Page  97 ï~~2007 THE MICHIGAN BOTANIST 97 ARBUSCULAR MYCORRHIZAL FUNGAL ROOT COLONIZATION AND INOCULUM PROPAGULES IN DECIDUOUS FLOODPLAIN FORESTS OF SOUTHWESTERN MICHIGAN, USA Peter G. Avis Sara D. Foster Paul D. Olexia Department of Biology 3950 Lakeshore Drive Department of Biology Indiana University Northwest Chicago, IL, USA Kalamazoo College 3400 Broadway 1400 Academy Street Gary, IN, 46408, USA Kalamazoo, MI, 49007, USA Corresponding author: Peter G. Avis,, phone: 219-980-6717 ABSTRACT We examined the arbuscular mycorrhizal (AM) fungi of temperate deciduous floodplain forests in southwest Michigan, USA, and the relationship between soil moisture and AM fungal root colonization and inoculum. In four floodplain forests, AM root colonization was measured for roots collected from soil samples taken along soil moisture, micro-topography and vegetation gradients. An AM fungal inoculum bioassay was also conducted using these samples to measure the amount of infective propagules found across these gradients. Although soil moisture had no significant effect on either AM parameter measured, the effect of forest site was significant in most instances. We also found a significant positive relationship between AM infective propagules and root density. Overall, this study illustrates that AM fungi are distributed throughout these temperate floodplain forests even in the potentially most anoxic, highly saturated low lying portions of the floodplain. KEY WORDS: Arbuscular mycorrhizal fungi, floodplain forests, riparian, soil moisture INTRODUCTION Floodplain forests in much of the world (Terborgh and Andresen 1998) consist of trees that associate with arbuscular mycorrhizal (AM) fungi. As in most forests, AM fungi of floodplain forests likely serve roles in the mineral nutrition of host plants and belowground carbon allocation. In addition, floodplain AM fungi appear to be especially important in riparian habitats as they can link terrestrial and aquatic ecosystems by contributing significant amounts of protein to the nutrient budgets of rivers (Harner et al. 2004). However, how AM fungi are distributed across the complex soil hydrological, geo-morphological, and vegetation gradients that characterize floodplain forests is not clear. Floodplain forests exhibit a dynamic soil hydrology and flood pulses can differentially impact soil moisture levels (Mitsch and Gosselink 1993) resulting in soils that are mosaics of saturation in space and time. Soil saturation, and conversely soil aeration, may be a key factor affecting AM fungi as anoxic conditions are considered detrimental to AM fungal growth (Mosse et al. 1981; Pawlowska and Charvat 2002; Entry et al. 2002) and soil flooding can limit the colonization of AM fungi in grasses of oligotrophic shrub wetlands (Miller 2000; Miller and Sharitz 2000). However, other studies have shown higher AM

Page  98 ï~~98 THE MICHIGAN BOTANIST Vol. 46 fungal spore counts in poorly drained soils (Khalil and Loynachan 1994; Troeh and Loynachan 2003) and no relationship between soil moisture and AM fungal colonization in prairie pothole wetlands (Wetzel and van der Valk 1996) or native plants of calcareous wetlands in New York (Van Hoewyk et al 2001). In flooding pampas grasslands of Argentina, duration of flooding impacted root colonization but not spore density (Escudero and Mendoza 2005). These latter studies suggest that factors such as vegetation distribution or soil nutrient levels may determine the distribution of AM fungi in floodplains. Since the AM fungi of floodplain forests have been little studied and these forests differ from other wetlands such as by having a preponderance of woody plant root systems, the factors that influence AM fungi in these important ecosystems remain unclear. Therefore, the objective of this study was to examine how AM fungi are distributed in temperate deciduous floodplain forests of the Kalamazoo River in southwest Michigan, USA, and to test the prediction that the colonization of roots by AM fungi and their abundance of infective propagules should be low in low-lying often saturated soils of floodplain forests but higher in more aerated zones. Alternatively, if factors other than soil moisture such as vegetation or soil nutritional gradients are more important, then clear relationships between soil moisture and AM fungi may not exist. To test these predictions, we examined the relationships between AM fungal root colonization and inoculum propagules, root density, soil moisture, and a multi-factor ecological zone concept characterized by micro-topography and vegetation. METHODS Sites We examined four study areas in temperate floodplain forests along the Kalamazoo River in Kalamazoo County, Michigan, USA (in order from upstream to downstream): the Augusta Floodplain Reserve (AUG; 42019'N and 85021'W); the Galesburg-Augusta High School River Preserve (GAL; 42016'N and 85025'W); the Comstock Preservation Area (COM; 42017'N and 85031'W) and the Kalamazoo Nature Center (KAZ; 42021'N and 85034'W). These forests consist of deciduous tree species such as Acer saccharinum L., Fraxinus spp., and Ulmus americana L. as well as the less prevalent Plantanus occidentalis L. and Gleditsia tricanthos L. that grow along a complex soilhydrologic and geo-morphological gradient. This gradient ranges from low-lying moist, potentially anoxic soils that are often annually flooded and are virtually devoid of herbaceous ground cover to higher areas containing somewhat drier soil, abundant herbaceous vegetation and more diverse tree communities. Conceptually this mosaic of vegetation and geomorphology resembles models proposed to describe other riparian ecosystems (Lindsey et al. 1961; Gregory et al. 1991; Merrit and Lawson 1992; Mitsch and Gosselink 1993). Mean annual precipitation is about 100 cm and mean annual temperature is about 9.4Â~C. The closest US Geological Survey river hydrology gauging station to these sites has measured 30 historic crests (near bank-full to exceeding flood stage) in the last 70 years ( Soil sample locations We collected soil samples from three ecological zones that characterize these floodplain forests. In the lowest lying areas, where soil appeared to be nearly saturated or had been inundated for the longest period, herbaceous vegetation was generally absent except for an occasional aquatic macrophyte. The tree canopy here was dominated by A. saccharinum. These micro-topographic depressions were often inundated during over-bank water flows or rises in the water table. We designated this habitat as "wet zone." The other clearly distinguishable habitat was higher and contained drier soils, dense herbaceous vegetation, a varying degree of shrub cover and canopy dominants that in

Page  99 ï~~2007 THE MICHIGAN BOTANIST 99 cluded A. saccharinum as well as Fraxinus spp., U. americana, and Tilia americana. Here, flooding is thought to occur only during the greater flood pulses. We designated this habitat as "dry zone." Between these two habitats we typically found areas of transition that were variable in shape and scale relative to the other zones but did not always exist if topography changed abruptly. These habitats contained sparse herbaceous vegetation, moderately moist soil and a sloping topography. We designated these zones as "transition." Samples were collected from 13-20 October 1994. Within each floodplain location (AUG, GAL, COM, KAZ) two sampling transects were selected that contained the three distinct ecological zones described above. Within each transect, two soil samples were extracted from each of the three zones. Thus, a total of 48 samples (4 locations x 2 sampling transects x 3 zones x 2 soil samples/zone) were collected. Sampling sites within an area were selected at points 1/3 and 2/3 along the length of the wet zone. Sampling sites for the transition and dry zones were then taken from points perpendicular to those in wet zone sites, approximately in the middle of each respective zone. If the dry zone often extended quite far (e.g., >20m), the sample site was marked >5m beyond the transition zone. Soil sampling Soil samples at each site were extracted with a "cup cutter" (a tool used by golf course greens keepers to make holes) with a diameter of 10.8 cm and set to a depth of -14.5 cm. The extracted cores were placed in a polyethylene bag and stored at 4Â~C. Each core was divided into three parts: Half for a mycorrhizal inoculum bioassay; one quarter to measure AM root colonization of in situ roots; and one quarter to determine soil moisture content and soil organic matter content, the latter measured as a part of a parallel study focused on describing the vegetation patterns in these forests and that also measured several other environmental features of each sampling site including the relative elevation and plant community composition and biomass (Avis P.G., Foster S. and Olexia P.D., unpublished). Soil moisture Soil moisture content was determined gravimetrically (Brower and Zar 1977). Thirty grams of homogenized soil sample were dried at 1050C for 24 h. The dry mass was determined and the percent soil moisture was calculated relative to the wet mass. Root clearing and staining of AM fungi Fine roots (<2 mm diameter) were autoclaved in 10% KOH for 10 minutes at 1210C to clear root cells of cytoplasm and rinsed with deionized water several times. Darkly pigmented roots required additional clearing and were bleached with 50% bleach acidified with a few drops of 5N HCl (Bevege 1968) and then acidified overnight in a 1% HCl solution (Koske and Gemma 1989). Fine roots were stained in 0.02% trypan blue in a 1:2:2 solution of lactic acid, glycerol and deionized water (Kormanik and McGraw 1982) in the autoclave for 15 min at 1210C and destained in 50% glycerol. Root density Total root length for each core was estimated by a modified Newman technique described by Tennant (1975). All roots were extracted from the sub-samples and fine roots were separated from larger diameter roots. The lengths of larger diameter roots were measured directly with a ruler. Stained fine roots were then cut into fragments not exceeding 2.54 cm, and placed in petri dishes containing destain solution. Root-line intersections were counted and counts were done five times for each sample and were then converted to root length (cm) using the formula described by Tennant (1975). Root density was determined by dividing root lengths by the volume of sample soil used, which had been determined by measuring how much water of a known volume the original soil core had displaced. In situ root colonization by AM fungi Root length colonization by AM fungi was determined on -10 cm of randomly selected root fragments from each sample following a slightly modified technique reported by Kormanik et al. (1980). Roots with 1-33% of the width colonized by AM hyphae, hyphal coils, vesicles and/or arbuscules

Page  100 ï~~100 THE MICHIGAN BOTANIST Vol. 46 were scored as "1", 34-66% colonization were scored as "2" and 67-100% colonization were scored as "3." Only roots with cortex were scored. AM fungal inoculum bioassay An AM fungal inoculum bioassay was conducted to measure the arbuscular mycorrhizal fungal propagules of sample soils following a similar protocol as in Corbin et al. (2003). Soil was homogenized by hand (shaking the bag) and, in duplicate, -135 ml of each sample were placed into conetainers (Stuewe and Sons, Inc., Corvallis, Oregon, USA) that were surface sanitized (soaked in 10% bleach for 10 minutes). Controls (sand, Consumers Sand and Concrete, Comstock, Michigan, USA; and autoclaved surplus sample soil) to check for contamination were set up similarly. Un-germinated hybrid Zea mays seeds (Early Golden Bantam, American Seed, South Easton, Massachusetts, USA) were placed in the top 1 cm of each filled conetainer and allowed to grow in a greenhouse. Plants were grown under ambient light supplemented by full spectrum growth lights (cycles of 16 hours of light, 8 hours of dark) and watered to field capacity approximately every other day. Plants were grown for three weeks and then roots were harvested, cut to 1 cm lengths, and stored in isopropyl alcohol before staining. AM fungi associated with bioassay roots were stained following a method modified after Brundrett et al. (1994). Cytoplasm of root cells was cleared by autoclaving at 1210C at 20 psi for 3 min in 10% KOH. After rinsing twice with deionized water, roots were stained for 24 hrs at room temperature (200C) in 0.3% Chlorozol black E in a 1:1:1 solution of lactic acid, glycerol and deionized water. To destain, roots were transferred to a 50% glycerol solution for 24 hrs at room temperature. For each bioassay sample, we analyzed from 40-100 cm of root for the number of colonization units formed from single entry points per length of root observed. The total root length examined varied since the samples did not have equal amounts of root growth. Ten of the stained 1 cm root fragments were placed parallel to each other in 50% glycerol on a microscope slide and between 4 and 10 slides were made per plant depending on the amount of root produced by the time of harvest. Slides were examined microscopically at 40x and 100x. Colonization units were determined as in Corbin et al. (2003), which were counted as the number of separate clusters of appresoria, hyphae, arbuscules, and/or vesicles developed during primary colonization of the root. Statistics After checking for normality and log-transforming non-normal data, we tested the effect of zone, soil moisture, site and transect on root density, mycorrhizal score and bioassay root colonization by analysis of covariance. Zone, site and transect were considered fixed effects and soil moisture and the log of root density (where it was added as an independent variable) as continuous covariates. Models were developed to examine the joint and separate impacts of zone and soil moisture while controlling for site and transect. We also tested the impact of root density in regressions of mycorrhizal score and bioassay root colonization. For our analyses we used PROC MIXED in SAS (v8.2; SAS Institute Inc., Cary, North Carolina, USA); the correlations among root density, mycorrhizal score and bioassay root colonization were also computed with PROC CORR. RESULTS Soil moisture was not significant in any of the regressions we tested (Table I). However, soil moisture was significantly correlated to zone (mean gravimetric soil moisture for all sites was 54.5, 46.2, and 38.7% for the wet, transition and dry zones, respectively; Pearson correlation = -0.58, p < 0.0001) and zone had a significant effect on log in situ root colonization when moisture was removed from regression and when root density was included (Table I). For log root density, log in situ root colonization and log bioassay root colonization, the effect of site was significant in most regressions but the effect of transect was significant only for log root density and log in situ root colonization

Page  101 ï~~2007 THE MICHIGAN BOTANIST 101 TABLE I. Regression statistics for log root density, log in situ root colonization and bioassay root colonization. Values are those from regression models that include all effects listed except for root density or unless indicated otherwise. Where log root density is listed as an effect, values include those models that included root density, site, transect and zone. Regressions significant in complete model were also significant when regressions did not include root density and zone (if not included).,*,, *** indicate p-values <0.05; <0.01; <0.001, respectively. Parameter Effect df F p-value Log root density Soil moisture 1,36 2.33 0.1355 Zone 2,36 0.44 0.6494 Site 3,36 7.01 0.0008*** Transect(Site) 4,36 2.92 0.0342* Log in situ root colonization Soil moisture 1,34 0.24 0.6281 Zone 2,34 2.31 0.1146 Zone 2, 35 3.38 0.0454* (soil moisture not included) Zone 2,34 4.03 0.0268* (root density included but not soil moisture) Site 3,34 1.98 0.1357 Site 3,34 3.29 0.0322* (root density included) Transect(Site) 4,34 3.29 0.0222* Transect(Site) 4,34 4.64 0.0043** (root density included) Log root density 1,34 3.20 0.0825 Log bioassay root colonization Soil moisture 1,31 0.33 0.5713 Zone 2,31 0.50 0.6138 Site 3,31 4.60 0.0089** Transect(Site) 4,31 1.28 0.2973 Log root density 1,30 15.78 0.0004*** (Table I). Figure 1 illustrates the impact of site on the distribution of these parameters in relation to zone. Each site appeared to have distinctive root densities that varied independent of zone. In situ root colonization and bioassay root colonization tended to increase from wet to upland in most sites but sites generally had different levels of each of these parameters. For example, the COM site had the highest and the KAZ site the lowest overall root density with GAL and AUG sites intermediate. In contrast, KAZ and AUG sites had the highest but COM and GAL the lowest in situ root % AM fungal colonization; and the COM, GAL, and AUG sites had higher bioassay root colonization than the KAZ site. Root density was significantly positively correlated to bioassay root colonization (Pearson correlation = 0.70, p < 0.0001) but in situ root colonization was not significantly correlated to root density or bioassay root colonization. DISCUSSION AM fungi colonized roots and AM fungal inoculum was found in soils collected from all three ecological zones of the floodplain forests we examined even in the low-lying, relatively saturated "wet" zone. Although soil moisture was strongly correlated to the zone gradient along which we sampled and mycor

Page  102 ï~~102 THE MICHIGAN BOTANIST Vol. 46 102 THE MICHIGAN BOTANIST Vol. 46 "--*- AU iCT 0.1.4 0 1.2 I6 0 4 - oT S2.0 - 064 E0 - 25 -S5 ___ -"- COM ---- GAL KAZ vet-transihcvn upk i Zone FIGURE 1. Mean root density (cm/cm-3), in situ percent AM fungal root colonization (mean mycorrhizal score) and inoculum bioassay root colonization (infection units/cm root) in each site. Bars represent s.e. rhizal parameters tended to be higher in the "dry" zone, soil moisture had no statistically significant effect on any of the mycorrhizal parameters tested. Therefore, our study provides little support for the hypothesis that saturated soils inhibit root colonization or inoculum production by AM fungi in these forests. Two caveats are important to consider, though. We measured soil moisture and mycorrhizal parameters at only one time and the impact of saturation may be more evident at other times of the year such as during spring floods. Also, our bioassay only tested for the presence of inoculum and not the direct influence of soil saturation on the production and/or infectiveness of inoculum. But, even if

Page  103 ï~~2007 THE MICHIGAN BOTANIST 103 soil moisture does impact AM fungi at times other than when we measured, inoculum was present in the most saturated soil we sampled possibly because flood pulses moved propagules from higher to lower areas of the floodplain landscape (Miller 2000; Wetzel and van der valk 1996; Kahlil and Loynachan 1994) and/or when stressed AM fungi produced more inoculum (Rickerl et al. 1994). As a result, this inoculum could ensure colonization of host roots when moisture levels become favorable for AM development. The floodplain forest AM fungi studied here are similar to those in other wetlands (e.g. Wetzel and van der valk 1996) in that they appear to be impacted by factors other than soil saturation. In our study, site and ecological zone appeared to be important. Site had a consistent effect on all parameters indicating that the floodplain forests from which we sampled were quite different environments for AM fungi. For example, despite having relatively high in situ root colonization, the KAZ site had very low colonization in inoculum bioassays. These site differences may result from differences in the component species of the AM fungal communities and/or vegetation as vegetation patterns differ between sites (Avis, Foster and Olexia, unpublished). Similarly, within site differences may also explain the significant effect of transect on root density and in situ root colonization. Zone also had a significant effect on in situ root colonization and this effect was even more pronounced when root density was included in the regression. Our zonal concept encompassed multiple factors. We believe that at least for AM fungal colonization of floodplain roots, the zone concept provided a surrogate measure for factors such as host plant type and root growth characteristics which are likely important to AM fungal distribution in these forests regardless of soil moisture levels. Other important factors this concept may incorporate include soil nutrient and/or soil contamination (e.g. PCB's from industrial activity along the river). We were not able to measure either of these factors but both may vary along these same gradients. An interesting and unexpected result we found was the significant positive relationship between root density and AM inoculum. This suggests that roots serve as important vectors of AM inoculum in these forests which is consistent with the understanding that AM fungal inoculum includes spores as well as hyphae from infected roots (Harley and Smith 1983). However, the lack of relationship between root density and in situ root colonization suggests that the infective component of roots is extra-radical (e.g. hyphae attached to but outside the root) and that the abundance of infective extra-radical hyphae was not directly related to the AM fungal colonization structures we examined in the cortex of the roots. AM fungi are distributed throughout the floodplain forests we studied but a clear relationship to soil moisture was not found. As a result, the potential influence of floodplain AM fungi to plant growth and carbon sequestration in floodplains and to link terrestrial and aquatic components of riparian ecosystems appears widespread. ACKNOWLEDGMENTS We thank T. Darcy, L. Zimmerman, P. Sotherland, and T. Straw for assistance and advice during the completion of this study; D. Johnson for statistical consultation; the Kalamazoo Country Club for use of their cup cutter; and Kalamazoo College's Diebold Fellowship for financial support.

Page  104 ï~~104 THE MICHIGAN BOTANIST Vol. 46 REFERENCES Bevege, D.I. (1968). A rapid technique for clearing tannins and staining intact roots for detection of mycorrhizas caused by Endogone spp. Trans. Brit. Mycol. Soc. 51(5):808-811. Brower, E.J., Zar, J.H. (1977). Field and laboratory methods for general ecology. WC Brown Company Publishers, Dubuque, Iowa, USA. Brundrett, M., Melville, L., Peterson, L. (1994). Practical methods in mycorrhiza research. Mycologue Publications, Ontario, Canada. Corbin, J.D., Avis, P.G., Wilbur, R.B. (2003). The role of phosphorus availability in the response of soil nitrogen cycling, understory vegetation and arbuscular mycorrhizal inoculum potential to elevated nitrogen inputs. Water, air and soil pollution 147:141-161. Entry, J.A., Rygiewicz, P.T., Watrud, L.S., Donnelly, P.K. (2002). Arbuscular mycorrhizal response to adverse soil condition. In Arbuscular Mycorrhizae Interactions in Plants, Rhizosphere and Soils edited by Sharma, A.K., Johri, B.N. Science Publishers, Enfield, New Hampshire, USA, pp 135-158. Escudero, V., Mendoza, R. (2005). Seasonal variation of arbuscular mycorrhizal fungi in temperate grasslands along a wide hydrologic gradient. Mycorrhiza 15(4):291-299. Gregory, S.V., Swanson, F.J., McKee, W.A., Cummins, K.W. (1991). An ecosystem perspective of riparian zones. Bioscience 41(8): 540-551. Harley, J.L., Smith, S.E. (1983). Mycorrhizal Symbiosis. Academic Press, London, United Kingdom. Harner, M.J., Ramsey, P.W., Rillig, M.C. (2004). Protein accumulation and distribution in floodplain soils and river foam. Ecology Letters 7(9): 829-836. Khalil, S., Loynachan, T.E. (1994). Soil drainage and distribution of VAM fungi in two toposequences. Soil Biol. Biochem. 26(8): 929-934. Kormanik, PP., Bryan, W.C., Schultz, R.C. (1980). Procedures and equipment for staining large numbers of plant root samples for endomycorrhizal assay. Can. J. Microbiol. 26:536-538. Kormanik, P.P, McGraw, A.C. (1982). Quantification of vesicular-arbuscular Mycorrhizae in lant roots. In Methods and Principles of Mycorrhizal Research edited by Schenck, N.C. The American Phytopathological Society, St. Paul, Minnesota, USA. Koske, R.E., Gemma, J.N. (1989). A modified procedure for staining roots to detect VA mycorrhizas. Mycol. Res. 92(4):486-488. Lindsey, A.A., Petty, R.O., Sterling, D.K., VanAsdall, W. (1961) Vegetation and environment along the Wabash and Tippecanoe Rivers. Ecological Monographs 31(2):105-156. Merrit, R.W., Lawson, D.L. (1992). The role of leaf litter macroinvertebrates in stream-floodplain dynamics. Hydrobiologia 248:65-77. Miller, S.P. (2000). Arbuscular mycorrhizal colonization of semi-aquatic grasses along a wide hydrologic gradient. New Phytologist 145:145-155. Miller, S.P., Sharitz, R.R. (2000). Manipulation of flooding and arbuscular mycorrhiza formation influences growth and nutrition of two semiaquatic grass species. Functional Ecology 14:738-748. Mitsch, W.J., Gosselink, J.G. (1993). Wetlands 2nd ed, Van Nostrand Rienhold, New York, USA. Mosse, B.D., Stribley, D.P., LeTacon, F. (1981). Ecology of Mycorrhizae and mycorrhizal fungi. In Advances in Microbial Ecology edited by Alexander, M. Plenum Press, New York, pp 137-210. Pawlowska, T.E., Charvat, I. (2002). Influence of edaphic and environmental factors on arbuscular Mycorrhizae. In Arbuscular Mycorrhizae Interactions in Plants, Rhizosphere and Soils edited by Sharma, A.K., Johri, B.N. Science Publishers, Enfield, New Hampshire, USA, pp 105-134. Rickerl, D.H., Sancho, F.O., Ananth, S. (1994). Vesicular-arbuscular endomycorrhizal colonization of wetland plants. Journal of Environmental Quality 23:913-916. Tennant, D. (1975). A test of a modified line intersect method of estimating root length. Journal of Ecology 63:995-1001. Terborgh, J., Andresen, E. (1998). The composition of Amazonian forests: patterns at local and regional scales. Journal of Tropical Ecology 14:645-664. Troeh, Z.I., Loynachan, T.E. (2003). Endomycorrhizal fungal survival in continuous corn, soybean, and fallow. Agron. J. 95:224-230. Van Hoewyk, D., Wigand, C., Groffman, P.M. (2001). Endomycorrhizal colonization of Dasiphora floribunda, a native plant species of calcareous wetlands in eastern New York State, USA. Wetlands 21(3):431-436. Wetzel, P.R., van der Valk, A.G. (1996). Vesicular-arbuscular Mycorrhizae in prairie pothole wetland vegetation in Iowa and North Dakota. Canadian Journal of Botany 74:883-890.