Yield associated traits correlate with cytokinin profiles in developing pods and seeds of field-grown soybean cultivars

Field Crops Research 214 (2017) 175–184

Contents lists available at ScienceDirect

Field Crops Research journal homepage: www.elsevier.com/locate/fcr

Yield associated traits correlate with cytokinin profiles in developing pods and seeds of field-grown soybean cultivars

MARK

Shrikaar Kambhampatia, Leonid V. Kurepina, Anna B. Kisialab, Kahlan E. Bruceb, Elroy R. Coberc, ⁎ Malcolm J. Morrisonc, R.J. Neil Emeryb, a b c

Department of Biology, University of Western Ontario, 1151 Richmond Street, London, ON N6A 5B7, Canada Biology Department, Trent University, 2140 East Bank Drive, Peterborough, ON K9J 7B8, Canada Ottawa Research and Development Centre, Agriculture and Agri-Food Canada, 960 Carling Ave, Ottawa, ON K1A 0C6, Canada

A R T I C L E I N F O

A B S T R A C T

Keywords: Cytokinin Developmental stages Seed set Seed filling Soybean Thousand seed weight Yield

While lab and greenhouse based studies have long indicated that cytokinins (CK) promote yield increases in soybean (Glycine max L.), it is not known if the relationship would be valid under more complex field conditions. Thus, an ambitious CK metabolite analysis was undertaken involving long-term field trials of commercial and historical soybean varieties to determine if differences between CK profiles are related to variation in yield performance. Twenty-seven cultivars were evaluated in this study representing a wide range of performance and assessed for 12 agronomically important plant and seed parameters under field conditions. Identification and quantification of 14 forms of CK was undertaken by high-performance liquid chromatography tandem mass spectrometry (HPLC–MS/MS) at three stages of reproductive development that are critical for yield determination (R4–R6). This revealed a substantial increase in CK levels, especially the highly metabolically active free bases, in the high yielding cultivars of soybean. Significant correlations between yield components and the most active CK form, tZ (trans-Zeatin), were detected at both R4 and R5 stages. A similar trend was observed for cZ (cis-Zeatin), indicating a possible role of both zeatin isomers in pod and early seed development. Positive, significant relationships between yield and cytokinins were maintained also at R6 stage; however, a switch in hormone profiles and increased levels of isopentenyl adenine (iP) types of CK in high yielding cultivars suggested that the presence of iP derivatives allowed developing seeds to maintain their active role in sink organs and attract assimilates during the seed filling phases, when the metabolism of the maturing plant was generally slowing down. Results suggested that cytokinin metabolites or their associated genes, may serve as a valuable, early indicators of yield performance in marker-assisted breeding programs for soybean or be manipulated through gene editing techniques.

1. Introduction Plant hormones are signal molecules that are produced within the plant and regulate critical developmental processes. A group of hormones, the cytokinins (CKs), play a major role in cell division and differentiation. CKs control a set of metabolic processes like flowering, fruit set, seed filling and many other related functions including inhibition of senescence (Mok and Mok, 2001; Jameson and Song, 2016). In terms of crop plants, they most critically regulate resource partitioning among different plant organs, a phenomenon referred to as source-sink relationships. As the amount of CKs rises in the sink organs, the demand for nutrients increases leading to an increase in sink strength and accumulation of assimilates (Emery et al., 2000; Götz

et al., 2007). Seeds are major organ sinks that impact economic performance since their elevated sink strength leads to a greater seed set and filling that is in turn reflected in higher yields (Brugière et al., 2008). Regulation of the source-sink relationship is a major function of CKs that the current research aims to elucidate. Structurally, CKs are adenine derivatives that are classified as isoprenoid or aromatic depending on the type of side chain substitute at the N6 position. Four major types of isoprenoid CKs occur in nature, namely isopentenyl adenine (iP), trans-zeatin (tZ), cis-zeatin (cZ) and dihydrozeatin (DZ) (Kamada-Nobusada and Sakakibara, 2009; Spichal, 2012; Yamburenko et al., 2017). Although, many of these are generally considered as bioactive CK types (Hirose et al., 2008), the occurrence and the relative physiological activity of the different forms of CKs can

⁎

Corresponding author. E-mail addresses: [email protected] (S. Kambhampati), [email protected] (L.V. Kurepin), [email protected] (A.B. Kisiala), [email protected] (K.E. Bruce), [email protected] (E.R. Cober), [email protected] (M.J. Morrison), [email protected] (R.J.N. Emery). http://dx.doi.org/10.1016/j.fcr.2017.09.009 Received 17 July 2017; Received in revised form 5 September 2017; Accepted 7 September 2017 Available online 17 September 2017 0378-4290/ © 2017 Elsevier B.V. All rights reserved.

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

family members expressed during seed development (Jameson and Song, 2016) and potential linkages to molecular markers such as has been reported in rice (Ashikari et al., 2005).

significantly vary among different plant species and among developmental stages (Emery et al., 2000). For example, cis-isomers (cZR (zeatin riboside), and cZRNT (zeatin nucleotide)), along with their isopentenyl precursors were found to be the major CK forms in developing embryos of pea (Pisum sativum). Additionally, the total CK quantities peaked during cell division and closely correlated with high rates of sugar metabolism (Quesnelle and Emery, 2007). In barley (Hordeum vulgare), Powell et al. (2013) studied field growth of high and low yielding lines and reported highest concentrations for cis- and transforms of zeatin with the former associated with floret setting and the latter during the grain filling stage. Moreover, both endogenous CK levels, and exogenous application of CKs at critical stages of seed development has been shown to have considerable influence on yield. Early studies with lupine (Lupinus angustifolius) showed an increased pod initiation upon exogenous CK application during flower initiation (Atkins and Pigeaire, 1993). In cereals, it was firmly established that CKs regulate kernel yields (Ashikari et al., 2005; Zalewski et al., 2010). This appears to be applicable to eudicot crops as well and several strategies have been proposed to target CK accumulation at reproductive phases of growth to enhance yields (Jameson and Song, 2016). Considering the importance of CKs in plant reproductive development, we aimed to identify the forms and abundances of CKs during peak reproductive developmental stages of soybean (Glycine max L. Merr.) and to correlate the hormone profiles with selected yield parameters of this economically important legume crop. Soybean is a species of legume that is widely cultivated across the world. The high oil and protein concentration makes it a valuable crop with multiple uses both in food and biomaterial industry (Nguyen et al., 2016). Despite its importance, a major problem in soybean cultivation is its flower and pod abortion. Soybean in general produces numerous floral buds of which a considerable proportion, 26–72% abscise before fertilization (Abernathy et al., 1977). Linkages between CK and yields of soybean have been suggested for quite some time and a review of the evidence is presented in Kokubun (2011). Among the reproductive stages of soybean, R1–R8, R1 being the flower initiation and R8 being full maturity of pods, studies have shown that stages R3 and R4 which correspond to stages from early pod set until full pod formation are the critical stages for yield determination as most of the abortion occurs during this time (Carlson et al., 1987; Peterson et al., 1990). Earlier studies also reported the role of CKs in regulating the abortion rates and number of pods being set (Carlson et al., 1987; Dyer et al., 1987; Peterson et al., 1990; Nagel et al., 2001). This was later linked to floral positions within the raceme and the probability of abortion at different locations (Kokubun and Honda, 2000). More recently, improved soybean productivity, as measured by the number of seeds per plant, seed weight and seed diameter, was achieved along with reduced abortion rates upon treatment with benzyladenine, a synthetic CK (Larissa et al., 2014). The sum of all the evidence clearly indicates that increased CK level at critical stages of reproductive development is positively reflected in soybean yield. However: moving the application of this knowledge to cropping conditions requires field-based studies. As pointed-out by Kokubun (2011), simple approaches involving exogenous CK application have only been successful in pot-grown plants, but their effects are obscured in field-grown plants. To verify whether a positive effect of CK on soybean yield can be expected under the complexity of field conditions an approach was used like that undertaken for barley by Powell et al. (2013). To that end, 27 soybean varieties were sampled from longterm field trials. These varieties were selected to represent a wide range of yield performance and were sampled across three reproductive stages. We report profiles of fourteen different CK types that include nucleotide, riboside and free base forms at R4–R6 stages that correspond to full pod, seed initiation and seed filling, stages critical for yield determination. Results provide insight into the clarity of the CK/yield associations and a baseline for further strategies for yield increase such as investigation of spatio-temporal expression patterns of CK gene

2. Materials and methods 2.1. Plant materials Soybean tissue samples for hormone analysis were collected in 2010 from two sets of soybean cultivars that were part of the long-term field trials conducted by Agriculture and Agri-Food Canada (AAFC). The first set included thirteen varieties developed by the University of Guelph and AAFC and evaluated in the Ontario 2600 CHU Conventional Soybean Variety Trial, while the second set included fourteen varieties evaluated in the Fifty Years Historical Variety Trial. Both trials were grown at the AAFC Ottawa Research and Development Centre (ORDC Central Experimental Farm, Ottawa, Canada (45°23′12.60” N lat, −75°42′12.54” W long)). All the cultivars used in our study had an indeterminate growth habit. Soybean planting, harvest and post-harvest processing of plant materials were performed by ORDC, following their guidelines for soybean cultivation. Five hundred seeds were planted in four row plots (1.8 × 5 m), with four experimental replicated plots per cultivar. To increase the precision of field trials involving large number of entries and to reduce the effect of within complete-block variation, alpha lattice block designs were used. The 27 cultivars evaluated in this study represented wide range of yield performance and were chosen based on yield data from previous field trials. Twelve agronomically important plant and seed parameters were evaluated in the field trials: yield [kg ha−1], thousand seed weight (TSW, g), days to maturity (DTM; the number of days a cultivar takes for 95% of the pods to ripen, with a moisture content in freshly matured pods of 35%), plant height [cm], lodging score (LS; 1 = completely erect to 5 = completely flattened), seed quality (SQ; 1 = excellent to 5 = poor) and seed composition [%]: protein, oil, sucrose, total sugars, raffinose/stachyose, and total carbohydrates. Seed composition was determined with a near infra-red whole grain analyser, (Infratec 1241, FOSS North America, Eden Prairie, MN, USA). This data was used for drawing statistical correlations with cytokinin levels obtained from hormone profiling by mass-spectrometry. Agronomical and hormone data of the cultivars from both trials were combined and analysed together to increase the sample size and provide a robust model to determine biological correlations. Table 1 includes the list of soybean cultivars and their yield characteristics that were revealed to be associated with cytokinin levels during plant generative development. The values of the remaining eight characteristics of soybean cultivars are presented in Supplementary Table S6. Three reproductive stages that critically correspond to soybean yield formation, pod and early seed development (Fehr and Caviness, 1977; Pedersen and Lauer, 2004) were identified and sampled in this study. Stage R4 is the latter one of the two stages that describe pod development. It begins about 20 days after flowering, when at least one full pod that is 3/4-in.-long at one of the four uppermost nodes on the main stem with a fully developed leaf. At this stage, rapid pod growth is occurring and seeds are starting to develop while flowering is still present on the upper branch nodes. At the second analysed stage – R5, seeds are 1/8-in.-long in the pods at one of the four uppermost nodes on the main stem. The R5 stage describes the initiation of seed development where root growth is slowing while rapid seed filling begins and dry weight and nutrients start being redistributed through the plant to the developing seed. The transition from R5 to R6, the latest stage used for CK profiling in this study, can take up to 15 days. The beginning of the R6 stage (over 40 days after flowering) corresponds to the time when pods contain green seeds that fill the pod to full capacity at one of the four uppermost nodes on the main stem. At that time, seeds of many sizes can be found on the plant. Nitrogen fixation continues all the way through R6 and 176

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

Table 1 Agronomic characteristics of soybean cultivars that expressed significant correlations with cytokinin levels during plant reproductive growth stages (means ± SE, n = 4). Cultivar

Yield ± SE [kg/ha]

1000 seed weight ± SE [g]

Days to maturity (DTM) ± SE

Plant height ± SE [cm]

Ontario 2600 CHU Conventional Soybean Variety Trail OAC Champion 3575 ± 84.8 OAC Blythe 3721 ± 122.1 OAC 08-11C 3848 ± 39.3 OAC Drayton 4578 ± 149.6 OAC Wallace 4434 ± 103.2 OAC Lakeview 3799 ± 88.3 OAC Wellington 3262 ± 80.9 Dares 3884 ± 93.5 OAC Purdy 3616 ± 87.9 OAC Madoc 4204 ± 171.0 DH618 4074 ± 135.3 OAC 07-26C 3911 ± 77.6 OAC Nation 3809 ± 58.5

234 222 200 233 237 233 230 259 242 235 235 230 260

± ± ± ± ± ± ± ± ± ± ± ± ±

3.6 5.5 2.3 2.3 2.0 3.6 4.0 4.0 2.1 1.7 2.4 2.2 2.1

133 139 125 140 140 131 132 138 138 125 133 139 140

± ± ± ± ± ± ± ± ± ± ± ± ±

1.4 0.3 0.7 0.4 1.0 1.3 1.1 1.4 1.1 0.8 1.8 0.7 0.9

108 ± 9.1 92 ± 4.1 99 ± 2.8 93 ± 3.4 100 ± 2.7 99 ± 7.2 99 ± 2.1 106 ± 1.1 106 ± 11.1 86 ± 5.3 88 ± 3.3 102 ± 8.3 98 ± 4.7

Fifty Years Historical Variety Trial Maple Arrow 3192 ± 23.2 Altona 2077 ± 68.7 AC Orford 2337 ± 84.8 Flambeau 1801 ± 35.5 Portage 1988 ± 125.2 Mandarin 2178 ± 120.8 AC Bravor 2553 ± 43.9 AC Harmony 2367 ± 124.8 Maple Glen 2324 ± 88.1 Maple Ridge 1606 ± 8.5 Crest 2144 ± 187.6 Pagoda 1265 ± 101.2 McCall 1838 ± 36.9 Dundas 3037 ± 44.5

217 193 231 200 202 251 226 166 232 162 190 172 170 228

± ± ± ± ± ± ± ± ± ± ± ± ± ±

1.8 1.9 3.7 4.9 4.6 3.0 3.6 1.3 2.5 1.7 2.7 3.3 4.8 2.0

127 115 104 112 115 115 127 104 112 100 127 104 104 127

± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

90 95 71 83 91 83 93 76 80 75 87 79 77 87

± ± ± ± ± ± ± ± ± ± ± ± ± ±

1.0 4.9 2.1 4.5 3.0 1.4 3.8 1.5 2.6 1.2 3.0 2.6 4.8 4.3

Pooled supernatants were evaporated in a speed vacuum concentrator (Savant SPD111 V, UVS400, Thermo Fisher Scientific, Waltham, MA) at ambient temperature. The dried sample was stored at −20 °C until further use. 1 mL of 1 M HCO2H was used to reconstitute the dried supernatant residue for complete protonation of CKs and samples were subjected to solid phase extraction (SPE) on a mixed mode, reverse phase/cation exchange cartridge (Oasis MCX 60 μm 6 cc; 150 mg; Waters, Mississauga, Canada). Prior to sample loading, the cartridges were activated with 5 mL CH3OH and equilibrated with 5 mL 1 M HCO2H. CKs were eluted based on their chemical properties with CK nucleotides forms eluted using 5 mL 0.35 M ammonium hydroxide (NH4OH) followed by riboside, free base and methylthiol forms eluted using 0.35 M NH4OH in 60% CH3OH. Collected fractions were evaporated to dryness and stored at −20 °C. Dried residues of CK nucleotide fractions were reconstituted in 1 mL of 0.1 M ethanolamine-HCL (pH 10.4) and dephosphorylated to form ribosides using 3.4 units of bacterial alkaline phosphatase (Sigma, Oakville, Canada) at 37 °C for 12 h, followed by evaporation to dryness in a speed vacuum concentrator at ambient temperature. Upon reconstitution with 1.5 mL MilliQ H2O, the dephosphorylated nucleotides were further purified using reverse-phased C18 SPE columns (Oasis C18 3 cc; Waters, Mississauga, Canada) that were activated and equilibrated using 3 mL CH3OH and 6 mL MilliQ H2O respectively. 3 mL MilliQ H2O was used to wash the sorbent bed followed by elution with 1.5 mL of 80% CH3OH. Samples were then evaporated to dryness in a speed vacuum concentrator at ambient temperature and the dried residue stored at −20 °C until further analysis. Purified and dried fractions of CK nucleotides, ribosides, free bases and methylthiols were reconstituted in 1.5 mL of initial HPLC mobile phase condition (95:5H2O: Acetonitrile (C2H3N) with 0.1% CH3CO2H), transferred to glass HPLC vials and proceeded for analyses by liquid chromatography tandem mass spectrometry.

large amounts of N are still being accumulated from the soil, directly to the seed. Tissue samples for cytokinin profiling were collected in triplicates from field plots used for evaluation of yield parameters of each of 27 tested soybean cultivars. Whole pods were collected for R4 and R5 stages while green seeds were used for CK quantification in stage R6, following visual inspection for the uniformity of the pods selected at each growth stage. Harvested samples were transported to the laboratory in electric cooler (−20 °C; Canadian Tire, Toronto, Canada) and stored in −80C freezer until CK extraction was performed.

2.2. Cytokinin extraction and purification Cytokinin extraction and purification was carried out as described in Farrow and Emery (2012) and Morrison et al. (2015a). Field-collected, previously frozen tissue was weighed (100 mg) and the mass recorded. The frozen tissue was placed in 1.5 mL Eppendorf tube with 1 mL cold (−20 °C) modified Bieleski #2 extraction buffer (CH3OH: H2O: HCO2H, (15:4:1 v/v/v)) and homogenized in a ball mill (25 Hz, 5 min, 4 °C; Retsch MM300, Haan, Germany) with zirconium oxide beads (Comeau Technique Ltd., Vaudreuil-Dorion, Canada). Following homogenization, 10 ng each of deuterated internal standard CK was added. The standards were: [2H6]iP, [2H6][9R]iP, [2H6][9RMP]iP, trans-[2H5]Z, trans-[2H5][9R]Z, trans-[2H5][9RMP]Z, [2H3]DZ, [2H3] [9R]DHZ, and [2H6][9R-MP]DHZ, [2H5]MeSZ, and [2H6]MeSZR (OlChemIm Ltd, Olomouc, Czech Republic). As deuterated standards of cisZ, cis-[9R]Z and cis-[9RMP]Z were not commercially available at the time of the study, the levels of these two compounds were quantified based on the recovery of the deuterated standards of the corresponding trans-compounds. Homogenized samples were subjected to sonication for 5 min, vortexed and allowed to extract passively overnight at −20 °C. Following overnight extraction, samples were centrifuged at 8400 × g for 10 min (Sorvall ST 16, Fisher Scientific) and the supernatant was collected. Pellets were used for re-extraction with Bieleski #2 buffer at −20 °C for 30 min followed by centrifugation as above. 177

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

Fig. 1. Regression analyses of yield vs. cytokinin forms that showed significant interactions during soybean reproductive development (insert: Pearson correlation coefficient at R4–R6).

178

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

Fig. 2. Regression analyses of Thousand Seed Weight (TSW) vs. cytokinin forms that showed significant interactions during soybean reproductive development (insert: Pearson correlation coefficient at R4–R6).

recovered internal standard. Data was analyzed using Analyst (v. 1.5.1.) software (AB SCIEX, Concord, Canada).

2.3. HPLC–MS/MS instrumentation Purified CK fractions were separated and analyzed using Waters 2680 Alliance HPLC system (Waters, Milford, Mass., USA) linked to a Quattro-LC triple quadrupole MS (Micromass, Altrincham, UK) equipped with a Z-electrospray ionization source (ESI). Positive-ion mode was used for all analyses. A 20 μL sample was injected on a Genesis C18 reversed-phase column (4 μm, 150 mm × 2.1 mm; Jones Chromatography, Foster City, Calif., USA), and the CKs were eluted with an increasing gradient of acetonitrile (A) and 0.1% formic acid in 20 mmol L−1 ammonium acetate at a pH adjusted to 4.0 (B) at a flow rate of 0.2 mL min−1. The initial conditions were 8% A and 92% B, changing linearly after 5 min to 15% A and 85% B for 2 min, followed by 100% A for 2 min, then returning to initial conditions for 2 min. The HPLC effluent was introduced into the electrospray source (source block temperature 80 °C, desolvation temperature 250 °C) using conditions specific for each CK where quantification was obtained by multiple reaction monitoring (MRM) of the parent ion and the appropriate product ion as described in Farrow and Emery (2012). Quantification of CKs was done using isotopic dilution method (Jacobsen et al., 2002). Analyte concentrations were determined based on direct comparison of the endogenous analyte peak area to that of the

2.4. Statistical analysis To determine statistically significant correlations between the agronomic characteristics and hormone profiles measured during soybean pod and seed development, Pearson correlations were calculated using PROC CORR implemented in SAS® v.9.1 and visualized using SigmaPlot software (Systat Software Inc.). Levels of statistical significance, in the form of p values, are expressed in the text and on the regression plots. 3. Results 3.1. Agronomical parameters and CK profiles In the presented work, plant hormone cytokinins were profiled across the three stages of soybean reproductive development. We report the thorough evaluation of fourteen different forms of cytokinin in 27 field-grown cultivars of soybean that varied in regards to yield performance along with other important agronomic traits. The extensive 179

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

height (R = 0.68; p < 0.001). Therefore, similar positive relationships as those detected between cytokinins and yield levels were observed for other yield components, although their significance levels were often lower (Fig. 2–4; Tables S3–S5). Moreover, it is worth pointing out that additional positive correlations (p < 0.05) were detected for two dihydrozeatin (DZ) forms during the R6 stage; namely, free base DZ was associated with DTM, while its riboside form (DZR) was positively correlated with the levels of both TSW and DTM.

analysis we have undertaken is one of the largest studies of the cytokinin metabolites ever performed in field-grown crop species. The presence of all the fourteen CK forms was detected in the pods of each of the tested soybean cultivars collected in R4 and R5 stages. In the R6 stage however, complete disappearance of all forms of trans- and cisisomers of zeatin was noted, suggesting limited hydroxylation of isopentenyl adenine in the maturing soybean tissues (Table S1). The main purpose of this investigation was to reveal the possible associations between cytokinin metabolites and plant agronomical performance. In total, 276 correlation matrixes were analysed separately for R4–R6 stages. The most ground-breaking finding of this investigation was that out of the twelve morphological and physiological traits of soybean, all four parameters that describe crop yield performance, namely yield, thousand seed weight (TSW), days to maturity (DTM) and plant height, turned out to be positively correlated with cytokinin levels during at least one of the three tested growth stages. The levels of these agronomic characteristics in 27 soybean cultivars can be found in Table 1. Regression analyses showing the fit of data along with Pearson correlation coefficient values, between four yield components and each of 14 CK forms, as well as the total levels of CK, free bases (FB), ribosides (RB), nucleotides (NT), methylthiols (MET), and total tZ, cZ, DZ and iP type cytokinins, at each of three stages tested, are presented in Figs. 1–4 and in Supplementary Tables S2–S5. On the contrary, the associations detected between cytokinin levels and lodging score, seed quality, content of: protein, oil, sucrose, total sugars, raffinose/stachyose, and total carbohydrates did not reveal trends as clear as those for yield and thus no clear biological significance was determined (data not shown).

4. Discussion Cytokinin distribution and signaling during plant vegetative and reproductive development differs significantly not only between dicot and monocot plant species (Li et al., 2013; Jiskrová et al., 2016), but also among the species of each of the two clades, including cereals (Ashikari et al., 2005; Brugière et al., 2008; Zalewski et al., 2010) or legumes (Emery et al., 1998; Emery et al., 2000). Moreover, recent works report different roles for CKs, even within cultivars of a single species (Powell et al., 2013). It highlights the need for not only speciesspecific studies, but also for the research that aims to investigate CK involvement in important aspects of plant metabolism even within closely related genotypes, such as in the present study. The critical role of phytohormones in the formation and abortion of reproductive organs in soybean was recognized decades ago. Among phytohormones, cytokinins are considered to play a vital role in soybean flower and pod development (reviewed in Kokubun, 2011). Endogenous CK levels were reported to be higher at the start of anthesis and this decreased as flowering progressed, along with the correlated decline of CK levels in xylem sap and the percentage of flowers that set (Carlson et al., 1987). Moreover, a significant loss of yield in soybean is commonly due to abortion and abscission of flowers at the late flowering and early pod set stages (Abernathy et al., 1977), particularly at the distal portions of the raceme compared to the proximal end (Huff and Dybing, 1980). This abscission was revealed to be correlated with the levels of CK in relation to the reproductive stages (Reese et al., 1995) and the proximity on raceme (Kokubun and Honda, 2000). It has been shown previously that stimulation of flowering, an increase in pod number and total seed weight can be obtained in soybean upon treatment with a synthetic CK, 6-benzylaminopurine (BA) (Crosby et al., 1981; Mosjidis et al., 1993). Despite the evidence above, such strong linkage between CK presence and determinants of yield such as flower and pod set did not translate into success under field conditions and the use of, rather imprecise, exogenously applied CKs (Nagel et al., 2001; Kokubun, 2011). Therefore, to better understand the relevance of CK in yield determination under field conditions, it is necessary to establish if endogenous CK levels correspond to yield traits or whether this effect is simply swamped out by the complexity of cropping conditions. In this study, during three stages of soybean reproductive development (R3, R4 and R5), we scanned for and reported on the presence of fourteen different CK types that included: the nucleotide, riboside and free base forms of isopentenyladenine (iP), trans-Zeatin (tZ), cisZeatin (cZ) and dihydrozeatin (DZ) along with methylthiol riboside and free base of trans-Zeatin (MeSZR and MeSZ). The distinct patterns of total isoprenoid CKs across various stages of pod and seed development have never been evaluated so thoroughly in any crop plants grown under the field conditions. We investigated whether the quantities of the individual CK types detected in a range of soybean cultivars, characterized by varying levels of agronomic performance, were associated with plant agronomic and physiological parameters. To obtain a broader picture of all the possible effects between cytokinin levels during soybean reproductive development and plant performance at harvest, we compared the hormone profiles with twelve agronomical characteristics, including: yield, thousand seed weight (TSW), days to maturity (DTM), plant height, lodging score, seed quality and seed composition of protein, oil, sucrose, total sugars, raffinose/stachyose, and total carbohydrates.

3.2. Correlations between CK levels during plant reproductive development and soybean yield performance The highest number of the strongest positive correlations (p < 0.001) with cytokinin forms and fractions during soybean reproductive development was observed for total plant yield as compared with the other three yield related agronomic parameters significantly associated with CKs (TSW, DTM and plant height). The highest total CK level and a positive correlation with yield was detected in R6 stage, while the most active, free base type CKs showed a positive correlation with yield at all three analysed stages (R4 and R5 at p < 0.001, and R6 at p < 0.05). Remarkably, different forms of cytokinins were detected as positively correlated with soybean yield at subsequent developmental stages (Fig. 1, Table S2). In the earlier stages that are more susceptible to pod abortion (R4 and R5), both isomers of the most biologically active CK form, free base zeatin were the most strongly associated with yield level (p < 0.001). The clear shift in the dominance between zeatin and iP types of cytokinin was noted for high yielding cultivars during the latter, seed filling stage (R6) when levels of all forms of zeatin (Z, ZR and ZNT) were below mass-spectrometer detection limits. This suggests that while zeatin CK forms are not being synthesized by the maturing seeds, the high yielding varieties retain a higher quantity of CK in the form of iP types during R6 stage. Interestingly, increased levels of methylthiols, especially MeSZ, were observed in high yielding cultivars, and their positive correlations with yield were detected across all three reproductive growth stages (p < 0.001 at R4, p < 0.01 at R5 and p < 0.05 at R6). This may suggest important function of these poorly understood CK types during soybean yield formation. 3.3. Correlations between CKs and other yield components (TSW, DTM, plant height) during soybean reproductive development The 27 soybean cultivars selected for this study showed a positive correlation between yield and three yield related agronomic traits, e.g. TSW (R = 0.68; p < 0.001), DTM (R = 0.87; p < 0.001) and plant 180

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

Fig. 3. Regression analyses of Days to Maturity (DTM) vs. cytokinin forms that showed significant interactions during soybean reproductive development (insert: Pearson correlation coefficient at R4–R6).

iP forms and yield were detected during the R4 stage, while they appeared to be the most abundant CK forms in the high yielding cultivars during the later, R6 stage, when seed filling takes place (Fig. 1, Tables S1, S2). Our results indicate a dramatic accumulation in all nucleotide, riboside and free base forms of iP-type CKs at the R6 stage in the higher yielding varieties, with the content of free base type iP lower compared to that of iPNT and iPR levels. Legumes in general are known to have higher nucleotide and riboside forms of CKs compared to their free base derivatives on this latter point during seed development (Emery et al., 1998; Emery et al., 2000; Quesnelle and Emery, 2007). Nucleotides are thought to indicate the level of de novo CK synthesis (Sakakibara, 2006). Riboside forms, the direct free base precursors, have been previously proven to be bioactive in bacterial assays (Romanov et al., 2006); although, more recent research suggests that only free bases can be considered as active cytokinins, able to interact with CK receptors in plant systems (Lomin et al., 2015). Although tZ is usually considered to be the most biologically active cytokinin, various bioassays shown that iP is yet another CK that expresses high affinity to CK receptors

Most remarkably, the present research revealed that the total plant yield and the three yield associated components: thousand seed weight (TSW), days to maturity (DTM) and plant height were all significantly associated with CK quantity and activity profiles. The analysis of correlations among plant yield and different cytokinin forms and types revealed very interesting patterns. We detected a significant shift in the CK composition that occurred between the R4 and R6 stages, with tZ and cZ free bases being the dominant, positively correlated with yield, forms of cytokinin during the two earlier stages, while their complete absence was observed in the latter, R6 stage (Table S1). Strong positive correlations of tZ and cZ type CKs with total yield level during the R4 and R5 stages suggest that the most active free base fraction of both zeatin isomers play an important role during the earlier stages of pod and seed set by increasing sink strength, ultimately leading to an increased yield (Fig. 1, Table S2). On the other hand, and contrary to zeatin trend, a complete reversal of roles during soybean reproductive development was observed for isopentenyladenine and its derivatives. Negative correlations between 181

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

Fig. 4. Regression analyses of plant height vs. cytokinin forms that showed significant interactions during soybean reproductive development (insert: Pearson correlation coefficient at R4–R6).

(DTM). In general, high yield directly depends on the length of vegetative growth; however, this usually is controlled by the fact that a prolonged growing season facilitates full seed maturation and allows for better yield quality at harvest. The observation that the number of DTM was positively correlated with cytokinins additionally suggests that the high yielding cultivars of soybean maintain their physiological activity longer, and more effectively divert assimilates to the maturing seeds, even during the later stages of reproductive development. Positive correlations similar to the ones described above were detected between cytokinins and plant height. The observed patterns are consistent since one of the main biological functions of CKs is stimulation of cell proliferation and plant productivity (Sakakibara, 2006). It can be expected that high yielding cultivars are characterized by the elevated CK levels, not only in reproductive organs, but also in their vegetative tissues which, in turn, is manifested as a larger plant size to accommodate more racemes, ultimately leading to an increased yield. Remarkably, methylthiol riboside and free base forms of transZeatin (MeSZR and MeSZ) were also detected in high quantities at all three stages of high yielding varieties (Table S1). In particular, a positive correlation of methylthiol free base forms with all yield components (Figs. 1–4), points towards a potentially significant role of methylthiol-CK forms in seed filling. So far, methylthiol types of CKs have only been studied in the context of plant-microbe interactions and were suggested to have pathogen stimulated origins and be responsible for tissue proliferation (Morrison et al., 2015a; Giron and Glevarec, 2014). Given that soybean is a symbiotic legume species, it is possible that these forms indeed originate from symbiotic bacteria and are transported to reproductive tissues. However, the origin of these CK forms in this system remains unknown. Another possible explanation for their high accumulation could be the inability of cytokinin oxidases (CKX) to degrade methylthiol- forms of CKs (Morrison et al., 2015b). Finally, dihydrozeatin riboside (DZR) and free base (DZ) forms showed positive correlations with TSW and DTM, although only in R6 stage (Fig. 2 and 3). This may suggest a role for DZ derivatives in seed filling, given that these forms of CKs are also known to be bioactive, with AHK3, a histidine kinase involved in cytokinin signalling, that has ligand binding activity towards DZ along with other forms of CKs

(Sakakibara, 2006; Kiba et al., 2013). Moreover, previous studies report that iP-type hormones are implicated in the control of ripening-related processes in various fruit-bearing plant species (Böttcher et al., 2015). Our results are consistent also with the findings that iP forms of CKs are most likely associated with cell expansion and seed filling in soybeans during the later stages of seed development, R6–R8 (Nguyen et al., 2016). In field grown barley cultivars, Powell et al. (2013) reported that, independently of the cultivar performance, cZ peaked early at the cell division and elongation phase of kernel development. In contrast, the increase in tZ levels was observed especially in high yielding barley cultivars during the later stages, when kernel filling occurs. Results of the present study provide new evidence that CKs are limiting factors for pod and seed set in soybean (Jameson and Song, 2016 and references therein). However, we propose that soybean developed a different mechanism of hormonal control over the plant reproductive development. High levels of the most physiologically active zeatin detected in high yielding cultivars at the R4 and R5 stages most likely originate from rapid hydroxylation of iP (Takei et al., 2004) during the early stages of pod and seed development when plants remain still highly metabolically active. Later, the sustained and continued de novo CK synthesis that results in increased levels of iP types of cytokinin at R6 stage, indicates that the high yielding soybeans maintain higher sink strength also as the reproductive stage progresses. This in turn, allows them to retain higher pod and seed set as well as prolongs seed filling into the maturation period when most other growth processes are finished, assuring higher yield and better seed quality. The other three agronomical parameters that revealed positive associations with CK levels during soybean pod and seed development (TSW, DTM and plant height) additionally reinforced the associations observed between cytokinins and plant yield (Figs. 2–4; Tables S3–S5). The first of these traits – Thousand Seed Weight (TSW), a direct indicator of seed size, revealed less clear trends while some reinforcement of yield patterns, such as positive correlation with tZ and cZ at R4, and with all the iP types at R6 stage, were still observed. In a similar manner, patterns to that of yield were observed for another of the analysed soybean characteristics, Days to Maturity 182

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

traits. Furthermore, manipulating CK gene family members using gene editing as a tool to increase CK content, is also of potential interest for the future research.

(Romanov et al., 2006; Lomin et al., 2015). The results of this study clearly indicate the existence of a strong connection between soybean yields and CK levels. Such an association seem to be relatively robust considering the fact, that the presented results were obtained not under the controlled, laboratory conditions, but instead, all the hormone profiles originate from plants grown in the field environment. Similar to the results of previous reports, for which continuous spraying with exogenous BA achieved increased pod set in soybean, higher yielding cultivars tested in our work demonstrated a high CK level, which was maintained as a continuous supply during the subsequent stages of reproductive development. It was shown that CK action in plants is controlled by CK quantity and that the physiological function of CK is modulated by chemical transformations between CK types (Kiba et al., 2013). Based on the observed correlations, we suggest that tZ and cZ type CKs play an important role in pod/seed set in soybeans, while iP derivatives may be involved in seed filling. Methylthiol riboside and free base forms of trans-Zeatin (MeSZR, MeSZ) were also found to be synthesized in high quantities, although, the source of their production and their role in achieving sink strength in seeds has not yet been fully revealed. It is important to note that all the soybean cultivars used in our research had an indeterminate growth habit. We assume that the extended vegetative growth period that overlaps with the reproductive development might be affecting the observed CK profiles and their implications toward seed formation, allowing for easier transportation of CK produced in actively growing stems and leaves directly to the sink reproductive organs. Therefore, the future studies could focus also on comparing cytokinin profiles between indeterminate and determinate soybean cultivars. It is possible that different patterns of CK production during the reproductive development are more typical for soybeans of determinate growth habit. The rate limiting step in CK biosynthesis is controlled by an enzyme Isopentenyl transferase (IPT) while CK breakdown is largely controlled by another enzyme, Cytokinin oxidase/dehydrogenase (CKX), which cleaves the isopentenyl side chain (Sakakibara, 2006). Lonely guy (LOG) is one more enzyme responsible for activation of CKs as it converts the nucleotide forms to the more active free base forms of CKs in a single step reaction (Kurakawa et al., 2007). Together, these three enzymes along with enzymes involved in CK inactivation (O-, N-glycosyl transferases) and reactivation (β-glucosidases) play a major role in maintaining CK homeostasis in plants (Jameson and Song, 2016). Several lines of evidence highlight the observation that increase in endogenous CK levels by altering CK biosynthesis or degradation genes by means of natural variation or genetic manipulation during essential stages of reproductive development results in increased yields. This was shown in several plant species; for example, strong expression of the ZmIPT2 gene and its product overlaid the change in CK levels in developing maize kernels suggesting a major role in CK biosynthesis for kernel development (Brugière et al., 2008). Genetically manipulating the expression of specific members of the CK gene families with a goal of increasing the hormone level has also been shown to improve yield parameters. The increased inflorescence meristems and higher grain number due to a loss of function mutation in OsCKX gene was reported in rice (Ashikari et al., 2005) and higher seed size and yield were observed in HvCKX1 knockout lines of barley (Zalewski et al., 2010). Several mutations in genes involved in CK biosynthesis and degradation leading to higher CK accumulation, that were shown to reflect in phenotypic traits related to yield in various plant species were reviewed by Jameson and Song (2016). The variation in cytokinin profiles detected among the soybean cultivars that differ in yield characteristics and the observed strong, positive correlations between the yield compounds and cytokinin levels suggest that high yielding soybeans could potentially encode natural variations within the cytokinin gene families that provide them with an agronomic advantage due to altered hormone levels. If identified, these variations can be used in marker assisted selection for breeding of soybean varieties with superior agronomic

Contributions EC, MM conducted soybean field trials and assisted with the manuscript. LK and KB were responsible for conducting the experiments and analysis of data. SK and AK were responsible for analysis of data and writing of the manuscript. NE was involved in planning, design of the experiments and writing of the manuscript. Acknowledgements The authors acknowledge the funding support of the Natural Sciences and Engineering Research Council of Canada (NSERC) for a Collaborative Research and Development Grant and Industry partner SEVITA International. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.fcr.2017.09.009. References Abernathy, R.H., Palmer, R.G., Shibles, R., Anderson, I.C., 1977. Histological observations on abscising and retained soybean flowers. Can. J. Plant Sci. 57, 713–716. Ashikari, M., Sakakibara, H., Lin, S., Yamamoto, T., Takashi, T., Nishimura, A., Angeles, E.R., Qian, Q., Kitano, H., Matsuoka, M., 2005. Cytokinin oxidase regulates rice grain production. Science 309, 741–745. Atkins, C., Pigeaire, A., 1993. Application of cytokinins to flowers to increase pod set in Lupinus angustifolius L. Aust. J. Agric. Res. 44, 1799–1819. Böttcher, C., Burbidge, C.A., Boss, P.K., Davies, C., 2015. Changes in transcription of cytokinin metabolism and signalling genes in grape (Vitis vinifera L.) berries are associated with the ripening-related increase in isopentenyladenine. BMC Plant Biol. 15, 223. Brugière, N., Humbert, S., Rizzo, N., Bohn, J., Habben, J.E., 2008. A member of the maize isopentenyl transferase gene family, Zea mays isopentenyl transferase 2 (ZmIPT2), encodes a cytokinin biosynthetic enzyme expressed during kernel development. Plant Mol. Biol. 67, 215–229. Carlson, D.R., Dyer, D.J., Cotterman, C.D., Durley, R.C., 1987. The physiological basis for cytokinin induced increases in pod set in IX93-100 soybeans. Plant Physiol. 84, 233–239. Crosby, K.E., Aung, L.H., Buss, G.R., 1981. Influence of 6-benzylaminopurine on fruit-set and seed development in two soybean, Glycine max (L.) Merr. genotypes. Plant Physiol. 68, 985–988. Dyer, D.J., Carlson, D.R., Cotterman, C.D., Sikorski, J.A., Ditson, S.L., 1987. Soybean pod set enhancement with synthetic cytokinin analogs. Plant Physiol. 84, 240–243. Emery, R.J.N., Leport, L., Barton, J.E., Turner, N.C., Atkins, C.A., 1998. Cis-isomers of cytokinins predominate in chickpea seeds throughout their development. Plant Physiol. 117, 1515–1523. Emery, R.J.N., Ma, Q., Atkins, C.A., 2000. The forms and sources of cytokinins in developing white lupine seeds and fruits. Plant Physiol. 123, 1593–1604. Farrow, S.C., Emery, R.N., 2012. Concurrent profiling of indole-3-acetic acid, abscisic acid, and cytokinins and structurally related purines by high-performance-liquidchromatography tandem electrospray mass spectrometry. Plant Methods 8, 42. Fehr, W.R., Caviness, C.E., 1977. Stages of Soybean Development. Iowa State University of Science and Technology, Ames, Iowa. Götz, K.-P., Staroske, N., Radchuk, R., Emery, R.J.N., Wutzke, K.-D., Herzog, H., Weber, H., 2007. Uptake and allocation of carbon and nitrogen in Vicia narbonensis plants with increased seed sink strength achieved by seed-specific expression of an amino acid permease. J. Exp. Bot. 58, 3183–3195. Giron, D., Glevarec, G., 2014. Cytokinin-induced phenotypes in plant-insect interactions: learning from the bacterial world. J. Chem. Ecol. 40, 826–835. Hirose, N., Takei, K., Kuroha, T., Kamada-Nobusada, T., Hayashi, H., Sakakibara, H., 2008. Regulation of cytokinin biosynthesis, compartmentalization and translocation. J. Exp. Bot. 59, 75–83. Huff, A., Dybing, C.D., 1980. Factors affecting shedding of flowers in soybean (Glycine max (L.) Merrill). J. Exp. Bot. 31, 751–762. Jacobsen, J.V., Pearce, D.W., Poole, A.T., Pharis, R.P., Mander, L.N., 2002. Abscisic acid, phaseic acid and gibberellin contents associated with dormancy and germination in barley. Physiol. Plant. 115, 428–441. Jameson, P.E., Song, J., 2016. Cytokinin: a key driver of seed yield. J. Exp. Bot. 67, 593–606. Jiskrová, E., Novák, O., Pospíšilová, H., Holubová, K., Karády, M., Galuszka, P., Robert, S., Frébort, I., 2016. Extra- and intracellular distribution of cytokinins in the leaves of monocots and dicots. New Biotechnol. 33, 735–742.

183

Field Crops Research 214 (2017) 175–184

S. Kambhampati et al.

and pod set in soybeans (Glycine max (L.) Merr.). Ann. Bot. 88, 27–31. Nguyen, T.Q., Kisiala, A., Andreas, P., Emery, N.R.J., Narine, S., 2016. Soybean seed development: fatty acid and phytohormone metabolism and their interactions. Curr. Genomics 17, 241–260. Pedersen, P., Lauer, J.G., 2004. Soybean growth and development in various management systems and planting dates. Crop Sci. 44, 508-515. Peterson, C.M., Williams, J.C., Kuang, A., 1990. Increased pod set of determinate cultivars of soybean, Glycine max, with 6-benzylaminopurine. Bot. Gaz. 151, 322–330. Powell, A.F., Paleczny, A.R., Olechowski, H., Emery, R.J.N., 2013. Changes in cytokinin form and concentration in developing kernels correspond with variation in yield among field-grown barley cultivars. Plant Physiol. Biochem. 64, 33–40. Quesnelle, P.E., Emery, R.J.N., 2007. cis-Cytokinins that predominate in Pisum sativum during early embryogenesis will accelerate embryo growth in vitro. Can. J. Bot. 85, 91–103. Reese, R.N., Dybing, C.D., White, C.A., Page, S.M., Larson, J.E., 1995. Expression of vegetative storage protein (VSP-β) in soybean raceme tissues in response to flower set. J. Exp. Bot. 46, 957–964. Romanov, G.A., Lomin, S.N., Schmülling, T., 2006. Biochemical characteristics and ligand-binding properties of Arabidopsis cytokinin receptor AHK3 compared to CRE1/ AHK4 as revealed by a direct binding assay. J. Exp. Bot. 57, 4051–4058. Sakakibara, H., 2006. Cytokinins: activity, biosynthesis, and translocation. Annu. Rev. Plant Biol. 57, 431–449. Spichal, L., 2012. Cytokinins – recent news and views of evolutionally old molecules. Funct. Plant Biol. 39, 267–284. Takei, K., Yamaya, T., Sakakibara, H., 2004. Arabidopsis CYP735A1 and CYP735A2 encode cytokinin hydroxylases that catalyze the biosynthesis of trans-zeatin. J. Biol. Chem. 279, 41866–41872. Yamburenko, M.V., Kieber, J.J., Schaller, G.E., 2017. Dynamic patterns of expression for genes regulating cytokinin metabolism and signaling during rice inflorescence development. PLoS One 12, e0176060. Zalewski, W., Galuszka, P., Gasparis, S., Orczyk, W., Nadolska-Orczyk, A., 2010. Silencing of the HvCKX1 gene decreases the cytokinin oxidase/dehydrogenase level in barley and leads to higher plant productivity. J. Exp. Bot. 61, 1839–1851.

Kamada-Nobusada, T., Sakakibara, H., 2009. Molecular basis for cytokinin biosynthesis. Phytochemistry 70, 444–449. Kiba, T., Takei, K., Kojima, M., Sakakibara, H., 2013. Side-chain modification of cytokinins controls shoot growth in Arabidopsis. Dev. Cell 27, 452–461. Kokubun, M., Honda, I., 2000. Intra-raceme variation in pod-set probability is associated with cytokinin content in soybeans. Plant Prod. Sci. 3, 354–359. Kokubun, M., 2011. In: Ng, Prof. Tzi-Bun (Ed.), Physiological Mechanisms Regulating Flower Abortion in Soybean. Soybean – Biochemistry, Chemistry and Physiology. InTech. http://dx.doi.org/10.5772/15694. Kurakawa, T., Ueda, N., Maekawa, M., Kobayashi, K., Kojima, M., Nagato, Y., Sakakibara, H., Kyozuka, J., 2007. Direct control of shoot meristem activity by a cytokinin-activating enzyme. Nature 445, 652–655. Larissa, P.B., Hilton, D.T.J., GdFS, Priscilla, 2014. Does benzyladenine application increase soybean productivity? Afr. J. Agric. Res. 9, 2799–2804. Li, J., Nie, X., Tan, J.L., Berger, F., 2013. Integration of epigenetic and genetic controls of seed size by cytokinin in Arabidopsis. Proc. Natl. Acad. Sci. U. S. A. 110, 15479–15484. Lomin, S.N., Krivosheev, D.M., Steklov, M.Y., Arkhipov, D.V., Osolodkin, D.I., Schmülling, T., Romanov, G.A., 2015. Plant membrane assays with cytokinins receptors underpin the unique role of free cytokinins bases as biologically active ligands. J. Exp. Bot. 66, 1851–1863. Mok, D.W., Mok, M.C., 2001. Cytokinin metabolism and action. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52, 89–118. Morrison, E.N., Emery, R.J.N., Saville, B.J., 2015a. Phytohormone involvement in the Ustilago maydis −Zea mays pathosystem: relationships between abscisic acid and cytokinin levels and strain virulence in infected cob tissue. PLoS One 10, e0130945. Morrison, E.N., Knowles, S., Hayward, A., Thorn, R.G., Saville, B.J., Emery, R.J., 2015b. Detection of phytohormones in temperate forest fungi predicts consistent abscisic acid production and a common pathway for cytokinin biosynthesis. Mycologia 107, 245–257. Mosjidis, C.O.H., Peterson, C.M., Truelove, B., Dute, R.R., 1993. Stimulation of pod and ovule growth of soybean, Glycine max (L.) Merr. by 6-benzylaminopurine. Ann. Bot. 71, 193–199. Nagel, L., Brewster, R., Riedell, W.E., Reese, R.N., 2001. Cytokinin regulation of flower

184