Table of Contents
Related Titles
Title Page
Copyright
Preface
List of Contributors
Part I: Methods for the Design and Optimizationof New Active Ingredients
Chapter 1: High-Throughput Screening in Agrochemical Research
1.1 Introduction
1.2 Target-Based High-Throughput Screening
1.3 Other Screening Approaches
1.4 In Vivo High-Throughput Screening
1.5 Conclusions
Acknowledgments
References
Chapter 2: Computational Approaches in Agricultural Research
2.1 Introduction
2.2 Research Strategies
2.3 Ligand-Based Approaches
2.4 Structure-Based Approaches
2.5 Estimation of Adverse Effects
2.6 In-Silico Toxicology
2.7 Programs and Databases
2.8 Conclusion
References
Chapter 3: Quantum Chemical Methods in the Design of Agrochemicals
3.1 Introduction
3.2 Computational Quantum Chemistry: Basics, Challenges, and New Developments
3.3 Minimum Energy Structures and Potential Energy Surfaces
3.4 Physico-Chemical Properties
3.5 Quantitative Structure-Activity Relationships
3.6 Outlook
References
Chapter 4: The Unique Role of Halogen Substituents in the Design of Modern Crop Protection Compounds
4.1 Introduction
4.2 The Halogen Substituent Effect
4.3 Insecticides and Acaricides Containing Halogens
4.4 Fungicides Containing Halogens
4.5 Plant Growth Regulators (PGRs) Containing Halogens
4.6 Herbicides Containing Halogens
4.7 Summary and Outlook
References
Part II: New Methods to Identify the Mode of Action of Active Ingredients
Chapter 5: RNA Interference (RNAi) for Functional Genomics Studiesand as a Tool for Crop Protection
5.1 Introduction
5.2 RNA Silencing Pathways
5.3 RNAi as a Tool for Functional Genomics in Plants
5.4 RNAi as a Tool for Engineering Resistance against Fungi and Oomycetes
5.5 RNAi as a Tool for Engineering Insect Resistance
5.6 RNAi as a Tool for Engineering Nematodes Resistance
5.7 RNAi as a Tool for Engineering Virus Resistance
5.8 RNAi as a Tool for Engineering Bacteria Resistance
5.9 RNAi as a Tool for Engineering Parasitic Weed Resistance
5.10 RNAi Safety in Crop Plants
5.11 Summary and Outlook
References
Chapter 6: Fast Identification of the Mode of Action of Herbicidesby DNA Chips
6.1 Introduction
6.2 Gene Expression Profiling: A Method to Measure Changes of the Complete Transcriptome
6.3 Classification of the Mode of Action of an Herbicide
6.4 Identification of Prodrugs by Gene Expression Profiling
6.5 Analyzing the Affected Metabolic Pathways
6.6 Gene Expression Profiling: Part of a Toolbox for Mode of Action Determination
References
Chapter 7: Modern Approaches for Elucidating the Mode of Action of Neuromuscular Insecticides
7.1 Introduction
7.2 Biochemical and Electrophysiological Approaches
7.3 Fluorescence-Based Approaches for Mode of Action Elucidation
7.4 Genomic Approaches for Target Site Elucidation
7.5 Conclusion
References
Chapter 8: New Targets for Fungicides
8.1 Introduction: Current Fungicide Targets
8.2 A Retrospective Look at the Discovery of Targets for Fungicides
8.3 New Sources for New Fungicide Targets in the Future?
8.4 Methods to Identify a Novel Target for a Given Compound
8.5 Methods of Identifying Novel Targets without Pre-Existing Inhibitors
8.6 Non-Protein Targets
8.7 Resistance Inducers
8.8 Beneficial Side Effects of Commercial Fungicides
8.9 Concluding Remarks
References
Part III: New Methods to Improve the Bioavailability of Active Ingredients
Chapter 9: New Formulation Developments
9.1 Introduction
9.2 Drivers for Formulation Type Decisions
9.3 Description of Formulation Types, Their Properties, and Problemsduring Development
9.4 Bioavailability Optimization
9.5 Conclusions and Outlook
References
Chapter 10: Polymorphism and the Organic Solid State: Influenceon the Optimization of Agrochemicals
10.1 Introduction
10.2 Theoretical Principles of Polymorphism
10.3 Analytical Characterization of Polymorphs
10.4 Patentability of Polymorphs
10.5 Summary and Outlook
Acknowledgments
References
Chapter 11: The Determination of Abraham Descriptors and Their Application to Crop Protection Research
11.1 Introduction
11.2 Definition of Abraham Descriptors
11.3 Determination of Abraham Descriptors: General Approach
11.4 Determination of Abraham Descriptors: Physical Properties
11.5 Determination of Abraham Descriptors: Examples
11.6 Application of Abraham Descriptors: Descriptor Profiles
11.7 Application of Abraham Descriptors: LFER Analysis
11.8 Application of Abraham Descriptors: Generality of Approach
Acknowledgments
References
Part IV: Modern Methods for Risk Assessment
Chapter 12: Ecological Modeling in Pesticide Risk Assessment:Chances and Challenges
12.1 Introduction
12.2 Ecological Models in the Regulatory Environment
12.3 An Overview of Model Approaches
12.4 Regulatory Challenges
References
Chapter 13: The Use of Metabolomics In Vivo for the Development of Agrochemical Products
13.1 Introduction to Metabolomics
13.2 MetaMap®Tox Data Base
13.3 Evaluation of Metabolome Data
13.4 Use of Metabolome Data for Development of Agrochemicals
13.5 Discussion
13.6 Concluding Remarks
References
Chapter 14: Safety Evaluation of New Pesticide Active Ingredients: Enquiry-Led Approach to Data Generation
14.1 Background
14.2 What Is the Purpose of Mammalian Toxicity Studies?
14.3 Addressing the Knowledge Needs of Risk Assessors
14.4 Opportunities for Generating Data of Direct Relevance to Human Health Risk Assessment within the Existing Testing Paradigm
14.5 Enquiry-Led Data Generation Strategies
14.6 Conclusions
References
Chapter 15: Endocrine Disruption: Definition and Screening Aspects in the Light of the European Crop Protection Law
15.1 Introduction
15.2 Endocrine Disruption: Definitions
15.3 Current Regulatory Situation in the EU
15.4 US EPA Endocrine Disruptor Screening Program and OECD Conceptual Framework for the Testing and Assessment of Endocrine-Disrupting Chemicals
15.5 ECETOC Approach
15.6 Methods to Assess Endocrine Modes of Action and Endocrine-Related Adverse Effects in Screening and Regulatory Contexts
15.7 Proposal for Decision Criteria for EDCs: Regulatory Agencies
References
Index
Related Titles
Krämer, W., Schirmer, U., Jeschke, P., Witschel, M. (eds.)
Modern Crop Protection Compounds
2012
ISBN: 978-3-527-32965-6
Filho, V. C.
Plant Bioactives and Drug Discovery
Principles, Practice, and Perspectives
2012
ISBN: 978-0-470-58226-8
Walters, D.
Plant Defense
Warding off attack by pathogens, herbivores and parasitic plants
2010
ISBN: 978-1-4051-7589-0
Tadros, T. F. (ed.)
Colloids in Agrochemicals
Colloids and Interface Science
Volume 5 of the Colloids and Interface Science Series
2009
ISBN: 978-3-527-31465-2
The Editors
Dr. Peter Jeschke
Bayer CropScience AG
BCS AG-R&D-CPR-PC-PCC-Chemistry 2
Bldg. 6510
Alfred-Nobel-Str. 50
40789 Monheim
Germany
Dr. Wolfgang Krämer
Rosenkranz 25
51399 Burscheid
Germany
Dr. Ulrich Schirmer
Berghalde 79
69126 Heidelberg
Germany
Dr. Matthias Witschel
BASF SE
GVA/HC-B009
67056 Ludwigshafen
Germany
All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.
Library of Congress Card No.: applied for
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
Bibliographic information published by the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at <http://dnb.d-nb.de>.
© 2012 Wiley-VCH Verlag & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany
All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form — by photoprinting, microfilm, or any other means — nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law.
Print ISBN: 978-3-527-33175-8
ePDF ISBN: 978-3-527-65593-9
ePub ISBN: 978-3-527-65592-2
mobi ISBN: 978-3-527-65591-5
oBook ISBN: 978-3-527-65590-8
Preface
Today, modern agriculture is facing an enormous challenge – namely, that it ensure that sufficient high-quality food is available to meet the needs of a continuously growing population.
In 2011, the world's population exceeded seven billion people, and a prognosis by the United Nations has suggested that by the year 2050 – assuming moderate birth rates – this will increase to as many as 9.1 billion.
Beyond that, losses of agriculturally usable land, climate change, and changes in the eating habits of the peoples of newly industrialized countries will require major improvements to be made in agricultural productivity. In addition to the increasing demand for food in general, people are today requesting a greater protein intake, especially in countries undergoing transition, and this in turn will lead to a higher consumption of the cereals required as feed used for meat production. Coincidentally, these changing food demands are meeting new requests for bioenergy to be produced via agriculture. Climatic changes that influence the distribution of weeds, pests, and diseases, and their prospective consequences for agriculture, represent a further challenge for crop protection. Change in seed breeding and genetically modified (GM) crops demonstrate progressive solutions for better supplies of food by employing technological innovations from both biochemistry and biotechnology. Nevertheless, the traditional research and development of crop protection compounds remains the most effective method for combating losses in agricultural yields. Currently, such losses are in the range of 14% due to competition by weeds, 13% due to damage by fungal pathogens, and 15% by insect damage.
Another very important reason for employing crop protection compounds is to improve the quality of food. For example, mycotoxins produced by species of Fusarium (a fungus that causes damage to the ears of wheat) lead to increasing problems in food production. In addition, changes in rainfall, temperature, and relative humidity can each favor the growth of fungi that produce mycotoxins, so that crops such as groundnuts, wheat, maize, rice, and coffee may become unsuitable for consumption by both humans and animals. Thus, the need for effective research into new crop protection compounds can be fulfilled only by introducing new scientific approaches within the methodology of seeking new active ingredients, by improving the identification process of new targets, by studying aspects of bioavailability, and by improving the tools applied to risk assessment studies of toxicological and ecotoxicological aspects, utilizing new technologies.
This book, which is based partly on Part IV: New Research Methods of the First Edition of the textbook Modern Crop Protection Compounds (Wiley-VCH, 2007), provides details of the progress that has been made during the past few years towards new methods in modern crop protection research. This includes progress not only in chemical synthesis but also in physico-chemical research, the use of biological research progress and the knowledge and application of genetics and proteomics, and the use of mathematical methods in the design and risk assessment of new active ingredients. Consequently, this book will reflect the exclusively broad field of research in the areas of chemistry, biology, biochemistry, formulation research, toxicology, and ecotoxicology that have been used to identify and develop new chemical tools, such that “green” technology can enjoy further success.
The book, which provides a broad overview of a range of current methods used in modern crop protection research, is divided into four Parts that incorporate 15 chapters, each written by renowned experts at the R&D divisions of major agrochemical companies.
Part I presents methods for the design and optimization of new active ingredients. By using modern research techniques and serendipitous, highly specific biological screening systems, significant progress has been achieved during the past 25 years in computational methods for lead identification and optimization, based on molecular structure information and/or quantum chemistry. Additionally, in-silico toxicology approaches to estimate specific risk profiles of agrochemicals will have an emerging impact in the future. In the search for a so-called “optimal product” in modern crop protection in terms of efficacy, environmental safety, user friendliness, and economic viability, the halogen substitution of active ingredients is increasingly recognized as a very important tool.
In Part II are described new methods for identifying the modes of action of active ingredients. Reverse-genetic approaches such as RNA interference (RNAi) offer useful tools to elucidate modes of action, to identify novel targets for exploitation, or to help create new generations of crop protection technologies. For several years, the rapid identification of herbicidal modes of action has been possible via gene expression profiling, using DNA chips. An elucidation of the target sites of neuromuscular insecticides at an early stage in their discovery and development can play an important role in the prioritization of selected candidates. However, despite great technological progress having been made, the targeted discovery of novel fungicides remains an immense challenge because of the restrictions that have been posed on new active ingredients by the obligatory physico-chemical properties permitting a sufficient bioavailability that will, in turn, guarantee fungicidal activity.
In Part III, new methods are examined to improve the bioavailability of the active ingredients. According to novel trends in application technologies, an innovative formulation comprises a mixture of various molecules, besides the active ingredient. In this context, the influence of polymorphism and the organic solid state on the quality and efficiency of agrochemicals plays an important role. Molecular descriptors, as defined by Abraham, can be used to set up linear free energy relationships (LFERs) of relevance to agrochemical research and environmental fate.
Finally, modern methods for risk assessment are addressed in Part IV. Today, many tools are available that can be used to assimilate the knowledge required to evaluate human health and environmental safety, such as exposure modeling, in vitro models to evaluate phenotypic and gene expression changes, computational toxicology, bioinformatics, and systems biology. Despite its complexity and a lack of experience of its use, environmental effect modeling has a great potential for regulatory risk assessments with modern crop protection products, although at present its use is not yet fully accepted. In Chapter 14, entitled Safety Evaluation of New Pesticide Active Ingredients: Enquiry-Led Approach to Data Generation, attention is focused heavily on advances in molecular biology and biotechnology, and how these may be used in conjunction with computational toxicology and bioinformatics to make toxicity testing more relevant to low-level human exposures, to reduce the need for in-vivo testing in animal models, and to make the whole process of hazard data generation quicker and less expensive. In parallel, an evaluation of the endocrine disruption definition and screening aspects in light of the European Crop Protection Law has led to a proposal for decision criteria for endocrine-disrupting compound (EDC) regulatory agencies. This aspect is discussed, taking into consideration the scientific needs of the near future.
We hope that this book will prove to be an invaluable source of information for all of those people working in crop protection science – whether as governmental authorities, as researchers in agrochemical companies, scientists at universities, conservationists, and/or managers in organizations and companies involved with making improvements in agricultural production – to help nourish a continuously growing world population, and to advance the production of bioenergy.
Within this book the authors have named the products/compounds preferably by their common names. Although, occasionally, registered trademarks are cited, their use is not free for everyone. In view of the number of trademarks involved, it was not possible to indicate each particular case in each table and contribution. We accept no liability for this.
May 2012
Peter Jeschke
Wolfgang Krämer
Ulrich Schirmer
Matthias Witschel
List of Contributors
Part I
Methods for the Design and Optimizationof New Active Ingredients
Chapter 1
High-Throughput Screening in Agrochemical Research
Efficient and economical agriculture is essential for sustainable food production fulfilling the demands for high-quality nutrition of the continuously growing population of the world. To ensure adequate food production, it is necessary to control weeds, fungal pathogens, and insects, each of which poses a threat of yield-losses of about 13–15% before harvest (Figure 1.1). Although a broad range of herbicides, fungicides and insecticides already exists, shifts in target organisms and populations and increasing requirements necessitate a steady innovation of crop-protection compounds.
Weeds, fungal pathogens and insects belong to evolutionary distinct organism groups (Figure 1.2), which makes it virtually impossible to have a single crop-protection compound capable of addressing all pest control problems. On closer examination, even the grouping of pests simply as insects, fungi and weeds is, in many cases, still an insufficient depiction. Although the term “insecticide” is often used for any chemical used to combat insects, spider mites or nematodes, the differences between these organisms are so significant that it is more precise to speak of insecticides, acaricides, and nematocides. Among plant pathogenic fungi, the evolutionary range is even much broader and oomycetes are not fungi at all, although oomyceticides commonly are also commonly referred to as “fungicides”. Hence, the agrochemical screening of fungicides and insecticides requires a substantial range of diverse species. The situation for herbicide screening is, in some ways, the reverse, but is no easier. Indeed, the close genetic similarity between crop and weed plants generates challenges with regards to the specificity of herbicidal compounds, in differentiating between crop and weed plants. This also results in a need to use a range of different crop and weed plants in screening programs.
In light of the above circumstances, agrochemical screening has employed, in both laboratory and glass-house trials, a wide spectrum of model and pest species. The recent developments described in this chapter, however, have allowed an even higher throughput not only in glass-house tests on whole organisms, but also the exploitation of biochemical (in vitro) target tests. Not surprisingly, the implementation of molecular screening techniques and the “omics” technologies – functional genomics, transcriptomics and proteomics, etc. – into agrochemical research has been a major challenge due to the high diversity of the target organisms [5].
Molecular agrochemical research with biochemical high-throughput target screening commenced with several model species, each of which was chosen mainly because of their easy genetic accessibility or specific academic interests. These first favorite model organisms of geneticists and molecular biologists were largely distinct from the most important pest species in agriculture, however. Nonetheless, recent progress in genome sequencing has led to a steadily growing knowledge about agronomically relevant organisms (Figure 1.3 and Table 1.1).
Organisms | ||
Plants | Fungi and oomycetes | Insects and nematodes |
Dicotyledonous plants | Ascomycetes | Diptera |
Arabidopsis thalianaa | Saccharomyces cerevisiaea | Drosophila melanogastera |
Brassica oleracea | Alternaria brassicicola | Musca domestica |
Glycine max | Aspergillus oryzaea | Aphids |
Lotus corniculatus | Botryotinia fuckeliana | Acyrthosiphon pisum |
Solanum tuberosuma | Gibberella zea | Lepidoptera |
Monocotyledonous plants | Magnaporthe griseaa | Bombyx moria |
Oryza sativaa | Mycosphaerella graminicola | Coleoptera |
Sorghum bicolor | Neurospora crassa | Tribolium castaneum |
Triticum aestivum | Podospora anserinaa | Hymenoptera |
Zea mays | Sclerotinia sclerotiorum | Aphis melifera |
Brachypodium distachyon | Ustilaginomycetae | Nematodes |
Setaria italica | Ustilago maydisa | Caenorhabditis elegansa |
Hordeum vulgare | Uredinomycetae | Meloidogyne incognita |
Puccinia graminis | ||
Phakopsora pachyrhizi | ||
Oomycetes | Basidiomycetes | |
Hyaloperonospora arabidopsis | Phanerochaete chrysosporium | |
Phytophtora infestansa | Laccaria bicolor | |
Pythium ultimum | Zygomycota | |
Rhizopus oryzae |
a.Completed or close to completion, otherwise: in progress.
The situation is relatively simple for weeds, as all plants are closely related (Figure 1.2). The first model plant to be sequenced, Arabidopsis thaliana, is genetically not very distinct from many dicotyledonous weeds, and the monocotyledonous crops are closely related to the monocotyledonous weeds which, in turn– starting several thousand years ago – formed the foundation for today's cereals species. The first sequenced insect genome of Drosophila melanogaster, a dipteran insect, was exploited extensively in both genetic and molecular biological research. To better reflect relevant pest organisms such as lepidopteran pests or aphids, species such as Heliothis virescens (tobacco budworm) and Myzus persicae (green peach aphid) have been investigated by the agrochemical industry, while Bombyx mori, Acyrthosiphon pisum and Tribolium castaneum have been sequenced in public projects (Table 1.1). Baker's yeast, Saccharomyces cerevisiae, has long been the most commonly used model fungus, while the ascomycete Magnaporthe grisea and the ustilaginomycete Ustilago maydis have been the first sequenced relevant plant pathogens. It is certain that, within the next few years, even the broad evolutionary range of the many different plant pathogenic fungi and oomycetes (see Figure 1.2) will be included in genome projects.
The progress of molecular biology of agronomically relevant organisms has enabled the introduction of target-based biochemical (in vitro) high-throughput screening (HTS), which has significantly changed the approach to the screening for agrochemicals during the past 15 years. Target-based HTS is a technology utilized in the agrochemical industry to deliver new actives with defined modes of action (MoA) [6].
Most major research-based agrochemical companies have established biochemical HTS, often conducted in cooperation with companies having special expertise in specific fields of biotechnology. The first wave of genomics – which included genome-wide knock-out programs of model organisms – indicated that about one-quarter of all genes are essential; that is, they were lethal by knock-out [6–8]. The resulting high number of potential novel targets for agrochemicals must be further investigated to clarify the genes' functions (reverse genetics) and to better understand their role in the organism's life cycle. Although the technology of genome-wide knock-out itself was highly efficient and well established, it transpired that even the knock-out of some known relevant targets were not lethal, either because of genetic or functional redundancy, counter-regulation, or because a knock-out does not perfectly mimic an agonistic drug effect on, for example, ion channels. Consequently, knock-out data are today reviewed critically with respect to as many aspects as possible of the physiological roles of potential targets and, as a result, they are taken as just one argument for a gene to be regarded as an interesting potential target. It must also be considered that clarification of the genes' functions is a challenging and resource-consuming task, and that attention is perhaps more often focused on targets with a sound characterization of their physiological role.
The best proof for an interesting agrochemical target is the “chemical validation” by biologically active compounds. This is true for all the established targets. However, new chemical hits acting on such targets must have an advantage over the already known compounds. This may be a chemical novelty, a novel binding site, an increased performance, or providing a means to overcome resistance. From the standpoint of innovation and the chance to open new areas, novel targets are of particular interest, especially when active compounds are already known, such as a natural product (e.g., the ryanodine receptor for insecticides). Most interesting are novel and proprietary targets which arise from genetics programs or from MoA discovery. MoA elucidation for biological hits has, therefore, become much more important.
Modern analytical methods such as high-performance liquid chromatography/mass spectrometry (HPLC/MS), electrophysiology, imaging, and others build a gateway to today's novel target discovery. The benefit of electrophysiology for clarifying neurophysiological effects is obvious. Cellular imaging techniques complement electrophysiology and are, furthermore, a general approach for MoA studies. For metabolic targets, such as those of sterol biosynthesis, direct target identification may be possible by metabolite analysis [9, 10]. For such compounds gene expression profiling has also proved to be a valuable tool for the MoA classification [11, 12]. When used as fingerprint methods, metabolite profiling and gene expression profiling allows a rapid and reliable detection method for known MoA, and a clear identification and classification of unknown modes of action. Yet, despite the extensive progress in technology, MoA elucidation of novel targets is still – and will be for the near future – a highly demanding challenge. Only the combination of all available methodologies, with emphasis placed on traditional careful physiological and biochemical examinations, will reveal a clearly identified novel molecular target [13].
During the past decades, the identification of resistance mutations to pesticides has provided one of the most clear-cut approaches to target clarification. Although the technological progress has considerably fostered throughput in screening for mutations with a certain phenotype – so-called “forward genetics” [14] – it yet does not seem to be a reliable source of novel targets.
Once a target has been envisaged, further criteria for a “good” target are applied. Clearly, the most important criterion is the druggability of a target, which means accessibility by agro-like chemicals (see below) [15]. It is no coincidence, that the best druggable targets have preexisting binding niches, favoring ligands that comply with certain physico-chemical properties. Furthermore, the target should be relevant during the damaging life phase of a pest, and the destructive effect on a weed or pest under practical conditions should occur within a short period of time after treatment.
Having cleared all of these hurdles, an interesting target must be assayable in order to be exploited, which in turn makes assay technology capabilities a crucial asset. Overall, the number of promising targets remaining is at least two orders of magnitude lower than the number of potential targets found by gene knock-out [6]. Yet, even after having made such great efforts it still difficult to predict whether or not a new active ingredient will be identified, and whether a novel target finally will be competitive in the market.
Often, pharmaceutical research is systematically concentrated towards particular target classes, an example being protein kinases in cancer research [16]. Thereby, know-how can be accumulated and specialized technology can be concentrated for a higher productivity [17]. A successful target triggers the attention to the next similar targets, leading to a considerable understanding of, for example, the human kinome [18, 19]. A similar approach in agrochemical research is of limited value, as there are no such privileged target classes (Figure 1.4). In fact, the common denominator of the diverse agrochemical targets often is the destructive character of the physiological consequences of interference with the target's function, sometimes even being a “side-effect,” such as the generation of reactive oxygen species (ROS) [6]. Nevertheless, there are exceptions – one of which is the class of protein kinases – which have been identified as a promising target class for fungicides [20–22].
In pharmaceutical research, HTS [24] has proven to be a major source of new lead structures [23], thereby motivating agrochemical research to incorporate – at least in part – this approach into the pesticide discovery process. At Bayer CropScience for example, the first HTS systems were set up during the late 1990s, after which the screening capacity expanded rapidly to more than 100 000 data points per day on a state-of-the-art technology platform. This included fully automated 384-well screening systems, a sophisticated plate replication and storage concept, a streamlined assay validation, and a quality control workflow. An expansion of the compound collection with the help of combinatorial chemistry and major investments in the development of a suitable data management and analysis system was also initiated.
The concept allows the screening of large numbers of compounds as well as large numbers of newly identified targets, thus yielding a corresponding number of hits. The simultaneously developed quality control techniques were able to separate valid hits from false-positives and/or uninteresting compounds due to various reasons (e.g., unspecific binding). Interestingly, several target assays deliver considerable numbers of in vivo active compounds, while for other in vitro HTS assays the often remarkable target inhibition was not transferred into a corresponding in vivo activity. In some cases, this can be attributed to an insufficient target lethality of more speculative targets or “Agrokinetic” factors for in vivo species. As discussed earlier, the value of a thorough validation of (i) targets, (ii) assays, and (iii) chemical hits becomes evident.
The extended target validation led to increased numbers of target screens with in vivo active compounds. Hence, even more time could be spent on the hit validation, namely the introduction of control tests to eliminate, for example, readout interfering compounds (i.e., hits that were found only due to their optical properties or chemical interference with assay components).
The process of continuous improvement has to date shifted to an extended characterization of hits with respect to reactivity, binding modes [25] (competitive/non-competitive, reversible/irreversible, and so on [26]), speed of action and erratic inhibition due to “promiscuous” behavior of the compound class among others [27]. At the same time – if feasible – the hits or hit classes are submitted to orthogonal assays such as electrophysiology in case of neuronal targets, that help to further classify and validate the hits independent of the readout.
During the late 1990s, Bayer CropScience followed the trend introduced by pharmaceutical companies of conducting genomic projects in collaboration with a biotech-partner. Unfortunately, however, although this genomic approach provided more than 100 new screening assays, it did not deliver the desired output.
Hence, about five years ago the target-based screening approach was redirected, with the new direction subsequently leading to the following favorable changes:
Of great interest also was the observation that the relative percentage of enzyme assays compared to cell-based assays has increased (Figure 1.5) over the past 10 years. This finding reflects not only technological progress that has been made, but also the increasing back-concentration on ion channel targets for insecticides, which are especially highly validated targets.
At the same time, the chemical libraries at Bayer CropScience became more diversity-oriented, with major efforts being made to further increase the quality of the compound collections (see below), with especially careful quality checks of the hit compounds.
All of these measures together have greatly increased the proportion of true hits, so that finally the chemistry capacities are concentrated on fewer, albeit well-characterized, hit classes with a clearly increased likelihood of a successful hit-to-lead optimization.
The huge amount of data and information generated during the various phases of HTS and subsequent validation processes has triggered the development of sophisticated data analysis tools [28] that help biologists and chemists to select and prioritize the most promising hits or hit classes (cluster of similar compounds) (Figure 1.6).
Biochemical in vitro screening may deliver compounds that, despite a clear target activity, are unable to exhibit in vivo activity due to, for example, unfavorable physico-chemical properties (lacking bioavailability), rapid metabolism, insufficient stability, or a poor distribution in the target organism. Nevertheless, these chemical classes are still of interest to chemists because such properties reflect the characteristics of the compounds that may be overcome by chemical optimization. As a consequence, “agrokinetics” has led to the identification of pure in vitro hits as such, and also helped to elucidate the reasons for failure in the in vivo test, thus guiding the in vitro to in vivo transfer of hit classes. Such observations underline the fact that the in vitro and in vivo screening processes can be complementary, and together can be used to broadly characterize the activity of test compounds within the early discovery process.
Currently, two trends can be observed among the high-throughput community: (i) miniaturization into the nanoliter dispensing regime; and (ii) new high-content screening (HCS) techniques. The small-volume screening (either on 1536-well plates or the recently introduced low-volume 384-well plates) clearly is also of interest for agrochemical research, since the enzymes and substrates of new target proteins are often difficult and costly to produce in larger quantities. Due to the above-mentioned screening strategy this process is not so much driven by the need to further increase the capacity, but rather by cost efficiency, the standard reaction volume having decreased from more than 50 µl to 5–10 µl (Figure 1.7). Moreover, further reductions are possible with new pipetting equipment having now reached a robust quality with inaccuracies of below 5% in the 1 µl range.
Other very important aspects of ion channel screening are the recently developed automated and medium-throughput patch clamping systems that perfectly meet the increased demand for in-depth hit characterization. Yet, the future role of HCS – fully automated confocal life cell microscopy imaging systems – is less clear than in pharmaceutical research, where it has become the validation and screening method development of the past few years [29]. Nonetheless, the applicability of HCS to agrochemical research will need to be evaluated in the future.
During the past two decades, computational chemistry has become a key partner in drug discovery. Indeed, one of its main contributions to high-throughput methods is that of virtual screening [24, 30], a computational method that can be applied to large sets of compounds with the goal of evaluating or filtering those compounds against certain criteria, prior to or in lieu of in vivo or in vitro testing. In this regard, some methods consider target structure information while others are based solely on ligand similarity to complex model systems. Additionally, when three-dimensional information is incorporated into an analysis, the calculation becomes more demanding, especially if a flexible target protein is considered. Although massive screening with fully flexible models is not yet feasible, the so-called flexible docking of huge (both real and virtual) compound collections into a rigid binding pocket has today become routine [31]. The most obvious advantage of the latter method over the relatively fast similarity searches is that any compound which has binding site complementarity will be identified, and that no similarity to a known ligand is needed. This stands in contrast to similarity-based screening, where completely new scaffolds are rarely found.
In order to have a reasonable hit enrichment when using docking methods, computational chemistry must start with high-resolution protein structures; if possible, more than one ligand co-crystal would be used to construct the binding domain. In addition, some programs are capable of handling a certain degree of target flexibility through ensemble formations of binding domains from various experimental structures [32]. Whilst the quality of the results will obviously improve, a greater computational effort will be required as a consequence.
Virtual-target-based screening can be applied in many ways, the most obvious being the screening of huge libraries in order to prioritize the synthesis, acquisition and/or biochemical screening, or to select reactants for combinatorial libraries that show highest hit likeliness. These applications do yield target-focused libraries, and can be extended to families of targets, such as kinases or G-protein-coupled receptors (GPCRs).
Since the very beginning of the search for new agrochemicals, in vivo screening has been the primary basis for agrochemical research, leading to the identification and characterization of new active chemistries and their subsequent optimization. In 1956, 1800 compounds needed to be evaluated for every one that became a product, a number that had risen to 10 000 by 1972 [33]. By 1995, the number has risen to more than 50 000, and today is about 140 000 compounds tested per product discovered [34]. In part, this rise is due to the increasing demands with regard to the need for increased biological activity, improved mammalian and environmental safety, as well as a variety of economic considerations. Beginning in the mid-1990s, most of the major agrochemical companies established in vivo HTS systems [15, 35–38], an interest which coincided with the development and expanding use of combinatorial libraries. In the HTS systems, the numbers of compounds screened each year are reported as ranging between 100 000 and 500 000, with most programs utilizing less than 0.5 mg of substance to produce relevant answers for a targeted set of plants, insects and fungi, using either 96-well or 384-well microtiter plates (MTPs) (Figure 1.8). Such HTS systems can produce a large number of hits, all of which are dependent on the screening dose, pass criteria, and the number and type of test species used. The quality of the hits from the HTS can be improved through the addition of extra dose rates and replicates [15], which can in turn improve the quality of the hits delivered to relevant follow-up screens.
HTS programs are based on automation, miniaturization, and often also the use of model organisms or systems which are easy to handle and adaptable to the MTP format. In pesticide discovery programs, model systems using Aedes aegypti, D. melanogaster, A. thaliana, Caenorhabditis elegans [39] or cell-growth-based fungicide assays can be successful in identifying a large number of hits. These model systems, using species that can be highly sensitive, are primarily intended to identify biological activity. However, in follow-up tests with agriculturally relevant species the number of interesting compounds often decreases dramatically due to a weak translation between the model organisms and the real pest species. As such, HTS systems with model organisms can potentially miss relevant hits (Figure 1.9).
As a consequence of this less-than-ideal translation, there has been an evolution among in vivo HTS systems to incorporate more relevant target organisms [40], particularly for insecticides and fungicides. For example, 96-well MTP assays involving pest lepidopteran larvae are widely used [15, 41, 42], while leaf-disc assays have been developed [6, 33, 43] that have been adapted by many companies for sap-feeding insects such as aphids.
HTS systems for fungicides utilize cell growth tests, but also cover only a part of the relevant target organisms; all obligate pathogens such as mildews or rusts cannot be tested. Additionally, such cell tests do not test the relevant phases of the development of fungal pathogens on living plant tissues. However, this gap can be closed by using leaf discs [44, 45] or whole plants with relevant fungal species.
The development and further improvement of the more relevant HTS assays using target pest species for insecticides and fungicides is an on-going challenge. In many cases, these assays can be significantly more complex, and the time and effort required to run target organism assays can be greater than was required for previous model systems. As such, the number of species screened in an in vivo HTS has often been reduced to just a few, with one or two model species as general indicators of biological activity, plus perhaps a couple of specific pests that represent major product areas. For example, in the case of insecticides many discovery programs focus on one or two lepidopteran species that serve as indicators for a broad range of chewing pests, and an aphid species that is an indicator for a broad range of sap-feeding insect pests. While these two product areas do not denote the total insecticide market, they do capture the largest segments. Thus, the use of these more complex HTS systems requires a balance relative to throughput and dedicated resources for an in vivo HTS program. The net result is that better-characterized compounds with a more relevant biological profile are derived from HTS programs that focus on representative pest insects.
Figure 1.10