Fruit abscission, occurring in advance of harvest, is characterized by the activation of a specific abscission zone (AZ) located between the pedicel and fruitlet, resulting in severe economic losses in many fruits [1,2]. The abscission is a highly regulated developmental process that is both influenced and activated by either internal cues or environmental conditions. Endogenous plant hormones have been implicated to contribute to the abscission of immature fruit, among which auxin is suggested to be a key regulator in fruit retention [3]. Spraying of auxins or synthetic auxins (e.g., 1-naphthaleneacetic acid (NAA) or 2,4-dichlorophenoxyacetic acid (2,4-D)) can efficiently reduce the fruit abscission by maintaining the cells adherence in the AZ and repress the activity of hydrolytic enzymes, such as cellulases which, in turn, promote fruitlet abscission by loosening the cell walls [4,5].

Auxin plays a pivotal role in the process of plant organ abscission. It is increasingly interesting to investigate the role of genes related to the auxin signal transduction pathway in this process. It is well documented that auxin stimulates the rapid transcription of a set of genes called primary or early auxin responsive genes including Aux/IAA, GH3, SAUR[6–8]. The AUX/IAA proteins are short-lived nuclear proteins that function as repressors of auxin-induced transcription, and auxin promotes the degradation of this large family of transcriptional regulators, leading to diverse downstream responses [9,10]. GH3 genes encode IAA-conjugating enzymes, which could act as feedback regulators by reducing free auxin levels [11], while SAUR genes code for highly conserved short-lived small proteins and likely function in cell expansion, hypocotyls and stamen filament elongation [12,13]. The promoter regions of these primary/early auxin-responsive genes contain one or more auxin-responsive elements (AuxRE) conferring auxin responsiveness. The identification of the AuxRE sequence led to the isolation of the Arabidopsis Auxin Response Factor1 (ARF1) gene which was shown to bind with specificity to AuxRE [14,15]. ARF transcription factors regulate many responses to auxin, including embryo patterning, vascular formation, phototropism and gravitropism [16,17]. Recently, it was found that expression of Ce-IAA in the AZ was negatively correlated with floret bud abscission in Cestrum cut flower [18]. Similar correlation between the effectiveness of auxin in delaying abscission and induction of Aux/IAA gene expression were also found in Cestrum elegans[19] and Mirabilis jalapa[20]. In tomato, SlIAA genes play important roles in delaying floral pedicel abscission, with SlIAA1, 9 and 12 required for the completion of ethylene-induced abscission and SlIAA4, 6, 7, 8, 14, 16, 17 and 29 important for calcium-delayed abscission [21]. Tomato GH3 effectively represses the delayed abscission induced by IAA, while SAUR may serve as a marker of IAA level throughout the abscission process [21]. In Arabidopsis, two ARF genes, ARF1 and ARF2, along with other ARF genes, Nonphototropic Hypocotyl4 (NPH4)/ARF7 and ARF19, might act together to regulate senescence and floral organ abscission [22].

The substantial production of Litchi (Litchi chinensis Sonn.), an important economic fruit crop widely grown in southern China and other southeast Asian areas [23], has been challenged by excessive fruit drop, one of the major factors causing irregular cropping or alternate bearing. There are two or three distinct abscission periods depending on cultivar during each reproductive cycle, typically a year [24]. The first period of abscission occurs at the end of full female bloom and lasts for about a month; the second abscission wave starts one or two weeks later and lasts for about two weeks [25]. The third drop wave, lasting for about one week, takes place two to three weeks before fruit maturation [24]. Generally, only 2% to 5% of the initial female flowers develop into mature fruits [25]. The peak and periodicity of fruit drop are affected by environmental and/or internal factors, for instance, spells of heavy overcast and rainy weather. Low temperature or drought during the flowering and fruit-setting period always induces sever fruit abscission. Yuan and Huang (1988) demonstrated that litchi fruit-set relied greatly upon current photosynthates and the fruit abscission multiwaves were in parallel with the upsurges of abscissic acid (ABA) in the seed, while IAA seemed to be antagonistic [26]. Basipetal IAA flux through AZ prevents abscission by rendering the AZ insensitive to ethylene [20,27]. Recently, microarray analysis revealed that expressions of abscission-related genes were initiated by flower removal and leading to pedicel abscission were inhibited by IAA application, confirming that ethylene sensitivity in the AZ is associated with altered expression of auxin-regulated genes resulting from auxin depletion [28]. In addition, auxin can also influence carbohydrate availability to affect fruit abscission. For example, application of synthetic auxin 3,5,6-TPA (3,5,6-trichloro-2-pyridil-oxyacetic acid) stimulated carbohydrate accumulation in developing fruitlets to reduce the dropping of citrus fruit [29]. Synthetic auxins including NAA, 2,4-D, 2,4,5-T (2,4,5-trichlorophenoxyacetic acid) and 2,4,5-TP (2-(2,4,5-trichlorophenoxy) propionic acid), were also widely used to prevent fruit abscission in litchi management practices [30,31]. As auxin response is known to be mediated in part by altering gene expression, characterization of genes related to the auxin signal transduction pathway is of particular significance for understanding the molecular basis of auxin action [18,20,22].

Therefore, in this study, genes related to auxin signaling were firstly isolated from litchi fruit, and their expression patterns in response to girdling plus defoliation treatment—an approach to completely block the carbohydrate transport to fruitlets and induce severe carbohydrate stress for fruit development [32,33]—were investigated. Our results expand our understanding of carbohydrate stress-induced fruitlet abscission at the molecular level and have theoretical implications for exploring effective strategies to regulate litchi fruit abscission.

Like many fruit trees species, litchi trees blossom with high profusion and thereafter exhibit massive fruitlet abscission, which leads to a low yield problem, thus limiting litchi industry development [42]. This process is generally under hormonal and metabolic regulation [43]. There is circumstantial evidence that developing fruitlets compete for photoassimilates, and the fruit abscission multiwaves are in parallel with the upsurges of abscissic acid (ABA), while IAA seems to be antagonistic [26]. Although the nature of the hormonal signals that link carbohydrate shortage and abscission are still unknown, it is commonly believed that the endogenous balance of ethylene and auxin in the AZ affects organ abscission. Application of exogenous ethylene accelerates abscission, whereas auxin treatment decreases cell sensitivity to ethylene and delays or inhibits the process [44]. Auxin exerts its effects by regulating the expression of numerous genes, including early auxin responsive genes and ARFs. In this paper, the isolation of LcAUX/IAA1, LcGH3.1, LcSAUR1 and LcARFs from litchi fruit was firstly reported. LcAUX/IAA1, LcGH3.1, LcSAUR1, LcARF1 and LcARF2 were predicted to encode proteins of 203, 613, 142, 792 and 832 amino acids, with calculated molecular weights of 22.67, 69.20, 11.40, 88.20 and 93.16 kDa, respectively. The genes isolated have shown the highest homology to the corresponding genes: LcAUX/IAA: 70% amino acid identity with RcIAA1 (XP_002517529.1 from castor bean); LcGH3.1: 87% amino acid identity with RcGH3.6 (XP_002533739.1 from castor bean); LcSAUR1: 49% amino acid identify with PtSAUR (XP_002305471.1 from Populus trichocarpa); LcARF1: 79% amino acid identity with VvARF4 (XP_002285019.2 from grape); LcARF2: 78% amino acid identity with VvARF8 (XP_002266678.2 from grape). LcAUX/IAA1, like its homologs from other plant species, contains all the four conserved domains (I, II, III and IV). It is generally accepted that domain I has a repressor function [15], and domain II contributes to the rapid degradation of the Aux/IAA proteins, while domains III and IV are responsible for homo- and heterodimerization among the various members of the Aux/IAA and auxin response factor (ARF) proteins [45]. LcAUX/IAA1 may be nuclear-localized proteins due to the existence of two types of NLS (Figure 1). The protein encoded by LcGH3.1 gene conserves three motifs essential for conjugating amino acid to excessive IAA (Figure 2) [11]. Apart from IAA, some GH3 genes from Arabidopsis are also known to adenylate other plant hormones such as jasmonic acid and salicylic acid [35]. The deduced ORF of LcSAUR1 encodes small proteins with molecular mass of 11.40 kDa, similar with those of OsSAURs, which ranged from 10 kDa for OsSAUR1 to 27 kDa for OsSAUR21 (Figure 3) [46]. The multiple sequence alignments of the full-length protein sequences showed that LcSAUR1 contains a conserved, SAUR-specific domain (SSD) of approximately 60 residues in the central region. In addition, alignment of the LcARFs amino acid sequences with that of other reported ARFs revealed that the predicted proteins shared typical features including an N-terminal DNA-binding domain and two C-terminal domains, “Box III” and “Box IV” (Figure 4), which contribute to protein dimerization [14,15].

It has been proposed that carbohydrates and hormones participate in a complex signal transduction system [33,47]. Previous work demonstrated that defoliation and girdling management could reduce the amount of resources available for current reproduction in individual branches, resulting in decrease of fruit production [32]. In addition, defoliation or shading reduces photosynthesis but also enhanced levels of abscisic acid (ABA) and 1-aminocyclo-propane-1-carboxylic acid (ACC) that positively correlated with abscission [26,47]. In the present work, we demonstrated that treatment of girdling plus defoliation reduced endogenous IAA concentration concomitant with increasing fruitlet abscission (Figures 5 and 6) and that, in litchi fruit, auxin has a key role in fruit retention. None of the litchi early auxin responsive genes or ARFs was AZ specific and they were also detected in the whole fruitlet. The expression patterns of auxin-related genes differed during litchi fruit AZ activation. The expression of LcAUX/IAA1, LcGH3.1 and LcSAUR1 increased in AZ and fruitlet after treatment with girdling plus defoliation which significantly induced litchi fruitlet shedding (Figure 7). By comparing the expression levels of the LcAUX/IAA1, LcGH3.1 and LcSAUR1 in the different tissues and IAA content, it is likely that the LcAUX/IAA1 and LcSAUR1 may play a more important role in abscission than the LcGH3.1, because the LcAUX/IAA1 and the LcSAUR1 were expressed mostly in AZ. However, these results differed from the finding of tomato flower pedicel abscission. During the first 4 h after flower excision, most SlIAA genes in the control reached their peak expression level at 1 h but then decreased sharply, which corresponded well to the changes of endogenous IAA level [21]. Moreover, tomato GH3 increased slightly after 8 h and maintained a low expression level during abscission, implying that it may be an effective negative regulator in IAA-induced delay in abscission, while small auxin-up RNA was suggested to be a marker of IAA level throughout the abscission process [21]. Additionally, application of 2,4-D, which delayed floret abscission, induced a higher expression of Aux/IAA genes in the floret AZ [21]. Our findings, together with previous reports, suggest that functional divergence underlying abscission processes may be present between litchi and tomato. Thus, further studies are needed to fully unravel the function of LcAUX/IAA1 and LcSAUR1 genes in regulating fruit abscission.

LcARF1 and LcARF2 exhibited similar expression patterns both in AZ and fruitlet tissues (Figure 7). In particular, LcARF1 was negatively correlated with fruitlet abscission in that its transcript decreased in AZ and fruitlet under the treatment of girdling and defoliation, which was contrary to the work of [22] showing that an increase of ARF2 mRNA in detached leaves in Arabidopsis thaliana following dark treatment, with a time course similar to that of SAG12 mRNA. Interestingly, the expression pattern in AZ during abscission of LcARF2 was decreased at 1 days and then increased at 2 days and finally decreased again at 4 days, which is opposite of that of LcAUX/IAA1, suggesting that a cascade of AUX/IAA-ARF might also exist in litchi fruit. The developmental specificity of auxin response is determined by the interacting pairs of ARFs and Aux/IAAs [48]. Among the auxin-signaling, AUX/IAAs are proposed to bind to ARFs and inhibit transcription of ARF target genes, while a degradation of AUX/IAAs in response to auxin treatment would lead to increase ARF activity [49–51]. It should be needed to examine whether LcAUX/IAA1 is able to interact with LcARF2 in the future.

In conclusion, one AUX/IAA (LcAUX/IAA1), one GH3 (LcGH3.1), one SAUR (LcSAUR1) and two ARF (LcARF1 and LcARF2) genes were isolated and characterized from litchi fruit for the first time. Expression studies of these genes revealed different expression patterns in response to girdling plus defoliation application which induced fruitlet abscission of litchi and reduced endogenous IAA content. By comparing these genes’ expression levels, IAA content and fruitlet abscission, it was postulated that LcAUX/IAA1, LcSAUR1 and LcARF1 may play a more important role in the abscission of litchi fruitlet. These results provide novel molecular insight into carbohydrate stress-induced fruitlet abscission and the auxin signal transduction pathway involved in litchi fruitlet abscission.