University of Florida Thesis Or Dissertation Formatting Template

THE OCCURRENCE OF ENTOMOPATHOGENIC NEMATODES AND THEIR NATURAL ENEMIES IN GREEK CITRUS ORCHARDS

ALEXANDROS DRITSOULAS

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2020

To my wife, Stella and Greece

ACKNOWLEDGMENTS

I would like to begin by wholeheartedly thanking my advisor Dr. Larry W. Duncan for his unceasing support of my doctoral research. I would like to express my gratitude and appreciation to him for his immense enthusiasm and motivation during my studies. His insightful guidance has helped me navigate the most challenging part of my life thus far. I am a better person for having known him.

I would also like to acknowledge Dr. Tesfamariam Mengistu for his advice, assistance, and consideration for me particularly during my first two years as a Ph.D. student. I am also indebted to the rest of my dissertation committee: Dr. Peter DiGennaro, Dr. Felipe Soto-Adames,

Dr. Lukasz Stelinski, and Dr. Nian Wang for their perceptive comments and for encouraging me to widen the scope of my research.

My fellow graduate students and lab mates have been an integral part of my grad-student life and I would like to thank them for their friendship, compassion, and camaraderie. Their inspiring words will accompany me throughout my career and life.

Last but not least, I would like to express my immense appreciation to my parents for standing by me, and my wife for all the sacrifices she made and her willingness to follow me on the path I have chosen.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 7

LIST OF FIGURES ...... 8

ABSTRACT ...... 10

CHAPTER

1 GENERAL INTRODUCTION ...... 12

Review of Literature ...... 14 Entomopathogenic Nematode Ecology ...... 14 Entomopathogenic Nematode Phylogeny ...... 16 Soil Microarthropods ...... 18 Molecular approach and characterization of the target population ...... 24 Conserving Services of Entomopathogenic Nematodes: An Example ...... 28

2 TOWARD MAXIMIZING TAXONOMIC COVERAGE IN METAGENOMIC STUDIES OF ENVIRONMENTAL SAMPLES ...... 31

Introduction ...... 32 Materials and Methods ...... 35 Sucrose Centrifugation and Flotation-Berlese-Flotation Comparison ...... 35 Sucrose Centrifugation, Berlese funnels and Heptane Flotation Comparison...... 36 Results...... 37 Sucrose Centrifugation and Flotation-Berlese-Flotation Fomparison ...... 37 Sucrose Centrifugation, Berlese Funnels and Heptane Flotation Comparison...... 37 Discussion ...... 38

3 COMPARING HIGH THROUGHPUT SEQUENCING AND REAL TIME QPCR FOR CHARACTERIZING ENTOMOPATHOGENIC NEMATODE BIOGEOGRAPHY ...... 47

Introduction ...... 48 Materials and Method ...... 50 Samples Collection ...... 50 Library Preparation ...... 51 Bioinformatics ...... 52 Statistical Analysis ...... 53 Phylogenetic Tree ...... 54 Results...... 55 Discussion ...... 57

4 NATURAL OCCURRENCE OF ENTOMOPATHOGENIC NEMATODES AND THEIR NATURAL ENEMIES IN GREEK CITRUS ORCHARDS...... 70

Introduction ...... 71 Materials and methods ...... 73 Sampling ...... 73 Soil analysis ...... 74 DNA extraction ...... 74 Library preparation ...... 74 Bioinformatics ...... 76 Real time PCR testing, cultures of EPN and standard curve preparation ...... 77 Gene copies by individuals’ volume and data correction...... 78 Phylogenetic analysis ...... 78 Statistical Analysis ...... 79 Results...... 79 Discussion ...... 82

5 CONCLUSION...... 101

LIST OF REFERENCES ...... 104

BIOGRAPHICAL SKETCH ...... 118

LIST OF TABLES

Table page

3-1 Detection frequency of entomopathogenic nematodes (EPNs) and species previously reported to EPN competitors ...... 62

3-2 Non-parametric Spearman's correlations between the species’ measurements from metabarcoding and qPCR ...... 62

3-3 Significant variables from stepwise multiple regression of two nematode species measured by high throughput sequencing or qPCR ...... 63

3-4 Significance of the Canonical Correspondence Analysis model, axes, and variables...... 63

3-5 S. feltiae ASVs table illustrating “head-tail” structure associated with the presence of within the species variation...... 64

4-1 Amplicon sequence variant (ASV) table illustrating “head-tail” structures associated with intraspecific variation ...... 88

4-2 Non-parametric (Spearman's) correlations between number of metabarcoding reads and the Ct values derived from qPCR measurements of the nine EPN species ...... 89

4-3 Significance of the Canonical Correspondence Analysis variables...... 90

4-4 Significant variables from stepwise multiple regression of nine nematode species measured by high throughput sequencing...... 91

LIST OF FIGURES

Figure page

2-1 Process flow diagrams for four extraction methods ...... 41

2-2 Efficiency of sucrose centrifugation (SC) compared to that of modification flotation- Berlese-flotation (FBF) ...... 42

2-3 Efficiency of sucrose centrifugation (SC) compared to that of Berlese funnels (BF) in extracting eight microarthropod taxa from 250cc mineral soil sample...... 43

2-4 Extraction efficiency of Berlese funnels (BF), sucrose centrifugation (SC) and heptane flotation (HF) ...... 44

2-5 Proportional composition of microarthropod taxa. Pie charts represent recovery from Berlese funnels (BF), sucrose centrifugation (SC) and heptane flotation (HF)...... 45

2-6 Ecological indices (species richness, S’ ; Shannon Diversity Index, H’ ; dominance, D’) from samples extracted with three methods ...... 46

3-1 Phylogenetic relationships of ASVs identified as S. feltiae based on sequencing reads of ITS_1 region as inferred by maximum likelihood ...... 65

3-2 The frequency of S. feltiae (percent of 56 samples) through different types of vegetation was detected by metabarcoding and qPCR ...... 66

3-3 Box plots of data from high throughput sequencing (A) and qPCR (B) measuring Steinernema feltiae populations in Portugal...... 67

3-4 Fit of Taylor’s Power Law to sample statistics for Steinernema feltiae populations measured using high throughput sequencing (A) and qPCR (B) ...... 68

3-5 Canonical correspondence analysis depicting biplots of the regional distribution and relationships between significant abiotic factors and soil organisms...... 69

4-1 Analysis of ASVs identified in the genus Steinernema based on sequencing reads of the ITS_2 region as inferred by the UPGMA method ...... 92

4-2 Phenetic relationships of ASVs identified in the genus Heterorhabditis based on sequencing reads of the ITS_2 region as inferred by the UPGMA method ...... 93

4-3 Detection frequency (proportion positive sites) ...... 94

4-4 Relative abundances of acari mite families ...... 95

4-5 Relationship between relative number of ribosomal DNA copies per IJ measured by qPCR and the relative body volume per IJ estimated by Andrassy formula for nine EPN species ...... 96

4-6 Relative abundances of the entomopathogenic nematode species ASVs detected by metabarcoding in the two ecoregions Argos and Chania...... 97

4-7 Results from redundancy analyses (RDA) depicting biplots of both citrus ecoregions survey Argos and Chania (n = 62) ...... 98

4-8 Principal component analysis of the samples collected from the two ecoregions Argos (n=32) and Chania (n=30) ...... 99

4-9 Ecological indices (species richness, S’ ; Shannon diversity index, H’ ; eveness, J’) of EPN species from samples extracted from two ecoregions, Argos and Chania ...... 100

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

THE OCCURRENCE OF ENTOMOPATHOGENIC NEMATODES AND THEIR NATURAL ENEMIES IN GREEK CITRUS ORCHARDS

Alexandros Dritsoulas

August 2020

Chair: Larry W. Duncan Major: Entomology and Nematology

Entomopathogenic nematodes (EPNs) are promising biological control agents of key insect pests. They are widely, commercially formulated and can be used as both classical and augmentation biological control agents. Soil management tactics have been identified to conserve and enhance the services of naturally occurring EPNs. Despite their utility in agriculture and ubiquitous occurrence in the soil, the biogeography of EPNs is poorly understood, as are the physical environmental conditions favoring each of the more than 100 known species of EPNs and their natural enemies. One such guild of natural enemies of EPNs found in all soil food webs are the soil microarthropods, comprising the acari mites and collembola. This dissertation explores the utility of metabarcoding to characterize EPNs and microarthropods in the citrus orchards of two Greek ecoregions.

A single extraction method was identified to optimize the recovery of both nematodes and microarthropods from environmental samples. Sucrose centrifugation (SC), a common method long used by nematologists, was compared to several conventional soil fauna extraction methods for their efficiency in recovering soil microarthropods. Although heptane flotation (HF) recovered more mites and collembola than did SC, it recovered no nematodes and produced samples requiring excessive time to identify the organisms among the organic debris. Moreover,

there was no difference between SC and HF in relative extraction efficiency of microarthropod taxa, and ecological indices derived from each method did not differ. Surprisingly, SC was much more efficient than methods using Berlese funnels, the mainstay method of acarologists for more than a century.

I found that high-throughput sequencing (HTS) detected EPNs in environmental samples with greater sensitivity than that of real-time PCR (qPCR). The fit of Taylor’s Power Law to data from each method obtained from DNA archived from a survey of Portuguese habitats, showed HTS to be more accurate than qPCR. Further, multivariate analyses of both data sets detected the same biotic and abiotic variables as being related to the EPN.

Using HTS, and improved primer design, I found that citrus orchards in Argos and

Chania support EPN communities with greater diversity than reported elsewhere to date, but with very low abundance. In order to quantify EPN species, I derived a significant linear relationship between the relative species differences detected by metabarcoding and the mean copy numbers/nematode (from qPCR standard curves), based on the average body volume of each species (from Andrassy’s formula). Multivariate analyses indicated that soil clay content and pH are related to EPN community structure and that tydeid mites and Oscheius sp. nematodes may be important antagonists of these EPN communities.

CHAPTER 1 GENERAL INTRODUCTION

Entomopathogenic nematodes (EPN) comprise one of the most well-studied groups of organisms in soil. They are used as biocontrol agents and they serve, in combination with their symbiotic bacteria, as model organisms for studying mutualism. Although EPNs must interact with multiple guilds of soil organisms, it is well-noted that little is known about the types, importance and dynamics of specific subterannean interactions, which could affect the post- application biology of commercial EPNs, including their efficacy as well as that of naturally occurring EPNs. Because of the importance of EPNs in pest managemet, some attention has been paid to soil microarthropods that prey on nematodes and their ability to regulate EPN populations.

The past decade saw a major improvement in methods used to study soil food webs with the wide adoption of real time PCR to identify and quanify subterranean species (Torr et al.,

2007). Species-specific primers-probe combinations simplified the identification of soil organisms, which traditionally required expertise in different fields and was laborious and challenging due to the the small size and tremendous diversity of soil animals. EPN identification is particularly difficult because the unique free-living stage of EPN in soil, the infective juvenile, has few morphological characters to permit a confident identification. The major limitation of qPCR is its ability to detect only those species for which primers-probe tools are utilized. Continued development and cost reduction of metagenomic tools has the potential to overcome that limitation – indeed metabarcoding can theoretically detect all species for which a given gene region is represented in the molecular databases such as Genbank. Nevertheless, the taxonomic accuracy with which metagenomic tools characterize soil communities from environmental samples has not been compared to that of qPCR.

Given the expanding capacity to study soil food webs with molecular tools, the lack of a universal method for recovering mesofauna from soil is a hindrance that increases the time and effort needed for sample preparation and complicates the comparison of studies that employ varying methods. Traditionally, nematologists and entomologists have relied on very different means of recovering the animals they study, without comparing the efficiency with which these methods collect both nematodes and microarthropods.

Studies in the field of entomopathogenic nematology in Greece are quite limited with no characterization of local EPN community structures. All European countries on the

Mediterranean, including Portugal, Spain, France, Italy and Turkey, with the exception of

Greece, have had several EPN biogeographical surveys. With its very diverse landscape, Greece could contain ecoregions with very different communities of EPNs and their natural enemies.

Based on the above observations, my dissertation has the following objectives:

• Determine an optimum method to recover nematodes and microarthropods from

soil samples by comparing the most widely used existing extraction methods.

• Determine if high throughput sequencing can quantify soil organisms with

comparable accuracy to that of qPCR.

• Characterize the communities of EPNs and microarthropods in Greek citrus

orchards and habitat properties that may influence them.

Review of Literature

Nematodes are non-segmented, elongated roundworms that are colorless, without appendages, and usually microscopic. Roundworms in the phylum Nematoda are animals adapted to nearly every ecosystem, inhabiting a broad range of environments from marine and freshwater, to terrestrial environments. Four out of every five multicellular animals on the planet are nematodes (Platt, 1994). Nematode species can be difficult to distinguish, and although over

25,000 have been described, the total number of nematode species has been estimated to be 1 million (Lambshead, 1993). Nematodes are classified along with insects, mites and other molting animals in the clade of Ecdysozoa, and, unlike flatworms, the nematode digestive system is a tube like structure within another tube of body wall with a hemolymph filled pseudocoelom, which houses the reproductive system and other organs. Nematodes lack a circulatory system, but the nervous system is well developed (Decraemer and Hunt, 2006).

Entomopathogenic Nematode Ecology

Nematodes are divided in two basic groups, free–living and parasitic. Some of the plentiful parasitic and predatory species are considered as beneficial organisms to humans because they can be used to manage pests that are important in agriculture, forestry and health

(Bedding et al., 1993). Among beneficial groups, such as slug-parasitic nematodes, entomophilic nematodes, predatory nematodes (for managing plant-parasitic nematodes and plant-pathogenic fungi), there are the so–called entomopathogenic nematodes (EPN). The word entomopathogenic has the Greek origin entomon (insect) and pathogenic (disease causing). EPN have been used in classical, conservation, and augmentation biological control programs (Grewal et al., 2005).

Since first described, there was interest in the commercial development of EPNs because they are easily mass produced, require no registration by the U.S. Environmental Protection

Agency (EPA) or others, no special application equipment is needed, and the apparent lack of resistance development by targets.

More than 120 species of EPNs have been described, of which at least 100 are steinernematids and 21 are heterorhabditids (Bhat et al., 2020). There is a high specificity between entomopathogenic nematodes and their bacteria symbionts; bacteria from the

Xenorhabdus spp. and Photorhabdus spp. (Gram–negative, γ–Proteobacteria) are associated with

Steinernema and Heterorhabditis, respectively (Lewis and Clarke, 2012). Together, nematodes and their symbionts form a complex that have insecticidal effect against a wide range of insect hosts (Kaya and Gaugler, 1993). After entering the insect’s hemolymph, the infective juveniles release bacteria symbionts and the host dies usually within 2 to 3 days. Inside the insect cadaver amidst an abundance of nutrients, the nematodes undergo normal development completing the life cycle where infective juveniles molt twice and become adults. The nematodes feed on the developing population of bacteria and host tissues while completing their development. Like all nematode genera, EPN have six stages, the egg, four juvenile stages and the adult stage.

Nematodes may carry out one to three generations inside the host (Lewis and Clarke, 2012).

When food resources become inadequate in the insect cadaver, nematodes undergo an alternative developmental stage, which can survive harsh conditions, called alternately “dauer” (DJ) or infective juvenile (IJ). The IJs exit en mass from the insect cadaver in a search of new hosts

(Lewis and Clarke, 2012). The main difference in lifecycle between the two genera of EPN is that the first generation of heterorhabditids is exclusively hermaphroditic and the following generations include males, females, and hermaphrodites (Strauch et al., 1994). In contrast, all

generations in steinernematid species are amphimictic forms (males and females), except one species with a first generation hermaphroditic stage, Steinernema hermaphroditum (Stock et al.,

2004). Thus, heterorhabditids require just one individual IJ to penetrate and develop in a host, while most steinernematids require at least two individuals for host colonization before the proliferating and egg laying by females in the host medium.

The dauer is a non-feeding stage with a closed buccal cavity and thick cuticle relying on lipid energy reserves (mainly triglycerides) and glycogen storage granules (Fitters et al., 1999).

The process from the previous host to the new one comprises four phases characterized by specific behaviors: dispersal, foraging, host discrimination and infection. EPN have developed two main foraging modes, the cruiser foragers and the ambush forages. Cruisers, roam the soil in a search for the next host, led by volatile signals. The IJs of ambushers, such as Steinernema carpocapsae, and Steinernema scapterisci remain near the soil surface (Georgis and Poinar,

1983) where they spring their bodies into the air, facilitating their attachment to passing insects

(Campbell and Gaugler, 1993).

Entomopathogenic Nematode Phylogeny

The term entomopathtogenic nematode refers to species of Steinernematidae and

Heterohabditidae families vectoring insect–pathogenic bacteria of genus Xenorhabdus and

Photorhabdus (Kaya and Gaugler, 1993). A few studies have characterized two Oscheius species, Oscheius caroliniesis (Rhabditida: Rhabditidae) and Oscheius chongmingensis (Rhabditida: Rhabditidae), and one Caenorhabditis sp., Caenorhabditis briggsae (Rhabditida: Rhabditidae), as insect pathogens because of their association with entomopathogenic Serratia bacteria (Abebe et al., 2010). Although these species can cause disease to a host insect as a result of their association with bacterial symbionts, their association with the bacteria is facultative. Thus, Dillman et al., (2012) suggest criteria based on

fundamental principles of the EPN lifestyle to differentiate EPNs from phoretic, necromenic, or other less virulent forms of parasitism. These criteria are: (1) nematodes may have a symbiotic relationship with bacteria to facilitate pathogenesis (their association may not necessarily be obligate, but it should not be transient), and (2) insect death should occur sufficiently rapidly, usually in less than 120 h.

According to Poinar, (1990) species delimitation in both EPN families was mainly based on comparison of morphological or morphometric data and crossbreeding tests, with the biological species concept forming the framework for their delimitation. However, crossbreeding tests are laborious and time consuming, and interpretation of morphological features for species identification in Steinernema requires substantial expertise to ensure accuracy. Taking into consideration all the available methods, analysis of nucleotide sequence data has proven useful not only for diagnostics at different taxonomic levels but has also provided valuable data for phylogenetic inference of EPN (Adams et al., 1998). Studies of Steinernema and Heterorhabditis have been conducted using molecular methods, such as random amplification of polymorphic

DNA (RAPD) (Liu and Berry, 1996) and restriction fragment length polymorphism (RFLP)

(Reid et al., 1997). Among nuclear genes, ribosomal genes have been used extensively at different taxonomic levels. Ribosomal genes include the 18S rRNA gene, the internal transcribed spacers (ITS1 and ITS2), the 5.8S and the 28S rRNA genes, which contain variable and conserved regions (Stock, 2009). The 5.8S rRNA gene is a highly-conserved region in contrast to the ITS1 and ITS2 regions, which evolve at a higher rate than the 18S and 28S rRNA genes, making them ideal for phylogenetic studies at species and population levels (Stock, 2009).

Sequences of the D2D3 region of the 28S rRNA gene were also used frequently to characterize

EPN populations (Stock et al., 2001) and are informative for species phylogeny. In addition,

mitochondrial cytochrome c oxidase subunit I (COXI) and cytochrome b (cytb) genes evolve more rapidly than any other loci of the ribosomal gene, giving a higher resolution and more information for phylogeny at this taxonomic level (Stock, 2009).

Moreover, quantitative-PCR, employing species–specific primers, has shown great potential to EPN ecology studies. Its high sensitivity has efficiently supported monitoring of

EPN establishment and distribution, population dynamics and interactions among soil dwelling organisms at different trophic levels (Raquel Campos-Herrera et al., 2011). Real time PCR assays have also overcome some of the limitations in conventional EPN species detection (e.g. isolation time from the soil, identification of a single species) methods like the baiting-method of Galleria mellonella (Campos-Herrera et al., 2011). Finally, next–generation sequencing coupled with powerful bioinformatics tools has made possible de novo acquisition of many nematode genomes (Schwartz et al., 2011).

Soil Microarthropods

Mites and collembola are frequently referred as ‘generalists’, ‘polyphagous’ or

‘omnivorous’, and many studies have reported their role in suppressing nematode populations.

These generalist predators dominate higher trophic levels in the soil food web and provide a background level of nematode suppression in all soils. Mites, collembolans and symphylans, together with a variety of small insects collectively comprise the community defined as microarthropods and they are ubiquitous in soil. They are particularly abundant in natural habitats such as grasslands and forests, where soil organic matter has accumulated over many years, root biomass is high, roots are permanently present and a layer of decaying litter always covers the surface of soil (Stirling, 2014).

Mites are the most abundant arthropods in soil and are found in four major taxonomic groups: the Oribatida formerly Cryptostigmata, Prostigmata, Mesostigmata and Astigmata. More

than 50,000 species have been described. Oribatids are by far the most numerous soil mites, but in contrast to other microarthropods, they reproduce relatively slowly and therefore are considered to be K strategists. Although mainly fungivores and detritivores, some are predacious. The Prostigmata are the most diverse group of soil mites, displaying a huge range of morphological and behavioral variation. Many species of this order are well known as serious plant pests, however below ground species are also important because of their wide distribution and large populations which occasionally transcend those of other mites. Soil dwelling prostigmatic mites generally feed on fungi, algae, or ingest particulate matter, but some are predators, feeding on other arthropods or nematodes. Mesostigmatid mites are not as numerous as oribatid or prostigmatid mites, nevertheless, are universally present in soil in different kind of habitats. A few species are fungal feeders, but most are predators. There are indications that mites of this order show specificity in their prey. Larger species tend to feed on small arthropods or their eggs, while the smaller species are mainly nematophagous. The Astigmata are the least common of the soil mites. In agroecosystems, they are most abundant in situations where crop residues or manures are aggregated into soil under moist conditions.

Collembolans (springtails) are basal hexapods that occasionally show antagonistic interactions with mites by their abundance in soil. Many species are opportunistic r-strategists capable of rapid population growth where organic matter decomposes and high microbial activity provide them with food. They primarily feed on fungi, but many species are omnivorous and some are known to consume nematodes. Symphylans, also known as garden centipedes or pseudocentipedes, are soil-dwelling arthropods of the class Symphyla which are relatively large white arthropods that resemble centipedes. They are widespread in soil, but they are not as common as collembolans. They are photophobic and can move rapidly. They hide in voids when

soil is disturbed and are, therefore, rarely seen. They feed on decomposing organic matter, and can also be crop pests when they consume plant roots. Walter et al. (1989) showed that some species are predators of arthropods and nematodes.

Microarthropods are grouped into five feeding categories (bacterivores, fungivores, plant feeders, omnivores and predators), but omnivory is common, even among animals that may exist primarily on one food source. Thus, many arthropod predators are generalists, opportunists and may attack nematodes, lumbricid and enchytraeid, oligochaetes, tardigrades, insects, myriopods and mites (Moore et al., 1988).

Walter and Ikonen, (1989) provided evidence of nematophagy from field observations of various groups of microarthropods. Despite focusing just on grassland soils of Colorado, the observations they made reveal a great deal about the types of microarthropods that prey on nematodes, and their feeding habits and behavior in soil. Studies of feeding by 14 species of mesostigmatic mites showed that they ate 4–8 nematodes (Acrobeloides sp.) per day, which equates to between 22% and 109% of their body weight. The amount of nematode biomass consumed was strongly related to the body weight of the predator. Another important assumption of Walter’s and Ikonen’s work was that mites did not reveal a significant preference when a choice of nematode prey was given. The most crucial nematophagous taxa in Colorado grasslands observed in this research were the Acari (mites) and Symphyla. All four main acarine sub-orders were found in grassland soils where nematophagy and fungivory occurred by different species across all four orders. This implies that mites’ feeding habits are a species-level phenomenon, hence, it is not always possible to infer specific feeding habits based on observations of mite morphology. Mesostigmatic mites were reported to be the most important group of nematophagous arthropods, as only six of 63 species tested did not readily feed on

nematode prey. In the presence of only nematode prey, most of the nematophagous species produced eggs and developed into adults. Nematophagous mesostigmatids had developed specialized mouthparts different from those with which generalist predators were equipped.

Observations on 30 mite species indicated that there was not a clear relationship between the structure of their chelicerae and feeding habits (Walter and Ikonen, 1989). Most mesostigmatic mites consumed nematodes and arthropod prey, and some were also able to feed on fungi. Thus, three categories of nematophagous arthropods were recognized: omnivores feeding on microbes and nematodes; general predators of soil invertebrates; and nematode specialists. However, it was recognized that the specialized group could attack a variety of prey. Santos and Whitford

(1981) confirmed nematophagy in tydeid mites in Chihuahuan desert environment. Species in families Rhodacaridae, Digamasellidae and Ascidae were considered to represent a

‘nematophagous mite guild’ characterized by: (i) an ability to maintain continuous cultures on nematode prey; (ii) the presence of chelicerae adapted to feeding on nematodes and arthropods; and (iii) a small body size and convergent body plan that allowed these animals to access prey residing in small pore spaces.

Thereafter, more mites have been shown to be predators of nematodes. For instance, mites from four different orders (Mesostigmata, Endeostigmata, Oribatida and Astigmata) were tested against entomopathoggenic nematodes in the laboratory and the majority of these mites were able to consume Steinernema feltiae and Heterorhabditis heliothidis (Epsky et al., 1988).

Although most species also fed on arthropods, some required nematode prey for completing the life cycle reproduction. Observations in a strawberry field in Turkey indicated that after entomopathogenic nematodes killed white grubs, mites of genus Sancassania (mites phoretic and necromenic associated with white grubs) molted to adult stage and fed on the host tissues, on

microbes associated with the cadavers, and on emerging IJ’s. In laboratory tests, this mite consumed about 40 nematodes per day. was more likely to consume S. feltiae than

H.bacteriophora, and were able to kill more nematodes in sand than in a loamy soil (Karagoz et al., 2007). Oliveira et al., (2007) reported that the oribatid mite Pergalumna sp. consumed an average 18 Meloidogyne javanica and 42 Pratylenchus coffeae in 24 hours after counting the stylets of the plant parasitic nematodes in the fecal pellets of the mites. Thus, all these studies infer that many mites are generalist feeders, but they are capable of killing significant numbers of nematodes in some situations.

Despite the limited research on the contribution of nematodes to the diet of Collembola, there is evidence that many species can consume large numbers of nematodes. According to

Gilmore, (1970) in an extensive laboratory study, 10 of 12 Collembola species tested fed on nematodes. In two of these species (Entomobryoides dissimilis and Sinella caeca), one individual was able to consume more than 25 nematodes in 24 hours, in arenas of mixed soil with vermiculite. In another experiment in which two species of Collembola (Tullbergia krausberi and Onychiurus armatus) were introduced into a sandy soil, the number of Tylenchorhynchus dubius on ryegrass seedlings was reduced by about 50% after 77 days (Sharma, 1971). Although both these studies were carried out in highly artificial environments, the high numbers of prey consumed suggests that nematodes are likely to be part of the diet of many collembolans and that texture of soil plays a crucial role in this interaction. However, consumption of nematodes in the laboratory does not necessarily infer measurable predation in soil. Proisotoma sp. reached high population densities in greenhouse cultures of root-knot nematode (more than 10,000 individuals/L soil) and was associated with decreased numbers of nematodes in pots, but

observations of gut contents revealed only decaying roots and associated microflora (Walter et al., 1993).

Numerous molecular approaches for detecting prey remains in the guts, feces and regurgitates of predators are now available in which we can measure predation by soil-dwelling arthropods on nematodes (Pompanon et al., 2012). A PCR-based approach was developed for detection of three species of entomopathogenic nematodes, in which springtails and mesostigmatids mite were employed to calibrate post-ingestion prey detection. Fields tests confirmed predatory activity by retrieving microarthropods from barley plots where the nematode Phasmarhabditis hermaphrodita had been applied 12 h earlier. Prey DNA was detected in three of the four most abundant microarthropod taxa recovered from the field, indicating that some species had fed on the introduced nematodes (Read et al., 2006).

Additionally, several microarthropod taxa that were normally considered detritivores were found to consume nematodes when they were exposed in a no-choice laboratory experiment (Heidemann et al., 2011). There are also experiments in which predators had the option of different prey. Phasmarhabditis hermaphrodita and Steinernema feltiae were added, alive and freeze-killed, to forest soil and mites were later retrieved and checked for the presence of the inoculated nematodes. Mesostigmatids of Gamasidae and Uropodidae families, consumed nematodes, but interestingly, several of the oribatid mite species also ate living nematodes, although they have the option of different food in the field. These studies suggest that many of these omnivores are opportunists feeding on nematodes and that a huge range of microarthropods are potentially predators, particularly in situations where nematodes are readily available or when generalist predators encounter specific soil characteristics.

Molecular approach and characterization of the target population

The DNA barcode or a ‘bar-code of life’ is a concept first described in 2003 at Cold

Spring Harbor Laboratory (Stoeckle, 2003). The DNA barcode is analogous to the universal product code (UPC) commonly used on retail products also known as the ‘zebra code’. Instead of a numeric code, the molecular barcode has a nucleotide sequence from a common gene that reveals information unique for every species in the planet capable of being discriminated by it.

The bar-code system approach is quite different from a phylogeny approach but can be used with pre-existing phylogenies. There is overlap between DNA bar-coding, taxonomy and systematics terms. Combining all these approaches together could give a better system for species identification, measuring biological diversity and assessing evolutionary relationships among and between various taxonomic ranks (Stock, 2009).

Despite arguments about the validity of a universal DNA bar-code system and how best it can be employed, numerous research projects tested the concept. According to (Blaxter, 2003) nematodes are among the first organisms used to test the barcode concept. For example, Floyd et al., (2002) created a phylogenetic tree from unknown nematodes sampled from Scottish upland grasslands by interpreting and comparing the sequences of an 18S ribosomal DNA (rDNA) barcode. More recently, 18S sequences were also considered as a ‘coarse diagnostic tool’ identifying 360 nematode species from a tropical rainforest in Costa Rica (Powers et al., 2009).

In that study, 18S rDNA sequences were generated via direct sequencing of the PCR products and sorted into molecular operational taxonomic units (MOTUs) based on primary sequence. A total of 167 unique nematode MOTUs were identified and compared with small subunit (SSU) sequences archived in GenBank to assess putative identifications and likely relationships.

As reported by Powers, (2004) there is not sufficient information in nematode databases for extensive delimitation of all species based on the 18S, however getting a glimpse of

phylogenetic trees gives an inference of the relative position in the tree of our unknown nematode samples. Expansion of the 18S nematode tree of life through collaboration of projects such as NemATOL (National Science Foundation (NSF)-funded nematode branch of the Tree of

Life Project, http://nematol.unh.edu/) will undoubtedly become a valuable resource to the DNA bar-code system of this Nematoda.

Considering collembola and mites in addition to nematodes, DNA metabarcoding has tremendous potential to characterize species and biodiversity in environmental DNA.

Metabarcoding of animals increasingly uses the cytochrome c oxidase subunit I (COI) gene, because no other genetic region has the breadth of taxonomically verified coverage. The focus on the COI gene is due to its conservative nature among protein-coding genes in the mitochondrial genome of animals (Brown, 1985). There are plentiful papers suggesting a comprehensive phylogenetic reconstruction of the major lineages of acariformes using sequences coding for the small subunit rRNA (18S rRNA) gene and a fragment of the cytochrome c oxidase subunit I

(COI).

Metagenomics refers to studies that focus on genetic analysis of samples recovered directly from the environment. As synonyms we could refer environmental genomics, ecogenomics or community genomics. Metagenomics has important application to ecology, evolutionary and conservation biology, where estimating species richness is a priority. It has great potential to fill many gaps in created in mesofaunal taxonomy related to morphological analysis through light microscopy. The belowground biodiversity is poorly known, where nematodes and microarthropods are expected to have a high species richness of which only a small proportion has already described (Lambshead D., 1993; Walter and Proctor, 2013). Not only is the analysis of morphological characters to delimit species time consuming but also

infeasible for many related species (Ristau et al., 2013). Metagenomics play a vital role in the field of biogeography. Many conventional reports detail mesofauna and meiofauna that have low dispersal abilities (Jörger et al., 2012), while other studies proved that they are cosmopolitan

(Curini-Galletti et al., 2012). These contrasting conclusions suggest the broad existence of cryptic species, which are morphologically similar but genetically distinct different species

(Pfenninger and Schwenk, 2007). To understand a diverse mesofauna, standard barcoding based on Sanger sequencing is not a perfect tool for investigating biodiversity at large spatial scales.

Next Generation Sequencing with the ability of mass molecular identification shows great potential to assess mesofaunal and meiofaunal diversity.

Roche 454 was the first commercial platform successfully used for metagenomic surveys of metazoan biodiversity (Creer et al., 2010). Taking into consideration marine nematodes surveys, Roch 454 is the most frequently used NGS platform. In contrast to all advantages that

454 platform has, there are some considerable drawbacks. The cost is high due to high cost of reagents and there is a reading error rate approximately 1% per base at a single read (Glenn,

2011). This is the reason why Illumina platform is increasingly used since the cost is much lower and error per base is lower, approximately 0.1% (Glenn, 2011). The most important for Illumina was that improving the sequencing protocols and libraries gives better-quality results. For instance, the quality and the length of the amplicons could be improved by “read merging” where each set of paired end reads is combined to a single contig due to the common overlapping area of these reads.

Handling a metagenomic sample we should take into consideration shortcomings and pitfalls of this approach. During library preparation, PCR amplification can introduce PCR artifacts like chimeras (V. Wintzingerode et al., 1997). Chimeras seem to have higher probability

in richer and more genetically diverse samples where two molecules with some similarities, but from different organisms, artificially combined in the same sample. Another thing we should be aware of is the speculative biodiversity sought by using the specific primers. According to Bik et al., (2012) amplification and sequencing of a single locus cannot recover all the biodiversity included in one sample. Therefore, in order to overcome this problem and be more sensitive identifying more taxa in a sample, it is suggested to use a “cocktail” of primer sets targeting alternative loci (Pfenninger and Schwenk, 2007). This would prevent underestimation of the actual biodiversity. For instance, if a single primer set did not match in a taxon (even this is a lower resolution taxon), output results would omit the whole taxon. Due to enormous number of raw reads, output filtering and clustering steps are required, in order to be able to cluster all these sequences in MOTU’s (molecular operational taxonomic units). MOTU approach requires to set up a threshold, a critical value of the proportional similarity of sequences, with which we can define the species limit. This critical value is called cut-off and the most common value is between 95% to 99%. An alternative way of clustering is the amplicon sequence variant (ASV) which refers to individual DNA sequences recovered from a high-throughput marker gene analysis following the removal of spurious sequences generated during PCR amplification and sequencing. In the last few years, ASVs have almost replace the OTUs, as the ASVs are thus inferred sequences of true biological origin without any arbitrary thresholds.

With respect to EPN biogeography, early studies inferred that, steinernematids are generally associated with temperate regions as opposed to heterorhabditids occurring mostly in warmer areas. However, Hominick, (2002) proposed that EPN distribution depends more on species rather than genus level adaptation. Variation of traits, within each genus, such as morphology (size, sheath retention, etc.), survival strategy, foraging behavior, and host

specialization are more likely to drive their speciation and global distribution (Hominick, 2002).

Confirming the previous assumptions, Spiridonov created phylogenetic trees based on ITS1-

5.8S-ITS2 sequences and showed that there is inconsistency in features such as habitat preference and geographical distribution patterns. In addition to climate, cropping systems affect habitat stability with respect to the crop longevity and its cultural practices. Forests, timber lots, pastures, orchards, and other perennial systems provide edaphic stability that does not exist in annual cropping systems, especially in arable crops where soil disturbance is higher (Tarasco et al., 2015). Tillage negatively affected EPN species in the southeastern and central USA

(Lawrence et al., 2006) and EPNs were rarely encountered in long–term tillage trials in

Switzerland (Campos-Herrera et al., 2015b). Neher (1999) concluded that tillage is the only crop management practice with such a negative impact that EPNs cannot persist. Texture of soil is one of the most crucial edaphic properties for EPN distribution, with higher EPN activity in sandy or loamy soils than in heavier clay soils. Also, salinity, pH, concentration of fertilizer elements and pesticide residues are just a few of the edaphic variables that influence EPN occurrence

(Barbercheck and Duncan, 2004).

Conserving Services of Entomopathogenic Nematodes: An Example

In Florida, citrus is grown in two ecoregions, the Central Ridge and the Flatwoods. Both regions have very sandy soil (often 85–98 % sand) although flatwoods soils are somewhat finer textured. Trees in flatwoods orchards are frequently grown on bedded soil to provide adequate rooting volume because soils are poorly drained due to shallow (<1 m) groundwater. By contrast, the root system depths of trees on the deep, coarse sands of the central ridge frequently exceed 5 m. The size of root weevil (Diaprepes abbreviatus) populations and the damage they cause to citrus trees differs in the two regions (Futch et al., 2005).

Diaprepes root weevil is a polyphagous invasive pest, originated from the Caribbean that was detected in Florida in 1964. Adult weevils feed on young leaves while the larvae feed on fibrous and large structural roots as they develop and grow giving access to plant parasitic

Phytophthora spp., creating a pest–disease complex (Graham et al., 2003). In flatwoods citrus orchards, root weevil population density can be high enough to kill large numbers of young trees, especially in poorly drained areas combined with high salinity (McCoy, 1999). Root weevils are less abundant and often go undetected for many years in orchards on the central ridge.

Autochthonous EPN were first reported to be major soil enemies of Diaprepes root weevils in citrus orchards by Beavers et al., (1983) who found as high as 70 % infection by nematodes of caged larvae that were buried for three weeks beneath citrus trees canopy in central Florida.

Duncan et al., (2003) observed high weekly mortality rates of caged, buried Diaprepes root weevil larvae as baits for EPN in orchards of central ridge and flatwoods. During two years of monitoring in both regions, differences in EPN species diversity and in mortality rates were documented. In the central ridge ecoregion native Steinernema diaprepesi, Heterorhabditis indica and Heterorhabditis zealandica were observed with a mortality rate for the baiting insects as high as 50% while in a flatwood site only H. indica was detected with a mortality rate that never exceeded 8%.

Due to the big differences between these two regions in weevil occurrence, Duncan et al.,

(2003) speculated that the regional patterns of Diaprepes root weevil abundance might be a result of natural control carried out by native EPN which appeared to be more abundant and in higher diversity on the central ridge. Campos-Herrera et al., (2013) confirmed that the central ridge citrus orchards supported greater EPN evenness, diversity and species richness. They confirmed also that H. indica dominated flatwoods communities in association with sporadic occurrance of

Steinernema khuongi, while communities with abundant S. diaprepesi, H. zealandica and H. indica occurred on the central ridge. Moreover, spatial patterns of S. diaprepesi were more aggregated regionally whereas H. indica was ubiquitous and often the dominant species. A second survey of EPN in 91 natural areas of Florida supported the habitat preferences of these species and showed that Florida citrus orchards are generally more conducive to the occurrence of EPNs than the more heterogeneous natural habitats. Differences in detection frequency between natural and managed perennial areas was similar to other reports showing that natural areas are less suited to EPNs than managed orchard habitats (Campos-Herrera et al., 2008;

Tarasco et al., 2015).

The original question that inspired this study was “if EPNs modulate root weevil abundance differently on the central ridge than in the flatwoods, what is the mechanism and can it be exploited for biological control?” Subsequent studies of EPN species that inhabit Florida citrus orchards demonstrated that Diaprepes root weevil is best managed by the closely related S. diaprepesi and S. khuongi (El-Borai et al., 2012) and that the two EPNs are adapted to the central ridge and flatwoods, respectively, based on their capacity to survive desiccation (El-Borai et al.,

2016). Effective conservation biocontrol tactics based on management of soil water potential

(Duncan et al., 2013) and soil pH (Campos-Herrera et al., 2019a, 2014) favor S. diaprepesi directly (Raquel Campos-Herrera et al., 2011) and indirectly by suppressing a bacterial natural enemy of the nematode (El-Borai et al., 2005). Use of particular citrus rootstocks shown to attract EPNs to root herbivores has been proposed (Ali et al., 2012). Accordingly, this dissertation addressed relationships between EPNs and microarthorpods in Greek orchards by characterizing the communities and their spatial patterns with regard to habitat properties.

CHAPTER 2 TOWARD MAXIMIZING TAXONOMIC COVERAGE IN METAGENOMIC STUDIES OF ENVIRONMENTAL SAMPLES

Chapter summary: Manipulating soil properties to modify the dynamics between nematodes and their natural enemies has been proposed to conserve services such as the biological control of insect pests by entomopathogenic nematodes. Many soil microarthropods including acari mites and collembola are natural enemies of nematodes; however, little is known about the naturally occurring assemblages of these two soil dwelling groups and how they might be influenced by soil conditions. Despite the ready availability of metagenomic tools to characterize soil communities of widely diverse taxa there is not a single method used that could recover both groups. Because samples of nematodes extracted from soil by sucrose centrifugation (SC) also contain soil mites, collembola, protozoans and fungal and bacterial propagules, the efficiency of SC to recover microarthropods was compared to more conventional methods such as heptane flotation (HF), Berlese funnels (BF) and a modified flotation Berlese method (FBF). Organisms were identified using an inverted microscope to class in one experiment and to order in a second. Significantly more microarthropods of all taxa were recovered by SC than with either Berlese method (BF or FBF). Forty percent more microarthropods comprising seven orders were recovered by HF compared to SC, but the difference was not significant. Ecological indices (diversity, richness and evenness) derived from HF and SC were congruent and significantly higher than those derived from BF. Compared to SC, BF and SBF, excessive organic matter in the HF extractions made mite detection and identification difficult and time consuming. Moreover, neither HF nor any Berlese method recovered nematodes. Accordingly, we found SC to be the most efficient method for microarthropod extraction, making it an ideal method for metagenomic studies of communities of nematodes and many of their natural enemies in the soil.

Introduction

Entomopathogenic nematodes (EPN) are among the most well studied soil guilds. These nematodes are ubiquitous across all continents except Antarctica. In many crops they decrease damage by insect plant pests (Duncan et al., 2013; Frank and Walker, 2006; Shields et al., 2009).

Several companies produce EPN products for use by growers; nonetheless, few reliable methods to recruit the services of naturally occurring nematodes to manage insect pests are known. Soil type and texture affects EPN performance (El-Borai et al., 2012; Koppenhöfer and Fuzy, 2006;

Nielsen and Lewis, 2012) and adding sand to tree planting holes increased EPN efficacy against a root weevil pest of citrus (Duncan et al., 2013). Cultural practices including application of soil amendments (Bednarek & Gaugler, 1997; Duncan et al., 2007), and tillage (Millar and

Barbercheck, 2002) had positive and negative effects on EPNs, respectively. The EPNs also occupy a trophic level in a wider food web that includes natural enemies such as nematophagous fungi (Kaya and Koppenhöfer, 1996), ectoparasitic bacteria (Enright and Griffin, 2005) and soil microarthropods (Walter and Ikonen, 1989) that regulate EPN populations. Surveys and experiments have identified physical soil properties, such as pH (Campos-Herrera et al., 2013a;

Hara et al., 1991), salinity (Nielsen et al., 2011) and ground water depth (Campos-Herrera et al.,

2013b), that potentially modulate EPN populations directly or indirectly by affecting their hosts

(Gazit et al., 2000) or natural enemies (Campos-Herrera et al., 2019).

Mites, springtails and other microarthropods are major components of soil biodiversity and food web function. Numerous reports indicate that these generalist predators dominate the higher trophic levels that regulate nematode populations. Poinar (1979) observed mesostigmatid mites in the genus Macrocheles consuming infecting juveniles (IJs) of Steinernema feltiae. Walter and

Ikonen (1989) described nematophagy in various groups of microarthropods. Epsky et al. (1988)

found that 12 of 13 mite species in the orders mesostigmata, endeostigmata, oribatida, astigmata fed on Steinernema carpocapsae, and Karagoz et al. (2007) reported that presence of

Sancassania sp. reduced EPN efficacy. Compared to laboratory experiments and field observations (Jabbour and Barbercheck, 2011; Wilson and Gaugler, 2004), there are fewer comprehensive studies of microarthropods in naturally occurring food webs. Affordable metagenomic tools now provide wider opportunity to study cryptic soil communities. Traditional methods of recovering organisms from soil vary by discipline. Estimates of optimum methods based on extraction efficiency and cost usually focused on one group of organisms among many that might be recovered. For the purpose of studying nematode and microarthropod communities we are unaware of any comparisons of extraction efficiencies of both groups by a given extraction method.

Microarthopods and nematodes can both be separated from soil by passive methods (flotation, rinsing, adhesion), or by allowing organisms to migrate actively from soil into a trapping device

(Berlese funnels or Baermann funnels) (Krantz and Walter, 2009; Southey, 1970). Nematodes are most commonly recovered using various modifications of the Baermann funnel (Townshend,

1963) or centrifugal flotation (Jenkins, 1964). In addition, live baiting techniques using sentinel insect larvae can be used in the laboratory or in the field for detection or estimation of entomopathogenic nematode populations (Fan and Hominick, 1991; Koppenhöfer et al., 1998).

Live baiting has the advantage of recovering all EPN life stages inside the insect, not only the dauer larvae that are the only life stage in soil. It is noteworthy that centrifugal flotation recovers not only nematodes but also other light components in mineral soil such as microarthropods and fungal spores.

The most commonly used procedure to recover microarthropods is Tullgren extraction using

Berlese funnels. In Tullgren extraction, the litter or soil samples are placed on a mesh screen inside a collection funnel. Light and heat is applied to the upper side of the sample, creating a temperature gradient which causes a progressive desiccation which drives microarthropods from the sample and into a collection vessel. Heptane flotation (HF) exploits the lipophilic nature of the microarthropods’ cuticle where the apolar waxy cuticle has affinity for the polar heptane and not to the apolar water (Aucamp and Ryke, 1964). The amount soil processed by these procedures is generally in the range of 100-250 g, but this volume is insufficient to capture arthropod diversity in the deeper, mineral soil fraction using Tullgren extraction. For this reason,

Arribas et al. (2016) added an extra flotation step (described below) to the Tullgren extraction protocol in order to improve the efficiency of the Berlese in recovering microarthropods from the extraction substrate.

Mites and collembola are commonly encountered in nematode samples extracted by sucrose centrifugation. Duncan et al. (2007) measured food web responses to bare and manure-mulched soil augmented with entomopathogenic nematodes, where nematodes, mites, collembola, enchytraeid worms, nematophagous fungi, and bacterial ectoparsites of EPNs were all extracted with SC. Sucrose centrifugation is also used for extracting mycorrhizae, by virtue of the spores

(Shamini and Amutha, 2014). Here we compared the efficiency of SC to three methods developed for capturing microarthropods. Our hypothesis is that SC, unlike other methods, is an efficient technique to recover both nematodes and microarthropods from mineral soil and is especially well-suited to metagenomic studies of nematodes and their natural enemies. To achieve these goals our objectives were to i) test SC extraction efficacy compared to that of

modified flotation-Berlese Funnel method, and ii) test the efficacy of SC to characterize the microarthropods community compared to Berlese funnels and heptane flotation.

Materials and Methods

Sucrose Centrifugation and Flotation-Berlese-Flotation Comparison

We compared the microarthropod extraction efficiency of a flotation-Berlese-flotation

(FBF) method (Arribas et al., 2016) to that of sucrose centrifugation (Jenkins, 1964; SF), using soil samples from an experimental citrus orchard adjacent to the University of Florida

Department of Entomology and Nematology. An auger was used to extract 24 cores (dia. 10.5cm x 23cm depth; ~2000 ml volume). Two cores collected from each of 12 trees were processed by either FBF or SC.

For FBF (Figure 2-1A), the large 2-liter soil sample was mixed vigorously in a bucket with 30 liters of water to dissolve soil aggregates and allow mineral material to sediment.

Immediately after mixing, the floating material was filtered through a 400-mesh sieve (38 microns) to obtain a bulk subsample of <250cc containing organic matter and soil mesofauna.

Subsamples were processed in a Berlese apparatus (Berlese 1905; Tullgren 1918; Southwood and Henderson 2000) where the sample rested on a layer of cheese-cloth placed over a plastic mesh in the 30.5 cm dia. funnel and exposed to a moderate vertical gradient of heat and light (25 watt lamp) until the organic material was completely dry (approx. 5 days). All mesofauna collected in the flask of 95% alcohol beneath the funnel were captured on a 400-mesh sieve and preserved in absolute ethanol in a 15ml tube.

Samples extracted by SC (Figure 2-1B) were processed initially in the same manner as for FBF, except that the organic subsample was collected in 2-4, 100-ml centrifuge tubes and centrifuged 1700 rpm (810 g) to precipitate nematodes and soil particles and remove organic debris in the decanted supernatant. The samples were then resuspended in a high-density sugar

solution (specific gravity = 1.10-1.18) to precipitate soil particles and suspend mesofauna in the supernatant for collection with a sieve.

Microarthropod specimens were identified at the level of subclass for Collembola and

Acari mites using a dissecting microscope. Significant differences between sampling methods in the frequency of specimens were determined by T-test.

Sucrose Centrifugation, Berlese Funnels and Heptane Flotation Comparison

Samples were taken from 15 sites (auger dimensions, dia. 2.5cm x 28cm depth) in the

Natural Area adjacent to the Department of Entomology and Nematology. Twelve cores were taken at each site and 4 randomly chosen cores were combined into three subsamples of 250cc.

One subsample from each of the 15 sites was processed either by BF, SC and HF. However, samples processed with HF produced an excess of organic matter making detection of microarthropods difficult. Consequently, two data sets were created as two independent experiments. The efficiency of BF compared to that of SC was determined using data from 15 sites. A comparison of all three methods was made using data from just 6 of the sites from which counts were also made for the HF method.

Fauna were extracted in Berlese funnels as described previously. The illumination of the apparatus was controlled by a potentiometer adjusted the first day to produce a weak vertical gradient of heat and dimmed light. The light brightness was increased chromatically each passing day and the samples remained in the funnels for 8 days.

Samples were processed by SC as in the previous experiment, except that the volume of organic matter was small enough to process each sample in a single tube. To extract finally the microarthropods by HF (Figure 2-1C), the 250cc sample of mineral soil was suspended in 4 liters water and then decanted through a 400-mesh sieve. The collected organic matter was washed into a round 1000 ml flask and resuspended in 500 ml. About 25 ml of heptane was added to the

soil-water mixture, stirred for 2 minutes to allow the microarthropods to come into suspension in the heptane layer. Distilled water was slowly added until the heptane was in the neck of the flask.

With a dipper-ladle, the organic phase and part of aqueous phase was collected and poured through a 400-mesh sieve. The collected fauna and organic matter were rinsed several times with ethanol 95% to remove the excess of Heptane and transferred to a vial in ethanol suspension.

Microarthropod specimens were identified with light microscopy at the level of subclass for Collembola, order for Protura and Diplura, and suborder for Acari mites. Three ecological indices were estimated from the six samples common to each method: species richness (number

′ 푠 of species, S); Shannon–Weaver diversity index, 퐻 = − ∑푛=1 푝퐼 푙표푔푒 푝퐼 where pi is the proportion of species i (Pielou, 1975); and Simpson's (1949) index of dominance, 퐷′ = 1 −

푠 2 ∑푛=1 푝퐼 . Comparison of mean differences for BF and SC (n=15) were by T- test. Wilcoxon

Sign Ranked Test was used to assess differences in means derived from all three extraction methods (n=6). Proportional representations of each taxa in a sample were created using

JMP® (SAS institute) software.

Results

Sucrose Centrifugation and Flotation-Berlese-Flotation Comparison

The microarthropod extraction efficiency of SC was 88% higher than that of FBF in the first experiment. Fifty-five percent more mites (P=0.003) and 177% more collembola

(P=0.0004) were recovered using SC than FBF (Figure 2-2).

Sucrose Centrifugation, Berlese Funnels and Heptane Flotation Comparison

Approximately five times fewer microarthropods were recovered by BF compared SC in the second experiment. Mesostigmata, Prostigmata, Oribatida, Endeostigmata, Protura, Diplura,

Collembola, and Astigmata were recovered by both methods. Sucrose centrifugation was more

efficient than BF (N=15; P=0.001) in recovering all but one of the eight orders identified (Figure

2-3).

In the third experiment in which subsamples from just 6 of the 15 sites sampled in experiment 2 were extracted, consistently more microarthropods were recovered by HF than by

SC or BF. Berlese funnels recovered fewer animals than HF in five of the eight taxa and fewer than both other methods in four taxa. Whereas HF recovered larger numbers of microarthropods than did SC for seven of the taxa, and twice as many overall, the differences were not significant

(Figure 2-4).

Linear regressions (N=48) of the numbers of mites of each taxon (log X+1) recovered by each method resulted in no relationship between BF and HF (P=0.721), a weak positive trend for

BF and SC (P=0.0325), and a very strong relationship between HF and SC (P<0.0001). This was reflected by the structure of the microarthropod communities derived from HF and SC, which were more similar to one another than to that from BF. In particular, BF failed to recover prostigmatids and especially endostigmatids that comprised 14% of communities recovered by both HF and SC (Figure 2-5). The ecological indices S’, H’ and D’ estimated with data from HF and SC were congruent, whereas those from BF were lower (S’, P=0.0033; H’, P=0.0044; D’,

P=0,0061) in all cases (Figure 2-6).

Discussion

A variety of methods exist to extract nematodes from soil, primarily to recover economically important species which differ in the efficiency with which they are recovered by a given method (Southey, 1970). To our knowledge, no extraction method has been recommended for recovery of both microarthropods and nematodes (McSorley and Walter, 1991). Of the four methods studied here, those used primarily by acarologists recovered almost no nematodes. By

contrast, the most commonly used nematode extraction method was surprisingly effective in recovering microarthropods.

According to Petersen and Luxton (1982), passive extraction methods are more efficient than active methods such as Berlese funnels, as they can extract also inactive and sluggish microarthropods. Nevertheless, Berlese funnels are the most frequently used extraction method for soil microarthropods. Over 90% of selected acarology studies utilized Berlese funnels, despite reportedly poor efficiency in recovering certain taxa and immature stages (Andre et al.,

2017). Most of those studies restricted sample depth to 10cm, because of low recovery efficiency in mineral soil compared to flotation methods (Ducarme et al., 1998). This is consistent with our finding that BF recovered about 20% as many microarthropods as did SC.

The inefficiency of Berlese funnels for recovery of microarthropods in deeper soil layers resulted in numerous modifications such as the FBF technique. Flotation-Berlese-flotation separates animals and organic matter from a large volume of mineral soil, and then relies on motility to recover just the microarthropods (Arribas et al., 2016). Consequently, in comparison to SC, FBF still suffers from the poor extraction efficiency of the simple Berlese funnel method.

Moreover, the FBF requirement of large (2000 ml) soil volumes creates practical problems in handling the samples compared to other methods.

Despite shortcomings, Berlese funnels produce clean samples for microscopy and molecular processing. The large quantity of organic debris in the HF product is a serious impediment, despite the exceptionally high recovery of animals with this technique. The excessive impurities not only make identification and counting laborious and time consuming, they are likely to interfere with using toolkits for DNA extraction.

The consistently higher numbers of microarthropods captured by HF compared to SC suggest that the lack of significant differences was due to inadequate replication. Nevertheless, the relative recovery of the various taxa by HF and SC was highly correlated, and the community structure and diversity reflected by the two methods were nearly identical. This analytical congruence, combined with cleaner samples that facilitate counting or molecular analysis and the fact that only SC will recover nematodes, make it an ideal extraction method for studying both groups of animals. Moreover, the loss of saprophytic soil fungal and bacterial propagules during the sieving and rinsing process of SC is advantageous because primarily fungi and bacteria closely associated with nematodes, mites and collembola are retained. This property has been exploited in studies that utilized qPCR tools to estimate the occurrence and dynamics of fungal and bacterial natural enemies of entomopathogenic nematodes (Campos-Herrera et al., 2014;

Pathak et al., 2017). Sucrose centrifugation appears to be uniquely suited to study species assemblages affecting soil nematodes, given the breadth of natural enemy guilds it can capture.

Figure 2-1. Process flow diagrams for four extraction methods. A) Flotation-Berlese-flotation and simple Berlese device, B) Sucrose centrifugation and C) Heptane flotation.

Figure 2-2. Efficiency of sucrose centrifugation (SC) compared to flotation-Berlese-flotation (FBF) for extracting Acari mites and Collembola from 2L mineral soil samples. Differences between taxa abundances evaluated by T-test (N=12. P<0.001,***). Data presented as mean ± standard error.

Figure 2-3. Efficiency of sucrose centrifugation (SC) compared to that of Berlese funnels (BF) in extracting eight microarthropod taxa from 250 cc mineral soil sample. Wilcoxon nonparametric multiple comparisons used to test the differences between the methods. Data are presented as mean ± standard error test (N=12; P<0.05,*; P<0.01,**; P<0.001,***).

Figure 2-4. Extraction efficiency of Berlese funnels (BF), sucrose centrifugation (SC) and heptane flotation (HF). Bars and error bars denote means and 95% confidence intervals respectively. Means that are significantly different in multiple comparisons using Wilcoxon test are represented by different letters above bars (P < 0.05).

Figure 2-5. Proportional composition of microarthropod taxa. Pie charts represent recovery from Berlese funnels (BF), sucrose centrifugation (SC) and heptane flotation (HF).

Figure 2-6. Ecological indices (species richness, S’; Shannon Diversity Index, H’; dominance, D’) from samples extracted with three methods; Berlese funnels (BF), sucrose centrifugation (SC) and heptane flotation (HF). Bars and error bars denote means and 95% confidence intervals respectively. Means that are significantly different in multiple comparison using Wilcoxon test are represented by different letters above bars (P < 0.05).

CHAPTER 3 COMPARING HIGH THROUGHPUT SEQUENCING AND REAL TIME QPCR FOR CHARACTERIZING ENTOMOPATHOGENIC NEMATODE BIOGEOGRAPHY

Chapter summary: Entomopathogenic nematodes (EPNs) are widely distributed in soils across all continents except Antarctica. Assessing the EPN community structure in an ecoregion can help reveal their biological control potential against important crop pests. Common methods for detecting EPNs in soil samples include baiting with sentinel insects, direct observation of extracted nematodes, or use of species-specific primer-probe combinations using qPCR. Less well studied is the use of high throughput sequencing (HTS), which has tremendous potential to characterize soil communities of EPNs and natural enemies of EPNs. Here, for the first time, we compared qPCR and HTS to characterize EPN food webs. The frequency and abundance of 10

EPN species and 13 organisms associated with EPNs from 50 orchard and natural area sites in two ecoregions of Portugal were evaluated using qPCR tools, and results were published in 2019.

We applied an HTS approach to analyze frozen DNA samples from 36 sites in that study.

Universal primers targeting ITS1 were used for nematode detection. All EPN species detected by qPCR were also detected by HTS. The EPN species and nearly all free-living nematodes detected by both processes were highly correlated (P < 0.01). Steinernema feltiae, the dominant

EPN species, was detected by HTS in 55% more sites than by qPCR. HTS also detected more

EPN species than did qPCR. Sample accuracy, measured by the fit of Taylor’s Power Law to data from each method, was significantly better using HTS (r2=0.95, P < 0.01) than qPCR

(r2=0.76, P < 0.01). The effect of biotic and abiotic variables on individual EPN species did not differ according to ANOVA and multiple regression analyses of both data sets while the drivers

Republished with permissions from: Dritsoulas, A., Campos-Herrera, R., Blanco-Pérez, R., Duncan, L.W., 2020. Comparing high throughput sequencing and real time qPCR for characterizing entomopathogenic nematode biogeography. Soil Biology and Biochemistry 145. doi:10.1016/j.soilbio.2020.107793

of EPN community structure did not differ when analyzing either data set with CCA. Our results combined with decreasing costs of metabarcoding, suggest that HTS may provide the most cost- effective and accurate means of assessing soil food webs of methods currently available.

Introduction

Entomopathogenic nematodes (EPN) in the genera Steinernema and Heterorhabditis, have been the subject of extensive research for more than a half century, due to their potential as biocontrol agents of many pest insects. Much of this work is oriented toward utilizing EPNs in either an inoculative (Parkman et al., 1993; Shields et al., 2009) or an inundate release strategy as a biopesticide (Duncan et al., 1996; Shapiro-Ilan et al., 2015). Development of mass production technology and easy-to-use formulations led to the expanded use of EPN and modest commercial successes in some markets (Dolinski et al., 2012; Georgis et al., 2006).

Each EPN species is symbiotically associated with a specific entomopathogenic bacterial species. These nematode-bacteria complexes have insecticidal effect against a broad range of insect hosts (Kaya and Gaugler, 1993). The nematodes infect the insects through body openings or by penetrating the cuticle, then release the symbionts from the nematode intestine. Insects are killed in a few days by septicemia, after which nematodes and bacteria reproduce within the cadaver. Similar to many nematode species, in response to harsh conditions such as overpopulation and resource depletion, EPN development arrests at a modified third-stage juvenile, a stress-resistant stage called “dauer” or “infective juvenile” (IJ). The IJ is the only free-living stage capable of exiting the cadaver in search of new hosts.

Use of EPNs in IPM programs requires proper fit of the nematode species to the cropping system and target pests (Shapiro-Ilan et al., 2006), preservation of genetic variability and properties such as persistence and virulence of the populations in cultures (Neumann and

Shields, 2011), and use of cultural practices beneficial to EPN functioning in the soil food web

(Stuart et al., 2008). However, due to the cryptic nature of soil communities, relatively little is known about the magnitude of biological control provided by naturally occurring EPNs, or methods to exploit these services. Recent surveys of native EPNs and their natural enemies, combined with field experiments, identified some physical (Campos-Herrera et al., 2014, 2013b;

El-Borai et al., 2016) and biotic (Campos-Herrera et al., 2019a) soil properties that potentially modulate the behavior of EPN populations and their contribution to pest suppression (Campos-

Herrera et al., 2019b, 2014, 2013b; El-Borai et al., 2016).

Correct identification of EPN species is critical to understand observations made in ecological studies; however, morphological diagnostic methods are laborious and require taxonomic expertise. Indeed, IJ EPNs generally do not have adequate morphological characters for absolute species identification. Sentinel insects are typically employed to recover EPNs from soil samples, but infection is dependent on the susceptibility of the insect to particular EPN species, as well as environmental conditions such as soil moisture, temperature and porosity

(Stuart et al., 2006). Real time, quantitative PCR has proven useful for studying EPN community structure directly from mass samples of nematodes extracted from soil (Campos-Herrera et al.,

2015, 2013a; Spiridonov et al., 2007). The main principles of the method are the design of species-specific primers/probe for each species, and development of standard curves from pure cultures for quantification (Braun-Kiewnick and Kiewnick, 2018; Torr et al., 2007).

Technological developments continue to shift the predominant approaches to species identification for soil communities. The rapidly decreasing cost of gene amplicon sequencing in high-throughput (HTS) or next-generation sequencing (NGS) has numerous applications in soil and nematode community analysis. HTS of nematode communities has the potential to provide increased taxonomic resolution and capture rare taxa that are missed using qPCR, or

misidentified through morphological analysis (Treonis et al., 2018). Nevertheless, diversity assessment with species-level resolution remains an unresolved aspect of HTS. Potential errors and artefacts can arise at any step of the process: DNA amplification is limited by primers design, amplification efficiency can be species-specific, and formation of chimeric molecules can occur, especially when data include large numbers of unknown sequences (Porazinska et al.,

2009). The resolution of different marker genes (Blaxter, 2003), clustering differences between bioinformatic pipelines (molecular operational taxonomic units “MOTUs” or amplicon sequence variants “ASVs”) and the availability of a high-quality reference database for species level taxonomic identification are additional challenges (Blaxter, 2003; Callahan et al., 2017; Qing et al., 2019).

The objective of this study was to compare two molecular approaches to detect EPN species: the qPCR and HTS using ITS1 sequences. We applied a HTS, metabarcoding approach using Illumina MiSeq sequencing platform to analyze frozen DNA samples from a previous survey (Campos-Herrera et al., 2019a). In that survey, the frequency and abundance of 10 EPN species and 13 organisms associated with EPNs from 50 orchard and natural area sites distributed between two ecoregions in Algarve (Portugal) were evaluated using qPCR tools. We tested the hypotheses that: (1) HTS can quantify soil organisms with comparable accuracy to that of qPCR and (2) the species detection threshold is lower for HTS than that for qPCR.

Materials and Method

Samples Collection

Soil survey methods and qPCR survey results were given by Campos-Herrera et al.

(2019). Briefly, 100 soil samples were recovered from 50 sites in the Portuguese Algarve region, comprising citrus, pine, palmetto and oak as the dominant plant species. Nematodes were extracted with sucrose centrifugation and then DNA was extracted using the PowerSoil® DNA

Isolation Kit (MoBio) for that survey purposes. The same DNA samples used by Campos-

Herrera et al. (2019) were shipped on dry ice to the University of Florida facilities where they arrived in chilled condition. Not all the frozen samples we received had enough DNA for the initial step of Library preparation and subsequently, only 56 samples from 36 sites with adequate total amount of extracted DNA from both or at least one of the two samples were processed

Library Preparation

For HTS purposes and the proper identification of target organism, ribosomal DNA and

ITS region were amplified (average length >730bp) from bulk DNA extractions using universal primers TW81 (5’- GTTTCCGTAGGTGAACCTGC-3’) as forward primer and AB28 (5’-

ATATGCTTAAGTTCAGCGGGT-3’) as reverse primer (Iqbal et al., 2016). Primers were modified to include an overhang adapter sequence to enable sequencing, following the Illumina protocol for the 16S rRNA gene sequencing in microbial samples (16S Library Preparation

Protocol at http://support.illumina.com).

Library preparation consist of four parts: (i) amplicon PCR, (ii) amplicon PCR cleanup,

(iii) index PCR, and (iv) index PCR cleanup. Primarily, all samples normalized in 5 ng/ml DNA concentration and amplicon PCR consisted of initial denaturation 95°C for 3 min, twenty-five cycles of denaturation at 98°C for 30s, annealing at 55°C for 30s, elongation at 72°C for 60s, and terminal elongation at 72°C for 10min. A single 25 μL PCR reaction contained 2.5 μL of template of 5 ng/μL (12.5 ng total), 12.5 μL of 2x KAPA HiFi HotStart ReadyMix (KAPA biosystems), 1μL of each 10 μM overhang primer, 8 μL of 10 mM Tris pH 8.5. Positive controls consisting of DNA extracted from a laboratory culture of the nematodes Steinernema feltiae and

Heterorhabditis bacteriophora was also amplified, and negative controls consisting of purified, nuclease-free water were included for each PCR reaction. PCR products were verified on 0.8% agarose gels after staining with SYBR™ Safe DNA Gel Stain. PCR products were purified with

1.0× Agencourt AMPure XP beads (Beckman Coulter, Brea, CA) incubated for 5 minutes at room temperature, washed twice with 80% ethanol, and eluted in 50μL of 10 mM Tris pH 8.5. A second PCR added a specific index sequence to the amplicons for sample discrimination.

Amplicons were used as template for a limited cycle PCR amplification to add dual-index barcodes and the P5 and P7 Illumina sequencing adapters (Nextera XT Index Kit [FC‐131‐1001];

Illumina, San Diego, CA, USA). The conditions were initial denaturation at 95°C for 3 min, 8 cycles of denaturation at 98°C for 30s, annealing at 55°C for 30s, and elongation at 72°C for 30 sec and a terminal elongation at 72°C for 10 min. Each 50 μL PCR reaction tube contained 5 μL of template, 25 μL of 2x KAPA HiFi HotStart ReadyMix (KAPA biosystems), 5 μL of Index

Primers (N7XX), 5 μL of Index 2 Primers (S5XX). Finally, for the second clean up, PCR products were purified with 1.1× magnetic beads, eluted in 25 μL of 10 mM Tris pH 8.5. After library preparation, each individual library was quantified using Qubit 3.0 fluorometer the dsDNA BR kit (Life Technologies) and according to the average library size, the libraries were normalized in equal molar concentrations of 4nM and pooled together in a single library in aliquots of 5μL. The library was sequenced using 2 × 300 bp paired-end Illumina sequencing on the MiSeq platform at the Interdisciplinary Center for Biotechnology Research (ICBR) of

University of Florida.

Bioinformatics

The data we received from the sequencing facility was already demultiplexed with the

Illumina adapters trimmed and data separated by barcodes into respective sample identification codes. Initial quality assessment of each read was checked FASTQC v0.11. (Andrews et al.,

2015), and all the quality information for individual read assessed were combined into a single viewable document using MULTIQC (Ewels et al., 2016). Due the average size of the target locus that was more 730bp, the subsequent reads did not meet the set merging criteria. We

therefore used only the forward reads R1 for the bioinformatic pipeline. R1 reads derive from

ITS 1 of ribosomal DNA which was used for the nematode identification. The resulting data set was de-replicated with the ASV-based approach, in which we used the DADA2 method for denoising, through QIIME2 v2019.4 pipeline, including removal of primer sequences, truncating sequences by length and removing chimeric sequences with a de novo approach too (Callahan et al., 2016), which resulted in a length of 250bp . We then generated count tables by mapping raw

ASVs, assigning taxonomy generating input files for taxonomy assignment in QIIME2 from the

NCBI database. The standalone database was generated including all the non-redundant nucleotide sequences from all traditional divisions of GenBank, EMBL, and DDBJ excluding

GSS,STS, PAT, EST, HTG, and WGS (ftp://ftp.ncbi.nlm.nih.gov/blast/db ; nt.*tar.gz) employing an NCBI command-line tool to run BLAST, called BLAST+, integrated directly into our workflow.

Statistical Analysis

Campos-Herrera et al. (2019a) analyzed the means of two composite samples at each site with an area of ∼0.5 ha. In the current study, samples from the 36 sites with adequate total amount of extracted DNA were analyzed differently according to whether one or two samples were available. The means were calculated for each site if both samples were available (20 sites), whereas information from a single sample was used for sites with just one sample (16 sites). The number of EPN copies measured in 12.5 ng DNA was adjusted based on the total amount of extracted DNA and then transformed to log (x + 1) for statistical analyses. Nematophagous fungi

(NF) data and bacterial data reported by Campos-Herrera et al. (2019a) were transformed to square root (x) and log (x+1), respectively. One-way ANOVA and Tukey’s HSD test were used to assess differences in the population densities of the most abundant EPN species found in the different types of vegetation. Taylor’s Power Law was fitted to data for the most abundant

nematodes by regressing the log-transformed sample variances against log-transformed means to assess sample measurement reliability for the two data sets (qPCR and HTS) using the 20 sites with adequate DNA from both samples for HTS (Duncan et al., 2001). Nonparametric

Spearman’s Correlations were calculated between abundant nematode species measured by each method (JMP® Pro, v14.1.0.; SAS Institute Inc., Cary, NC). Stepwise multiple regression

(backward elimination; alpha = 0.15 for removal) of nematode species against selected soil organisms and abiotic properties from (Campos-Herrera et al., 2019a) was performed using data from each method (Minitab, v. 17.3.1; State College, PA). Canonical correlation analysis was used to identify and measure the associations among explanatory and response variables setting orthogonal linear combinations of the variables within each set that best explain the variability.

First, we used Pearson correlations (R) to selected soil properties to avoid variables with strong collinearity. Selected abiotic factors were included as explanatory variables or predictors. Tests of dimensionality for the canonical correlation analysis, was employed to indicate the canonical dimensions that were statistically significant at the 0.05 level. The graphical results of the CCA were presented with bi–plot scaling (R Development Core Team, ‘Vegan’ package).

Phylogenetic Tree

For phylogenetic analysis, the newly obtained ITS1 sequences were aligned using the

CLUSTAL W multiple alignment program (Thompson et al., 1994). Maximum likelihood (ML) analysis was performed with the program PhyML (Guindon and Gascuel, 2003) provided on the

“phylogeny.fr” website (http://www.phylogeny.fr/). The probability of inferred branch was assessed by the approximate likelihood ratio test (aLRT), an alternative to the nonparametric bootstrap estimation of branch support (Anisimova and Gascuel, 2006). Steinernema citrae

(Steinernematidae; DDBJ/EMBL/GenBank accession no. MF536116.1) was used as an outgroup for the construction of a phylogenetic tree (Figure 3-1).

Results

Metabarcoding analysis after removing chimeras recovered 23578 unique amplicon sequence variants (ASVs) from twenty-one eukaryotic phyla, with 10.6% (2212 ASVs) assigned to the Nematoda. Unique nematode ASVs were found, encompassing 26 nematode families while more than half (1177) were identified to genus level or below, setting a threshold of 80% coverage. From all nematode ASVs, eighteen were identified as entomopathogenic nematodes.

Seven nematode species of interest were identified by HTS; four EPN species and four free living bacteriophagous nematodes (FLBN) in the genus Oscheius, two of which were associated with EPNs in previous studies (Campos-Herrera et al., 2015b; Ye et al., 2018).

Oscheius onirici and O. tipulae were detected by both HTS and qPCR, but O. dolichura and O. myriophyla were only detected using HTS, because primers-probes were not designed and not used for these two species. Eighty ASV’s matching species in the Acrobeloides genus were recovered by HTS; however, blast results did not provide unambiguous identification, even though the query sequence was 100% identical to the reference sequence of the database. HTS did not detect any of the Pristionchus species detected by qPCR (Table 3-1).

Four EPN species were detected by HTS, compared to just two using qPCR. It is unremarkable that Heterorhabditis megidis was found only using HTS, because primers-probe for this species were not employed by Campos-Herrera et al., (2019a). Of the remaining EPN species detected by HTS, S. affine, failed to amplify using qPCR. The relative population densities of the four nematode species detected by both methods (S. feltiae, H. bacteriophora, O. tipulate and O. onirici) were significantly (P < 0.0001) correlated, with between 69%-95% of the variability explained by the models (Table 3-2).

The detection frequency provided by HTS was generally higher than that from qPCR.

The most commonly encountered EPN S. feltiae was detected by HTS 62% more frequently (34

samples, 60.7%) than by qPCR (21 samples, 37.5%). NGS detected S. feltiae in all 13 samples from palmetto, 92.3% of 13 samples of oak, 41% of 7 citrus samples, and in 77% of 13 samples from pine. By contrast, qPCR detected S. feltiae in just 69% of oak habitats, in fewer than half and a quarter of the samples from citrus and pine, respectively, and did not detect S. feltiae in any palmetto habitat (Figure 3-2). Oscheius onirici was detected about twice as often (9 samples) by NGS compared to qPCR (4 samples) and O. tipulae, the most frequently encountered species was found in all but seven NGS samples (96%) compared to all but 6 qPCR samples (89%)

Despite the effect of detection frequency on the estimated relative abundance of the nematode in different vegetative habitats, results of ANOVA did not differ using either NGS or qPCR measurements (Figure 3-3). Plant species were shown to have a highly significant effect (P <

0.01) on S. feltiae abundance when data from both methods were subjected to one-way analysis of variance, with oak habitat supporting more S. feltiae than pine, palmetto or citrus orchards

(Figure 3-3).

Taylor’s law explained 95% of the variability in the S. feltiae variance-mean relationship derived from HTS sample statistics (HTS P19 < 0.001), compared to 76% for the qPCR-derived statistics (qPCR P19 < 0.001) (Figure 3-4). However, the O. tipulae variability explained by qPCR (68%) was more than 80% higher than that explained using NGS data (37%) (not shown).

The slope of the regression line using HTS indicated that S. feltiae population is highly aggregated whereas that using qPCR suggested a more random spatial pattern (Figure 3-5).

Data from the two detection methods resulted in similar regression models of EPN against the biotic and abiotic habitat variables. For S. feltiae, elevation, electrical conductivity, soil moisture, pH and P level were significant abiotic variables using either data set. The NF

Arthrobotrys oligospora was inversely related to S. feltiae measured by NGS and Hirsutella

rhossiliensis was inversely related to the nematode using both data sets. Oscheius onirici was positively associated with S. feltiae measured by qPCR. More variation in the data was explained by NGS measurements (60%) than by those of qPCR (35%) (Table 3-3). Regression models for

Oscheius tipulae were also very similar for the two data sets, with pH silt and clay, A. oligospora and Purpureocillium lilacinum being significant biotic and abiotic variables using either data set.

The models explained 59% of the variability for NGS data and 17% for those from qPCR.

Paenibacillus sp. was not found to be related to any EPN species, but the total amount of the bacterium and the bacterial encumbrance rate per S. feltiae (log Paenibacillus sp. – log S. feltiae) were positively associated with soil pH, which ranged between 4.2 and 8.18 (Table 3-3).

Canonical correspondence analysis (CCA) of data from qPCR explained a greater amount of the total variation than did data from HTS (Figure 3-3). The CCA identified three significant abiotic variables: pH, clay and elevation (P < 0.05). The same variables were most important for the HTS model, and model significance was nearly the same using either data set (Table 3-4).

Blasting results of HTS data revealed that species such as S. feltiae and O. tipulae comprise multiple ASVs (Table 3-5), exhibiting patterns referred to as “head-tail” by Porazinska et al. (2010).

Discussion

Characterizing EPN biogeography with a goal of conservation of biological control requires fine-scale taxonomic resolution, because closely related EPN species can exhibit highly divergent phenotypes for key properties such as habitat adaptation (El-Borai et al., 2016) and insect host specificity (Lewis et al., 2006; Peters, 1996; Simões and Rosa, 1996). Here we showed that HTS technologies, targeting the rDNA ITS region, can achieve the required resolution within this nematode guild. This is the first report in which HTS was used to identify natural communities of EPN species.

Both original hypotheses were supported by the data. Using DNA from nematodes and other organisms extracted from soil samples, the HTS measurements of S. feltiae population density were highly correlated with those obtained previously from qPCR (Campos-Herrera et al., 2019) and the detection frequency was significantly higher using HTS. Blasting the sequencing data did not reveal any of Pristionchus spp. reads, probably because of the primers’ limitations. In the case of Acrobeloides, multiple ASVs were identified as Acrobeloides sp.; however, blast results did not provide unequivocal identification of the genus, suggesting either that ITS1 is not an informative region or there are erroneous reference data. These results explain why Campos-Herrera et al., (2012) designed qPCR probes targeting 18S SSU region for

Acrobeloides-group detection. In the case of S. feltiae, a frequent lack of detection by qPCR resulted in apparent differences in the capacity of EPNs to colonize some vegetation habitats with few samples (e.g., palmetto); however, the differences between the two methods were not significant when discriminating the relative habitat preference among these plants. The intraspecific variability of the S. feltiae ITS region may account for the apparently lower detection limit of HTS compared to qPCR. Primers and probes designed for a species in one region may be relatively strain-specific (Spiridonov et al., 2004). Additionally, qPCR may be less sensitive than procedures using the Illumina MiSeq sequencer. Approximately 10 copies per reaction are required for detection by qPCR (Forootan et al., 2017), whereas the threshold for

HTS is undetermined. Other comparative studies have also found MiSeq superior to qPCR, as well as HTS platforms such as Ion Torrent PGM and Roche-454, for detection of pathogens in mock samples (Frey et al., 2014).

We found conflicting evidence that the HTS measurements were more reliable than those of qPCR. Taylor's power law is an empirical law in ecology relating the variance and mean of

the number of individuals of a species per unit area of habitat. The fit of the regression line to the data provides evidence of the measurement reliability and the regression slope is a quantifiable measure of population dispersion. The slope using HTS measurements indicates that S. feltiae tended to be highly aggregated, whereas that from qPCR measurements revealed a tendency toward randomness. The fit of the model to the different data sets (R2=0.96 vs 0.76) supported the interpretation provided by the HTS measurements as the more likely property of this species.

Nevertheless, despite the good fit to the S. feltiae HTS data set, it is not apparent why neither qPCR nor HTS measurements of O. tipulae were especially well described by the power law.

The relationships measured between habitat properties and the two most frequently encountered nematodes, S. feliae and O. tipulae, were very similar for the HTS and qPCR data sets; however, HTS data provided stronger support as measured by the coefficients of determination for stepwise multiple regression. An inverse relationship between the EPN and two common nematophagous fungi suggests that some habitat properties may favor the predaceous fungi at the expense of the EPN. For example, soil moisture was weakly inversely related to both fungal species (P < 0.1; not shown) and may partly account for greater abundance of S. feltiae in the wetter soils.

The multiple regression relationships between S. feltiae, Paenibacillus sp. and pH especially warrant additional study. Paenibacillus sp. was described as an ectoparasitic bacterium specific to Steinernema diaprepesi that exhibited properties of density dependent regulation of the nematode in laboratory experiments (El-Borai et al., 2005) and temporal surveys in the field (Campos-Herrera et al., 2019b). Basically, the bacterium adhered to the cuticle and impeded movement and host-finding of S. diaprepesi at high soil pH but detached from the cuticle at low pH. Campos-Herrera et al., (2019a) speculated that the detection of the

bacterium in Algarve in the absence of S. diaprepesi indicated a lack of specificity by the primers-probe set or that the bacterium is less host specific than reported. If this bacterium in the

Algarve is associated with EPNs, S. feltiae would be a possibility. The observation that qPCR detected S. feltiae in only half of the sites inhabited by Paenibacillus sp. does not support the likelihood of a close association between the two, whereas HTS detected S. feltiae at all but one of the 16 sites where the bacterium was found. Moreover, both S. feltiae and the bacterium were associated with soil pH in the same manner as reported in previous laboratory and field experiments (Campos-Herrera et al., 2019b, 2014, 2013b; El-Borai et al., 2005). EPN abundance was inversely related to soil pH while the bacterium was highly positively associated with pH both in total abundance and when expressed as spore abundance per S. feltiae abundance (i.e. spore encumbrance rate).

Based on our comparative results, HTS seems preferable to qPCR for community analyses for multiple reasons. The cost of HTS here was about twenty percent higher than that of qPCR. However, HTS potentially reveals everything in a soil sample that can be amplified by universal primers, whereas qPCR found only that which was sought. Heterorhabditis megidis was found only by HTS, because qPCR was not attempted for this species. H. megidis is a cosmopolitan species, occurring worldwide in temperate regions from North America to Asia, with an apparent preference for turf and weedy habitats (Stock and Kaya, 1996; Stuart and

Gaugler, 1994). . It has been reported also in some Mediterranean countries like Greece (Menti et al., 1997), Turkey (Yilmaz et al., 2009) and Israel (Glazer et al., 1993), but this is the first report of H. megidis in Portugal.

The specificity of qPCR primers-probes is probably sometimes excessive for communities with significant intraspecific variability. Blasting the results of HTS suggested that

most species consist of multiple ASVs. Porazinska et al., (2010), described species such as S. feltiae as exhibiting “head-tail” patterns where a single head ASV perfectly matches a NCBI reference sequence and comprises the majority of the sequencing reads (“head” formed 99% of the S.feltiae reads) with tail ASVs represented by just a few sequences¸ revealing the real ITS 1 variation of a species. A phylogenetic tree verified that the eight S. feltiae ASVs probably belong to the same species because they are more closely related to each other than to any other closely related species. Literature and NCBI database survey, suggested that the closest related species to S. feltiae is the Steinernema citrae which was used as a root to our phylogenetic tree (Figure

3-1).

EPNs are a well-studied guild and the information included in the NCBI database, especially for the ribosomal gene, is relatively good compared to other groups. Nevertheless, an ongoing challenge of HTS and all molecular survey methods is the quality of reference databases, which contain mistaken identities and taxonomic gaps for known and undescribed species. HTS reveals these questionable sequences for further study, whereas they remain undetected by qPCR. In this study, S. affine consists of two ASVs which exhibited 98.1% and

98.0% identity to a reference sequence. Further study of the populations can resolve whether the

2% difference between the query and the reference sequences reveal an undescribed species or intraspecific variation of S. affine.

The congruence of results from those species relevant to both the qPCR and HTS tools used here, support the use of HTS for soil community analyses at the species level. The capacity of HTS to measure infinitely more species at a lower cost than can be achieved by qPCR will ensure wide adoption of metagenomic methods and hasten our understanding of EPN biogeography and those factors that modulate EPN presence and abundance.

Table 3-1. Detection frequency of entomopathogenic nematodes (EPNs) and species previously reported to EPN competitors which were detected by metabarcoding and qPCR, (percent of 56 samples). Metabarcoding qPCR EPN Steinernema feltiae 60.7 37.5 Steinernema affine 3.5 Heterorhabditis bacteriophora 7.1 3.5 Heterorhabditis megidis 1.7 Competitors Acrobeloides-group 62.5 Oscheius onirici 16 7.1 Oscheius tipulae 87.5 89.2 Pristionchus maupaci 3.5 Pristionchus pacificus 30.3

Table 3-2. Non-parametric Spearman's correlations between the species’ measurements from metabarcoding and qPCR Species Spearman ρ Prob>|ρ|

Steinernema feltiae 0.8737 <.0001 Heterorhabditis bacteriophora 0.7201 <.0001 Oschieus tipulae 0.9026 <.0001 Oschieus onirici 0.6393 <.0001

Table 3-3. Significant variables from stepwise multiple regression of two nematode species measured by high throughput sequencing or qPCR and a bacterial species measured by qPCR, regressed against soil properties and potential biotic antagonists. Paenibacillus Paenibacillus Species S. feltiae O. tipulae sp. abundance sp. encumbrance Approach HTS qPCR HTS qPCR R2 / R2 adjusted 0.70/0.60 0.50/0.35 0.69/0.59 0.33/0.17 0.61/0.50 0.69/0.55 A. oligospora -0.02 0.037 0.005 H. -0.01 -0.04 -0.004 0.01 rhossiliensis P. lilacinus 0.001 0.06 O. onirici 0.05 Elevation -0.001 -0.008 -0.001 0.001 Paenibacillus 0.021 0.014 sp

H20 0.05 0.015 0.003 -0.04 -0.02 EC 0.001 0.01 pH -0.02 -0.005 0.026 0.01 0.01 Sand 0.002 0.058 Silt 0.007 0.028 Clay -0.05 P 0.001 0.06 0.01 -0.007 K 0.005 0.001 Note: Independent variables included Oschieus tipulae, Oschieus onirici, Arthrobotrys oligospora, Arthrobotrys dactyloides, Purpureocillium lilacinum, Hirsutella rhossiliensis, Peanibacillus sp., elevation, soil moisture, pH, electrical conductivity, percent clay, silt, sand and organic matter, K, P, Mg, Zn, Fe.

Table 3-4. Significance of the Canonical Correspondence Analysis model, axes, and variables.

Metabarcoding qPCR Pr(>F) Pr(>F) Model 0.043 * 0.03 * CCA1 0.086 . 0.143 CCA2 0.53 0.198 pH 0.011 * 0.01 * Clay 0.064 . 0.014 * Elevation 0.163 0.023 * P 0.074 . 0.462 Note: Significance is also indicated by P<0.05=* ; P<0.1= ·.

Table 3-5. S. feltiae ASVs table illustrating “head-tail” structure associated with the presence of within the species variation. # identity % coverage e-value Description title No. reads Length 1 100 100 1.94E-126 Steinernema feltiae isolate H9 50100 250 Head 2 100 100 1.94E-126 Steinernema feltiae isolate I2 2415 250 Tail 3 99.6 100 9.02E-12 Steinernema feltiae isolate H9 1082 250 4 99.6 100 9.02E-12 Steinernema feltiae isolate 11A 686 250 5 99.6 100 9.02E-12 Steinernema feltiae isolate H9 517 250 6 97.22 100 2.54E-115 Steinernema feltiae isolate DONR 355 250 7 99.6 100 9.02E-12 Steinernema feltiae isolate H9 291 250 8 100 100 1.94E-126 Steinernema feltiae isolate 11A 76 250

Figure 3-1. Phylogenetic relationships of ASVs identified as S. feltiae based on sequencing reads of ITS_1 region as inferred by maximum likelihood. Steinernema citrae (MF536110) was used as an outgroup.

Figure 3-2. The frequency of S. feltiae (percent of 56 samples) through different types of vegetation was detected by metabarcoding and qPCR.

Figure 3-3. Box plots represent logarithmic number of reads detected by high throughput sequencing (A) and logarithmic number of individuals detected by qPCR (B) measuring Steinernema feltiae populations in Portugal. Differences in means are designated by boxes without the same letters and were determined by Tukey’s HSD test (P≤0.05) using log (X+1) transformed data.

Figure 3-4. Fit of Taylor’s Power Law to sample statistics for Steinernema feltiae populations measured using high throughput sequencing (HTS) (A) and qPCR (B).

Figure 3-5. Canonical correspondence analysis depicting biplots of the regional distribution and relationships between significant abiotic factors and soil organisms.

CHAPTER 4 NATURAL OCCURRENCE OF ENTOMOPATHOGENIC NEMATODES AND THEIR NATURAL ENEMIES IN GREEK CITRUS ORCHARDS

Chapter summary: Characterizing entomopathogenic nematode (EPN) biogeography with a goal of conservation biological control requires fine-scale taxonomic resolution, because closely related EPN species can exhibit divergent phenotypes for key properties such as habitat adaptation and insect host specificity. We used high throughput sequencing (HTS) to measure

EPNs and natural enemies of EPNs in 62 citrus orchards in 2 ecoregions in Greece. We designed improved primers targeting the ITS2 rDNA to discriminate EPN species, and previously published primers targeting 28S to identify acari mites. We also characterized 16 soil physico- chemical properties of each Greek orchard. The same nine EPN species (Steinernema glaseri, S. diaprepesi, S. khuongi, S. carpocapsae, S. feltiae, S. riobrave, Heterorhabditis bacteriophora, H. indica, H. floridensis) were detected in two Greek ecoregions (northeastern Peloponnese and northwestern Crete) . Species specific primers-probes were used to confirm HTS results.

However, qPCR amplification occurred only after an enrichment step was applied to the original

DNA samples, indicating the detection limit is lower with HTS than qPCR and the abundance of the majority of EPNs in Greece is very low. A significant linear relationship derived between the relative species differences of mean copy numbers from qPCR standard curves and the body volume of the nine EPN species (from Andrassy’s formula) provided a basis to estimate the numbers of individuals from the HTS copy numbers. Acaridae, Ascidae, Tarsonemidae and

Tydeidae were the most dominant families in the acari mite community structure in both regions, and Redundancy Analysis (RDA) revealed Tydeidae as a significant biotic variable for explaining variability of the EPN communities. A demonstrated lower detection limit than qPCR, and the capacity of HTS to detect all species in a sample without species specific tools make it an ideal method to study EPN biogeography.

Introduction

Entomopathogenic nematodes (EPN) in the families Heterorhabditidae and

Steinernematidae are obligate insect parasitoids and, as such, are a widely studied soil guild

(Kaya and Gaugler, 1993). They have been used in classical, conservation, and augmentative biological control programs targeting many insect species(Grewal et al., 2005) (Grewal, Ehlers,

& Shapiro-Ilan, 2005). For all that is reported about EPNs, relatively little is known about their natural occurrence in many parts of the world, or how those endemic species might be exploited to control insect pests. A few studies have documented increased biological control by endemic

EPN following soil modification (Duncan et al., 2013) or by augmentation using locally-adapted, persistent EPN (Shields et al., 2009). However, there are still fundamental questions to be answered about EPN performance in agricultural and other ecosystems. Most reviews emphasize the inconsistent efficacy of augmented EPNs against soil-dwelling pests in agricultural systems

(Georgis et al., 2006; Shapiro-Ilan et al., 2009), suggesting a dearth of knowledge about EPN interactions with soil properties and food webs. Several recent natural surveys and experiments have tried to address those deficiencies by using molecular tools to detect and quantify native

EPNs and some of their natural enemies in varied physical habitats (Campos-Herrera et al., 2013,

2014, 2019; El-Borai et al., 2016).

The study of EPN biogeography is often focused on characterizing naturally occurring populations that are locally adapted and therefore best suited to regulate insect pests (Hominick,

2002; Hominick et al., 1996; Stuart et al., 2006). Like most soil organisms, local EPN spatial patterns are typically patchy; however, they are widely distributed and virtually ubiquitous in soils across all continents except Antarctica. Hominick (2002) observed that EPN biogeography is characterized more accurately by species than genus. Global inventories currently show many species that are regionally restricted like S. cubanum and S. oregonense and others such as

Steinernema feltiae and Heterorhabditis indica with worldwide distribution. However, the known EPN geographic distribution is heavily biased by varying sampling efforts, with Europe being by far the most intensively sampled continent (San-Blas, 2013; Stock, 2005). Numerous

EPN surveys have been reported from the European Mediterranean countries of Portugal, Spain,

France and Italy, with the notable exception of Greece, where just Steinernema feltiae and

Heterorhabditis megidis have been reported.

The EPN detection frequency from environmental samples generally varies from less than 10% to 40%, although significantly higher occurrence has been reported (Raquel Campos-

Herrera et al., 2013; Emelianoff et al., 2008; Stuart and Gaugler, 1994) Much of this variation is likely a result of methodology rather than EPN abundance (Hominick, 2002). EPNs are commonly detected either by microscopy, baiting with host insects, or molecular techniques.

Identification through microscopy of nematodes extracted from soil, especially juvenile EPNs, which are particularly non-descript, requires deep expertise and abundant time. Baiting techniques using sentinel insect larvae have the advantage of recovering all EPN stages inside the cadaver, but is biased by host specificity and soil conditions (Fan and Hominick, 1991;

Koppenhöfer et al., 1998; Stuart et al., 2006). Real time, quantitative PCR has been used to detect EPNs in samples of nematodes extracted from soil for more than a decade (Torr et al.,

2007) and large numbers of species-specific primers are now available (Braun-Kiewnick and

Kiewnick, 2018). Although qPCR is very sensitive, primers and probes designed for a species in one region might be relatively strain specific. Recently, high throughput sequencing techniques were shown to detect EPNs from ITS 1 rDNA in environmental samples with greater sensitivity than that of qPCR (Dritsoulas et al., 2020). However, a drawback of the method was an inability to quantify the absolute number of individuals rather than their relative abundance.

High throughput sequencing tools such as metabarcoding also greatly facilitates characterizing soil food webs. Soil microarthropods including acari mites and collembola include natural enemies of nematodes that are relatively little studied, but which may significantly modulate the services of EPNs. The present paper describes the use of metabarcoding to characterize the distribution of EPNs and some of their natural enemies, including microarthropods, in citrus orchards in two Greek ecoregions. The objectives of this project were to 1) provide a more comprehensive map of EPN biodiversity in two important agricultural regions of Greece, and 2) investigate relationships between EPN species and biotic and abiotic soil properties that might affect their occurrence or abundance.

Materials and methods

Sampling

Sixty-two citrus orchards were surveyed during the summer of 2017 in two regions of

Greece - 32 sites from the region around Argos in the Peloponnese and 30 from the region around Chania on Crete. Two composite samples were collected from approximately 0.5ha at each orchard. Each sample comprised 15 cores (2.5cm dia x 30cm depth) collected from under the canopy of 15 trees (ca. 2200cm3 per sample). The samples were transferred in coolers to the laboratory where they were stored at 10 °C and processed within 7 days.

Each composite sample was gently mixed and then nematodes, soil microarthropods, and microorganisms present in or on the nematodes or other mesofauna were extracted by sucrose centrifugation from 250cm3 of soil (Jenkins, 1964). The contents from the two samples per site were combined, to provide organisms extracted from 500cm3 soil per site. Samples were concentrated in 95% ethanol in 15-ml Falcon™ 15mL conical centrifuge tubes, which were transported to the University of Florida and stored at 4°C until DNA extraction.

Soil analysis

A subsample of 250cc from each soil sample was analyzed (Waters Agricultural

Laboratories, Camilla, GA, USA) for pH, soil organic matter (OM), percentage sand, silt and clay, electric conductivity (EC), cation-exchange capacity (CEC) and soil nutrient elements (P,

K, Mg, Ca, and B). A second cluster of the same volume from each sample, was separated analyzed at soil science and agriculture chemistry laboratory of agriculture University of Athens

(AUA) for soil texture giving the percentage of sand, loam and clay.

DNA extraction

Based on the observation that samples in ethanol yield low DNA, following centrifugation and aspiration of excess ethanol, the tubes were refilled with 1xPBS (phosphate buffer saline) and incubated overnight at 4 °C. After a second centrifugation and aspiration of excess PBS, DNA was extracted with DNeasy® PowerSoil Kit (Qiagen).

Library preparation

The DNA extract concentrations of each sample were measured using the Qubit® dsDNA

High Sensitivity Assay Kit (Thermo Fisher Scientific, USA). Two groups of libraries were created for the entomopathogenic nematodes and microarthropods. For nematodes, the primers targeted 5.8S rDNA amplifying the ITS 2 region from bulk DNA using a de novo designed forward primer with average amplicon length 450bp for steinernematids and 350bp for heterorhabditids. The universal primers are AD58F (5’- TCGATGAAAAACGCGGCAA-3’) as forward primer and AB28R (5’-ATATGCTTAAGTTCAGCGGGT-3’; Curran et al., 1994) as reverse primer. For microarthropods, universal primers were used, targeting the D3-D5 region of

28S rDNA with forward primer 28Sa 5’- GACCCGTCTTGAAGCACG-3’ and reverse primer

28Sbout 5’-CCCACAGCGCCAGTTCTGCTTACC-3’ (Tully et al., 2006). Primers were modified to include an overhang adapter sequence to enable sequencing, following the Illumina

protocol for the 16S rRNA gene sequencing in microbial samples (16S Library Preparation

Protocol, http://support.illumina.com).

According to Illumina protocols, library preparation comprises four parts: (i) amplicon

PCR, (ii) amplicon PCR cleanup, (iii) index PCR, and (iv) index PCR cleanup. Samples were standardized at 5 ng/ml DNA concentration. For the EPN, samples were amplified with the following conditions: initial denaturation 95°C for 3 min, 28 cycles of denaturation at 98°C for

30s, annealing at 56°C for 30s, elongation at 72°C for 60s, and terminal elongation at 72°C for

10min. For microarthropods the conditions were initial denaturation 95°C for 3 min, 25 cycles of denaturation at 98°C for 30s, annealing at 55°C for 30s, elongation at 72°C for 60s, and terminal elongation at 72°C for 10min. For all libraries a single 25 μL PCR reaction containing

2.5 μL of template of 5 ng/μL (12.5 ng total), 12.5 μL of 2x KAPA HiFi HotStart ReadyMix

(KAPA biosystems), 1μL of each 10 μM overhang primer, 8 μL of 10 mM Tris pH 8.5. Positive controls consisting of DNA extracted from a laboratory culture of the nematodes Steinernema feltiae and Heterorhabditis bacteriophora and negative controls consisting of purified, nuclease- free water were included for each set of PCR reactions. PCR products were verified on 1% agarose gels after staining with SYBR™ Safe DNA Gel Stain. All PCR products were purified with 1.0× Agencourt AMPure XP beads (Beckman Coulter, Brea, CA) and eluted in 50μL of 10 mM Tris pH 8.5. At the following index PCR, amplicons were used as template for a limited cycle amplification to add dual-index barcodes, P5 and P7 Illumina sequencing adapters using

Nextera XT Index Kit [FC‐131‐1004] for EPN and XT Index Kit [FC‐131‐1001] for microarthropods (Illumina, San Diego, CA, USA). The index PCR conditions were initial denaturation at 95°C for 3 min, 8 cycles of denaturation at 98°C for 30s, annealing at 55°C for

30s, and elongation at 72°C for 30 sec and a terminal elongation at 72°C for 10 min. Each 50 μL

PCR reaction tube contained 5 μL of template, 25 μL of 2x KAPA HiFi HotStart ReadyMix

(KAPA biosystems), 5 μL of Index Primers (N7XX), 5 μL of Index 2 Primers (S5XX). The total number of 124 index PCR products purified with 1.1× magnetic beads, eluted in 25 μL and quantified using Qubit 3.0 fluorometer. Finally, libraries were normalized in equal molar concentrations of 4nM and pooled together in a single library in aliquots of 5μL. The library was sequenced using MiSeq 2 × 300 bp paired-end Illumina at the Interdisciplinary Center for

Biotechnology Research (ICBR) of University of Florida.

Bioinformatics

ICBR delivered raw data in fastq format which were demultiplexed and separated into respective sample identification codes. FASTQC v0.11 (Andrews et al., 2015) was used for quality assessment of each read, and then all the quality information was combined into a single viewable document using MULTIQC (Ewels et al., 2016). ITS 2 and LSU D3-D5 amplicons of ribosomal DNA was used for the nematode and microarthropods identification respectively. In both datasets, R1 and R2 reads combined and de-replicated with the ASV-based approach, in which DADA2 was the denoising method, through QIIME2 v2019.4 pipeline, including removal of primer sequences, truncating sequences by length and removing chimeric sequences with a de novo approach according to Callahan et al., ( 2016), which resulted in a length of 350-450bp for nematodes and 500bp for microarthropods. We then generated count tables by mapping ASVs, assigning taxonomy generating input files for taxonomy assignment in QIIME2 from the NCBI database. The standalone database was generated including all the non-redundant nucleotide sequences from all traditional divisions of GenBank (ftp://ftp.ncbi.nlm.nih.gov/blast/db/nr.gz) employing an NCBI command-line tool to run BLAST, called BLAST+, integrated directly into our workflow.

Real time PCR testing, cultures of EPN and standard curve preparation

Six of sixty-two sites were selected to be tested confirming the identification the metagenomic approach. The criterion of selection was that these samples comprise ASVs from all nine species which were identical to GenBank reference. The six samples were tested with qPCR applying species specific primers and probes, primarily in the DNA extractions and secondarily in enriched DNA after applying twenty cycles PCR with universal primers and cleaned up with magnetic beads.

Pure cultures of the nine EPN species detected through metabarcoding, were maintained in in the laboratory using Galleria mellonella as host; some them isolated from Florida citrus orchards and other were donated by other laboratories. The nematodes harvested in deionized water and stored at 15°C. For the standard curves preparation, exactly 300 infective juveniles

(IJs) were collected by pipetting and transferred to a 1.5ml Eppendorf tube. To avoid loss of IJs on the inner pipette surfaces, the tips were first washed in 2% Triton X-100. DNA of 300 IJs was extracted by using the DNeasy® Blood & Tissue extraction toolkit (Qiagen). The quality and quantity of DNA samples were measured using the Nanodrop System. DNA dilutions corresponding to 100, 30, 10, 3 and 1 IJs, were used to make standard curves to determine the number of sequence copies per IJ for each of the species.

The nine EPN species-specific primers and probes used for qPCR were previously reported (Campos-Herrera et al., 2011a,b)(Sup. Table 4-2). Primers and probes were synthesized by Integrated DNA Technologies Inc. (IDT, San Diego, CA). Real-time PCR was performed in optical 96- well reaction plates (USA Scientific, Orlando, FL, USA) on an ABI Prism 7000

(Applied Biosystem). All reactions were performed in final volume of 20 μL, contained 10 μL of

TaqMan Universal PCR Master Mix, 1 μL of DNA template and 200 nM of the corresponding

TaqMan® Probe for all the nematodes. Annealing temperature for all reactions was 59°C except

for H. floridensis which was 57°C. Primer concentrations were 400 nM for all steinernematids,

250 nM for H. bacteriophora and H. indica and 600 nM for H. floridensis.

Gene copies by individuals’ volume and data correction.

Real-time PCR results of the 5 DNA dilutions corresponding to 100, 30, 10, 3 and 1 IJs, were used to calculate a relative number of copies per IJ, where 2 was raised to Ct value power, the numbers inverted, and divided by the respective number of IJs. The mean of the numbers resulted the relative number of copies per IJ for the respective species. The body mass of nematodes was calculated using the Andrassy (1956) formula W = (L*D2)/(1.6*106) where W is the mass (as fresh weight (μg) per individual, L is the nematode length (μm) and D is the greatest body diameter (μm). The morphometric parameters were derived from literature (Nguyen and

Hunt, 2010). All numbers (copies and body size) were devided by the smallest value for both variables revealing the relative differences. The two variables were regressed, and the derived equation was used to adjust the numbers of ASVs to reflect the abundance of IJs by applying the relative species differences of body volume as values of the explanatory variable. The relative differences on body volume and the mean of rDNA copies were obtained by dividing the species values by the highest value.

Genetic analysis of identified ASV

Phylogenetic analysis was carried out in GENEIOUS 10.1.3 software. Each of the identified EPN species was evaluated using a unique tree derived from the metabarcoding process. Sequences were aligned using ClustalW alignment method on default settings, and subsequently UPGMA trees were constructed while the robustness of clades of UPGMA trees was assessed using 1000 bootstrap replications (supplementary files). Nine sequences of the closest related species to the proposed species identities were used as outgroups, confirming the identifications of ASVs retrieved from NCBI GenBank. The outgroup sequences were selected

according to previous studies describing these EPN relationships (Iqbal et al., 2016; Spiridonov et al., 2004; Stock et al., 2018). Two phylogenetic trees were selected, one for each of the

Steinernematidae and Heterorhabditidae (Figure 4-1,4-2), using the 5 most frequently detected

ASVs for each species (Table 4-1).

Statistical Analysis

All biotic variables analyzed were detected in more than 10% of the localities. Biotic factors were log transformed while the abiotic factors were standardized by dividing by the highest values obtained for the variable before analysis. Regional differences in ecological indices were evaluated by T-test (JMP® Pro, v14.1.0.; SAS Institute Inc., Cary, NC) (Figure 4-

8). Stepwise regressions of EPN against soil properties and all EPN competitors were also performed in (JMP) employing the mixed type of variable selection (table 4-4). Multivariate analyses of selected soil organisms and soil properties were performed using the software R (R

Development Core Team, ‘Vegan’ package). Principal component analysis (PCA) was used to reveal EPN, acari mite and soil properties that contribute significantly to the total spatial variability of the two ecoregions (Figure 4-8). Detrended Canonical Correspondence Analysis

(DCCA) was used to estimate the length of the system. A value <3.0 suggested that the community is homogeneous; therefore, Redundancy Analysis was used as the most appropriate constrained analysis, applying Monte Carlo permutation for significant environmental variables at the 0.05 level. The graphical results of the RDA were presented with bi–plot scaling (R

Development Core Team, ‘Vegan’ Package) (Figure 4-7).

Results

The high-throughput sequencing produced two datasets, one based on ITS2 targeting nematodes and the other on the D3-D5 region of 28S rDNA targeting microarthropods. The ITS2 revealed 9,109,218 reads of which 56% (5,118,191) passed the quality filters and were denoised,

merged and characterized as non-chimeric. Further analysis recovered 39717 unique amplicon sequence variants (ASVs) from forty-nine phyla, with 8.3% (3261 ASVs) assigned to Nematoda.

Fifty-five percent (1816) of the unique ASVs from 35 nematode families were identified to genus level or below. By setting a threshold of 80% coverage, 412 ASVs were identified as entomopathogenic nematode species. The D3-D5 dataset yielded 9,637,729 reads that reduced to

3,627,990 (37%) after filtering, denoising, merging and chimera removal. The total number of unique ASVs was 8728 from 33 phyla, of which 1402 belong to arthropods. Setting a threshold of 80% coverage and eliminating all the ASVs that had fewer than 10 reads, revealed 580 ASVs from 42 families in the class Arachnida. Relative occurrence (Figure 4-3) and relative abundance

(Figure 4-4) of the detected families, showed Acaridae, Ascidae, Tarsonemidae and Tydeidae as dominating the acari mite’s community structure in both regions, occurring in almost every site.

Entomopathogenic nematodes were recovered from all 62 (100%) soil samples collected.

Steinernema spp. were detected from all sites and Heterorhabditis spp. from 43 sites (69%)

(Figure 4-3). Steinernematid ASVs were identified as Steinernema carpocapsae (in 37% of the sites), Steinernema diaprepesi (25.8%), Steinernema feltiae (82.2%), Steinernema glaseri (93,5),

Steinernema riobrave (32.2%) and Steinernema khuongi (33.8%). Heterorhabditids included

Heterorhabditis bacteriophora (35.4%), Heterorhabditis floridensis (37%), Heterorhabditis indica (24.1%) (Figure 4-3)

Phylogenetic analysis characterized 412 ASVs derived from ITS2 sequences as S. carpocapsae (17 ASVs), S. diaprepesi (6), S. feltiae, (23), S. glaseri (293), S. riobrave (19), S. khuongi 20), H. bacteriophora (2), H. floridensis (10) and H. indica (11). The inter- and intra- species distances in the steinernematid and heterorhabditid phylogenetic trees support the validity of the ASV designations, based on the consensus topology of the species from previous

phylogenetic studies (Figure 4-1,4-2 ). All ASV coverages were 100% and the vast majority had

>99% identity to GenBank reference sequences with only 14 S. khuongi ASVs having identities between 99% - 97,94% (Sup. Table 1).

The blast results from metabarcoding data revealed that all nine species tend to exhibit patterns referred to as “head-tail” by Porazinska et al. (2010), with some (S. glaseri and S. feltiae) more definite than others (H. indica and S. khuongi) (Table 4-1).

Real-time PCR failed to detect nematodes in raw DNA, with the exception of S. feltiae and S. carpocapsae from two different sites having a large number of metabarcoding reads.

Following an enrichment step and clean up of the PCR product with magnetic beads, each species was detected by qPCR at least once for 6 of the 9 tested species. The qPCR results were highly correlated with reads numbers for all species except H. indica, S glaseri, and S. riobrave

(Table 4-2).

The regression of relative numbers of rDNA gene copies per IJ on the relative body volume revealed that copies per IJ increased at 8.074 times the rate of the body volume

(r2=0.764, P9=0.0021) (Figure 4-5). The estimated EPN community structure changed markedly when ASV abundance was adjusted for numbers of gene copies per IJ. For example, S. glaseri abundance changed from 89% to 62% and 62% to 14% in Argos and Chania, respectively, whereas in Chania S. carpocapsae increased from 11% to 63% (Figure 4-6).

Stepwise multiple regression derived relationships between EPN species and three groups of biotic and abiotic factors as independent variables. Phosphorus was inversely associated with four of the nine species, clay was positively associated with three species and pH with two.

Among the acari mite group, the Tydeidae were directly related to five EPN species. Oscheius

sp., among the other organisms known to compete with or prey on EPNs, was associated with 6

EPN species (Table 4-4).

Redundancy Analysis (RDA) identified loam, P, Mg, B and pH as associated with the communities of acari mites (P<0.05). Three mite families, Trombiculidae, Oehserchestidae and

Tydeidae, were related to the structure of the EPN communities as were clay content, soil pH and manganese. Finally, Oscheius sp., Hirsutela sp., and P. entomophagus, reported elsewhere to be natural enemies of EPN, were found to explain significant variation in EPN communities (Figure

4-7, Table 4-3).

Substantial differences in the principal component analysis of soil properties between the two ecoregions were less apparent for the EPN communities and even less for the acari mites

(Figure 4-8). This outcome was supported by the ecological indices for EPN diversity, richness and evenness which were higher in Chania than Argos (Figure 4-9), whereas there were no differences between those indices for mites (data not shown). Tydeidae and Tarsonemidae and S. glaseri, S.carpocapsae, S. diaprepesi and S. riobrave had the greatest influence on Acari mite and EPN differentiation of the two regions, respectively (Figure 4-8).

Discussion

EPN biogeography is studied increasingly, but the factors that regulate species occurrence and abundance remain poorly understood (Campos-Herrera et al., 2013, 2019a;

Garcia Del Pino and Palomo, 1996; Mráček et al., 1999; Page, 2015; Tarasco et al., 2015;

Valadas et al., 2014). Although the biological and physical complexity of habitats obscures key processes affecting EPNs, the increased accessibility of metagenomic tools will vastly increase the resolution at which soil food webs are characterized. Here we have created a comprehesive inventory of EPNs and microarthropods in the citrus orchards of two Greek ecoregions using environmental DNA recovered in a one-step process (Dritsoulas and Duncan, 2020). Our results

showed that it is unlikely that other molecular tools such as qPCR could detect many of the species in these orchards, even if were it affordable, due to their low abundance (Dritsoulas et al.,

2019). Moreover, the DNA used here can be probed with additional barcodes to extend the inventory of organisms in these food webs indefinitely (Pathak et al., 2017; Campos-Herrera et al., 2019).

The EPN detection frequency and species richness in these orchards are the highest reported to date. Only one other survey, also from citrus orchards but using species-specific qPCR primers and probes, has reported EPNs in all sites sampled (Campos-Herrera et al., 2013).

However, the use of DNA rather than sentinel insects to detect EPN spatial patterns provides greater sensitivity and precludes meaningful comparison of recent surveys to previous studies.

Indeed, the superior sensitivity of barcoding poses similar problems for comparison of results to those from qPCR.

The sequence analysis and the constructed trees depict ASV cohesion around the identified species, and separation from closely related species, supporting their designation as intraspecific variants (Figure 4-1,4-2). All the identified species consisted of multiple ASVs that, in many cases, exhibit “head-tail” patterns where a single head ASV perfectly matches a

NCBI reference sequence and comprises the majority of the sequencing reads, with tail ASVs represented by just a few sequences (Porazinska et al., 2010) (table 4-1). However, the most abundant Steinernema feltiae ASV does not perfectly match any NCBI reference even though other tail-ASVs are identical to reference sequences, and S. riobrave does not include a single perfectly identified ASV (Sup. table 4-1). Several species detected here, were previously only reported from the New World. Steinernema diaprepesi has been found in Argentina (Del Valle et al., 2014),Venezuela (Spiridonov et al., 2004) and USA (Nguyen and Duncan, 2002), while S.

khuongi was recently described from Florida (Stock et al., 2019), and S. riobrave is known only from USA and Mexico (Stuart et al., 2006). Only S. riobrave among these three species has been produced commercially, although there are no reports of its use in Greece.

The comparison of species detected by HTS and qPCR in this study confirmed not only the superior sensitivity of HTS (Dritsoulas et al., 2020), but that the soils in these orchards have

EPNs in numbers vastly lower than previously reported from environmental samples. Real Time

PCR estimated 2.5 S.carpocapsae IJs per 250 cm3 soil in just one site where the reads from metabarcoding was relatively high; an average number of IJs based on ASV proportions in the sites where the nematode was detected only by metabarcoding was 0.04 IJ’s. DNA lost during cleanup following the PCR amplification prior to both HTS and qPCR processing confounds estimates of the species abundance here, but the lack of detection of most species in raw DNA from six sites with abundant ASVs for all species indicates that EPN communities in these orchards, while diverse, are smaller than typically reported. Whether the situation in these orchards is exceptional remains to be seen from similar surveys that employ metabarcoding.

Also unknown are whether a lack of EPN abundance reflects a paucity of soil arthropods (Gazit et al., 2000), and how metapopulation dynamics in these sites compares to other systems

(Campos-Herrera et al., 2010).

The significant correlations between qPCR estimates and ASV numbers for most species supports the validity of those identifications. A lack of congruence observed for some species likely reflects an inability of qPCR to reflect the entire intraspecific genome variation.

Steinernema glaseri was detected in all sites with qPCR but at levels that were unrelated to the numbers of ASVs. However, 300 different ASVs were identified as S. glaseri and it is likely that

some were not amplified by the qPCR primers-probe sequences (Dritsoulas et al., 2020;

Kiewnick et al., 2015; Torr et al., 2007).

The use of qPCR-derived correction factors to quantify communities detected by metabarcoding could be especially applicable to EPNs, since only the infective stage is encountered in soil. Variation in the rDNA copy number is regulated by multiple factors including developmental stage and body size (Cunha et al., 1999), as well as unknown causes of interspecific, intraspecific and even intragenomic variability (Bik et al., 2013; Pereira and

Baldwin 2016). Darby et al., (2013) demonstrated that reads from high-throughput amplicon sequencing, uncorrected for copy number, cannot accurately measure the proportional abundance of specimens in a sample. The strength of the regression of copy numbers on body size in this study suggests that adjusting metabarcode quantities based on body volume greatly improved the estimate of species abundance in these environmental samples. Similarly, clay and pH were identified by RDA as potential EPN community drivers, only after adjusting the data, thereby conforming with previous reports of the significance of clay and pH (Campos-Herrera et al.,

2019a, 2013; El-Borai et al., 2016; Hara et al., 1991). A similar approach using multiple isolates of the different species would likely provide a clearer picture of the error associated with the corrected estimates and whether HTS approaches can provide reliable population estimates.

The most important families for the acari mite community structure in both regions were

Acaridae, Ascidae, Tarsonemidae and Tydeidae representing a large part of the total abundance

(Figure 4-7). Numerous studies have tested species of Acaridae (Sarcoptiformes: Astigmata) and

Ascidae (Mesostigmata) for nematophagy (Epsky et al., 1988; Karagoz et al., 2007). There are fewer reports for Prostigmata families with absence of evidence for Tarsonemidae, however

there are extensive studies on significant levels of nematophagy involving Tydeidae, mites which are often thought to be primarily fungivorous (Santos et al., 1981; Santos and Whitford, 1981).

Among the potential predators and competitors of EPNs, Oscheius sp. and tydeid mites merit further attention for their relationships here with EPN community structure. Tydeid mites were detected with relatively high numbers of reads at most (87%) of the sites. It was the most significant species differentiating microarthropod communities in the two ecoregions and explained a significant amount of variation of EPN communities using RDA. Oscheius sp. by contrast was very patchy (24% frequency) in both ecoregions, and also explained a significant amount of variation in EPN community structure using RDA. Multiple regression showed that tydeid mites and Oscheius sp. each explained a significant amount of variability (P<0.10) in the abundance of 5 EPN species, but the relationships with EPNs were negative for Oscheius sp. and positive for the Tydeidae. Those relationships might be expected between EPNs and antagonistic organisms in geospatial surveys, depending on degree of aggregation of the natural enemy. For highly aggregated antagonists, other things being equal, expected EPN numbers would be least in sites that are suitable for the natural enemy. The mean abundance of the total EPN community was more than twice as large (P=0.02) in sites without Oscheius sp. than when the nematode was present. However, where antagonists occur widely across most of the sites, habitats that favor prey abundance could be reflected by increased numbers of natural enemies. Numerous direct relationships between EPNs and nematophagous fungi occurred in geospatial surveys of citrus orchards (Pathak et al., 2017) and natural areas (Campos-Herrera et al., 2016), suggesting that r- selected prey species, such as EPN IJs emerging from cadavers, could reliably increase the abundance of ubiquitous predators such as nematode-trapping fungi and certain microarthropod species. Indeed, an abundance of insects at some sites might support large numbers of EPNs and

their natural enemies compared to insect depauperate sites, resulting in geospatially positive correlations between predatory mites and their nematode prey.

Metabarcoding revealed that citrus orchards in Argos and Chania support EPN communities with greater diversity than reported elsewhere to date, but with very low abundance. The capacity of metabarcoding to measure infinite taxa with potential as EPN enemies and as EPN prey offers a promising new tool to understand EPN biogeography.

Advancing the utility of metabarcoding to improve the utility of EPNs will depend greatly on improving standard universal primers as well as the breadth and quality of poorly informed databases for the detection of the natural enemies of EPNs.

Table 4-1. Amplicon sequence variant (ASV) table illustrating “head-tail” structures associated with intraspecific variation. Accession Identified as Matcing Coverage ASV Argos Chania Code Heterorhabditis bacteriophora MH333236.1 H. bacteriophora strain BED301 100 100 5172 2 19 >Hb_2 MH333236.1 H. bacteriophora strain BED301 99.72 100 17 0 1 >Hb_1 Heterorhabditis floridensis DQ372922.1 H. floridensis strain Fl-332 100 100 711 8 8 >Hf_7 DQ372922.1 H. floridensis strain Fl-332 99.71 100 85 1 1 >Hf_2 DQ372922.1 H. floridensis strain Fl-332 99.71 100 40 1 0 >Hf_3 DQ372922.1 H. floridensis strain Fl-332 99.71 100 36 1 0 >Hf_9 DQ372922.1 H. floridensis strain Fl-332 99.71 100 35 0 1 >Hf_5 Heterorhabditis indica MF187650.1 H. indica isolate 3988 100 100 205 2 4 >Hi_5 MF187650.1 H. indica isolate 3988 99.71 100 151 1 1 >Hi_11 MF187650.1 H. indica isolate 3988 99.42 100 58 0 1 >Hi_7 MF187650.1 H. indica isolate 3988 99.71 100 40 0 1 >Hi_6 KJ938571.1 H. indica 100 100 39 1 1 >Hi_1 Steinernema carpocapsae MH231235.1 S. carpocapsae voucher UNPR52 99.77 100 2982 0 6 >Sc_8 MH231235.1 S. carpocapsae voucher UNPR52 100 100 284 2 4 >Sc_15 MH231235.1 S. carpocapsae voucher UNPR52 99.77 100 61 0 1 >Sc_16 MH231235.1 S. carpocapsae voucher UNPR52 99.54 100 56 0 1 >Sc_7 MH231235.1 S. carpocapsae voucher UNPR52 99.77 100 42 1 0 >Sc_9 Steinernema Diaprepesi GU173996.1 S. diaprepesi strain Hancock 31 100 100 211 3 8 >Sd_6 GU173996.1 S. diaprepesi strain Hancock 31 99.32 100 46 1 0 >Sd_1 GU173996.1 S. diaprepesi strain Hancock 31 99.55 100 35 0 1 >Sd_5 GU173996.1 S. diaprepesi strain Hancock 31 99.77 100 28 1 1 >Sd_2 GU173996.1 S. diaprepesi strain Hancock 31 99.77 100 17 0 1 >Sd_3 Steinernema feltiae KM016419.1 S. feltiae strain Jakutsk 98.61 100 7299 8 3 >Sf_23 JN886618.1 S. feltiae isolate H9 100 100 3180 15 7 >Sf_7 JN886618.1 S. feltiae isolate H9 99.77 100 2720 11 6 >Sf_9 JN886618.1 S. feltiae isolate H9 99.77 100 345 5 1 >Sf_8 JN886594.1 S. feltiae isolate 11A 99.77 100 277 2 1 >Sf_2 MG952288.1 S. feltiae isolate Iso17 100 100 212 1 6 >Sf_19 Steinernema glaseri GU173999.1 S. glaseri strain NJ 100 100 107148 28 26 >Sg_192 GU173999.1 S. glaseri strain NJ 99.77 100 20718 22 21 >Sg_5 GU173999.1 S. glaseri strain NJ 99.77 100 240 0 3 >Sg_282 GU173999.1 S. glaseri strain NJ 99.77 100 224 1 1 >Sg_126 GU173999.1 S. glaseri strain NJ 99.54 100 185 1 1 >Sg_59 Steinernema riobrave GU174001.1 S. riobrave strain Lennon 4 99.78 100 350 6 3 >Sr_18 GU174000.1 S. riobrave strain Bartow 99.78 100 120 3 2 >Sr_5 GU174001.1 S. riobrave strain Lennon 4 99.33 100 60 1 0 >Sr_14 GU174001.1 S. riobrave strain Lennon 4 99.78 100 54 1 0 >Sr_13 GU174001.1 S. riobrave strain Lennon 4 99.55 100 50 1 0 >Sr_10 Steinernema Khuongi GU174002.1 S. sp. Arcadia 99.31 100 181 2 4 >Sk_9 GU174002.1 S. sp. Arcadia 100 100 124 1 3 >Sk_4 GU174002.1 S. sp. Arcadia 99.54 100 66 1 0 >Sk_12 GU174002.1 S. sp. Arcadia 98.62 100 51 0 1 >Sk_2 GU174002.1 S. sp. Arcadia 99.77 100 43 1 0 >Sk_1 Note: Each ASV has a unique code (asv code) and is identified individually, indicating the accession number, the description of the NCBI reference sequence (identified as), the coverage and the identity score of the blasting results (matching). Also indicated are the abundance (ASV copies), and the occurrence of each ASV in the two ecoregions (Argos and Chania)

Table 4-2. Non-parametric (Spearman's) correlations between number of metabarcoding reads and the Ct values derived from qPCR measurements of the nine EPN species found in this study. The DNA samples derived from six sites containing ASVs from these nine species which were identical to GenBank reference sequences. HTS_reads/qPCR_copies Spearman ρ Prob>|ρ| H floridensis 0.938 0.0057 H. indica - - H. bacteriophora 1 <.0001 S. feltiae 0.9992 <.0001 S. carpocapsae 0.9856 0.0003 S. glaseri - - S. diaprepesi 0.9429 0.0048 S. khuongi 0.9426 0.0049 S. riobrave - -

Table 4-3. Significance of the Canonical Correspondence Analysis variables. RDA comparison Variable Pr(>F) Acari families vs soil properties (A) Loam 0.085 *

P 0.018 * Mg 0.005 ** pH 0.008 ** B 0.099 . Mn 0.058 . EPN species vs soil properties (B) Clay 0.037 *

Mn 0.076 . pH 0.010 ** EPN species vs Acari families (C) Oehserchestidae 0.072 .

Trombiculidae 0.046 * Tydeidae 0.002 ** EPN species vs FLN & NF (D) O.sp 0.001 *** P.entomophagus 0.025 * Hirsutella sp. 0.020 *

Note: See Table 4-4 footnote for definition of abbreviations.

Table 4-4. Significant variables from stepwise multiple regression of nine nematode species measured by high throughput sequencing, regressed against soil properties and potential biotic antagonists comprising acari mite families, free living nematodes and nematophagus fungi.

H.bacteriophora H.floridensis H.indica S.carpocapsae S.diaprepesi S.feltiae S.glaseri S.riobrave S.Khuongi

R-square 0.117308 0.183572 0.139426 0.109325 0.049485 0.226452 0.184745 0.279369 0.175954

B* - - - (-)0.0255 - - - - - CEC - - (-)0.0958 - - - - - (-)0.0008 Clay ------0.0435 0.0348 0.0087 Cu - 0.0721 (-)0.0437 ------K ------Loam ------(-)0.041 - - Mg - - - - - 0.033 - - - Mn (-)0.0068 - - - - 0.0209 - - - OM 0.0594 ------P - 0.0111 (-)0.0258 - - - (-)0.0045 (-)0.0487 - pHw ------(-)0.01 (-)0.0151 - S - 0.0154 ------Zn - (-)0.0194 ------

RSquare 0.236227 0.2467 0.136397 0.107819 0.104548 0.340651 0.096047

Hirsutella sp. - - 0.0042 - - - 0.001 - - O.sp (-)0.0856 - - (-)0.0052 (-)0.0199 0.0663 (-)0.0001 (-)0.0482 - O.tipulae - - - 0.087 0.0739 - - - - P.lilacinum 0.0344 ------(-)0.0431 - P. entomophagus 0.0343 - - - - - (-)0.0054 - -

RSquare 0.323 0.145 0.226 0.229 0.098 0.367 0.251 0.329

Ascidae - - (-)0.0211 - - - - - 0.0623 Chortoglyphidae ------(-)0.0667 (-)0.0526 - Eupodidae ------(-)0.0845 Laelapidae ------0.0502 0.0267 - Nothridae ------0.0273 Oehserchestidae - - 0.0744 ------Penthaleidae - - - - - (-)0.0662 0.0638 - - Rhodacaridae - (-)0.0888 ------(-)0.0178 Tarsonemidae - - - 0.0056 - - - - - Trombiculidae 0.0015 (-)0.0764 - (-)0.0458 - - - - - Tydeidae 0.0589 - 0.0109 0.0713 - - <.0001 0.0992 -

*Independent variables included acari families of Ascidae, Haemogamasidae, Laelapidae, Rhodacaridae, Acaridae, Chortoglyphidae, Histiostomatidae, Nanorchestidae, Nothridae, Oehserchestidae, Eupodidae, Penthaleidae, Tarsonemidae, Trombiculidae, Tydeidae; free living nematodes Oscheius sp., Oscheius tipulae, Pristionchus entomophagus; nematophagous fungi Hirsutella sp., Purpureocillium lilacinum, and soil properties, clay, loam, phosphorus (P), potassium (K), magnesium (Mg), soil pH, cation exchange capacity (CEC), sulfur (S), boron (B), zinc (Zn), manganese (Mn), iron (Fe), copper (Cu), and organic matter (OM).

Figure 4-1. Phenetic relationships of ASVs identified in the genus Steinernema based on sequencing reads of the ITS_2 region as inferred by the UPGMA method. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The evolutionary distances were computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site. Reference sequences of five species (S. cubanum, S. brazilense, S. citrae, S.backanense and S. pakistanense) were used as outgroups of each clade and C. elegans as a global outgroup. All reference sequences are indicated by the NCBI accession numbers. The rest of reference sequences were employed to confirm the right identification of each ASV in the phylogram.

Figure 4-2. Phenetic relationships of ASVs identified in the genus Heterorhabditis based on sequencing reads of the ITS_2 region as inferred by the UPGMA method. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The evolutionary distances were computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site. Reference sequences of five species (H. georgiana, H. baujardi and H. pakistanense) were used as an outgroup of each clade and C. elegans as a global outgroup. All reference sequences are indicated by the NCBI accession numbers. The rest of reference sequences were employed to confirm the right identification of each ASV in the phylogram.

Figure 4-3. Detection frequency (proportion positive sites) of A) entomopathogenic nematode species, B) acari mite families, C) free living nematodes and nematophagus fungi, in two Greek ecoregions, Argos (32 samples) and Chania (30 samples). Values represent average ± SEM

Figure 4-4. Relative abundances of acari mite families. Pie charts represent the proportional composition of the acari mite ASVs in the two ecoregions Argos and Chania.

Figure 4-5. Relationship between relative number of ribosomal DNA copies per Infective Juvenile (IJ) measured by qPCR and the relative body volume per IJ estimated by Andrassy formula for nine EPN species (H. bacteriophora (Hb), H.indica (Hi), H. floridensis (Hf), S. carpocapsae (Sc), S.diaprepesi (Sd), S.feltiae (Sf), S. glaseri NC (Sg NC), S. riobrave (Sr) and S. khuongi (Sx).

Figure 4-6. Relative abundances of the entomopathogenic nematode species ASVs detected by metabarcoding in the two ecoregions Argos and Chania before correction (A), and number of individual EPNs of each species following correction (B) using the equation y=-4.239+8.074*x, where x = the relative body volume per IJ and y = the DNA copies (ASVs) per IJ.

Figure 4-7. Redundancy analyses (RDA) depicting biplots of both citrus ecoregions survey Argos and Chania (n = 62) distribution and relationships between significant abiotic factors and soil organisms. A) Relationships between EPN species and soil properties as explanatory variables. B) Relationships between acari mite families and soil properties. C) Relationships between EPN species and acari mite families as explanatory variables. D) Relationships between EPN species and five EPN natural enemies. Three free living nematode species Oscheius tipulae, Oscheius sp., Pristionhus entomophagus and two nematophagus fungi species Purpuleocillium lilacinum)

Figure 4-8. Principal component analysis of the samples collected from the two ecoregions Argos (n=32) and Chania (n=30) based on, A) soil properties, B) EPN species where ASV copy number was corrected to nematode counts, C) ASV numbers of acari mite families. Corrplots on the right side of each PCA biplot shows the contribution of each variable on the dimensions of the biplots

Figure 4-9. Ecological indices (species richness, S’ ; Shannon diversity index, H’ ; eveness, J’) of EPN species from samples extracted from two ecoregions, Argos and Chania. Bars and error bars denote means and 95% confidence intervals, respectively. Means that are significantly different in multiple comparisons using Wilcoxon test are denoted by * (P ≤ 0.05).

100

CHAPTER 5 CONCLUSION

The aim of this research was to characterize the communities of entomopathogenic nematodes, soil microarthropods, and habitat properties that may influence them using novel molecular technologies. The study was based on four key elements: i) biogeography of entomopathogenic nematodes, as promising biocontrol agents, is poorly understood, ii) soil microarthropods are well-noted natural enemies of EPN, iii) soil properties affect EPN services, therefore cultural practices could exploit them iv) advances in high throughput sequencing permits the characterization of soil communities of widely diverse taxa. Another objective of this study was to create appropriate protocols and methods that could be easily employed in nematology and acarology assays.

The first question I asked was whether a single extraction method is appropriate for studying both EPNs and soil microarthropods. Sucrose centrifugation (SC), a common method used by nematologists, was compared to conventional acarology methods for their efficiency in recovering both kinds of animal. SC was more efficient than methods using Berlese Funnels. SC was not more efficient than heptane flotation (HF); however, the indices and relationships between microarthropod taxa derived from both methods were congruent, and HF recovered no nematodes. Thus, SC is an efficient method for extracting nematodes and soil microarthropods.

To determine whether high throughput sequencing can characterize entomopathogenic nematodes at an adequately fine-scale of taxonomic resolution, it was necessary to design novel primers and then compare the results from metabarcoding to those from quantative real time

PCR. This research demonstrated a lower detection threshold of metabarcoding, and generally greater reliability as measured by the fit of Taylor’s Power Law to EPN means derived by each method from pairs of samples. While not identical, multivariate analysis of results from each

101

approach identified similar biotic and abiotic factors that may influence the EPN community.

Nevertheless, HTS detected many, but not all, of the EPN natural enemies (free living nematode competitors of EPNs) detected by qPCR, suggesting a mismatch of the primers to the undetected species. As far as we are aware, there are no perfect universal primers that target all groups of nematodes, so primer optimization would be a fruitful line of future research (Griffiths et al.,

2018).

Using these improved protocols for metabarcoding with DNA samples from two Greek ecoregions, we showed that citrus orchards in Argos and Chania supported EPN communities with greater diversity than reported elsewhere to date, but with very low abundance, confirming the low detection threshold of the method. We adjusted the number of reads per species to number of IJs using a correction derived from the linear relationship between numbers of copies/IJ (from qPCR) and the average body volume of each species (from Andrassy’s formula).

Redundancy analysis indicated that soil clay content and pH, two properties identified in previous studies (Campos-Herrera et al., 2019a; El-Borai et al., 2016), were related to EPN community structure, only after adjusting the data. Intraspecific variability of copy numbers per

IJs merits additional study to understand the error bounds in this type of ASV correction. Tydeid mites and Oscheius sp. nematodes were also identified as potentially important natural enemies of EPNs.

EPNs are a widely studied soil guild and well-represented in GenBank databases. As we reported here, limitations in qPCR to reveal the total intraspecific variation of key gene sequences can be overcome with HTS using GenBank information. In the case of mocroarthropods, there are few metagenomic studies with no consensus regarding appropriate primers. This issue occurs mainly because current reference databases are incomplete, however

102

representation improves each year. Missing taxa for DNA metabarcode analysis results from a paucity of research using the method, but also from primer bias, highlighting the value of supplementing studies with local DNA barcoding and continuing a focus on designing universal primers.

The capacity of metabarcoding to measure virtually infinite taxa with potential to drive

EPN biogeography and local population biology ensures its importance as a research tool. This study shows metabarcoding aptness for regions lacking EPN reports. This study also identified several such potential key species that merit further investigation. Four of these species have not been previously reported in Europe or anywhere in the countries adjascent to Greece, thus requiring additional confirmation of these findings. It is also noteworthy that EPN species commonly reported in Mediterranean countries such S. feltiae, S. carpocapsae and S. glaseri were also detected in Greece, completing the species distribution map and showing potential for these native poulations to be employed against key crop pests in Greece such as olive fruit fly

(Sirjani et al., 2009) and mediterranean fruit fly (Gazit et al., 2000).

103

LIST OF REFERENCES

Abebe, E., Jumba, M., Bonner, K., Gray, V., Morris, K., Thomas, W.K., 2010. An entomopathogenic Caenorhabditis briggsae. The Journal of Experimental Biology 213, 3223–9. doi:10.1242/jeb.043109

Adams, B.J., Burnell, A.M., Powers, T.O., 1998. A phylogenetic analysis of heterorhabditis (nemata: rhabditidae) based on internal transcribed spacer 1 DNA sequence data. Journal of Nematology 30, 22–39.

Alexander, P.R., Hominick, W.M., Briscoe, B.R., 1997. Molecular taxonomy and phylogeny of entomopathogenic nematode species (Rhabditida: Steinernematidae) by RFLP analysis of the ITS region of the ribosomal DNA repeat unit. Systematic Parasitology 37, 187–193. doi:10.1023/A:1005843320923

Ali, J.G., Alborn, H.T., Campos-Herrera, R., Kaplan, F., Duncan, L.W., Rodriguez-Saona, C., Koppenhöfer, A.M., Stelinski, L.L., 2012. Subterranean, Herbivore-Induced Plant Volatile Increases Biological Control Activity of Multiple Beneficial Nematode Species in Distinct Habitats. PLoS ONE 7, e38146. doi:10.1371/journal.pone.0038146

Andrassy, I., 1956. Die rauminhalts-und gewichtsbestimmung der fadenwürmer (Nematoden). Acta Zoologica Hungarica 2.1, 1–5.

Andre, H.M., Ducarme, X., Lebrun, P., 2017. Nordic Society Oikos Soil Biodiversity : Myth , Reality or Conning ? Author ( s ): Henri M . André , Xavier Ducarme and Philippe Lebrun Published by : Wiley on behalf of Nordic Society Oikos Stable URL : http://www.jstor.org/stable/3547272 JSTOR is a not- 96, 3–24.

Andrews, S., Krueger, F., Seconds-Pichon, A., Biggins, F., Wingett, S., 2015. FastQC. A quality control tool for high throughput sequence data. Babraham Bioinformatics [WWW Document]. Babraham Institute. doi:v.0.11.5

Anisimova, M., Gascuel, O., 2006. Approximate Likelihood-Ratio Test for Branches: A Fast, Accurate, and Powerful Alternative. Systematic Biology 55, 539–552. doi:10.1080/10635150600755453

Arribas, P., Andújar, C., Hopkins, K., Shepherd, M., Vogler, A.P., 2016. Metabarcoding and mitochondrial metagenomics of endogean arthropods to unveil the mesofauna of the soil. Methods in Ecology and Evolution 7, 1071–1081. doi:10.1111/2041-210X.12557

Aucamp, J.L., Ryke, P.A.J., 1964. Preliminary report on a grease film extraction method for soil microarthropods. Pedobiologia 4, 77–79.

Beavers, J.B., McCoy, C.W., Kaplan, D.T., 1983. Natural Enemies of Subterranean Diaprepes abbrevialus (Coleoptera: Curculionidae) Larvae in Florida1. Environmental Entomology 12, 840–843. doi:10.1093/ee/12.3.840

104

Bedding, R., Akhurst, R., Kaya, H.K., 1993. Nematodes and the biological control of insect pests. CSIRO Australia.

Bednarek, A., Gaugler, R., 1997. Compatibility of Soil Amendments with Entomopathogenic Nematodes, Journal of Nematology.

Berlese, A., 1905. Apparachio per raccogliere presto ed in gran numero piccoli Artropodi. Redia 2, 85–89.

Bhat, A.H., Chaubey, A.K., Askary, T.H., 2020. Global distribution of entomopathogenic nematodes, Steinernema and Heterorhabditis. Egyptian Journal of Biological Pest Control. doi:10.1186/s41938-020-0212-y

BIK, H.M., SUNG, W., DE LEY, P., BALDWIN, J.G., SHARMA, J., ROCHA-OLIVARES, A., THOMAS, W.K., 2012. Metagenetic community analysis of microbial eukaryotes illuminates biogeographic patterns in deep-sea and shallow water sediments. Molecular Ecology 21, 1048–1059. doi:10.1111/j.1365-294X.2011.05297.x

Blaxter, M., 2003. Counting angels with DNA. Nature 421, 122–123. doi:10.1038/421122a

Braun-Kiewnick, A., Kiewnick, S., 2018. Real-time PCR, a great tool for fast identification, sensitive detection and quantification of important plant-parasitic nematodes. European Journal of Plant Pathology. doi:10.1007/s10658-018-1487-7

Callahan, B.J., McMurdie, P.J., Holmes, S.P., 2017. Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. The ISME Journal 11, 2639– 2643. doi:10.1038/ismej.2017.119

Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W., Johnson, A.J.A., Holmes, S.P., 2016. DADA2: High-resolution sample inference from Illumina amplicon data. Nature Methods 13, 581–3. doi:10.1038/nmeth.3869

Campbell, J.F., Gaugler, R., 1993. Nictation behaviour and its ecological implications in the host search strategies of entomopathogenic nematodes (Heterorhabditidae and Steinernematidae) [WWW Document]. Behaviour. doi:10.1163/156853993X00092

Campos-Herrera, R., Ali, J.G., Diaz, B.M., Duncan, L.W., 2013a. Analyzing spatial patterns linked to the ecology of herbivores and their natural enemies in the soil. Frontiers in Plant Science 4, 378. doi:10.3389/fpls.2013.00378

Campos-Herrera, R., Blanco-Pérez, R., Bueno-Pallero, F.Á., Duarte, A., Nolasco, G., Sommer, R.J., Rodríguez Martín, J.A., 2019a. Vegetation drives assemblages of entomopathogenic nematodes and other soil organisms: Evidence from the Algarve, Portugal. Soil Biology and Biochemistry 128, 150–163. doi:10.1016/j.soilbio.2018.10.019

Campos-Herrera, R., El-Borai, F.E., Duncan, L.W., 2012. Wide interguild relationships among entomopathogenic and free-living nematodes in soil as measured by real time qPCR. Journal of Invertebrate Pathology 111, 126–135. doi:10.1016/j.jip.2012.07.006

105

Campos-Herrera, R., El-Borai, F.E., Ebert, T.E., Schumann, A., Duncan, L.W., 2014. Management to control citrus greening alters the soil food web and severity of a pest- disease complex. Biological Control 76, 41–51. doi:10.1016/j.biocontrol.2014.04.012

Campos-Herrera, Raquel, El-Borai, F.E., Stuart, R.J., Graham, J.H., Duncan, L.W., 2011. Entomopathogenic nematodes, phoretic Paenibacillus spp., and the use of real time quantitative PCR to explore soil food webs in Florida citrus groves. Journal of Invertebrate Pathology 108, 30–39. doi:10.1016/J.JIP.2011.06.005

Campos-Herrera, R., Gómez-Ros, J.M., Escuer, M., Cuadra, L., Barrios, L., Gutiérrez, C., 2008. Diversity, occurrence, and life characteristics of natural entomopathogenic nematode populations from La Rioja (Northern Spain) under different agricultural management and their relationships with soil factors. Soil Biology and Biochemistry 40, 1474–1484. doi:10.1016/j.soilbio.2008.01.002

Campos-Herrera, R., Jaffuel, G., Chiriboga, X., Blanco-Pérez, R., Fesselet, M., Půža, V., Mascher, F., Turlings, T.C.J., 2015a. Traditional and molecular detection methods reveal intense interguild competition and other multitrophic interactions associated with native entomopathogenic nematodes in Swiss tillage soils. Plant and Soil 389, 237–255. doi:10.1007/s11104-014-2358-4

Campos-Herrera, R., Johnson, E.G., EL-Borai, F.E., Stuart, R.J., Graham, J.H., Duncan, L.W., 2011. Long-term stability of entomopathogenic nematode spatial patterns in soil as measured by sentinel insects and real-time PCR assays. Annals of Applied Biology 158, 55–68. doi:10.1111/j.1744-7348.2010.00433.x

Campos-Herrera, R., Pathak, E., El-Borai, F.E., Schumann, A., Abd-Elgawad, M.M.M., Duncan, L.W., 2013b. New citriculture system suppresses native and augmented entomopathogenic nematodes. Biological Control 66, 183–194. doi:10.1016/J.BIOCONTROL.2013.05.009

Campos-Herrera, Raquel, Pathak, E., El-Borai, F.E., Stuart, R.J., Gutiérrez, C., Rodríguez- Martín, J.A., Graham, J.H., Duncan, L.W., 2013. Geospatial patterns of soil properties and the biological control potential of entomopathogenic nematodes in Florida citrus groves. Soil Biology and Biochemistry 66, 163–174. doi:10.1016/j.soilbio.2013.07.011

Campos-Herrera, R., Půža, V., Jaffuel, G., Blanco-Pérez, R., Čepulyte-Rakauskiene, R., Turlings, T.C.J., 2015b. Unraveling the intraguild competition between Oscheius spp. nematodes and entomopathogenic nematodes: Implications for their natural distribution in Swiss agricultural soils. Journal of Invertebrate Pathology 132, 216–227. doi:10.1016/j.jip.2015.10.007

Campos-Herrera, R., Stuart, R.J., Pathak, E., El-Borai, F.E., Duncan, L.W., 2019b. Temporal patterns of entomopathogenic nematodes in Florida citrus orchards: Evidence of natural regulation by microorganisms and nematode competitors. Soil Biology and Biochemistry 128, 193–204. doi:10.1016/j.soilbio.2018.10.012

106

Creer, S., Fonseca, V.G., Porazinska, D.L., Giblin-Davis, R.M., Sung, W., Power, D.M., Packer, M., Carvalho, G.R., Blaxter, M.L., Lambshead, P.J.D., Thomas, W.K., 2010. Ultrasequencing of the meiofaunal biosphere: Practice, pitfalls and promises. Molecular Ecology 19, 4–20. doi:10.1111/j.1365-294X.2009.04473.x

Cunha, A., Azevedo, R.B.R., Emmons, S.W., Leroi, A.M., 1999. Variable cell number in nematodes. Nature 402, 253. doi:10.1038/46211

Curini-Galletti, M., Artois, T., Delogu, V., De Smet, W.H., Fontaneto, D., Jondelius, U., Leasi, F., Martínez, A., Meyer-Wachsmuth, I., Nilsson, K.S., Tongiorgi, P., Worsaae, K., Todaro, M.A., 2012. Patterns of Diversity in Soft-Bodied Meiofauna: Dispersal Ability and Body Size Matter. PLoS ONE 7, e33801. doi:10.1371/journal.pone.0033801

Curran, J., Driver, F., Ballard, J.W.O., Milner, R.J., 1994. Phylogeny of Metarhizium: analysis of ribosomal DNA sequence data. Mycological Research 98, 547–552. doi:10.1016/S0953- 7562(09)80478-4

Decraemer, W., Hunt, D.J., 2006. Structure and classification, in: Perry, R.N., Moens, M. (Eds.), Plant Nematology. CABI, pp. 3–32.

Del Valle, E.E., Balbi, E.I., Lax, P., Rondán Dueñas, J., Doucet, M.E., 2014. Ecological aspects of an isolate of Steinernema diaprepesi (Rhabditida: Steinernematidae) from Argentina. Biocontrol Science and Technology 24, 690–704. doi:10.1080/09583157.2014.890171

Dillman, A.R., Chaston, J.M., Adams, B.J., Ciche, T.A., Goodrich-Blair, H., Stock, S.P., Sternberg, P.W., 2012. An entomopathogenic nematode by any other name. PLoS Pathogens 8. doi:10.1371/journal.ppat.1002527

Dolinski, C., Choo, H.Y., Duncan, L.W., 2012. Grower acceptance of entomopathogenic nematodes: Case studies on three continents. Journal of Nematology 44, 226–235.

Dritsoulas, A., Campos-Herrera, R., Blanco-Pérez, R., Duncan, L.W., 2020. Comparing high throughput sequencing and real time qPCR for characterizing entomopathogenic nematode biogeography. Soil Biology and Biochemistry 145. doi:10.1016/j.soilbio.2020.107793

Ducarme, X., André, H.M., Lebrun, P., 1998. Extracting endogeous microarthropods: A new flotation method using 1,2-dibromoethane. European Journal of Soil Biology 34, 143–150. doi:10.1016/S1164-5563(00)88651-3

Duncan, L.W., Graham, J.H., Dunn, D.C., Zellers, J., McCoy, C.W., Nguyen, K., 2003. Incidence of endemic entomopathogenic nematodes following application of Steinernema riobrave for control of Diaprepes abbreviatus. Journal of Nematology 35, 178–186.

Duncan, L.W., Graham, J.H., Zellers, J., Bright, D., Dunn, D.C., El-Borai, F.E., Porazinska, D.L., 2007. Food Web Responses to Augmenting the Entomopathogenic Nematodes in Bare and Animal Manure-Mulched Soil.

107

Duncan, L.W., McCoy, C.W., Stansly, P.A., Graham, J.H., Mizell, R.F., 2001. Estimating the Relative Abundance of Adult Citrus Root Weevils (Coleoptera: Curculionidae) with Modified Tedders Traps. Environmental Entomology 30, 939–946. doi:10.1603/0046- 225X-30.5.939

Duncan, L.W., McCoy, C.W., Terranova, A.C., 1996. Estimating sample size and persistence of entomogenous nematodes in sandy soils and their efficacy against the larvae of diaprepes abbreviatus in Florida. Journal of Nematology 28, 56–67.

Duncan, L.W., Stuart, R.J., El-Borai, F.E., Campos-Herrera, R., Pathak, E., Giurcanu, M., Graham, J.H., 2013. Modifying orchard planting sites conserves entomopathogenic nematodes, reduces weevil herbivory and increases citrus tree growth, survival and fruit yield. Biological Control 64, 26–36. doi:10.1016/j.biocontrol.2012.09.006

El-Borai, F., Killiny, N., Duncan, L.W., 2016. Concilience in entomopathogenic nematode responses to water potential and their geospatial patterns in Florida. Frontiers in Microbiology 7, 1–12. doi:10.3389/fmicb.2016.00356

El-Borai, F.E., Duncan, L.W., Preston, J.F., 2005. Bionomics of a Phoretic Association Between Paenibacillus sp. and the Entomopathogenic Nematode Steinernema diaprepesi. Journal of Nematology 37, 18–25.

El-Borai, F.E., Stuart, R.J., Campos-Herrera, R., Pathak, E., Duncan, L.W., 2012. Entomopathogenic nematodes, root weevil larvae, and dynamic interactions among soil texture, plant growth, herbivory, and predation. Journal of Invertebrate Pathology 109, 134– 142. doi:10.1016/j.jip.2011.10.012

Emelianoff, V., Le Brun, N., Pagès, S., Stock, S.P., Tailliez, P., Moulia, C., Sicard, M., 2008. Isolation and identification of entomopathogenic nematodes and their symbiotic bacteria from Hérault and Gard (Southern France). Journal of Invertebrate Pathology 98, 211–217. doi:10.1016/j.jip.2008.01.006

Enright, M.R., Griffin, C.T., 2005. Effects of Paenibacillus nematophilus on the entomopathogenic nematode Heterorhabditis megidis. Journal of Invertebrate Pathology 88, 40–48. doi:10.1016/J.JIP.2004.10.002

Epsky, N.D., Walter, D.E., Capinera, J.L., 1988. Potential Role of Nematophagous Microarthropods as Biotic Mortality Factors of Entomogenous Nematodes (Rhabditida: Steinernematidae, Heterorhabditidae). Journal of Economic Entomology 81, 821–825. doi:10.1093/jee/81.3.821

Ewels, P., Magnusson, M., Lundin, S., Käller, M., 2016. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics (Oxford, England) 32, 3047–8. doi:10.1093/bioinformatics/btw354

Fan, X., Hominick, W.M., 1991. Efficiency of the Galleria (wax moth) baiting technique for recovering infective stages of entomopathogenic rhabditids (Steinernematidae and Heterorhabditidae) from Sand and soil, Revue Nématol.

108

Fitters, P.F.L., Patel, M.N., Griffin, C.T., Wright, D.J., 1999. Fatty acid composition of Heterorhabditis sp. during storage. Comparative Biochemistry and Physiology - B Biochemistry and Molecular Biology 124, 81–88. doi:10.1016/S0305-0491(99)00102-9

Floyd, R., Abebe, E., Papert, A., Blaxter, M., 2002. Molecular barcodes for soil nematode identification. Molecular Ecology 11, 839–50. doi:10.1046/j.1365-294X.2002.01485.x

Forootan, A., Sjöback, R., Björkman, J., Sjögreen, B., Linz, L., Kubista, M., 2017. Methods to determine limit of detection and limit of quantification in quantitative real-time PCR (qPCR). Biomolecular Detection and Quantification 12, 1–6. doi:10.1016/j.bdq.2017.04.001

Frank, J.H., Walker, T.J., 2006. Permanent Control of Pest Mole Crickets (Orthoptera: Gryllotalpidae: Scapteriscus ) in Florida. American Entomologist 52, 138–144. doi:10.1093/ae/52.3.138

Frey, K.G., Herrera-Galeano, J.E., Redden, C.L., Luu, T. V, Servetas, S.L., Mateczun, A.J., Mokashi, V.P., Bishop-Lilly, K.A., 2014. Comparison of three next-generation sequencing platforms for metagenomic sequencing and identification of pathogens in blood. BMC Genomics 15, 96. doi:10.1186/1471-2164-15-96

Futch, S., Duncan, L., Soc, M.Z.-P.F.S.H., 2005, U., 2005. Validation of an area-wide extension program to estimate the seasonal abundance of adult citrus root weevils with un-baited pyramidal traps. Proceedings of the Florida State Horticultural Society 143–147.

Garcia Del Pino, F., Palomo, A., 1996. Natural occurrence of entomopathogenic nematodes (Rhabditida: Steinernematidae and Heterorhabditidae) in Spanish soils. Journal of Invertebrate Pathology 68, 84–90. doi:10.1006/jipa.1996.0062

Gazit, Y., Rössler, Y., Glazer, I., 2000. Evaluation of entomopathogenic nematodes for the control of Mediterranean fruit fly (Diptera: Tephritidae). Biocontrol Science and Technology 10, 157–164. doi:10.1080/09583150029297

Georgis, R., Koppenhöfer, A.M., Lacey, L.A., Bélair, G., Duncan, L.W., Grewal, P.S., Samish, M., Tan, L., Torr, P., van Tol, R.W.H.M., 2006. Successes and failures in the use of parasitic nematodes for pest control. Biological Control 38, 103–123. doi:10.1016/J.BIOCONTROL.2005.11.005

Georgis, R., Poinar, G.O., 1983. Effect of Soil Texture on the Distribution and Infectivity of Neoaplectana carpocapsae (Nematoda: Steinernematidae). Journal of Nematology 15, 308– 11.

Gilmore, S.K., 1970. Collembola predation on nematodes. Search Agriculture, Entomology and Limnology 1, 1–12.

GLAZER, I., LIRAN, N., POINAR, G., SMITS, P., 1993. Identification and biological activity of newly isolated heterorhabditid populations from Israel. Fundamental and Applied Nematology 16, 467–472.

109

Glenn, T.C., 2011. Field guide to next-generation DNA sequencers. Molecular Ecology Resources 11, 759–769. doi:10.1111/j.1755-0998.2011.03024.x

Graham, J.H., Bright, D.B., McCoy, C.W., 2003. Phytophthora - Diaprepes weevil complex: Phytophthora spp. relationship with citrus rootstocks. Plant Disease 87, 85–90. doi:10.1094/PDIS.2003.87.1.85

Grewal, P.S., Ehlers, R.U., Shapiro-Ilan, D.I., 2005. Nematodes as biocontrol agent, Nematodes as Biocontrol Agent. doi:10.21273/hortsci.46.2.333a

Griffiths, B.S., de Groot, G.A., Laros, I., Stone, D., Geisen, S., 2018. The need for standardisation: Exemplified by a description of the diversity, community structure and ecological indices of soil nematodes. Ecological Indicators 87, 43–46. doi:10.1016/j.ecolind.2017.12.002

Guindon, S., Gascuel, O., 2003. A Simple, Fast, and Accurate Algorithm to Estimate Large Phylogenies by Maximum Likelihood. Systematic Biology 52, 696–704. doi:10.1080/10635150390235520

Hara, A.H., Gaugler, R., Kaya, H.K., Lebeck, L.M., 1991. Natural populations of entomopathogenic nematodes (Rhabditida: Heterorhabditidae, steinernematidae) from the Hawaiian Islands. Environmental Entomology 20, 211–216. doi:10.1093/ee/20.1.211

Heidemann, K., Scheu, S., Ruess, L., Maraun, M., 2011. Molecular detection of nematode predation and scavenging in oribatid mites: Laboratory and field experiments. Soil Biology and Biochemistry 43, 2229–2236. doi:10.1016/j.soilbio.2011.07.015

Hominick, W.M., 2002. Biogeography., in: Entomopathogenic Nematology. CABI, Wallingford, pp. 115–143. doi:10.1079/9780851995670.0115

Hominick, W.M., Reid, A.P., Bohan, D.A., Briscoe, B.R., 1996. Entomopathogenic nematodes: Biodiversity, geographical distribution and the convention on biological diversity. Biocontrol Science and Technology 6, 317–331. doi:10.1080/09583159631307

Iqbal, M., Ehlers, R.-U., Waeyenberge, L., 2016. Molecular characterisation of novel isolates of entomopathogenic nematodes. Nematology 18, 277–291. doi:10.1163/15685411- 00002959

Jabbour, R., Barbercheck, M.E., 2011. Soil microarthropod response to the application of entomopathogenic nematode-killed insects in maize and flower strip habitats. Pedobiologia 54, 243–251. doi:10.1016/j.pedobi.2011.04.001

Jenkins, W.R., 1964. A rapid centrifugal-flotation technique for separating nematodes from soil. Plant Disease Reporter 48, 692. doi:10.1007/s10333

Jörger, K.M., Norenburg, J.L., Wilson, N.G., Schrödl, M., 2012. Barcoding against a paradox? Combined molecular species delineations reveal multiple cryptic lineages in elusive meiofaunal sea slugs. BMC Evolutionary Biology 12, 245. doi:10.1186/1471-2148-12-245

110

Karagoz, M., Gulcu, B., Cakmak, I., Kaya, H.K., Hazir, S., 2007. Predation of entomopathogenic nematodes by Sancassania sp. (Acari: Acaridae). Experimental and Applied Acarology 43, 85–95. doi:10.1007/s10493-007-9105-y

Kaya, H.K., Gaugler, R., 1993. Entomopathogenic Nematodes. Annual Review of Entomology 38, 181–206. doi:10.1146/annurev.en.38.010193.001145

Kaya, H.K., Koppenhöfer, A.M., 1996. Effects of microbial and other antagonistic organism and competition on entomopathogenic nematodes. Biocontrol Science and Technology 6, 357– 371. doi:10.1080/09583159631334

Kiewnick, S., Frey, J.E., Braun-Kiewnick, A., 2015. Development and Validation of LNA-Based Quantitative Real-Time PCR Assays for Detection and Identification of the Root-Knot Nematode Meloidogyne enterolobii in Complex DNA Backgrounds. Phytopathology 105, 1245–1249. doi:10.1094/PHYTO-12-14-0364-R

Koppenhöfer, A.M., Campbell, J.F., Kaya, H.K., Gaugler, R., 1998. Estimation of entomopathogenic nematode population density in soil hy correlation hetween hait insect mortality and nematode penetration. Fundamental and Applied Nematology 21, 95–102.

Koppenhöfer, A.M., Fuzy, E.M., 2006. Effect of soil type on infectivity and persistence of the entomopathogenic nematodes Steinernema scarabaei, Steinernema glaseri, Heterorhabditis zealandica, and Heterorhabditis bacteriophora. Journal of Invertebrate Pathology 92, 11–22. doi:10.1016/j.jip.2006.02.003

Krantz, G.W. (Gerald W., Walter, D.E., 2009. A manual of acarology. Texas Tech University Press.

Lambshead D., P.J., 1993. Recent developments in marine benthic biodiversity reserch. Oceanis 19, 5–24.

Lawrence, J.L., Hoy, C.W., Grewal, P.S., 2006. Spatial and temporal distribution of endemic entomopathogenic nematodes in a heterogeneous vegetable production landscape. Biological Control 37, 247–255. doi:10.1016/j.biocontrol.2006.02.002

Lewis, E.E., Campbell, J., Griffin, C., Kaya, H., Peters, A., 2006. Behavioral ecology of entomopathogenic nematodes. Biological Control 38, 66–79. doi:10.1016/j.biocontrol.2005.11.007

Lewis, E.E., Clarke, D.J., 2012. Nematode parasites and entomopathogens, in: Insect Pathology. pp. 395–424. doi:10.1016/B978-0-12-384984-7.00011-7

Liu, J., Berry, R.E., 1996. Steinernema oregonensis n. sp. (Rhabditida : Steinernematidae) from Oregon, U.S.A [WWW Document]. Fundamental and Applied Nematology. URL https://www.documentation.ird.fr/hor/fdi:010006263 (accessed 6.2.20).

McCoy, C.W., 1999. Arthropod pests of citrus roots. Citrus Health Management. APS Press 149–156.

111

McSorley, R., Walter, D.E., 1991. Comparison of soil extraction methods for nematodes and microarthropods. Agriculture, Ecosystems and Environment 34, 201–207. doi:10.1016/0167-8809(91)90106-8

Menti, H., Wright, D.J., Perry, R.N., 1997. Desiccation survival of populations of the entomopathogenic nematodes Steinernema feltiae and Heterorhabditis megidis from Greece and the UK. Journal of Helminthology 71, 41–46. doi:10.1017/s0022149x00000778

Millar, L.C., Barbercheck, M.E., 2002. Effects of tillage practices on entomopathogenic nematodes in a corn agroecosystem. Biological Control 25, 1–11. doi:10.1016/S1049- 9644(02)00042-7

Moore, J.C., Walter, D.E., Hunt, H.W., 1988. Arthropod regulation of micro- and mesobiota in below-ground detrital food webs, Annual review of entomology. Vol. 33.

Mráček, Z., Bečvář, S., Kindlmann, P., 1999. Survey of entomopathogenic nematodes from the families steinernematidae and heterorhabditidae (Nematoda: Rhabditida) in the Czech Republic. Folia Parasitologica 46, 145–148.

Neher, D.A., 1999. Nematode communities in organically and conventionally managed agricultural soils. Journal of Nematology 31, 142–154.

Neumann, G., Shields, E.J., 2011. Field persistence of steinernema carpocapsae Weiser (NY001), Steinernema feltiae Filipjev (Valko) and Heterorhabditis bacteriophora Poinar (Oswego) in Alfalfa Fields. Great Lakes Entomologist 44, 42–52.

Nguyen, K., Hunt, D., 2010. Entomopathogenic Nematodes: Systematics, Phylogeny and Bacterial Symbionts, Entomopathogenic Nematodes: Systematics, Phylogeny and Bacterial Symbionts. Brill. doi:10.1163/ej.9789004152939.i-816

Nguyen, K.B., Duncan, L.W., 2002. Steinernema diaprepesi n. sp. (Rhabditida: Steinernematidae), a parasite of the citrus root weevil Diaprepes abbreviatus (L) (Coleoptera: Curculionidae). Journal of Nematology 34, 159–170.

Nielsen, A.L., Lewis, E.E., 2012. Designing the ideal habitat for entomopathogen use in nursery production. Pest Management Science 68, 1053–1061. doi:10.1002/ps.3267

Nielsen, A.L., Spence, K.O., Nakatani, J., Lewis, E.E., 2011. Effect of soil salinity on entomopathogenic nematode survival and behaviour. Nematology 13, 859–867. doi:10.1163/138855411X562254

Oliveira, A.R., De Moraes, G.J., Ferraz, L.C.C.B., 2007. Consumption rate of phytonematodes by Pergalumna sp. (Acari: Oribatida: Galumnidae) under laboratory conditions determined by a new method. Experimental and Applied Acarology 41, 183–189. doi:10.1007/s10493- 007-9062-5

Page, S., 2015. ANNUAL Review of Current Management Options. Therapeutics, Advanced Veterinary.

112

Parkman, J.P., Hudson, W.G., Frank, J.H., Nguyen, K.B., Smart, G.C., 1993. Establishment and Persistence of Steinernema scapterisci (Rhabditida: Steinernematidae) in Field Populations of Scapteriscus spp. Mole Crickets (Orthoptera: Gryllotalpidae). Journal of Entomological Science 28, 182–190. doi:10.18474/0749-8004-28.2.182

Pathak, E., Campos–Herrera, R., El–Borai, F.E., Duncan, L.W., 2017. Spatial relationships between entomopathogenic nematodes and nematophagous fungi in Florida citrus orchards. Journal of Invertebrate Pathology 144, 37–46. doi:10.1016/j.jip.2017.01.005

Peters, A., 1996. The natural host range of Steinernema and Heterorhabditis spp. and their impact on insect populations. Biocontrol Science and Technology 6, 389–402. doi:10.1080/09583159631361

Petersen, H., Luxton, M., 1982. Quantitative Ecology of Microfungi and Animals in Soil and Litter, Source: Oikos.

Pfenninger, M., Schwenk, K., 2007. Cryptic animal species are homogeneously distributed among taxa and biogeographical regions. BMC Evolutionary Biology 7, 121. doi:10.1186/1471-2148-7-121

Pielou, E.C., 1975. Ecological diversity. Wiley.

Poinar, G.O., 1990. Taxonomy and biology of steinernematidae and heterorhabditidae. Entomopathogenic Nematodes in Biological Control 23–62. doi:10.1201/9781351071741

Poinar, G.O., 1979. Natural Enemies of Entomogenous Nematodes, in: Nematodes for Biological Control of Insects. CRC Press, pp. 215–221. doi:10.1201/9781351074957-7

Pompanon, F., Deagle, B.E., Symondson, W.O.C., Brown, D.S., Jarman, S.N., Taberlet, P., 2012. Who is eating what: Diet assessment using next generation sequencing. Molecular Ecology. doi:10.1111/j.1365-294X.2011.05403.x

Porazinska, D.L., Giblin-Davis, R.M., Faller, L., Farmerie, W., Kanzaki, N., Morris, K., Powers, T.O., Tucker, A.E., Sung, W., Thomas, W.K., 2009. Evaluating high-throughput sequencing as a method for metagenomic analysis of nematode diversity. Molecular Ecology Resources 9, 1439–1450. doi:10.1111/j.1755-0998.2009.02611.x

Porazinska, D.L., Giblin-Davis, R.M., Sung, W., Thomas, W.K., 2010. Linking operational clustered taxonomic units (OCTUs) from parallel ultra sequencing (PUS) to nematode species. Zootaxa 55–63. doi:10.11646/zootaxa.2427.1.6

Powers, T., 2004. NEMATODE MOLECULAR DIAGNOSTICS: From Bands to Barcodes. Annual Review of Phytopathology 42, 367–383. doi:10.1146/annurev.phyto.42.040803.140348

113

Powers, T.O., Neher, D.A., Mullin, P., Esquivel, A., Giblin-Davis, R.M., Kanzaki, N., Stock, S.P., Mora, M.M., Uribe-Lorio, L., 2009. Tropical nematode diversity: Vertical stratification of nematode communities in a Costa Rican humid lowland rainforest. Molecular Ecology 18, 985–996. doi:10.1111/j.1365-294X.2008.04075.x

Qing, X., Wang, M., Karssen, G., Bucki, P., Bert, W., Braun-Miyara, S., 2019. PPNID: a reference database and molecular identification pipeline for plant-parasitic nematodes. Bioinformatics. doi:10.1093/bioinformatics/btz707

Read, D.S., Sheppard, S.K., Bruford, M.W., Glen, D.M., Symondson, W.O.C., 2006. Molecular detection of predation by soil micro-arthropods on nematodes. Molecular Ecology 15, 1963–1972. doi:10.1111/j.1365-294X.2006.02901.x

Ristau, K., Steinfartz, S., Traunspurger, W., 2013. First evidence of cryptic species diversity and significant population structure in a widespread freshwater nematode morphospecies ( Tobrilus gracilis ). Molecular Ecology 22, 4562–4575. doi:10.1111/mec.12414

Santos, P.F., Whitford, W.G., 1981. The Effects of Microarthropods on Litter Decomposition in a Chihuahuan Desert Ecosystem. Ecology 62, 654–663. doi:10.2307/1937733

Schwartz, H.T., Antoshechkin, I., Sternberg, P.W., 2011. Applications of high-throughput sequencing to symbiotic nematodes of the genus Heterorhabditis. Symbiosis 55, 111–118. doi:10.1007/s13199-012-0151-9

Shamini, S., Amutha, K., 2014. Techniques For Extraction of Arbuscular Mycorrhizal Fungi Spores. International Journal of Frontiers in Science and Technology 2, 1–6.

Shapiro-Ilan, D.I., Cottrell, T.E., Mizell, R.F., Horton, D.L., Zaid, A., 2015. Field suppression of the peachtree borer, Synanthedon exitiosa, using Steinernema carpocapsae: Effects of irrigation, a sprayable gel and application method. Biological Control 82, 7–12. doi:10.1016/J.BIOCONTROL.2014.12.003

Shapiro-Ilan, D.I., Gouge, D.H., Koppenhöfer, A.M., 2009. Factors affecting commercial success: case studies in cotton, turf and citrus., in: Entomopathogenic Nematology. CABI, Wallingford, pp. 333–355. doi:10.1079/9780851995670.0333

Shapiro-Ilan, D.I., Gouge, D.H., Piggott, S.J., Fife, J.P., 2006. Application technology and environmental considerations for use of entomopathogenic nematodes in biological control. Biological Control 38, 124–133. doi:10.1016/J.BIOCONTROL.2005.09.005

Sharma, R.D., 1971. Studies on the plant parasitic nematode Tylenchorhynchus dubius [WWW Document]. Meded. Landb. Hogesch. Wageningen. URL https://library.wur.nl/WebQuery/wurpubs/525767 (accessed 6.2.20).

Shields, E.J., Testa, A., Neumann, G., Flanders, K.L., Schroeder, P.C., 2009. Biological Control of Alfalfa Snout Beetle with a multi-species application of locally adapted persistent entomopathogenic nematodes: The first success. American Entomologist 55, 250–257. doi:10.1093/ae/55.4.250

114

Simões, N., Rosa, J.S., 1996. Pathogenicity and host specificity of entomopathogenic nematodes. Biocontrol Science and Technology 6, 403–411. doi:10.1080/09583159631370

Simpson, E.H., 1949. Measurement of diversity [16]. Nature. doi:10.1038/163688a0

Sirjani, F.O., Lewis, E.E., Kaya, H.K., 2009. Evaluation of entomopathogenic nematodes against the olive fruit fly, Bactrocera oleae (Diptera: Tephritidae). Biological Control. doi:10.1016/j.biocontrol.2008.11.002

Southey, J.F., 1970. Laboratory methods for work with plant and soil nematodes. H.M. Stationery Off.

Southwood, T.R.E., Henderson, P.A., 2000. Ecological Methods. Blackwell Publishing, Oxford.

Spiridonov, S.E., Moens, M., Wilson, M.J., 2007. Fine scale spatial distributions of two entomopathogenic nematodes in a grassland soil. Applied Soil Ecology 37, 192–201. doi:10.1016/J.APSOIL.2007.06.009

Spiridonov, S.E., Reid, A.P., Podrucka, K., Subbotin, S.A., Moens, M., 2004. Phylogenetic relationships within the genus Steinernema (Nematoda: Rhabditida) as inferred from analyses of sequences of the ITS1-5.8S-ITS2 region of rDNA and morphological features. Nematology 6, 547–566. doi:10.1163/1568541042665304

Stirling, G.R., 2014. Biological products for nematode management., in: Biological Control of Plant-Parasitic Nematodes: Soil Ecosystem Management in Sustainable Agriculture. CABI, pp. 342–389. doi:10.1079/9781780644158.0342

Stock, S. P., 2009. Molecular approaches and the taxonomy of insect-parasitic and pathogenic nematodes., in: Insect Pathogens: Molecular Approaches and Techniques. CABI, Wallingford, pp. 71–100. doi:10.1079/9781845934781.0071

Stock, S. Patricia, 2009. Molecular approaches and the taxonomy of insect-parasitic and pathogenic nematodes, in: Insect Pathogens: Molecular Approaches and Techniques. CABI Publishing, pp. 71–100. doi:10.1079/9781845934781.0071

Stock, S.P., Campbell, J.F., Nadler, S.A., 2001. Phylogeny of Steinernema Travassos, 1927 (Cephalobina: Steinernematidae) Inferred From Ribosomal DNA Sequences and Morphological Characters. Https://Doi.Org/10.1645/0022- 3395(2001)087[0877:POSTCS]2.0.CO;2 87, 877–889. doi:10.1645/0022- 3395(2001)087[0877:POSTCS]2.0.CO;2

Stock, S.P., Campos-Herrera, R., El-Borai, F.E., Duncan, L.W., 2018. Steinernema khuongi n. sp. (Panagrolaimomorpha, Steinernematidae), a new entomopathogenic nematode species from Florida, USA. Journal of Helminthology 1–16. doi:10.1017/S0022149X18000081

115

Stock, S.P., Griffin, C.T., Chaerani, R., 2004. Morphological and molecular characterisation of Steinernema hermaphroditum n. sp. (Nematoda: Steinernematidae), an entomopathogenic nematode from Indonesia, and its phylogenetic relationships with other members of the genus [WWW Document]. Nematology. doi:10.1163/1568541042360555

Stock, S.P., Kaya, H.K., 1996. A Multivariate Analysis of Morphometric Characters of Heterorhabditis Species (Nemata: Heterorhabditidae) and the Role of Morphometrics in the Taxonomy of Species of the Genus. The Journal of Parasitology 82, 806. doi:10.2307/3283895

Stoeckle, M., 2003. Taxonomy, DNA, and the Bar Code of Life. BioScience 53, 796–797. doi:10.1641/0006-3568(2003)053[0796:tdatbc]2.0.co;2

Strauch, O., Stoessel, S., Ehlers, R.-U., 1994. Culture conditions define automictic or amphimictic reproduction of entomopathogenic rhabditid nematodes of the genus {IHeterorhabditis}. Fundamental and Applied Nematology 17, 575–582.

Stuart, R.J., Barbercheck, M.E., Grewal, P.S., Taylor, R.A.J., Hoy, C.W., 2006. Population biology of entomopathogenic nematodes: Concepts, issues, and models. Biological Control 38, 80–102. doi:10.1016/J.BIOCONTROL.2005.09.019

Stuart, R.J., Gaugler, R., 1994. Patchiness in Populations of Entomopathogenic Nematodes. Journal of Invertebrate Pathology 64, 39–45. doi:10.1006/jipa.1994.1066

Tarasco, E., Clausi, M., Rappazzo, G., Panzavolta, T., Curto, G., Sorino, R., Oreste, M., Longo, A., Leone, D., Tiberi, R., Vinciguerra, M.T., Triggiani, O., 2015. Biodiversity of entomopathogenic nematodes in Italy. Journal of Helminthology 89, 359–366. doi:10.1017/S0022149X14000194

Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22, 4673–4680. doi:10.1093/nar/22.22.4673

TORR, P., SPIRIDONOV, S.E., HERITAGE, S., WILSON, M.J., 2007. Habitat associations of two entomopathogenic nematodes: a quantitative study using real-time quantitative polymerase chain reactions. Journal of Animal Ecology 76, 238–245. doi:10.1111/j.1365- 2656.2006.01196.x

Treonis, A.M., Unangst, S.K., Kepler, R.M., Buyer, J.S., Cavigelli, M.A., Mirsky, S.B., Maul, J.E., 2018. Characterization of soil nematode communities in three cropping systems through morphological and DNA metabarcoding approaches. Scientific Reports 8, 2004. doi:10.1038/s41598-018-20366-5

Tullgren, A., 1918. Ein sehr einfacher Ausleseapparat fuer terricole Tierfaunen. Zeitschrift Fur Angewandte Entomologie 4, 149–150.

116

Tully, T., D’Haese, C.A., Richard, M., Ferrière, R., 2006. Two major evolutionary lineages revealed by molecular phylogeny in the parthenogenetic collembola species Folsomia candida. Pedobiologia 50, 95–104. doi:10.1016/j.pedobi.2005.11.003

V. Wintzingerode, F., Göbel, U.B., Stackebrandt, E., 1997. Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis. FEMS Microbiology Reviews 21, 213–229. doi:10.1111/j.1574-6976.1997.tb00351.x

Valadas, V., Laranjo, M., Mota, M., Oliveira, S., 2014. A survey of entomopathogenic nematode species in continental Portugal. Journal of Helminthology 88, 327–341. doi:10.1017/S0022149X13000217

Walter, D.E., Ikonen, E.K., 1989. Species, guilds, and functional groups: taxonomy and behavior in nematophagous arthropods. Journal of Nematology 21, 315–27.

Walter, D.E., Kaplan, D.T., Davis, E.L., 1993. Colonization of Greenhouse Nematode Cultures by Nematophagous Mites and Fungi. Journal of Nematology 25, 789.

Walter, D.E., Proctor, H.C., 2013. Mites and Biological Diversity, in: Mites: Ecology, Evolution & Behaviour. Springer Netherlands, pp. 447–459. doi:10.1007/978-94-007-7164-2_11

Walter, E.W., Moore, J.C., Loring, S.J., 1989. Symphylella sp. (Symphyla: Scolopendrellidae) predators of arthropods and nematodes in grassland soils [WWW Document]. Pedobiologia. URL https://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=6725640 (accessed 6.2.20).

Wilson, M., Gaugler, R., 2004. Factors limiting short-term persistence of entomopathogenic nematodes. Journal of Applied Entomology 128, 250–253. doi:10.1111/j.1439- 0418.2004.00814.x

Ye, W., Foye, S., MacGuidwin, A.E., Steffan, S., 2018. Incidence of oscheius onirici (Nematoda: Rhabditidae), a potentially entomopathogenic nematode from the marshlands of Wisconsin, USA. Journal of Nematology 50, 9–26. doi:10.21307/jofnem-2018-004

Yilmaz, H., Waeyenberge, L., Demir, I., Moens, M., Demirbaǧ, Z., 2009. A new entomopathogenic nematode species for Turkey, Heterorhabditis megidis Poinar, Jackson & Klein 1987 (Rhabditida: Heterorhabditidae). Turkish Journal of Agriculture and Forestry 33, 385–391. doi:10.3906/tar-0808-26

117

BIOGRAPHICAL SKETCH

Alexandros Dritsoulas was born in Athens Greece in 1986. He graduated in 2010 from the Department of Crop Science in the Agriculture University of Athens with the specialization of Plant Protection and Environment with research focused on biological control using predatory mites. During the final year of his degree he was awarded the European Union’s “Erasmus” fellowship to complete a six-month program of study at the Corvinus University of Budapest in

Hungary. Alexandros graduated Valedictorian in 2013 with a Master of Science Degree in

Integrated and Organic Crop Production. His thesis focused on organic propagation of tobacco.

In 2016, Alexandros received an assistantship for a Ph.D. in Entomology and

Nematology and graduated from the University of Florida in 2020. His Ph.D. research at the

University of Florida has a multi-faceted focus. He compared different sampling methods, maximizing taxonomic coverage in metagenomic studies of environmental samples. He also assessed the potential of next-generation sequencing tools to characterize the natural community structure of entomopathogenic nematodes and soil microarthropods . He improved protocols for the use of this new approach and demonstrated its contribution to nematode biogeography in the first survey of EPNs in a croppnig system in Greece.

Alexandros plans to work as postdoctoral researcher to broaden his experience in his chosen subject area, before returning to Greece to work in agricultural nematology though an academic position. He considers himself fortunate to have trained with exceptional scientists in one of the most prestigious graduate programs in nematology. He hopes to deliver the same enthusiasm and excitement for knowledge to his future students and colleagues that he experienced from his professors in the U.S.A.

118