/3’? 0'5/ ”TH; PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5108 KIProj/AccAPdelRC/Datoouejndd EPIDEMIOLOGY OF SCLEROTINIA HOMOEOCARPA IN MICHIGAN: GEOSTATISTICAL AND POPULATION BIOLOGICAL APPROACHES By Brandon Joseph Horvath A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Plant Pathology 2003 ABSTRACT EPIDEMIOLOGY OF SCLEROTINIA HOMOEOCARPA IN MICHIGAN: GEOSTATISTICAL AND POPULATION BIOLOGICAL APPROACHES By Brandon Joseph Horvath Dollar spot is a severe turfgrass pathogen in North America, and in particular in Michigan. The disease is caused by the pathogen Sclerotinia homoeocarpa F .T. Bennett. This fungus is not known to produce sexual or asexual spores, and therefore, its primary mode of transport is via infected grass clippings on equipment and humans. As fungicide use becomes more restricted, it is important to have a basic understanding of the epidemiology of the major turfgrass pathogens. With'this goal in mind, this study used both geostatistical and population biological approaches to better understand the epidemiology of this pathogen. The objectives for this project were: 1) to determine if isolates of S. homoeocarpa from golf courses in Michigan could be differentiated using amplified fragment length polymorphism (AFLP) markers and vegetative compatibility groups (VCGs), and 2) to quantify the spatial structure of dollar spot incidence and determine its temporal stability. Five populations were sampled for this study. Using 32 isolates subsampled from these five populations, AF LPs and VCGs were defined for each isolate. I found no relationship between an isolate’s AFLP fingerprint and its VCG. Nor was there any apparent relationship between isolates based on geographic location since some isolates from opposite parts of the state shared the same fingerprint. A total of 889 isolates were collected from three of the populations for further study of the VCG distributions in populations. While 860 isolates fit into the 6 known VCGs present in Michigan, there were also 29 isolates that did not fit in these 6 groups raising the possibility of the presence of additional groups. Chi-square analysis revealed significant differences between VCG distributions between years and locations. At two of the three locations, differences were also found between fairways within a location indicating that each fainNay behaved independently. The geostatistical study was established on a 9.1 m X 18.3 m area of creeping bentgrass (Agrostis palustn's Huds.) and annual bluegrass (Poa annua L.) at the Robert Hancock Turfgrass Research Center in E. Lansing, MI. The study area was subdivided into 223 areas where dollar spot foci were counted over the entire season. Variograms of disease incidence were constructed for each date and showed clear spatial structuring at relatively small scales (~0-1O m). Closer examination of the variogram model parameters showed that the nugget and sill parameters scaled with each other while the range parameter remained fairly constant within each season. Between 50 and 60 percent of the total population variance in each year was spatially structured. This indicates that the spatial structure of dollar spot remains relatively unchanged regardless of disease severity, suggesting that the factor primarily responsible for the spatial pattern is one that does not move about in space. Copyright by Brandon Joseph Horvath 2003 To all those along the way who believed that this day would come. ACKNOWLEDGMENTS I always read with interest other thesis and dissertation acknowledgements. | find myself wondering where they were and what they were thinking about when they set down for posterity those names of people that they would like to thank. I also find myself thinking about what this will mean to me in 30 years when I open it and read it again but instead of being 30 I will be 60 and in the twilight of my career, not setting off on the beginning of it. The first group of people I would like to say thank you to is my family. Mom and Dad, you have been there when I needed you most. Through all the rough spots when I felt like I was going it alone you were there. Nate, you’re my brother, confidant, and one of the best human beings I know. I am proud of all the accomplishments you have made. You truly have defined who you are as a person. Grandpa, your life ended before I could finish this project, but I was very proud of the fact that I wasn’t the first person in our family to pursue an agricultural education and I hope you are proud of my accomplishments. Grandma, not everyone can say they have had the relationship with their grandma that l have had with you. Knowing l was just a phone call away from a hot, home-cooked meal in college will mean more than you’ll ever know. Vi, Mim, Mac, and Ash- I can’t describe what it has meant to me personally to have such good, close relationships with you all. I don’t have enough room to mention all the people whom I have had the distinct privilege of calling my friends. Jason, you have become a true friend, helping me selflessly when my chips were down, no thank you will ever be enough. Ron and vi Cindy Detweiler, if you hadn’t opened your home and hearts to me twelve months ago, I don’t think I would be writing this today. A thank you doesn’t seem like enough. Nancy, thanks for all the little tips and tricks that have helped me navigate through graduate school, I will miss the lunches. Phil, thanks for the game it was fun. My committee, Dr. Crum, Dr. Robertson, and Dr. Jarosz thank you for helping me to steer a pretty clean course through my PhD. I learned a lot about how to do good research from all of you. Dr. Sasha Kravchenko, without your help and instruction I wouldn’t understand what geostatistics were let alone be able to explain how to do them. Dr. Kurt Lamour and Dr. Mary Hausbeck, your assistance and collaboration allowed me to complete a section of this dissertation on AFLPs without having to reinvent the wheel. Joe, you have been my mentor and have treated my like a colleague. I plan on being your colleague for a long time to come. I would also be remiss if I didn’t mention a few select people that are what I would call “shadow mentors”, people that have had an immeasurable impact on your life without being asked. Dr. Jim Miller, your class changed my worldview of science forever and I am eternally grateful for those interactions. Dr. Ken Poff, thank you for taking an interest in me and passing on some of your wisdom, you probably don’t realize the impact it has had on my view of science. Doubt, in the sense of not knowing what the future holds is a good thing. Jennifer, I struggled with what to say here that would be worthy of how I feel, and I decided on this: I love you. I don’t know what the future holds, but I do hope that you will be in it with me. vii TABLE OF CONTENTS LIST OF TABLES ix LIST OF FIGURES x QtiAEEEBJ VARIATION OF SCLEROTINIA HOMOEOCARPA WITHIN AND AMONG GOLF COURSES IN MICHIGAN INTRODUCTION 1 MATERIALS AND METHODS 4 RESULTS 10 DISCUSSION 15 LITERATURE CITED 21 CHAEIERZ GEOSTATISTICAL ANALYSIS OFDOLLAR SPOT EPIDEMICS IN MICHIGAN INTRODUCTION 24 MATERIALS AND METHODS 28 RESULTS 32 DISCUSSION 39 LITERATURE CITED 45 APPENDICES 48 APPENDIX 1- Raw indicator semivariograms of isolate VCGS 49 APPENDIX 2- Raw semivariograms for all dates from 2000-2002 50 viii LIST OF TABLES CHAELEBJ TABLE 1- Sampled isolates for AF LP fingerprinting TABLE 2- Vegetative compatibility group (VCG) distributions of Sclerotinia homoeocarpa isolated in 2000 & 2001 from four fairways at each location TABLE 3- Results of chi-square analyses of VCG distributions of Sclerotinia homoeocarpa from 3 populations in Michigan QtIAflEBZ TABLE 1- Variogram model parameters (nugget, sill, range), goodness of fit (F), and residual sum of squares (RSS) for the nine experimental variograms shown in Figure 3 TABLE 2- Variogram model parameters for each rating date in 2000-2002 showing nugget, sill, and proportion of structural variance values 11 14 37 38 LIST OF FIGURES QtIAEIEBJ FIGURE 1- Map depicting geographical location of 5 populations of S. homoeocarpa populations sampled for VCG and AFLP analysis FIGURE 2- Images showing surface (top) and reverse (bottom) views of mycelial interactions of a S. homoeocarpa isolate (center of image) against tester isolates (surrounding center) in a petri dish FIGURE 3- Genetic similarity of 32 S. homoeocarpa isolates sampled from five populations in Michigan based on 15 polymorphic AF LP markers CHARTER] FIGURE 1- Schematic layout of study area at the Robert Hancock Turfgrass Research Center (E. Lansing, MI) showing overall arrangement of sampling locations FIGURE 2- Graph comparing total disease progress from 2000- 2002 as measured by counts of total disease foci in all sample locations FIGURE 3- Experimental variograms from 2000-2002 measuring spatial dependence from the early, middle, and late phases of the dollar spot epidemic FIGURE 4- Graph showing the proportion of structural variance (C/Co+C) over time for exponential variogram models from 2000-2002 13 14 29 34 35 36 Chapter 1 VARIATION OF SCLEROTINIA HOMOEOCARPA WITHIN AND AMONG GOLF COURSES IN MICHIGAN Introduction Dollar spot disease of turfgrasses is caused by the pathogen, Sclerotinia homoeocarpa F .T. Bennett (3). The disease is common throughout the world and is destructive to both cool and warm season grasses (20, 21, 23). In North America, with the exception of the Pacific Northwest, dollar spot is the most important pathogen of most cultivated fine turfgrass species (20, 21, 23). In Michigan, the disease is a major problem for most golf courses where the epidemic begins in June and can continue into late September causing extensive damage if left untreated. The disease can blight large areas of turf as a result of coalescing disease foci. Diseased turf has a poor aesthetic appearance, impairs the playing surface by creating depressions that affect ball roll, and leaves areas of bare soil where weed species can encroach on the area (20, 23). The pathogen is not known to produce conidia or undergo sexual reproduction in North America (1, 10, 11). Jackson found that while United Kingdom isolates of S. homoeocarpa will undergo sexual reproduction, no isolate from the US has been known to develop fertile apothecia (11). However, Hsiang and Mahuku (10) reported that some populations in Southern Ontario had random amplified polymorphic DNA (RAPD) patterns consistent with recombination events within a local population. It is commonly assumed that S. homoeocarpa is disseminated via direct transfer of mycelium from infected leaves (7, 20, 21, 23). Control of dollar spot via fungicide application is generally accomplished using the contact fungicide, chlorothalonil. However, the EPA is expected to restrict the use of this fungicide on golf courses. In the event that a limited amount of fungicide is available to a golf course it is critical that superintendents be able to apply chlorothalonil judiciously. Other single-site mode of action fungicides are available to control dollar spot, but fungicide resistanceis a problem in many dollar spot populations (6, 9, 25). Vegetative compatibility is the ability of the pathogen to form a stable heterokaryon as a result of a self/nonself genetic recognition event when two individual strains fuse (8, 16). The systems can be allelic or non-allelic in nature. Fungi that have an allelic compatibility system determine whether two strains are compatible via identity of alleles at a particular compatibility locus. In contrast, a non-allelic system usually involves alleles at multiple loci interacting to determine compatibility (16). Studies of compatibility are useful for studying diversity in populations, detecting new lineages in a local area, and observing population dynamics. Aspergillis flavus was examined in a cotton field using vegetative compatibility groups (VCGs) as a measure of genetic diversity (2). Large numbers of VCGs were identified and the distribution changed from year to year over the three-year study. The large number of VCGs suggests a large change in the genetic makeup of the population each season. The authors suggested the observed diversity could be a result of the migration of conidia from other locations and/or a seasonal change in the number of strains making up each VCG. Kohn et al. (15) studied mycelial compatibility, a specific component of vegetative compatibility, in Sclerotinia sclerotiorum. Using DNA fingerprinting techniques, they found that the mycelial compatibility group (MCG) diversity was high and that MCGs made up genotypically distinct lineages. The observed diversity was attributed to be due to the occasional outbreeding event and the migration of new strains into populations. Powell and Vargas (18) identified 6 VCGs from isolates sampled from creeping bentgrass and annual bluegrass from 8 locations in Michigan and the Midwest. They found that the VCG distributions at a location change over a season. They also reported that isolates from the same VCG could be isolated from both creeping bentgrass and annual bluegrass indicating that host specificity is not associated with particular VCGs. Using the sequence of the nuclear internal transcribed spacer region 1 (ITS1) they also found that all sampled isolates shared the same sequence and were from the same species. Raina et al. (19) studied the genetic variability of S. homoeocarpa using RAPDs and found that isolates of dollar spot from the midwest and northeastern United States were very similar. Both of these studies support the empirical evidence that S. homoeocarpa is a clonal pathogen. In contrast, Sonoda (22) identified more than 54 VCGs of S. homoeocarpa isolated from bermudagrass (Cynodon dactylon). One hundred nineteen isolates were collected from three locations; nearly 50% represented VCGs, indicating a significant amount of genetic exchange or migration. Hsiang and Mahuku’s (10) study using RAPDs of dollar spot isolates from eight populations in Southern Ontario supported random mating in three of the eight populations sampled. Many different molecular tools are available for the study of plant pathogens (5, 10, 12, 15, 17, 19). lsozymes are relatively inexpensive, but problems often occur in generating enough polymorphic markers to be of use. RFLPs (restriction fragment length polymorphisms) can often be very informative, however suitable DNA probes must be available. The widely used RAPDs (random amplified polymorphic DNA) suffer most generating reproducible results because of sensitivity to running conditions. AFLPs (amplified fragment length polymorphisms) are noted for their ability to rapidly generate large numbers of reproducible and neutral (not under independent selection) markers at independent loci (17, 24). AFLPs avoid the problems inherent in most other tools used for fungal genetic analysis. The primary drawback to AFLPs is the relatively high startup cost. Cilliers et al. (5) used AFLP analysis to differentiate isolates and MCGs of Sclerotium rolfsii from South Africa. They identified 9 MCGs in a collection of 73 isolates from 10 locations in South Africa. Isolates were identified with a specific MCG using AFLPs. The objective of this study was to determine if isolates of S. homoeocarpa from golf courses in Michigan could be differentiated using AFLP markers and VCGs Materials and Methods Isolation and Culture. Isolates were collected from symptomatic plants infected with S. homoeocarpa. Three different locations in Michigan were sampled in July 2000 and 2001 (Fig. 1). Four fairways at each location were selected for sampling from which symptomatic leaves of 50 infection centers along a transect running the length of the fairway were individually collected in paper coin envelopes. Two to three small segments of leaf tissue displaying lesions were placed on acidified water agar plates (10mL lactic acid/L) and allowed to grow for 2-3 days at 25°C. Hyphae growing out of the leaf tissue were then isolated onto potato dextrose agar (PDA) plates and allowed to grow for about 5 days at 25°C. Using a modified method of Boesewinkel (4), ten agar plugs were removed from the PDA plates using a sterile coffee stirrer and placed in 1.5 mL microfuge tubes containing 1 mL sterile H20 for long term storage at room temperature. David Gilstrap provided additional isolates from 2 golf courses (Fig. 1) for the AF LP evaluation of the genetic diversity of S. homoeocarpa. They were stored in the same manner as the other isolates. VCGS. All isolates were paired with six tester isolates representing the six known VCGs in Michigan (18) using a method modified from Kohn (15). Sets of four isolates were paired against all six tester isolates in 24 well culture plates. Each well contained 1 mL of PDA amended with 5 drops/L of McCorrnlck’s Red Food Coloring to highlight antagonistic zones. Each isolate was also paired with itself as a control. Isolates that were not classified in the first screening were then paired with all six tester isolates on 100 X 15 mm petri dishes to clarify the interactions between the isolate and tester strains. Chi-square (X2) analyses were performed on the observed frequency distributions of the three most frequent VCGs (A, B, C) using the null hypothesis that there were Legend Population ID Location Name City/State 6C1 Alpine GC Grand Rapids, MI GC2 The Emerald GC St. Johns, MI 603 Maple Lane GC Warren, MI Used for AFLP analysis (provided by Gilstrap) / 6C4 Lochmoor GC Grosse Pointe Woods, MI 605 Hancock Turfgrass E. Lansing, MI Research Center 6:1 ec2 . eca O O 605 O GC4 Figure 1. Map depicting geographical location of 5 S. homoeocarpa populations sampled for VCG and AFLP analysis. Legend indicates population ID, name of location, and city/state. no differences in the frequency distributions of these VCGs between fairways within golf courses, within golf courses, or between years. Isolates provided by Gilstrap were also classified into VCGs using the same techniques described above for the other isolates. However, these isolates were not subjected to chi- square analysis due to a different sampling scheme. DNA extraction and AFLP fingerprinting. A subset of isolates from the three populations sampled for this study and the additional isolates provided by Gilstrap were fingerprinted using the AFLP technique (Table 1). Isolates were grown in approximately 20 mL of potato dextrose broth in 100 x 15 mm petri dishes for seven days at 23 to 25°C. Mycelial mats were washed with distilled water and dried briefly under vacuum before being frozen to -20°C and Iyophilized. Lyophilized mats were ground with a sterile mortar and pestle. Whole genomic DNA from approximately 50 mg of ground mycelium was extracted using a QIAGEN Dneasy Plant Mini Kit (QIAGEN lnc., Valencia, CA) according to the manufacturer’s directions. DNA was quantified by comparing the intensity of illumination of a 1 uL drop on 1.5% agarose gels amended with ethidium bromide and viewed under UV light to known standards ranging from 10 to 250 ng/uL. Approximately 100 ng of DNA was then subjected to a restriction/ligation reaction, pre-selective amplification, and selective amplifications using the PCR core mix, adaptor sequences, core primer sequences and fluorescence labeled primers provided in the AFLPTM Microbial Fingerprinting Kit (Perkin-Elmer Corp., Isolate ID Year Isolated Population-VCG IDa A14-8-00 2000 GC1-D A16—1-01 2001 GC1-D ML7-29-00 2000 GC3-D ML8-46-00 2000 GC3-E ML11-19-01 2001 GC3—D ML12-40-00 2000 GC3-F E17-14-01 2001 GCZ-A ML7-7-00 2000 GC3-C A16-16-01 2001 GC1-E 1-7003-SH-R 1994 GC4-E 7039-SH-S 1998 GCS-E 1-7016-SH-R 1994 GC4-E ML7-11-01 2001 GCB-A 1-7024-SH-R 1994 GC4-C 1-7018-SH-R 1994 GC4-C 1-7021-SH-R 1994 GC4-C 1-7008-SH-R 1994 GC4-C 1-7005—SH-R 1994 GC4-E 1-7013-SH-R 1994 GC4-E 1-7004-SH-R 1994 GC4-C 7041-SH-S 1998 GC5-C A9-36-01 2001 GC1-B 7034-SH-S 1998 GC5—A E4-3-01 2001 GC2-C E4-1-00 2000 GC2-B 1-7015-SH-R 1994 GC4-E A9-10-01 2001 GC1-C ML7-2-01 2001 GC3-B 7043-SH-S 1998 GC5-B 7033-SH-S 1998 GC5-B 7040-SH-S 1998 GC5-B 7036-SH-S 1998 GCS—‘B a Designation of isolate in Fig. 3 listed by population ID and VCG. Table 1. Sampled isolates for AFLP fingerprinting. Isolates are listed in the order from top-bottom as they appear in Fig. 3 . Foster City, CA) and performed exactly as described in the PE/ABI AFLP Microbial Fingerprinting protocol parI# 402977 Rev A. All PCR reactions were performed using an MJ Research Minicycler (MJ Research Inc., Waltham, MA) in 0.2 mL tubes according to the cycling parameters outlined in the microbial fingerprinting protocol. An initial optimization set of reactions was performed using pre-selective products from two randomly chosen isolates. Amplifications with the selective primers EcoRI-AA, AC, AG and AT were performed in all 16 combinations with the Msel-CA, CC, CG and CT selective primers. EcoRl selective primers were labeled at the 5’ end with either carboxyfluorescein (FAM), oarboxytetramethyrhodamine (TAMRA), or carboxy-4',5'-dichloro-2',7'- dimethoxyfluorescein (JOE) fluorescent dyes. The fluorescent dyes were excited by laser radiation and visualized by their characteristic absorption-emission frequencies. Only the fragments containing an EcoRI restriction site were resolved. Selective amplification AFLP products and a carboxy-X-rhodamine (ROX) size standard were loaded into each lane on a denaturing polyacrylamide gel and the fragments resolved in an ABI 3700 DNA Sequencer. Results were prepared for analysis in the form of electropherograms using GeneScan Analysis software (PE/ABI). AFLP fragments were scored manually as present = 1 or absent = 0 using Genotyper software (PE/ABI). Only DNA bands that consistently exhibited unambiguous presence/absence profiles were scored. Using the program NTSYS—pc (Rohlf, F. J. 1993. NTSYS-pc - Numerical Taxonomy and Multivariate Analysis System, Version 2.02k. Applied Biostatistics Inc.), the combined on data matrix for isolates was used to construct a genetic similarity matrix of all possible pairwise comparisons of individuals using Jaccard’s similarity coefficient: GS(ij) = a/(a + b + c). GS(ij) is the measure of genetic similarity between individuals iand j, where a is the number of polymorphic bands shared by iand j, b is the number of bands present in iand absent in j, and c is the number of bands present in j but absent in i. Trees were constructed using unweighted pair group with mathematical averaging (UPGMA) cluster analysis to provide a graphical representation of the relationships among isolates. Results Isolation and Culture. A total of 1200 samples of S. homoeocarpa were collected from three golf courses in Michigan. Of the 1200 samples, 889 isolates were placed in pure culture and stored. The collection efficiency (% success in obtaining an isolate from a sampled spot) was 74.1%. Isolates that were stored in H20 at room temperature have been routinely recovered over the entire course of our study. VCGS. 860 isolates were placed into one of the six VCGs (Table 2) described by Powell and Vargas (18). Isolates were scored as compatible when 10 .5382 some Fm £9.53 50.. E0: 88 w 88 c. 85.8. 388052. m.=...eo.ow 3 22.3.56 30>. 955 3.3.858 9:933 .N «33 8» «N 3 Na FF EN 84 FF .83 220 SN FF a F F 8 E. mml I .83 as... can: No F N o F. m NF .N I .83 88 F N F o o N F FF NF 8 FF N o F o n F FF a F. o o N c o N a 2 o o o F c o n F I68 I F.FF o W NF N 8 8 F I .83 88 NF. o N m o 2 FN o NF 8 o o n F FF 9 o FF BF o F N F N 2 o m 9. o o F F N FF F F °8N 2.: 2%... on." FF N a o 2. NNN FF I I .53 2285. 9.: o N F o 8 FF. FF .83 88 F.. o o e o FF N e F 8 F F F c F FN N FF 3 N o N o m 8 N B NF. B N o o 2 F N 4 I58 1 BF N o N 8 FN F: o .83 88 9. F o o o NF N» o FF 8 o o o o N F» o FF 3 o o o o S 8 o m 3 F o N o m 8 c e 88 2825 8..” F F o F. FF 2: Fe .83 2...? m: o o o N NF FFI NF .83 88 NF. o o F N o ON 2 BF 3. o o F o c 2 N 3 NF. o o N o F. 8 8 NF 8 o o N o B NF 3 m IFBN l 8F F F o N B B: N .53 88 3 o F o F NF 8 F 2 3 o o o F FF 8 F. 3 8 o o o o FF 8 o NF 3 F o o o 8 FN F m 88 22.2 .83 .880 55.0 n. m o o m < Ania“. =3 5.38.. 9.20 00> 11 there was no noticeable barrage zone between an isolate and a tester strain. Isolates were scored as incompatible when a barrage zone was formed between an isolate and a tester strain (Fig. 2). Twenty-nine isolates either did not fit in one of these six groups or the incompatibility reaction was indistinguishable, and were classified as “other” (Fig. 2). These isolates were ignored when isolates were sampled for the AFLP analysis and they were not subjected to chi-square analysis. Over the entire course of the study isolates from VCGs A, B, and C were found in all fairways of each golf course. VCGs D, E, and F were either absent completely or present at very low frequencies in each population. Of the three major VCGs, group B was present in the highest frequency over both years, followed by group C, and group A. Over all locations in 2000, group A was much less prevalent than in 2001. The reverse was true for group C where it was more frequent in 2000 than in 2001. Overall, fewer isolates were collected in 2001. Chi-square analysis showed there were significant differences in VCG frequency distributions between fainlvays within a golf course at the Maple Lane and Alpine locations (Table 3). The analysis also showed there were significant differences between locations and between years. AFLP Genotyping. The EcoRl + AC/ Msel + CA primer combination resolved the greatest number of clear fragments of the selective primers tested and resulted in more than 80 clearly resolved AF LP fragments in each of the 32 isolates analyzed. In total, 100 AFLP fragments were resolved with 15 being 12 Figure 2. Images showing surface (top) and reverse (bottom) views of mycelial interactions of a S. homoeocarpa isolate (center of Image) against tester isolates (surrounding center) in a petri dish. Comparison Chi-square d.f. P-value Within Maple Lane GC fairways 14.29 6 .0266 Within Emerald GC fairways 8.99 6 .1741 Within Alpine GC fairways 31.73 6 <.0001 Between Locations 79.16 4 <.0001 Between 2000 and 2001 150.16 2 <.0001 Table 3. Results of chi-square analyses of VCG distributions of Sclerotinia homoeocarpa from 3 populations in Michigan. 14 present in some isolates and absent in others (polymorphic). Isolates were from 55 to 100% similar (Fig. 3). Isolates with identical AFLP profiles did not necessarily come from the same location or have the same VCG. Overall, isolates from the same location or with the same VCG were not more similar. Discussion Several Our evaluation of as many as 185 isolates from a single sampling at one location represents the largest sample of isolates of S. homoeocarpa ever examined for VCG diversity at a location. We found clear evidence that VCG distributions can vary within a golf course, among golf courses, and over time. These data adds to the findings of Powell and Vargas (18) who found that there were differences in the distribution of VCGs over time and location. Other studies have attempted to understand the structure of S. homoeocarpa populations (10, 18, 19, 22). It thus appears plausible that VCG distributions on each fairway within a golf course operate as independent populations, each with a unique distribution of VCGs. One exception in this study was the Emerald location in St. John's, Ml. Chi-square analysis for isolates within a fairway at this location revealed no significant differences in the VCG distributions between fairways. This golf course was completely redesigned and renovated in 1996 making it much younger than both Alpine and Maple Lane golf courses that are well established and have been in play for at least 25+ years. The predominance of the V065 A, B, and C in the populations sampled are similar to the results found 15 ‘ ccsie I I I I l T I I l I V I U I I I 0.35 0.66 ' ' 0.117 0.89 1.'oo Genetic simuhrity Figure 3. Genetic similarity of 32 S. homoeocarpa isolates sampled from five populations in Michigan based on 15 polymorphic AFLP markers. Populations are designated GC1-GC5 followed by the VCG. 16 for one of the fields sampled by Kohn et al.(15) for MCG diversity in S. sclerotiorum where they hypothesize that the relative lack of diversity in the field was indicative of the diversity that was initially introduced into the area or as a result of selection of strains from an initially diverse population. The evolutionary forces of drift and migration as well as the putative lack of sexual recombination in S. homoeocarpa can limit the number of VCGs found in a population (16, 18). Also, age of the golf courses, cultural practices, fungicide management regimes, and environmental conditions may all be potential factors in the development and distribution of VCGs of S. homoeocarpa. Further research should focus on developing testable theoretical models that seek to explain the variation in VCG distribution that has been observed within sampling locations, between sampling locations, and over time. It would also be worthwhile to investigate the possibility of bias in the sampling scheme used by both this study and Powell and Vargas (18) that is based on selecting a single isolate from a few infected lesions that were cultured from a single dollar spot. An exhaustive sampling scheme that characterizes the presence of all strains of S. homoeocarpa growing in a single dollar spot would serve to close this question. Finally, examining the VCG diversity that is present in the less highly maintained areas of a golf course may also aid in our understanding of the factors responsible for the distribution of VCGs present in different populations. The use of molecular markers for the study of fungal plant pathogen populations is well documented (2, 5, 10, 13, 14, 15, 17, 19). Recently, these techniques have been used with greater frequency for examinations of turfgrass 17 pathogens. RAPDs were used to examine genetic variation present in a collection of 26 isolates from the northeastern and midwestem areas of the US (19). Raina et al. found a very high level of genetic similarity between isolates regardless of location, indicating a strong clonal population structure. However, a limitation of their study was the small number of isolates from a single location, making inferences about population structure difficult. Hsiang and Mahuku (10) also used RAPDs to assess variation in S. homoeocarpa populations in Southern Ontario. They sampled populations of S. homoeocarpa more intensely than Raina et al. (19), collecting over 20 isolates per population. They found that 5 of the 8 populations exhibited significant linkage disequilibrium Indicating a clonal population structure. The remaining 3 populations had linkage disequilibria consistent with a random mating system. In the populations that they studied, they did not perform any VCG comparisons to corroborate their results of random mating. This could have provided crucial information about the disease cycle of Sclerotinia homoeocarpa. Most of the genetic variation was found between populations and very little variation was found within populations. Corroborating the findings of Raina et al. (19), our results support a clonal population structure in the S. homoeocarpa populations sampled because of the low amount of genetic diversity present. AF LP fingerprints were not able to resolve isolates based on VCG or geographic location. An isolate from Grand Rapids (Alpine 60) had the same AFLP fingerprint as an isolate from one of the Detroit locations (Maple Lane). Also, isolates from Maple Lane were present in all of the major branches of the 18 tree. This indicates a significant amount of the genetic variation observed in this study is present within a population. The lack of a pattern between AFLP genotypes and independent measures such as geographic location and VCG is interesting because these results point to a fairly recent introduction of the pathogen into Michigan. The construction and development of golf courses in Michigan is an activity that has taken place over the last century and so the introduction of the pathogen on golf course turf presumably would have occurred at some point over this period. Another possibility that could explain these data is regular migration between populations so that there is no differentiation of the populations. Migration seems to be a less likely scenario because of the large distances between the populations sampled and the lack of any evidence for a spore forming stage that could be aerially disseminated. One other possibility is that selection could be a factor involved in the lack of diversity present at the sampled populations. Kohn et al. (15) suggested that diversifying selection (26) was an important driver of diversity because it predicts that a mosaic of pathogen genotypes that are specialized for differing conditions are favored in the absence of other selective factors. Certainly the presence of diverse microclimates, different management practices, and cultivar selections on today’s golf courses would provide a similar disturbed environment compared to the environments discussed by Kohn et al.(15). This type of selection would also fit well with the data generated by Powell and Vargas (18) who found that the VCGs at a location change over time and hypothesized that the change could be the result of environmental conditions. 19 Future research should use both molecular as well as VCG characters to test the hypotheses generated by this research. How does dollar spot first appear on a golf course? This question is important to understanding and elucidating the population structure of this pathogen. The question could be approached by monitoring a population over time on a newly established golf course using the techniques applied in this study. Also, research determining the mode and survival of overwintering inoculum of S. homoeocarpa would also help to shed light on the recalcitrant population structure of this pathogen. 20 Literature Cited 1. Baldwin, NA. and Newell, A.J. 1992. Field production of fertile apothecia by Sclerotinia homeocarpa in Festuca turf. J. Sports Turf Res. Inst. 68: 73-76. Baymen, P., and Cotty, P.J. 1991. Vegetative compatibility and genetic diversity in the Aspergillis flavus population of a single field. Can. J. Bot. 69: 1707-1711. Bennett, F .T. 1937. Dollar spot of turf and its causal organism Sclerotinia homoeocarpa n. sp. Ann. Appl. Biol. 24: 236-257. Boesewinkel, H.J. 1976. Storage of fungal cultures in water. Trans. Br. Mycol. Soc. 66: 183-185. Cilliers, A.J., Herselman, L., Pretorius, Z.A. 2000. Genetic Variability Within and Among Mycelial compatibility Groups of Sclerotium rolfsii in South Africa. Phytopathology 90: 1026-1031. Detweiler, A.R., Vargas, J.M., Jr., and Danneberger, T.K. 1983. Resistance of Sclerotinia homoeocarpa to iprodione and benomyl. Plant Dis. 67: 627- 630. Fensterrnacher, J.M. 1980. Certain features of dollar spot disease and its causal organism Sclerotinia homoeocarpa. In: Advances in Turfgrass Pathology. Eds. P.O. Larsen and BC. Joyner, Harcourt, Brace, Jovanovich, Duluth, MN, pp. 49-53. Glass, ML, and Kuldau, GA. 1992. Mating type and vegetative incompatibility in filamentous ascomycetes. Annu. Rev. Phytopathol. 30: 201-224. Golembiewski, R.C., Vargas, J.M.,Jr., Jones, AL, and Detweiler, AR. 1995. Detection of demethylation inhibitor (DMI) resistance in Sclerotinia homoeocarpa populations. Plant Dis. 79: 491-493. 10. Hsiang, T., and Mahuku, GS. 1998. Genetic variation within and between southern Ontario populations of Sclerotinia homeocarpa. Plant Pathology 48: 83-94. 11.Jackson, N. 1973. Apothecial production in Sclerotinia homeocarpa F .T. Bennett. J. Sports Turf Res. Inst. 49: 58-63. 21 12.Jones, C.J., Edwards K.J., Castaglione, S., Winfield, M.O., Sala, F., van de Wiel, C., Bredemeijer, G., Vosman, B., Matthes, M., Daly, A., Brettschneider, R., Bettini, P., Buiatti, M., Maestri, E., Malcevschi, A., Marmiroli, N., Aert, R., Volckaert, G., Rueda, J., Linacero, R., Vazquez, A., and Karp, A. 1997. Reproducibility testing of RAPD, ALFP, and SSR markers in plants by a network of European laboratories. Molecular Breeding 3: 381-390. 13. Kohli, Y., Brunner, L.J., Yoell, H., Milgroom, M.G., Anderson, J.B., Morrall, R.A.A., and Kohn, L.M. 1995. Clonal dispersal and spatial mixing in populations of the plant pathogenic fungus, Sclerotinia sclerotiorum. Molecular Ecology 4: 69-77. 14.Kohli, Y., Morrall, R.A.A., Anderson, J.B., and Kohn, L.M. 1992. Local and trans-Canadian clonal distribution of Sclerotinia sclerotiorum on canola. Phytopathology 82: 875-880. 15. Kohn, L.M., Stasovski, E., Carbone, I., Royer, J., and Anderson J.B. 1991. Mycelial incompatability and molecular markers identify genetic variability in field populations of Sclerotinia sclerotiorum. Phytopathology 81: 480- 485. 16. Leslie, J.F. 1993. Fungal vegetative compatibility. Annu. Rev. Phytopathol. 31: 127-150. 17. Majer, D., Mithen, R., Lewis, B.G., Vos, R, and Oliver, RP. 1996. The use of AFLP fingerprinting for the detection of genetic variation in fungi. Mycol. Res. 100: 1107-1111. 18. Powell, J.F., and Vargas, J.M., Jr. 2001. Vegetative compatibility and seasonal variation among isolates of Sclerotinia homoeocarpa. Plant Dis. 85: 377-381. 19. Raina, K., Jackson, N., and Chandlee, J.M. 1997. Detection of genetic variation in Sclerotinia homoeocarpa isolates using RAPD analysis. Mycol. Res. 101: 585-590. 20. Smith, JD, Jackson, N., and Woolhouse, AR. 1989. Fungal Diseases of Amenity Turf Grasses. E. and F. Spon, New York. 21.Smiley, R.W. 1992. Compedium of Turfgrass Diseases 2"" Edition. American Phytopathology Society Press, St. Paul, MN. 22. Sonoda, RM. 1988. Vegetative compatibility groups among Sclerotinia homoeocarpa from leaves of Paspalum notatum. Proc. Soil Crop Sci. Soc. Fla. 48: 35-36. 22 23.Vargas, J.M., Jr., 1994. Management of Turfgrass Diseases. Lewis Publishers, Ann Arbor, MI. 24.Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., Homes, M., Frijters, A., Pot, J., Peleman, J., Kuiper, M., and Zabeau, M. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Reserch 23: 4407- 4414. 25.Warren, C.G., Sanders, P., and Cole, H. 1974. Sclerotinia homoeocarpa tolerance to benzimidazole configuration fungicides. Phytopathology 64: 1 139-1 142. 26.Worrall, J.J. 1999. Structure and Dynamics of Fungal Populations. Kluwer Academic Publishers, Dordrecht, The Netherlands. 23 Chapter 2 GEOSTATISTICAL ANALYSIS OF DOLLAR SPOT EPIDEMICS IN MICHIGAN Introductlon Dollar spot disease of turfgrasses is caused by the pathogen, Sclerotinia homoeocarpa F.T. Bennett (2). The disease is commBn throughout the world and is destructive to both cool and warm season grasses (18, 19, 21). In North America, with the exception of the Pacific Northwest, dollar spot is the most important pathogen of most cultivated fine turfgrass species (18, 21). In Michigan, the disease is a major problem for most golf courses where the epidemic begins in June and can continue into late September. The disease can blight large areas of turf as a result of coalescing disease foci and can cause extensive damage if left untreated. Diseased turf impairs the playing surface by creating depressions that affect ball roll and areas of bare soil where weed species can encroach on the area (18, 21). The biology of dollar spot has not been studied extensively due to the relative ease with which this disease can be controlled with fungicides. Outside of Great Britain, dollar spot has not been reported to form sexual or asexual spores (1, 13, 18, 19). Hsiang and Mahuku (11) reported that a population from Ontario exhibited DNA fingerprints that were consistent with sexual reproduction, but no fruiting bodies or spores were found. Nitrogen fertility is an important factor in the management of dollar spot (4, 14, 18, 19, 21, 25). Most reports support the view that increased fertility leads to a reduction in the severity of the disease (14, 18, 19, 21, 25). However, Couch and Bloom (4) reported that the susceptibility of 24 Poa pratensis was actually increased in higher fertility treatments in the greenhouse, but that the effect would be masked in the field because symptoms would not appear before the rapidly growing, infected leaf blades were mown off. Spread of the disease is widely believed to be the result of direct movement of the pathogen on infected blades during mowing operations, and human transport of infected clippings on shoes, balls, etc. (7, 18, 19, 21, 25). Little or no research exists on inoculum sources, but some reports contend that stromatized infected tissue is the primary inoculum source for the pathogen (7, 25). Control of dollar spot via fungicide application is generally accomplished using the contact fungicide, chlorothalonil. However, the EPA has recently banned the use of this fungicide on home lawns, and future restrictions on commercial use of chlorothalonil is expected. Because the amount of fungicide available to an individual golf course is expected to be limited, it will become critical for superintendents to be able to apply chlorothalonil judiciously in areas that require treatment. Other fungicides are available for control, but fungicide resistant dollar spot populations have been reported for most of them including, the demethylation inhibiting (DMI) (8), benzimidazole (5, 23), and dicarboximide (5) fungicides. Studying the epidemiology of a plant disease traditionally calls for the assessment of the severity or occurrence of disease over some area of interest. These data are generally collected over an area using some sampling scheme. However, traditional statistical techniques require adherence to assumptions such as the independence of samples and normality of data. It makes intuitive 25 sense that plants located closer in space to a diseased plant have a higher probability of becoming infected than plants located farther away. Recently, using the location of these samples in space as an additional datum has led to the description of disease epidemics over space and time using geostatistical techniques (20, 24, 26, 27). Important epidemiological questions arise from spatially explicit descriptions of disease: Is the disease appearance clustered in space? If so, what factors contribute to this clustering? What practices might ameliorate the effects of those factors? Can we predict both when and where disease is going to occur so that fungicide applications can be targeted? One method available to address these questions is geostatistics. Originally developed for the study of geological phenomena, geostatistics have found wide application in a number of fields including phytopathology (20, 24, 26, 27). The primary goal of these techniques is to explain how a variate of interest (e.g. disease severity) at a location in space is correlated with all the other points where the van'ate has been measured (9, 10, 12, 15, 17). Further treatments allow for the prediction of the variate at unmeasured locations, and the assessment of prediction confidence. These techniques also allow for the analysis of nominal values such as genotypes or size classes in a similar manner (9, 10). Geostatistics provides a powerful set of tools that can yield insight into the dynamic nature of plant disease. Research in plant disease epidemiology using geostatistical tools is relatively new. Geostatistics have been used to study the spatial pattern of 26 disease incidence and severity (20, 26) and inoculum levels (24, 27). These tools have also been used to study physico-chemical properties of soils (10), plant distributions and ecology (15, 17), and microbial distributions (20, 26, 27). Variography was used by Wollum and Cassel (26) to study the spatial variability of Rhizobium japonicum in soil planted with soybeans. They concluded that geostatistics were a good tool for studying the dynamics of microbial populations. Stein et. al. (20) examined the spatio-temporal development of Peronospora parasitica epidemics in cabbage (Brassica oleracea). They were able to determine that spatial variability of the fields was dependent on disease incidence. They also found spatial dependence when fields were recovering from disease. Xiao et. al. (27) studied the spatial patterns of Verticillium dahliae microsclerotia in the soil, and verticillium wilt in cauliflower using geostatistics. They found that the spatial structure of microsclerotia in the soil was not very strong. The structure of the microsclerotia did not play a role in the clustered appearance of disease. They attributed this to a very high amount and fairly uniform distribution of microsclerotia. They concluded other factors affect the appearance of wilt. However, they did find that the severity of the wilt was associated positively with the weak spatial pattern for the presence of microsclerotia. The information generated by a geostatistical approach to the study of plant disease epidemics allows one to develop or refine management strategies, and help to determine the contributing factors that deserve further study. Ultimately, this information can be used to design models that can be used 27 to predict the occurrence of plant disease, helping to minimize pesticide applications, or make cultural practices more effective. We had 3 objectives: 1) to observe dollar spot epidemics and determine if disease occurrence has a spatial structure, 2) if a spatial structure is present, determine the geostatistical parameters associated with the spatial structuring, 3) Determine if the spatial structure changes during an epidemic, or over seasons. Materials and Methods Sampling. The study site was established at the Robert Hancock Turfgrass Research Center on a 9.1 m X 18.3 m area of creeping bentgrass (Agrostis palustris Huds.) and annual bluegrass (Poa annua L.). The study site received no fungicide applications from 2000-2002. The study site was divided into 200 0.3 m2 areas on a regular grid at 0.9 m intervals in 2000. In 2001, 23 additional 0.30 m2 areas were established at random locations within the study site, and all 223 areas were subdivided into four 0.15 m subareas (Fig. 1). These additional locations were added and all areas subdivided to increase the number of data pairs at small lag distances. Each subarea’s x,y coordinates were recorded using its center. A 0.3 m2 wooden frame that was divided into quadrants was used to delineate the subareas. Two points at each sampling location were marked with marking paint in order to place the frame at the same location at each sample time. Isolates were also collected from an arbitrarily selected dollar spot at each location in July 2000 and the vegetative compatibility group (VCG) for each isolate was determined in order to assess if clustering was present in VCGs (Appendix 1). 28 18.3 m 9.1m 33’ . l3\ .00 \: 0.3m 0.15m Figure 1. Schematic layout of study area at the Robert Hancock Turfgrass Research Center (E. Lansing, MI) showing overall arrangement of sampling locations. 29 Data Collection. Dollar spot foci were counted three times per week in 2000 and twice per week in 2001 and 2002 from each of the locations in the study area. Foci were counted when they reached a size large enough to observe. This helped to minimize the possibility of accidentally counting an area that was not a true dollar spot. Counting of the study area was stopped in each year when disease became prevalent enough in a location that there were too many disease fool to count accurately. Geostatistical analysis. Spatial continuity was measured using the variogram, that describes how a variate changes over space. Intrinsic to geostatistical analysis is the hypothesis that the expectation of differences between any pair of points depends solely on the distance (h) between the points (6, 9, 12). Therefore, to estimate the variogram from disease incidence data with the intrinsic hypothesis in mind; 1 2 If (h) — “2N(h) (1%? - V!) where y is the semivariance, h is the average separation distance between pairs of points, N is the number of data pairs, and v| and VI are the ith and jth data values at separation distance h. This results in estimates of semivariance (7) that are plotted as a function of separation distance (h). Once a variogram is calculated, a model that fits the observed data can be fit to the experimental data. A variogram model defines three key parameters: nugget, sill, and range. The nugget is the result of the discontinuity that occurs when the semivariance, which is defined to be 0 when lag distance, h=0, jumps to some value > 0 at a 30 very small distance away. The nugget is the result of a combination of sources of variation including experimental error and short-scale variability (9, 12). Generally, as the distance between pairs of points increases, the semivariance will also increase. Therefore as pairs of points are separated by larger distances they become less correlated. However, at some separation distance between pairs the semivariance will reach a plateau where increases in separation distance do not result in a change in the semivariance. This distance where semivariance reaches a plateau is the range, and is also the boundary between spatially dependent and spatially independent variation. The third parameter calculated for a variogram model is the sill. It is defined as the semivariance value reached at the range. If the sampling design has accounted for most of the variation in the system, then the sill value is often very similar to the overall sample variance (5’). The total variance can be ascribed to three catagories: nugget (Co), structural or spatial variance (C), and sill variance (Co+C). The proportion of the total variance that is accounted for by structural variance, or the proportion of structural variance, can be calculated by, C/Co+C, and is often expressed as a percentage of the total variance that is spatially structured. When this value approaches 1, a large proportion of the total sample variance is spatially dependent. When the value approaches 0, spatial dependence is low. If the sill value (Co+C) is not similar to the total sample variance (3’), this indicates that there may be further structure at scales larger than those sampled. Variograms were calculated using the windows interface WinGSLIB (Statics, LLC., San Francisco, CA) for the geostatistical software package, 31 GSLIB (version 9) (6). The choice of the number of lags, lag interval, and lag tolerance were defined iteratively based on the number, interval and tolerance for lags that yielded a smooth, well-behaved variogram. Ultimately, 15 lags with a 0.61 m lag interval, and 0.30 m lag tolerance were the parameters that gave the most well-behaved result for all three years. The experimental variograms were then modeled using the geostatistical software package, GS+ version 3.0 (Gamma Design Software, Plainwell, MI). The GS+ software package allows one to automatically fit models to the experimental variogram, and then chooses the best fit based on the model with the smallest unweighted least squares value. Any additional changes that were necessary were made by hand to the initial fit provided by the program. The experimental variograms used for modeling were generated in the 68+ program using a 12.2 m lag distance and 1.5 m lag interval as parameters for 2000, and a 6.1 m lag distance and .61 m lag interval for 2001 and 2002. These distances were chosen to avoid over-fitting the models to the increasing semivariance values at larger lag distances seen in the experimental variograms shown in Figure 3. Results Sampling and Data Collection. Disease was observed and the number of dollar spot foci counted at each location until August 25th in 2000, September 9'“ in 2001, and September 13th in 2002. Total disease progress curves for each of the three years appear in Figure 2. The epidemic in 2002 was the most severe followed closely by the epidemic in 2000. The epidemic in 2001 was much less severe than either 2000 or 2002. The total disease progress curves for all three 32 years were similar in shape. Each progress curve had an early season outbreak that was not as severe as the late season outbreak that began in early August and continued into September. Geostatistics! analysis. Variograms were calculated for each date disease counts were taken in 2000-2002. Nine variograms, each representing a variogram from the early, middle, and late phases of the epidemic for each year are shown in Figure 3. The remaining variograms from other sample dates are presented in Appendix 2. Anisotropy was examined at several dates and no anisotropic trends were apparent (Data not shown). The variograms in all three years show clear spatial structuring that occurs at smaller lag distances (<09 m), particularly as disease incidence increases over time. Throughout 2001 and 2002 the model that best fit the data was an exponential model (Table 1). In 2000, the first three dates showed a nugget effect indicating no spatial structuring, and then for the remainder of the season both spherical and exponential models were defined. The proportion of structural variance (C/Co+C) was about 0.5 for 2000 and about 0.6 for both 2001 and 2002 (Table 2). This value means that about 50% of the total variation in 2000 and about 60% of the total variation in 2001 and 2002 is spatially structured. The range parameter that was calculated for each variogram model in 2000 had larger values and a wider range of variation as compared to the smaller and less variable range parameters calculated in 2001 and 2002. 33 65:80. 99:3 __w c. .08 0836 .99 Fo $5.8 E 8530:. mm Nooméoom E9: 3285 ommomfi .32 actmano FEED -N 2:9“. cunn 893 8m-.. 92-8 373 3?: SEN .57. PI F _ I o I. 8.: m w NSNIT SON m. 82+ a 82+ ooom m a O m. ooov /\. ooom magma...— came»:— 34 .o_Eoc_oo Foam .28 65 Fe 3me 92 new .ofiEE .25.... on. Eat 85288 $5QO 05539: Noowéoom E9.— mEEmotm> EcoEtoaxw .m 2:9... 38:838.. a o F o n F. n N F o 4 IJIOI‘I‘ICIT‘LIIF- 144 C 8.391 388.39.. a o F o m e n N F o hq .choEtoaxo cc... 9.. Lo. Ammm. moLmacm Fo Ezm _m:c.mo. new .ANFV E Fo mmocooom .395. .__.m .6835 EeFoEEma _oooE Em.mo..m> .3 2an 83... F8... Ne... F8... :3 azcmcoaxm 918 88 N48... .8... 5... 8.”... F3: @8890 .38 88 8-m.Nm.N 8F... mNN EN... FF... ascecoaxm mareN N8N FFFo... F5... F... 9.3 eFeN 3.2886 m=<.F.N F8N 8-m.mo.N N8... 8... Fem... mFNd azcmcoaxm. 5..-? 58 3mg... 8.... an... new... oFF... assocoem. €72 F8N QBF Sm... F08 2.2 a...” @2885. .33.... SON FNF... 8.... SN 89 SF 3.88%.”. 5..-: SON SN... Fm... NF.FN 3o... 8F 8:888“. .5..-NN SON mam N. .8. 8:3. 6+8. _..m .8. 88:2 2.3 .28: 38 a3 37 .828 88...? 6.5828 .6 8.885 new .__.m .882: 0526...... Nooméoom c. 28 959 some .2 EQoEEmQ 38E Em..mo..m> -N 28... 3.0 000! NF... 5.: an... 91.8 88 8... «FF 3.... 91.8 88 8... «SF 98 91.FF 88 Fm... 5.: 8... NF.NFF 8 91.3 88 8... 8...... NEF 8m... 88 8... 8.8 F8 3.... o8N 8... F8.» 85 Q84 58 No... BF. 8.: 91F 8 Fm... 88 BFN 91.8 F8N N8 F.8F 9F... 8.3. 88 8... =8: 8... 98... 9.3 918 F8 8... 8.8 a8 9.8 8 Fm... as... Fe... 88F 88 8... 9.3. vFvN 91.8 38 8... 88 No.9 3...; 88 8... F8... 3..., 88.9 88 8... 89F 8... 81.8 58 NF... 8N N3... .378 88 8... .8... F»... 8w... 88 8... F8... 3:... 91.: 58 8... BF... F8N 378 88 Fm... NF... 89 91.8 88 8... F8... 22. 91.? F8 8... 8m... 8F.F s78 88 8... Ba... 8... 91.8 88 8... £3 F8... 91... 58 on... F88 3... s78 88 3... n3... 8... 91.8 88 8... m5... F9... 91... 58 8... 8.8 88F .37.: 88 8... F8... F8.F 918 88 on... 83 2.8 9:8 58 8... F98 8NF 5...: 88 8... 98+ NVFF 91.2 88 8... 3...... N3... 57F” F8 8... :8 8... 373 88 9.... we? 83 91.9 88 8... N8... 82. 378 F8 8... 8.8 8.: stF 88 8... 83 NFF 91... 88 8... 8o... :8 .378 F8N NF... «NF 36 57°F 88 Fe... 889 83 91... N8 8... 8m... mFN... 33F F8N 85 8.9 o...” 32 88 Ne... 83 8F.N 928 88 8... N8... SN... 33F 88 N8 8»... NF? 33 88 8... n8... F8.F .378 88 N2. 8...... N8... .2..NF F8N 8F... o8... NS... .32. 88 S... F8N 2F .378 N8 8... N8... 8...... .33 58 N8 83 Fed .578 88 8... 83 8...... s78 88 OF... F2... 8.... 3...... F8N Fm... 23. NNFN .578 88 8... 38 8F... .372 N88 8... 88... 8o... EN 58 Fm... «8.4 8F .578 88 8... F3... 8F... .37.: 88 8... m8... 8o... F578 68 F2. 8o... 8o... .52: 88 8... Fe... Fe... 37NF N8 8... F9... 8:. .578 58 3... F8.F N8... 53F 88 9.... N... N8... .2... 88 8... 8...... N2... .578 F8N 8... :3 FFF... S7NF 88 FF... 82. FF... EN 88 8... 9...... 83 53F F8N 8... 83 8F... .52.. 88 FF... 8N... F... .578 88 8... N8... 8o... 53F 58 8... N8... N8... .52. 88 680.0. 8+8. .8. 8 =3 8.8.0. No.88 N8. 88 .8» .8.% 498 .8. 88 23> 8.5.; .__m .89.: 8.5.; ...m .89.: 85:; _..m .89.: 5338 .8325» .8325» 38 Discussion Although the total disease severity during the epidemics in each of the three years was different, there were also several similarities between the epidemics. First, the disease progress curves were similar in shape (Fig. 2). All three progress curves showed that dollar spot first appears in early to mid June with a minor outbreak and then becomes much more severe during the months of August and September. Our data agree with those of other researchers that have shown the later season phase of the epidemic is most damaging to a turf area (7, 16, 18, 21). Disease progress was similar in each year despite differences in disease severity indicating that environmental parameters play a large role in the overall severity and timing of dollar spot outbreaks. In all three years dollar spot was observed to decrease in presence during July, presumably because of the hot, humid growing conditions present during that time. These results support the view that environmental parameters are primarily responsible for disease appearance and resulting overall severity. The experimental variograms that were calculated for each date were also similar over time. Once a variogram is calculated for each date in the study, a model is calculated that fits the observed data. One reason to fit models to the data is so that the key model parameters the nugget, range, and sill, can be compared to observe how they change over time. The sampling design determines the smallest scale at which spatial relationships can be resolved. In 2000 the design consisted of 200 0.3 m2 areas spaced on 0.9 m centers. The limitations of this design is that the smallest lag interval for the calculated 39 variogram was 0.9 m, and there was no information about the disease at smaller scales. If spatial dependence exists at a scale smaller than the smallest sampled interval, then it would become a part of the nugget variance and would not be accounted for in the variogram. In 2000, the calculated variograms displayed spatial dependence, and about 50% of the total variation was spatially structured (Table 2). In 2001 and 2002 23 additional areas were added randomly to the study site, and the 0.3 m2 areas were subdivided into four 0.15 m2 subareas to protect against the possible problem of the scale of spatial dependence. Increasing the resolution of the sampling design increased the information about small-scale variability. As the smallest sampling interval was 0.15 m, the addition of the random locations was important to be able to evaluate lag distances between 0.15 m and 0.9 m. These changes in the sampling design resulted in a gain of information as reflected by a 10% increase in the proportion of structural variance from 50% in 2000 to 60% in 2001 and 2002 (Table 2). This increase in spatial resolution at the smaller scales is why there is a much stronger spatial dependence observed for the 2001 and 2002 data as compared to the 2000 data. Based on the experimental variograms calculated for all three years we conclude that dollar spot incidence is spatially correlated in our study area, and that the spatial correlation is present on a small scale. Other locations should be included in future studies to determine if dollar spot incidence at other locations is similarly spatially correlated. Interestingly, the nugget and sill values for variogram models from each date scale with one another indicating that the spatial structure is relatively stable 40 over time. The stability of this relationship can also be seen in the stability of the proportion of structural variance (C/Co+C) over time (Figure 4). The range parameter is also relatively stable further confirming a structure that is stable and relatively constant over time. While the range did fluctuate in 2000, the smallest lag interval was only 0.91 m, and these fluctuations could be a function of the lack of small-scale sampling. The higher resolution in 2001 and 2002 decreased the fluctuations in the range parameter where the overall change in the range from low to high in both years was between two and three meters. Exponential variogram models were defined in all three years. These results clearly show that as disease intensity increases over a season, the spatial structure that is present is stable and doesn’t change much over time. One possible explanation for the observed spatial structure is that areas with more disease increase at the same relative amount as areas with less disease. If this were not the case, then one would expect the spatial structure, as measured by the proportion of structural variance (C/Co+C), to change as disease increased over the season. However, the structure that was observed remained stable over the season. The distance between areas with similar disease intensities (i.e. range) also remains relatively stable over time indicating that these areas are not shifting within an epidemic or among epidemics. Because one would expect different locations to behave differently as a result of either micro- or macroclimatic changes that occur over an epidemic, these results support the view that environmental parameters are not a major factor in the spatial structuring at the scales observed. 41 The literature provides much speculation on the mode of spread for this non-spore forming pathogen (7, 11, 18, 25). These reports range from the movement of mycelial fragments on diseased tissue via human and mechanical transport (7, 18, 25) to the production of an undiscovered spore that is produced (11). Data from this study disagree with both of these possibilities. If the pathogen were transferred via mechanical means, then one would expect the spatial structure of disease incidence to change over time because mycelial fragments would be distributed over the area via regular, uniform mowing practices. If the pathogen was transferred via human means, then the spatial structure should be indicative of a pattern similar to a pattern of movement over the area by people. If this pathogen produced some unknown spore, then one would expect that the dispersal of such a spore would occur such that the spatial structure of the disease would change with the release of spores. However, none of these possible outcomes were observed in this study. Rather, our results indicate that the primary factor governing the spatial structure is one that doesn’t move in space and whose spatial structure is relatively constant regardless of the intensity of disease. One hypothesis that would fit these data is that the host and/or pathogen are important in the spatial structuring. The predominant grasses found on golf course putting surfaces are creeping bentgrass (A. palustn‘s) and annual bluegrass (P. annua). Both of these grasses are non-uniform in their susceptibilities to S. homoeocarpa (3, 22). The breeding strategy employed for creeping bentgrass results in the production of a synthetic cultivar, meaning that 42 each seed is genetically distinct. This results in a range of variation in susceptibility/resistance to dollar spot. Annual bluegrass is a non-cultivated grass that invades putting surfaces as a weed, and also is known to be genetically variable (21 ). The area we studied was at least 10 years of age and was a mixed sward of creeping bentgrass and annual bluegrass. Over time the competitiveness of each seedling would govern those genotypes of grasses found in a site. These successful genotypes would then be more or less susceptible to dollar spot, and this would be observed as a mosaic of disease incidence with a spatial structure corresponding to the spatial structure of the grasses. This hypothesis would also predict that the inoculum density of the pathogen would also follow this spatial structure because areas with previous higher disease incidence would produce more infested tissue, which is believed to be the primary inoculum source for dollar spot. Overall, these data support the view that there is a relatively stable spatial structure governing disease incidence that is unaffected by disease severity. Furthermore, the results support a theoretical model that the host and pathogen are involved in the observed spatial structure over the scales assessed by this study, and that environmental parameters appear to be most important in overall disease severity and timing of disease outbreaks. Future research in this area should include the evaluation of other locations to determine if the observed spatial structure is ubiquitous, and the testing of the theoretical models posed by this research to confirm or exclude factors associated with the spatial structuring of dollar spot incidence. These 43 research areas would provide the information that is needed to begin developing predictive models that can predict the incidence and location of dollar spot based on a knowledge of the environmental and geospatial parameters that govern where and when dollar spot occurs. Once predictive models become available it would then be possible to implement precision fungicide applications for the control of this disease. Literature Cited 1. Baldwin, NA. and Newell, A.J. 1992. Field production of fertile apothecia by Sclerotinia homeocarpa in Festuca turf. J. Sports Turf Res. lnst. 68: 73-76. Bennett, F.T. 1937. Dollar spot of turf and its causal organism Sclerotinia homoeocarpa n. sp. Ann. Appl. Biol. 24: 236-257. Cole, H., Duich, J.M., Massie, LB, and Barber, WD. 1969. Influence of fungus isolate and grass variety on Sclerotinia dollar spot development. Crop Science 9: 567-570. Couch, H.B., and Bloom, JR. 1960. Influence of turfgrasses. II. Effect of nutrition, pH, and soil moisture on Sclerotinia dollar spot. Phytopathology 50: 761-763. Detweiler, A.R., Vargas, J.M., Jr., and Danneberger, T.K. 1983. Resistance of Sclerotinia homoeocarpa to iprodione and benomyl. Plant Dis. 67: 627-630. Deutsch C.V., and Journel, AG. 1998. GSLIB: Geostatistical Software Library and User’s Guide 2"d Edition. Oxford University Press, New York. Fensterrnacher, J.M. 1980. Certain features of dollar spot disease and its causal organism Sclerotinia homoeocarpa. In: Advances in Turfgrass Pathology. Eds. P.O. Larsen and 8.6. Joyner, Harcourt, Brace, Jovanovich, Duluth, MN, pp. 49-53. Golembiewski, R.C., Vargas, J.M.,Jr., Jones, A.L., and Detweiler, AR. 1995. Detection of demethylation inhibitor (DMI) resistance in Sclerotinia homoeocarpa populations. Plant Dis. 79: 491-493. Goovaerts, P. 1997. Geostatistics for Natural Resources Evaluation. Oxford University Press, New York. 10. Goovaerts, P. 1998. Geostatistical tools for characterizing the spatial variability of microbiological and physico-chemical soil properties. Biol. Fertil. Soils 27, 315-334. 11.Hsiang, T., and Mahuku, GS. 1998. Genetic variation within and between southern Ontario populations of Sclerotinia homeocarpa. Plant Pathology 48: 83-94. 12. Isaaks, E.H., and Srivastava, RM. 1989. An Introduction to Applied Geostatistics. Oxford University Press, New York. 13.Jackson, N. 1973. Apothecial production in Sclerotinia homeocarpa F.T. Bennett. J. Sports Turf Res. Inst. 49: 58-63. 45 14. Markland, R.E., Roberts, E.C., and Frederick, LR. 1969. Influence of nitrogen fertilizers on Washington creeping bentgrass, Agrostis palustris Huds. ll. Incidence of dollar spot, Sclerotinia homoeocarpa infection. Agron. J. 61: 701 - 705. 15. Oliver, MA. and Webster, R. 1986. Combining nested and linear sampling for determining the scale and form of spatial variation of regionalized variables. Geographical Analysis 18: 227-242. 16. Powell, J.F., and Vargas, J.M., Jr. 2001. Vegetative compatibility and seasonal variation among isolates of Sclerotinia homoeocarpa. Plant Dis. 85: 377-381. 17. Rossi, R.E., Mulla, D. J., Joumel, A. G., and Franz, E. H. 1992. Geostatistical tools for modeling and interpreting ecological spatial dependence. Ecological Monographs 62: 279-314. 18. Smith, JD, Jackson, N., and Woolhouse, AR. 1989. Fungal Diseases of Amenity Turf Grasses. E. and F. Spon, New York. 19. Smiley, R.W. 1992. Compedium of Turfgrass Diseases 2"" Edition. American Phytopathology Society Press, St. Paul, MN. 20.Stein, A., Kocks, C.G., Zadoks, J.C., Frinking, H.D., Ruissen, MA, and Myers DE. 1994. A geostatical analysis of the spatio-temporal development of downy mildew epidemics in cabbage. Phytopathology 84: 1227-1239. 21 .Vargas, J.M., Jr., 1994. Management of Turfgrass Diseases. Lewis Publishers, Ann Arbor, MI. 22.Vincelli, P., and Doney J.C.,Jr. 1997. Variation among creeping bentgrass cultivars in recovery from epidemics of dollar spot. Plant Dis. 81: 99-102. 23.Warren, C.G., Sanders, P., and Cole, H. 1974. Sclerotinia homoeocarpa tolerance to benzimidazole configuration fungicides. Phytopathology 64: 1139-1142. 24.Webster, R., and Boag B. 1992. Geostatistical analysis of cyst nematodes in soil. Journal of Social Science 43: 583-595. 25.Williams, D.W., Powell, A.J., Jr., Vincelli, P., and Daugherty, CT. 1996. Dollar spot on bentgrass influenced by displacement of leaf surface moisture, nitrogen, and clipping removal. Crop Sci. 36: 1304-1309. 26. Wollum, A.G.,Il, and Cassel, OK. 1984. Spatial variability of Rhizobium japonicum in two North Carolina soils. Soil Science Soc. Am. J. 48, 1082- 1086. 46 27.Xiao, C. L., J. J. Hao, and K. V. Subbarao. 1997. Spatial patterns of microsclerotinia of Verticillium dahliae in soil and verticillium wilt of cauliflower. Phytopathology 87 (3), 325-331. 47 APPENDICES 48 APPENDIX 1 RAW INDICATOR SEMIVARIOGRAMS OF ISOLATE VCGS 80 m iv. rie nee VCG A 0.3 0.25 0.2 0.15 0.1 0.05 Continuance 0.35 9 .° n . N at 0.15 0.1 0.05 V66 3 Leg Distance (in) o o o 10 20 30 40 so 0 10 20 30 40 50 L09 Distance (1!) La. Distance (to VCG D VCG E 0.04 0.02 0.035 0 3 0.03 8 0.015 g 0.025 E . . g 0.02 E 0.01 i 0.015 I . O . 9 ' 0.01 ° 0.005 ° ° a C o 0.005 0 0 0 o 10 20 30 40 so 0 10 20 30 4o 50 Leg om.- ncc (1'!) Ion iverienco VCG P 0.014 0.012 0.01 0.000 0.006 0.004 0.002 o 10 20 30 40 50 Leg Distance (in 49 APPENDIX 2 RAW SEMIVARIOGRAMS FOR ALL DATES FROM 2000-2002 Juno 0. 2000 June 9, 2000 0.1 0 0.2 0 0 °. 0 0 . o o o 0.08 o g . . g 0.15 O . £006 2 ........... h- . I 0 1 3 0.04 .2 f e a 0.02 g 0.05 o o 10 20 30 40 50 o 10 20 30 40 so he Old-nu (h) Leg counc- (II) June 12. 2000 June15, 2000 0.14 0.12 a a c 0“ a 8 0.00 E C I .2 0.00 z 5 0.04 E ' 0.02 o 10 20 30 4o 50 0 1o 20 30 40 50 Leg Diane. (11) Leg Dlmnco (in) June 19. 2000 June 22. 2000 8 3 e c I I '1': t I I z z: E E I O I I 10 20 30 40 so 0 10 20 30 4o 50 Log Distance (11) Leg Distance (in) June 20, 2000 June 20, 2000 g 8 7 s 8 § 6 5 ‘ .3 5 E 3 I 4 z z 3 E s 2 ,-, 2 C 1 1 o 0 o 10 20 30 40 so 0 10 20 30 40 50 Log Distance (h) Lag Distance (to) 50 Cornlvsrlence July 3,2000 20 30 40 50 Leg Dial-nee (h) Continuance 0. July 5. 2000 20 30 40 50 Leg Distance (to) July 7, 2000 July 10. 2000 I I 0 0 c c I I '1': t a a 2 2 E E I I a a 0 10 20 30 40 50 0 10 20 30 40 50 Leg Distance (to) Log Dist-nee (in) July 12. 2000 July 14. 2000 2 I = E 2 .2 I E 8 8 0 10 20 30 ‘0 50 0 1o 20 30 40 50 Leg Distance (11) Le. Did-no. (In) July 17. 2000 July 10. 2000 o o 0 I e c .3 E o a .2 .2 i E o o 00 a 20 30 40 50 Leg Distance (11) 20 30 4o 50 Leg Dish use (11) 51 July 21. 2000 July 24. 2000 8 8 I I I I ‘1: 1: 5 E E E I I o a 0 10 20 30 40 50 10 20 30 40 50 Lsg Dldsncs (11) L09 DIstsncs (hi July 20. 2000 July 28. 2000 I I 8 . o 0 ° 2 I I t t 5 3 E E .1 .1 o 10 20 30 40 50 10 20 30 40 so Lsg Old-nee (h) Lsg Olslsncs (In July 31,2000 Augus12.2000 I I 0 I I I 1! 2 I- b I I 2 .2 E E I I a 00 0 10 20 30 40 50 10 20 30 40 so Ls. Distsncs (1:) Ln. Distance (11) August 4. 2000 August 7, 2000 140 120 s 8 g 100 g 12 00 E I I .2 60 3 E 40 5 a a 20 0 0 10 20 30 4o 50 L09 Distsncs (h) 20 30 40 so Ls. Dislsncs (h) 52 August 0. 2000 August 14. 2000 I I I I s s t a I I 2 2 E I I I 00 a 0 10 20 30 40 50 10 20 30 40 50 Leg Distsnss (h) Lsg Distsncs (11) June 15. 2001 June 19. 2001 0.35 0.3 3 a c 0.25 c 5 0.2 g I I 2 0.15 2 5 0.1 5 0.05 0 10 20 30 40 50 10 20 30 40 50 Lsg Distsncs (h) Lsg Distsncs (h) June 22. 2001 June 25. 2001 0.7 s 8 0.0 3 . 0.5 g 13 0.4 2 2 0.3 E E 11.2 0.1 0 10 20 30 40 50 10 20 30 40 50 Lsg Distsnss (ti) Lsg Dlmncs (ll) Juns 20. 2001 July 2. 2001 0.4 0.35 3 0.3 1" 8 0 25 g a: 0 2 ‘3 5 ’ 2 5 0.15 s . 0.1 “ 0.05 0 20 30 Lsg Dlsts ncs (I1) 40 50 20 30 40 50 Leg DIstsncs (in) 53 July 5. 2001 July 0,2001 e e e o c c e e t t e e 2 2 E E e e a a 10 20 30 40 50 10 20 30 40 50 Leg Distance (in) Lag Distance (11) July 12,2001 July 10,2001 3 a : : t t S S E E e e a a 10 20 30 40 50 10 20 30 40 50 Leg Distance (11) Leg Diuance (11) July 10.2001 July 23,2001 e e 8 8 I I t t S S E E e e a u 10 20 30 40 50 10 20 30 40 50 Leg Distance (it) Leg Distance (11) July 20.2001 July 31,2001 8 8 c c e e t t 3. 5 5 E a 8 20 30 Leg Distance (h) 40 50 20 30 40 50 Lag Distance (11) 54 August 6.2001 August!.2001 1.4 1.2 O O 3‘ 8 ..°oee°eeeOO I I t 0.0 t C C .2 0.6 2 5 0.. E ' 0.2 " 0 0 10 20 30 40 50 10 20 30 40 50 Leg Distence (it) Leg Distance (in) August13.2001 August17.2001 8 8 8 c 2 .2 b 5 E 3 E E O O a I 0 1O 20 30 40 50 1o 20 30 40 so Leg Distance (it) Leg Distsnce (it) August 20.2001 August 24,2001 7 8 ° 3 2 c 5 a a 4 I I .2 2 3 5 5 2 o a 1 0 0 10 20 30 40 50 0 10 20 30 40 50 Leg Diasnce (ii) Leg Distsnce (In) August 27,2001 August 30.2001 0 12 7 10 a 5 5 ° e a s 0 .2 2 E 3 c 4 e 2 e n ' 2 1 0 0 0 10 20 30 ‘0 50 Leg Distance (h) 20 30 Leg Dlstsnce (h) 40 50 55 September 4, 2001 September 0. 2001 7 6 3 ° 3 5 s 5 = 4 t 4 ‘3 s e 3 2 3 2 5 2 5 2 a a 1 1 O 0 0 10 20 30 40 50 0 10 20 30 40 50 Leg Distance (it) Leg Distance (h) June 24. 2002 July 2. 2002 O O 0 II c c 2 2 h. h I I 2 2 E E O O a a 0 10 20 30 40 50 0 10 20 30 40 50 Leg Distance (ll) Lag Distance (in) July 3. 2002 July 12. 2002 0.25 e 02 e g 2 .5 0.15 ‘2 I 3 0.1 2 5 E a 0.05 3 0 0 10 20 30 40 50 0 10 20 30 ‘0 50 Lag Distance (in) Lag Distance (In) July 10.2002 July 10,2002 0.5 e 0 0 0 e OJ . e 0 0 = s g 0.3 'c = : - 0.2 - E E O O a 0.1 o 0 30 40 50 20 Lag Dlstence (h) 30 40 50 20 Leg Distance (h) 56 July 23. 2002 July 25. zoo: 10 20 30 Leg Didance (h) 40 50 2.5 3 3 2 3 2.5 c o e e e e 0 ° ° 0 g 1.5 ’ . O . ° 5 2 b s 1 s 1.5 s 05 s ‘ . “ 0.5 o o 10 20 30 40 50 0 10 20 30 ‘0 50 Leg Didsnce (it) Leg Distance (11) July 30. 2002 August 2. 2002 7 o 0 e . 5 2 5 3 . I E ‘ 3 z 3 5 3 1 " 1 0 0 o 10 20 30 40 50 o 10 20 so 40 so Leg Distance 01) Leg Distance (M August 0. 2002 August 0. 2002 0 0 e 5 . 5 a . 3 1 3 3 5 3 E 5 E 2 E 2 O I so 1 ' 1 0 0 10 20 30 40 50 0 10 20 30 40 50 Leg Dldence (i1) Lag Distance 01) August 13. 2002 August 10. 2002 0 0 e 5 . 5 g e°°eeOeeO°° OeeeeO° g s ‘ . 4 t 3 t 3 5 2 E 2 E 2 I I e o 0 10 20 30 40 50 Leg Dluence (h) 57 August 20. 2002 August 23. 2002 0 7 0 e u 0 u 3 s 5 " t 3 ‘ = 5 3 . a 2 3. 1 0 0 10 2O 30 40 50 10 20 30 40 50 Leg Distance (it) Leg Distsnce (it) August 27.2002 August 30,2002 e 3 3 e . e : = 2 2 e0 E E I I a a 10 20 30 40 50 1O 20 30 40 50 Leg Distance (h) Leg Distance (M September 3. 2002 September 10. 2002 10 12 a a g 10 g «E a t I I 2 0 2 2 0 10 20 30 40 50 10 20 30 40 50 Leg Distance ('1) Leg Distance (in) September 13. 2002 I 0 c 2 b I 2 E I a 20 30 Leg Distance (n) 40 50 58 IIIIIIIIIIIIIIIIIIIIIIIIII {Will/1M!(HI/lllfl/IHIHIIII/mill!!!”Ill/Hill”?!ll 3 1293 02585 5622