Legume protein precipitable phenolic and nutrient concentration responses to defoliation and ontogeny

Abstract

A study was conducted to determine how total phenolic (TP), protein precipitable phenolic (PPP), C, and N concentrations, and amount of protein bound (PB) by PPP in leaves of Desmodium paniculatum (panicled tick-clover; PTC) and Lespedeza cuneata (sericea lespedeza; SL) were affected by simulated herbivory and plant ontogeny. All PTC treatments resulted in a decrease (P ≤ 0.05) in TP, PPP, C, and N concentrations and PB between vegetative and seed set stages. All SL treatments resulted in increased (P ≤ 0.05) or stable TP and PPP concentrations from vegetative stage to seed set. The amount of PB was greatest (P ≤ 0.05) in SL plants submitted to 25% defoliation, and flowering and seed set stages had greater (P ≤ 0.05) PB than the vegetative stage. Ontogenesis and defoliation did not (P > 0.05) affect SL N and C concentrations. The protein binding characteristics of PPP from PTC, but not that of SL, appear to be altered in response to stress. Results might correspond with seed dispersal strategies of the two species, with PTC's epizoochory making increased palatability at seed set beneficial.

Keywords:

• condensed tannins
• legumes
• protein precipitable phenolics
• bound protein

1. Introduction

A resurgence of interest in environmentally friendly farming practices and a push for the development of sustainable animal disease control systems which combine non-chemical methods and minimal drug use, have led to research interest in the use of forages containing condensed tannins (CTs). CTs, if used properly, have the potential to suppress ruminant gastrointestinal nematode infections, reduce the need for nitrogen fertilizer inputs, and decrease ruminal methane production (Waller et al. 2001; Waghorn 2008).

The biological activity of particular interest to this study is the ability of CTs to bind to proteins. CTs improve ruminant protein nutrition by binding to plant proteins in the rumen and preventing microbial degradation thereby increasing amino acid flow to the small intestine (Min & Hart 2003). Naumann et al. (2013a) found protein precipitable phenolic (PPP) concentrations and percent total CTs from warm-season legumes were positively correlated (R² = 0.81). Therefore, the PPP concentrations reported in this study may infer percent total CTs.

CTs are hypothesized to be a plant defense against herbivory, and as such, herbivory intensity is thought to play a role in inducing CT concentration changes (Rohner & Ward 1997; Boege 2005; Kohi et al. 2009). Kohi et al. (2009) reported that intermediate herbivory pressure (50% defoliation) yielded mopane (Colophospermum mopane (J. Kirk ex Benth.) J. Kirk ex J. Léonard) regrowth with the greatest CT concentration (1.1 mg/g) of all defoliation treatments, while severe herbivory (75% and 100% defoliation) had no effect on CT concentration. In another study, Rohner and Ward (1997) reported light defoliation, 12 herbivores/km², had no effect on CT concentration of twisted acacia (Acacia raddiana Savi), while heavier defoliation, 26 herbivores/km², increased CT concentration. Defoliating 75% of leaves had no effect on mopane CT concentration (Kohi et al. 2009) but decreased CT concentration in smooth casearia (Casearia nitida (L.) Jacq.; Boege 2005). Decreased CT concentration combined with increased leaf nitrogen concentration (du Toit et al. 1990) resulting from heavy browsing pressure may increase palatability of CT-containing forages.

Native and naturalized North American legumes containing CTs are of interest for ruminant production systems because they are adapted to local environmental conditions (Muir & Bow 2008). Sericea lespedeza (Lespedeza cuneata (Dum. Cours.) G. Don) (SL) and panicled tick-clover (Desmodium paniculatum (L.) DC.) (PTC) are warm-season perennial herbaceous legumes which produce biologically active CTs. Both SL and PTC are effective as natural anthelmintics (Min et al. 2004; Terrill et al. 2009; Cherry et al. 2012), reduce ruminal methane production (Puchala et al. 2005; Animut et al. 2008; Naumann et al. 2013d), and bind moderate to high amounts of protein (64.7 and 78.1 g PB kg⁻¹ for PTC and SL, respectively; Naumann et al. 2013b) making these species suitable candidates for incorporation into pasture and rangeland ruminant production. However, little is known about the effects of repeated defoliation or defoliation intensity on their nutritive value or CT concentrations and properties, that is protein-binding efficiency, as a mechanism of defense following herbivory stress.

Therefore, the objectives of this study were to evaluate how simulated herbivory and plant ontogeny affect (1) C, N, PPP, and total phenolic (TP) concentrations; (2) ratio of PPP to TP (PPP:TP); (3) ratio of C to N (C:N); (4) protein bound (PB) by PPP (g PB kg⁻¹); and (5) binding efficiency (ratio of PB:PPP) of PPP in PTC and SL.

2. Materials and methods

2.1. Plant species

Plant material consisted of leaves from two warm-season perennial herbaceous legumes, PTC and SL. PTC is a North American native (Diggs et al. 1999), and SL, native to Asia, is naturalized throughout the southern USA (Cummings et al. 2007). PTC and SL seeds were germinated in growth chambers in March 2012, and, after three weeks, plants were transplanted in a greenhouse into 1.9 L containers filled with topsoil (A and B horizons) collected at the Texas A&M AgriLife Research Center, Stephenville, TX (32° 15′ N, 98° 12′ W, altitude 395 m). The soil used in the experiment was originally a Windthorst fine sandy loam soil (fine, mixed, active, thermic Udic Paleustalf; pH 6.6, 10 mg kg⁻¹ nitrate-N, 13 mg kg⁻¹ P, 206 mg kg⁻¹ K, 1416 mg kg⁻¹ Ca, 247 mg kg⁻ Mg, 15 mg kg⁻¹ S, 152 mg kg⁻¹ Na, 10.12 mg kg⁻¹ Fe, 0.59 mg kg⁻¹ Zn, 1.10 mg kg⁻¹ Mn, and 0.22 mg kg⁻¹ Cu using Mehlich III extraction; Mehlich 1984). All plants were watered twice daily by an automatic irrigation system for a total of 10 mm/d. Plants received 24-8-16 (24% elemental N, 3.5% elemental P, and 13.3% elemental K) Miracle-Gro^® water soluble fertilizer (The Scotts Company, LLC., Marysville, OH, USA) twice (9 July and 16 July) at 0.19 g kg soil⁻¹.

A total of 172 plants of each species were used in this experiment. Plants were placed randomly by species into one of the five treatments which were arranged randomly on greenhouse tables: undefoliated and three successive defoliations of leaf growth and regrowth from 25%, 50%, 75%, and 100% of the plant canopy height taken from the uppermost portion of the canopy. For each species, the 25%, 50%, 75%, and 100% defoliation treatments each contained 38 plants. Ten plants were treated as undefoliated controls at each of the second and third defoliation events, for a total of 20 undefoliated control plants. At the vegetative defoliation event, all plants were naive, and thus there was no need to designate a group of plants as previously undefoliated. For all defoliated treatments, six or seven individual plants served as an experimental unit giving a total of six experimental units (replications) per defoliated treatment combination. Within treatment, leaves from the six or seven plants were composited in order to accrue sufficient dry matter material to perform laboratory analyses. These composites formed the replicates for the experiment. Once assigned to a composite group, plants remained in their respective composite group throughout the experiment. Dead plants were removed from the experiment before each defoliation event and were not replaced for the remainder of the experiment. At the flowering (second) and seed set (third) defoliation events, 100% of leaves were removed from 10 previously undefoliated control plants. Because fewer control plants were required to accumulate sufficient material for laboratory assay, two control plant composites contained three plants each and one composite contained four plants at each event. Control plants were removed from the experiment once defoliated.

Defoliation took place at 145 (1 August 2012), 181 (6 September 2012), and 222 (17 October 2012) days after seeding, approximately equivalent to vegetative, early reproductive (flowering), and seed set stages. For each defoliation event (1 August, 6 September, and 17 October 2012), leaves were hand plucked according to treatment defoliation percentage. Inclusion of only leaves, rather than including clipped stems, more closely paralleled insect or ruminant browsing. Assays of this material should better reflect forage ingested by small ruminants than whole-plant samples (Muir et al. 2008). At each defoliation event, plant height was measured from the base to the stem apex and leaves were removed from the appropriate percentage of the total height from the uppermost unfurled leaf downward to simulate herbivory by browsing ruminants. The defoliated stem was allowed to remain in place, and at the second and third defoliation events, leaf samples were collected from regrowth from axillary-type buds along the previously defoliated stem in addition to new apical-type growth. Leaves were dried at 55 °C in a forced air oven for 48 h or until weight change ceased, then ground to pass a 1-mm screen in a sheer mill (Wiley Arthur H. Thomas Co., Philadelphia, PA, USA) and stored for later analysis at the Stephenville herbage lab. Leaf material was analyzed for C, N, TP, and PPP concentrations, and the amount of PB per unit of leaf tissue. Percent N and C were estimated using a Vario MACRO C-N Analyzer (Elementar Americas, Inc., Mt. Laurel, NJ, USA). No analyses were completed for 25% defoliation treatment in PTC at seed set, due to insufficient plant material.

2.2. CT purification

CTs were purified using Sephadex LH-20 (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) for subsequent use as a standard from each species by following Naumann et al.'s (2013b) modification of the method described by Wolfe et al. (2008). Briefly, plant tissue (20 g was extracted with 250 ml of 7:3 (v/v) acetone:water). The aqueous portion containing the CTs was retained, and residual acetone was removed by evaporation under reduced pressure. The extract, along with enough 1:1 (v/v) methanol:water to form a slurry, was mixed with Sephadex LH-20, and the slurry was repeatedly washed with 1:1 (v/v) methanol:water until the absorbance at 280 nm was negligible (absorbance ≤ 0.10). CTs bound to the Sephadex were released by washing with 7:3 (v/v) acetone:water, followed by evaporation of residual acetone by air stream/vacuum. The aqueous phase containing CTs was frozen at −80°C and lyophilized.

2.3. Protein precipitability

Hagerman and Butler's (1978) scaled down method was modified to determine protein precipitability of CTs in two duplicate crude plant extracts. Crude plant extracts were prepared by extracting 50 mg of plant tissue with 1 mL 1:1 (v/v) methanol:water on a G10 Gyrotory^® shaker (New Brunswick Scientific Co., Inc., Edison, NJ, USA) for 30 min followed by centrifugation at 16,070× g for 5 min.

To determine PPP, 50 µl of supernatant from crude plant extracts were combined with 250 µl buffer A (0.20 M acetic acid, 0.17 M sodium chloride, pH 4.9), 50 µl bovine serum albumin, and 50 µl 1:1 (v/v) methanol:water and incubated at room temperature for 30 min prior to centrifuging at 16,070× g for 5 min. Supernatant was removed by vacuum aspiration and the protein-phenolic pellet was washed with 250 µl buffer A before re-centrifuging and aspirating. The protein-phenolic pellet was dissolved in 800 µl of SDS/TEA (sodium dodecyl sulfate [1% w/v]-triethanolamine [5% v/v]) before adding 200 µl FeCl₃ (0.01 M FeCl₃ in 0.01 M HCl). Absorbance was read at 510 nm after 30 min and quantified via external standards.

Following the method described by Naumann et al. (2013b) to determine PB, the procedure was carried out as described above, but the protein-phenolic pellet was analyzed for N to quantify precipitated protein. Rather than dissolving the protein-phenolic pellet in SDS/TEA, the pellet was dissolved in 500 µl of buffer A, and the solution was transferred into a foil cup and allowed to dry. A Vario MACRO C-N Analyzer was used to analyze the dried protein-phenolic residue for percent N, which was multiplied by 6.25 (Van Soest 1994) to calculate the amount of PB.

To determine TP, 50 µl of supernatant from the crude plant extract was combined with 850 µl of SDS/TEA before adding 200 µl of FeCl₃. Absorbance at 510 nm was read after 30 min and quantified via external standards as described for the PPP assay.

2.4. Statistical analyses

Data for C, N, TP, PPP, PB, and ratios were analyzed using the GLIMMIX procedure of SAS (SAS Inst., Inc., Cary, NC, USA). The effects of ontogeny, level of defoliation, and their interaction were included in the model. There were six replications per defoliated/ontogeny treatment combination and three per undefoliated treatment. When significant effects were detected in the model, the LSMEANS statement was used to estimate means and mean separations were generated using the LINES option. Differences were considered significant at P ≤ 0.05 unless otherwise stated.

3. Results and discussion

3.1. Carbon concentration

Carbon concentrations were similar for all PTC treatments at vegetative stage (Table 1). No treatments changed C concentration from vegetative to flowering stage. At flowering, plants with 25% defoliation had greater C concentration than undefoliated control and 100% defoliation plants, while plants subjected to 50% and 75% defoliation had intermediate C. Carbon concentrations decreased in all PTC treatments from vegetative to seed set stage, except the undefoliated control in which C concentration did not change. At seed set, C concentration was similar for all PTC treatments. Some of the decrease in PPP concentration from vegetative to seed set stage may be a result of the decrease in C (R² = 0.52) because C comprises a portion of PPP structure. Neither ontogenesis nor defoliation had an effect on SL C concentration.

Table 1. Concentration of C (g kg⁻¹) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	464.1^Footnoteb,Footnotex	462.0^Footnoteb,Footnotex	464.5^FootnoteaFootnoteb,Footnotex	467.1^Footnotea,Footnotex	464.1^Footnoteb,Footnotex	2.9
Flowering	458.0^FootnotebFootnotec,Footnotex	468.0^Footnotea,Footnotex	466.5^Footnotea,Footnotex	461.2^Footnoteb,Footnotex	456.2^Footnotec,Footnotex	2.9
Seed set	447.6^FootnoteaFootnoteb,Footnotey	–	448.9^Footnotea,Footnotey	445.1^Footnoteb,Footnotey	447.5^FootnoteaFootnoteb,Footnotey	2.9
SL^Footnote2
Vegetative	442.3	456.8	448.9	447.1	442.4	10.3
Flowering	448.3	456.5	458.2	444.5	447.8	10.3
Seed set	450.6	453.7	449.7	443.1	442.3	10.3

¹ Ontogenetic stage × treatment interaction, P = 0.0014.

² No effect of ontogenetic stage (P = 0.15), treatment (P = 0.34) or ontogeny × treatment interaction (P = 0.15).

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

3.2. Nitrogen concentration

Nitrogen concentration decreased in undefoliated PTC plants across defoliation events (Table 2). From the vegetative to flowering stage, N concentration increased in PTC plants defoliated at 25%, 50%, and 100% but was unchanged for 75% defoliation. At vegetative stage, N concentration of plants defoliated at 25% was similar to that of 50% defoliation but greater than N concentrations of plants subjected to 75% and 100% defoliation. Undefoliated plants had the lowest N concentration of all defoliation levels at flowering. Results suggest PTC N concentration decreases as leaves age, possibly due to N mobilization to younger, growing tissue. Nitrogen concentration decreased in PTC from flowering to seed set, and at seed set N concentration of leaves from each defoliation treatment was less than the N concentration of that treatment at the vegetative stage. Crude protein in PTC reached a peak of 212 g kg⁻¹ at flowering and a low of 140 g kg⁻¹ at seed set. PTC CP concentrations are similar to or greater than those reported for other native herbaceous legume species in the region (Muir et al. 2005).

Table 2. Concentration of N (g kg⁻¹) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	28.86^Footnoteb,Footnotex	31.23^Footnotea,Footnotey	30.73^Footnotea,Footnotey	29.74^Footnoteb,Footnotex	28.86^Footnoteb,Footnotey	0.5
Flowering	25.15^Footnoted,Footnotey	33.87^Footnotea,Footnotex	32.38^Footnoteb,Footnotex	30.82^Footnotec,Footnotex	31.30^Footnoteb,Footnotec,Footnotex	0.5
Seed set	22.36^Footnotec,Footnotez	-	23.29^Footnoteb,Footnotec,Footnotez	24.07^Footnoteb,Footnotey	26.69^Footnotea,Footnotez	0.5
SL^Footnote2
Vegetative	19.68	22.34	21.06	19.52	19.68	2.8
Flowering	18.68	23.28	19.18	21.52	21.80	2.8
Seed set	17.69	23.22	23.15	22.44	23.48	2.8

¹ Ontogenetic stage × treatment interaction, P < 0.0001.

² No effect of ontogenetic stage (P = 0.11), treatment (P = 0.08) or ontogeny × treatment interaction (P = 0.11).

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y,z Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

At seed set, PTC plants defoliated at 100% had greater N concentration than plants subjected to 50 and 75% defoliation. PTC is characterized by terminal panicled inflorescences on long peduncles (Diggs et al. 1999; Muir & Bow 2008). Seeds usually contain the greatest proportion of legume N (Sinclair & de Wit 1975), and N might have been mobilized from the tissue closest to inflorescences (50% and 75% defoliation leaves) first to provide N in seed, possibly accounting for the lower N concentration in PTC plants subjected to 50 and 75% defoliation than 100% defoliation. Neither ontogenesis nor defoliation had an effect on SL N concentration. Crude protein concentration ranged from 111 to 147 g kg⁻¹ in SL.

3.3. Carbon to nitrogen concentration ratio

The C:N ratio in PTC increased in undefoliated plants, decreased in 25% and 100% defoliation plants and was unchanged for 50% and 75% defoliation plants from vegetative to flowering stage (Table 3). Results can be attributed to the respective changes in N concentrations because C concentration did not change in any treatments from vegetative to flowering stage. The C:N ratio of undefoliated plants and plants undergoing 100% defoliation was similar to those in 75% defoliation but greater than the C:N ratio of plants undergoing 25% and 50% defoliation at vegetative stage. Undefoliated plants had the greatest C:N ratio at flowering, likely a result of having the lowest N concentration at this ontogenic stage. From flowering to seed set, the C:N ratio increased for all PTC treatments. Both C and N concentrations decreased in PTC at seed set, but C:N ratios suggest N concentration decreased to a greater degree than C concentration at seed set when C:N ratio in undefoliated plants was similar to that of 50% defoliation but greater than those plants undergoing 75% and 100% defoliation.

Table 3. Ratio of the concentration of C to N (g C:g N) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	16.10^Footnotea,Footnotez	14.81^Footnoteb,Footnotex	15.15^Footnoteb,Footnotey	15.74^Footnotea,Footnotey	16.10^Footnotea,Footnotey	0.3
Flowering	18.22^Footnotea,Footnotey	13.86^Footnotec,Footnotey	14.48^Footnoteb,Footnotey	14.97^Footnoteb,Footnotey	14.58^Footnoteb,Footnotez	0.3
Seed set	20.06^Footnotea,Footnotex	–	19.29^Footnotea,Footnoteb,Footnotex	18.52^Footnoteb,Footnotex	16.83^Footnotec,Footnotex	0.3
SL^Footnote2
Vegetative	22.51^Footnotea,Footnotey	20.48^Footnoteb,Footnotex	21.36^Footnoteb,Footnotex	23.02^Footnotea,Footnotex	22.51^Footnotea,Footnotex	0.9
Flowering	24.27^Footnotea,Footnotex,Footnotey	19.64^Footnoteb,Footnotex	20.22^Footnoteb,Footnotex,Footnotey	20.88,^FootnoteaFootnoteb,Footnotex,Footnotey	20.65^Footnotea,Footnoteb,Footnotey	0.9
Seed set	26.04^Footnotea, Footnotex	19.84^Footnoteb,Footnotex	19.52^Footnoteb,Footnotey	19.79^Footnoteb,Footnotey	18.85^Footnoteb,Footnotez	0.9

¹ Ontogenetic stage × treatment interaction, P < 0.0001.

² Ontogenetic stage × treatment interaction, P = 0.005.

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y,z Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

The C:N ratio in SL was unchanged in all treatments from vegetative to flowering stage (Table 3). Likewise, from flowering to seed set, the C:N ratio did not change for SL in all defoliation levels. However, from vegetative stage to seed set, C:N ratio in SL increased in undefoliated plants, decreased in plants undergoing 75% and 100% defoliation, and was unchanged for 25% and 50% defoliation. At both flowering and seed set, undefoliated SL had the greatest C:N ratio of all treatments, perhaps due to the increase in carbon-based PPP over time in undefoliated SL leaves.

3.4. PPP concentration

Concentrations of PPP in PTC were similar for all levels of defoliation at the vegetative stage and remained stable or increased in all levels of defoliation from the vegetative to flowering stage, with the exception of plants undergoing 100% defoliation in which PPP decreased 37% to the lowest of all defoliation levels at flowering (Table 4). At all defoliation levels, PTC PPP concentration decreased from flowering to seed set, with 50% defoliated plants undergoing the greatest decrease, 221%. Undefoliated control PTC plants had the greatest PPP concentration at seed set, up to 129%, the PPP concentration of all other treatments at seed set. A decrease in C concentration from vegetative to seed set stage suggests repeatedly defoliated PTC plants might have lacked the resources at seed set to produce well-defended leaves through PPP accumulation, due to resources required to replace defoliated leaves and set seed.

Table 4. Concentration of PPPs (g kg⁻¹) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	95.2^Footnotea,Footnotey	92.6^Footnotea,Footnotex	89.2^Footnotea,Footnotey	96.9^Footnotea,Footnotey	95.2^Footnotea,Footnotex	6.2
Flowering	119.7^Footnotea,Footnotex	92.4^Footnotec,Footnotex	110.2^Footnotea,Footnoteb,Footnotex	101.9^Footnoteb,Footnotec,Footnotex	69.3^Footnoted,Footnotey	6.2
Seed set	62.4^Footnotea,Footnotez	-	34.3^Footnoteb,Footnotez	27.3^Footnoteb,Footnotez	35.8^Footnoteb,Footnotez	6.2
SL^Footnote1
Vegetative	81.5^Footnotea,Footnotey	47.0^Footnoteb,Footnotey	58.3^Footnoteb,Footnotey	65.8^Footnotea,Footnoteb,Footnotex	81.5^Footnotea,Footnotex	6.8
Flowering	114.3^Footnotea,Footnotex	89.9^Footnoteb,Footnotex	91.9^Footnoteb,Footnotex	70.0^Footnotec,Footnotex	70.5^Footnotec,Footnotex	6.8
Seed set	130.3^Footnotea,Footnotex	60.3^Footnotec,Footnoted,Footnotey	82.1^Footnoteb,Footnotex	53.8^Footnoted,Footnotex	71.9^Footnoteb,Footnotec,Footnotex	6.8

¹ Ontogenetic stage × treatment interaction, P < 0.0001.

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y,z Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

Concentration of PPP in SL increased from vegetative to flowering stage in plants that were undefoliated and those defoliated at 50% and remained elevated through seed set (Table 4). SL plants subjected to 25% defoliation increased PPP concentration between vegetative and flowering stage but decreased 49% at seed set to a concentration similar to that at vegetative stage. Concentration of PPP in SL plants subjected to either 75% or 100% defoliation remained unchanged across defoliation events. At the vegetative stage, undefoliated plants had greater PPP concentrations than those subjected to 25% and 50% defoliation. Samples from undefoliated and 100% defoliation plants contained the oldest leaves, found near the plant base, while samples from plants subjected to 25% and 50% defoliation contained younger leaves of recent growth from the upper canopy. At vegetative stage, all leaves were naive, having not yet experienced defoliation. Results suggest SL leaves accumulate PPP as they age. At flowering, plants subjected to 25% and 50% defoliation had greater PPP concentrations than those subjected to 75% and 100% defoliation. The difference appears to be due to defoliation since all leaves sampled at this point were new regrowth tissue. At each sampling time, undefoliated plants consistently had among the greatest PPP concentrations of all treatments. Undefoliated SL plants had up to 73%, 63%, and 142% greater PPP concentration compared to defoliated plants at vegetative stage, flowering, and seed set, respectively. Results suggest that PPP concentration in SL increases with time in undefoliated plant leaves, and repeatedly defoliated SL may lack the resources to produce well-defended leaves through PPP accumulation.

3.5. TP concentration

As expected, TP concentrations followed the same pattern as PPP concentrations in PTC given that PPP composed a large proportion, as much as 89%, of TP in PTC (Table 5). TP concentrations in PTC were similar for all treatments at the vegetative stage. TP concentrations remained stable or increased in all treatments from the vegetative to flowering stage, with the exception of plants exposed to 100% defoliation in which TP concentration decreased 39% and was the lowest of all treatments at flowering. All PTC treatments experienced a decrease in TP concentrations at seed set, likely as a result of the concurrent decrease in PPP. Undefoliated plants had the greatest TP concentration of all PTC treatments at seed set.

Table 5. Concentration of TPs (g kg⁻¹) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	116.4^Footnotea,Footnotey	128.1^Footnotea,Footnotex	118.3^Footnotea,Footnotey	121.3^Footnotea,Footnotex	116.4^Footnotea,x	6.9
Flowering	149.6^Footnotea,Footnotex	131.6^Footnotea,Footnoteb,Footnotex	139.1^Footnotea,Footnotex	114.9^Footnoteb,Footnotex	83.7^Footnotec,Footnotey	6.9
Seed set	80.3^Footnotea,Footnotez	-	44.5^Footnoteb,Footnotez	45.7^Footnoteb,Footnotey	52.8^Footnoteb,Footnotez	6.9
SL^Footnote1
Vegetative	91.4^Footnotea,Footnotez	53.6^Footnotec,Footnotey	64.3^Footnoteb,Footnotec,Footnotey	75.5^Footnotea,Footnoteb,Footnotex	91.4^Footnotea,Footnotex	10.1
Flowering	143.5^Footnotea,Footnotey	117.2^Footnoteb,Footnotex	120.5^Footnotea,Footnoteb,Footnotex	87.2^Footnotec,Footnotex	78.7^Footnotec,Footnotex	10.1
Seed set	174.0^Footnotea, Footnotex	100.9^Footnoteb,Footnotex	119.5^Footnoteb,Footnotex	70.5^Footnotec,Footnotex	95.6^Footnoteb,Footnotec,Footnotex	10.1

¹ Ontogenetic stage × treatment interaction, P < 0.0001.

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y,z Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

TP concentration in SL followed a similar pattern as PPP with a few exceptions (Table 5). PPPs composed as much as 90% of TP in SL. The concentration of TP increased in undefoliated SL plants at each defoliation event. Concentration of TP in SL increased in plants exposed to 25% and 50% defoliation between vegetative and flowering stage and remained elevated through seed set. Concentration of TP in SL was unchanged across defoliations for 75% and 100% defoliated plants. At each sampling time, undefoliated plants consistently had among the greatest TP concentrations of all treatments. Undefoliated SL plants had up to 71%, 82%, and 147% greater TP concentrations than defoliated plants at vegetative stage, flowering, and seed set, respectively. Concentrations of TP in SL plants subjected to 25% and 50% defoliation were similar to undefoliated plants at flowering, likely due to the increase in PPP, a portion of TP, in new leaves following defoliation. Similar to PPP, results suggest that TP concentration in SL increased with time in undefoliated plants.

3.6. PPP to TP concentration ratio

In an effort to determine if plants undergoing simulated herbivory preferentially produced PPP as a mechanism of defense, the ratio of PPP to TP (PPP:TP) was determined. The PPP:TP ratio did not change in PTC from vegetative to flowering stage (Table 6). At seed set, PPP:TP ratio remained stable in the undefoliated PTC plants but decreased in plants subjected to 50%, 75%, and 100% defoliation, likely as a result of the considerable decrease in PPP concentration at seed set. Results indicate that although both PPP and TP decreased at seed set in undefoliated PTC plants, undefoliated plants experienced less of a decrease in PPP as a proportion of TP than defoliated plants at seed set.

Table 6. Ratio of the concentration of PPPs to TPs (g PPP:g TP) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	0.814^Footnotea,Footnotex	0.727^Footnoteb,Footnotex	0.753^Footnoteb,Footnotex	0.798^Footnotea,Footnotex	0.814^Footnotea,Footnotex	0.04
Flowering	0.801^Footnoteb,Footnotec,Footnotex	0.703^Footnoted,Footnotex	0.792^Footnotec,Footnotex	0.885^Footnotea,Footnotex	0.831^Footnoteb,Footnotex	0.04
Seed set	0.766^Footnotea,Footnotex	–	0.599^Footnotea,Footnoteb,Footnotey	0.556^Footnoteb,Footnotey	0.682^Footnotea,Footnoteb,Footnotey	0.04
SL^Footnote2
Vegetative	0.893^Footnotea,Footnotex	0.887^Footnotea,Footnotex	0.890^Footnotea,Footnotex	0.873^Footnotea,Footnotex	0.893^Footnotea,Footnotex	0.03
Flowering	0.803^Footnoteb,Footnotey	0.766^Footnoteb,Footnotey	0.769^Footnoteb,Footnotey	0.805^Footnoteb,Footnotex,Footnotey	0.898^Footnotea,Footnotex	0.03
Seed set	0.749^Footnotea,Footnotey	0.605^Footnotec,Footnotez	0.691^Footnoteb,Footnotez	0.769^Footnotea,Footnotey	0.753^Footnotea,Footnotey	0.03

¹ Ontogenetic stage × treatment interaction, P = 0.001.

² Ontogenetic stage × treatment interaction, P = 0.003.

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y,z Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

The PPP:TP ratio decreased across all defoliation events in SL plants exposed to 25% and 50% defoliation (Table 6). The PPP:TP ratio of undefoliated SL plants decreased between vegetative and flowering stage and was unchanged at seed set. The PPP:TP ratio of SL plants exposed to 100% defoliation was unchanged from vegetative to flowering stage but decreased from flowering to seed set. The PPP:TP ratio of 75% defoliated SL plants decreased from vegetative to seed set stage. The decrease in PPP:TP ratio of all treatments from vegetative to seed set stage indicates SL plants produced other phenolics preferentially with ontogenesis.

3.7. Protein binding ability of PPPs

PPPs may improve protein nutrition in ruminants. PPPs bind to plant proteins in the rumen and prevent microbial degradation thereby increasing amino acid flow to the small intestine (Min & Hart 2003). The amount of PB by PPP in PTC followed a similar pattern as TP and PPP concentrations (Table 7). The amount of PB increased in undefoliated and 50% defoliated PTC plants and was unchanged in 25%, 75%, and 100% defoliated PTC plants from vegetative to flowering stage. Amount of PB was lowest in PTC plants exposed to 25% defoliation treatment at flowering. All PTC treatments experienced a decrease in PB between flowering and seed set, likely a result of the decrease in concentration of protein binding agents during that time. Undefoliated PTC plants bound at least 94% more protein than defoliated plants after two defoliations, suggesting repeated defoliation may reduce PTC ability to contribute to ruminant nutrition by binding protein in the rumen with resulting rumen-escape protein benefits (Min & Hart 2003).

Table 7. Amount of PB (g kg⁻¹) by PPPs in PTC at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	72.6^Footnotea,Footnoteb,Footnotey	69.3^Footnoteb,Footnotex	72.0^Footnotea,Footnoteb,Footnotex	83.8^Footnotea,Footnotex	72.6^Footnotea,Footnoteb,Footnotex	4.3
Flowering	91.8^Footnotea,Footnotex	63.6^Footnotec,Footnotex	87.5^Footnotea,Footnoteb,Footnotex	93.1^Footnotea,Footnotex	76.7^Footnoteb,Footnotex	4.3
Seed set	51.5^Footnotea,Footnotez	–	22.0^Footnoteb Footnotey	26.5^Footnoteb,Footnotey	25.1^Footnoteb,Footnotey	4.3

¹ Ontogenetic stage × treatment interaction, P < 0.0001.

^a,b,c Means within row followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y,z Means within column followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

The amount of PB was greatest in SL plants exposed to 25% defoliation. The amount of PB for SL plants subjected to 100% defoliation was similar to that of 75% defoliated plants but less than PB for undefoliated, 25%, and 50% defoliated plants. Flowering and seed set stages had greater PB than the vegetative stage, suggesting that PB in SL increases as plants age.

3.8. PB by PPPs to PPP concentration ratio

It is not known if plants have the ability to change the properties of PPP in regrowth in response to stress. In an attempt to measure this response, the protein binding ability of the PPP produced by plants in each defoliation treatment of the study was estimated. The ratio between PB and PPP indicates the protein binding efficiency of the CTs, with a large ratio implying an efficient PPP. All PTC treatments had similar PB:PPP ratio at the vegetative stage (Table 8). The PB:PPP ratio was unchanged for all PTC treatments from vegetative to flowering stage, with the exception of plants subjected to 100% defoliation in which PB:PPP ratio increased by 44%. From flowering to seed set in PTC, PB:PPP ratio was unchanged for undefoliated plants, decreased in plants exposed to 50% and 100% defoliation, and increased in 75% defoliated plants. Although 100% defoliated PTC plants had the least PPP and a moderate PB at flowering, PB:PPP ratio was equal to or greater than all other treatments at flowering, suggesting that at flowering, PPP in regrowth of PTC plants exposed to 100% defoliation changed their protein binding qualities following defoliation stress, becoming more efficient at binding protein. In a similar manner, 75% defoliated PTC plants had among the lowest PPP and PB at seed set, but had the greatest PB:PPP ratio, suggesting that plants exposed to 75% defoliation had very efficient PPP at seed set and PPP attributes were altered from flowering to seed set. This may have been a coincidence, however, because PB:PPP ratio decreased in PTC plants subjected to 50% and 100% defoliation at seed set. Both 50% and 100% defoliated plants decreased in PB and PPP from flowering to seed set, so the decrease in PB:PPP ratio for PTC plants exposed to 50% and 100% defoliation at seed set was due to a greater decrease in PB than PPP.

In SL at vegetative stage, PB:PPP ratio for 25% defoliated plants was similar to 50% defoliated plants but greater than all other treatments (Table 8). PB to PPP ratio decreased in SL plants exposed to 25% and 50% defoliation and was unchanged in undefoliated, 75%, and 100% defoliated SL plants from vegetative to flowering stage. From flowering to seed set, PB:PPP ratio increased in 25% defoliated plants and was unchanged in all other SL treatments. At each sampling time, undefoliated SL plants consistently had among the lowest PB:PPP ratios of all treatments. Results appear to be attributed to concentrations of PPP in SL and their respective changes in quantity rather than protein binding efficiency of PPP. Following stress by defoliation, PTC appears to change the protein binding properties of its PPP while SL does not. CT structural analysis may help confirm changes in PPP protein binding characteristics.

Table 8. Ratio of the concentration of PB by PPPs to PPPs (g PB:g PPP) in PTC and SL at vegetative, flowering, and seed set stages undergoing 0%, 25%, 50%, 75%, or 100% defoliation.

	Level of defoliation, %
	0	25	50	75	100	SEM
PTC^Footnote1
Vegetative	0.783^Footnotea,Footnotex	0.758^Footnotea,Footnotex	0.815^Footnotea,Footnotex	0.914^Footnotea,Footnotey	0.783^Footnotea,Footnotey	0.08
Flowering	0.769^Footnoteb,Footnotec,Footnotex	0.693^Footnotec,Footnotex	0.805^Footnoteb,Footnotec,Footnotex	0.934^Footnoteb,Footnotey	1.128^Footnotea,Footnotex	0.08
Seed set	0.841^Footnoteb,Footnotex	–	0.442^Footnotec,Footnotey	1.196^Footnotea,Footnotex	0.701^Footnoteb,Footnotec,Footnotey	0.08
SL^Footnote2
Vegetative	0.713^Footnotec,Footnotex	1.528^Footnotea,Footnotex	1.262^Footnotea,Footnoteb,Footnotex	1.058^Footnoteb,Footnotec,Footnotex	0.713^Footnotec,Footnotex	0.1
Flowering	0.607^Footnotec,Footnotex	0.892^Footnotea,Footnoteb,Footnotey	0.763^Footnoteb,Footnotec,Footnotey	1.001^Footnotea,Footnotex	0.935^Footnotea,Footnoteb,Footnotex	0.1
Seed set	0.614^Footnotec,Footnotex	1.340^Footnotea,Footnotex	0.794^Footnoteb,Footnotey	1.257^Footnotea,Footnotex	0.960^Footnoteb,Footnotex	0.1

¹ Ontogenetic stage × treatment interaction, P = 0.0002.

² Ontogenetic stage × treatment interaction, P = 0.0011.

^a,b,c Means within row within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

^x,y Means within column within species followed by a common superscript letter do not differ (P ≤ 0.05) according to a least significant difference multiple mean separation.

3.9. General discussion

Concentrations of PPP, TP, and PB decreased in all PTC treatments from flowering to seed set. In contrast, concentrations of PPP and TP were unchanged or increased for all SL treatments, except plants exposed to 25% defoliation, in which PPP decreased, from flowering to seed set. Concentration of PB was similar between flowering and seed set stages in SL as well. Results might correspond with divergent seed dispersal strategies of the two species. Seed dispersal strategy in PTC depends primarily on epizoochory by adhering to herbivore hair, whereas SL depends on a combination of barochory and anemochory through simple seed drop. Decreasing chemical defenses in leaves, and hence attracting herbivory, at seed maturity would be beneficial for PTC but not for SL.

Similar evidence for seed dispersal strategy affecting CT concentration was presented by Wrangham and Waterman (1981). Vervet monkeys preferred the seed of umbrella thorn acacia (Acacia tortilis (Forsk.) Hayne) over fever tree (Acacia xanthophloea Benth.) seed, which had more than 400% greater CT concentration than the former. Umbrella thorn acacia fruit is indehiscent, depending on endozoochory, while fever tree fruit dehisces, scattering seed by other means (Lamprey 1967). Umbrella thorn acacia appears to depend on animals, like the vervets, to disperse seeds. By contrast, fever tree seeds do not appear to be adapted for animal dispersal, relying instead on wind or water dispersal (Wrangham & Waterman 1981).

Differences in SL PPP concentrations between defoliation treatments at the vegetative stage reflect differences in SL leaf age. At the vegetative stage when all leaves were naive, SL plants exposed to 100% defoliation had greater PPP concentrations than plants experiencing 25% and 50% defoliation. Samples from plants subjected to 25% and 50% defoliation contained the most recent plant growth, the uppermost leaves of the canopy, while samples from 100% defoliated plants contained the young leaves of the upper canopy as well as the oldest leaves, found near the plant base. Results suggest that in the absence of defoliation stress, SL leaves accumulate PPP. These results were reinforced at each sampling time as undefoliated SL plants consistently had among the greatest PPP concentrations of all treatments. All PTC treatments experienced stable or increased N concentrations from vegetative to flowering stage, and all treatments underwent a decrease in N concentration between flowering and seed set. In contrast, ontogenesis and defoliation did not affect SL N concentration. Results suggest PTC might have experienced nutrient stress after the second defoliation and was unable to compensate in subsequent leaf development for N lost to defoliation. Seeds usually contain a large proportion of legume N (Sinclair & de Wit 1975), and in a N-stressed situation, PTC might have mobilized N from vegetative tissue to reproductive organs to meet N demands for seed fill. In previous studies, N concentration has been shown to decrease in vegetative tissue as N-stressed plants invest in reproductive tissue. Selamat and Gardner (1985) reported a decrease in vegetative N as N was redistributed to seeds in N-stressed non-nodulating peanut (Arachis hypogaea L.) during fruiting, while Saulnier and Reekie (1995) reported a decrease in N of common evening primrose (Oenothera biennis L.) leaves during reproduction when N availability was limiting. By maintaining stable leaf N concentrations across multiple defoliation events, SL appears to be able to utilize N resources more efficiently than PTC at seed set when demands for N are high.

Differences in CT structure between the two species may help explain differences in CT concentrations, including PPP. Heavier molecular weight has been proposed as an indicator of greater CT biological activity (Naumann et al. 2013c). However, Huang et al. (2010) and Naumann et al. (2013b) found CTs with lower molecular weights exhibited protein binding abilities as strong or stronger than CTs with greater MW. Additional structural characteristics, including stereochemistry, hydroxylation pattern, interflavan linkages, and functional groups may account for different biological activities among different species (Naumann et al. 2013b). Anthocyanidin monomer analysis revealed PTC CTs are 41.42% delphinidin and 58.58% cyanidin. SL CTs are 80.36% delphinidin, 19.49% cyanidin, and 0.15% pelargonidin (2013 personal observation by C.E. Cooper; unreferenced).

4. Conclusions

Results from this experiment indicate that seed dispersal strategy and stress from herbivory may play distinct roles in determining leaf PPP and PB concentrations. For both PTC and SL, undefoliated plants contained leaf PPP concentrations up to 129% and 142% greater than defoliated treatments after two defoliations, suggesting repeatedly defoliated PTC and SL plants may be less well equipped to defend against further herbivory compared to undefoliated plants.

Repeated defoliation reduced PB in PTC and thus PTC's ability to increase amino acid flow to the small intestine and improve ruminant nutrition (Min & Hart 2003). Concentration of PB increased in SL as plants aged, but repeated defoliation decreased SL's concentration of PPP and the concentration of PB by SL as a result, which in turn may have implications for ruminant nutrition. Further research is needed to determine if PPP concentration reduction at seed set would affect other beneficial biological activities of PTC and SL PPP, including gastrointestinal nematode control and rumen methane production.

Ontogeny and stress from simulated herbivory may exert dissimilar effects on the nutritive value of these species. Regenerated PTC leaves may be less attractive to herbivores due to increased CT and lower N concentrations. However, the decrease in leaf chemical defenses in PTC at seed set may actually increase palatability, negating the effects of decreased N concentrations on herbivore preference during autumn when most other associated plants around PTC are senescing and losing nutritive value. By maintaining stable leaf nutrient concentrations across multiple defoliation events, SL is able to utilize nutrient resources more efficiently than PTC at seed set when nutrients are prioritized for reproduction; decreased palatability at a time when herbivory is not desirable is an added benefit. Different nutrient allocation and seed set strategies provide managers with varying challenges and opportunities that are legume species specific.