167]Table S8). The analysis highlighted nine predominant TF families
   among these, namely bHLHs, ERFs, NACs, MYB-related, MYBs, G2-likes
   (GLK), C2H2s, FAR1s, and WRKYs ([168]Figure 9). The bHLH family was
   most abundant, comprising 29 TFs and representing 9.3% of the total.
   Out of these, 16 TFs displayed decreased expression, while 13 TFs
   demonstrated increased expression in the mutant. Additionally, after
   bHLH, the ERF (26 TFs, 8.4%), NAC (22 TFs, 7.1%), MYB-related (21 TFs,
   6.8%), MYB (18 TFs, 5.8%), and GLK (17 TFs, 5.5%) families displayed
   the next highest representation ([169]Figure 9, [170]Table S8). It is
   worth mentioning that 13 GLK TFs demonstrated decreased expression,
   encompassing GLK1 and GLK2, whereas 4 TFs displayed increased
   expression in the mutant. These results suggest potential disruptions
   in chloroplast development, which could play a role in the
   manifestation of the observed yellow phenotype in the mutant.

Figure 9.

   [171]Figure 9
   [172]Open in a new tab

   DEGs involved in transcription factors between YT and WT.

2.10. qRT-PCR Validation of the Candidate DEGs

   The DEGs identified through RNA-Seq analysis were validated using the
   qRT-PCR method. We selected a range of genes implicated in Chl
   metabolism, Car metabolism, PSI, PSII, and TFs for validation
   ([173]Table S9). A total of 12 DEGs were examined, and their expression
   patterns aligned with the transcriptome data ([174]Figure S3,
   [175]Tables S4 and S9). This consistency between the qRT-PCR results
   and transcriptome data affirmed the credibility of the transcriptome
   data and validated its use for this study.

3. Discussion

   Mutants lacking Chl offer essential contributions to scientific
   research and practical applications, especially concerning woody plant
   mutants with altered leaf coloration [[176]48,[177]49]. Holly leaf
   color mutants have received significant interest owing to their
   distinct foliage hues and economic and practical utility [[178]42]. In
   this study, we investigated a significant Chl-deficient holly cultivar
   exhibiting a yellow leaf phenotype. Through comprehensive cytological,
   physiological, and transcriptomic analyses, we aimed to elucidate its
   distinctive characteristics.

   The differences in coloration between the mutant leaves and those of
   the wild type highlight the exceptional nature of this leaf color
   mutation, reflecting the diverse leaf coloration observed among the
   plants [[179]50]. The significant reductions in Chls contents in yellow
   leaves signify a substantial impairment in Chl biosynthesis.
   Specifically, the decrease in Chl b exceeded that of Chl a, classifying
   this mutant as a Chl b-deficient type according to Falbel’s
   classification [[180]51]. Similar observations have been reported in a
   yellow mutant of Q. shumardii, which exhibited higher L* and C values,
   lower Chl content, and a pronounced decline in Chl b compared to Chl a
   [[181]37]. The significantly higher ratios of Car/Chls in yellow leaves
   could be the direct reason for the mutant’s yellow appearance, as seen
   in yellow mutants of G. biloba [[182]15], Brassica napus [[183]14], and
   L. indica [[184]52].

   The process of Chl biosynthesis is essential for both photosynthesis
   and plant development, given the critical role of Chl molecules. They
   play pivotal roles in absorbing light energy, facilitating electron
   transfer within the photosynthetic electron transport chain, and
   initiating reactions that separate charges within reaction centers
   [[185]53,[186]54]. Photosynthesis, in turn, is fundamental for
   synthesizing organic compounds essential for plant growth and
   development [[187]55]. This process occurs on the thylakoid membranes
   of chloroplasts, which are structured into regularly stacked grana in
   higher plants, optimizing light harvesting and energy conversion
   efficiency [[188]56]. Therefore, leaf color mutants often exhibit
   structural alterations in thylakoid organization, such as reduced grana
   stacks and irregular arrangement, indicating an underdeveloped
   chloroplast ultrastructure and potential loss of photosynthetic
   capacity [[189]26,[190]57]. For instance, in the pgb tea Chl-deficient
   mutant, mature leaves exhibited significantly fewer chloroplasts with
   defective morphology, lacking stacked thylakoid structures but
   containing numerous giant plastoglobuli, along with reduced osmiophilic
   granules and absence of starch granules in the plastids [[191]58].
   Similarly, in this study, the swollen appearance, absence of stromata
   and starch grains, and presence of numerous vesicles and plastoglobuli
   indicate severe disruption of chloroplast structure in the mutant.
   Therefore, the reduction in Chl content and the occurrence of aberrant
   structural alterations in chloroplasts should be the primary factors
   contributing to our study’s observed decrease in photosynthetic
   capacity. Comparable findings were observed in the yellow mutant of U.
   pumila [[192]38]. As a consequence of compromised photosynthesis,
   mutant leaves also exhibit deficiencies in carbon skeletons and major
   carbon metabolism end-products, including flavonoids, leading to
   delayed plant growth [[193]59].

   Leaf mutations arise from a complex interaction of environmental
   conditions and genetic determinants [[194]6]. Within the framework of
   this investigation, the simultaneous occurrence of green-leaf plants
   and yellow-leaf mutants suggests the importance of genetic
   modifications. RNA-Seq analysis is instrumental in identifying DEGs in
   various developmental stages and physiological conditions. Recent
   studies on Hemerocallis spp. [[195]60] and Primulina pungentisepala
   [[196]61] have utilized RNA-Seq technology to explore the mechanisms
   underlying leaf mutation. Likewise, our study utilized Illumina
   sequencing to identify 3024 DEGs in the yellow and green leaves. We
   primarily concentrated our data analysis on DEGs related to porphyrin
   and Chl metabolism, Car biosynthesis, photosynthesis, photosynthetic
   antenna proteins, and TFs that contribute to leaf coloration.

   The pathway of Chl synthesis involves a cascade of enzymatic reactions
   catalyzed by over twenty enzymes [[197]16]. Examining Chl precursors
   helps identify potential impediments in the Chl synthesis pathway
   [[198]62]. In the yl1 rice mutant, there was a significant increase in
   Proto IX content, while the levels of Mg-Proto IX and Pchlide were
   notably reduced compared to the WT. This indicates a potential
   impairment in the synthesis of Mg-Proto IX from Proto IX, contributing
   to the yellow leaf phenotype observed in the mutant [[199]63]. Compared
   to wild-type plants, the levels of ALA and PBG exhibited
   non-significant variations, while the levels of Urogen III, Coprogen
   III, Proto IX, Mg-Proto IX, and Pchlide displayed a significant
   decrease. These findings indicated that Chl biosynthesis in the mutant
   should be blocked during the conversion of PBG to Urogen III in our
   study. Small alterations in the expression levels of the 27 genes,
   which encode 15 enzymes responsible for Chl synthesis in A. thaliana,
   had the potential to disturb Chl metabolism and concentration, thereby
   causing variations in leaf color [[200]64]. For example, if HEMA, which
   encodes glutamyl-tRNA reductase (GluTR) responsible for catalyzing
   glutamyl-tRNA into glutamate-1-semialdehyde (GSA) and initiating Chl
   biosynthesis, is down-regulated, it can result in reduced Chl content
   [[201]65,[202]66]. The down-regulation of HEMD, encoding
   uroporphyrinogen III synthase and cyclizing and isomerizing
   hydroxymethylbilane to produce uroporphyrinogen III, was associated
   with the decrease in Chl levels observed in pink leaves compared to
   variegated pink-green leaves in ornamental kale [[203]67]. Magnesium
   chelatase (MgCh) was crucial in regulating Chl biosynthesis through its
   catalytic function, inserting Mg^2+ into protoporphyrin IX and
   converting it into Mg-protoporphyrin IX [[204]68,[205]69]. Comprising
   three subunits—D, I, and H—MgCh’s functionality relies on their
   synergistic interaction, with these subunits being conserved across
   plant species [[206]6]. Reduced MgCh activity in leaf mutants was
   associated with a reduction in Chl content [[207]37]. Studies have
   shown that suppressing the CHLI gene in strawberries led to yellow
   leaves and disrupted Chl-heme synthesis [[208]70]. The mutant gene
   GaCHLH in Gossypium arboreum failed to interact effectively with
   GaCHLD, thereby hindering MgCh assembly and disrupting Chl synthesis
   [[209]71]. In our investigation, the up-regulation of CHLH and CHLD may
   disrupted the synergy among the three subunits, ultimately leading to
   decreased Chl biosynthesis in the mutant. Both Chl and heme originate
   from a shared precursor molecule, glutamyl-tRNA^glu, constituting
   tetrapyrrole molecules. Because their synthesis competes for the same
   substrate, changes in heme metabolism can impact the rate of Chl
   synthesis [[210]16]. Feedback regulation from excessive heme inhibits
   ALA synthesis, disrupting Chl metabolism and causing changes in leaf
   color [[211]72]. In investigating the grc1 rice mutant, researchers
   identified a 45-kb insertion within the gene LOC_Os06g40080, encoding a
   heme oxygenase (HO1), negatively impacting Chl synthesis [[212]73]. In
   a similar vein, the yellow-green leaf mutant of B. napus showcased
   decreased Chl biosynthesis and increased heme production due to an
   Ile320Thr mutation in BnaC08g34840D [[213]74]. We found increased
   expression of heme synthase pathway genes (COX10), which had disrupted
   Chl metabolism in the mutant.

   Changes in leaf coloration can also stem from mutations influencing the
   Chl degradation process. The initial step in Chl, a breakdown
   facilitated by Mg-dechelatase, involved the removal of the central Mg
   ion and was encoded by the Stay-Green (SGR) gene [[214]75]. Arabidopsis
   mutants lacking pheophorbide oxygenase displayed stay-green leaves,
   while overexpression of SGR in rice resulted in a yellowish-brown
   phenotype [[215]76]. In line with these findings, the yellow type of I.
   × attenuata ‘Sunny Foster’ showed significantly reduced Chls
   concentrations. Key DEGs related to Chl synthesis were suppressed,
   while Chl degradation was enhanced, resulting in reduced Chl
   accumulation and the observed lower Chls levels.

   Cars, acting as accessory light-harvesting pigments, capture and
   transfer light energy to Chls, playing a crucial role in
   photoprotection within photosystems, particularly under high light
   conditions [[216]77]. They contributed to the yellow-to-red hues
   observed in leaves [[217]78]. Down-regulation of genes related to Car
   biosynthesis could lead to leaf yellowing [[218]79]. In our analysis,
   we identified six DEGs associated with Car biosynthesis. Among them,
   four showed significant down-regulation in the mutant plant, namely
   CCS, NCED1, NCED2, and CYP707A1. The Car content typically experienced
   a less pronounced reduction than Chl content in many yellow-leaf
   mutants, although both notably decreased [[219]51]. Consequently, the
   ratio of Car/Chl significantly increased, suggesting an elevated Car
   content that should contributed to the yellow leaf phenotype.

   Light energy absorption occurs through pigments, initiating its
   transfer to the reaction centers within the thylakoid membrane
   [[220]80], encompassing PSII, cytochrome f/b6 (cytf/b6), PSI, the Chl
   a/Chl b light-harvesting complex (LHC), and ATP synthase [[221]81]. The
   reaction centers are crucial in facilitating electron transfer driven
   by light, absorbing and harnessing its energy [[222]82]. In
   Chl-deficient rice mutants, crucial PSI core proteins (psaA and psaB),
   subunits of ATP synthase, cytochrome, and the LHC, were significantly
   decreased [[223]83]. The down-regulation of genes involved in
   photosynthesis impaired the function of these complexes, hindering
   photosynthesis [[224]84]. Significant alterations in the
   phosphorylation levels of PsbC, PsbO, PsbP, LHCA2, and LHCA3 proteins
   were key factors contributing to the observed photosynthetic
   disparities between the yellowing mutant yl1 and the wild type
   [[225]85]. Previous studies had also proposed a connection between
   diminished expression of photosynthesis-related DEGs, like psbC, petD,
   and atpA, and compromised light absorption, leading to inefficiencies
   in energy transfer to reaction centers [[226]15,[227]86]. In our study,
   the yellow mutant exhibited reduced expression levels of five
   photosynthesis-related DEGs (two psbC, two petH, one petF), in contrast
   to the wild-type plants. This observation aligned with the diminished
   photosynthetic capacity and abnormal chloroplasts observed in the
   mutant. The decreased activity of photosynthesis genes likely
   contributed to the noticeable decline in photosynthetic potential
   observed in the yellow mutant.

   TFs play pivotal roles in orchestrating the precise temporal and
   spatial expression patterns of target genes, exerting control over
   their transcriptional activity through activation or repression
   mechanisms [[228]87]. They exert influence or control over numerous
   molecular functions and biological processes, including pigment
   biosynthesis and chloroplast development [[229]88,[230]89]. The
   differentially expressed TFs identified in our study belonged to gene
   families such as bHLHs, NACs, MYB-related, ERFs, B3s, C2H2s, GRASs,
   LBDs, MYBs, WRKYs, and GLKs. Similar findings were reported in a
   yellow-leaf mutant of G. biloba [[231]28]. The suppression of
   bHLH71-like in pepper plants led to elevated levels of zeaxanthin and
   antheraxanthin, implying a role for bHLH71-like as a promoter of Car
   biosynthesis in the yl1 yellow pepper mutant [[232]90]. Moreover, bHLHs
   also regulate the formation and development of stomata. Knocking out
   the bHLH-type TFs OsICE1 and OsICE2 in A. thaliana resulted in the loss
   of stomata [[233]91]. Studies have demonstrated that ANAC046 positively
   regulates Chl biosynthesis in A. thaliana, interacting with the
   promoter regions of key genes involved in Chl synthesis, including
   NYC1, SGR1, SGR2, and PAO [[234]92]. Gene families like MYB-related
   [[235]93], ERFs [[236]94], WRKYs [[237]95], and MYBs [[238]96] have
   been documented to exert significant influence on Chl metabolism and
   consequently impact leaf coloration. The GLK TFs, found within the
   newly classified GARP superfamily in plants, were known to participate
   in a range of biological processes, including the formation of
   chloroplasts [[239]97,[240]98]. In flowering plants, a typical
   occurrence involved a pair of GLK genes, namely GLK1 and GLK2,
   expressed in photosynthetic tissues and performing overlapping
   functions [[241]98]. In rice, enhancing chloroplast formation in both
   callus and roots had been observed through the overexpression of OsGLK1
   [[242]99]. In Arabidopsis, Atglk double mutants displayed pale green
   leaves, with chloroplasts measuring approximately half the size in the
   cross-sectional area compared to wild-type ones [[243]100]. Research
   suggested that BpGLK1 is critical in regulating chloroplast development
   in the yellow-green mutant by binding directly to the promoters of
   genes linked to BpCHLH and BpCRD1 [[244]1]. Overexpression of MdGLK1 in
   the MdGLK double mutant alleviated the abnormalities in chloroplast
   structure and function, leading to the up-regulation of genes
   associated with Chl biosynthesis (CHLH, CAO) and photosynthesis
   (LHCB1-5) [[245]101]. Similarly, lower expression of GLKs had been
   observed in yellow leaves of G. biloba compared to green leaves
   [[246]15]. In our investigation, seven GLK TFs, including GLK1 and
   GLK2, showed decreased expression levels in the yellow mutant,
   underscoring a tight association between changes in leaf color and
   chloroplast development. These discoveries provide further insights
   into the involvement of TFs’ role in modulating pigment metabolism,
   chloroplast maturation and influencing leaf coloration in I. ×
   attenuata ‘Sunny Foster’.

4. Conclusions

   This study thoroughly examined cytological, physiological, and
   transcriptomic variances between wild-type and yellow mutant leaves.
   The yellow leaf mutant exhibited aberrant chloroplast morphology,
   notable discrepancies in pigment levels, limited crucial Chl precursors
   synthesis, and diminished photosynthetic efficiency compared to its
   wild type. The decrease in Chl b content exceeded that of Chl a,
   classifying this mutant as Chl b-deficient. Transcriptome analysis
   elucidated a spectrum of DEGs linked to pigment metabolism in the
   yellow mutant. The decreased expression of CHLD and CHLI, indicating
   unsynergistic interactions with CHLH, as well as the diminished
   expression of HEMD and the heightened expression of COX10, disrupted
   Chl synthesis while favoring heme synthesis.

   Furthermore, the Chl degradation process was heightened by the
   increased expression of SGR. As a result, these combined effects
   significantly reduced Chl concentration within the mutant.
   Additionally, a reduction in the activity of genes involved in Car
   biosynthesis was noted, corresponding with the lower carotenoid
   content. The down-regulation of TFs such as GLK1, GLK2, and bHLHs
   contributed to the defective chloroplasts. These effects ultimately
   culminated in a decrease in Chl content and an elevation in the Car/Chl
   ratios, manifesting as the yellow leaf phenotype ([247]Figure 10). In
   conclusion, our findings offer meaningful gene resources that could
   inform plant breeding efforts to enhance leaf color traits.

Figure 10.

   [248]Figure 10
   [249]Open in a new tab

   A proposed model for potential molecular leaf mutations in I. ×
   attenuata ‘Sunny Foster’. Gene upregulation is indicated by the
   vertical red arrow; gene downregulation is denoted by the vertical blue
   arrow. The green leaf originates from the wild-type plant (I. ×
   attenuata ‘Foster#2’), and the yellow leaf is derived from the mutant
   plant (I. × attenuata ‘Sunny Foster’).

5. Materials and Methods

5.1. Plant Materials

   The study analyzed green leaves from the cultivar I. × attenuata
   ‘Foster #2’ and yellow leaves from its naturally occurring bud mutant
   cultivar, ‘Sunny Foster’, characterized by the presence of newly
   emerged yellow foliage, as shown in [250]Figure 1. Formal
   identification of these plant materials was conducted by Professor
   Donglin Zhang from the University of Georgia. Voucher specimens of
   these plant materials have been deposited in the Nanjing Forestry
   University Ilex Germplasm Resource Bank with the deposit numbers
   DI00025 and DI00026. Both varieties were 2-year-old cutting liners,
   planted in one-gallon plastic containers. Each pot was filled with a
   substrate composed of a 3:1 volume ratio mixture of peat and perlite,
   with the substrate pH value set at 6.5. A cohort of 200 two-year-old
   plants constituted the experimental population. These specimens were
   grown under standardized environmental conditions at Nanjing Qingzhuguo
   Landscape Horticulture Co., Ltd. (Nanjing, China).

   On a cloudless day in June 2023, green leaves from the WT cultivar and
   yellow leaves from the YT cultivar were collected. Three leaves were
   collected from each individual plant of both varieties. For each
   biological replicate, three randomly selected plants from both the wild
   type and mutant type were utilized. Three measurements were conducted
   for each accession to ensure robustness of the data. Following
   collection, the harvested leaves were immediately flash-frozen in
   liquid nitrogen and stored at −80 °C for subsequent analyses. All leaf
   samples were collected with the approval of the owner of Nanjing
   Qingzhuguo Landscape Horticulture Co., Ltd. (Nanjing, China).

5.2. Color Indice Measurement

   A handheld precision colorimeter (CR-8, Shenzhen 3nh Technology Co.,
   Ltd. Shenzhen, China) was employed to record three color parameters:
   L*, a*, b* values. Following this, the hue angle (H° = arctan (b*/a*),
   color ratio (a*/b*), and chroma (C = (a*^2 + b*^2)^1/2) were computed
   based on previously reported formulas [[251]102]. The parameter L*
   indicates brightness, measured on a scale ranging from 0 to 100, while
   a* and b* represent chromaticity without predefined numerical
   boundaries. Negative values of a* indicate green hues, while positive
   values suggest red. Similarly, negative b* values correspond to blue,
   and positive values signify yellow. The parameter C quantifies color
   intensity. The hue angle (H°) is utilized to characterize colors, with
   red associated with a hue angle of 0°, yellow with 90°, green with
   180°, and blue with 360° [[252]103]. Three separate plants were
   measured for each variety, with three random points selected on each
   leaf from the upper, middle, and lower regions and their average values
   were calculated.

5.3. SPAD, Pigment and Chl Precursors Content Measurements

   The SPAD value, serving as an indicator of Chl content, was determined
   utilizing a SPAD-502 Chl meter (SPAD, Konica Minolta Sensing, Tokyo,
   Japan). For Chl extraction, 0.2 g of both green and yellow leaves were
   finely chopped and immersed in 95% ethanol overnight. Subsequently, the
   extracted Chl and Car were quantified at wavelengths of 470 nm, 649 nm,
   and 665 nm using a UV-1800 spectrophotometer (DU-800, Beckman Coulter,
   Brea, CA, USA), following the method described by Lichtenthaler
   [[253]104]. Anthocyanins were extracted in a solution consisting of
   methanol and HCl (99:1, v/v) at 4 °C in a light-free environment for 24
   h, as outlined by Hughes et al. and Manetas et al. [[254]105,[255]106].
   The anthocyanin content was determined by measuring the absorbance at
   530 nm, with cyanidin-3-glucoside as the calibration standard.
   Flavonoid content was assessed by referencing a previous method
   [[256]107]. Briefly, 0.5 mL aliquots of properly diluted extracts or
   standard solutions were dispensed into 15 mL conical tubes filled with
   2 mL of double-distilled water. Afterward, 0.15 mL of a 5% solution of
   sodium nitrite (NaNO[2]) was incorporated, allowing for a 5 min
   reaction period. This was followed by the addition of 0.15 mL of 10%
   solution of aluminum chloride hexahydrate (AlCl[3]·6H[2]O), followed by
   an additional 5 min incubation period. After adding 1 mL of a 1 M
   sodium hydroxide (NaOH) solution, the resulting mixture was vigorously
   agitated and then incubated for 15 min. The absorbance of the solution
   was measured at a wavelength of 415 nm.

   The procedure involved pulverizing 50 mg of fresh leaves into a fine
   powder using liquid nitrogen. The powder was combined with 450 mL of
   0.01 mol L^−1 PBS solution (pH 7.4) and subjected to centrifugation at
   12,000× g for 10 min to obtain the supernatant. Plant Enzyme-Linked
   ImmunoSorbent Assay kits (#YJ503247, Yuanju, Shanghai, China) were
   utilized to quantify the levels of seven Chl precursor compounds,
   including ALA, PBG, Urogen III, Coprogen III, Proto IX, Mg-Proto IX,
   and Pchlide, in both types of leaves.

5.4. Leaf Anatomical Structure Observation

   Leaf samples were sliced into segments measuring 1 cm × 2 cm and
   immersed in a fixative solution comprising formalin, alcohol, and
   glacial acetic acid (in a ratio of 90:5:5 by volume) for a minimum of
   48 h. After fixation, the leaves underwent dehydration using a series
   of ethanol solutions, followed by two xylene rinses, and subsequently
   encased within paraffin wax, as previously outlined [[257]108]. Samples
   were permeated and paraffin-embedded. These samples were double-stained
   with safranin and fast green counterstain. Photomicrographs were
   captured using an optical microscope (BX53, Olympus, Tokyo, Japan).

5.5. Leaf Epidermis Microstructure Observation

   Cryogenic Scanning Electron Microscopy (CryoSEM) was used to analyze
   leaf epidermis microstructure as detailed previously [[258]109]. Fresh
   leaf fragments with 1 cm × 1 cm were firmly affixed onto 2 cm × 4 cm
   copper plates using computer thermal compound paste. The samples were
   initially frozen at −20 °C on a large metal holder before being coated
   with a layer of gold in an argon environment. The plate containing
   either the fresh, non-metalized sample or the frozen, metalized sample
   was securely attached to the cooling stage of a Deben Coolstage
   refrigerating unit (Deben, Suffolk, UK) set at −30 °C. These specimens
   were then examined using a Quanta 200 scanning electron microscope
   (FEI, Hillsboro, OR, USA) in high vacuum mode at 20 kV, with a
   backscattered electron detector QBSD and an 8–12 mm working distance.

5.6. Chloroplast Ultrastructure Observation

   Following the previously described methodology [[259]110], 1 mm × 2 mm
   segments were prepared for ultrastructural examination and fixed with
   2.5% glutaraldehyde in 0.1 mol L^−1 phosphate buffer (with a pH of
   7.2). After three 15 min washes in phosphate buffer, they underwent a 2
   h post fixation with 1% OsO[4] in phosphate buffer (with a pH of 7.0).
   After three additional washes in phosphate buffer, the samples were
   dehydrated using a gradient series of ethanol (30%, 50%, 60%, 70%, 80%,
   90%, and 95%), with each step lasting approximately 15 to 20 min.
   Subsequently, they were immersed in pure acetone for a duration of 1 h,
   with a total of seven washes. For infiltration, the samples underwent a
   series of steps: they were initially immersed in a 1:1 mixture of pure
   acetone and the final spurr resin mixture for 1 h at room temperature.
   Subsequently, they were transferred to a 1:3 mixture of pure acetone
   and the final resin mixture for 3 h. Finally, they were placed in the
   pure spurr resin mixture overnight. The embedded sample was then
   sectioned into thin slices (50–80 nm) using a diamond knife and an
   ultra-microtome (UC6/FC6, Leica, Wetzlar, Germany). These sections
   underwent staining with lead acetate and uranium dioxide for 15 min
   each. Observation and photography of the ultrastructure of chloroplasts
   were performed using a JEM 1400 transmission electron microscope (JEOL,
   Tokyo, Japan).

5.7. Photosynthetic Parameters Measurements

   Between 09:00 and 11:00, a portable photosynthesis system (LI-6400,
   LICOR, Lincoln, NE, USA) was utilized to measure the net photosynthetic
   rate (P[n]), stomatal conductance (G[s]), intercellular CO[2]
   concentration (C[i]), and transpiration rate (E). During the
   measurements, the CO[2] concentration was held at 400 μmol mol^−1, the
   leaf chamber temperature was set at 25 °C, and the light density in the
   sample chamber was set to 1000 μmol m^−2 s^−1 in the sample chamber.

   On the same day, between 10:00 and 12:00, Chl fluorescence
   characteristics were evaluated using a pulse amplitude modulation (PAM)
   fluorometer (Junior-PAM-II, Heinz Walz GmbH, Effeltrich, Germany),
   following the procedure described previously [[260]111]. Chl
   fluorescence parameters were recorded following a previously published
   method [[261]24]. Before measurements, the leaves were subjected to a
   30 min dark adaptation period. The photosynthetically active radiation
   (PAR) was maintained at 125 μmol m^−2 s^−1. To ensure that the Ft
   fluorescence level remained between 200 and 400 counts, adjustments
   were made to the measuring light intensity, ranging from 7 to 11, with
   a frequency set at 1 or 2. Using the WinControl-3 software (version
   3.32), parameters such as fluorescence origin (Fo), fluorescence
   maximum (Fm) and variable fluorescence (Fv = Fm − Fo), the maximum
   quantum yield of photosystem II (Fv/Fm), actual photosynthetic
   efficiency of photosystem II (ΦPSII), non-photochemical quenching
   (NPQ), regulatory energy dissipation quantum yield (Y (NPQ)), and
   non-regulatory energy dissipation quantum yield (Y (NO)) were
   automatically captured and calculated. Each biological replicate
   consisted of three leaves sampled from individual plants of each
   variety, with three randomly selected plants from both the wild and
   mutant types.

5.8. Plant Height, Biomass, Soluble Sugar, and Soluble Protein Measurements

   Plant height for each variety was measured in triplicate, with nine
   plants in each replicate. From each group, three randomly selected
   plants were excavated, washed with deionized water, and wiped clean to
   determine both fresh and dry mass. The fresh mass of each plant was
   recorded individually before being subjected to drying in an oven
   (Model 202-0B, Shangyi, Shanghai, China) at 65 °C for 12 h to obtain
   the dry mass for each treatment. Soluble sugar content was assessed via
   anthrone colorimetry [[262]112], whereas quantification of soluble
   protein content followed the Bradford method [[263]113], with bovine
   serum albumin serving as the standard.

5.9. RNA Extraction and Preparation of cDNA Library

   Total RNA extraction from the green leaves of WT and the yellow leaves
   of YT was carried out using the NEBNext^® UltraTM Plant RNA Extraction
   Kit (TaKaRa, Dalian, China). Each RNA sample underwent three biological
   replicates. Evaluation of RNA sample quantity and quality involved 1%
   agarose gel electrophoresis and a Nanodrop 1000 spectrophotometer
   (Nanodrop, Wilmington, DE, USA). Assessment of RNA integrity and
   concentration was performed using an Agilent 2100 Bioanalyzer (Agilent
   Technologies, Inc., Santa Clara, CA, USA). Subsequently, mRNA isolation
   was conducted using the NEBNext Poly (A) mRNA Magnetic Isolation Module
   (E7490; NEB, Ipswich, MA, USA), followed by fragmentation into 200 nt
   short RNA inserts. The fragmented mRNA underwent first and
   second-strand cDNA synthesis. Following end-repair, dA-tailing, and
   adaptor ligation on the double-stranded cDNA, PCR amplification was
   performed to enrich appropriate fragments.

5.10. Illumina Deep Sequencing and Data Analysis

   Sequencing of the cDNA libraries obtained from the samples was
   conducted on an Illumina NovaSeq 6000 platform (Illumina, San Diego,
   CA, USA). Processing the raw data (raw reads) involved the utilization
   of custom Perl scripts to filter out adaptors, N bases, and low-quality
   reads, resulting in the acquisition of clean reads. Quality metrics
   such as Q20, Q30, and GC content were assessed for the clean data.
   Following this, alignment of the clean reads to the I. latifolia
   reference genome ([264]https://ngdc.cncb.ac.cn/gwh, accessed on 1
   December 2023, GWHBIST00000000) was performed using HISAT software
   (version 2.0.4) with default parameters. Calculation of read counts for
   each gene was accomplished using HTSeq (version 0.6.1), and gene
   expression levels were quantified in terms of fragments per transcript
   kilobase per million fragments mapped (FPKM) values.

5.11. Identification and Functional Analysis of DEGs

   DESeq2 R software (version 1.20.0) was employed to detect DEGs in
   libraries derived from both green and yellow leaves. This software
   offers statistical procedures tailored for digital gene expression data
   analysis using a negative binomial distribution model. To control false
   discovery rates, the resulting p-values underwent adjustment using
   Benjamini and Hochberg’s method. Genes with a Padj ≤ 0.05 and
   |log[2](Fold Change)| ≥ 1 were deemed significantly differentially
   expressed. Additionally, we utilized the cluster Profiler R package
   (Version 3.5.0) to conduct enrichment analyses for Gene Ontology (GO)
   and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Gene Set
   Enrichment Analysis (GSEA) was performed to assess whether predefined
   gene sets exhibited significant and consistent differences between the
   two biological states. For GSEA, the local GSEA analysis tool
   ([265]http://www.broadinstitute.org/gsea/index.jsp, accessed on 5
   December 2023) was utilized with the KEGG dataset. Enrichment analysis
   revealed significant enrichment of all DEG sets, including both
   up-regulated and down-regulated DEGs, across all differential
   comparison combinations.

5.12. qRT-PCR Verification

   Gene-specific primers were designed with Primer Premier 6.0 (Premier
   Biosoft Inc., Palo Alto, CA, USA), utilizing the Actin gene as the
   internal reference [[266]114]. Total RNA was carried out using the
   RNAprep Pure Plant Plus Kit (#DP432, Tiangen, Beijing, China), followed
   by cDNA synthesis using the HiScript III 1st Strand cDNA Synthesis Kit
   (#R312-01, Vazyme, Nanjing, China). qRT-PCR was performed using AceQ
   Universal SYBR qPCR Master Mix (#Q511-02, Vazyme, Nanjing, China). Each
   sample underwent technical triplicates with three biological
   replicates. Relative expression levels of target genes were assessed
   using the 2^−ΔΔCt comparative Ct method [[267]115].

5.13. Statistical Analysis

   GraphPad Prism 9.0 software (GraphPad Inc., La Jolla, CA, USA) and SPSS
   19.0 software (SPSS Inc., Chicago, IL, USA) were employed for
   statistical analysis. Significance levels were determined via the
   t-test, where significance was denoted by asterisks as follows: **** p
   < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05, while non-significant
   results were labeled as “ns”. Error bars represent the standard
   deviation of the mean (SD). Image analyses were carried out using
   ImageJ ([268]https://imagej.net/ij/, accessed on 15 December 2023). For
   cytological images, 10 randomly selected views were analyzed.

Acknowledgments