Impact of nitrogen fixing and plant growth-promoting bacteria on a phloem-feeding soybean herbivore

Applied Soil Ecology 86 (2014) 71–81

Contents lists available at ScienceDirect

Applied Soil Ecology journal homepage: www.elsevier.com/locate/apsoil

Impact of nitrogen fixing and plant growth-promoting bacteria on a phloem-feeding soybean herbivore S.M. Brunner a,b , R.J. Goos c , S.J. Swenson a , S.P. Foster a , B.G. Schatz d, Y.E. Lawley d,e, D.A. Prischmann-Voldseth a, * a

North Dakota State University, Department of Entomology 7650, PO Pox 6050, Fargo, ND 58108-6050, United States North Dakota Department of Agriculture, 600 E Boulevard Ave Department 602, Bismarck, ND 58505-0020, United States North Dakota State University, Department of Soil Science, PO Pox 6050, Fargo, ND 58108-6050, United States d North Dakota State University, Carrington Research Extension Center, 663 Hwy 281 N, PO Box 219, Carrington, ND 58421-0219, United States e University of Manitoba, Plant Science Department, 222 Agriculture Building, 66 Dafoe Road, Winnipeg, MB R3T 2N2, Canada b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 30 April 2014 Received in revised form 2 October 2014 Accepted 6 October 2014 Available online xxx

In soybean (Glycine max L.) production systems, growers often inoculate seeds with the symbiotic N-fixing species Bradyrhizobium japonicum along with other bacterial species or chemicals intended to enhance plant growth and yield. However, microbes associated with plant roots can also impact the biology of above-ground insect herbivores through influencing various aspects of plant physiology, with effects dependent on the bacterial species. Because rhizobial seed inoculants can potentially affect densities of soybean herbivores, we investigated the performance of phloem-feeding soybean aphids (Aphis glycines Matsumura) on: (1) soybeans receiving one of four commercially available rhizobial seed inoculants, (2) non-inoculated soybeans associated with existing soil bacteria, or (3) non-inoculated soybeans receiving high levels of fertilizer to suppress N-fixation while still providing adequate N as nitrate. We quantified effects of inoculants and aphid presence on parameters associated with plant growth, N-fixation, and foliar N levels, and explored relationships between aphid densities and these plant parameters. Regardless of inoculation treatment, aphid presence negatively affected plant biomass, pod density, and total N concentration in aerial plant tissues, although, effects on the concentration of ureide N (primary products of N-fixation) were not significant. Inoculant identity significantly impacted aphid populations in both short- (1 week) and longer-term experiments (2 months), with pest densities negatively related to the number of root nodules per plant. This research indicates that nodulation status of soybeans can influence above-ground herbivores. ã 2014 Elsevier B.V. All rights reserved.

Keywords: Aphis glycines Soybean Nitrogen fixation Plant growth-promoting bacteria

1. Introduction Plants often form close mutualistic associations with soil microorganisms such as bacteria and fungi, with the relationship between legumes (Fabaceae) and nitrogen (N) fixing bacteria being one of the most well known. In this case, both types of organisms directly benefit from the relationship, however, there can be indirect effects on organisms at different trophic levels, such as on herbivorous insects (Dean et al., 2009; Thamer et al., 2010 Ballhorn et al., 2013). Root-associated bacteria (i.e., rhizobia) and mycorrhizal fungi can aid plants by increasing resistance to aboveground herbivores via defensive compounds; however, concomitant increases in plant growth and nutrition due to these

* Corresponding author. Tel.: +1 701 231 9805; fax: +1 701 231 8557. E-mail address: [email protected] (D.A. Prischmann-Voldseth). http://dx.doi.org/10.1016/j.apsoil.2014.10.007 0929-1393/ ã 2014 Elsevier B.V. All rights reserved.

associations may make plants more susceptible to other herbivores (Koricheva et al., 2009; Pineda et al., 2010). Commercial soybean (Glycine max L.) production requires less N input than non-leguminous crops due to the plant’s ability to obtain biologically fixed N via a symbiotic relationship with Bradyrhizobium japonicum (Bradyrhizobiaceae), a bacterial species in a group of N-fixing taxa associated with legumes collectively referred to as rhizobia (van Rhijn and Vanderleyden, 1995). The bacteria are located within root nodules and convert atmospheric N2 into nitrogenous compounds called ureides (Schubert, 1986). In general, ureide N is correlated with N-fixation (van Berkum et al., 1985; Schubert, 1986). Several factors can impact the strength of rhizobial-plant associations and subsequent nodulation and N-fixation rates, most notably soil N (e.g., nitrate) levels (Streeter and Wong, 1988). When a soil contains adequate N, either from organic matter or synthetic fertilizer, a plant will use these sources before associating with rhizobia (Evans, 1982; Ohyama et al., 2009).

72

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

The source of N used by soybeans (from fertilization or N-fixation) can impact the identity and concentration of nitrogenous compounds within plant tissues (Pate, 1980; McNeil and LaRue, 1984), with the total N within xylem sap of nodulated soybeans consisting of 60–95% ureides (McClure and Israel,1979). The other major form of N translocated from roots to the shoots is asparagine (ASN), which is an amino acid present within the xylem, regardless of whether the plants obtain N from N-fixation or fertilizer, although more of this amino acid results from the latter scenario (Streeter, 1972; McClure and Israel, 1979; Schubert, 1986; Shelp and Da Silva, 1990). Rhizobia are an integral part of the soil microbial community and can remain viable in soil for years, even when their legume host is not present (Bottomley, 1992). However, because the most suitable species may not occur in high densities, many soybean growers use commercially available rhizobial seed inoculants to increase N-fixation and boost yield (Bottomley, 1992; Keyser and Li, 1992). These commercial inoculants, typically applied to seeds before planting, contain B. japonicum, but often have additional components intended to increase yield (e.g., other biological organisms or growth-promoting factors). The other chemical and biotic components in these inoculants may affect plant physiology in various ways, including by enhancing seedling emergence, root nodulation, N-fixation, and plant growth (Dénarié ., Dobbelaere et al et al., 1996Dénarié ., Dobbelaere et al et al., 1996Dénarié ., Dobbelaere et al., 2003Dénarié ., Dobbelaere et al., 2003 Cassán et al., 2009; Rodriguez-Navarro et al., 2011; Sindhu et al., 2010). Rhizobial species identity, root nodulation, and N-fixation impact the behavior and biology of above-ground herbivorous arthropods, including those that feed on legumes (Wilson and Stinner, 1984; Dean et al., 2009; Kempel et al., 2009; Thamer et al., 2011; Katayama et al., 2011a). In addition, in some legume systems, rhizobia can alter induced plant defenses, i.e., the production of volatile organic compounds associated with herbivore plant preference (Ballhorn et al., 2013). Effects of rhizobia on herbivores may depend on the latter’s feeding habit, as chewing herbivores seem to be affected to a greater degree than sucking herbivores, such as phloem-feeding aphids (Kempel et al., 2009; Katayama et al., 2011a). Although there is mixed evidence for effects of rhizobia on aphids (Kempel et al., 2009; Katayama et al., 2011a), one study indicated that results might depend on the species of rhizobia associating with the plant (naturally occurring versus commercially available; Dean et al., 2009). Thus, it is possible that specific inoculants may differentially affect aphids that feed on legumes, including soybean aphids (Hemiptera: Aphididae: Aphis glycines Matsumura). Soybean aphids overwinter on buckthorn (Rhamnaceae: Rhamnus), and in the spring, winged females migrate to soybean, the preferred summer (secondary) host on which females reproduce asexually and give birth to live nymphs (Ragsdale et al., 2004). These reproductive strategies facilitate rapid aphid population growth, and densities can quickly become high enough to surpass economic thresholds (Ragsdale et al., 2007). Several plant factors and aphid behaviors are involved in host plant selection and acceptance (reviewed by Powell et al., 2006), including visual, olfactory, and gustatory cues. To investigate the impact of rhizobial seed inoculants on soybean aphid establishment and reproduction on soybean plants, we conducted greenhouse and field studies with six treatments: four different commercially available inoculants, a non-inoculated control in which plants associated with existing soil rhizobia, and a high N treatment, in which additional fertilizer was used to suppress nodulation and N-fixation while still providing adequate N (as nitrate) for plant growth. Because aphid feeding can impact plant growth and physiology, in field experiments two cages were erected in each plot (with and without aphids) in order to explore main and interactive effects of inoculant treatments and aphids on plant parameters associated with growth, foliar N levels, and

reliance on N-fixation. We also used correlation analysis and multiple linear regression to assess relationships between aphid density and various plant parameters related to plant growth, N-fixation and foliar N content, in order to explore what parameters were the most important drivers of aphid populations. 2. Methods 2.1. Field experiment 2.1.1. Experimental design Field experiments were conducted in 2010 and 2011 in Carrington, ND at the North Dakota State University (NDSU) Research and Extension Center. The exact location of experimental plots (1.5  7.6 m) within the research farm changed between years. Within a replicate, plots were separated by same-size soybean buffer plots, with replicates separated by 1.5 m fallow strips. In late June of both years, ammonium sulfate (2.3 l/ha) was mixed with glyphosate (3.5 l/ha) and applied to plots for weed control. Plots were not sprayed with any other chemicals for the duration of the experiment. There were six experimental treatments: CNTL (a noninoculated control), N (a non-inoculated high nitrogen treatment), and four types of commercially available seed inoculants, B (B. japonicum, N-DURE, peat formulation, INTX Microbials, LLC, Kentland, IN), B + D [B. japonicum + Delftia acidovorans (Comamonadaceae), BioBoost1 Plus, liquid formulation, Brett Young, Winnipeg, MB, Canada], B + L (B. japonicum + lipo-chitooligosaccharides, Optimize1 400, liquid formulation, Novozymes, Brookfield, WI) and B + A [B. japonicum + Azospirillum brasilense (Rhodospirillaceae), PRIMO, liquid formulation, INTX Microbials]. Azospirillum brasilense is a known nitrogen fixer and research has indicated it can enhance the effectiveness of B. japonicum (Groppa et al., 1998) and increase nodulation and N-fixation of several legumes (Dobbelaere et al., 2003). In addition, some A. brasilense strains can produce phytohormones and enhance growth of soybean seedlings (Cassán et al., 2009). Delftia spp. (syn. Comamonas) are known as bioremediators of environmental pollutants (Ubalde et al., 2012). Delftia acidovorans is considered a plant growth promoter and can produce phytohormones, such as indole-3-acetic acid (Barazani and Friedman, 1999), and protect grape roots from harmful nematodes (Aballay et al., 2013). Related species have also been shown to fix N (Han et al., 2005) and increase the effects of some legume inoculants (Morel et al., 2011). Lipo-chitooligosaccharides, or Nod factors, are bacteria-produced plant signaling molecules essential to establishing the symbiotic relationship between N-fixing bacteria and legume roots (Dénarié et al., 1996). They have been associated with plant growth promotion, including enhanced germination and root growth of soybean (Souleimanov et al., 2002; Prithiviraj et al., 2003) and protection against root pathogens (Duzan et al., 2005). In 2010, each treatment was replicated six times, while in 2011 there were only five replicates of each treatment. Treatments were randomly assigned to plots within replicates. Soybean seeds (indeterminate, Dairyland 401RR, Dow AgroSciences, Indianapolis, IN) were inoculated according to package instructions (ml product kg 1 seed): B, 74 per 23 (2  108 viable B. japonicum cells g 1); B + D, 104 per 45 (2  109 viable B. japonicum and 1 107 D. acidovorans g 1; B + L, 83 per 45 (includes liquid additive, 5  109 viable B. japonicum cells g 1, 2  10 7 % lipo-chitooligosaccharide); B + A, 125 per 45 (includes liquid additive, 3  109 viable B. japonicum cells and 1 108 A. brasilense g 1). Seeds were planted on 20 May 2010 and 2 June 2011 using a Hege 1000 research plot planter at a rate of approx. 543,630 viable seeds ha-1, with 17.8 cm between-row and 10.4 cm within-row seed spacing. For high N plots, 67.3 kg N fertilizer ha 1 was applied shortly after planting by

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

hand broadcasting urea across the plot. Rainfall events were relied upon for incorporating urea into the soil. 2.1.2. Aphid infestation On 23 June 2010 and 18 July 2011, two mesh field cages (0.6 w  0.6 d  1.4 m h; Bioquip, Rancho Dominguez, CA) were erected in each experimental plot and the bottom edge buried 8 cm underground. Cages were centered in plots so that approximately 0.3 m separated cages from each other and from plot edges. Plants were thinned to two per cage and sprayed with PyGanic1 (pyrethrin; REI = 12 h, 6.57 g/l; MGK, Minneapolis, MN) on 23 June 2010 and 18 July 2011 in order to eliminate any resident arthropods. One cage in each plot was infested with soybean aphids (designated as +Aphid cages) on 28 June 2010 and 21 July 2011. Five adult soybean aphids were added to each plant within a cage by draping an infested leaf piece over the second trifoliate. Aphids originated from a colony maintained at NDSU on potted soybean plants (RG607 RR) at 25  2  C and 16:8 L:D. 2.1.3. Data collection In both years, soybean aphid densities and plant growth stages were non-destructively assessed (via visual inspection of all above-ground plant structures; i.e., abaxial and adaxial leaf surfaces, stems, petioles, pods) 24 h after infestation to determine the rate of insect establishment on host plants (i.e., 29 June; 2010 22 July 2011). In 2010, aphid density was also quantified non-destructively for both plants within +aphid and no-aphid cages on 6 July, 20 July, 4 August, and 19 August. In 2011, aphid density was non-destructively assessed on, 4 August, and 23 August. In both years, final aphid counts were conducted after plants had been destructively sampled and frozen (on 28 August 2010; and 8 September 2011). In 2010, insects other than soybean aphids were rarely observed, and when encountered, were removed by hand. In 2011, there was a natural infestation of soybean aphids in the general region, and some aphids remained in field cages after spraying with PyGanic1. Therefore, no-aphid cages were resprayed with PyGanic1 (6.57 g/l) on 27 July. In addition, the presence of an unidentified aphid pathogen was noted in all experimental cages late in the season (23 August 2011). Both plants in each cage were destructively sampled on 28 August 2010 and 8 September 2011, which involved counting the number of pods, cutting stems at the soil surface, and placing each plant in a separate plastic bag. In 2010, data on number of root nodules per plant were obtained from non-caged, non-aphid infested plants (n = 5 per plot, sampled on 29 September) within adjacent plots (6 replicates). In 2011, root nodules were collected from experimental plants (from 4/5 replicates). After the above-ground foliage was removed, the top 15 cm of the root system was dug out of the soil, and all soil and root material placed in a self-sealing plastic bag and stored at 4  2  C for a maximum of 1 week. To assess nodules, each sample was placed in a sieve (710 mm opening, U.S.A standard testing sieve, no. 25) to remove excess soil and find any loose root nodules. All roots and root nodules were then washed and nodules still attached to the roots were removed by hand. Above-ground plant material was placed in paper bags and dried for one week (120  C in 2010, 96  C in 2011). In 2011, mold was discovered on stored plant samples approx. 1week after being removed from the drying ovens. We concluded that plants had not been completely dried (perhaps due to equipment malfunction), or had been stored improperly. Therefore, plants were returned to drying ovens (120  C) for another week. Dried plant material (above-ground vegetative and reproductive tissue) was weighed on a digital scale (Sartorius type 1412) and stored in

73

self-sealing plastic bags placed inside black plastic garbage bags at room temperature (20  C) until nitrogenous compounds were assessed. 2.2. Greenhouse experiment Because prolonged soybean aphid feeding can reduce the volume of soybean root nodules and lower the rate of N-fixation, presumably by removing photosynthate from the phloem (McClure and Israel, 1979; van Berkum et al., 1985; Riedell et al., 2009), correlations between aphid densities and plant parameters from the field experiment may be indicative of aphid effects on plants rather than plant effects on aphids. Therefore, we conducted a greenhouse experiment to examine how inoculation treatments impacted aphid and plant parameters when effects of aphids on plants were minimized (i.e., by limiting the duration of the experiment). Treatments paralleled those in the field experiment. There were nine replications of each treatment, with pot location randomized within each replicate. 2.2.1. Planting procedures Plants were grown in plastic pots (7.5 cm h  11.5 cm d) lined with sterile plastic bags at 24  4  C, 40–80% RH and 16:8 L:D. The soil used was collected from Streeter, ND (Renshaw sandy loam, Calcic Hapludolls) and was naturally low in N (4.2  0.3 kg ha 1 nitrate N, n = 3). Soil was mixed in a 1:1 ratio with silica sand (2040 medium, Twin Cities Concrete, West Fargo, ND) and both were pasteurized in an autoclave at 121  C for 30 min. Each pot received 600 g of the sand-soil mix and a basal nutrient solution (10 mg P as potassium phosphate, 10 mg S as potassium sulfate, and 25 mg N as calcium nitrate tetrahydrate per pot). The high N treatment received 50 mg N at planting, and 21 days after planting (DAP) received an additional 10 mg per pot on a weekly basis until the end of the experiment. Soybean seeds (Dairyland 401) were surface sterilized by soaking in a 5% bleach solution for 10 min then rinsed several times with distilled water. All commercial inoculants were mixed according to package instructions and the following amounts added to 15 g of seed: B, 0.050 g product; B + D, 0.034 ml product; B + L, 0.030 ml product; B + A, 0.040 ml product. Seeds in the control were inoculated with soil removed from the roots (rhizosphere soil) of field-grown soybean plants collected from Carrington, ND (0.05 g soil per 15 g seed) in order to parallel more closely conditions experienced by control plants in the field experiment. Each pot initially received three seeds per pot, and seedlings were thinned to one per pot 15 DAP. Pots were watered gravimetrically on a daily basis. Because thrips (Thysanoptera) were present in the greenhouse, all plants were sprayed with PyGanic1 (6.57 g/l) 18 DAP. 2.2.2. Aphid infestation and data collection Soybean plants were infested with 10 adult soybean aphids 26 DAP, when plants were at the V2–V3 growth stage (second or third trifoliate fully expanded). Aphids were added to the adaxial leaf surface of the second trifoliate of each plant using a small paintbrush. Plants were set within a larger plastic pot and covered with a mesh drawstring bag (48  71 cm; DC3148-P, Bugdorm, Megaview Science Co., Ltd., Taiwan) supported by crisscrossed wires. Aphid densities and plant growth stages were assessed after 24 h (adults and immatures counted separately) and 7 days after infestation (33 DAP). Plants were destructively sampled 39 DAP, which involved cutting stems at the soil surface, extracting and washing roots, counting all root nodules, and drying all plant material in an oven (48 h at 70  C). Roots and shoots (i.e., aboveground biomass) were weighed separately on in order to calculate root:shoot mass ratios.

74

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

2.3. Analysis of nitrogenous compounds within above-ground plant tissue Due to problems with mold on plants sampled in 2011, nitrogenous compounds were only assessed in plants sampled in 2010. Dried plant tissue was ground using a coffee grinder (F203, Krups, Millville, NJ) and stored in airtight containers. A small sample (0.2–0.4 g) from each plant was sent to the NDSU Soil Testing lab for analysis of total nitrogen (by the Kjeldahl method). 2.3.1. Nitrate and ureide analyses Samples were digested by placing approximately 0.4 g (weighed to 0.001 g) of ground plant material and 20 ml of distilled water in a 40 ml screw top vial (Pyrex, VWR International), and heating at 90  C for 30 min in a water bath. After cooling to room temperature, samples were filtered (Whatman paper filter, size 2, VWR) into a plastic container containing 2 g of H+ resin beads (Dowex1 MarathonTM C hydrogen form, Sigma–Aldrich, St Louis, MO), which were recharged after use. Nitrate content was determined according to the salicylic acid method (Cataldo et al., 1975). Initially, 0.8 ml of a salicylic:sulfuric acid mixture (3 g in 60 ml; Sigma–Aldrich) was added to a vial and 0.2 ml of filtered sample mixed in and left to react for at least 10 min. Then, 19 ml of a 2 M NaOH solution was added to a tube and allowed to cool to room temperature, after which 1 ml of the sample was pipetted into an acrylic cuvette. Absorbance at 420 nm was recorded in a UV–vis spectrophotometer (Beckman Coulter DU 530). The concentration of nitrate in each sample was determined from a calibration line constructed from the absorbances of six standards (0, 5, 10, 15, 20, and 40 mg/kg of nitrate) prepared similarly to the samples. Ureide content in soybean tissue was determined by the method of Patterson et al. (1982). Briefly, 1 ml of a filtered sample (see above) was placed in a glass vial along with 1.0 ml of 0.2 M potassium hydrogen phthalate buffer (Sigma–Aldrich) and 0.5 ml of dilute bleach solution (1:3 v/v), and mixed by shaking. After 5 min, 2 ml of color-developing solution (15 ml of 0.5 M NaOH and 40 ml of 4.1 M phenol in 30:70, water: methanol) was added to each vial and left to react for 10 min, after which 5.5 ml of distilled water was added. Approximately 1 ml of the sample was transferred into an acrylic cuvette and its absorbance at 625 nm wavelength recorded in the UV–vis spectrophotometer. The concentration of each sample was determined from a calibration line constructed from the absorbances of six standards (0, 5, 10, 15, 20, or 40 mg/kg of the ureide allantoin, purchased from Sigma– Aldrich) prepared similarly to the samples. Relative abundance of nitrate was calculated as [(nitrate N/ureide N + nitrate N)  100] (adapted from Herridge, 1982). 2.3.2. Amino acids Concentrations of free amino acids were quantified within above-ground soybean tissue (one plant from each +aphid cage). We particularly focused on ASN because this is the primary amino acid translocated from soybean roots to the shoots (McClure and Israel, 1979). First, 1.0 g dried plant material was combined with 15 ml 70% (v/v) ethanol in a glass test tube and mixed at 200 RPM at 30  C for 1 h (VWR 1585 shaking incubator). Then, 1 ml of the mixture was pipetted into a 1.5 ml microtube and centrifuged (10,000  g, Eppendorf 5415 R) for 10 min at room temperature, before 200 ml of the supernatant was combined with 20 nmol (in 100 ml) of the internal standard norvaline) and derivatized using the EZ:faastTM kit for free (physiological) amino acids (Phenomenex1, Torrance, CA). One protocol modification was that amino acids were redissolved in 30 mL of a 80:20% heptane:dichloromethane solution instead of Reagent 6. The calibration solution, containing norvaline and all three amino acids standards

(200 nmol/ml of each amino acid) provided with the kit, was derivatized by the same method. Samples were manually injected into a gas chromatograph (GC; 5890 Series II Plus, Hewlett– Packard) coupled to a quadrupole mass spectrometer (MS; 5972 Series, Hewlett–Packard) equipped with a 10 m  0.25 mm i.d. ZB-AAA column (Phenomenex). The MS was operated in the electron ionization mode with an ionizing voltage of 70 eV. The amounts of amino acids in samples were calculated using peak areas of the amino acids and norvaline internal standard from the total ion chromatogram and allowing for the differential responses of the MS to the different compounds (determined from the standards); i.e., the amount of amino acid in a sample = (peak area of the amino acidsample)/(peak area of norvalinesample)  (peak area of norvalinestandard)/(peak area of amino acidstandard)  200 nmol]. The concentration of total free amino acids was calculated by summing the concentrations of the following amino acids: AAA (a-aminoadipic acid), ABA (a-aminobutyric acid), bAIB (b-aminoisobutyric acid), ALA, APA (a-aminopimelic acid), ASN, ASP, HIS, GLN, GLU, GLY, aILE (allo-isoleucine), LEU, LYS, MET, PHE, PRO, TRP, TYR, VAL. 2.3. Data analysis All data were analyzed using SYSTAT1 (SYSTAT Software Inc., 2007). Histograms and Levene’s test for equality of variance were to determine if data met assumptions necessary for parametric statistics. Data points that were identified by the statistical program as outliers or that were due to specific methodological issues (e.g., high predator densities within a cage, contamination by natural aphid populations) were removed from analyses. 2.3.1. Field experiment For the field experiment, data from the two plants within each cage were averaged for analyses, and data on mean aphid density per plant were log (X + 1) transformed. We used repeated measures ANOVA with aphid treatment as the independent variable and aphid density as the dependent variable in order to assess differences in aphid populations between purposefully infested +aphid cages and no-aphid cages. We also used repeated measures ANOVA to assess the impacts of inoculant treatment on aphid populations, with aphid density as the explanatory variable. Due to differences in timing of sampling between years, data were analyzed separately for each year. The number of vegetative (V) soybean growth stages is partially dependent on environmental conditions, i.e., the number of trifoliate leaves produced prior to the first reproductive stage (i.e., R1, flowering) can vary (Pedersen and Elbert, 2009). Although we assessed plant growth stage on each sampling date, we did not count the number of trifoliates on R1 plants, and so do not know exactly how many V growth stages each experimental field plant went through. We therefore used data from dates where all plants were in V or R growth stages to analyze effects of aphids and inoculants on this parameter. Data were converted to numeric values for analysis (e.g., R1 = 1) and averaged across both plants within a cage. Data from explanatory variables that were assessed once at the end of both seasons (i.e., change in growth stage, dry weight of above-ground biomass, number of pods per plant) were analyzed using factorial ANOVA with year, aphid infestation, and inoculant treatment as the independent variables. Data on parameters only assessed for one year (2011 root nodule density, N content of above-ground biomass in 2010; concentration of total N, ureide and nitrate, and the relative abundance of nitrate) were assessed using factorial ANOVA with aphid infestation and inoculant treatment as the independent variables. Data on 2010 nodule density from adjacent non-experimental plots, lacking the aphid

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

infestation treatment, were analyzed using ANOVA with inoculant treatment as the sole independent variable. The concentration of free amino acids (ASN and all amino acids combined) was only assessed in +Aphid plants in 2010; thus, data were analyzed using ANOVA with inoculant treatment as the independent variable after a log transformation. For all analyses using ANOVA, if the overall P-value was significant (defined as P  0.05), Fisher’s LSD test was used for mean separation among treatments. Strength of linear correlations between aphid densities and plant parameters were determined using Pearson correlation coefficients and R2 values (coefficient of determination). The significance of the correlations (i.e., the likelihood that the correlation coefficient would occur if there were no relationship between the variables) was determined using Bonferroni probabilities as part of the Pearson correlation analysis. We focused on the number of nodules per plant and concentration of N compounds in the above-ground tissue as the most relevant indicators of host plant quality. 2.3.2. Greenhouse experiment ANOVA and Fisher’s LSD post hoc test with inoculant treatment as the independent variable were used to analyze treatment effects on aphid densities at 24 h (adults and juveniles per adult) and 7 days (adults + juveniles) and on the following plant parameters: number of root nodules, root:shoot ratios, dry weight of above-ground biomass, concentration of total N, nitrates and ureides, and the relative abundance of nitrate in the above ground plant tissue, with each dependent variable analyzed separately. No plants reached an R stage during the course of the experiment, and therefore, we were able to calculate the change in growth stage from 24 h to 7 days, and used ANOVA to determine if inoculant treatment had a significant impact on this parameter. Because of the short duration of the experiment, effects of aphids on the plants were minimized, and thus, we used least squares multiple linear regression to determine relationships between final aphid densities at day 7 (i.e., the dependent variable) and independent variables associated with plant growth and N status. We eliminated variables that were highly correlated with each other (correlation coefficient > 0.7). We then used best subsets regression and associated statistics to select the most parsimonious model from the remaining variables [density of adult aphids at 24 h, nymphs per adult at 24 h, growth stage at 24 h, change in growth stage from 24 h to 7 d, dry weight of above ground biomass, root to shoot ratio, density of root nodules, and concentration of ureide N, nitrate N (log transformed), and total N in aboveground biomass] (Quinn and Keough, 2002). We then used the subset of explanatory variables from the best subsets analysis in an additive general linear model. 3. Results 3.1. Field experiment 3.1.1. Aphid density In 2010, aphid treatments were established, with essentially no aphids detected in the no-aphid cages (24 h: A = 0  0, +A = 4.1  0.3). The difference in aphid densities was more pronounced at the end of the season, leading to a significant time  aphid interaction (time  aphid, df5,345,P < 0.001; aphid, df1,69, P < 0.001; 61 days: A = 19.3  3.3, +A = 12.9  104  2.9 4  10 ). However, in 2011, despite insecticide applications, no-aphid cages were contaminated with soybean aphids. Initially, densities were similar between cage types on the first sampling date (24 h: P = 0.831, A = 31.2  7.5, +A = 21.4  3.6), but became significantly higher in +aphid cages for the remainder of the season (time  aphid, df3,174, P < 0.001; aphid, df1,58,; 14 days: P < 0.001,

75

A = 62.4  11.2, +A = 826.3  84.7; 33 days: P < 0.001, A = 687.5  76.3, +A = 1502.2  106.5; 49 days: P = 0.030, A = 630.6  71.1, +A = 953.9  85.5). Even though the difference was significant, aphid densities in no-aphid cages in 2011 were still well above zero and were likely impacting plant growth and physiology (Ragsdale et al., 2007; Beckendorf et al., 2008; Riedell et al., 2009). In 2010, aphid densities increased through time (time, df5,145, P < 0.001), although the impact of inoculant treatment on aphid densities in +aphid cages was not consistent over time (time  inoculant, df25,145, P = 0.012; Fig. 1a). Initially, inoculant treatment did not impact aphid densities on experimental plants (inoculant, df5,30, 24 h: P = 0.775; 8 days: P = 0.484). However, by 22 days there were significant differences in aphid densities among treatments (inoculant, df5,30, P = 0.046), with lower aphid densities on B and B + D plants compared to control, high N and B + A plants. This pattern continued throughout the season (inoculant, df5,30, 36 days: P = 0.072; 56 days: P = 0.045; 61 days: P = 0.032; Fig. 1b). In 2011, aphid densities increased through time (time, df3,66, P < 0.001), but in contrast to 2010, the impact of inoculant treatment on aphid densities was consistently non-significant (time  inoculant, df15,66, P = 0.966; inoculant, df15,22, P = 0.675; Fig. 1c–d). However, 33 days after infestation, a large proportion of dead aphids was found in all +aphid cages, which likely influenced aphid densities prior to discovery and contributed to the decline in populations seen on day 49. 3.1.2. Plant growth parameters In 2010, at the start of the experiment (28 June), all experimental plants were pre-reproductive (9.0% V2, two fully expanded trifoliate leaves; 88.2% V3, three trifoliates; 2.8% V4, four trifoliates). On 6 July (8 days after aphid infestation), 67.4% of plants were still in a V stage (V4–V13), whereas 32.6% were at R1 (beginning bloom). On 22 July (22 days), 97.2% of plants were at R2 (full bloom) and 2.8% were at R3 (beginning pod). On 4 August (36 days), 9.0% were at R3 and 91.0% were at R4 (full pod). On 19 August (56 days), 2.8% were at R4, 68.7% were at R5 (beginning seed), and 28.5% were at R6 (full seed). On the last sampling date (28 August 2010, 61 days), all plants were at the R6 stage. Neither aphid infestation nor inoculant treatment significantly impacted plant growth stage on 28 June (aphid  inoculant, df5,60, P = 0.130; aphid, df1,60, P = 0.821; inoculant, df5,60, P = 0.450) or 19 August (aphid  inoculant, df5,60, P = 0.179; aphid, df1,60, P = 0.457; inoculant, df5,60, P = 0.343; Table 1). In 2011, experimental plants were in a mix of V (22.5%, V4–V8) and R stages (77.5% all R1) 24 h after aphid infestation (22 July). On 4 August (14 days), 10.8% were at the R2 growth stage, 84.2% were at R3 and 5.0% were at R4 (Table 1). Neither experimental treatment had a significant impact on plant growth stage on this date (aphid  inoculant, df5,46, P = 0.977; aphid, df1,46, P = 0.438; inoculant, df5,46, P = 0.213). On 23 August (33 days), 99.2% were in the R5 stage and on 8 September (49 days) 100% were in the R6 stage. In 2010, within adjacent non-experimental plots (i.e., plants not caged, no aphids purposefully added), inoculant treatment had a significant impact on the mean number of root nodules per plant (inoculant, df5,30, P = 0.008), with the highest density occurring on plants in the B and B + D treatments, and the lowest in the high N and control treatments (CNTL, i.e., plants associated with naturally-occurring rhizobia; Table 1). In 2011, there were no significant impacts of aphid infestation or inoculant treatment on the mean number of nodules per plant (aphid  inoculant, df5,34, P = 0.911; aphid, df1,34, P = 0.662; inoculant, df5,34, P = 0.791). The latter may be related to denitrification or the urea being leached out of the root zone of the young soybean plants, as there were several major rain events in 2011, with plots receiving 14.2 cm of rainfall within a 2 week period (14–27 June, North Dakota

76

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

6

6

a) 2010 CNTL N B B+D B+L B+A

5

Mean aphids per plant (log X+1) ± SEM

4 3 2 1 0 6

*

ns

*

4

*

a

ns

a

a b

3

b

ab

2

ns

1

0 0

10

20

30

40

50

CNTL

60

N

B

B+D

B+L

B+A

6

c) 2011

d) 2011, d49

5

5

4

ns

ns

4

ns

3 2

b) 2010, d61 5

3

ns

2

1

a

a

a

a

a

a

N

B

B+D

B+L

B+A

1

0

0

0

10

20

30

40

50

Days after aphid infestation

60

CNTL

Inoculant treatment

Fig. 1. Soybean aphid density per plant (log transformed) throughout the season in 2010 (a) and 2011 (b) (non-destructively sampled) and at the end of the experiment in 2010 (b) and 2011 (d) (last date destructively sampled) for each inoculant treatment. CNTL: non-inoculated control where plants associated with existing soil microbes; N: non-inoculated high soil nitrogen treatment; B: plants inoculated with Bradyrhizobium japonicum; B + D: plants inoculated with B. japonicum + Delfia acidovorans; B + L: plants inoculated with B. japonicum + lipo-chitooligosaccharides; B + A: plants inoculated with B. japonicum + Azospirillum brasilense. Data are means  standard errors and are from +Aphid cages only. Asterisks and different letters indicate significant differences within sample dates at P  0.05.

Agricultural Weather Network, http://ndawn.ndsu.nodak.edu/) that resulted in standing water within the plots. The impact of aphid infestation on total above-ground biomass varied by year (year  aphid  inoculant, df5,103, P = 0.748; year  aphid, df1,103, P = 0.005; year  inoculant, df5,103, P = 0.240; year, df1,103, P < 0.001). In 2010, aphid infestation negatively impacted total above-ground biomass regardless of inoculation treatment (aphid  inoculant, df5,57, P = 0.510; aphid, df1,57, P < 0.001; A = 95.00  2.83, +A = 75.36  3.11). Additionally, inoculation treatment had a significant impact on above-ground biomass (inoculant, df5,57, P = 0.018), which was due to significantly lower biomass in the high N and B + D treatments compared to the others (Table 1). However, in 2011, there was no impact of either aphid infestation or inoculant on the total above-ground plant biomass (aphid  inoculant, df5,46, P = 0.993; aphid, df1,46, P = 0.441; inoculant, df5,46, P = 0.936; A = 66.83  3.38, +A = 63.70  2.48; Table 1). Effects of aphid infestation on the number of pods per plant paralleled results for plant biomass and also varied from year to year (year  aphid  inoculant, df5,106, P = 0.631; year  aphid, df1,106, P = 0.001; year  inoculant, df5,106, P = 0.307; year, df1,106, P = 0.202). In 2010, the number of pods per plant was lower on plants infested with aphids, regardless of inoculant treatment (aphid  inoculant, df5,60, P = 0.540; aphid, df1,60, P < 0.001; A = 131.01  4.12, +A = 102.51  3.93). In addition, there was a significant effect of inoculant on pod density (inoculant, df5,60, P = 0.038) that was primarily driven by lower numbers of pods in the high N treatment (Table 1). In 2011, there were no significant effects of aphid or inoculant treatments on pod densities (aphid  inoculant, df5,46, P = 0.901; aphid, df1,46, P = 0.655; inoculant, df5,46, P = 0.709; A = 113.00  3.83, +A = 110.02  3.11; Table 1).

3.1.3. Nitrogenous plant compounds Nitrogenous compounds within above-ground plant tissue were only assessed in 2010 (Table 1), and not in 2011, because of problems with mold on stored plant samples. No-aphid plants had a significantly higher total N concentration (g/kg) compared to +aphid plants (aphid, df1,59, P = 0.010; A = 24.90  0.76, +A = 22.38  0.54). However, inoculant treatment did not have a significant impact on total N (inoculant  aphid, df5,59, P = 0.904; inoculant, df5,59, P = 0.660). Effects of aphids on the concentration of ureides and nitrates (g/kg), and the relative abundance of nitrates in above-ground plant tissue were not significant (ureides: aphid, df1,60, P = 0.242, A = 1.02  0.06, +A = 0.92  0.06; nitrates: Aphid, df1,59, P = 0.171, A = 0.34  0.02, +A = 0.38  0.03; relative abundance of nitrates: Aphid, df1,60, P = 0.201, A = 27.38  2.03, +A = 31.29  2.18). Likewise, there was no impact of inoculant on these parameters (ureides: inoculant  aphid, df5,60, P = 0.928; inoculant, df5,60, P = 0.874; nitrates: inoculant  aphid, df5,59, P = 0.088; inoculant, df5,59, P = 0.056; relative abundance of nitrates: inoculant  aphid, df5,60, P = 0.682; inoculant, df5,60, P = 0.461). There was no impact of inoculant treatment on the amount of free ASN in above-ground plant tissue within +aphid plants (df5,30, P = 0.476), although, concentrations of all free amino acids were significantly lower in B and B + D plants compared to high N and B + A (df5,30, P = 0.038; Table 1). 3.1.4. Relationship between aphid densities and plant parameters In 2010, there was a negative correlation between aphid density per plant (log transformed, 61 days) in each treatment and the average number of root nodules per plant in each treatment on a

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

77

Table 1 Data on plant parameters from field and greenhouse experiments for each inoculant treatment. Experiment

Year

Field Growth stage 6–28 Growth stage 8–19 Growth stage 8–4 # root nodules/planta

[Total N] (g/kg) [Ureide] (g/kg) [Nitrate] (g/kg) Relative nitrate (%) [Free ASN] nmol/g [All free amino acids] mmol/g

2010 2010 2011 2010 2011 2010 2011 2010 2011 2010 2010 2010 2010 2010 2010

Greenhouse Growth stage 24 h Growth stage day 7 Change in growth stage (24 h to day 7) # root nodules/plant Root:shoot ratio Foliar biomass/plant (g) [Total N] (g/kg) [Ureide] (g/kg) [Nitrate] (g/kg) Relative nitrate (%)

n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

Above-ground biomass/plant (g) # pods/plant

Inoculant treatment CNTL

N

B

B+D

B+L

B+A

V2.9  0.1 a R5.1  0.1 a R2.8  0.1 a 31.6  3.7 a 72.4  5.8 a 88.2  6.3 a 65.7  4.9 a 122.6  7.9 a 109.0  5.2 a 23.9  1.4 a 0.99  0.11 a 0.33  0.03 a 27.0  2.7 a 99.6  9.4 a 2.6  0.4 ab

V2.8  0.1 a R5.5  0.1 a R3.2  0.1 a 24.0  1.9 a 81.6  10.7 a 70.5  6.4 b 67.0  6.0 a 98.6  8.0 b 111.4  5.7 a 22.0  1.1 a 0.87  0.10 a 0.45  0.05 a 36.0  4.7 a 155.9  46.5 a 2.8  0.3 a

V2.9  0.1 a R5.3  0.1 a R3.0  0.1 a 45.2  4.3 bc 68.7  9.8 a 90.3  4.7 a 68.2  4.2 a 127.5  7.0 a 121.0  5.9 a 24.3  1.5 a 0.97  0.09 a 0.32  0.03 a 26.5  3.6 a 98.6  10.2 a 2.0  0.3 b

V2.9  0.1 a R5.3  0.1 a R2.9  0.1 a 49.4  4.6 c 82.7  10.6 a 79.9  3.9 b 60.4  6.9 a 110.3  5.5 ab 108.7  7.6 a 23.0  1.2 a 0.93  0.11 a 0.33  0.03 a 30.6  4.6 a 91.2  12.0 a 2.0  0.2 b

V3.0  0.1 a R5.2  0.1 a R3.0  0.1 a 34.1  5.9 ab 77.2  8.7 a 90.3  5.3 a 66.1  5.8 a 119.0  7.3 a 112.5  7.3 a 24.1  1.0 a 1.02  0.09 a 0.39  0.03 a 28.3  1.9 a 211.8  84.4 a 2.6  0.3 ab

V3.0  0.0 a R5.3  0.1 a R3.0  0.1 a 35.7  6.4 ab 75.1  14.4 a 90.2  7.2 a 63.6  4.5 a 122.7  10.5 a 108.1  5.3 a 24.4  1.0 a 1.05  0.13 a 0.35  0.04 a 27.6  3.8 a 119.6  11.6 a 3.0  0.8 a

V2.1  0.1 a V3.6  0.2 a 1.4  0.2 a 53.3  3.4 c 0.42  0.02 b 2.21  0.09 b 17.9  1.3 bc 0.65  0.06 d 0.44  0.04 a 41.3  3.5 b

V2.3  0.2 a V3.1  0.1 a 0.78  0.2 b 43.0  5.1 c 0.37  0.02 b 2.21  0.12 b 15.4  0.8 bc 0.35  0.06 b 0.37  0.04 a 52.1  5.1 b

V2.3  0.2 a V3.3  0.2 a 1.0  0.0 b 19.9  7.1 b 0.40  0.04 b 2.14  0.08 ab 14.1  1.8 b 0.17  0.08 a 0.47  0.08 a 76.4  10.0 a

V2.1  0.1 a V3.1  0.1 a 1.0  0.2 b 00 a 0.30  0.02 a 1.85  0.11 a 24.6  1.5 a 0.40  0.03 b 1.61  0.18 b 79.8  1.3 a

V2.1  0.1 a V3.2  0.2 a 1.1  0.1 ab 41.8  4.7 c 0.42  0.02 a 1.87  0.13 a 17.6  1.4 c 0.60  0.08 cd 0.49  0.06 a 46.3  4.9 b

V2.3  0.2 a V3.3  0.2 a 1.0  0.0 b 44.7  4.5 c 0.39  0.01 b 2.40  0.13 b 16.5  0.8 bc 0.42  0.06 bc 0.44  0.03 a 52.5  3.8 b

Data are means  standard errors. Means within the same row with different letters are significantly different at P  0.05. CNTL: control where plants were associated with existing/native soil microbes; N: non-inoculated high soil nitrogen treatment; B: plants inoculated with Bradyrhizobium japonicum; B + D: plants inoculated with B. japonicum + Delfia acidovorans; B + L: plants inoculated with B. japonicum + lipo-chitooligosaccharides; B + A: plants inoculated with B. japonicum + Azospirillum brasilense. Includes: AAA, ABA, ßAIB, ALA, APA, ASN, ASP, HIS, GLN, GLU, GLY, aILE, LEU, LYS, MET, PHE, PRO, TRP, TYR, VAL. a 2010 data from adjacent plots, 2011 data from experimental plants.

Mean aphids per plant (log X+1)

plot-wide scale (R2 = 0.755, x2 = 4.922, P = 0.027; Fig. 2a). In 2011, a similar pattern occurred when using aphid (log transformed, 49 days) and nodule data from experimental plants (R2 = 0.623, x2 = 19.025, P < 0.001; Fig. 2b). Correlations between aphid densities and ureide, ASN, and total amino acid concentrations in above-ground tissue were not significant (ureides: R2 = 0.032, x2 = 1.048, P = 0.306; ASN: R2 = 0.019, x2 = 0.638, P = 0.424; all amino acids: R2 = 0.035, x2 = 1.175, P = 0.278; data not shown). However, there was a weak positive relationship between aphid densities and nitrate concentrations (R2 = 0.126, x2 = 4.386,

5.6

P = 0.036) and the relative abundance of nitrates in foliar plant tissues (R2 = 0.118, x2 = 4.084, P = 0.043; data not shown). 3.2. Greenhouse experiment 3.2.1. Aphid density Inoculation treatment did not impact the establishment of adult aphids on host plants (inoculant, df5,48, P = 0.083), and there were no treatment effects on nymphs per adult after 24 h (inoculant, df5,48, P = 0.203) or total aphid densities (adults + immatures) after

3.6

CNTL N 3.4 B B+D B+L 3.2 B+A

a) 2010

5.2 4.8 4.4

b) 2011

3.0 4.0 3.6

2.8

R2 = 0.755 P = 0.027

3.2 20

25

30

35

40

45

50

2.6 55 0

R2 = 0.623 P < 0.001 20

40

60

80

100

120

140

Mean root nodules per plant Fig. 2. Correlations between (a) plot-level correlations between mean soybean aphid densities (log transformed) on 28-Aug 2010 and mean density of root nodules per plant. Root nodule data from adjacent non-experimental plots (i.e., uncaged plants not infested with aphids), (b) mean soybean aphid densities (log transformed) on 23-Aug 2011 and mean root nodules per experimental plant. Data from +Aphid cages only.

78

80

7d 70

60

a a

a

ab

50

40

b

b

B+D

B+L

30 CNTL

N

B

B+A

Inoculation treatment Fig. 3. Mean soybean aphid densities per plant 7 days after inoculation in a greenhouse experiment. CNTL = control inoculated with native soil microbes; N: non-inoculated high soil nitrogen treatment; B: plants inoculated with Bradyrhizobium japonicum; B + D: plants inoculated with B. japonicum + Delfia acidovorans; B + L: plants inoculated with B. japonicum + lipo-chitooligosaccharides; B + A: plants inoculated with B. japonicum + Azospirillum brasilense. Data are means  standard errors. Different letters indicate significant differences at P  0.05.

24 h (inoculant, df5,48, P = 0.075). However, at the end of the experiment (day 7) total aphid densities were significantly lower on B + D and B + L plants compared to those in control, high N and B + A treatments (inoculant, df5,48, P = 0.025; Fig. 3). 3.2.2. Plant growth parameters The majority of plants were at the V2 growth stage 24 h after aphid infestation and at the V3 stage on d 7 (Table 1). Inoculation treatment did not significantly impact growth stage on either sampling date (inoculant, df5,48 24 h, P = 0.598; day 7, P = 0.286). However, B + D plants were progressing through growth stages faster than those in the other treatments (inoculant, df5,48, P = 0.016; Table 1). Nodulation was completely suppressed on plants in the high N treatment and the number of root nodules on control plants inoculated with field soil was significantly lower than plants receiving commercial inoculants (inoculant, df5,48, P < 0.001; Table 1). Root to shoot ratios were significantly lower in the high N treatment (i.e., roots were proportionally smaller than shoots) compared to other plants (inoculant, df5,48, P = 0.002). Inoculant treatment also significantly impacted the dry weight of aboveground plant tissue (inoculant, df5,48, P = 0.005), with plants in the high N and B treatments being lighter than the others. 3.2.3. Nitrogenous plant compounds Total N concentrations were greatest in high N plants, followed by B + D and B plants (inoculant, df5,48, P < 0.001; Table 1). Inoculant treatment also had a significant impact on the concentration of ureide N in above-ground plant tissue (inoculant, df5,48, P < 0.001), with levels highest in B and B + D plants and lowest in the control (Table 1). Nitrate concentrations were highest in the high N treatment and similar among inoculated and control plants (inoculant, df5,48, P < 0.001). Plants in the high N treatment had a greater relative abundance of nitrate in their aerial tissues than other plants, as did control plants inoculated with field soil (inoculant, df5,48, P < 0.001). 3.2.4. Relationship between aphid densities and plant parameters Based on the corrected AIC, Schwarz’s BIC, and Mallows’ Cp, the best subsets model that most parsimoniously explained variation

in total aphid density on day 7 consisted of the following four explanatory variables: density of root nodules, concentration of nitrate in the above-ground tissue, change in growth stage from 24 h to day 7, and adult aphid density at 24 h. Models based on the uncorrected AIC and adjusted R2 included those four variables, as well as dry weight of above-ground biomass and concentration ureide N in the above-ground plant tissues. These six variables were used in a general linear model. There was a weak non-linear relationship (i.e., polynomial, 2nd order, adjusted R2 = 0.140, P = 0.022) between ureide N concentration and aphid density. However, the model fit was not improved by the inclusion of the second order term, even if it was standardized or centered, and both the first and second order terms were not significant (P > 0.05), so these parameters were not included in the final model. The remaining five variables explained approximately 47% of variability in total aphid densities at the end of the experiment (adjusted R2 = 0.474, n = 53, SE of estimate = 12.97). Nodule density was negatively related to aphid densities (Fig. 4), and was the best predictor of aphid densities (standardized regression coefficient (b) = 0.661, t = 4.88, P < 0.001), followed by the concentration of nitrate N in the aerial tissue (b = 0.483, t = 3.54, P = 0.001), change in growth stage (b = 0.353, t = 3.42, P = 0.001), densities of adults after 24 h (b = 0.266, t = 2.61, P = 0.012) and dry weight of above-ground biomass (b = 0.238, t = 2.07, P = 0.044). Results were similar if root nodule weight was used instead of nodule density, although slightly less variation was explained (adjusted R2 = 0.415, n = 53, SE of estimate = 13.67). 4. Discussion Microorganisms associated with plant roots have important impacts on plant growth and physiology, including enhancing absorption or availability of nutrients, and affecting the production of compounds involved in defense against insect herbivores (Pineda et al., 2010). Soybean producers often use seed inoculants that contain N-fixing and plant growth-promoting bacteria to enhance plant growth, N-fixation, and yield. Understanding how rhizobia impact organisms at higher trophic levels is important for pest management, especially because effects can depend on bacterial species (Dean et al., 2009). In this study we examined how soybean seed inoculants with different N-fixing and plant growth-promoting bacteria impacted densities of pestiferous soybean aphids, soybean growth, nodulation, and foliar N levels,

Mean aphids per plant (7 d)

Mean aphids per plant ± SEM

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

120

R2 = 0.202 P = 0.002

100 80

CNTL N B B+D B+L B+A

60 40 20 0

20 40 60 Mean root nodules per plant

80

Fig. 4. Relationships between mean soybean aphid densities (7 days) and mean root nodules per plant from a greenhouse experiment. CNTL: control inoculated with native soil microbes; N: non-inoculated high soil nitrogen treatment; B: plants inoculated with Bradyrhizobium japonicum; B + D: plants inoculated with B. japonicum + Delfia acidovorans; B + L: plants inoculated with B. japonicum + lipochitooligosaccharides; B + A: plants inoculated with B. japonicum + Azospirillum brasilense.

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

and examined effects of aphids on plant parameters in the field. We then explored relationships between aphid densities and factors related to soybean N-fixation and foliar N content. In 2011, non-significant aphid and inoculant treatment effects in the field were likely due to three factors: (1) contamination of no-aphid cages with high densities of aphids, (2) the widespread presence of an aphid pathogen, and (3) the failure of additional fertilizer to suppress nodulation in the high N treatment. Therefore, the discussion primarily focuses on data from 2010 and the greenhouse experiment. In the 2010 field experiment, aphid presence negatively impacted plant parameters associated with growth (i.e., aboveground biomass, number of pods) and total N in above-ground plant tissue, which is not surprising as negative effects of soybean aphids on plant growth and yield have been documented (Costamagna et al., 2007; Ragsdale et al., 2007; Beckendorf et al., 2008; Riedell et al., 2009). Although concentrations of ureide N and nitrate N in above-ground foliage were not significantly affected by aphid infestation, at the end of the season the latter tended to be higher within aphid-infested plants. This is somewhat similar to other field experiments (Riedell et al., 2013a), in which there were no effects of soybean aphid infestation on ureide concentrations in foliage, but nitrate concentrations were significantly higher. In our study, the most obvious effects of inoculant treatments on plant parameters were driven by differences between plants receiving additional fertilizer to suppress root nodulation and other treatments. In general, in addition to having fewer nodules and lower levels of N-fixation and ureides in the xylem (Streeter and Wong, 1988; Shelp and Da Silva, 1990), plants primarily getting N from fertilizer are shorter and have decreased biomass and grain yield versus plants that associate with N-fixing bacteria (McClure and Israel, 1979; Herridge et al., 2008; Salvagiotti et al., 2008). In our experiments, plants in the high N treatment still had considerable concentrations of ureides, even in plants in the greenhouse that lacked nodules. Research has shown that ureides can reach high levels in non-nodulated legumes (McNeil and LaRue, 1984), which could be due to remobilization of N and/or de novo synthesis from existing nucleotides (Polayes and Schubert, 1984; Diaz-Leal et al., 2012). Generally, there were few significant differences in plant parameters among soybeans associated with existing soil bacteria and plants inoculated with commercial rhizobia (i.e., B, B + D, B + L, B + A), although, in the greenhouse differences were more pronounced and control plants appeared to be less reliant on N-fixation than were other inoculated plants. One notable difference was that in field plants the concentration of total free amino acids in aerial tissue was lowest in soybeans inoculated with B. japonicum alone (B) and B. japonicum + D. acidovorans (B + D). Inoculant treatment did not have strong effects on aphid establishment on plants, although we used no-choice experiments and did not assess aphid attraction to, or preference for, plants in specific treatments. However, inoculant treatment significantly impacted aphid populations in both field and greenhouse experiments, with inoculant effects apparent after 2 weeks in the field and after 7 days in the greenhouse. Aphid densities on plants in the high N treatment were similar to those on plants associated with existing soil rhizobia (control) and plants inoculated with B. japonicum + A. brasilense (B + A). Plants with B. japonicum + D. acidovorans (B + D) had the lowest aphid populations, while densities on plants inoculated with B. japonicum alone (B) and B. japonicum + lipo-chitooligosaccharides (B + L) were low to intermediate, depending on the experiment. Dean et al. (2009) also found that rhizobial inoculants can impact soybean aphid populations. However, in contrast to our study, they found that the lowest aphid densities occurred on plants associated with

79

naturally occurring rhizobia as opposed to plants with a N fertilizer treatment and a commercial inoculant (B. japonicum; HiStick 2, Becker-Underwood, IA). When examining relationships between aphid densities and plant parameters, there was a strong negative relationship between the former and the number of root nodules, both in the field and the greenhouse. Although one greenhouse study showed soybean aphid feeding reduced nodule volume and N-fixation, there was no impact on root nodule density (Riedell et al., 2009). Therefore, it seems likely that in our experiments root nodulation was influencing soybean aphid responses, rather than the converse. We also found that aphid densities were positively related to how fast plants were growing and negatively related to the biomass of aerial tissue. Relationships between foliar nitrate N concentration and aphid densities were not clear cut, as the relationship was positive in the field and negative in the greenhouse. We did not find a strong relationship between ureide N concentrations and soybean aphid densities. This is in contrast to Riedell et al. (2013b), who found a positive association between concentrations of ureide N and soybean aphid populations, as well as noting that aphid populations in the field peaked at the same time as ureide levels. However, the former analysis included samples from across an entire growing season (i.e., the age of the plant was not standardized), and the latter could be due to increased availability of amino N in the phloem, rather than ureides per se, as the authors pointed out. Similarly, Johnson (2008) found that percent leaf N was negatively related to aphid density on evening primrose (Oenothera biennis L.), and speculated that this may have been due to a lack of correlation between leaf N and amino N in the phloem. Compounds in phloem sap, especially amino acids, are particularly relevant for soybean aphids. Although we did not quantify amino acids in the phloem, it is interesting to note that in 2010 the two treatments with the lowest concentrations of free amino acids in aerial tissue also had the lowest densities of soybean aphids, although, correlations between the latter and aphid densities were non-significant. Chiozza et al. (2010) proposed that differences in amino acids influenced host plant resistance against soybean aphid. The negative relationship between N-fixing root nodules and aphid densities could also be related to excess N being used for the production of defensive compounds. As has been previously suggested, differences in legume root nodulation and factors associated with N-fixation could potentially impact herbivorous arthropods, both by altering the amount, identity, and availability of nitrogenous compounds and defensive chemicals within plant tissues. In experiments in which nodulation (and presumably N-fixation) of soybeans were lowered by adding nitrate fertilizer, densities, larval development time, and foliar damage of a chewing herbivore (Mexican bean beetle, Epilachna varivestis Mulsant) decreased as nitrate addition increased (Wilson and Stinner, 1984). Several studies have compared herbivore performance on legumes exposed to rhizobia that were able to nodulate normally and genetically related mutants that lacked the ability to form nodules. On clover (Trifolium repens L.), densities of green peach aphids (Myzus persicae Sulzer) were only marginally higher on nodulated plants (Kempel et al., 2009), while a study on soybean found no effect of rhizobia on soybean aphid densities (Katayama et al., 2011a). With regard to chewing insects and spider mites, several studies showed that herbivore fitness (e.g., abundance, fecundity, and/or weight gain) and diversity were higher on nodulated plants that were more robust and/or had higher foliar N (Kempel et al., 2009; Katayama et al., 2010, 2011a,b). In one study with nodulated soybean, the plants also had lower levels of defensive phenolic compounds (Katayama et al., 2010). However, experiments with clover (Kempel et al., 2009) and lima beans (Phaseolus lunatus L.; Thamer et al., 2011) that used

80

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81

genetically-related lines that either did or did not produce nitrogen-based cyanogenic defensive compounds showed that positive effects of rhizobia on herbivores were lessened or negated in plants producing defensive compounds, and in the latter study herbivore damage was significantly reduced on inoculated plants even though the plants were larger. In summary, this research demonstrates that the identity of rhizobial seed inoculants can affect soybean aphid densities. Additionally, there was a strong negative relationship between the latter and the number of root nodules, although it is unclear if results were due to alterations in host plant nutrition (specifically reduced levels of amino acids in the phloem), increases in plant defensive compounds, or both. Understanding more about how rhizobia impact plants and above-ground herbivores is important for arthropod pest management, especially since numerous land management practices impact the soil microbial community. Acknowledgements Thank you to the anonymous reviewers, technicians at the Carrington ND Research and Extension Center for planting and plot maintenance, Aaron Badillo, Caitlin Berschneider, Jade Monroe, and Bijaya Rai for help with sample collection and processing, the North Dakota Soybean Council and North Dakota State Board of Agricultural Research and Education for funding, and to the NDSU Advance FORWARD program sponsored by the National Science Foundation (HRD-0811239). References Aballay, E., Ordenes, P., Mårtensson, A., Persson, P., 2013. Effects of rhizobacteria on parasitism by Meloidogyne ethiopica on grapevines. Eur. J Plant Pathol. 135, 137–145. Ballhorn, D.J., Kautz, S., Schadler, M., 2013. Induced plant defense via volatile production is dependent on rhizobial symbiosis. Oecologia 172, 833–846. Barazani, O., Friedman, J., 1999. Is IAA the major root growth factor secreted from plant-growth- mediating bacteria? J. Chem. Ecol. 25, 2397–2406. Beckendorf, E.A., Catangui, M.A., Riedell, W.E., 2008. Soybean aphid feeding injury and soybean yield, yield components, and seed composition. Agron. J. 100, 237–246. Bottomley, P.J., 1992. Ecology of Bradyrhizobium and Rhizobium. In: Stacey, G., Burris, R.H., Evans, H.J. (Eds.), Biological Nitrogen Fixation. Chapman and Hall, New York, NY, pp. 293–348 943 pp. Cassán, F., Perrig, D., Sgroy, V., Masciarelli, O., Penna, C., Luna, V., 2009. Azospirillum brasilense Az39 and Bradyrhizobium japonicum E109, inoculated singly or in combination, promote seed germination and early seedling growth in corn (Zea mays L.) and soybean (Glycine max L.). Eur. J. Soil Biol. 45, 28–35. Cataldo, D.A., Haroon, M., Schrader, L.E., Youngs, V.L., 1975. Rapid colorimetric determination of nitrate in plant tissue by nitration of salicylic acid. Commun. Soil Sci. Plant Anal. 6, 71–80. Chiozza, M.V., O’Neal, M.E., MacIntosh, G.C., 2010. Constitutive and induced differential accumulation of amino acid in leaves of susceptible and resistant soybean plants in response to the soybean aphid (Hemiptera: Aphididae). Environ. Entomol. 39, 856–864. Costamagna, A.C., Landis, D.A., DiFonzo, C.D., 2007. Suppression of soybean aphid by generalist predators results in a trophic cascade in soybeans. Ecol. Appl. 17, 441–451. Dean, J.M., Mescher, M.C., De Moraes, C.M., 2009. Plant-rhizobia mutualism influences aphid abundance on soybean. Plant Soil 323, 187–196. Dénarié, J., Debellé, F., Promé, J.-C., 1996. Rhizobium lipo-chitooligosaccharide nodulation factors: signaling molecules mediating recognition and morphogenesis. Annu. Rev. Biochem. 65, 503–535. Diaz-Leal, J.L., Galvez-Valdivieso, G., Fernandez, J., Pineda, M., Alamillo, J.M., 2012. Developmental effects on ureide levels are mediated by tissue-specific regulation of allantoinase in Phaseolus vulgaris L. J. Exp. Bot. 63, 4095–4106. Dobbelaere, S., Vanderleyden, J., Okon, Y., 2003. Plant growth-promoting effects of diazotrophs in the rhizosphere. Crit. Rev. Plant Sci. 22, 107–149. Duzan, H.M., Mabood, F., Zhou, X., Souleimanov, A., Smith, D.L., 2005. Nod factor induces soybean resistance to powdery mildew. Plant Physiol. Biochem. 43, 1022–1030. Evans, J., 1982. Response of soybean-rhizobium symbioses to mineral nitrogen. Plant Soil 66, 439–442. Groppa, M.D., Zawonzink, M.S., Tomaro, M.L., 1998. Effect of co-inoculation with Bradyrhizobium japonicum and Azospirillum brasilense on soybean plants. Eur. J. Soil Biol. 34, 75–80. Han, J., Sun, L., Dong, X., Cai, Z., Sun, X., Yang, H., Wang, Y., Song, W., 2005. Characterization of a novel plant growth-promoting bacteria strain Delftia

tsuruhatensis HR4 both as a diazotroph and a potential biocontrol agent against various plant pathogens. Syst. Appl. Microbiol. 28, 66–76. Herridge, D.F., 1982. Relative abundance of ureides and nitrate in plant tissues of soybean as a quantitative assay of nitrogen fixation. Plant Physiol. 70, 1–6. Herridge, D.F., Peoples, M.B., Boddey, R.M., 2008. Global inputs of biological nitrogen fixation in agricultural systems. Plant Soil 311, 1–18. Johnson, M.T., 2008. Bottom-up effects of plant genotype on aphids, ants, and predators. Ecology 89, 145–154. Katayama, N., Nishida, T., Zhang, Z.Q., Ohgushi, T., 2010. Belowground microbial symbiont enhances plant susceptibility to a spider mite through change in soybean leaf quality. Popul. Ecol. 52, 499–506. Katayama, N., Zhang, Z.Q., Ohgushi, T., 2011a. Community-wide effects of belowground rhizobia on above-ground arthropods. Ecol. Entomol. 36, 43–51. Katayama, N., Zhang, Z.Q., Ohgushi, T., 2011b. Belowground rhizobia positively affect abundances of aboveground sap feeding and leaf chewing herbivores. J. Plant Interact. 6, 173–174. Kempel, A., Brandl, R., Schadler, M., 2009. Symbiotic soil microorganisms as players in aboveground plant-herbivore interactions- the role of rhizobia. Oikos 118, 634–640. Keyser, H.H., Li, F., 1992. Potential for increasing biological nitrogen fixation in soybean. Plant Soil 141, 119–135. Koricheva, J., Gange, A.C., Jones, T., 2009. Effects of mycorrhizal fungi on insect herbivores: a meta-analysis. Ecology 90, 2088–2097. McClure, P.R., Israel, D.W., 1979. Transport of nitrogen in xylem of soybean plants. Plant Physiol. 64, 411–416. McNeil, D.L., LaRue, T.A., 1984. Effect of nitrogen source on ureides in soybean. Plant Physiol. 74, 227–232. Morel, M.A., Ubalde, M.C., Braña, V., Castro-Sowinski, S., 2011. Delftia sp: JD2: a potential Cr(VI)- reducing agent with plant growth-promoting activity. Arch. Microbiol. 193, 63–68. Ohyama, T., Ohtake, N., Sueyoshi, K., Tewari, K., Takahashi, Y., Ito, S., Nishiwaki, T., Nagumo, Y., Ishii, S., Sato, T., 2009. Nitrogen Fixation and Metabolism in Soybean Plants. Nova Science Publishers, New York. Pate, J.S., 1980. Transport and partitioning of nitrogenous solutes. Ann. Rev. Plant Physiol. 31, 313–340. Patterson, T., Glenister, R., LaRue, T., 1982. Simple estimation of ureides in soybean tissue. Anal. Biochem. 119, 90–95. Pedersen, P., Elbert, B., 2009. Soybean Growth and Development. Iowa State University, Extension Publication PM1945, Ames, IA. Pineda, A., Zheng, S.-J., van Loon, J.J.A., Pieterse, C.M.J., Dicke, M., 2010. Helping plants to deal with insects: the role of beneficial soil-borne microbes. Trends Plant Sci. 15, 507–514. Polayes, D.A., Schubert, K.R., 1984. Purine synthesis and catabolism in soybean seedlings: the biogenesis of ureides. Plant Physiol. 75, 1104–1110. Powell, G., Tosh, C.R., Hardie, J., 2006. Host plant selection by aphids: behavioral, evolutionary, and applied perspectives. Annu. Rev. Entomol. 51, 309–330. Prithiviraj, B., Zhou, X., Souleimanov, A., Kahn, W.M., Smith, D.L., 2003. A hostspecific bacteria-to-plant signal molecule (Nod factor) enhances germination and early growth of diverse crop plants. Planta 216, 437–445. Quinn, G.P., Keough, M.J., 2002. Experimental design and data analysis for biologists. Cambridge University Press, New York, NY. Ragsdale, D.W., Voegtlin, D.J., O’Neil, R.J., 2004. Soybean aphid biology in North America. Ann. Entomol. Soc. Am. 97, 204–208. Ragsdale, D.W., McCornack, B.P., Venette, R.C., Potter, B.D., Macrae, I.V., Hodgson, E. W., O’Neal, M.E., Johnson, K.D., O’Neal, R.J., Difonzo, C.D., Hunt, T.E., Glogoza, P.A., Cullen, E.M., 2007. Economic threshold for soybean aphid (Hemiptera: Aphididae). J. Econ. Entomol. 100, 1258–1267. Riedell, W.E., Catangui, M.A., Beckendorf, E.A., 2009. Nitrogen fixation, ureide, and nitrate accumulation responses to soybean aphid injury in Glycine max. J. Plant Nutr. 32, 1674–1686. Riedell, W.E., Beckendorf, E.A., Catangui, M.A., 2013a. Soybean aphid injury effects on shoot nitrogen components in Glycine max. Crop Sci. 53, 232–239. Riedell, W.E., Beckendorf, E.A., Catangui, M.A., 2013b. Relationships between soybean shoot nitrogen components and soybean aphid populations. Arthropod Plant-Interact. 7, 667–676. Rodriguez-Navarro, D.N., Oliver, I.M., Contreras, M.A., Ruiz-Sainz, J.E., 2011. Soybean interactions with soil microbes, agronomical and molecular aspects. Agron. Sustain. Dev. 31, 173–190. Salvagiotti, F., Cassman, K.G., Specht, J.E., Walters, D.T., Weiss, A., Dobermann, A., 2008. Nitrogen uptake, fixation and response to fertilizer N in soybeans: a review. Field Crop Res. 108, 1–13. Schubert, K.R., 1986. Products of biological nitrogen fixation in higher plants: synthesis, transport, and metabolism. Ann. Rev. Plant Physiol. 37, 539–574. Shelp, B.J., Da Silva, M.C., 1990. Distribution and metabolism of xylem-borne ureido and amino compounds in developing soybean shoots. Plant Physiol. 94, 1505–1511. Sindhu, S.S., Dua, S., Verma, M.K., Khandelwal, A., 2010. Growth promotion of legumes by inoculation of rhizosphere bacteria. In: Khan, M.S., Zaidi, A., Musarrat, J. (Eds.), Microbes for Legume Improvement. Springer, Vienna, Austria, pp. 195–235. Souleimanov, A., Prithiviraj, B., Smith, D.L., 2002. The major Nod factor of Bradyrhizobium japonicum promotes early growth of soybean and corn. J. Exp. Bot. 53, 1929–1934. Streeter, J.G., 1972. Nitrogen nutrition of field-grown soybean plants: I. Seasonal variations in soil nitrogen and nitrogen composition of stem exudate. Agron. J. 64, 311–314.

S.M. Brunner et al. / Applied Soil Ecology 86 (2014) 71–81 Streeter, J., Wong, P.P., 1988. Inhibition of legume nodule formation and N2 fixation by nitrate. Crit. Rev. Plant Sci. 7, 1–23. SYSTAT Software Inc, 2007. SYSTAT112 Statistics–I. SYSTAT, Software, Inc., San Jose, CA. Thamer, S., Schadler, M., Bonte, D., Ballhorn, D.J., 2011. Dual benefit from a belowground symbiosis: nitrogen fixing rhizobia promote growth and defense against a specialist herbivore in a cyanogenic plant. Plant Soil 341, 209–219. Ubalde, M.C., Braña, V., Sueiro, F., Morel, M.A., Martinez-Rosales, C., Marquez, C., Castro-Sowinski, S., 2012. The versatility of Delftia sp. isolates as tools for bioremediation and biofertilization technologies. Curr. Microbiol. 64, 597–603.

81

van Berkum, P., Sloger, C., Weber, D.F., Cregan, P.B., Keyser, H.H., 1985. Relationship between ureide N and N2 fixation aboveground N accumulation, acetylene reduction, and nodule mass in greenhouse and field studies with Glycine max L. (Merr). Plant Physiol. 77, 53–58. van Rhijn, P., Vanderleyden, J., 1995. The rhizobium-plant symbiosis. Microbiol. Rev. 59, 124–142. Wilson, K.G., Stinner, R.E., 1984. A potential influence of rhizobium activity on the availability of nitrogen to legume herbivores. Oecologia 61, 337–341.