The authors have declared that no competing interests exist.
Conceived and designed the experiments: L. Shalom AS. Performed the experiments: L. Shalom SS NZ L. Shlizerman HZ. Analyzed the data: L. Shalom RO AS. Contributed reagents/materials/analysis tools: EB. Wrote the paper: L. Shalom AS.
Alternate bearing (AB) is the process in fruit trees by which cycles of heavy yield (ON crop) one year are followed by a light yield (OFF crop) the next. Heavy yield usually reduces flowering intensity the following year. Despite its agricultural importance, how the developing crop influences the following year's return bloom and yield is not fully understood. It might be assumed that an ‘AB signal’ is generated in the fruit, or in another organ that senses fruit presence, and moves into the bud to determine its fate—flowering or vegetative growth. The bud then responds to fruit presence by altering regulatory and metabolic pathways. Determining these pathways, and when they are altered, might indicate the nature of this putative AB signal. We studied bud morphology, the expression of flowering control genes, and global gene expression in ON- and OFF-crop buds. In May, shortly after flowering and fruit set, OFF-crop buds were already significantly longer than ON-crop buds. The number of differentially expressed genes was higher in May than at the other tested time points. Processes differentially expressed between ON- and OFF-crop trees included key metabolic and regulatory pathways, such as photosynthesis and secondary metabolism. The expression of genes of trehalose metabolism and flavonoid metabolism was validated by nCounter technology, and the latter was confirmed by metabolomic analysis. Among genes induced in OFF-crop trees was one homologous to
Alternate bearing (AB) is the process by which cycles of heavy yield (ON crop) one year are followed by a light yield (OFF crop) the next (reviewed in
The mechanism(s) by which the developing crop influences return bloom and yield the following year is not fully understood. Two hypotheses have been suggested. The “nutritional” hypothesis holds that return bloom and yield are proportional to tree carbohydrate status. Lack of carbohydrate in the ON year directly or indirectly reduces flowering the following year
The floral induction period in citrus starts in mid-November and lasts until approximately the end of December to mid-January (
Fruit presence inhibits return flowering. However, it is not clear at which stage the fruit exerts its inhibitory effect: at flowering induction, transition of the shoot apical meristem to floral meristem, or subsequent stages of floral development and bud break. Moreover, the nature of the signal (‘AB signal’) and the organ or tissue from which it originates, be it the fruit itself or the leaf which senses fruit presence, are not known. Regardless of the source tissue for the AB signal, it must be received, directly or indirectly, at the bud, and more specifically, at the apical meristem which has to “decide” whether to develop into an inflorescence or remain a vegetative meristem. Therefore, following perception of the signal, the bud must undergo a series of events which depend on fruit load. In the current work, we analyzed changes in global gene expression during bud development in ON and OFF trees, to identify metabolic and controlling pathways that play a role in bud fate. To determine the earliest time point for the transcriptome analysis, we first analyzed changes in bud morphology during its development, and changes in the expression of key flowering control genes. Based on those results, global gene-expression analysis was carried out at a few key time points in the buds, which receive the ‘AB signal’, and in leaves and stems, which might play a role in generating and transporting the signal.
Plant material was collected from a commercial orchard of 10-year-old Murcott mandarin (
Buds were collected and fixed in an FAA solution [10 formaldehyde:5 acetic acid:85 ethanol (70%), v/v]. Fixation was followed by an ethanol dilution series and subsequent stepwise exchange of ethanol with Histoclear (xylem substitute). Samples were embedded in paraffin and cut by microtome (Leica RM2245) into 12-µm sections. Sections were stained with safranin and fast green
Total RNA was extracted from buds and from leaves and stems (LS) using the CTAB extraction method
The RNA levels of trehalose biosynthetic genes, flavonoid biosynthetic genes,
For global gene expression, the citrus GeneChip (Affymetrix, Inc., Santa Clara, CA) carrying 30,171 probes was used. The array is estimated to represent about 15,500 genes. RNA samples were processed as recommended by the Affymetrix GeneChip Expression Analysis Technical Manual at the Center for Genomic Technologies of the Hebrew University of Jerusalem. Total RNA was quantified and then adjusted to a final concentration of 1 µg/µl. Single-stranded and then double-stranded cDNA was synthesized from total RNA (0.5 µg total RNA for each reaction) using oligo-dT primer and the Affymetrix One-Cycle Labeling Kit and control reagents. The resulting double-stranded cDNA was column-purified and then used as a template to generate biotin-tagged cRNA from an
Gene ontology (GO) analysis was performed using the AgriGo interface (
Buds (200 to 300 mg) were pulverized in liquid nitrogen using a mortar and pestle and the powder was transferred to a 15-ml tube. Three volumes of water-saturated n-butanol were added, and the mixture was vortexed for a few minutes, then incubated under shaking (200 rpm) for 12 h at room temperature. Following short centrifugation (15,000 RPM at room temperature) and phase separation, the upper phase was collected into a fresh tube, and incubated at room temperature for 1 h to allow the butanol to evaporate. Samples were filtered through a Millex-HV Durapore (PVDF) membrane (0.22 µm) before injection into the LC-MS instrument. MS analyses were carried out by the ultraperformance LC-quadrupole time of flight (UPLC-QTOF) instrument (Waters Premier QTOF, Milford, USA), with the UPLC column connected on-line to a PDA detector (Waters Acquity), and then to the MS detector equipped with an electrospray ion (ESI) source (performed in ESI-positive mode). Separation was performed on a 2.1×50 mm i.d., 1.7-µm UPLC BEH C18 column (Waters Acquity). The chromatographic and MS parameters were as follows: the mobile phase consisted of 0.1% formic acid in water (phase A) and 0.1% formic acid in acetonitrile (phase B). The linear gradient program was: 100% to 95% A over 0.1 min, 95% to 5% A over 9.7 min, held at 5% A for 3.2 min, then returned to the initial conditions (95% A) in 4.2 min. The flow rate was 0.3 ml/min, and the column temperature was kept at 35°C. Masses of the eluted compounds were detected with a QTOF Premier MS instrument. The following settings were applied during the UPLC-MS runs: capillary voltage of 3.2 kV, cone voltage of 30 eV, collision energy of 5 eV, and argon as the collision gas. The following settings were applied during the UPLC-MS/MS run: capillary spray of 3.2 kV, cone voltage of 30 eV, collision energies of 15 to 25 eV, and argon as the collision gas. The m/z range was 70 to 1,000 D. The MS system was calibrated using sodium formate, and Leu-enkephalin was used as the lock mass. MassLynx software version 4.1 (Waters Inc.) was used to control the instrument and calculate accurate masses.
ANOVA test for qPCR results, bud measurements and metabolomic data was conducted using the JMP® version 10 software (SAS Institute Inc. Cary, NC).
Normally, fruit load status in an AB variety is similar among most of the trees in an orchard in a given year, i.e., most trees either bear a heavy crop (ON-crop year) or a low crop (OFF-crop year). A few trees, however, show the opposite trend, allowing the collection of samples from both AB states from nearby trees. Buds, leaves and stems of heavy-loaded and low-loaded Murcott trees from the same orchard were collected from May, soon after fruit set, until January, the end of the flowering induction period. Flowering intensity of these trees was assessed the following spring (
Vegetative shoots, generative inflorescences containing only flower buds, and mixed inflorescences containing flowers and leaves, were counted during flowering peak in trees which carried heavy yield (ON) and light yield (OFF) during the previous year. Mean number of three biological replicates ± SE. Stars denote a significant difference between ON and OFF buds (P<0.05).
Buds were collected from ON- and OFF-crop trees in mid-July (A) and during the indicated months (B), fixed, dissected, dyed and photographed. Bud width and length were measured following photography (B). Mean values of 50 buds ± SE. Stars denote a significant difference between ON and OFF buds during the same time point (P<0.02).
The mRNA levels of key flowering genes were measured in buds of ON- and OFF-crop trees at a few time points: mid-May—immediately after fruit set, mid-July—1 month after natural fruit thinning (June drop), and mid-September—the last time point at which fruit removal during an ON-year reverses the AB trend. In addition, samples were collected from mid-November until mid-January, considered the flowering induction period (
mRNA levels of the indicated genes in ON and OFF buds were determined by real-time PCR during the indicated months. Mean values of three independent biological replicates ± SE. Stars denote a significant difference between the expression of the gene in ON and OFF buds during the same time point (P<0.05).
The mRNA levels of
In May, the transcript levels of
The expression of
The expression of
The above results showed that there was already a clear difference in the sizes of ON and OFF buds in May. Moreover, the mRNA levels of four key flowering control genes,
Transcriptome analysis was carried out with the above samples using the Citrus Genome Array (Affymetrix) containing 30,171 probes and estimated to represent about 15,500 genes (
Hierarchical cluster analysis of global gene expression in buds and leaves+stems (LS) in ON-crop (On) and OFF-crop (Off) trees at the indicated times (A). Venn diagrams of differentially expressed probes, induced in buds and LS of ON- and OFF-crop trees during the indicated months.
As a comparison of the two AB states was the main target of this study, the number of DEPs in ON and OFF buds and ON and OFF LS at the various time points is presented in
Considering the high number of DEPs in May relative to the other time points, GO and other analyses were performed only for ON and OFF trees in May. Overall, 1767 (out of 2205) and 2359 (out of 3087) DEPs induced in the ON year and OFF year, respectively, were GO annotated (
None of the biological processes which were induced in OFF LS could be identified by SEA—only those in the buds were (
Differentially expressed probes were analyzed using MapMan. Blue squares represent genes induced in ON buds and red squares represent genes induced in OFF buds. A description of the specific genes and their fold change is provided in
GO term | Description | % in input list | % in BG/Ref | p-value | FDR | Fold enrichment |
GO:0019748 | secondary metabolic process | 5.5336 | 1.2429 | 7.00E-10 | 6.60E-07 | 4.45 |
GO:0009698 | phenylpropanoid metabolic process | 3.1621 | 0.5005 | 2.20E-08 | 6.70E-06 | 6.32 |
GO:0009812 | flavonoid metabolic process | 2.5692 | 0.3149 | 2.00E-08 | 6.70E-06 | 8.16 |
GO:0009699 | phenylpropanoid biosynthetic process | 2.5692 | 0.4707 | 2.20E-06 | 0.00052 | 5.46 |
GO:0009813 | flavonoid biosynthetic process | 1.9763 | 0.3016 | 6.30E-06 | 0.0011 | 6.55 |
GO:0010017 | red or far-red light signaling pathway | 1.1858 | 0.0895 | 7.30E-06 | 0.0011 | 13.25 |
GO:0009639 | response to red or far-red light | 1.7787 | 0.2519 | 9.70E-06 | 0.0013 | 7.06 |
GO:0009821 | alkaloid biosynthetic process | 1.1858 | 0.1061 | 2.10E-05 | 0.0024 | 11.18 |
GO:0009820 | alkaloid metabolic process | 1.9763 | 0.3546 | 2.60E-05 | 0.0027 | 5.57 |
GO:0019438 | aromatic compound biosynthetic process | 2.9644 | 0.8120 | 5.10E-05 | 0.0048 | 3.65 |
GO:0009585 | red, far-red light phototransduction | 0.9881 | 0.0795 | 6.00E-05 | 0.0051 | 12.42 |
GO:0009583 | detection of light stimulus | 0.9881 | 0.0862 | 9.00E-05 | 0.0065 | 11.47 |
GO:0007602 | phototransduction | 0.9881 | 0.0862 | 9.00E-05 | 0.0065 | 11.47 |
GO:0006725 | cellular aromatic compound metabolic process | 4.5455 | 1.7699 | 0.00017 | 0.011 | 2.57 |
GO:0009582 | detection of abiotic stimulus | 0.9881 | 0.0994 | 0.00018 | 0.011 | 9.94 |
GO:0009581 | detection of external stimulus | 0.9881 | 0.1027 | 0.00022 | 0.013 | 9.62 |
GO:0051716 | cellular response to stimulus | 3.3597 | 1.1435 | 0.00024 | 0.013 | 2.94 |
GO:0042398 | cellular amino acid derivative biosynthetic process | 2.7668 | 0.8584 | 0.00032 | 0.017 | 3.22 |
GO:0009809 | lignin biosynthetic process | 1.1858 | 0.1823 | 0.00046 | 0.022 | 6.50 |
GO:0009808 | lignin metabolic process | 1.1858 | 0.1823 | 0.00046 | 0.022 | 6.50 |
GO:0006575 | cellular amino acid derivative metabolic process | 3.3597 | 1.2495 | 0.00064 | 0.028 | 2.69 |
GO:0009628 | response to abiotic stimulus | 4.7431 | 2.0815 | 0.00067 | 0.028 | 2.28 |
GO:0009791 | post-embryonic development | 2.5692 | 0.8352 | 0.00078 | 0.03 | 3.08 |
GO:0009416 | response to light stimulus | 2.5692 | 0.8319 | 0.00075 | 0.03 | 3.09 |
GO:0051606 | detection of stimulus | 0.9881 | 0.1359 | 0.00082 | 0.031 | 7.27 |
GO:0015979 | photosynthesis | 1.7787 | 0.4574 | 0.00096 | 0.034 | 3.89 |
GO:0009314 | response to radiation | 2.5692 | 0.8651 | 0.0011 | 0.037 | 2.97 |
GO:0009805 | coumarin biosynthetic process | 0.9881 | 0.1525 | 0.0014 | 0.045 | 6.48 |
GO:0009804 | coumarin metabolic process | 0.9881 | 0.1525 | 0.0014 | 0.045 | 6.48 |
BG, background; Ref, reference; FDR, false discovery rate.
Only one process was induced in LS during the ON year—Cell Wall Organization (GO:0009664)—which showed about 18-fold enrichment (
GO term | Description | % in input list | % in BG/Ref | p-value | FDR | Fold enrichment |
GO:0010252 | auxin homeostasis | 0.9804 | 0.0331 | 6.10E-07 | 0.0005 | 29.58 |
GO:0006073 | cellular glucan metabolic process | 3.1373 | 0.7292 | 7.80E-06 | 0.0011 | 4.30 |
GO:0009312 | oligosaccharide biosynthetic process | 1.1765 | 0.0829 | 6.50E-06 | 0.0011 | 14.20 |
GO:0044042 | glucan metabolic process | 3.1373 | 0.7292 | 7.80E-06 | 0.0011 | 4.30 |
GO:0005992 | trehalose biosynthetic process | 0.9804 | 0.0464 | 4.50E-06 | 0.0011 | 21.13 |
GO:0046351 | disaccharide biosynthetic process | 1.1765 | 0.0762 | 3.80E-06 | 0.0011 | 15.43 |
GO:0044264 | cellular polysaccharide metabolic process | 3.1373 | 0.7524 | 1.20E-05 | 0.0014 | 4.17 |
GO:0005976 | polysaccharide metabolic process | 3.5294 | 0.9380 | 1.40E-05 | 0.0014 | 3.76 |
GO:0005991 | trehalose metabolic process | 0.9804 | 0.0597 | 1.80E-05 | 0.0017 | 16.43 |
GO:0005984 | disaccharide metabolic process | 2.5490 | 0.5966 | 5.60E-05 | 0.0046 | 4.27 |
GO:0009311 | oligosaccharide metabolic process | 2.5490 | 0.6099 | 7.00E-05 | 0.0053 | 4.18 |
GO:0042221 | response to chemical stimulus | 6.8627 | 2.9598 | 0.00011 | 0.0079 | 2.32 |
GO:0016137 | glycoside metabolic process | 2.5490 | 0.6529 | 0.00014 | 0.0088 | 3.90 |
GO:0005985 | sucrose metabolic process | 2.3529 | 0.5734 | 0.00015 | 0.0091 | 4.10 |
GO:0005982 | starch metabolic process | 2.3529 | 0.6032 | 0.00025 | 0.014 | 3.90 |
GO:0044262 | cellular carbohydrate metabolic process | 4.3137 | 1.6075 | 0.00027 | 0.014 | 2.68 |
GO:0016138 | glycoside biosynthetic process | 1.1765 | 0.1624 | 0.00035 | 0.017 | 7.24 |
GO:0010035 | response to inorganic substance | 1.5686 | 0.3281 | 0.00068 | 0.031 | 4.78 |
GO:0010038 | response to metal ion | 1.3725 | 0.2585 | 0.00078 | 0.034 | 5.31 |
GO:0009733 | response to auxin stimulus | 1.9608 | 0.5204 | 0.001 | 0.042 | 3.77 |
BG, background; Ref, reference; FDR, false discovery rate.
As already mentioned, an
Fold change (FC) between OFF and ON buds and leaves+stems (LS) of microarray probe Cit corresponding to
Global gene-expression analysis showed that probes encoding trehalose metabolism enzymes are induced in ON buds (
mRNA levels of trehalose phosphate phosphatase (TPP) and trehalose phosphate synthase (TPS) were measured in ON and OFF buds in May. Fold change (FC) between ON and OFF buds in their corresponding microarray probes are shown in the lower panel. Mean number of three biological replicates ± SE. Stars denote a significant difference in the expression of the gene between ON and OFF buds (P<0.05).
Probes for six genes of the flavonoid metabolic pathway, 4-coumarate:coenzyme A ligase (
The mRNA levels of phenylalanine ammonia-lyase (PAL), chalcone synthase (CHS), cinnamate 4-hydroxylase (
The flavonoid biosynthetic pathway was further investigated by metabolomic analysis of a few flavonoids in ON and OFF buds during May using UPLC-QTOF-MS. The following compounds were identified by accurate mass, fragmentation pattern and a few standards: naringin/narirutin, hesperidin/neohesperidin, poncirin/didymin (flavonones), diosmin (flavone). In agreement with the gene-expression analyses, the intensities of all tested compounds were higher in OFF buds than ON buds, although with varied significance (
The indicated flavonoids were measured in ON and OFF buds using LC-MS. Y axis indicates the intensity of each compound, RU relative units. Mean number of three biological replicates ± SE. Stars denote a significant difference between ON and OFF buds (P<0.05).
The effect of year 1 yield on the return bloom of year 2 was as expected: heavy yield resulted in a lower number of flowers and higher number of vegetative buds, whereas the opposite was true following a light yield (
In general, bud morphology and anatomy did not change significantly from May to January. This is in agreement with Lord and Eckerd's
To date, the expression of flowering control genes has been mostly investigated in leaves and stems. To the best of our knowledge, this work provides the only report describing the expression of flowering control genes in citrus buds. We recently demonstrated that during the flowering induction period, the mRNA levels of
One of the outcomes of the genomic analysis, validated by real-time PCR, was the induction in OFF buds and LS of the
The clustering analysis demonstrated that the impact of the three tested conditions, time, tissue type and AB state, on the level of similarity between the expression profiles follows the order: AB state>tissue type>time, i.e., ON and OFF organs showed more similar patterns than under the effect of time or tissue type (
Among the three analyzed time points, the largest number of DEPs between ON and OFF trees was detected in May, while in September and between November and January, their number was relatively low. Moreover, in May, the number of DEPs was much higher in buds than in LS. Taken together with the finding that the maximal difference in bud width develops between May and July, these results suggest that AB signal, if present, is generated much earlier than the flowering induction period, and causes the above changes. Alternatively, changes in gene expression in May might reflect changes in resource allocation: when fruit is present (ON trees), buds are deprived of photoassimilates, which directly reduces their size in comparison to the case in OFF trees.
As already noted, the highest number of DEPs was evident during May in the buds. Obviously, we cannot cover all of the metabolic processes which are induced at this time point, and we therefore briefly discuss three of these processes: two of them, flavonoid biosynthesis and photosynthesis, are induced in OFF buds, and one, trehalose metabolism, is induced in ON buds. The expression of genes belonging to two of these processes, flavonoid biosynthesis and trehalose metabolism, were also validated by nCounter technology.
ON buds showed increased expression of the two genes of trehalose metabolism,
Genes belonging to the three components of photosynthesis—light reactions, Calvin cycle and photorespiration—are induced in OFF buds (
Genes of a few pathways of secondary metabolism were induced in OFF buds, including flavonoids, phenylpropanoids, alkaloids and lignin (
In summary, results of this work show that a relatively long time before the flowering induction period, fruit load affects many regulatory and metabolic processes in the bud. Obviously, it should be considered that this and other conclusions of the work are based on a single cropping year. Although the expression of some of the flowering control genes was partially investigated during another year, with similar results (data not shown), year to year environmental and other external variations might affect the results, and therefore the conclusions. It should also be mentioned that the nature of the AB signal, and whether it is produced that early, remain open questions. Even if produced in May, or earlier, the signal must be reversible, as fruit thinning or complete removal from ON trees reverses the AB state. Ongoing studies in our laboratory include analyses of buds following fruit removal in September, when the number of differentially expressed genes between ON and OFF buds was very low. These analyses are expected to clarify which of the processes induced in ON and OFF buds are directly affected by fruit load. We are also investigating the possibility of
(TIF)
(TIF)
(TIF)
(TIF)
(TIF)
(TIF)
(XLSX)
(XLSX)
(XLSX)
(XLSX)
(XLSX)
(XLSX)