The nitrogen-fixing Frankia significantly increases growth , uprooting resistance and root tensile strength of Alnus formosana

Restoration of Alnus formosana (Burk.) Makino on landslide areas is important for agroforestry, forestry and soil erosion control in Taiwan. To ensure successful reforestation, A. formosana seedlings have to develop strong root system for nutrient and water acquisition as well as anchorage. Inoculating of A. formosana with symbiotic nitrogen-fixing actinobacteria Frankia may help mitigate drought and nutrient deficiencies on landslide sites. However, the effects of Frankia inoculation on growth, root architecture and mechanical properties of A. formosana seedlings are not well understood. In this research, a Frankia strain AF1 was isolated from actinorhizal nodules of local A. formosana and recognized as Frankia species, and its influences on growth performance and root mechanical properties of A. formosana seedlings were examined and analyzed. The results showed that the inoculated seedlings had significantly larger height and root biomass, longer root length, and more root tip number than that of the non-inoculated controls. Consistently, the inoculated seedlings had statistically significant higher uprooting resistance, root tensile resistance force and tensile strength than the controls. The results reveal that this native Frankia strain can promote growth performance, root system architecture, anchorage ability and root tensile strength of A. formosana.


INTRODUCTION
Landslide hazards and their large negative impacts on human lives, economies and infrastructure have become a growing challenge globally (Dai et al., 2002).Due to the fragile geology, steep terrain and torrential rains brought by typhoons, landslides have become one of the most severe disasters in Taiwan.Vegetation landslide engineering has become increasingly important for landslide prevention and rehabilitation in recent years (Chen et al., 2014).In general, trees and forests can play an important role in preventing and rehabilitating landslides.Actinorhizal trees are pioneer species and can improve tree growth and survival on degraded landslide soils (Diagne et al., 2013).Alnus formosana (Alnus formosana (Burk.)Makino), belonging to the family Betulaceae, is a native nitrogen-fixing actinorhizal woody species, widely distributed throughout the island of *Corresponding author.E-mail: jtlee@mail.ncyu.edu.tw.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License Taiwan in landslide sites (Liao, 1996).It has high potential for agroforestry, forestry, lumber production and landslide restoration.It can establish symbiosis with nitrogen-fixing Frankia and form actinorhizal root nodules in which Frankia provides fixed nitrogen to the host plant for growth and development (Lee, 1986;Lin, 1992).Several previous studies have demonstrated that Frankia can improve the establishment and growth of Alnus in degraded lands (Lefrançois et al., 2010;Santi et al., 2013;Bissonnette et al., 2014;Põlme et al., 2014).Inoculation with Frankia significantly increases seedling growth, biomass and root nodules of Alnus crispa and Alnus sieboldiana (Yamanaka et al., 2005;Quoreshi et al., 2007).
Morphological types of tree root system architecture have been classified into heart system, plate system and taproot system (Stokes and Mattheck, 1996).Stokes et al. (2009) indicated that taproot length, amount of lateral roots and root architecture affects uprooting resistance of plants.Orfanoudakis et al. (2010) showed that inoculation of Alnus glutinosa with Gigaspora rosea and Frankia improves ramous root.Past studies on root morphological characteristics and biomechanical properties of Alnus species were focused on Alnus subcordata and Alnus viridis.In A. viridis, the maximum root area ratio (RAR) values were located in the upper 30 cm soil layer and the maximum rooting depth was about 1 m, whereas the root tensile strength decreased with diameter (Bischetti et al., 2005).In A. subcordata, the root density, root number and RAR decreased with increasing depth and the maximum rooting depth was 1 m, while the root tensile strength decreased with increasing root diameter (Naghdi et al., 2013).However, there were few studies that tried to investigate the effects of inoculation of Frankia on growth, root morphological characteristics and mechanical properties of A. formosana.Therefore, the purposes of this research were: (1) to isolate the Frankia strain from A. formosana, and (2) to examine the influences of Frankia inoculation on growth performance, uprooting resistance and tensile strength of A. formosana seedlings in order to provide strategy for landslide prevention and erosion control practices.This study focused on the application of nitrogen-fixing Frankia to alder seedling production in order to improve seedling growth performance and enhance root mechanical properties, which is important in prevention of landslide hazards.

Sample collection
An elite tree of A. formosana was selected from the natural forest stand located at Fenqihu Township, Chiayi County, Taiwan (219417.59E,2599775.50N,TWD 97) in October 2014.Actinorhizal nodules were gathered from roots at 6 to 30 cm deep in soil, kept in sealed plastic bags and transported to laboratory in a cold box for Frankia strain isolation (Lin, 1992).Cones were also collected from upper crown of the same tree.Cones were sun-dried in trays for seed release.Seeds were extracted, cleaned and freeze-stored in polyethylene bags at -20°C.

Frankia strain isolation and purification
Actinorhizal root nodules were washed in distilled water to remove soil particles.Single nodule lobes were cut 2 mm from the tip with scalpel, ultrasonically cleaned, surface-sterilized with 75% ethanol for 10 min, 15% NaOCl for 10 min, and 30% hydrogen peroxide for 10 min.The endosymbiont Frankia in nodule lobes were isolated aseptically on QMOD medium (Lalonde and Calvert, 1979).

DNA extraction, sequencing, and gene sequence similarity analysis
Frankia genomic DNA was extracted with Puregene DNA Purification Kit (QIAGEN, Pleasanton, CA, USA), and subsequently subjected to 1.2% agarose gel electrophoresis.The primers used for PCR of rDNA were primer FGPL2054 (5'-CCGGGTTTCCCCATTCGG-3') and primer FGPS989e (5'-GGG GTC CTT AGG GGC T-3') (Daniel et al., 1999).Then, the amplified samples were analyzed by gel electrophoresis and the particular PCR products were sequenced.The DNA sequences were submitted to NCBI to access GenBank for sequence similarity analysis of the Frankia rDNA gene sequences.

Seedling preparation
Seeds of A. formosana were surface cleaned with tap water, sterilized 2 times with 10% NaOCl solution for 5 min and washed with sterile water, and then germinated in autoclaved peat moss and vermiculite mixtures (1:1, v/v) in October 2015.The wooden boxes (l×w×h, 30 cm × 30 cm × 60 cm) were used for transplanting.The boxes were sterilized with 10% NaOCl solution, and the sandy loam soils collected from the same natural forest stand were autoclaved and then fumigated with 200 g Basamid fumigant per cubic meter of soil, and the soil surface was sealed with polyethylene sheets for 14 days to prevent the toxic gas from escaping.Then, the boxes were filled with the sterilized soils.When seedlings attained a height of 5 cm, they were individually transplanted to the boxes, and watered regularly.Twenty-eight seedlings in boxes were arranged randomly into two individual plastic houses.

Inoculum preparation
The isolated Frankia strain was successively cultured twice for 30 days.Inocula were prepared by concentrating Frankia cultures in sterile 1 ml tubes (15,000 rpm for 10 min at 4°C) using a Hitachi centrifuge (HIMAG Centrifuge CR 15T, Rotor RT15A, Tokyo, Japan).The pellets were then homogenized with glass tissue homogenizer (Wheaton 25802a, Millville, NJ, USA) in sterile BAP medium and sonicated for 1 min using an ultrasonicator (Biologics 150VT, Manassas, VA, USA) on ice, and used as inoculum.The protein concentration of the homogenized inoculum was assayed with the Bradford method (Bradford, 1976).The protein concentration of the Frankia inoculum was 2.8±0.02µg ml -1 .

Inoculation test
Four weeks after transplanting, 14 plants in one plastic house were inoculated with the isolated Frankia strain.A 5 ml suspension of the strain (protein concentration 2.8±0.02µg ml -1 ) was dripped into five small holes near the seedling.The process was duplicated after fourteen days for ensuring high rate of root colonization.Another 14 non-inoculated control seedlings were treated with sterilized water.The boxes of inoculated and control seedlings were placed individually in two separate plastic houses.The seedlings were grown at 26±4°C during daytime and 18±5°C at night time, with 60 to 80% relative humidity, and 1000±200 μmoles photon m -2 s -1 photosynthetic photon flux density during daytime.Eight months later, the seedlings were sampled for measurements of growth performance, root morphological characteristics, uprooting resistance and tensile strength.

Plant growth performance and root morphological characteristics
After 8 months of cultivation, seven inoculated and seven noninoculated control plants were randomly chosen, separately.The height and stem-base diameter of seedlings were measured with ruler and caliper.The root systems were carefully excavated by hand with trowel (Böhm, 1979).The root length and root numbers were recorded.Images of seedling roots were captured for analysis of root architecture and morphological characteristics.The root morphological characteristics analysis was conducted using a WinRHIZOPro analysis system (Regent Instruments, Quebec, QC, Canada) (Bouma et al., 2000).However, root volume was evaluated with water displacement technique due to large quantity of roots (Pang et al., 2011).Biomass of leaf, root, stem and root nodule was estimated by drying in a hot air oven at 75°C until a constant weight was obtained.Root functional characteristics, that is, root mass density (g dm -3 ), root length density (m dm -3 ), tissue mass density (g cm -3 ), specific root length (m g -1 ), and root to shoot ratio were computed (Stokes et al., 2009;Burylo et al., 2012;Gould et al., 2016).Live roots were also collected from sampled seedlings and prepared for subsequent tensile testing.

Vertical uprooting test
After 8 months of cultivation, 7 inoculated and 7 non-inoculated control seedlings were randomly sampled for vertical uprooting test, respectively.The soil material was categorized as sandy loam soil (containing 65.2% sand, 27.4% silt and 7.4% clay).At first, seedling height and stem-base diameter were recorded.The seedling stem was removed from 15 cm above the stem base and the bark was peeled away to prevent slide of the pulling fixture.The uprooting test was conducted using an in situ pullout instrument (U-Soft USPA-003, U-Soft Technology Co., Taiwan) fitted with a 5T load cell (Kyowa, Tokyo, Japan) attached to loading recorder and constant control unit, and erected on a triangular steel frame.Then, the instrument was attached to the pulling fixture.The uprooting force was applied perpendicular to the soil surface at a constant rate of 2 mm min -1 .The data of uprooting resistance force and displacement were recorded with load cell and displacement transducer, and stored on a laptop computer.The uprooting test was terminated once the resisting force dropped sharply.

Root tensile test
After root excavation, live roots of different diameter classes (0 to 1, 1 to 2, 2 to 5, and 5 to 10 mm) at a depth of 30 cm below the soil surface were collected randomly from the sampled seedlings, respectively.The roots were kept in separate sealed plastic bags to prevent drying of root tissues (De Bates et al., 2008) and transported with a cold box (Bischetti et al., 2005).Root samples were then immersed in a 15% ethanol solution at 4°C in order to Lee and Tsai 215 conserve root tensile strength (Bischetti et al., 2009).Tensile tests were performed in the laboratory using a tensile-testing machine (U-Soft USPT-003, U-Soft Technology Co., Taipei, Taiwan).The load cell (Kyowa LCN-A, Tokyo, Japan; sensor resolution 0.1 N, maximum force 500 N) was connected to a loading recorder and control unit.The data of tensile force and displacement were compiled on a portable computer.A total of 110 root segments were randomly sampled from the inoculated seedlings.Another 110 root segments were also randomly collected from the non-inoculated seedlings.All root segments were washed and cut to 60 mm in length, and the root segments were clamped with sand paper during testing to prevent slippage.Then, the root segments were pulled vertically at a constant speed of 4.7 mm min -1 until the resisting force dropped sharply.The root tensile strength (Ts, MPa) was calculated using the following formula (De Baets et al., 2008;Osman et al., 2011;Zhang et al., 2012): where Fmax is the maximum force (N) at rupture and di is the mean root segment diameter (mm) measured at three points, that is, near the upper clamp, halfway and near the bottom clamp, using a digital caliper (accuracy of 0.01 mm).

Data analysis
T-test in SPSS 22.0 software (Chicago, IL., USA) was used for analyzing variations in growth performance and morphological characteristics data between inoculated and non-inoculated control seedlings.The relationships between uprooting resistance, root tensile resistance, tensile strength and morphological characteristics were evaluated using Microsoft Excel regression analysis.

Actinobacterial strain isolation and sequencing
The actinobacterial strain was isolated and purified on QMOD medium and classified as AF1.Molecular analysis showed that the rDNA gene sequence of the isolated AF1 has 100% similarity to that of Frankia species genus (Figure 1).The local strain was recognized as Frankia spp.AF1.Inoculation test showed that the isolated Frankia strain can induce nodule development in the roots of A. formosana (Figure 2).

Root system architecture
Results of the investigation revealed that inoculated A.     formosana plants developed larger root systems than the non-inoculated ones (Figure 3).The taproots of inoculated plants grew to 43 cm deep.In addition, the inoculated plants developed about 85% of the root matrix in the top 40 cm soil, and its lateral roots extended profusely, with many nodules (Figure 3a).Conversely, the taproots of non-inoculated ones grew only to 20 cm deep.Also, the non-inoculated plants developed about 90% of the root matrix in the top 30 cm soil, and its lateral roots grew sparsely, without any nodules (Figure 3b).The types of root architecture of inoculated and noninoculated A. formosana seedlings were categorized to the heart root system according to Stokes and Mattheck (1996).WinRHIZO analysis of root morphological characteristics revealed that Frankia inoculation significantly influenced all morphological characteristics.Generally, the inoculated plants developed larger total root length (60%), root surface area (61%), root volume (55%), and root tip number (72%) than the noninoculated ones (Table 2).Statistical analysis of root functional characteristics exhibited that the root density and root length density of inoculated plants were higher than the controls.Generally, the inoculated A. formosana seedlings had higher root density (60%) and root length density (67%) than the non-inoculated ones (Table 3).

Uprooting resistance
In this study, seven uprooting tests were conducted to examine the uprooting resistance of the inoculated and non-inoculated seedlings.The results revealed that the average maximum uprooting resistance of the inoculated seedlings (1.09±0.40kN) was significantly higher than the non-inoculated ones (0.60±0.17 kN) (Table 4).The uprooting resisting force increased with displacement up to the peak and then decreased sharply as the roots broke (Figure 4).Regression analysis exhibited the significant linear positive correlations between the maximum uprooting resistance force and morphological characteristics, that is, tree height, stem-base diameter, taproot length, root biomass, and shoot biomass (Figures 5,6,7,8,and 9).

Root tensile strength
In total, 220 tests were performed to investigate the root tensile strength of the inoculated and non-inoculated A. formosana seedlings.Among them, 86 root tensile tests of the inoculated seedlings and 102 root tensile tests of the non-inoculated controls were successful.The average root tensile resistance force of inoculated seedlings (81.1±19.0N) was significantly higher than that of the controls (68.0±23.5 N).The mean value of root tensile strength of inoculated seedlings (17.45±3.36MPa) was significantly higher than that of the non-inoculated controls (11.42±1.83MPa).On the other hand, the average root diameter of non-inoculated seedlings (2.88±1.28mm) was significantly higher that the inoculated ones (2.42±1.20 mm) (Table 5).Regression analysis revealed that root tensile resistance force increased with increasing root diameter in accordance with a positive power function correlation (Figure 10), whereas the root tensile strength decreased with increasing root diameter in accordance with negative logarithmic function correlation (Figure 11).Furthermore, the maximum root tensile resistance and tensile strength of inoculated seedlings were significantly higher than that of the non-inoculated controls.

DISCUSSION
The results showed that the native symbiotic         actinobacterium strain in nodules of A. formosana was isolated and recognized as Frankia by 16S rDNA gene similarity analysis.Inoculation test showed that this Frankia strain can induce actinorhizal nodule development in the roots of A. formosana seedlings.A number of studies indicated that Frankia strains associate with Alnus spp.(that is, Alnus acuminata, A. crispa, A. glutinosa, Alnus nepalensis, Alnus rubra and Alnus sieboldiana, respectively) (Benson, 1982;Carlson and Dawson, 1985;Hooker and Wheeler, 1987;Vendan et al., 1999;Carú et al., 2000;Oliveira et al., 2005;Yamanaka et al., 2005).Faure-Raynaud et al. (1991) analyzed the diversity of Frankia strains isolated from single nodules of A. glutinosa and showed no divergence among strains isolated from the same nodule.McEwan et al. (2015) also demonstrated that a single ribotype of Frankia is the major bacterium in single lobe from a nodule of A. glutinosa.
The research revealed that A. formosana seedlings inoculated with the native Frankia had significantly higher growth performance than the non-inoculated ones.Many previous studies have shown that inoculation with Frankia strains could significantly increase growth and development of alder seedlings (Prat, 1989;Wheeler et al., 1991;Kendall et al., 2003;Martin et al., 2003;Schrader and Graves, 2008;Bissonnette et al., 2014;Yamanaka et al., 2009).Lumini et al. (1994) showed that inoculation of selected Frankia strains and arbuscular mycorrhizal fungi along with sterilized media developed significantly higher shoot biomass than the noninoculated controls.Quoreshi et al. (2007) demonstrated that A. crispa inoculated with Frankia had significantly higher biomass, nodule lobes, and nodule weight than the controls.Moreover, Vendan et al. (1999) showed that Nepalese alder (A.nepalensis) plants inoculated with Frankia have a higher shoot length, root length, and biomass, while the Frankia strain AVC-II exhibited better infectivity and productivity of Nepalese alder than other strains tested.They also clearly indicated the potentiality of utilizing the Alnus-Frankia specificity for higher productivity through effective symbiosis.Schrader and Graves (2008) demonstrated that alder seedlings inoculated with species-specific Frankia strain grew larger, and gained more biomass than the crossinoculated ones.Consistently, the study also demonstrated the positive effect of local Frankia strain on growth performance of A. formosana seedlings.
The root system architecture of A. formosana seedlings inoculated and non-inoculated with Frankia were similar to the heart root system (Stokes and Mattheck, 1996).The seedlings inoculated with Frankia had deeper taproot and more profuse roots than the non-inoculated ones.The inoculated seedlings had significantly higher total root length, root surface area, root volume, and root tip number than the controls.Moreover, seedlings inoculated with Frankia also had significantly higher root density and root length density than the non-inoculated ones.Wheeler et al. (1979) demonstrated that alder inoculated with Frankia developed more lateral root primordia than the non-inoculated ones, suggesting that Frankia can induce lateral root formation.De Bates et al. (2006) indicated that root density is a pertinent parameter to estimate the erosion-reducing efficacy.Stokes et al. (2009) also showed that higher root length density increases the uprooting resistance of plants.
This study demonstrates that A. formosana seedlings inoculated with Frankia have significantly higher uprooting resistance than the controls, indicating a higher anchorage capability in roots of the inoculated seedlings.Also, there were strong linear positive correlations between uprooting resistance and tree height, stem-base diameter, taproot length, root biomass and shoot biomass.Additionally, the inoculated A. formosana seedlings have longer taproot and more profuse lateral roots than the non-inoculated ones.Hence, inoculation with Frankia significantly augmented the numbers of lateral roots, which consequently stimulate seedling anchorage capability and uprooting resistance.Stokes et al. (2005) also indicated that heart root system is more resistant to uprooting than taproot system and plate root system.
The results of root tensile tests showed that the root tensile resistance force increased with increasing root diameter, whereas the root tensile strength decreased with increasing diameter.The findings are congruent with many other studies (Bischetti et al., 2005;Genet et al., 2005;De Baets et al., 2007;Normaniza et al., 2011;Nyambane et al., 2011;Abdi et al., 2014;Mohammed and Normaniza, 2014;Capilleri et al., 2016).Root chemical components, such as cellulose, lignin, hemicellulose and holocellulose, are closely related to root mechanical properties.Genet et al. (2005) found that root cellulose content increased with increasing root diameter and increasing tensile strength in both Pinus pinaster and Castanea sativa.Several studies have shown that the root cellulose content increases and lignin content decreases with an increase in root diameter, and decrease in tensile strength (Lv et al., 2013;Zhang et al., 2014;Yang et al., 2016).In addition, the results from the present study also demonstrated that the average root tensile resistance force and tensile strength of inoculated seedlings were significantly higher than that of the noninoculated controls.This suggests that inoculation with Frankia promotes root tensile resistance force and tensile strength of A. formosana seedlings.

Conclusions
Collectively, results of the present study clearly show that the native Frankia strain significantly enhances growth performance, root system architecture, uprooting resistance and root tensile strength of A. formosana seedlings.The findings of this study are of great importance in the application of Frankia in alder seedling production for landslide reforestation and soil conservation practices.This is the first report to demonstrate that inoculation with Frankia can significantly enhance growth, root system architecture, anchorage capability and root tensile strength of A. formosana seedlings.Additionally, further studies on the diversity of native Frankia strains and its symbiotic compatibility with alder species are needed for seedling production.Also, researches on the effects of alder roots on hillslope stability and erosion control are useful in ecological restoration and soil conservation in landslide areas.
the mean ± standard error of 7 replicates.Values in the same column followed by different superscript letters significantly differ at 5% significant level.
the mean ± standard error of 86 root segments and 102 root segments of inoculated and non-inoculated A. formosana, respectively.Values in the same column followed by different superscript letters significantly differ at 5% significant level.

Table 1 .
Growth performance of A. formosana seedlings inoculated and non-inoculated with Frankia after 8 months of cultivation.

Table 2 .
Root morphological characteristics of A. formosana seedlings inoculated and non-inoculated with Frankia after 8 months of cultivation.

Table 3 .
Root functional characteristics of A. formosana seedlings inoculated and non-inoculated with Frankia after 8 months of cultivation.
a All values are the mean ± standard error of 7 replicates.Values in the same column followed by different superscript letters significantly differ at 5% significant level.

Table 4 .
Uprooting resistances of inoculated and non-inoculatedA.formosana seedlings after 8 months of cultivation.
bAll values are the mean ± standard error of 7 replicates.Values in the same column followed by different superscript letters significantly differ at 5% significant level.

Table 5 .
Average root diameter, tensile resistance force and tensile strength of inoculated and non-inoculated A. formosana seedlings after 8 months of cultivation.