The Genetic Regulation of Gene Expression, Transcriptional Networks, and Leaf Development in the Perennial Model Plant P...

Permanent Link: http://ufdc.ufl.edu/UFE0041074/00001

Material Information

Title: The Genetic Regulation of Gene Expression, Transcriptional Networks, and Leaf Development in the Perennial Model Plant Populus
Physical Description: 1 online resource (165 p.)
Language: english
Creator: Drost, Derek
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009


Subjects / Keywords: Plant Molecular and Cellular Biology -- Dissertations, Academic -- UF
Genre: Plant Molecular and Cellular Biology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Most heritable phenotypic traits are affected by the concurrent inheritance of alleles at multiple loci. As opposed to simple patterns of single gene inheritance, polygenic traits result in complex patterns of segregation, which are frequently assessed through quantitative trait loci (QTL) mapping methods. While QTL detection is a commonplace task in modern biology, cloning specific genetic variants underlying QTL is not ? the task remains a challenge often associated with laborious and time-consuming fine mapping. Recently, one novel method has been proposed to speed the identification of candidate genes and polymorphisms from QTL. ?Genetical genomics? relies on allying traditional quantiative trait analysis to higher-level genomic data, including whole-genome expression or proteomic data. By considering these genomic ?phenotypes? together with morphological or molecular phenotypes of interest, the role of genetics in these traits' regulation can be compared. QTL coordinately affecting phenotypes and transcript or protein abundance indicate a putative causal relationship between the phenotype and gene(s) represented by those transcripts or proteins, which can be functionally tested. Genetical genomics has recently demonstrated significant promise to speed QTL cloning, and has implicated co-expressed gene networks underlying disease and complex phenotypes in model species. Here, I directly implement genetical genomics methods aimed to reduce the challenges associated with moving from QTL to causal polymorphisms in poplar. First, I demonstrate a novel genotyping and genetic mapping approach for outcrossing species, by leveraging RNA-based molecular markers for high-resolution genetic analysis in Populus. These results lay a solid foundation for QTL and expression QTL (eQTL) mapping by increasing marker coverage and reducing breadth of QTL confidence intervals. Secondly, I apply whole-genome eQTL data to identify co-expressed transcriptional networks in three diverse plant tissues. These results build upon the understanding of transcriptional regulation in all plant species. Finally, I utilize genetical genomics to discover a candidate gene for leaf morphological variation in Populus, a high value phenotype for photosynthetic productivity and a vital trait for evolutionary classification of genus members into evolutionary sections. Taken together, these results paint a favorable picture for additional forward-genetic, genome-level studies to characterize complex phenotypes in poplar and other forest trees.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Derek Drost.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Kirst, Matias.
Local: Co-adviser: Peter, Gary F.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041074:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041074/00001

Material Information

Title: The Genetic Regulation of Gene Expression, Transcriptional Networks, and Leaf Development in the Perennial Model Plant Populus
Physical Description: 1 online resource (165 p.)
Language: english
Creator: Drost, Derek
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009


Subjects / Keywords: Plant Molecular and Cellular Biology -- Dissertations, Academic -- UF
Genre: Plant Molecular and Cellular Biology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Most heritable phenotypic traits are affected by the concurrent inheritance of alleles at multiple loci. As opposed to simple patterns of single gene inheritance, polygenic traits result in complex patterns of segregation, which are frequently assessed through quantitative trait loci (QTL) mapping methods. While QTL detection is a commonplace task in modern biology, cloning specific genetic variants underlying QTL is not ? the task remains a challenge often associated with laborious and time-consuming fine mapping. Recently, one novel method has been proposed to speed the identification of candidate genes and polymorphisms from QTL. ?Genetical genomics? relies on allying traditional quantiative trait analysis to higher-level genomic data, including whole-genome expression or proteomic data. By considering these genomic ?phenotypes? together with morphological or molecular phenotypes of interest, the role of genetics in these traits' regulation can be compared. QTL coordinately affecting phenotypes and transcript or protein abundance indicate a putative causal relationship between the phenotype and gene(s) represented by those transcripts or proteins, which can be functionally tested. Genetical genomics has recently demonstrated significant promise to speed QTL cloning, and has implicated co-expressed gene networks underlying disease and complex phenotypes in model species. Here, I directly implement genetical genomics methods aimed to reduce the challenges associated with moving from QTL to causal polymorphisms in poplar. First, I demonstrate a novel genotyping and genetic mapping approach for outcrossing species, by leveraging RNA-based molecular markers for high-resolution genetic analysis in Populus. These results lay a solid foundation for QTL and expression QTL (eQTL) mapping by increasing marker coverage and reducing breadth of QTL confidence intervals. Secondly, I apply whole-genome eQTL data to identify co-expressed transcriptional networks in three diverse plant tissues. These results build upon the understanding of transcriptional regulation in all plant species. Finally, I utilize genetical genomics to discover a candidate gene for leaf morphological variation in Populus, a high value phenotype for photosynthetic productivity and a vital trait for evolutionary classification of genus members into evolutionary sections. Taken together, these results paint a favorable picture for additional forward-genetic, genome-level studies to characterize complex phenotypes in poplar and other forest trees.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Derek Drost.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Kirst, Matias.
Local: Co-adviser: Peter, Gary F.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041074:00001

This item has the following downloads:

Full Text




2 2009 Derek R. Drost


3 To Mom, Dad, and Abigail


4 ACKNOWLEDGMENTS I would like to gratefully acknowledge the contributions of my graduate committee supervisors, Dr. Matias Kirst and Dr. Gary Peter, for their guidance, support, and committment to the success of my graduate education. Furthermore, I am grateful to Dr. Alic e Harmon and Dr. Marta Wayne for their advice and service as members of my supervisory committee. I would also recognize the contributions of several University of Florida post -doctoral scholars, research scientists, and graduate students to the success of this research: Evandro Novaes, Carolina Novaes, Chris Dervinis, Dr. Catherine Benedict, Brianna Miles, Dr. Alison Morse and Ryan Brown. Additionally, I thank the faculty staff, and students of the PMCB program for both their friendship and for lively dis cussions about science. I also thank Dr. Donald J. Lee (Professor of Agronomy, University of Nebraska -Lincoln) and Dr. Jerry Tuskan (Distinguished Scientist, Oak Ridge National Laboratory) for their continuing mentorship of my academic and professional dev elopment. Finally, and most importantly, I am indescribably grateful to my parents, Dirk and Sue, and my sister Abigail, for their steadfast love and support throughout all aspects of my education and life.


5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 8 LIST OF FIGURES ............................................................................................................................ 10 ABSTRACT ........................................................................................................................................ 12 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW ................................................................. 14 Poplar as a Reference for Tree Biology and Genetics .............................................................. 15 Quantitative Genetics in the Populus System ........................................................................... 18 Currently Established Genetic Marker Resources ............................................................. 18 Linkage/QTL Population Structures and Associated Analysis Methods ......................... 21 Quantitative Genetic Analyses of Gene Expression: Genetical Genomics ...................... 23 Leaf Morphological Variation in Populus ................................................................................. 26 Evolutionary Conservation and Previous Analyses of L eaf Morphology ........................ 26 Candidate Genes and Pathways Discovered in Arabidopsis ............................................. 27 Project Objectives ....................................................................................................................... 30 2 A MICROARRAY -BASED GENOTYPING AND GENETIC MAPPING APPROACH FOR HIGHLY HETEROZYGOUS OUT CROSSING SPECIES ENABLES LOCALIZATION OF A LARGE FRACTION OF THE UNASSEMBLED Populus trichocarpa GENOME SEQUENCE ......................................................................................... 38 Introduction ................................................................................................................................. 38 Materials and Methods ................................................................................................................ 40 Plant Growth Conditions and RNA Isolation .................................................................... 40 Microsatellite (SSR) Genotyping and Framework Map Construction ............................. 41 Microarray Analysis of Parental Genotypes ...................................................................... 42 Microarray Analysis of Family 52 124 .............................................................................. 43 Grouping, Ordering, and Mapping of SSRs, GEMs, and SFPs to Linkage Groups ........ 44 SequenceLevel Characterization of SFP Alleles .............................................................. 46 Results .......................................................................................................................................... 46 SSR Framework Map of Genotype 52225 ........................................................................ 46 Identification of Probes for Genotyping Family 52124 ................................................... 47 Identification of Probes for Transcript Profiling Family 52 124...................................... 49 Genotyping SFP and GEM Probes in the Progeny of Family 52 124.............................. 50 Genetic Mapping of Genotype 52225 ............................................................................ 51 Physical Orientation of the 52225 Genetic Map ........................................................... 52 Genetic Mapping of the Unassembled Populus Genome .................................................. 52 Verification of Map Position for Unassembled Sequence Scaffolds ............................... 53


6 Characterization of Sequence -Level Allelic Variation Represented By Mapped SFPs .................................................................................................................................. 54 Discussion .................................................................................................................................... 55 3 THE ONTOGENY OF THE GENETIC REGULATION OF GENE EXPRESSION AND TRANSCRIPTIONAL NETWORKS IN THE WOODY PERENNIAL MODEL Populus ........................................................................................................................................ 72 Introduction ................................................................................................................................. 72 Materials and Methods ................................................................................................................ 74 Plant Material and Growth Conditions ............................................................................... 74 RNA Isolation and Microarray Analysis ............................................................................ 74 eQTL Analysis ..................................................................................................................... 75 eQTL Hotspot Detec tion and Analysis .............................................................................. 76 Hotspot Based Co -Expression Network Construction ...................................................... 76 GO Annotation and Enrichment Testing ............................................................................ 76 Cis -Element Detection and Enrichment Testing ............................................................... 77 Results .......................................................................................................................................... 78 eQTL Detection and Genome Distribution ........................................................................ 78 Genetic Regulation of Gene Expression Is Largely Tissue Specific ................................ 79 Identification of Tissue -Specific eQTL Hotspots .............................................................. 80 Construction of Tissue Specific, Hotspot Based Co Expression Networks .................... 80 Gene OntologyBased Annotation of Tissue Specific Gene Co Expression Networks........................................................................................................................... 81 eQTL Based Prediction of Putative Network Regulators ................................................. 82 Enrichment of Transcription Factor Binding Sites in Co -Expression Networks ............ 83 Transcriptional Networks Shared Between Tissues Are Re gulated By Distinct Loci .... 85 Discussion .................................................................................................................................... 87 4 UTILIZING GENETICAL GENOMICS TO IDENTIFY AN ADP -RIBOSYLATION FACTOR, P t ARF1 AS A CANDIDATE GENE FOR LEAF SHAPE VARIATION IN Populus ...................................................................................................................................... 106 Introduction ............................................................................................................................... 106 Materials and Meth ods .............................................................................................................. 108 Plant Material and Phenotyping ........................................................................................ 108 Genotyping and Genetic Mapping of Progeny ................................................................ 108 Microarray Analysis .......................................................................................................... 109 QTL and eQTL Analysis ................................................................................................... 109 ARF1 cDNA Cloning and Sequencing ............................................................................. 110 ARF1 Promoter Isolation and Sequencing ....................................................................... 111 ARF1 Mutagenesis ............................................................................................................ 111 Nisqually Leaf Disc Expression Experiment ................................................................... 112 Verification of Allele -Specific Expression Effects in Segregating Population Subset 113 Results ........................................................................................................................................ 113 Identification of a Majo r QTL for Leaf Blade Width ...................................................... 113


7 Gene Expression Analysis of Leaf Tissue Identifies PtARF1 as a Candidate Gene for Lamina Shape ........................................................................................................... 114 Isolation of Interspecific ARF1 Coding and Promoter Polymorphisms ......................... 117 Localized Expression of ARF1 in Expanding Leaves ..................................................... 118 Verification of Allele -Specific Expression Effects for ARF1 ......................................... 118 Discussion .................................................................................................................................. 119 5 CONCLUSIONS ....................................................................................................................... 138 LIST OF REFERENCES ................................................................................................................. 146 BIOGRAPHICAL SKETCH ........................................................................................................... 165


8 LIST OF TABLES Table page 1 1 Range, leaf morphology, and species count associated with sections of genus Populus ................................................................................................................................... 36 1 2 Frequency of polymorphism across chromos omes in the genome of Populus trichocarpa, clone Nisqually 1 .............................................................................................. 37 2 1 Primer sequences for amplificatio n of scaffold anchored SSR loci .................................... 65 2 2 Summary of F tests for fixed effects in the mixed ANOVA conducted on parent tree microarray data ....................................................................................................................... 66 2 3 Primer sequences for amplification of sequence -verified SFP alleles ................................ 67 2 4 Summary statistics for P. trichocarpa X P. deltoides clone 52225 microa rray and SSR -based linkage map ......................................................................................................... 68 2 5 Summary of WGS scaffold sequences placed based on SFP and GEM markers, a nd resultant estimated coverage .................................................................................................. 69 2 6 Verification of scaffold map location for nine sequence scaffolds using SSR ma rkers and the framework SSR map ................................................................................................. 70 2 7 Verification of scaffold map location for six sequence scaffolds originally mapped based on GEM markers .......................................................................................................... 71 3 1 Summary of eQTL detected for each of the three poplar tissues. ....................................... 97 3 2 Significant eQTL hotspots by linkage group in each of the three tissues. ......................... 98 3 3 Summary of tissue -specific eQTL hotspot -based co -expression networ k construction in Family 52 124 .................................................................................................................... 99 3 4 Summary of tissue -specific eQTL hotspot -based co -expression network construction in Family 52 124. ................................................................................................................. 100 3 5 Tissue -specific eQTL hotspots detected in leaf tissue among Family 52 124 and characteristics of their as sociated co -expression networks ............................................... 101 3 6 Annotation of members and putative regulators of a leaf -specific coexpression network enriched for chloroplast biogenesis and functionality. ........................................ 102 3 7 Poplar genes from the sugar associated co -expression network on linkage group IX in xylem that were also regulated by exogenous sucrose in a separate study conducted in Arabidopsis .................................................................................................... 105


9 4 1 Primers utilized to amplify microsatellite loci across the major linkage group X QTL for leaf lamina characters in Family 52 124. ...................................................................... 131 4 2 Primer sequences utilized for cloning and qRT -PCR experiments. .................................. 132 4 3 Primer sequences utilized for Genome Walker promoter PCRs ....................................... 133 4 4 Summary of QTL detected for leaf lamina shape characters in P. trichocarpa X P. deltoides X P. deltoides Family 52124.............................................................................. 134 4 5 Phenotypic -expression correlation for genes with significant cis and trans -eQTL regulated by t he major lamina shape QTL locus ................................................................ 135 4 6 Genes with eQTL in major lamina character QTL region that are among the top 5000 correlated genes in genome ................................................................................................. 136 4 7 Expression characteristics of ARF -type gene family members relative to ARF1 ........... 137


10 LIST OF FIGURES Figure page 1 1 Overview of the pseudobackcross pedigree construction method for outcrossing forest trees ............................................................................................................................... 31 1 2 Hypothetical examples of cis and trans acting eQTL regulation ...................................... 32 1 3 A hypothetical example of candidate gene identification by the genetical genomics method ..................................................................................................................................... 33 1 4 A hypothetical example of a mechanism of network construction and phenotypic association from genetical genomics eQTL data in a segregating population ................... 34 1 5 Leaf morphologies assoicated with the four most populated evolutionary sections of genus Populus ......................................................................................................................... 35 2 1 Framework SSR map of P. trichocarpa X P. deltoides clone 52225 ............................. 60 2 2 Examples of significant fixed effects detected by analysis of variance of microarray data from the parents of Family 52 124................................................................................ 61 2 3 Microarray and SSR -based genetic map of P. trichoc arpa X P. deltoides 2 225 ........ 62 2 4 Allelic variations characterized by sequencing genomic DNA regions cor responding to mapped SFP probes ........................................................................................................... 64 3 1 Global distribution of eQTL across linkage groups for xylem, leaf, and root tissues expressed as the fraction of mapped gene models ............................................................... 92 3 2 Global distribution of all eQTL (ambiguous and unambiguous) across linkage groups for xylem, leaf, and r oot tissues ............................................................................................ 93 3 3 Overlap between probes and eQTL detected among the three tissues considered. All subparts exclude pr obes with ambiguously positioned eQTL ............................................. 94 3 4 Genome -wide linkage scan of expression traits and demarcation of eQTL hotspots produci ng co -expressed gene networks in leaf tissue .......................................................... 95 3 5 Leaf co -expression network constructed from the "blue" eQTL hotspot in Figure 3 4 and enriched for chloroplast (CP) related Gene Ontogeny categories ............................... 96 4 1 Leaf lamina shape variation among parents and progeny of the P. trichocarpa X P. deltoides pseudobackcross pedigree Family 52124 .......................................................... 124 4 2 Genome -wide composite interval mapping scan for leaf lamina shape characters in Family 52 124 ...................................................................................................................... 125


11 4 3 Fine -scale mapping of the major lamina width QTL on LG X in the segregating pedigree ................................................................................................................................. 126 4 4 Sequence alignment for ARF1 upstream regions isolated from P. trichocarpa and P. deltoides ................................................................................................................................ 127 4 5 Quantitative PCR analysis of regional ARF1 expression in differently staged expanding leaves of P. trichocarpa .................................................................................... 128 4 6 Quantitative RT PCR verification of genetic effects on ARF1 expression in the segregating population ......................................................................................................... 129 4 7 A hypothesized molecular mechanism for how differential ARF1 expression affects leaf lamina phenotypic variation in P. trichocarpa and P. deltoides ................................ 130


12 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 GENETIC REGULATION OF GENE EXPRESSION, TRANSCRIPTIONAL NETWORKS, AND LEAF DEVELOPMENT IN THE PERENNIAL MODEL PLANT Populus By Derek R. Drost December 2009 Chair: Matias Kirst Co -chair: Gary Peter Major: Plant Molecular and Cellular Biology Most heritable phenotypic traits are affected by the concurrent inheritance of alleles at multiple loci. As opposed to simple patterns of single gene inheritance, polygenic traits result in complex patterns of segregation, which are frequently assessed through quantitative trait loci (QTL) mapping methods. While QTL detection is a commonplace task in modern biology, cloning specific genetic variants underlying QTL is not the task remains a challenge often associated with laborious and time -consuming fine mapp ing Recently, one novel method has been proposed to speed the identification of candidate genes and polymorphisms from QTL. Genetical genomics relies on allying traditional quantiative trait analysis to higher level genomic data, including whole genome expression or proteomic data. By considering these genomic phenotypes together with morphological or molecular phenotype s of interest, the role of genetics in these traits' regulation can be compared. QTL coordinately affecting phenotype s and transcript or protein abundance indicate a putative causal relationship between the phenotype and gene(s) represented by those transcripts or proteins, which can be functionally tested Genetical genomics has recently demonstrated


13 significant promise to speed QTL clo ning and ha s implicated co -expressed gene networks underlying disease and complex phenotypes in model species Here, I directly implement genetical genomics methods aimed to reduce the challenges associated with moving from QTL to causal polymorphisms in poplar. First, I demonstrate a novel genotyping and genetic mapping approach for outcrossing species, by leveraging RNA based molecular mar kers for highresolution genetic analysis in Populus These results lay a solid foundation for QTL and expression QTL (eQTL) mapping by increasing marker coverage and reducing breadth of QTL confidence intervals Secondly, I apply whole -genome eQTL data to identify co -expressed transcriptional networks in three diverse plant tissues. These results build upon the understanding of transcriptional regulation in all plant species. Finally, I utilize genetical genomics to discover a candidate gene for leaf morph ological variation in Populus a high value phenotype for photosynthetic productivity and a vital trait for evolutionary classification of genus members into evolutionary sections. Taken together, these results paint a favorable picture for additional forw ard -genetic, genome level studies to characterize complex phenotypes in poplar and other forest trees.


14 CHAPTER 1 INTRODUCTION AND LIT ERATURE REVIEW The importance of forest ecosystems in the health and sustainability of life on Earth are irrefutable. Fores t trees cover 30% of earth's terrestrial surface and forest ecosystems are responsible for harboring the majority of terrestrial biodiversity [1]. In addition to essential ecological roles tree species provide substantial socioecono mic benefits in the forms of timber, cellulosic bio energy atmospheric carbon fixation and sequestration, and pulp and paper production [1]. Among the most ecologically and economically important forest tree species in the northern h emisphere are poplars; genus Populus sp. Poplars inhabit nearly 80 million hectares worldwide, including nearly 70 million hectares of indigenous forests in North America and Russia [2]. Various species of genus Populus are managed for wood and biomass production worldwide, while others serve as ecological foundation species of riparian and deciduous forest ecosystems throughout North America and Europe [3]. The recent sequencing and annotation of the Populus trichocarpa genome [4] provides new opportunit ies to understand how genetic diversity and genetic mechanisms contribu te to forested ecosystems by enabling the identification of genes, m arkers, and regulatory networks that affect traits with significant roles in adaptation, ecology, and production. Identifying these linkages between genetics and phenotypic traits facilitates a better understanding of the role of genetics in forest tree an d ecosystem function. Concurrently, this knowledge can direct tree improvement and breeding efforts toward more productive or environmentally fit genotypes with improved yield, wood quality, disease resistance, or environmental adaptation. Finally, and per haps most importantly, knowledge of the role of genetic mechanism s in forest health and productivity can promote improved science -based policies for forest resource management and forest conservation on a global scale.


15 Poplar as a Reference for Tree Biolo gy and Genetics In addition to their role in terrestrial ecosystems and silviculture several attributes of the poplar species have contributed to their emergence as a model genus for tree genetics, genomics, and molecular biology [5 7] First, poplars are among the most rapidly growing hardwood trees when cultivated under i ntensive management, with fieldand greenhouse -grown genotypes frequently achieving as much as three meters of growth per year [8]. Secondly, f lowering and reproduction are also rapid these processes normally occur within 3 6 years under field conditions, but transgenic genotypes overexpressing FT or LFY transgenes will flower in less than one year [9 11] Third, m ost poplar species are routinely propagated from vegetative cuttings, allowing specific genotypes to be maintained indefinitely and studied over long periods of time or in different environments [12] Fourth, s everal poplar genotypes and species are also amenable to stable, agrobacterium -mediated genetic transformation and regeneration [13] a protocol routinely performed in a number of different laboratories [4]. The transformation process facilitates reverse-genetic approaches to analyze and verify function of specific genes or regulatory elements of interest. In addition to favorable gr owth and developmental characteristics the facile interspecific hybridization among species within genus Populus benefits both fundamental scientific discovery and breeding of superior hybrids for deployment in managed forests Populus is composed of six evolutionary sections, which vary extensively in their morphology, adaptation, and natural distribution ( [12] Table 1 1 ). A number of interspecific segregating populations have been produced within and between these sections, most notably between P. trichocarpa and P. deltoides P. n igra and P. deltoides P. tremula and P. tremuloides and P. angustifolia and P. fremontii [12] These interspecific populations present several benefits for genetic analysis. First, alleles for certain genes or gene combinations may commonly be fixed within a single species. In


16 this case, progeny popula tions derived from intraspecific hybridizations will not produce variation for those genes or the phenotypes they influence In contrast, interspecific hybrid progeny produced from species fixed for alternative alleles will vary for phenotypes affected by those genes. Secondly, i n cases where alleles for genes affecting production or adaptation are fixed in a given species, interspecific crosses with an alternative species could be used to introgress new genetic combinations, increasing fitness or yie ld in the species of interest. Finally, interspecific pedigrees are useful to discover naturally occurring, specifically tuned variants for differential gene expression or function. Superior variants are rarely produced through biotechnology, since functional study of genes through genetic engineering is generally only effective to constitutively initiate or repress gene expression. However, gene discovery through naturally occurring variation provides the opportunity to identify alleles that have been subjecte d to the forces of evolution in a genomic context, finely tuning expression or activity of gene product s for specific ecological niches These novel combinations are likely to be superior in function relative to simple constitutive constructs produced thro ugh genetic engineering in a laboratory setting. Poplar's position as a model hardwood species was solidified with the release of the genome sequence of the P. trichocarpa genotype Nisqually 1 in 2006 [4, 6] P. trichocarpa was only the third plant species, and the first woody perennial to be sequenced, following the release of the Arabidopsis thaliana genome in 2000, and the rice genome in 2002 [14 16] The P. trichocarpa genome sequencing project produced an excellent draft sequence that accelerates genetic, molecular, biochemical and evolutionary studies in poplar. Whole -genome Sanger based shotgun sequencing and reassembly resulted in over 470 million base pairs (Mbp; estimated genom e size) of sequence distributed in 22,136 scaffolds with a 7.5 average


17 sequencing depth. Approximately 385Mbp of the total 470Mbp of sequence was anchored to the 19 chromosomes of poplar on the basis of sequence tagged sites converted from previously iden tified simple -sequence repeat (SSR) microsatellite markers. Initial gene identification efforts, based on a compendium of previously generated expressed sequence tags (ESTs) and computational gene modeling algorithms, tallied ~45,600 genes -about two -fol d more than Arabidopsis (~27,000) and comparable to the ~41,000 identified in rice [4]. Also contributing to its place as a model hardwood, a vast collection of gene expression data is available for a number of species in genus Populus Early gene expression analyses were based on transcript counting in EST libraries, or spotted cDNA microarrays produced from EST resources. These analyses generally focused on tissue-specific expression cataloging [17 25] or classifying transcriptional response to specific biological [22, 26 29] abiotic and environmental [30 33] or mechanical stimuli [18, 24, 34] Subsequent studies have incorporated advance s in modern microarray technology using oligonucleotide probes generated by in situ synthesis to obtain refined estimates of gene expression and catalog comparative patterns of gene expression between different organ types to produce a transcriptional "atl as" for P. trichocarpa [35] Since all poplar species are expected to contain genes very similar (or identical) to that of sequenced P. trichocarpa [36] differences in expression caused by regulatory and expression level polymorphisms are proposed to be critical for inter and intraspecific variation within the genus. Identifying these putative interspecific differences in gene expression, through whole -genome expression analysis and quantitative genetics, forms the central focus of the study described herein.


18 Quantitative Genet ics in the Populus System Currently Established Genetic Marker Resources Reliable methods to genotype DNA sequence polymorphisms are essential for quantitative genetic studies, as they contribute to the production of genetic maps -the genomic groundwork upon which trait data is analyzed via forward genetic analysis. In poplar, several types of genetic markers are available and have been utilized to characterize genetic variation. Early population genetic studies in P. fremontii and P. angustifolia utili zed restriction fragment length polymorphism (RFLP) [37] However, RFLPs were largely abandoned after the development of polymerase chain reaction (PCR) -based markers due to the low multiplex ratio of RFLP assays and the high cost and labor required per genotypic data point. Development of PCR -based markers such as random amplified polymorphic DNA (RAPD) and amplified fragment len gth polymorphism (AFLP) facilitated the first high -coverage genomic surveys of polymorphisms in Populus [38 40] Still, dominant inheritance and low transferability among genotypes and species made these markers less suitable for adoption by the scientific community Anchoring these markers to the genome sequence can also be challenging [41] yet sequence tagged markers were shown to be crucial for assembling the P. trichocarpa Nisqually 1 genome sequence [4]. Co dominant markers such as microsatellites (or simple sequence repeats SS Rs) and single nucleotide polymorphism (SNPs) have become more widely used in animal and plant genetics. The main advantages of SSRs are their multi allelic nature and high levels of heterozygosity; SNPs also bear these advantages, in addition to being the most abundant type of DNA sequence varia nt Since probes or primers used to assay SSRs and SNPs are designed specifically for their targeted locus, they are frequently transferable within and (sometimes) across species.


19 The International Populus Genome Co nsortium (IPGC) provides a large repository (4166) of known poplar SSRs with most (84%) being identified from the genome and physical map end sequences of Nisqually 1 [4, 42, 43] The remaining SSRs were g enerated from a series of other studies [44 47] One third (1,395) of the IPGC SSRs have been tested in at least one Populus segregating pedigree, where 553 of them (40%) were successfully mapped. In addition, a set of 92 of the SSR markers were tested for their transferability in 23 species, including five from the related Salix genus. The SSR resource is being extensively utilized for genetic mapping and analysis of QTLs. From a practical standpoint, SSRs have already been applied in breeding programs for fingerprinting and ancestry determination of cultivars [4 8, 49] While SSRs have been implemented frequently in poplar research, the use of SNPs is only in its infancy. The largest SNP discovery for the genus came as a by -product of the Populus trichocarpa genome sequence [4, 42] Since poplars are obligate outcrossing species, the sequencing and assembly of the heterozygous diploid genotype Nisqually 1 identified over one million SNPs and 162,000 single nucleotide insertion/deletion (indel) polymorphisms, with 20% of all polymorphisms falling withing gene coding sequences (Table 1 2) As expected by the action of purifying selection, nucleotide diversity of SNPs was lower (1.4) in genes when compared to the diversity found in the whole genome. Selection is apparently even stronger against indels within gene sequences, where they are 1.8-fold less ab undant than in the remainder of the genome. SNP frequency is variable between scaffolds of Nisqually1, with the highest abundance found on LG_V (average of 1 SNP every 330 bp) and the lowest on LG_XVI (average of 1 SNP every 484 bp). SNP frequency in gene sequences ranged from 591 bp per SNP on LG_XII to 433 bp per SNP on LG_XIII. However, variation in SNP diversity between scaffolds should be analyzed with caution, as differences in sequencing de pth were not


20 considered in this estimate In another study 1,635 and 610 SNPs were identified in P. tremula and P. trichocarpa, respectively, by aligning approximately 70,000 ESTs [50]. Asi de from these two large scale studies, additional poplar SNPs have been discovered and characterized in only a few targeted genes. These genes were sequenced in a sample of genotypes in studies aimed at estimating population genetic parameters, searching f or departures from neutral evolution, or contributing to candidate gene association studies [51 55] The main limitation of SNPs identified in these studies lies in the fact that forest species generally harbor an excess of low frequency alleles [56, 57] As a result, many SNPs identified in a given study may not be sui table as molecular markers in different genetic backgrounds. However, as sequencing technologies advance [58, 59] polymorphisms will be discovered at markedly increased rates and will reveal specific SNPs with higher allele frequencies and utility for marker -based forward genetic studies. While SSRs and SNPs represent the most useful ma rker classes in poplar today, other classes of sequence anchored molecular markers have been described that could be incorporated into the poplar genetic system. Principal among these are single -feature polymorphisms (SFPs) and expression level polymorphis ms (ELPs), also called gene expression markers (GEMs). SFPs were first described by researchers in the yeast and Arabidopsis genetic systems, who hybridized labeled genomic DNA to oligonucleotide microarrays and noted predictable signal variance at certain probes for different genotypes [60, 61] Sequencing revealed that SFPs may correspond to SNPs, small insertion/deletion polymorphisms, or small genomic rearrangem ents. Subsequently, expression studies based on hybridization of RNA and cDNA also demonstrated the ability to yield an abundance of SFP markers, and coordinately identify GEMs based on Mendelian segregation of divergent gene expression values between two parents of a mapping population [62, 63] While these markers appear to be abundant and highly robust in haploid


21 models and inbred plant species [62 64] it remains to be discovered whether these small polymorphisms detected by differential hybridization or expression will segre gate discernibly in a highly heterozygous, obligate outcrossing genetic system such as Populus Linkage/QTL Population Structures and Associated Analysis Methods The previously mentioned intersp ecific populations of Populus have proven to represent a stron g framework for QTL mapping, since variation between individuals of different species is expected to be far greater than variation between individuals of the same species. Thus, individual species, while potentially heterozygous for specific alleles, can be treated as inbred lines relative to one another and traditional inbred line mapping approaches can be adapted for genetic mapping and QTL analysis [12] A number of pedigrees have been established based on this assumption, using two primary and distinct strategies for population structuring and genet ic map production. Early efforts were directed towards development of an inbred F2 type pedigree system, wherein an interspecific cross was produced between two grandparents, one P. trichocarpa and one P. deltoides which were expected to be inbred relative to one another. Two resulting F1 full -sib progeny were selected and mated, producing a segregating progeny population wherein of the resulting progeny were expected to be homozygous for either P. trichocarpa or P. deltoides alleles, with the remaining bearing a heterozygous F1type genotype [65] The resulting population can be analyzed as an inbred F2 in standard QTL detection analysis [66, 67] QTL for a number of traits, including stem growth characteristics, crown architecture, leaf morphology, and disease resistance were subsequently detected on the basis of this procedure [68 72] Later efforts refined these mapping strategies to account for the outbred population structure and improve the mapping methods [73, 74] Yet these initial studies provided proof


22 that genetic mapping and QTL detection were feasible in the highly heterozygous poplar genetic system. More recent efforts have focused on F1 pseudotestcross and F2 pseudobackcross pedigree structures for genetic mapping and QTL detection in outcrossing forest species [73, 74] Pseudotestcross F1 p opulations are dependent on identifying markers where one parent is homozygous for a given allele at a marker, while the other is heterozygous. Such an arrangement of alleles will result in F1 progeny segregation of 1:1, allowing the population to be analyzed as a testcross in mapping and QTL detection software [12] However, the arrangement of alleles is known only a posteriori and can vary between parents from locus to locus, hence the utilization of the term "pseudo" -testcross. This fact dictates that pseudotestcross analysis will result in two single tree genetic maps for the pedigree, both of which are tested for QTL. Alternatively, pseudobackcross F2 populations can be generated, whereby an interspecific F1 hybrid can be backcrossed to a genotype of one of the F1 donor species ( Figure 1 1 ). General ly this is a different genotype than the one utilized to create the F1 (hence the pseudo -backcross), owing to the self -incompatibility barriers common within forest trees. Since the two species are expected to be inbred relative to one another, many ma rkers will again segregate at a ratio of 1:1. Fully informative markers (i.e. those which are biallelic in both the F1 hybrid and the recurrent pseudobackcross parent) can be utilized to create a separate genetic map for the recurrent parent background, pr omoting QTL detection both with respect to the interspecific differences fixed between the two species (based on the hybrid parent map) and intraspecific differences within the recurrent species background (based on the map derived from the recurrent speci es). This strategy has been extremely effective and popular for genetic mapping and QTL detection in forest tree species including Populus where rare alleles are common


23 within species and alleles affecting important morphological differences between speci es are often fixed. Thus, this strategy has contributed to the detection of numerous QTL for plant and organ morphology, biomass, phenology, disease resistance, and even molecular metabolites [75 77] Quantitative Genetic Analyses of Gene Expression: Genetical Genomics Analogous to the process of analyzing phenotypic trait QTL, wherein variation in phenotype among members of a segregating population is statistically associated to prevalence of a DNA marker genotype at a specific locus, variation in the abundance of an mRNA can be associated with marker genotypes through the process of expression QTL (eQTL) ma pping [78] Classical insight into the role of genetic variation in gene expression regulation was made by King and Wilson [79] followed by the assessment of protein abundance by Damerval et al. [80] However, without technology to concurrently analyze the abundance of large numbers of transcripts or proteins, these studies were limited in their scope to specific suites of well classified genes. DNA microarray technology [81, 82] facilitated the analysis of thousands of genes in parallel to reveal that transcript abundance is hi ghly variable between individuals in yeast [62] mice [83] plants [84, 85] and humans [86, 87] Furthermore, these differences behave as traditional quantitative phenotypes, segregating predictably under quantitative genetic models in progeny derived from crosses between divergent parents [88] Global studies of gene expression diversity among segregating populations were first proposed in 2001, yet work toward these ends was clearly underway at this point as the seminal articles describing true genome -wide linkage analysis of expr ession traits appeared in 2002 [62, 89] In agreement with the original propositions of Jansen and Nap in 2001 [78] these studies showed that variation in expression phenotypes is complex, with genes being regulated by combinations of local, cis acting variants and distant, trans acting variants ( Figure 1 2 ). In the


24 majority of genetic systems considered, distant trans acting variation seems to be more common, yet these loci tend to explain smal ler portions of heritable variation than their cis acting counterparts [62, 87, 89 91] Furthermore, often only small portions of a transcripts he ritable variation are captured in an eQTL experiment, presumably due to the large number of small effect loci affecting transcript abundance that even the largest experiments lack the statistical power to resolve into significant eQTLs [88] Most transcripts, in addition to being subjected to multiple points of cis and trans acting regulation, demonstrate nonadditivity and transgressive segregation, and pleiotropic loci that coordinately regulate the abundance of multiple tra nscripts are quite common [91 93] As indicated by the initial proposition of the genetical genomics concept, eQTL data gathered from segregating populations has been exceedingly useful to understand phenotypic variance, both at the level of single gene:phenotype relationships, and more complex biological processes [94] Initial efforts to associate individual, differentially expressed genes with QTL phenotypes were carried out primarily in the murine genetic system. Candidate genes for complex clinical diseases including diabetes [95] asthma [96] and obesity [83] were discovered by assessing the correlations between expression levels and clinical phenotypes for genes with eQTL co -locating to the position of phenotypic QTLs ( Figure 1 3 ). Efforts in Drosophila melanogaster Saccharomyces cerevisiae and Arabidopsis valida ted these initial findings, associating candidate genes with variation in traits of interest [89, 97, 98] Later studies began to address the task of linking known biochemical pathways to phenotypic variation through eQTL data. Since genes encoding proteins involved in a common biochemical pathway are functi onally connected through the metabolite pools of the pathway to which they belong, they are frequently co -regulated at the level of gene expression [99] Therefore, alterations in the level of express ion


25 for a single gene in a pathway can feed back to other members through these coregulatory mechanisms, leading to correlated expression for genes in that pathway ( Figure 1 4 ). This fact was first exploited to demonstrate the role of differential pheromone response pathway regulation with a variation in a G -protein coupled receptor in yeast [98] Subsequently, variation for growth and lignin traits in a Eucalyptus segregating population was shown to be associated with coordinate transcriptional regulation of the lignin biosynthetic pathway enzymes [100] A number of similar findings have been described for several fundamental biochemical and regulatory pathways in Arabidopsis [93, 101] These demonstrated cases of alleledependent differential regulation of genes in known biochemical and signaling pathways has broadened to a global approach to analysis of genome wide eQTL data. Specifically, demonstrated co -expression of genes that are members of conserved pathways suggests that identifying the correlation structure among whole -genome ex pression data may identify regulatory connections between biochemically distinct pathways, as well as implicate novel pathway members and regulators that have not been previously annotated for their role in an established pathway. So -called whole genome c o -expression network analysis has begun to successfully accomplish these goals in a number of organisms [102] Initial efforts to describe larger scale co -expressed networks were fundamentally based upon wel l characterized biological processes, such as the cell cycle [103] or from analyzing the transcriptomic response to directed genetic or environmental perturbations [104] Nonetheless, these initia l experiments clearly demonstrated the regulatory interplay between known pathways with seemingly distinct roles in cellular function. Subsequent efforts sought to overcome the barriers of chemical and directed genetic perturbation by implementing genome -w ide genetic perturbation leveraging segregating


26 populations constructed from divergent progenitors, wherein evolution and natural selection have produced simultaneous genetic variation at multiple loci [105] The aforeme ntioned interconnectedness of the transcriptome, coupled with independent assortment of loci in a QTL population, enables the discovery of relationships between expressed genes without preconditioning biochemical or functional connections [106, 107] These a posteriori generated co -expression networks have proven to be a powerful tool to elucidate signatures and pathways associated with complex phenotypes and disease traits in model and non -model organisms [101, 106, 108113] Post hoc incorporation of a priori knowledge about network ed genes (i.e. transcription factor binding sites, gene annotations and ontology, protein -protein interactions, and/or protein cellular localization) has substantially aided in the interpretation of these co expressed networks, by providing further connectio ns between data points that could not be inferred simply from their co -expression patterns [114, 115] The ability to predict whole -system response to genetic, environmental, or pharmacological perturbation is a key goal of modern biology, and these whole -genome -based network studies have shown initial promise towards realization of this goal. Leaf Morphological Variation in Populus Evolutionary Conservation and Previous Analyses of Leaf Morphology Variation for leaf morphology, specifically leaf blade shape, is substantial among the members of genus Populus Leaf shapes range from cordate to lanceolate and while shape can be somewhat variable within species, leaf morphology and shape are generally conserved across evolutionary sections [116, 117] In fact, leaf morphology has often been utilized to diagnose evolutionary relationships between the species/sections from the fossil record [118] Furthermore, leaf morphological characteristics in juvenile trees have been shown to be indicative of longer term productivity and yield [119, 120] Variation in leaf shape among


27 poplars is also heritable and relatively immune to the effects of environmental deviations [70, 121] Taken together, this evidence make s leaf morphological variation an attractive candidate for quantitative trait analysis, leading to previous efforts to understand the genetic components controlling these traits through QTL -based methods [69, 70, 122124] Leaf shape has been generally measured by blade length and width, as well as their ratio [69, 70] since these measures differ predictably and consistently between a lanceolate leaf shape as in P. trichocarpa or P. angustifolia versus a deltoid or cordate leaf shape common to P. deltoides P. fremontii, or P. nigra (Figure 1 5 ). Several early QTL studies identified loci for these traits among interspecific pedigrees of P. trichocarpa and P. deltoides [70, 124] However, due to changes in linkage group naming convention and the popularity of in-house anonymous DNA markers (AFLP and RFLP), it is difficult to determine whether loci detected in different pedigrees represent conserved genetic architecture or variation due solely to a set o f rare alleles restricted to one parent of the cross. More recent studies of sex-determination [125] and plant growth and biomass accumulation [75, 122, 126, 127] comprising multiple pedigrees constructed from several different interspecific combinat ions suggest that QTL detected in these pedigrees are often associated with evolutionarily conserved differences rather than rare alleles. Nonetheless, limited mapping resolution in QTL populations [128] coupled with the unavailability of genome sequence, has until recently hindered the purs uit of candidate genes for leaf morphological traits in Populus Recently, novel genetic and transcriptomic approaches [123] have shown promise to implicate candidate genes for leaf morphological variation a concept further explored within the work described herein. Candidate Genes and Pathways Discovered in Arabidopsis As a well -established model angiosperm plant Arabidopsis lends several clue s to genes, p athways, and gene networks that, if conserved, may underlie variation in leaf morphology


28 observed within Populus In angiosperms generally and Arabidopsis specifically, i t is thought that growth and division at the leaf margins results in leaf blade initiation and may be the result of the junction formed between adaxial and abaxial cell fates [129] Subsequent blade expansion is a result of cell divisions and expansion in the plate meristem, which gives rise to leaf cell layers and two dimensional (longitudinal and lateral) 'plates' of c ells [130, 131] Abundant variation for both blade length and width (and subsequently, their ratio) indicates that plate meristem processes are specifically regulat ed both in the longitudinal (length) and lateral (width) planes. Previous studies have implicated genes involved in both polar cell division and polar cell expansion that play a role in leaf shape variation. Among the first genes isolated that affect leaf blade expansion, specifically in the longitudinal plane, was ROTUNDIFOLIA3 ( ROT3 ). Aberrant cellular expansion in rot3 -1 the null allele homozygote resulted in an overall stunted plant with leaves stunted for growth specifically in the longitudinal plan e, whereas overexpression of ROT3 triggered hyperexpansion in the longitudinal direction and elongated leaves [132] Subsequent biochemical and bioinformatic evidence indicated that ROT3 is a cytochrome P450type gene that likely functions in the brassinosteriod signaling pathway, possibly in a leaf specific manner. A similar phenotype was obtained for mutation s in another ROTUNDIFOLIA gene, ROT4 [133] Despite the naming convention ROT4 and ROT3 are not functionally or biochemically related. ROT4 is an intronless 53 amino acid peptide that includes a 29 amino acid domain of unknown function. The original mutation was generat ed in an enhancer trap system, where overexpression of the ROT4 gene reduced cell division frequency in the longitudinal direction, decreasing the number of cells deposited in this plane of the leaf [133] At least 22 homologs of ROT4 exist in Arabidopsis and at least two of these show overexpression phenotypes similar to that of ROT4 suggesting similar


29 function during leaf development perhaps mediated by the highly conserved domain of unknown function. In addition to genes that affect leaf longitudinal growth, at least two other genes have been identified that when mutated result in aberrant leaf lateral growth. Analogous to the ROT3 and ROT4 mutations, these genes function to alter either cellular expansion or cellular division in the lateral plane. The ANGUST IFOLIA ( AN ) gene was identified as a mutant with decreased cell expansion in the lateral plane, resulting in a narrowed-leaf phenotype [134] AN is a functional homolog of animal carboxyl terminal binding proteins (CtBPs) or brefeldin -A -ADP ribosylated s ubstrate proteins (BARS), which are indistinguishable in function at the primary amino acid sequence level but function in diverse molecular roles [135] CtBPs generally function as nuclear transcriptional co repressors [136] while BARS function in membrane and vesicle fission dynamics, particularly at the Golgi complex [135] Recent evidence has shown that AN possesses no CtBPs or transcriptional co repressor function, indicating that it likely functions either as a BARS (which is thought to be vertebrate specific) or in some function that has yet to be attribut ed to CtBP/BARS superfamily proteins [137] It is know n, however, that AN affects microtubule dynamics, yet the mechanism of this activity remains unknown [134] In addition to AN which controls cell expansion, the ANGUSTIFOLIA3 (AN3 ) gene also affects leaf shape, with null mutants res ulting in a an-like phenotype [138] The an3 mutant, however, functions to decrease cell proliferation in the leaf width di rection, resulting in elongated, narrow leaves. AN3 encodes a protein with sequence similarity to the human transcriptional co activator and chromatin remodeling complex interactor SYNOVIAL SARCOMA TRANSLOCATION AN3 interacts with GROWTH REGULATING FACTOR 5 (GRF5) in vitro a gene for which altered


30 expression mimics a mild an3 phenotype [138] GRF5 functions in cell cycle cont rol, suggesting that an3 muta nts may be compromised for cell cycle progression at the leaf plate meristem. Project Ob jectives Despite the recent boom in research interest surrounding the Populus genetic system, fueled primarily by the aforementioned release of the full genome sequence of P. trichocarpa little remains known about the role of transcriptional variation in shaping diversity between species of genus Populus Within this project, I begin to address this shortcoming by leveraging whole -genome microarray expression data derived from a pseudobackcross of P. trichocarpa and P. deltoides to characterize novel aspects of transcriptional regulation in the Populus system. In Chapter 2, I des cribe the utilization of microarray data to discover, genotype, and genetically map two novel classes of Populus molecular markers GEMs and SFPs and in conjunction with SSR data, show the unexpected utility of these markers to improve the map-based sho tgun sequence assembly of the P. trichocarpa genome. In Chapter 3, this map data is utilized to characterize the genome -wide genetic diversity for expression traits (eQTL) in three distinct tissues (leaf, root, and xylem) and eQTL results are subsequently employed to construct and annotate a posteriori defined co -expression networks for each of the three tissues. Chapter 4 describes the integration of map and eQTL data from leaves, in a genetical genomics context, to implicate an ADP ribosylation factor GTP ase as a candidate gene for evolutionarily conserved variation in leaf shape between P. trichocarpa and P. deltoides Finally, in Chapter 5, I summarize the findings of this research and suggest avenues for future studies to better understand the role of t ranscriptional diversity in shaping phenotypic variation in Populus


31 Figure 1 1. Overview of the pseudobackcross pedigree construction method for outcrossing forest trees. An interspecific hybrid F1 population is constructed from a controlled cross of t wo elite genotypes. An individual is selected from this population and backcrossed to an individual of the same species, but different genotype, as one of the elite F1 progenitors The majority of the genetic variance segregating in the pedigree can be accou nted for due to differences between P. trichocarpa and P. deltoides


32 Figure 1 2 Hypothetical examples of cis and trans acting eQTL regulation. A) An eQTL for gene A is detected in cis -, since the position of the significance of the eQTL peak is located within the marker interval (M4 -M5) known to physically contain gene A on the basis of a physical map or genome sequence assembly. Cis acting eQTL may be caused by polymorphisms in a re gulatory region for the gene in question, a mutation that leads to missense -mediated transcript decay, or a trans acting regulatory factor tightly linked to the gene in question. B) An eQTL for gene B is detected in trans since the significance for the eQTL peak is located within a marker interval (M4 -M5) different from the interval known to contain gene B (M1 -M2). Trans acting eQTL may be located on the same linkage group as the gene (depicted above) or on a different linkage group (not depicted), an d may be the result of a sequence or regulatory polymorphism affecting a regulatory factor governing the expression of the gene in question


33 Figure 1 3. A hypothetical example of candidate gene identification by the genetical genomics method. A phenotypic trait of interest is measured and mapped as a quantitative trait using standard QTL mapping approaches. A significant QTL for this trait is identified near marker s "M4" and "M5" (red LOD curve). Whole transcriptome microarray analysis identifies a cis a cting eQTL for gene "A" (blue LOD curve) underlying the trait QTL. Statistical analysis indicates that the expression of gene "A" and the phenotype are negatively correlated at a statistically relevant threshold. The coincidence of QTL peaks, the physical position of the gene, and the correlation between trait and transcript abundance provide initial circumstantial evidence that differential transcriptional regulation of gene "A" may play a role in the phenotype in question


34 Figure 1 4. A hypothetical ex ample of a mechanism of network construction and phenotypic association from genetical genomics eQTL data in a segregating population. eQTL detected for three genes ("green", "blue", and "red") share similar eQTL LOD curve profiles, suggesting their underl ying regulation is shared. An arbitrary phenotypic trait LOD curve (black) also has a similar profile. Statistical correlations between the trait and the genes are highly significant, suggesting that these genes may act in a pathway that influences varian c e in the phenotype in question


35 Figure 1 5. Leaf morphologies assoicated with the four most populated evolutionary sections of genus Populus Narrow elliptical or lanceolate leaf shapes are observed among members of section Tacamahaca, whereas broader ovate, deltoid, or cordate shapes are observed among members of sections Aigeiros, Leucoides, and Leuce. Section Abaso is not diagrammed.


36 Table 1 1. Ra nge, leaf morphology, and species count associated with sections of genus Populus Sectiona Geographic ( c limatologic) range Leaf morphology Number of s pecies a Aigeiros N. America, Asia, W. Europe ( t emperate) Cordate/Deltoid 3 Leuce Circumpolar ( s ubarctic and c ool t emperate) Oval/Deltoid 10 Leucoides N. America, E. Asia ( w arm t emperate) Ovate 3 Tacamahaca N. America, Asia ( cool t emperate) Elliptical/Lanceolate 9 Turanga S.W. Asia, E. Africa (subtropical to t ropical) Lanceolate/Cordate 3 Abaso Mexico (tropical, subtropical, a rid) Deltoid/Lanceolate 1 a.Sections and species as recognized by Eckenwalder in [8].


37 Table 1 2. Fre quency of polymorphism across chromosomes in the genome of Populus trichocarpa, clone Nisqually1. Genome Genes a Scaffolds Number of polymorphisms Diversity b (bp / polymorphism) Number of polymorphisms Diversity b (bp / polymorphism) Indel SNPs Indel SNPs Indel SNPs Indel SNPs LG_I 13,344 85,410 2,410.01 376.53 2,132 16,772 4,409.58 560.53 LG_II 11,811 70,007 1,984.01 334.73 1,759 13,769 4,151.55 530.36 LG_III 7,767 46,293 2,246.28 376.88 1,340 9,721 3,990.46 550.07 LG_IV 5,798 39,659 2,600.94 380.25 953 7,180 4,244.67 563.39 LG_V 8,116 51,571 2,099.02 330.33 1,060 8,383 4,501.44 569.19 LG_VI 7,623 46,963 2,316.31 375.98 1,217 9,816 4,676.68 579.82 LG_VII 5,175 32,691 2,300.35 364.15 764 6,730 4,788.73 543.62 LG_VIII 6,940 42,758 2,223.37 360.87 1,184 10,362 4,678.92 534.63 LG_IX 5,589 33,874 2,220.73 366.41 1,005 7,980 4,321.07 544.19 LG_X 7,773 47,329 2,471.87 405.96 1,375 11,533 4,908.94 585.26 LG_XI 4,792 32,942 2,748.75 399.85 854 6,502 4,217.58 553.95 LG_XII 4,875 31,368 2,671.80 415.23 766 5,963 4,599.58 590.86 LG_XIII 4,979 32,977 2,308.85 348.60 921 8,179 3,844.72 432.94 LG_XIV 5,574 34,953 2,454.74 391.46 914 7,437 4,559.25 560.33 LG_XV 4,059 25,327 2,510.66 402.37 653 5,418 4,808.18 579.50 LG_XVI 4,458 26,454 2,875.34 484.55 842 6,465 4,219.58 549.56 LG_XVII 2,238 16,291 2,432.56 334.18 344 2,808 3,855.58 472.34 LG_XVIII 4,568 29,995 2,722.86 414.67 752 5,976 4,638.20 583.66 LG_XIX 3,644 24,744 2,806.48 413.30 636 5,832 4,378.72 477.51 Unmapped scaffolds 42,767 327,755 3,343.42 436.27 6,951 64,742 4,653.55 499.63 Total or Average 161,890 1,079,361 2,487.42 385.63 26,422 221,568 4,422.35 543.07 a. Genes are composed of 45,555 currently annotated gene models plus 10,288 additional less supported models. The whole gene sequence was considered, including its non translated porti ons (introns, 3UTR and 5UTR). b. Sequence gaps of known size (symbolized in the genome sequence with Ns) were excluded from the calculation of genomic diversity.


38 CHAPTER 2 A MICROARRAY -BASED GENOTYPING AND GENETIC MAPPING APPROACH FOR HIGHLY HETEROZYGOUS OUTCROSSING SPECIES ENABLES LOCALIZATION OF A LARGE FRACTION OF THE UNASSEMBLED Populus trichocarpa GENOME SEQUENCE This Chapter has been published in The Plant Journa l 2009; 58(6):10541067 (PMID: 19220791) Introduction Microarrays revolutionized the study of gene expression and have recently been applied for high -throughput genotyping of sequence and expressionlevel polymorphisms. Single feature polymorphisms (SFPs) detected by differential hybridization of genomic DNA to whole -genome microarrays were reported initially in yeast [61, 62] Arabidopsis [60, 64] and later in rice [139] Subsequent reports showed that h ybridization of RNA could also identify SFP in haploid yeast [140] and several inbred plants [91, 141145] while co ncurrently generating estimates of gene expression from segregants [91, 140] Utilizing RNA to characterize SFPs also creates the opportunity to identify gene expression markers (GEMs) genes that are differentially expressed between parents of mapping populations and show Mendelian segregation of expression values within progeny [145] Generating genotypic and gene expression data in a common assay establishes a framework for powerful forwar d genetic approaches, including genetical genomics studies [78] However, while microarray -based mapping has been successfully applied to haploid or homozygote lines, the approach has yet to be demonstrated in outcrossing plant species with high genetic diversity, where up to four alleles can segregate for each locus in a full -sib pedigree. RNA -based SFP genotyping requires robust separation of microarray signal variance associated with differential hybridization kinetics between alleles from variance du e to differences in mRNA abundance [140] Previous studies in species with limited genetic diversity have relied on short ( -mer) oligonucleotide probes to detect genetic variants because a unique single nucleotide polymorphism (SNP) can result in different ial hybridization and detection of


39 SFP [146] Short oligonucleotide based microarrays typically utilize multiple probes per gene (a probeset) to estimate gene expression. Thus, SFP -containing probes can be detected by comparing individual probe signal to the signal m easured across the probeset. Probes for which the signal deviates significantly from the probeset mean in a subset of the segregating population suggest the presence of a segregating SFP, while the remainder of the probeset provides an estimate of gene exp ression [140, 143, 145] However, in outcrossing species with extensive genetic diversity, abundant SN P variation and heterozygosity can result in significant bias for estimates of gene expression, since SFP may be present within many probes in a given probeset [146] Such biases render platforms that utilize short probes less reliable for concurrent analysis of gene expression and genetic polymorphisms in these experimental settings. Utilizing long oligonucleotide probes may improve estimates of gene expression in these cases. However, approaches to select optimal long -oligonucleotide probes for gene expression analy sis in highly diverse species or across multiple related species and their hybrids are lacking. Similarly, the capacity of longer probes to detect a useful quantity of segregating polymorphisms for genetic mapping has yet to be demonstrated. Our objective in this study was to develop an approach to select optimal long oligonucleotide probes for gene expression analysis and microarray -based genotyping in a highly heterozygous population. We utilized an interspecific pseudo -backcross of P. trichocarpa X P. d eltoides and a long oligonucleotide (>50-mers) microarray platform to develop a two -step method to discover candidate SFP in parent lines, then genotype sequence and expression -based polymorphic features in the progeny. We show that genotypic data generat ed by this method can contribute to the development of an accurate highdensity, gene based genetic map. Additionally, the value of these markers is demonstrated by the positioning of almost half of the previously


40 unassembled whole -genome shotgun (WGS) sequence scaffolds within the complex and highly heterozygous genome of P. trichocarpa. The results we describe provide a flavor for both the challenges and opportunities presented when undertaking a microarray -based genetic mapping study in a gen etically div erse plant species. Therefore, we believe the techniques we present can provide a strong framework for future microarray based genotyping in crops, forest tree species, and other complex plant genomes. Similarly, our approach for optimal probe selection for gene expression analysis within or between highly diverse species may prove useful for other agricultural and forest tree species with similar levels of genetic diversity. Material s and M ethods Plant Growth Conditions and RNA I solation A pseudo-backcross population (Family 52 124) derived from the cross of a female P. trichocarpa P. deltoides hybrid (genotype 52 225) and a male P. deltoides (genotype D124) was obtained from the Department of Forestry at the University of Minnesota Duluth as hardwood cuttings. After rooting, bud break and shoot elongation, fresh softwood terminal cuttings were harvested and placed in rooting media pellets ( Jiffy Forestry Products, Lorain, OH) for two weeks. Rooted cuttings were planted in two gallon pots and grown for six weeks in ebb and -flow benches in a greenhouse under long day conditions (16h light:8hr dark) and standard nutrient regime (Hocking s Modif ied Complete Fertilizer [147] ) supplemented with 25mM nitrogen (NH4NO3). Plants were distributed in the greenhouse in a partially balanced incomplete block design with three biological replications per genotype. At harvest, the main plant organs (stems, roots, leaves and sylleptic branches) were collected separately. Stems were further dissected into secondary xylem tissue and phloem/bark/immature xylem. Leaf, secondary xylem, and root tissue from two biological repl icates of each genotype were selected for gene expression analysis. All tissue was flash -frozen in liquid nitrogen immediately after harvest and stored at -


41 80 Celsius (C) prior to lyophilization and subsequent RNA isolation [148] RNA samples were treated with RQ1 DNase (Promega USA, Madison, WI), purified in RNeasy Plant Mini Kit columns (Qiagen USA, Valencia, CA), and integrity eva luated on 1% w/v agarose gels. Microsatellite (SSR) G enotyping and F ramework M ap C onstruction Parent trees and 418 progeny of Family 52124 were genotyped for 167 framework SSR loci ( [46, 47, 149] ). DNA was isolated from leaf samples using the Qiagen DNeasy Mini Kit (Qiagen USA) following the manufacturers protocol. PCR reagents and concentrations were as described [149] except that SSR loci were amplified from 7.5ng genomic DNA and amplified fragments were labeled by incorporation of fluorescein 12-dUTP (Roche Diagnostics, Germany) included at 8M. Amplification conditions were 94C initial denaturing for 5min; 9 cycles of touchdown: denaturing at 94C for 15sec, annealing 59C 50C one cycle each in one degree increments for 15sec, extension at 72C for 30sec; followed by 21 cycles of denaturing at 94C for 15sec, annealing at 50C for 15sec, and extension at 72C for 30sec; with a final extension at 72C for 3min. Fragments were detected as described [149] except in an Applied Biosystems (ABI; Foster City, CA) Prism 3730xl DNA analyzer. Alleles were identified and genotyped in GeneMapper 4.0 (ABI) and/or GeneMarker 1.5 (SoftGenetics LLC, State College, PA). Single -tree framework maps were constructed in MapMaker v.3.0 [150] as described [73, 151] and were anchored to the P. trichocarpa genome assembly v.1.1 through a BLASTN analysis [152] of PCR primer sequences for each marker. Proper placement of markers was confirmed by comparison of sequence predicted and experimentally determined P. trichocarpa SSR amplicon lengths. SSR used to confirm map position sequence scaffolds were ident ified using MsatFinder v.2.0 based on scaffold sequences from v.1.1 of the P. trichocarpa genome sequence. Primers were designed within the MsatFinder interface (Table 2 1 ) and SSR loci were amplified from 96


42 Family 52 124 progeny as described above. Thirt een of the sixteen loci segregated highly heterozygous alleles between the P. trichocarpa and P. deltoides backgrounds and were genotyped using agarose gel electrophoresis (1% w/v). The remaining three loci were scored using polyacrylamide gel electrophore sis as described [153] Microarray Analysis of Parental Genotypes RNA extracted from root, leaf, and secondary xylem of the parent s of Family 52124 was converted to double -stranded (ds ) c DNA (SuperScript Double Strand cDNA Synthesis Kit, Invitrogen USA, Carlsbad, CA) with oligo -dT primer (Promega USA) according to the manufacturers protocol, except that synthesis of first and second strands were extended to 16h. Resultant ds -cDNA was labe led using cy3 tagged random 9mers and Klenow fragment for 2h at 37C, denatured at 95C for 5min and hybridized to custom in situ synthesized oligonucleotide microarrays (produced by NimbleGen, Madison, WI) at 42C overnight (16-20h). Microarray probe desi gn A total of 55 793 gene models derived from the annotation of the P. trichocarpa genome sequence were represented in the microarray used in the analysis of the two parents of Family 52124. 60mer oligonucleotide probes were designed based on NimbleGen s tandard procedures that optimize uniqueness of the targeted genomic region and GC content, while minimizing self complementarities and homopolymer runs. The highest ranking 6 7 probes (probeset) were selected to represent each gene model, with optimal prob e spacing leading to uniformly distributed, non -overlapping coverage. Twenty negative control probes utilized in previous studies [4] were also included to serve in background quantification. Statistical analyses Raw signal data from all hybridizations was background subtracted, log2 transformed and quantile normalized [154] Normalized signal detected for each probe was ce ntered to zero and


43 analyzed in a gene -by -gene mixed ANOVA model (Equation 2 1) in SAS 9.1 (SAS Institute, Cary, NC) with genotype i (1 degree of freedom [df]), tissue j (2 df), tissue i by genotype j interaction (2 df), probe k (5 6 df) and genotype i by probe k interaction (5 6 df) as fixed effects. yijkli jk ij ik + eijkl (2 1) F tests for all fixed effects, as well as least -square mean estimates were obtained and correction for multiple tests was made u sing a modified false discovery rate (FDR) threshold (FDR < .025, Table 2 2 [155] ). Normalized log2 transformed signal values from m icroarrays derived from differentiating xylem tissue samples were analyzed separately using a similar model that excluded tissue effects. Pairwise t tests were implemented to contrast least-square means estimates of the interaction detected between the two parents for each probe in a probeset. Resulting p -values were corrected for multiple testing as above (FDR<.1). Microarray Analysis of Family 52 -124 Based on the probes selected from the parent tree data a modified microarray was designed for analysis of the progeny of Family 52 124. The modified microarray was comprised of 67, 897 probes including the pre -selected 55 793 gene expression probes and 12, 084 SFP genotyping probes, plus 20 controls [4]. Microarrays were synthesized in NimbleGens four -plex platform and utilized for analysis in the progeny. RNA isolated from one biological replicate of secondary xylem in 154 progeny genotypes was converted to ds -cDNA, labeled, and hybridized as described above. All 67 877 experimental probes were evaluated for Mendelian segregation in the progeny based on k -means clustering procedures modified from those described previously [143] Briefly, quantile normalized, log2transformed signal values detected for each probe in the progeny of Family 52 124 were separated in to two clusters using Proc Fastclus in SAS 9.1. Cluster membership was tested for the expected 1:1 segregation using a 2-test. Probes for which cluster


44 fre quencies deviated significantly (2 df=1> 3.84, p < 0.05) from the expected segregation were discar ded. Subsequently, the probability that an individual assigned to one cluster is not a member of the other cluster was evaluated by calculating the p value ( pi) associated with the modified normal deviate (Equation 2 2) : zi = | (xi mj)/ sj| (2 2) Where xi is the signal at a given probe for an individual assigned to cluster i, and mj and sj are the mean and standard deviation of signal at that probe for all individuals assigned to cluster j [143] We used zi > 1.96 ( pi < 0.05) as evidence that the two allelic classes were clearly distinguishable, and scored individuals below this threshold as missing data. Probes resulting in >10% missing data (n Grouping, O rdering, and M apping of SSRs, GEMs, and SFPs to Linkage G roups Selected GEM and SFP markers in conjunction with SSR markers utilized for the framework mapping, were grouped and ordered using MadMapper V248 linkage mapping software essentially as described [145] However, because MadMapper scripts were developed for marker grouping and orderin g in advanced generation Arabidopsis recombinant inbred lines, estimates of pairwise recombination frequency provided differ from those experimentally observed in a first generation backcross pedigree structure [156] In addition, only microarraybased markers grouping together with at least one SSR from the established framework map were included, as p robes not linked to the framework are likely to have an excess of genotyping error. Markers were re -genotyped after localization of recombination breakpoints using a structural change analysis method with the Strucchange statistical module in R [157] in a


45 strategy initially described by Singer et al [64] Structural change analysis detects large pattern shifts in a dataset based on a Bayesian Informati on Criterion (BIC) statistical threshold, and can be used to detect overall change between phases of alleles that are characteristic of recombination breakpoints. To contribute to the Strucchange analysis of breakpoint positioning, the p -value ( ps) assoc iated with the standard normal distribution (Equation 2 3) for the cluster of assignment was determined: zs= | (xi mi) / si| (2 3) The p -values for each distribution (zi and zs, Equations 2 2 & 2 3 ) were compared by the calculating the ratio R (Equation 2 4) which has a range from zero to one, analogous to the procedure described previously [64] : R = pi /( pi + ps) (2 4 ) If the alleles are highly distinct (i.e., clearly form separate distributions), individuals from the population return values of R very close to zero or one, depending on their allele. However, markers with little allelic distinction accumulate individuals at intermediate levels of R Utilizing a continuously distributed allele score such as R also provides a direct assessment of confidence associated with an assigned genotype on an individual -by -individual basis and thereby contributes to more concretely defined breakpoints in the Strucchange analysis. To verify proper placement of recombination breakpoints, agreement between Strucchange genotypic results and raw SSR genotypes were inspected. Additional breakpoints supported by the Strucchange minimum BIC statistic, but not present in the SSR data, were accepted if they included at least 3 microarray -based markers. Subsequently, genetic distances for the corrected genotypes were estimated with MapMaker v.3.0 [150]


46 Sequence -Level C haracterization of SFP A lleles A subset of mapped SFPs was arbitrarily selected for sequencelevel characterization in each parent of Family 52 124. PCR primers were designed from genome sequence surrounding five mapped SFP ( Table 2 3 ). Alleles were amplified from each parent tree using 50ng xylem ds -cDNA, 200M dNTPs, 2L 10X Advantage 2 PCR buffer and 0.4L Advantage 2 polymerase mix (Clontech Laboratories, Inc., Mountain View, CA) in a total volume of 20L. PCR was conducted in a two -step procedure with identical amplification conditi ons for each step: 95C initial denaturing for 5min, 30 cycles of 95C denaturing for 30sec, 58.5C annealing for 30sec, 72C extension for 1min 45sec, with a final extension of 72C for 7min. Secondary PCR was conducted using identical reagent concentrati ons, except that a 1:25 dilution of the primary PCR was substituted as template. Amplicons from the secondary reaction were gel purified in 1% w/v agarose and cloned into pGEM T (Promega USA) following the manufacturers protocol. Eight to ten independent clones per construct were isolated using the QIAprep miniprep kit (Qiagen USA) and sequenced bidirectionally from the SP6 and T7 promoters in an ABI Prism 3730xl. Resulting sequences were aligned and analyzed in Sequencher v.4.6 (Gene Codes Corporation, An n Arbor, MI) and ClustalW v.2.0 [158] Results SSR Framework Map of Genotype 52-225 We constructed a single -tree framework microsatellite map ( Figure 2 1 ) for the maternal P. trichocarpa X P. deltoides hybrid parent (genotype 52 225) of Family 52 -124 based on 167 SSR markers using a pseudotestcross strategy [73, 151] The framework map represented the 19 consensus linkage groups (LG) of poplar [159] though an unresolved gap remained in LG_VI due to a lack of informative markers in this region. Markers shared with the genetic map of genotype 52 225 produced for a different population (Family 13, for which the genotype also


47 serves a s the maternal parent; [40] ) were largely collinear. Framework SSR loci represented a subset of the sequence tagged sites used to assembl e the P. trichocarpa WGS contigs and scaffolds into chromosome -scale linkage groups [4]. Based on that information, we anchored and oriented the framework map relative to the genome assembly. The framework map spanned 2970cM with average marker intervals comprising 17.8cM and served as the basis for subsequent grouping of SFP and GEM markers into linkage groups. Identification of Probes for Genotyping Family 52-124 A microarray analysis was initially conducted in each parent line to (1) identify candidate SFP probes segregating in the pedigree and (2) identify a single optimal probe for gene expression analysis in the progeny. To develop a microarray platform that could be used for concurrent genotyping and transcript profiling the progeny of Family 52 124, we began by testing 6 7 probes per gene in the two parents. The custom platform was comprised of 384, 287 60mer features re presenting 55, 793 annotated gene models (probesets) from the sequenced genome of P. trichocarpa. This gene set included 45, 555 predicted gene models reported previously plus 10 238 ESTs and less supported gene models with transcriptional evidence [4]. For the probe selection study, RNA extracted from root, leaf, and secondary xylem of each parent of Family 52 124 was converted to double -stranded (ds ) cDNA labeled, and hybridized to the microarrays. After normalization, the data was assessed in an analysis of variance (ANOVA) with genotype, tissue -bygenotype interaction, probe and genotype -by-probe interactio n effects. Genotype effect accounts for overall differences in signal in a probeset between the two parental genotypes, and primarily reflects a difference in gene expression level between them ( Figure 2 2 C, D ). The tissue effect accounts for differences in expression detected by a probeset between different tissues, regardless of the genotype being profiled. Probe effect detects specific properties of probe that distinguish it from others in a probeset, independent of


48 parent genotype ( Figure 2 2 A, D ). Fi nally, the genotype -byprobe interaction accounts for specific properties of a probe that distinguish it from the rest of the probeset, dependent on the genotype being analyzed. Dependence on genotype suggests that these probes contain SFP between the pare ntal genotypes that may segregate in the progeny ( Figure 2 -2 B, D ). To identify candidate probes for SFP genotyping, two separate analyses were conducted. In the first, a t test was used to contrast least -square mean estimates of the interaction between th e two parental genotypes at each probe, across all tissues. A probe within a probeset may be biased towards one or the other parent due to differential hybridization (i.e. an SFP) and therefore is a candidate to be tested for segregation in the progeny. Fu rthermore, only probes for which the difference in least -square means between the parental lines exceeded an arbitrary four fold threshold were selected. We identified 2875 probes meeting these criteria (FDR < .1; p < .0085). When more than one probe from a probeset was identified we selected the most significantly interacting probe. In total, candidate SFP probes were identified for 912 genes. Among these, 770 exhibited hybridization bias favoring the 52 225 hybrid parent, while 142 demonstrated stronge r hybridization in the D124 P. deltoides parent. These results are expected because microarray probes were designed based on the genome sequence of P. trichocarpa [4], one of the species contributing to the hybrid parent. Therefore we hypothesized that the majority of candidate SFP may be explained by species -level polymorphism between P. trichocarpa and P. deltoides alleles. Based on this hypothesis and the interspecific pseudo -backcross pedigree structure, comprised of one P. trichocarpa and three P. deltoides alleles, we expected that most SFP and GEM alleles showing simple Mendelian inheritance should segregate at a ratio of 1:1. To identify additional candidate SFP probes for genotyping and mapping in the progeny, we reanalyzed the parental expression data derived from secondary xylem in a separate ANOVA.


49 Similar to the previous analysis, we contrasted each parent s interaction with individual probe s within a probeset, and selected those that were significant (FDR <.1, p <.0051) with at least a three -fold difference in least -square means estimates. The separate analysis focusing on xylem tissue was conc eived after previous work showed this tissue to be among the most transcriptionally diverse in Populus [4]. From this dataset, we initially identified 13, 191 additional candidate SFP probes, including 8986 with hybridization bias favoring the hybrid parent and 4205 showing hybridization bias favoring the P. deltoides parent. By again selecting only the most si gnificantly interacting probe in each probeset, we identified an additional 11, 172 genes harboring candidate SFP. In total, our two analyses identified single, specific probes from 12, 084 genes containing candidate SFP, which were subsequently carried forw ard for analysis of the progeny. Identification of Probes for Transcript Profiling Family 52 -124 A second objective of the microarray analysis of parental genotypes was to identify a single, optimal probe for expression analysis of the 55, 793 gene models in the 52124 progeny. To identify probes unbiased for gene expression analysis in both parental species backgrounds, we assumed that the probeset mean best represents the true expression value in each parent. Therefore, c ontrary to the previous analysis, the goal was to select the probe that performs most consistently within the probeset in both parents To select the optimal probe for gene expression analysis, an iterative selection process was implemented. (1) For each gene, probes were ranked based on the deviation of the least -square mean estimate of each probe effect, relative to the probeset mean. Pr obes that do not deviate significantly from the probeset mean suggest that inherent properties of the probe do not contribute bias to the signal detected at that probe. (2) Next, the highest ranking probe was analyzed for its sequence alignment uniqueness scores assigned during probe design. Only


50 probes with no more than one unique match to the Populus genome sequence were further considered. (3) Finally, probes were evaluated for significant (FDR < .1) genotype by-probe interaction. In cases where the prob e was not unique or showed a significant genotype -by-probe interaction, the next highest ranked probe was evaluated (i.e., next step of the iteration). After seven iterative rounds of selection, all probes had been considered by these criteria, and probes to measure gene expression were selected for 46, 001 genes. The selection for the remaining 9792 genes was based on a rank variable provided by NimbleGen. The rank variable concurrently accounts for probe chemical properties and probe uniqueness characteri stics. The highest ranked probe for each gene exhibiting non-significant probe effect and genotype -by -probe interaction effect was selected. For 149 genes, all probes in the probeset exhibited a significant probe effect or genotype -by-probe interaction. Si ngle probes were chosen for these genes solely on the basis of the NimbleGen rank variable. Genotyping SFP and GEM Probes in the P rogeny of Family 52-124 To evaluate the candidate SFP probes identified in the parent genotypes, we assayed RNA abundance in xylem tissue from 154 progeny of Family 52124. A modified microarray was designed, comprising the single selected expression probe per gene for each of 55, 793 gene models and the 12, 084 candidate SFP probes. Loci were genotyped using a k -means clustering a llele -calling procedure ( Materials and Methods ) Normalized data for each of the 67, 877 experimental probes was grouped into two separate clusters, and frequency of cluster membership was tested for 1:1 segregation ( 2 df=1< 3.84, p > .05). A total of 12, 680 features followed the expected Mendelian segregation pattern, including 9782 probes selected for gene expression analysis (17.5%), and 2898 of the candidate SFP probes (24%). Gene expression


51 probes that segregate in the mapping population may be utilized as gene expression markers (GEMs) and therefore were considered in further analyses. Next, signal separation between allelic classes was evaluated through a modified normal deviate and probes resulting in >10% ambiguous allele assignments were removed. R eliable genotypes in >90% of the progeny were obtained for 1733 probes, including 1014 GEMs and 719 SFPs (1.8% and 6% of total, respectively). The 1733 segregating features correspond to 1610 independent gene models segregating probes corresponding to bo th GEM and SFP were identified for 123 gene models. Genetic M apping of G enotype 52-225 The 1733 candidate SFP and GEM probes were utilized to generate a genetic map of genotype 52 225. Marker grouping, ordering and mapping was carried out as described previously [145] with slight modifications ( Materials and Methods). To corr ect for genotypic errors and ambiguities in the resulting linkage groups markers were re -genotyped after localization of recombination breakpoints using a structural change analysis (Materials and Methods ). In addition to the 167 framework SSRs, we unambiguously placed 324 SFP and 117 GEM loci in the map of 52225 ( Table 2 4, Figure 2 3 ). For most linkage groups, and the genome as a whole, average marker intervals were <5cM. The total genome length was 2798.5cM, in good agreement with recently publ ished genetic maps for interspecific crosses of Populus [40] Overall rate of marker placement error was low: for genes known to be physically located on specific chromosomes in the P. trichocarpa WGS sequence assembly, ten were not placed in their predicted linkage group an error r ate of 3.52% (10/284). Of the misplaced markers, seven corresponded to SFPs and three to GEMs These ten markers were subsequently excluded from the map.


52 Physical O rientation of the 2 -225 G enetic Map We oriented and aligned the 52 225 genetic map to t he chromosome level WGS assembly of P. trichocarpa Nisqually 1 based on physical positions of genes interrogated by SFP and GEM probes [4] and our previously anchored SSR loci. In large measure, the predicted genetic orientation and physical orientation were collinear several small inversions were detected that could be the result of error in map ordering or represent true differences in gene order between different P. trichocarpa clones or between P. trichocarpa and P. deltoides (data not shown) Slight variations in map order between Nisqually 1 and 52225 have been reported elsewhere [125] On average, predicted physical intervals between ordered markers contain 8 4.4 genes; however, the range is wide from one to 624 genes. Average physical distance spanned by marker intervals is 725 thousand base pairs ( kb ) and ranges from 146 base pairs ( bp ) to 5.31 million bp (Mbp) Genetic M apping of the U nassembled Populus G eno me Approximately 7700 sequence scaffolds from the WGS assembly are not assigned to specific linkage groups in v.1.1 of the P. trichocarpa genome sequence. These scaffolds vary in size from <100bp to >3.5Mbp (mean 16.8kb ) and represent 75Mbp of unplaced sequence [4]. Much of this sequence was postulated to be heterochromatic or derived from substantially divergent haplotypes in the sequenced clone [4, 42] O ur micro array based mapping results provided an unprecedented opportunity to anchor a large amount of this unplaced sequence to potential genomic l ocations in P. trichocarpa based on the genes physically localized within these sequence scaffolds. Among our 1733 candidate GEM and SFP markers, 783 were contained in genes residing in 492 sequence scaffolds. We successfully mapped 167 of these 783 loci, thereby placing 116 sequence scaffolds to unique genetic locations in linkage groups (Table 2 4 ). Five remaining scaffolds demonstrated linkage to other markers in the map, but could not


53 be unambiguously placed within a single linkage group (data not shown). This error rate associated with scaffold mapping (4.13%; 5/121) is congruent with the mapping error rate observed for markers with known position in the linkage group WGS assembly (see above). The 116 sequence scaffolds placed to the genetic map corresp ond to 35.7Mbp of WGS sequence assembly, or nearly 50% of the unlinked sequence ( Table 2 5 ). Among these mapped scaffolds, 34 (representing 23.3Mbp) could be linked by two or more markers, providing an orientation of sequence strands comprising the scaffol ds. Verification of M ap Position f or U nassembled S equence S caffolds To confirm that our assembly of genomic scaffolds using SFP and GEM markers was reliable, we verified the position of a subset of mapped scaffolds using SSRs. From the P. trichocarpa v.1.1 sequence scaffolds [4], we identified SSR loci within ten distinct scaffolds mapped using GEM and SFP mar kers and designed PCR primer s in their flanking sequences. Nine of the ten SSR segregated discernably, and a fter amplification and genotyping, we mapped these loci on the basis of only the original framework SSR map, to eliminate any bias that could be int roduced due to genotyping error in linkage group anchored SFP and GEM alleles. For eight of the nine scaffolds, we successfully verified the putative map location of the scaffold sequence with respect to the framework SSRs ( Table 2 6, Figure 2 1 ). Relative genetic distance between scaffold anchored markers in both the SFP/GEM -based map and SSR framework map were also in agreement ( Table 2 6 ). The lone scaffold (scaffold_121) for which we could not verify map position using this technique was placed on the basis of a single GEM to LG_XVIII, whereas data from two SSR consistently positioned it within LG_VI. Therefore, we speculated that this result was attributable to strong trans acting regulation from LG_XVIII acting on the gene characterized as a GEM. Since GEM may be the result of either cis or trans acting variation, we were interested


54 to determine if scaffolds mapped based on single GEM loci were less reliable for their genetic positioning. We considered the remaining 13 scaffolds that were placed on the basis of single GEMs in our map, and identified informative SSR in six of these 13 scaffolds. Using the framework SSR map, we successfully verified the predicted genetic placement for five of these six GEM anchored scaffolds ( Figure 1, Table 2 7 ). The sin gle unverified scaffold (Scaffold_250) localized to the same linkage group, but a different SSR bin, than predicted by the GEM locus (Table 2 7 ). Characterization of S equence -Level A llelic V ariation R epresented B y M apped SFPs SFP detected by short ( -mer) oligonucleotide probes often correspond to one or few SNPs or small indels [143, 146, 160] However, the implication of sequence mismatches on signal detected from long oligonucleotide probes has only recen tly been described [161] Thus, we characterized the allelic variations present in a sample of mapped SFP probes from the microarray platform. Using ds cDNA produced from xylem for each of the parent trees, we amplified, cloned, and sequenced regions corresponding to five mapped SFP loci and assayed polymorphisms between the alleles. We identified sequence level variation ranging from single SNP in the 60 -mer region to large in -del polymorphisms affecting >10bp ( Figure 2 4 B E ). Of the five SFP we characterized, one exhibi ted no variation between alleles within the sequence interrogated by the genotyping probe, though sequence variation between the alleles and probe was observed ( Figure 2 4 A ). Therefore, this probe may correspond to a n actual GEM that was mischaracterized as an SFP, as previously described [143] As we hypothesized, SFP we detected are primarily due to species -level sequence polymorphisms between P. trichocarpa and P. deltoides though multiple haplotypes w ere identified at two of the five probes ( Figure 2 4 B,E ).


55 Discussion Parallel genotyping and expression quantification using mRNA microarray hybridization data require accurate classification of differences in signal intensity arising from DNA sequence va riants versus transcript level abundance [140] To separate genetic polymorphism from differences in transcript abundance, candidate genotyping probes can subsequently be detected by identifying individual probes that deviate significantly from the probe set mean signal (which provides a balanced measure of expression), and that segregate in the progeny. Although first demonstrated in populations with simple genetic segregation patterns (i.e. hapl oid, recombinant inbred line and doubled haploid) and species with limited genetic diversity, we have extended mRNA -based microarray genotyping to a highly heterozygous, outcrossing plant species for which low resolution at the genotype level has often ham pered forward genetic gene discovery methods. Contrary to previous studies, which relied on microarray platforms comprising multiple (11 30) short probes ( -mer) per gene [63, 140, 143] we adopted a long-oligonucleotide microarray platform f or use in our study. Furthermore, our analysis relied on single optimal genotyping and gene expression probes selected by analyzing the parental individuals before characterizing the segregating population. A set of six or seven probes per gene was first s creened in the parental genotypes, and an analysis of variance was applied to identify probes interrogating potential polymorphisms and optimal probes for measuring transcript levels [142, 144] Next, the microarray platform was re -designed to comprise a single optimal gene expression pro be for e ach transcriptional unit and 12, 084 candidate SFP probes for analysis of 154 segregating progeny. From this analysis, we identified 1733 segregating features with reasonably low levels of ambiguous data (<10%). After applying a statistically based genotyping correction described previously [64] we successfully mapped 441 of these segregating features


56 (25.4%). Our mapped features include probes that were pre -selected for gene expression analysis and those pre -selected for SFP genotyping, corresponding to 117 GEM and 324 SFP markers. The sample of sequenced SFP regions indicates that our data analysis approach robustly detected sequence variants from RNA -based microarray data. T ogether with 167 framework SSR markers, our map represents one of the highest resolution genetic maps derived from a single pedigree in the Populus genus. Markers from the framework SSR map represent an important tool to delineate true versus spurious linkage of GEM and SFP to linkage groups in the genome, analogous to the situation described when mapping largely homozygous barley RILs [143] Nonetheless, we have demonstrated that GEM and SFP mapping in h ighly heterozygous species is both beneficial and feasible, and may serve as a supplement to traditional DNA -based markers. Our study focused on an inter -specific cross, in which sequence and gene expression variation may be extraordinary. However, estimat es of genetic variation and nucleotide diversity within individual species of the Populus genus [54] and other economically significant outcrossing plants [59, 162164] suggest that our analysis ap proach could also be adapted to identify genetically informative variants from diverse intra specific accessions. However, it is expected that variables including probe length and statistical thresholds associated with allele calling may require optimizati on, and that the abundance of SFP and GEM detected may be lower. Establishing a high -density, gene -based genetic map also provided an opportunity to position previously unlinked sequence scaffolds from the WGS sequence assembly of P. trichocarpa to putativ e genomic locations. The existing genome assembly comprises 410 Mbp of a total estimated genome size of 485 Mbp [4], but there is substantial variation in estimated chromosome sequence coverage, from 56% (chromosome XVII) and 65% (chromosome XIX) to


57 estimated completion (chromosomes X, XII and XIII). Of the 492 unplaced scaffolds in which we identified a segregating GEM or SFP marker, we unambiguously positioned 116 on our genetic map (23.6%). Scaffold sequences mapped using our GEM and SFP markers represent over 35 Mbp of previously unanchored sequence f rom the WGS assembly of P. trichocarpa, including more than 23 Mbp localized by at least two independent markers in the same scaffold. Of a sample of 15 putatively placed genomic scaffolds, the placement of 13 could be verified using independent SSR marker s, lending a good degree of confidence to our map-based re assembly of nearly 50% of the P. trichocarpa scaffold sequence. In addition, 18 scaffolds that we mapped using SFPs or GEMs have been previously mapped using SSRs and amplified fragment length polymorphisms by other research groups (A. Rohde, Institute for Agriculture and Fisheries Research University of Ghent; personal communication). Our microarraybased markers verified the genetic position for 17 of these 18 scaffolds. Misplacement of sampled s caffolds based on microarray marker data was generally due to mapping based on single GEM loci. Because GEMs can result from segregating cis or trans acting regulatory variation, scaffolds mapped based only on GEMs should be verified for their position us ing SSRs where possible. Despite this fact, localization of a large proportion of the previously unplaced genome sequence is a high -impact result for the Populus genomics community, even given the small degree of potential error in placements. Interestingl y, the newly mapped scaffolds are predominantly located in chromosomes with low sequence coverage, where larger gaps exist in the current assembly. It is unclear why there is bias towards mapping scaffolds in chromosomes with poor assembly. There may be a higher probability of mapping unassembled scaffolds to them simply because of their higher expected abundance there. Alternatively, smaller unmapped scaffolds could be more prevalent in


58 chromosomes that are populated by large numbers of hypervariable regions, as high levels of polymorphism are not favorable to long -range WGS assembly of a consensus haplotype [42] Such an observation was recently made in the sex determining telomeric region of Populus chromosome XIX [125] Furthermore, chromosome XVII, which has the lowest estimated percen tage of sequence fully assembled (56%, [4]), has the fourth highest rate of sequence polymorphism (unpublished data), and has the highest number of scaffolds mapped and tot al sequence added in our study. Although we can only speculate as to the basis for this phenomenon, our study provides a significant improvement to the WGS assembly of the P. trichocarpa sequence. Additional mapping studies using SFP and GEM markers that we have identified, and focusing on variation in the sequenced clone Nisqually 1, could shed light on the structural genomic nature of these scaffold sequences and their proper designation in the genome assembly as alternative haplotypes or bona fide unplaced WGS sequence segments. De novo sequencing and assembly of other P. trichocarpa and P. deltoides genotypes will also provide better indications of whether specific regions exist that are hypervariable within and between species haplotypes, and their genome location. Perhaps most importantly, our effort demonstrates the power that microarraybased mapping may bring to future map-based WGS reassemblies. We have shown that mapping based on physically positioned genes can rapidly localize and orient large amounts of WGS -derived sequence within the context of a physical assembly, even when the sequence is scattered amongst a number of small er scaffolds whose assembly is not supported by traditional WGS computerized assembly techniques or anonymous sequence marker anchoring methods. Thus, further application of microarray based mapping in genetically diverse species will not only increase res olution at the level of genotype for forward-genetic analyses, but may drastically


59 improve the initial quality of draft WGS assemblies to the community as a whole. In addition, providing a putative location for an unplaced sequence can identify candidate g enes affecting quantitative phenotypes that would otherwise go unconsidered if relying only upon the chromosome -level sequence assembly for characterization of a genomic interval.


60 Figure 2 1. Framework SSR map of P. trichocarpa X P. deltoides clone 52 225. Black loci: 167 framework markers genotyped in 418 segregating progeny. Red loci: SSR markers that verify positioning of scaffold sequences that were placed on the basis of genotyping SFP/GEM alleles in the mapping population. Pink loc i r egular font: SSR loci verifying positioning of scaffold sequences placed on the basis of single GEM loci ( Table 2 7 ). Pink loci, italicized font: SSR loci that fail to verify predicted positioning for scaffolds 121 and 250 based on GEM loci.


61 Figure 2 2. Examples of significant fixed effects detected by analysis of variance of microarray data from the parents of Family 52 124. Normalized, zero -centered signal measured in seven probes for each parent ( P. deltoides D124 = black line, P. deltoides X P. trich ocarpa 52225 = grey line) in two biolog ical replicates are displayed. A) Significant probe effect (gene ID grail3.0028018001) reflected by wide variation in measured signal intensity around t he probeset mean (probes 2,6,7) B) Significant genotype*probe e ffect (gene ID gw1.XVIII.2378.1) revealed by difference in signal intensity across a probeset within one genotype (probes 1,2 for genotype 52-225; probes 6,7 for genotype D124). C) Significant genotype effect (gene ID gw1.XII.1836.1) represents a property of the probeset as a whole, and is reflected by relatively constant signal variance between genotypes at each probe across the probeset. D) Significant genotype (probes 2 6), genotype*probe (probe 7), and probe effects (probe 1) within a single gene (gene ID eugene3.02350016).


62 Figure 2 3. Microarray and SSR -based genetic map of P. trichocarpa X P. deltoides 2 225. Colors and font styles represent marker types and genomic sequence locations: Black Framework SSR markers; Green SFP markers; Blue GEM markers; Italicized GEM/SFP markers contained within unplaced WGS scaffold sequences in v.1.1 of the P. trichocarpa genome; Plain font GEM/SFP markers with known linkage group anchored genomic coordinates. Maps were generated with publicly available MapChart software, v.2.1 [165]




64 Figure 2 4. Allelic variations characterized by sequencing genomic DNA regions corresponding to mapped SFP probes. Among sequenced clones, haplotypes are shown as detected for P. trichocarpa X P. deltoides clone 52 225 and P. deltoides clone D124. Variations between alleles or between detected sequence and probe sequence are depicted in red. A) No variation detected between parent trees for estExt_Genewise1_v1.C_LG_III2262. B) Ext ensive SNP and in-del polymorphism between haplotypes in grail3.0016013002. C) A 12bp deletion polymorphism in P. deltoides estExt_Genewise1_v1.C_LG_XVII1215. D) A single SNP distinguishes alleles of grail3.0005006601. E) Multiple SNP detected for extExt_Genewise1_v1.C_LG_XVIII1445.


65 Table 2 1. Primer sequences for amplification of scaffold anchored SSR loci. UFLA l ocus ID Primer name v.1.1 s caffold P. trichocarpa product sizea SSR m otif Motif repeats Primer s equence UFLA_29 Scaffold_29_ac_18_F Scaffold_29 246 ac 18 GGATCCGTGACCAAGTTCAG Scaffold_29_ac_18_R GAAGGTGCTCTTTTCCATGC UFLA_166 Scaffold_166_ac_9_F Scaffold_166 257 ac 9 ACACACCATGTGGCTCGTAG Scaffold_166_ac_9_R AGCATCGCATCACATCACAG UFLA_181 Scaffold_181_ag_8_F Scaffold_181 250 ag 8 TCGGTTGGTCTGTATCATCC Scaffold_181_ag_8_R ACGCAATGAGAGGTTTCTGG UFLA_118 Scaffold_118_ag_12_F Scaffold_118 249 ag 12 TCACGCCAGTAACCTTGTTG Scaffold_118_ag_12_R TTCTCGAGGTAAGGTGTCAGG UFLA_130 Scaffold_130_ac_5_F Scaffold_130 246 ac 5 GGTTGGCAATCTACCCTAGC Scaffold_130_ac_5_R GCTATGCGTAACCTGGAAGG UFLA_170 Scaffold_170_ag_5_F Scaffold_170 253 ag 5 CTTCTGGCCTCTATCATGCTG Scaffold_170_ag_5_R CTGCTGACTCCAGCTCAATG UFLA_147 Scaffold_147_ag_9_F Scaffold_147 251 ag 9 ACGAAACCTGGAAAAGGTTG Scaffold_147_ag_9_R AGCAACGCGTAATGTAAAGC UFLA_121 Scaffold_121_ag_11_F Scaffold_121 245 ag 11 CCCTGCTTCATGTCATTTCTG Scaffold_121_ag_11_R TCTACCACAAGGGATTCTTGC UFLA_121_b scaffold_121.102938.at.9F Scaffold_121 280 at 9 AAACTTTGCAACCTTGTCCA scaffold_121.102938.at.9R TCATAACTCGATTTTGAATCCCTA gw97_2 gw1.97.119.1_2L Scaffold_97 300 n/a n/a GCGAATAATTGGAGAACCC gw1.97.119.1_2R CCAATCTCGTCATCAACCTT UFLA_122 scaffold_122.698025.ct.12F Scaffold_122 240 ct 12 CACCATGCCAAGCATCATAG scaffold_122.698025.ct.12R TCCATCATTTGTGTGTGTGC UFLA_250 scaffold_250.163555.gat.5F Scaffold_250 210 gat 5 CTCATGGTATTGGTGAGGGAAT scaffold_250.163555.gat.5R CAGAGGTAGGGTCGGATTCA UFLA_269 scaffold_269.103504.tc.5F Scaffold_269 245 tc 5 TAATTCCACGGAATGGATGG scaffold_269.103504.tc.5R TGAATTCTCTCGCTTAGCTTTG UFLA_395 scaffold_395.21632.ta.5F Scaffold_395 237 ta 5 CGCGACTCGAATCATGAAAT scaffold_395.21632.ta.5R CATATTCACCTGCATGAAAGC UFLA_728 scaffold_728.5575.ct.12F Scaffold_728 256 ct 12 ACCGAAATGTGGGCTATGAA scaffold_728.5575.ct.12R TTGTCAGGCTCAATCTATGATG UFLA_13508 scaffold_13508.142.ctt.8F Scaffold_13508 246 ctt 8 ACAGGTTGTGGGAGGCTGAT scaffold_13508.142.ctt.8R TGCTTTCCACTTTTGATCCAG a. Predicted SSR product size derived from the Nisqually 1 genome sequence are provided.


66 Table 2 2 Summary of F -tests for fixed effects in the mixed ANOVA conducted on parent tree microarray data. Genotype Tissue Tissue* g enotype Probe Genotype* p robe Significant a 7,909 34,326 18,470 51,821 3,355 Non Significant 47,884 21,467 37,323 3,972 52,438 a. Significance judged at FDR < .025.


67 Table 2 3. Primer sequences for amplification of sequence-verified SFP alleles. Gene models are as reported in v.1.1 of the P. trichocarpa Nisqually 1 genome sequence. P. trichocarpa gene m odel Primer forward s eq uence Primer reverse s eq uence estExt_Genewise1_v1.C_LG_III2262 GGTTGGTTCGGTATTGCTGT GTATCGCACAACAGGCATTG estExt_Genewise1_v1.C_LG_XVII1215 GTTCGGGTTATGGGAGGAAT AGTGCCATGAATCCCATTTT estExt_Genewise1_v1.C_LG_XVIII1445 CTCTGGTGGAAGGCTCAAAG GCTGCTCAACTGGAAAAATCA grail3.0005006601 CACATGGCTGGACACAAAAC TTGGCTGGTCACTCCTCTCT grail3.0016013002 TTGAATCTGGTGGTGGTGAA AAGGTGACAACGAGCAGAGAA


68 T able 2 4 Summary statistics for P. trichocarpa X P. deltoides clone 52 225 microarray and SSR -based linkage map. Linkage g roup Framework SSR l oci GEM l oci m apped SFP l oci m apped v.1.1 s caffolds m apped Total m apped l oci Map l ength (cM) Average m arker s pacing (cM) Average r ecombinations (Std. Dev.) LG_I 19 6 25 16 68 415.0 6.10 4.09 (1.51) LG_II 11 10 25 4 52 193.2 3.72 2.18 (1.05) LG_III 11 3 8 6 30 150.5 5.02 1.83 (0.87) LG_IV 7 2 10 7 29 144.5 4.98 1.72 (0.79) LG_V 11 2 12 9 41 154.3 3.76 1.74 (0.84) LG_VI 8 4 11 4 31 189.1 6.10 1.96 (0.83) LG_VII 7 3 6 8 31 118.6 3.83 1.57 (0.84) LG_VIII 14 5 13 2 34 147.7 4.34 1.66 (0.79) LG_IX 8 4 7 10 30 111.7 3.72 1.5 0 (0.64) LG_X 10 8 23 4 45 157.7 3.50 1.84 (0.93) LG_XI 8 3 11 10 33 131.6 3.99 1.73 (1.03) LG_XII 4 4 5 3 20 84.3 4.22 1.41 (0.58) LG_XIII 7 4 10 3 25 115.9 4.64 1.58 (0.69) LG_XIV 10 0 6 7 27 141.8 5.25 1.65 (0.81) LG_XV 6 1 3 6 20 117.7 5.89 1.43 (0.60) LG_XVI 9 3 7 0 19 95.7 5.04 1.38 (0.59) LG_XVII 5 1 5 10 28 132.1 4.72 1.69 (0.81) LG_XVIII 4 1 9 2 16 76.1 4.76 1.62 (0.93) LG_XIX 8 3 11 5 29 121.0 4.17 1.58 (0.86) Unlinked Scaffolds n/a 50 117 n/a n/a n/a n/a n/a Genome Total 167 117 324 116 608 2798.5 4.62 1.80 (0.59)


69 Table 2 5 Summary of WGS scaffold sequences placed based on SFP and GEM markers, and resultant estimated coverage. Linkage Group Assembled s ize (Kbp)a Estimated c overage (%) Scaffold s equence a dded (Kbp) Estimated c overage revised (%) LG_I 35,500 80 5146 .0 91.6 LG_II 24,500 91 98.5 91.4 LG_III 19,100 79 1526 .0 85.3 LG_IV 16,600 95 1387.0 102.9 LG_V 18,000 78 2834 .0 90.3 LG_VI 18,500 92 4295 .0 113.4 LG_VII 12,800 85 582.4 88.5 LG_VIII 16,100 73 8.4 73.0 LG_IX 12,500 85 136.2 86 .0 LG_X 21,100 100 137.8 100.6 LG_XI 15,100 82 1386 .0 89.5 LG_XII 14,100 102 703.4 107.1 LG_XIII 13,100 107 2908.6 130.7 LG_XIV 14,700 85 3162.8 103.3 LG_XV 10,600 79 1792.2 92.3 LG_XVI 13,700 81 0. 0 81.0 LG_XVII 6,000 56 5601 .0 108.3 LG_XVIII 13,500 77 983.4 82.6 LG_XIX 12,000 65 2424.4 78.1 Mean n/a 83.8 1848.1 94.5 a.Original assembled size and estimated coverage as reported [4]. Revised estimated coverage based on these previously reported statistics. Revised estimated coverages exceeding 100% may be due to erroneous estimation of linkage group size due to the assumption of uniform genetic:physical distance ratio, or may result from the map -based linear reassembly of high ly divergent haplotypes that should, in fact, be collinear and distinct.


70 Table 2 6 Verification of scaffold map location for nine sequence scaffolds using SSR markers and the framework SSR map. JGI v.1.1 s equence s caffold Mapped SFP/GEM genes SFP/GEM location Anchored SSR f lanking s caffold in GEM/SFP m ap Verification SSR ID Verification SSR l ocation in f ramework m ap Anchored SSR f lanking s caffold in f ramework m ap Scaffold_29 eugene3.00290072 LG_I, 85.6cM G833, G2837 UFLA_29 LG_I, 119.7cM G833, G3784 estExt_fgenesh4_pg.C_290162 LG_I, 86.3cM Scaffold_130 gw1.130.59.1 LG_IV, 90.3cM G1809, O545 UFLA_130 LG_IV, 109.2cM G1809, O545 eugene3.01300051 LG_IV, 96.8cM Scaffold_166 eugene3.01660055 LG_IV, 0.0cM O349 UFLA_166 LG_IV, 0.0cM O349 Scaffold_181 eugene3.01810009 LG_VII, 65.0cM G354, P2794 UFLA_181 LG_VII, 52.4cM G354, P2794 Scaffold_118 fgenesh4_pg.C_scaffold_118000002 LG_III, 79.7cM G1629, P2611 UFLA_118 LG_III, 97.4cM G1629, P2611 eugene3.01180022 LG_III, 81.8cM Scaffold_170 eugene3.01700010 LG_XVII, 0.0cM G125 UFLA_170 LG_XVII, 0.0cM G125 eugene3.01700027 LG_XVII, 0.0cM fgenesh4_pg.C_scaffold_170000022 LG_XVII, 7.2cM eugene3.01700033 LG_XVII, 13.7cM Scaffold_147 estExt_Genewise1_v1.C_1470180 LG_VI, 189.9cM O50 UFLA_147 LG_VI_b, 94.5cM O50 Scaffold_121 gw1.121.10.1 LG_XVIII, 76.1cM O534, G79 UFLA_121 LG_VI_b, 7.1cM P2221, W12 UFLA_121_b LG_VI_b, 7.1cM P2221, W12 Scaffold_97 gw1.97.119.1 LG_I, 134.2cM G3784, G937 gw97_2 LG_I, 142.5cM G3784, G2837


71 Table 2 7 Verification of scaffold map location for six sequence scaffolds originally mapped based on GEM markers. Verification was conducted using sequence identified SSR markers and the framework SSR map. JGI v.1.1 s equence s caffold Mapped GEM g enes GEM l ocation Anchored SSR f lanking s caffold in GEM/SFP m ap Verification SSR ID Verification SSR l ocation in f ramework m ap Anchored SSR flanking scaffold in framework m ap Scaffold_250 fgenesh4_pg.C_scaffold_250000006 LG_I, 397.5cM G3688 UFLA_250 LG_I, 342.6cM P2786b, P2385 Scaffold_395 fgenesh4_pg.C_scaffold_395000004 LG_II, 9.0cM G734 UFLA_395 LG_II, 0.0cM G734 Scaffold_13508 eugene3.135080001 LG_V, 152.4cM G1255 UFLA_13508 LG_V, 166.6cM G1255 Scaffold_269 fgenesh4_pg.C_scaffold_269000014 LG_XIV, 18.8cM G1866, O59 UFLA_269 LG_XIV, 1.2cM G1866, O59 Scaffold_728 estExt_Genewise1_v1.C_7280001 LG_XV, 0.0cM G1454 UFLA_728 LG_XV, 7.4cM G1454, G1245 Scaffold_122 gw1.122.45.1 LG_XV,26.0cM G1454, G1245 UFLA_122 LG_XV, 28.0cM G1454, G1245


72 CHAPTER 3 THE ONTOGENY OF THE GENETIC REGULATION OF GENE EXPRESSION AND TRANSCRIPTIONAL NETWORKS IN THE WOODY PERENNIAL MODEL Populus This chapter will be submitted to a general biology or plant sciences journal for publication Introduction In humans [87] mice [83] yeast [62, 92] and several higher plants [84, 90, 91, 93, 166] l arge scale eQTL mapping studies demonstrated that genetic control of gene expression is highly complex, with many genes being regulated by a combination of cis acting loci of generally large effect and numerous trans acting elements wit h smaller contributions to mRNA abundance T hese studies have al so detected trans -eQTL hotspots regions of the genome coordinately contributing to varia ble expression in large numbers of unlinked transcripts. T he underlying genetic basis of eQTL hotspots has been a topic of current discussion [167] leading to hypotheses that hotspots may correspond to key regulators of gene expression, biochemical pathways, or developmental transitions. For example, i n yeast and mouse, trans acting regulators of small suites (<100) of genes have been cloned ver ified [98, 106] and shown to control expression of genes in common signaling, growth, and metabolic pathways eQTL mapping, in conjunction with traditional trait QTL analysis, has also identified polymorphisms responsib le for phenotypic variation in several species [83, 112, 168] Such results reinforce the role of transcriptional regulation in evolution [36] and suggest that pleiotropic trans acting eQTL hotspots may be critical for intra and inter -specific diversity in both gene expression and whole plant phenotype s G enome -wide gene expression and eQTL mapping studies have also been leveraged to reconstruct transc riptional networks contributing to biochemical and developmental pathways [78] Network analysis can identify regulators of pathway flux and implicate previously uncharacterized members. Analysis of a priori defined pathways has demonstrated extensive


73 genetic control of underlying gene expression networks in Arabidopsis [169] a finding intricately explored f or the glucosinolate biosynthesis [97, 101] and flowering time [93] T ranscriptional networks can also be identified a posteriori from eQTL data and their biological roles defined by identifying overrepresented metabolic and regulatory functions among networ k members [106] A posteriori approaches have proven particularly useful to demonstrate metabolic and molecular signatures associated with complex phenotypes and disease in model species [83, 110, 112] and we have previously utilized them to define a role for the lignin biosynthesis pathway in tree growth [84, 100] Additional genomic information, including metabolomic, transcription factor binding site (TFBS), and protein protein interaction data have been incorporated in to a few transcriptional ne twork studies [97, 114, 115] increas ing the power to identify key network members [115] Despite the insights into the role of transcriptional networks in shaping plant diversity and evolution, eQTL and transcriptional network studies have generally focused either on the genetic control of gene expression variation in the entire plant [85, 91] or in only single tissues/organs [84, 90, 170] Thus, the ontogeny and diversity of the genetic control of transcription and gene expression networks between tissues in plants (and with few exceptions also in animals [109, 171, 172] ), is largely unknown. The perennial woody plant Populus is a n ideal model to compare the gene tic architecture of gene expression between tissue types because organs are highly differentiated (for instance, woody stems vs. leaves ) and the species has a rapidly expanding genomic toolbox founded upon the genome sequence of Populus trichocarpa [4]. In this study, we employ a pseudo-backcross pedigree of P. trichocarpa and Populus deltoid e s to analyze genome -wide gene expression variation among three distinct tissues ( differentiating xylem, expa nding leaves, and mature roots) Genetic regulation of mRNA abundance is highly


74 di fferentiated among the tissues, and variation in tissue -specific expression programs is common We utilize tissue -specific eQTL to generate co transcriptional networks a posteriori on the basis of eQTL hotspots and predict network regulators on the basis of cis acting expression regulation. We demonstrate these networks to be significantly enriched for biologically coherent groups of genes and well -characterized promoter cis -elements that may play key roles in tissue specific developmental programs and phe notypic diversity between P. trichocarpa and P. deltoides Materials and Methods Plant Material and Growth Conditions A pseudo-backcross progeny (Family 52 124) of 396 individuals from a cross of P. trichocarpa X P. deltoides (genotype 52 225) and P. delto ides (genotype D -124) were propagated and grown as described [75] From a common set of 192 randomly selected individuals grown under high nitrogen conditions (25mM N as NH4NO3; [75]), we collected 180 samples of differentiating xylem, 183 expanding leaves and 163 whole -root s for gene expression analysis. Collected tissues were immediately flash -frozen in liquid nitrogen and stored at 80oC until lyophilization and RNA extraction. We favored using of a single biological replicate of progeny to maximize the size of the population and meiotic events sampled [128] Because th e experiment reflects the analysis of a segregating population, each allele is biologically replicated in approximately half of the individuals of the population [78] RNA I solation and M icroarray A nalysis RNA was extracted from each lyophilized sample by a standard protocol [148] converted to double -stranded cDNA, labeled with cy3, and hybridized to microarray s [173] Hybridizations were carried out using a previously described four -plex NimbleGen (Madison, WI) microarray platform (Gene Expression Omnibus Accession# GPL7234) using probes designed to minimize


75 the effects of sequence polymorphism on the estimates of gene expression [173] The microa rray comprised one probe per gene for 55,793 previously describe d gene models derived from the annotation of the genome sequence of P. trichocarpa clone 'Nisqually 1' (version 1.1; [4]) and a set of nonannotated ESTs. Raw data from all hybridizations was background subtracted, l og2 transformed, and quantile normalized as described [173] eQTL Analysis Each quantile -normalized gene expression value was analyzed as a q uantitative trait using composite interval mapping analysis [66, 67] implemented in QTL Cartographer ([174] ; walk speed = 2cM), based on our genetic map of Family 52 124 that is physically linked and oriented to the genome sequence of 'Nisqually1' [75] Significance of Log of Odds (LOD ) values was estimated for each tissue using a global permutation threshold [91] Global permutation thresholds for each tissue are reported in the footnotes of Table 3 1. eQTL were declared on the basis of a strategy wherein eQTL composed of unimodal LOD curves are located by the peak position [90] Polymodal peaks were declared as sepa rate eQTL if the trough between them exceeded 2 LOD. The eQTL were classified as c is or trans regulated based on co -localization of the eQTL LOD peak for the gene model with the genetic map marker bin containing the gene model in the 'Nisqually 1' sequenc e. While the Family 52 124 map encompasses >85% of the assembled genome sequence [75] 23,116 of the 55,793 gene models/ESTs assessed by our microarray are located on unassembled genomic scaffolds (17,726) or assembled chromosome telomeres outside of the covera ge of the genetic map (5 390). eQTL for these probes were designated as "ambiguous" for the purposes of declaring cis vs. trans regulation.


76 eQTL H otspot D etection and A nalysis To identify significant eQTL hotspots, we permuted the per -bin total eQTL peak counts for each tissue across the 1840 ~2cM bins of the genetic map 1000 times, and determined the 95th percentile of these permutations. Each 2cM bin with a total absolute eQTL peak count greater than this permutation threshold was declared an eQTL hotsp ot. Bins with eQTL counts surpassing the permutation thresholds calculated independently for each respective tissue were compared and classified as "unique" or "shared" eQTL hotspots between the three tissues. To eliminate differential gene density as an e xplanatory factor for eQTL hotspots (i.e. more genes per genetic distance ) we implemented a Chi -squared testing strategy [90] When applied, 238/255 bins continued to be significantly enriched for eQTL (or 93.3% of bins identified based on permutation threshold approa ch). Hotspot -Based Co -Expression N etwork C onstruction We constructed co -expression networks conditioned on the bins declared as eQTL hotspots in the previous analysis. For each ~2cM map bin identified as an eQTL hotspot we selected all genes whose eQTL LOD values surpassed the tissue specific permutation thresholds for eQTL significance (Table 1 footnotes). From the l og2 transformed normalized expression values for these genes in th e respective tissue of interest, we computed pairwise Pearson correlations of the expression values Networks were declared when no fewer than 1 0 genes in a given hotspot bin demonstrated a Pearson correlation of | r | > .80 (uncorrected P < 5.5 x 103 for n = 10) Network edges were constructed and tallied between network members displaying correlations surpassing this threshold. GO A nnotation and Enrichment T esting For each co -expression network constructed, we annotated member genes for Gene Ontology (GO ) [175] categories by conducting a BLASTx search of poplar gene model


77 transcripts against The Arabidopsis Information Resource (TAIR) proteins Release v.8.0. A significant BLAST match was declared at an E value threshold of < 1x105, with transcripts returning E > 1x105 annotated as "no hits". The GO annotation of the closest Arabidopsis ortholog w as assigned to the respective poplar gene. We identified putative orthologs for 45 648 genes on the micro array, of which 36 688 included at least one GO designation in TAIR's database. Overrepresentation of GO categories was tested within each network by producing 2x2 contingency tables for each GO category represented within the network followed by a right tailed Fishers Exact Test (calculating the probability of observing an equal or higher frequency of the category in the network, relative to the genome frequency of the GO category). Because each network was tested for a distinct number of GO enrichments, a Bonferroni correction for multiple testing was applied separately for each network tested, which was computed using the formula Pcorr = .05/ n for networks comprising n distinct GO categories. To further control the frequency of Type I error we did not consider GO categories "enriched" when only one gene in a network was assigned to that category, even if the enrichment was determined to be significant with respect to Pcorr. Cis -Element D etection and Enrichment Testing To annotate the presen ce and absence of common plant cis acting elements in the promoters of the gene models from the P. trichocarpa genome, we extracted the promoter sequences upstream of the start codon for the 55, 793 genes represented in the micro array. U ninterrupted sequenc e of 1,500 bp of length could be isolated for 49 066 genes as the position of the remaining 6 727 was less than 1 500bp from an unresolved sequence region, whole genome shotgun scaffold or contig end. To avoid bias associated with cis -motifs that may be located at preferential distances from the start codon, we did not consider these 6,727 genes in


78 our statistical analysis We downloaded the Plant Cis -Acting Regulatory Element (PLACE) sequence database [176] and determined the presence and absence of motifs within all 49, 066 gene promoters by using Patmatch [177] Among 469 cis -eleme nts deposited in the PLACE database, we detected 360 in at least one gene promoter region included in our analysis. For each of the 360 motifs, we tested each co-expression network for enrichment of genes bearing the motif in question using a right tailed Fishers Exact test. Multiple testing was corrected using a Bonferroni threshold of P < .05/360 = 1.389x104 to judge significance of resulting enrichments. Results eQTL D etection and Genome Distribution T o analyze the role of interspecific variation on the genetic architecture of gene expression in Populus mRNA levels in 180 xylem, 183 leaf, and 163 root tissue samples isolated from 192 progeny of Family 52 124 [75] were assayed using customized microarrays [173] comprised of one 60 -mer probe that represents each of the 45,555 gene models deri ved from the genome sequence of P. trichocarpa 'Nisqually 1' [4] and 10,238 additional EST sequences QTL analysis of normalized signal intensities identified 36,071 significant eQTL in xylem, 13,403 eQTL in leaf, and 9 137 eQTL in roots representing 30,313, 12,392, and 8, 534 genes/ESTs, respectively (Table 3 1) eQTL were classified as cis or trans acting, contingent upon the overlap of the eQTL peak with the marker interval to which the gene model represented by that eQTL was located in the genome. We could not determine cis or trans regulation for genes located in scaffolds that have not been mapped to chromosomes in the physical assembly, and therefore classified them as "ambiguously" regulated (Table 3 1) W ithin each of the three tissues c is acting eQTL were detected at a relatively constant rate (~8 10% of genes) independent of the linkage group or tissue when normalized to account for the varying number of gen es per


79 chromosome ( Figure 3 1 ). In contrast, trans acting eQTL frequency varied widely between different linkage groups and tissues. For instance, in xy lem the number of eQTL rang ed from a low of 76 trans eQTL on LG XII to a high of 20,935 on LG IX suggesting that some chromosomes contain major loci implicated in regulation of mRNA level that segregate in this poplar hybrid population. When eQTL corresponding to ambiguously positioned genes were included, linkage group dependent differences in eQTL frequency increased even more so (Figure 3 2 ). Genetic Regulation of Gene Expression I s Largely Tissue Specific Given the large degree of cellular differentiation among vegetative tissues in poplar, yet the seemingly small degree of differences in expressed genes separating them [4, 35] varying mechanism s of transcriptional regulation are likely to be an important component of tissue specific morphological and developmental differences [79] To address this ascertation we determined the degree of overlap between genes with eQTL in multiple tissues and the location of their eQTL peaks (Figure 3 3 ). In gen eral, sharing of eQTL across tissues was infrequent despite the fact that 4,631 unambiguoulsly placed gene models produced eQTL in both leaf and xylem ( Figure 3 3A ) only 1,389 of these were regulat ed by the same genomic interval in the two tissues ( Figure 3 3B ) The degree of eQTL sharing between xylem and roots was higher, as expected by the fact that both share similar types of tissue (e.g. secondary xylem). A mong 2,105 unambiguously localized genes producing eQTL in both xylem and roots 1,317 wer e regulat ed by the same genomic region. Cis acting eQTL (Figure 3 3C ) w ere shared between tissues at a substantially higher rate than trans acting eQTL ( Figure 3 3D ) in all pairwise comparisions with the exception of xylem and root where sharing was appro ximately equivalent for both eQTL types This outcome is consistent with the hypothesis that local cis acting variants have a more significant effect on gene expression control throughout development than second order, trans -


80 acting effects [91] Contrastingly, our data suggest that trans acting regulatory mechanism s might more common ly govern tissue specific gene expression and subsequently, developmental differentiation. Identification of Tissue -S pecific eQTL H otspots G enomic regions regulating accumulation or turnover of large numbers of transcripts or eQTL hotspots were id entified based on permutation threshold s derived for each tissue Map b ins containing 39 eQTL peaks in xylem, 21 eQTL peaks in leaf, and 16 peaks eQTL in root were identified as significant hotspots Using these critieria, w e detected 67 unique bins corresponding to statistically significant eQTL hotspots in xylem, 97 in leaf, and 88 in root (Table 3 2 ). While a large number of unique bins were enriched for eQTL relative to chance, many of these bins were adjacent to one another ( Figure 3 4 ) and thus likely correspond to a single hotspot a result of the limited mapping resolution of QTL in small populations [128] The eQTL hotspots we detected resulted primarily from the hyper accumulation of trans acting eQTL, and were generally not enriched for cis acting eQTL ( Table 3 3 ). Furthermore, t o test that hotspots did not correspond to regions of high gene density we used a strategy proposed to normalize for number of genes per map bin [90] and found that >90% of the original bins corresponding to eQTL hotspots remained significantly enriched for eQTL (data not shown). Construction of Tissue S pecific, H otspot -B ased C o -Expression N etworks Some eQTL hotspots have been shown to correspond to co transcribed gene sets that are enriched for common functional groups or known biochemical and regulatory pathways [98, 106] Therefore, the se hotspots can serve as a foundation to build transcriptional networks that are connected by common genetic regulation. To create networks, the mRNA abundance data for genes in each eQTL hotspot were used to generate co -expression matrices based on correlations among gene expression measurements across the population. Among the 97 leaf eQTL hotspots


81 detected, we constructed 51 gene co -expression networks within 38 of the original genetic map bins (Table 3 4) The leaf co -expression networks encompassed 1678 distinct genes and ranged in size from 11 to 945 genes (media n = 36 genes). While 38 bins were found to contain expression networks, many of these bins neighbored one another in the genetic map, and resulting networks were highly redundant [128] Nonetheless, at least nine independent leaf coexpression networks were detected within seven bona fide uni que loci (Figure 3 4 Table 3 5 ). Similar results were obtained for xylem and root tissues (Table 3 4) Gene O ntology -B ased Annotation of T issue -S pecific Gene C o -Expression N etworks eQTL hotspots in the segregating population revealed a number of co expre ssion networks, but their biological relevance remained largely unknown. To elucidate how global gene expression variation might translate into biologically meaningful differences in tissues and genotypes, we tested each co -expression network for enrichment of GO categories represented within each network. P oplar genes were categorized based on the GO designation of the most similar Arabidopsis homolog (BLASTx E -value cutoff < 1x105 for significant hits to TAIR release v8.0) resulting in the GO annotation of 36,688 of the 55,793 gene models represented in the microarray For the 51 co-expression networks identified i n leaf tissue, we identified 42 with at least one significant GO category enrichment (median Penrichment = 7.35x107). Analogous results were obtained for xylem ( 75/98 networks, Penrichment(median)= 1.24x105) and root (63/ 75 networks, Penrichment(median)= 9.58x106). Among the most significant enrichments the "blue" hotspot locus on linkage group (LG) VI (Figure 3 4 Bin 734) revealed a network with an overrepresentation of genes associated with chloroplast biogenesis and function ( Penrichment = 1.37x1040). Of the 49 genes present within the co -expressed network, 35 (68.6%) were GO annotated as being localized to the chloroplast (Figure 3 5, Table 3 6 ), a >5.5x enrichment over the number of chloroplast -localized genes expected by chance in a network of this size


82 (4504/36,688 Penrichment = 3.16x1022), thylakoid lumen ( Penrichment = 1.97x106), and chloroplast envelope ( Penrichment = 5.68x1010) cellular components were also significantly enriched in this network ( Table S4 ), reinforcing the notion that this LG VI locus plays an important role in regulating genes related to chloroplast biogenesis and/or function. Using this strategy on all 180 networks with at least one GO categorical enrich ment, we identified 1 83 distinct GO category enrichments, representing 1212 combinations of significant tissue -specific netw orks and enriched GO categories eQTL Based P rediction of P utative N etwork R egulators Developing transcriptional networks represents an initial step towards understanding the relationships between genes in a biological system. However, if one wishes to modify regulatory relationships influencing phenotypes key network regulator s must be identified. Co -expression networks developed on the basis of eQTL hotspots facilitate regulator identification, because differential transcript accumulation is predicted to occur due to a genetic variant underlying the eQTL hotspot position. Therefore, cis regulated g enes belonging to a network defined by an eQTL hotspot represent strong a priori candidate regulators. While this strategy is incomplete in that it will not identify network regulator s differentially controlled outside the realm of transcript abundance, it has previously offered direct evide nce of putative network regulators for downstream investigation [115, 171] We identified putative network regulators for 43 of the 62 leaf co -expression net works, 38 of the 75 root networks, and 50 of the 98 xylem networks. Frequently, more than one putative regulator was identified for each network. For the LG VI leaf co -expression network previously shown to be associated with chloroplast function, we ident ified six network members with gene models and eQTL in cis to the eQTL hotspot. Among these six genes, five code for chloroplast -localized protein products, three of which are known structural components of the chloroplast. These three genes represent the best candidates for


83 regulato rs of this co transcriptional network ( Figure 3 5 Table 3 6 ). Of particular interest is Fts Z2 a gene with a well described role in chloroplast structure, biogenesis, and division [178, 179] In some cases, however, no candidate regulator was associated with GO categories enriched within the network in question. It is possible that a variant not regulated at the level of transcription (i.e., structural post -transcriptional, or DNA sequence -level) is responsible for the eQTL hotspot and co -expression network. Our method of analysis would not i dentify the regulator in this case. Alternatively candidate regulators might not be included in a GO category due to lack of BLAST similarity with the Arabidopsis gene set used for GO annotation of the poplar genes. A number of the co -expression networks include candidate regulatory genes with no hits, indicating this possibility may be quite common. Finally, incomplete GO annotation of genes may mean that candidate regulators identified by our method, while correct, have yet to be assigned a GO annotation with a network associated process because of lack of evidence Enrichment of Transcription Factor Binding Sites in C o -E xpression N etworks In addition to the information obtained from GO annotation enrichment and putative network regulator analysis detaile d above, the functional roles of transcriptional networks may be inferred from conserved TFBS in co regulated genes [115] We utilized the publicly available Plant Cis acting Regulatory Element (PLACE; [176] ) database to define presenc e or absence of 360 cis element motifs in the 1500bp upstream of the start codon in 49, 066 gene models represented on the microarray. Subsequently, each co expression network was tested for significant enrichment of genes bearing each of the PLACE motifs (Fisher's exact test P -value < 1.389x104, after Bonferroni correction for 360 tests per network). Enrichment was detected for 27 motifs in 35 of the 62 leaf gene co-expression networks, 32 motifs in 21 of the 75 root networks, and 36 motifs in 29 of the 98 xylem networks Networks were significantly enriched


84 for as few as one to as many as 26 cis -element s (median of three motifs enriched per network). These conserved m otifs encompass highly variable functions, including light -induced transcriptional modulat ion, calcium responsive expression, sucrose responsive elements, and hormone -responsive elements. Additionally, elements known to bind specific families or classes of transcription factor (Myb-domain, bZIP, AP2) were enriched among several networks. A tota l of 419 combinations of 85 tissue -specific networks and 72 enriched motifs were detect ed amongst the dataset To determine whether additional biological insight could be gained based on motif enrichment, we considered a network detected in xylem on LG IX (map bin 1064) comprised of 110 genes. Among the 108 network member genes for which upstream sequence could be characterized, w e detected significant enrichment for 18 motifs, including motifs common to cytokinin -enhanced binding sites ( P = 4.84x105, 84/ 108 genes), a binding site for ATHB-2 (an Arabidopsis homeobox ZIP transcription factor; P = 1.18x104, 35/108 genes), and two W -box motifs previously associated with ERF3 ( P = 5.73x105, 105/108 genes) and SUSIBA2 ( P = 1.30x105, 101/108 genes) transcript ion factors. SUSIBA2 is a WRKY family transcription factor originally described in barley [180] that is known to regulate amylases and other genes in the starch catabolism pathway whose transcription are responsive to sugars. Among the 101 genes carrying the SUSIBA2 W -box motif several are associated with suga r metabolism, including two amylase homologs, a hexose transporter, and a pyruvate kinase homolog. Further supporting the hypothesis that this network function s in sugar metabolism, the Arabidopsis homologs of 25 of the genes exhibiting the SUSIBA2 binding motif were previously found to be differentially expressed in response to exogenous sucrose ( [181] Tab le 3 7 ). While sucrose associated GO categories were not significantly overrepresented in this network, chloroplast associated


85 functions ( P = 5.30x1012) and amylase activities ( P = 5.38x105) were enriched. Therefore, although the biological interpretatio n of this and other networks remains complex, additional levels of data beyond the information on gene expression can add to the analysis of co regulated eQTL hotspot networks by identifying common links in otherwise seemingly disconnected data. These link s will be vital for understanding network functions in the poplar genome, which is still undergoing refined functional annotation Similarly, in the network affecting co -expression of the chloroplast structural genes (LG VI ) four TFBS were enriched, inclu d ing a pyrimidine box implicated in sugar induced transcriptional repression present in 47 genes ( P = 1.012x104) and a light -responsive transcription element represented in each gene in the network ( P = 1.178x104). These results are perhaps not surprising given the network s close association with chloroplast GO categories and the established relationship of chloroplast function and activity to feedback regulate chloroplast associated gene expression through the metabolic products of photosynthesis [182] Transcriptional N et works Shared Between T issues A re Regulated B y D istinct Loci The ontogeny and conservation of transcriptional networks among different plant tissues is largely unknown. Interestingly, by contrasting the localization of eQTL hotspots among vegetative tissues and we observed that hotspot overlap was minimal only 9 hotspot bins were conserved between leaf and xylem, 1 1 between xylem and root, and 9 between root and leaf (Table 3 3 ). Only two hotspot bins were shared between all three tissues suggesting that large transcriptional networks are generally tissue specific Furthermore, very few genes exhibited eQTLs controlled by the same hotspot in different tissues ( Table 3 3 ). Therefore, eQTL hotspots could correspond to important genetic regulators governing m ajor differences in structure or development between tissues in poplar


86 Our analysis of eQTL and eQTL hotspot sharing between tissues suggested that individual genes and networks are commonly regulated by tissue -specific loci. However, even if networks are regulated by distinct loci in separate tissues, they could still share network members. Significant gene sharing between tissue specific networks was tested with a Chi -squared analysis, accounting for the number of probes in each network and shared between networks. Several instances of shared network membership were observed after correction for multiple testing at a Bonferroni threshold of .01/ n for n cross tissue network -to -network comparisons where at least 5 gen es were shared between networks For instance, a large leaf network of 428 genes on LG XI (interval 1226, network 1) shared 171 gene member s with networks in xylem. Among the set of shared genes, 91 were located in a LG I network of 928 gene s in xylem (interval 252, network 1; P An additional 53 genes in the leaf network on LG XI were located in a distinct xylem co -expression network on LG XV (interval 1557, network 1, 200 genes; P etwork 1, 1057 genes; P < 1.832x109). Thus, a total of 156 of the 171 shared network genes from the leaf LG XI network were restricted to just three networks in xylem, including 13 leaf network genes appearing in both the LG XV and LG I xylem networks. In total, we found 414 significant network pairs that were statistically enriched for shar ed genes among leaf and xylem. Similar patterns of cross tissue network member conservation were obtained for the comparisons of leaf with root and xylem with root. Int erestingly, however, these comparisons identified only 31 networks in leaves significantly overlapping for membership with 33 networks from xylem. Furthermore, the 31 networks from leaf were restricted to four genomic regions residing on linkage group s V I, IX, XIV and XVI. Similarly, the 32 xylem networks were derived from only six unique genomic regions containing networks ranging from 30 to 5787 genes These outcomes are mirrored in the


87 other pairwise comparisons between tissues our analysis identified significantly overlapping network memberships between 31 leaf, 64 root, and 43 xylem networks. These results suggest that a handful of distinct trans acting factors might control expression of coordinate groups of genes depending on the tissue being considered, and also indicate that different combinations of biological subnetworks could be combined to potentially drive tissue diversification. Discussion Unraveling the orchestrated action of genes and modeling their interactions in a biologi cal system is among both the most significant challenges and ultimate goals of biology. We ll described gene networks should help predict growth and development, as well as the outcome of perturbations from biotic and abiotic stresses [183] Here, a traditional quantitative genetic analysis was used to assemble networks that describe gene expression patterns and infer biological function, as well as mechanisms of regulation by identifying putative regulators and overrepresented cis elements Networks were identified within three developmentally diverse plant tissues by analyzing a n interspecific pseudo-backcross progeny of the model hardwood forest tree genus Populus Our results indicate that genetic variation o f gene expression accounted for by allelic differences between P. trichocarpa and P. deltoides are complex and abundant with a large number of transcripts regulated by combinations of local, cis acting variants of generally large effects, and distant trans acting variants of generally less significant effects. Interestingly, among leaf, root, and xylem the prevalence of eQTL was markedly distinct, with approximately three -fold more genes exhibiting genetic control over gene expression in xylem than either root or leaf. This is consistent with our previous observation that stem tissues have the great est diversity of expressed genes among poplar vegetative tissues [35] Gen es that are tissue or organ -specific may also evolve more rapidly, making them more likely to undergo


88 diversification in expression [36, 184] However, such factors alone cannot fully explain the vast increase in eQTL prevalence within xylem, since previous analysis showed < 3500 genes expressed exclusively i n stem tissue at a common FDR threshold [35] The di fference b etween eQTL frequencies among tissues was almost completely accounted for by a prevalence of trans acting eQTL in xylem compared to leaf and root (Table 3 1) However, this striking difference noted in xylem is primarily due to eQTL comprising a few major hotspots, particularly one on LG IX encompassing >14,000 genes, or around 25% of the coding genome We inte rpret the results associated with this particular hotspot cautiously, as other major pleiotropic loci detected within this study resulted in multiple subnetworks which were enriched for numerous, distinct GO categories. However, the major xylem locus on LG IX produced a single connected network comprising nearly 5,800 genes ( Table 3 4 ), and with relatively few enriched GO categories given its size. Nonetheless, hotspots of similar orders of magnitude (comprising up to 10 15% of the coding genome) are detect ed on chromosome 2 of Arabidopsis in crosses between accessions Bayreuth 0 and Shahdara [91] and Landsberg erecta x Cape Verde Islands [114] One of these hotspots has been attributed to the ERECTA locus, a known pleiotropic regulator of morphology and development that varies between different accessions [114] Our pedigree represents an interspecific cross from a dioecious and obligate outcross ing organism, encompassing levels of sequence variation and genome complexity (i.e., chromosome structural variation, [125] ) that may not be adequately modeled by intraspecific pedigrees produced in self pollinating model p lants such as Arabidopsis Therefore, we believe our overall results support the view that most eQTL hotspots detected in Populus are biologically relevant. It is likely that fundamental differences in gene expression regulatory cascades especially those mediated through trans acting factors, play key role s in development and morphological differences in Populus Still, the


89 specific mechanisms by which these effects occur clearly requires a more detailed case by-case investigation A second and related out c ome of our eQTL analysis was the surprising frequency with which tissue -specificity was associated with gene expression regulation. While thousands of genes exhibited eQTL in multiple tissues, the conservation associated with the location of the genomic r egion leading to differential regulation was limited, ranging from ~30.3 60.3% depending on the tissues considered (Figure 3 3) Conserv ed regulation between tissues was strongly biased towards cis acting eQTL, whereas trans -eQTL were significantly more li kely to be tissue -specific. Interestingly, this phenomenon extended not only to the analysis of eQTL hotspots, but also to gene expression networks. We found that a small number of tissue -specific hotspots frequently regulated expression subnetworks that, in other tissues, were regulated by distinct and unlinked loci. These results build upon those generated for barley, wherein limited pleiotropy associated with tissue -specific cis acting eQTL was investigated [185] and found to be relatively common Here, we have demonstrated that the tissue -specific modulation of single genes, as well as interconnected networks and biologically cohe rent co transcriptional modules, may represent a key component of differentiation between tissues and individuals in higher plants. This finding greatly increases the potential for regulatory complexity to play a key role in diversification of species and tissu es. A seemingly small number of genes, each regulated in a tightly controlled, tissue -specific manner by complex assortments of multiple cis acting elements and trans acting regulatory factors exponentially increase the number of combinations that can act in concert to generate phenotypic diversity. The regulatory landscape becomes even more complex when we consider that each trans acting factor is likely subjected to similar patterns and modes of regulation. These results and hypotheses are in sound agreem ent with previous


90 findings, which demonstrated the overall presence or absence of expression of specific genes is fundamentally similar between tissues in poplar and Arabidopsis [35, 186] despite the fact that molecular signatures associated with transcript abundance estimates cle arly distinguish organ and tissue types [186] Our approach in this study be gins to reconcile regulatory mechanisms by which these results may be explained. As information for the Populus model genetic system continue to grow, additional levels of data including miRNA activity, proteinprotein interaction, and epigenetic regulatio n will continue to improve our understanding of regulatory mechanisms driving tissue specialization in higher plants. Nonetheless, our current effort provides a foundation upon which these new levels of data can be laid. A key goal of systems biology is th e ability to generate testable predictions of system -wide behavior in response to specific perturbation(s). Generating networks of co -expressed genes represents an initial step toward this goal in Populus as co -expressed gene sets are expected to be respo nsive to perturbations among highly connected network members and putative network regulators. Our co assembly of gene expression QTL into hotspots has revealed no less than 50 tissue -specific co -expression networks and associated candidate regulators Pre vious studies in the simpler yeast genetic system showed that coexpression networks developed from gene expression and eQTL data are more predictive than those produced based on expression data alone [115] For example, eQTL -based networks in this yeast study more accurately pre dicted associations between genes regulated by common transcription factors, and showed increased ability to predict expression signatures associated with genetic and pharmacological perturbation. Accordingly, we anticipate that the networks produced in ou r study are similarly predictive of system behavior in response to specific genetic disruptions. A number of testable hypotheses can be generated from these data, including the response of specific co-expression networks to


91 perturbation of their putative regulator(s), the role of previously described transcription factors in coordinating networks exhibiting enrichment for specific cis -element motifs, and the contribution of co-expressed networks to variation in phenotypic traits associated with network en riched GO categories. We have explored only a very limited sample of these hypotheses within the context of this article, by considering the role of statistically enriched cis elements in a xylem network that may be regulated by sugar signaling, and identi fying a network heavily enriched for genes that are likely to be associated with variation in chloroplast function, morphology, and/or biogenesis. While the experiments to directly test and verify individual hypotheses are beyond the scope of the current e ffort, the results enhance our understanding of the interconnectedness of the genome, transcriptome, and regulatory elements in the poplar genetic system. Furthermore, experiments to validate the role of specific candidate regulators and networks on molecu lar and morphological phenotypes will be the focus of subsequent efforts from our group.


92 Figure 3 1. Global distribution of eQTL across linkage groups for A) xylem, B) leaf, and C) root tissues expressed as the fraction of mapped gene models. In cases where values on the eQTL/mapped genes axis exceed 70, values are provided in brackets. See Table 1 1, footnotes b and c for description of cis and trans -eQTL categorization procedure.


93 Figure 3 2 Global distribution of all eQTL (ambiguous and unambigu ous) across linkage groups for A) xylem, B) xylem rescaled for comparison with C) leaf, and D) root tissues. See Table 3 1 footnotes for a description of cis trans and ambiguous eQTL categorization procedure.


94 Figure 3 3 Overlap between probes and eQTL detected among the three t issues considered. All figure subparts exclude probes with ambiguously positioned eQTL. A) Overlap of probes (gene models) producing eQTL in each of the three tissues B) Cross tissue conservation of the genomic regulatory region (eQTL location) for probes in A. C) Cross tissue conservation of genomic regulatory position for cis -eQTL in B D) Cross tissue conservation of genomic regulatory position for trans -eQTL in B. Totals for sections in C a nd D do not equal that in B, as a small number of probes produce conserved eQTL s mapping in both cis and trans


95 Figure 3 4 Genome -wide linkage scan of expression traits and demarcation of eQTL hotspots producing co -expressed gene networks in leaf tissue. Similar results were obtained for root and xylem tissue, detailed in Table 3 4 (linkage scans for other tissues not shown).


96 Figure 3 5 Leaf co -expression network constructed from the "blue" eQTL hotspot in Figure 3 4 and enriched for chloroplast (CP) related Gene Ontogeny categories. Hu b colors indicate annotation nature of network members, and edges indicate Pearson correlation between co expressed genes of | r | > .80. Gene IDs relate to gene models presented in Table 3 6


97 Table 3 1 Summary of eQTL detected for each of the three poplar tissues. Tissue Category Total numbera Number c isb Number transc Number a mbiguous d Xylem Genes with eQTL 30,313 2,398 16,353 11,562 eQTL 36,071 2,408 18,684 14,979 Leaf Genes with eQTL 12,392 3,509 4,178 4,705 eQTL 13,403 3,528 4,500 5,375 Root Genes with eQTL 8,534 2,156 3,120 3,258 eQTL 9,137 2,161 3,321 3,655 a. The permutation -derived LOD thresholds for eQTL significance were 2.89 for xylem, 2.92 for leaf, and 2.93 for root. b. A gene was declared as having a cis -eQTL if the eQTL peak overlapped at any point with the genetic interval containing the corresponding gene model in the genome, as defined by its flanking SSR, SFP or GEM markers. c. A gene was declared as having a trans eQTL if the eQTL did not overlap at any point with the genetic interval containing the corresponding gene model in the genome, as defined by its flanking SSR, SFP or GEM markers. d. A gene was declared as having an ambiguous eQTL if the position of the corresponding gene model was not flank ed on both sides by SSR, SFP, or GEM markers, or if the gene resided on an unplaced genomic sequence scaffold from the genome assembly.


98 Table 3 2. Significant eQTL hotspots by linkage group in each of the three tissues. Linkage group Total gene m odels (GMs) Total GMs covered by map Total xylem hotspot binsa Total leaf hotspot binsb Total root hotspot binsc 1 4181 3469 12 10 6 2 2940 2856 1 2 0 3 2097 1854 0 2 3 4 2040 1881 0 0 2 5 2402 2081 3 2 5 6 2526 2412 3 6 5 7 1466 1302 0 1 0 8 2169 2148 1 2 3 9 1749 1667 21 9 11 10 2586 1749 0 6 1 11 1624 1300 3 14 3 12 1478 648 0 1 6 13 1703 2323 0 5 9 14 2391 1916 6 14 10 15 1377 1076 10 4 0 16 1528 1501 0 14 9 17 1074 489 7 1 8 18 1425 736 0 1 1 19 1311 1269 0 3 6 Total 38067 32677 67 97 88 a. The permutation -derived hotspot threshold value for xylem was 39 eQTL. b. The permutation derived hotspot thresh old value for leaf was 21 eQTL. c. The permutation-derived hotspot threshold value for root was 16 eQTL .


99 Table 3 3. Summary of tissue -specific eQTL hotspot -based co -expression network construction in Family 52 124. Linkage group Bin number Xylem eQTL Leaf eQTL Root eQTL Probes shared Cis Trans/Am Cis Trans/Am Cis Trans/Am Hotspots shared by xylem, leaf and root 5 708 7 205 4 21 0 24 0 8 1014 6 35 13 9 6 22 3 Hotspots shared by xylem and leaf only 1 1 6 77 10 204 2 2 276 4 59 9 23 4 9 1015 8 114 8 19 3 9 1020 1 52 2 20 0 9 1023 4 39 4 25 0 9 1028 2 61 4 29 0 9 1030 6 35 3 22 1 Hotspots shared by xylem and root only 1 131 6 34 1 41 0 9 1096 22 14132 0 65 14 11 1263 9 68 2 17 2 14 1395 13 799 1 57 0 14 1396 6 140 0 19 0 14 1397 1 76 2 23 0 14 1398 0 52 0 20 0 14 1399 0 180 0 19 0 17 1664 2 83 0 33 1 Hotspots shared by leaf and root only 5 682 10 12 16 1 0 7 838 9 16 3 15 1 11 1222 12 156 7 11 9 13 1333 9 27 3 63 1 14 1498 12 25 13 37 20 16 1628 9 21 7 14 3 19 1783 12 25 8 33 4

PAGE 100

100 Table 3 4. Summary of tissue -specific eQTL hotspot -based co -expression network construction in Family 52 124. Tissue eQTL h otspot b ins d etected Bin co expression n etworks Bins with n etworks Max network s ize Median n etwork s ize Total g enes in n etworks Minimum independent co -expression n etworks Minimum i ndependent g enomic regions Leaf 97 51 38 945 36 1678 9 7 Root 88 75 55 217 33 1188 16 11 Xylem 67 97 62 5787 99 9369 28 16

PAGE 101

101 Table 3 5 Network -producing t issue -specific eQTL hotspots detected in leaf tissue and characteristics of their associated co expression networks. Hotspot locus color Hotspot eQTL count Number of networks constructed Network s ize(s) Most significantly enriched GO category GO category type GO enrichment nominal P value Most significantly enriched cis element PLACE enrichment nominal P value LG Genomic physical i nterval Magenta 275 1 31 tubulin complex CC 1.615 x10 5 AGATC 1.236 x10 4 1 .22 1.52Mb Green 114 1 18 n/a n/a n/a n/a n/a 1 6.87 ?Mb Blue 239 1 51 chloroplast CC 1.370 x10-40 AATAAT 6.129 x10-5 6 2.27 3.63Mb Purple 892 2 428 chloroplast CC 5.703 x10-5 AGCGGG 1.979 x10-7 11 4.87 12.24Mb 36 protein amino acid phosphorylation MF 4.458 x10-8 TGCAAAG 1.923 x10-5 11 4.87 12.24Mb Teal 1421 2 945 DNA r ecombination MF 1.161 x10-18 CCACGTCATC 5.394 x10-6 14 4.717.1Mb 27 n/a n/a n/a CGCGGCAT 4.426 x10 5 14 4.71 7.1Mb Orange 316 1 21 chloroplast envelope CC 2.297 x10-4 AATAGAAAA 2.328 x10-5 16 .04.67Mb Red 126 1 11 manganese ion binding MF 1.043 x10-17 n/a n/a 19 02.31Mb

PAGE 102

102 T able 3 6 Annotation of members and putative regulators of a leaf -specific coexpression network enriched for chloroplast biogenesis and functionality. P. triochcarpa g ene m odel Network e dges Description/ f unction Chloroplast a ssociateda Cis regulated Candidate r egulator Gene ID number b estExt_Genewise1_v1.C_LG_I7940 4 glycyl tRNA synthetase / glycine -tRNA ligase N N 1 estExt_Genewise1_v1.C_LG_IV2768 2 RNA recognition motif (RRM) containing protein N N 2 estExt_Genewise1_v1.C_LG_IX1928 2 cytochrome c biogenesis protein family Y N 3 estExt_Genewise1_v1.C_LG_VI1166 1 FTSZ2 2 (FtsZ2 2); structural molecule Y Y 4 estExt_Genewise1_v1.C_LG_X3683 15 DEAD/DEAH box helicase Y N 5 estExt_Genewise1_v1.C_LG_XIII0363 21 similar to hypothetical protein [Vitis vinifera] Y N 6 estExt_Genewise1_v1.C_LG_XV1465 3 phosphoglycerate/bisphosphoglycerate mutase family protein Y N 7 estExt_fgenesh4_kg.C_LG_I0088 14 SWIB complex BAF60b domain containing protein Y N 8 estExt_fgenesh4_kg.C_LG_XV0016 7 GAMMA CAL2 (GAMMA CARBONIC ANHYDRASE LIKE 2); acyltransferase/ transferase Y N 9 estExt_fgenesh4_pg.C_LG_I2552 3 GTP binding protein LepA Y N 10 estExt_fgenesh4_pg.C_LG_VI0491 15 chloroplast thylakoid lumen protein Y Y 11 estExt_fgenesh4_pg.C_LG_VI1540 1 SQD1 (sulfoquinovosyldiacylglycerol 1); UDPsulfoquinovose synthase Y N 12 estExt_fgenesh4_pg.C_LG_XIX0756 2 50S ribosomal protein L21 Y N 13 estExt_fgenesh4_pg.C_LG_XVII0327 11 porin Y N 14 estExt_fgenesh4_pm.C_LG_IX0206 7 mannose 6 phosphate reductase (NADPH dependent) N N 15 estExt_fgenesh4_pm.C_LG_IX0706 11 SLP Y N 16 estExt_fgenesh4_pm.C_LG_V0614 1 similar to unnamed protein product [Vitis vinifera] (GB:CAO61051.1); contains InterPro domain Low temperature viability protein (InterPro:IPR007307) N N 17 eugene3.00011927 17 heavy metal associated domain containing protein Y N 18 eugene3.00031286 1 similar to unknown protein [Arabidopsis thaliana] (TAIR:AT5G27730.1); similar to unnamed protein product [Vitis vinifera] (GB:CAO68737.1) N N 19

PAGE 103

103 T able 3 6 Continued P. triochcarpa gene model Network edges Description/function Chloroplast associateda Cis regulated Candidate regulator Gene ID number b eugene3.00060837 5 similar to unknown [Populus trichocarpa x Populus deltoides] (GB:ABK96465.1) Y Y 20 eugene3.00111342 3 similar to hypothetical protein OsI_013284 [Oryza sativa (indica cultivar group)] (GB:EAY92051.1); similar to unknown protein [Oryza sativa (japonica cultivar group)] (GB:AAK09232.1); contains InterPro domain Protein of unknown function DUF858 Y N 21 eugene3.00190981 2 TIC110 Y N 22 fgenesh4_pg.C_LG_VI000277 9 thylakoid lumenal 20 kDa protein Y N 23 fgenesh4_pg.C_LG_VIII001084 20 RPL15 (ribosomal protein L15) Y N 24 fgenesh4_pm.C_LG_VIII00064 1 3 CAAX amino terminal protease family protein N N 25 grail3.0003038902 2 APG1 Y N 26 grail3.0013033101 8 oxygenevolving complex related Y N 27 grail3.0023008101 1 macrophage migration inhibitory factor family protein / MIF family protein Y Y 28 grail3.0024031501 2 ATSFGH (ARABIDOPSIS THALIANA S FORMYLGLUTATHIONE HYDROLASE); S formylglutathione hydrolase/ hydrolase N Y 29 gw1.I.1156.1 7 sugar transporter family protein Y N 30 gw1.I.5571.1 12 PRPL11 (PLASTID RIBOSOMAL PROTEIN L11); structural constituent of ribosome Y N 31 gw1.I.976.1 4 inorganic phosphate transporter Y N 32 gw1.IV.11.1 1 similar to hypothetical protein MtrDRAFT_AC147482g2v2 [Medicago truncatula] (GB:ABD32485.1) Y N 33 gw1.IV.1864.1 5 phosphoprotein phosphatase/ protein kinase Y N 34 gw1.IV.3048.1 22 ribosomal protein S5 family protein Y N 35 gw1.VI.2404.1 11 thioredoxin reductase Y Y 36

PAGE 104

104 Table 3 6 Continued P. triochcarpa gene model Network edges Description/function Chloroplast associateda Cis regulated Candidate regulator Gene ID number b gw1.VII.4002.1 2 similar to unknown protein [Arabidopsis thaliana] (TAIR:AT4G24530.1); similar to unnamed protein product [Vitis vinifera] (GB:CAO61608.1); similar to unknown [Populus trichocarpa] (GB:ABK95421.1); similar to hypothetical protein [Vitis vinifera] (GB:CAN78 778.1); contains InterPro domain Protein of unknown function DUF246 N N 37 gw1.X.2172.1 1 SCL30 (SC35 like splicing factor 30); RNA binding N N 38 gw1.X.2264.1 7 OVA1 (OVULE ABORTION 1); ATP binding / aminoacyl tRNA ligase Y N 39 gw1.XIII.1619.1 2 similar to unknown protein [Arabidopsis thaliana] (TAIR:AT5G28500.1); similar to unnamed protein product [Vitis vinifera] (GB:CAO65032.1) Y N 40 eugene3.00440183 5 similar to hypothetical protein [Vitis vinifera] (GB:CAN83711.1); similar to unnamed protein product [Vitis vinifera] (GB:CAO22167.1); similar to unknown [Populus trichocarpa] (GB:ABK94059.1) N N 41 eugene3.00860035 3 CPHSC70 1 (chloroplast heat shock protein 70 1); ATP binding / unfolded protein binding Y N 42 eugene3.01180078 6 BAM3 Y N 43 eugene3.01570036 13 lactoylglutathione lyase Y N 44 grail3.0107000401 15 chloroplast 30S ribosomal protein S20 Y N 45 gw1.16424.1.1 2 60S ribosomal protein L18A (RPL18aC) N N 46 gw1.210.7.1 1 protein arginine N methyltransferase family protein N N 47 gw1.129.1.1 3 similar to unknown [Populus trichocarpa x Populus deltoides] (GB:ABK96654.1); contains InterPro domain Protein of unknown function DUF1118 (InterPro:IPR009500) Y N 48 gw1.XVI.617.1 3 chloroplastic RNA binding protein P67 Y N 49 a Chloroplast association judged by membership in GO category GO:009507 (chloroplast) or its associated daughter terms in the G O heirarchy. b. Gene ID number corresponds to network gene identifiers shown in Figure 3 5.

PAGE 105

105 Table 3 7. Poplar genes from the sugar associated co -expression network on linkage group IX in xylem that were also regulated by exogenous sucrose in a previous study [181] conducted in Arabidopsis P. trichocarpa g ene m odel Arabidopsis homolog a Functional a nnotation fgenesh4_pg.C_LG_IX001309 AT1G07670 calcium transporting ATPase estExt_fgenesh4_pg.C_LG_VIII0870 AT1G48170 similar to expressed protein [Oryza sativa (GB:ABA95965.1 )] gw1.XIII.1587.1 AT1G64260 zinc finger protein related gw1.158.130.1 AT1G72440 EDA25 (embryo sac development arrest 25); binding grail3.0015003801 AT1G74560 NRP1 (NAP1 RELATED PROTEIN 1); DNA binding / chromatin binding / histone binding gw1.IX.4200.1 AT2G29050 ATRBL1 (ARABIDOPSIS THALIANA RHOMBOID LIKE 1) fgenesh4_pm.C_LG_X000013 AT3G06483 PDK (pyruvate dehydrogenase kinase) estExt_Genewise1_v1.C_LG_IX3773 AT3G47160 protein binding / zinc ion binding estExt_fgenesh4_pm.C_LG_VIII0091 AT3G52990 pyruvate kinase eugene3.00051316 AT3G54670 ATSMC1 estExt_Genewise1_v1.C_LG_II1288 AT3G61140 CSN1 estExt_fgenesh4_pm.C_LG_IX0496 AT3G63490 ribosomal protein L1 family protein estExt_Genewise1_v1.C_LG_I3371 AT4G13850 GRP2 eugene3.00100701 AT4G14660 RNA polymerase Rpb7 N terminal domain containing protein eugene3.00120829 AT4G24770 CP31 gw1.IX.3070.1 AT4G29430 RPS15AE (ribosomal protein S15A E); structural constituent of ribosome gw1.VII.1484.1 AT4G33470 HDA14 (histone deacetylase 14); histone deacetylase fgenesh4_pm.C_LG_IX000514 AT4G34890 ATXDH1 (XANTHINE DEHYDROGENASE 1); xanthine dehydrogenase grail3.0011007002 AT5G09250 KIWI; DNA binding / transcription coactivator estExt_fgenesh4_pm.C_LG_IX0400 AT5G09590 mtHSC70 2 estExt_fgenesh4_pm.C_LG_V0048 AT5G10360 EMB3010 (EMBRYO DEFECTIVE 3010); structural constituent of ribosome eugene3.00090483 AT5G11880 diaminopimelate decarboxylase estExt_fgenesh4_pg.C_LG_XI0654 AT5G39850 40S ribosomal protein S9 (RPS9C) grail3.0060007502 AT5G46160 ribosomal protein L14 family protein / huellenlos paralog (HLP) estExt_fgenesh4_pg.C_LG_IX1183 AT5G59030 COPT1 (COPPER TRANSPORTER 1); copper ion transmembrane transporter a. Arabidopsis homolog from genome annotation by The Arabidopsis Information Resource, v.8.0

PAGE 106

106 CHAPTER 4 UTILIZING GENETICAL GENOMICS TO IDENTIFY AN ADP -RIBOSYLATION FACTOR P t ARF1 AS A CANDIDATE GENE FOR LEAF SHAPE VARIATION IN P opulus Portions of this chapter will be submitted to a plant biology -related journal for publication Introduction Throughout the plant kingdom, leaf morphology demonstrates remarkable phenotypic diversity, making it a favorable system in which to study the evolution of variations in form. A number of genes and networks have been described that affect initial leaf development and pattern formation (reviewed thoroughly by Byrne [187] ) in both simple and complex leaves [188, 189] Similarly, mutagenesis screens have identified genes that appear to function in the regulation of leaf blade shape, specifically w idth [131, 134, 138] length [131133] and their ratios. An emerging paradigm from these discoverie s is that two -dimensional leaf shape can be regulated both by differential cell elongation (polarity) or differential cell proliferation favoring one dimension versus the other [129] Despite this molecular genetic framework of leaf initiation, development, and shape established through mutagenesis of model plant s comparably little is known about whether the same molecular mechanisms underlie evolutionary differences in leaf morphological variation. For instance, to our knowledge no study has directly addressed whether alternative alleles at these g enes underlie variation in leaf shape within or between different plant species (although Street et al. [123] take steps toward answering this q uestion in the Populus genetic system). In light of this shortcoming and the extensive diversity in leaf form, additional studies exploiting naturally occurring variation are needed to illuminate the role of previously discovered genes and networks in evol utionary variation for leaf traits. Quantitative trait locus (QTL) based approaches are one method that can be readily applied to identify genetic loci responsible for existing, natural diversity in leaf traits QTL have been identified for leaf morphologi cal traits in several species including tomato [190] poplar [70] oak

PAGE 107

107 [191] maize [192] and Arabidopsis [193] show ing tha t several aspects of leaf morphological variation are subject to multigenic control. Thus, it is apparent from these studies that loci thus far discovered on the basis of mutagenesis m ight only explain a portion the naturally occurring variation in leaf shape. However, molecular characterization of these regulatory loci has been generally prohibitive because of the histori cal challenges associated with moving from QTL to gene [194] Recently, novel experimental approaches developed from advances in genome sequence production and genomic technology have eased QTL cloning [78, 89, 195] making the QTL -based approach a powerful tool to elucidate molecular mechanisms underlying naturally occurring phenotypic variation [83, 100] The genus Populus is a particularly favorable system in which to apply these approaches, given its extensive genetic and phenotypic v ariation, the availability of several well established interspecific pedigrees, and rapidly growing genomic toolbox founded on the genome sequence of P trichocarpa [196] The genus is comprised of five evolutionary sections, and leaf morphology is widely regarded as diagnostic of evolutionary relationships at the sectional level [8]. In addition, several studies have shown that leaf morphological characters are predictive of long term clonal performance a nd growth [119, 121, 197] Therefore, detailed study of intersectional poplar hybrids may provide a n approach to identify loci associated variation in leaf morphology, and indirectly, clonal productivity Here, we analyze d an intersectional pseudo -backcross pedigree of narrow leaf P. trichocarpa and wide lea f P. deltoides for variation in leaf lamina shape From a major QTL for leaf lamina width and a genome -wide analysis of gene expression in expanding leaf tissue, we utilized expression QTL ( eQTL ) analysis to identify an ADP ribosylation factor (ARF) GTPase, which we designate ARF1 as a candidate gene regulating leaf morphology in this pedigree by a genetical genomics approach. We subsequently develop evidence to demonstrate the role of

PAGE 108

108 differential expression of ARF1 leaf width variation between P. trichocarpa and P. deltoides ARF proteins have been shown to control the localization of auxin efflux carrier s [198] which function to establish auxin gradients and apical basal cell polarity in developing plant organs suggesting that evolution of differential cellular polarity and auxin flux may play a significant role in leaf morphological variation observed in subgenera of genus Populus Materials and Methods Plant Material and Phenotyping A previously described [75, 173] pseudo -backcross pedigree Family 52 124 of P. trichocarpa X P. deltoides (clone 52 225) and P. deltoides (clone D 124) was utilized for phenotyping and subsequent identification of QTL and eQTL. Plants were grown as described [75, 173] Leaf morphological QTL were identified from traits measured digitally using Image Pro Plus software ( Media Cybernetics, Inc., Bethesda, MD) from a n image scan of the leaf closest to one -half the live crown height in three biological replicates of 396 individuals. Traits measured included leaf lamina length, width, and their ratio. Leaf length was measured along the midvein from the junction of the lamina and petiole to the distal tip of the leaf. Blade width was measured at the widest point of the lamina. An analysis of variance [75] was applied to the phenotypic measurements and least -square means (LSM) were estimated for each of the individuals in the population and were utilized for QTL analysis. Cl onal repeatability was calculated as described [75] except excluding the effects of row and column posi tion in the experimental design. Genotyping and Genetic Mapping of Progeny A previously described mi crosatellite (SSR) and microarray based genetic map [75] was u tilized for whole -genome QTL mapping of the leaf width trait. Additional SSR markers within the primary QTL interval were identified from the 'Nisqually 1' genome sequence (Table 4 1,

PAGE 109

109 [4]) using MsatFinder v.2.0 software. Primers were designed for these loci within the MsatFinder interface, and loci were amplified and genotyped in 96 recombinant progeny (as judged by flanking framework markers), using 1% agarose gels (w/v) as described [173] Microarray Analysis RNA was isolated [14 8] from one leaf immediately apical and basal to the phenotyped leaf in one biological replicate of 183 in dividuals. RNA was converted to double -stranded cDNA, labeled, and hybridized to a customized NimbleGen microarray platform. The RNA manipulations and microarray design have been previously described [17 3] Resulting signal data was quantile normalized and l og2 transformed [173] prior to being used as input for expression QTL analysis. QTL and eQTL A nalysis Leaf width trait QTL were initially identified using the LSM estimates for phenotypic measurements (described above) in QTL Cartographer v.4.0 [199] using composite interval mapping [66, 67] with a st andard threshold of the 95th percentile of 1000 permutations Subsequently, gene expression QTL were identified, measured for significance, and classified as cis or trans acting as described in Chapter 3 To determine the relationship between gene expression and phenotype, phenotypic values from the same biological replicate measured for gene expression were correlated with the normalized gene expression values using the m ultivariate Pearson correlation fun ction of JMP 7.0 (SAS Institute, Cary, NC) and significance measured by the associated t -statistic corrected for multiple testing as described in the Results Finally, trait QTL position was confirmed for the single biological replicate on which gene expre ssion was measured, using the composite interval mapping and permutation approach described above.

PAGE 110

110 ARF1 cDNA Cloning and S equencing Full length ARF1 cDNA sequence was isolated from Nisqually 1, clone 52 -225 and clone D 124 leaf cDNAs by PCR amplification. Briefly, RNA was extracted from intact, fully unfurled leaves of each genotype by a standard laboratory protocol [148] RNA was DN Ase treated following manufacturers protocol (Invitrogen USA, Carlsbad, CA) and purified in an RNeasy purification column (Qiagen USA, Valencia, CA). First -strand cDNA was produced from 2g purified RNA using: 500ng oligodT (Promega USA, Madison, WI), 100ng random hexamer (Promega), 1L dNTP mix (10mM), 4L first -strand synthesis buffer, 1L M MLV reverse transcriptase (Promega), and 1L RNAsin (Promega) in a total volume of 20L. cDNA was synthesized for 2hr at 37C followed by 15min at 70C to terminate s ynthesis. PCR was conducted using 5L cDNA template, 5L each of ARF1 forward and reverse primer (10mM, Table 4 2 ), 5L dNTP mix (10mM), 5L Advantage2 polymerase buffer (10x, Clontech Laboratories Inc., Mountain View, CA), 1L Advantage2 polymerase mix (C lontech), .65L DMSO, and 13.5L betaine (6.5M) in a total volume of 50L. The thermocyle utilized comprised initial denaturing at 93C for 5min, followed by denaturing at 93 C for 30sec, touchdown annealing for 30sec from 50 C 44 C (one cycle each), and ex tension at 68 C for 2min 30sec. Subsequently, 20 cycles of denaturing, annealing at 44 C and extension as above were utilized. A final extension of 68 C for 30min completed the thermocycle program. A secondary PCR was completed as above, replacing the cDNA template with an equivalent volume of primary PCR diluted in a 1:1 0 00 ratio. The gene -specific band resulting from the secondary PCR was isolated in a 1.2% w/v agarose gel and purified using the ZymoClean gel purification system (Zymo Research, Orange, CA ) as directed by manufacturers protocol. Purified PCR product was directionally cloned into the pENTR/d TOPO vector (Invitrogen) and transformed into TOP 10 competent cells (Invitrogen) per manufacturers protocol s Selected colonies were grown overnight in liquid

PAGE 111

111 culture and prepared for sequencing from the M13 forward and reverse promoters using the Qiaprep Miniprep Spin Kit (Qiagen). Resulting sequences were reassembled and analyzed using Sequencher (Gene Codes Corporation, Ann Arbor, MI) software. ARF1 Promoter Isolation and S equencing To generate promoter sequence information, Genome Walker (Clontech ) libraries were constructed against the DNA of the parent trees of Family 52 124 according to manufacturers protocol, using Eco RV, Pvu II, Sma I, and Sca I restriction endonucleases. Sequences were amplified according to manufacturers protocol with the primary reaction consisting of 7 cycles of denaturing at 94 C for 25sec and annealing/extension at 70 C for 3min, 32 cycles of denaturing at 94 C for 25sec an d annealing/extension at 65 C for 3min, and a fi nal extension at 65 C for 7min. The secondary reaction consisted of 5 cycles of denaturing at 94 C for 25sec and annealing/extension at 72 C for 4min, 20 cycles of denaturing at 94 C for 25sec and annealing/e xtension at 67 C for 4min, with a final extension at 67 C for 4min. Primers utilized in the genome walker steps can be found in Table 4 3. Products were visualized in 1% agarose gels stained with ethidium bromide, and resulting bands were purified using Zy moclean Gel DNA Recovery Ki t (Zymo Research) following the manufacturers protocol Gel -purified PCR products were cloned into the pGEM T vector (Promega) and transformed into E. coli competent cells (Invitrogen), purified, and sequenced from the SP6/ T7 promoters as described above. ARF1 M utagenesis The full -length ARF1 cDNA clone was subjected to site directed mutagenesis using the Stratagene QuikChange site-directed mutagenesis kit, in order to introduce amino acid substitutions known to inhibit the exchange of GTP and/or GDP by ARF interacting proteins [200, 201] 50ng pENTR/dTOPO entry vector containing full -length ARF1 was PCR amplified

PAGE 112

112 with mutagenic primers (Table 4 2) as directed by the manufacturers protocol, using 15 cycles of primer extension at 68 C for 3min 30sec. Resulting mutagenized sequences were treated with Dpn I restriction endonuclease to degrade methylated template plasmid, and reactions were subsequently transformed into E. coli XL 1 blue cells as directed. Putatively mutagenized clones were verified by sequencing from the vectors M13 forward and reverse promoters and were subsequently recombined and replicated in the pZKY 1 overexpression binary vector using Invitrogen LR clona se enzyme mix and Invitrogen TOP 10 competent cells, as directed by manufacturers protocols. Nisqually Leaf Disc Expression Experiment To determine whether ARF1 expression varied across the leaf or plant within a uniform genetic background, we analyzed t he quantitative expression of ARF1 in Nisqually 1 within specific regions of several leaves at various stages of expansion. Five clones of Nisqually1 were grown as described [75] Whole, intact leaves were harvested and flash frozen in liquid nitrogen. After l yophilization, leaf discs (~25mg dry weight) were extracted from eight predetermined positions (pooled across the leaf blade midvein) in the first, third, sixth, and ninth unfurled leaves from three of the five clones. RNA was extracted from each disc usin g a cetyltrimethylammonium bromide/chloroform extraction [148] followed by purification of the aqueous phase in a Qiagen RNeasy c olumn (Qiagen USA) as directed by manufacturers protocol. Purified RNA (500ng) was converted to first strand cDNA using Promega M MLV reverse transcriptase, random hexamers, and oligodT as described above. Each resulting sample was analyzed for expressio n of ARF1 by real time PCR (primers in Table 4 2) using .5uL of first strand cDNA template, 10x Brilliant SYBR Green qPCR Master Mix (Stratagene), and Mx3000P thermo cycler (Stratagene) as directed by manufacturers protocol, in a total volume of 10uL.

PAGE 113

113 Simi larly, expression was determined for the control genes PtACT2 PtUBQ and Pt18S for each sample. Resultant data were analyzed by an efficiency -corrected relative quantification method as described [202] normalized to the geometric mean of the three internal control genes [203] Verification of Allele Specific Expres sion Effects in Segregating Population Subset To verify the effect of allele combinations at the ARF1 locus in ARF1 expression values, we grew a subset of 60 progeny (30 each heterozygous and homozygous for P. deltoides ARF1 allele, as judged by flanking S SR genotypes) from the segregating population under conditions identical to those of the original study [75] We harvested the fourth and fifth fully unfurled leaves for ARF1 expression analysis. RNA was extracted by the standard protocol [148] converted to cDNA using M -MLV, random primers, and oligo -dT, and subjected to qPCR analysis as described for the leaf disc experiment above. Results Identification of a Major QTL for Leaf Blade Width We analyzed an interspecific pseudo backcross pedigree (Family 52 124) for variation in leaf lamina shape (measured by lamina length, width, and their ratio) segregating between the narrow -leaf donor parent species P. tr ichocarpa and the broad leaf recurrent parent P. deltoides (Figure 4 1) by composite interval QTL analysis and a previously established genetic map [75] One to five significant QTL were identified for the various traits (Figure 4 2, Table 4 4) including a major locus on linkage group (LG) X that regulated both lamina width (phenotypic variation explained [PVE] = 6.0%) and length:width ratio (PVE = 14.2%; Figure 42). The major QTL interval, constrained by sequence -linked SSR markers PMGC_2855 and G CPM _2122, encompassed 625 genes and spanned ~3.5Mb of uninterrupted sequence in the genome sequence of P. trichocarpa [4]. To reduce the number of potential candidate genes in the trait QTL (tQTL) interval, mapping resolution was increased by genotyping recombinant individuals (n=96) for

PAGE 114

114 seven add itional SSR markers identified from the genome sequence ( Table 4 1 ). A dditional mapping decreased the QTL interval to ~3.0Mb and 450 candidate genes (Figure 4 3). Due to the high proportion of phenotypic variance explained and quality of the genome assembl y in this tQTL region, we elected to pursue the major locus on LG X for further characterization and QTL cloning. Gene Expression A nalysis of Leaf T issue Identifies PtARF1 as a Candidate Gene for Lamina Shape As additional recombinant progeny could not be identified to increase map resolution across the target interval, we elected to move forward with an integrative genomics approach to further reduce the pool of candidate genes. Because of the prominent role of transcriptional diversity in evolutionary dis tinction of species [36] we hypothesized that interspecific differences in expression of a key regula tor of leaf expansion or cell division may explain the leaf lamina variation observed in the progeny of Family 52 124. To evaluate this hypothesis, we measured genome -wide gene expression in expanding leaf tissue from one biological replicate of 183 segreg ants and mapped the resulting transcript abundance as eQTL. Our genome -wide analysis identified 13,403 statistically significant eQTL representing 12,392 unique gene models (described fully in Chapter 3). Since our hypothesis was that lamina shape variation was a product of differential gene regulation arising from interspecific polymorphism(s) within the tQTL interval, we narrowed our focus to only the genes with eQTL surpassing the significance threshold in this region. Our analysis identified 161 eQTL which we classified as cis (n =116) or trans ( n =45 ) acting, contingent on the physical position of the gene model in the genome assembly We classified all eQTL arising from gene models on unassembled genomic scaffolds ( n =19) as trans acting, since the Populus genome sequence assembly in the tQTL region was predicted to be contiguous [4].

PAGE 115

115 We expected that if differential transcript accumulation of key regulatory factor s account ed for phenotypic variation in leaf shape, abundance of such transcripts should exhibit a significant statistical correlation with leaf lamina shape. Furthermore, differential regulation of these transcripts should be accounted for by the same region regulating variation in the phenotypes i.e., they should demonstrate eQTL controll ed by the tQTL interval. Thus, we utilized a standard multivariate correlation to determine the relationship between the leaf lamina phenotypes for each transcript with an eQTL in the interval of interest Among the 161 transcripts with eQTL co -localized t o the trait interval, we identified only two with statistically significant (Bonferroni = .05, P < 3.106x 104) correlation s to lamina width (Table 4 5 ). No genes were correlated with lamina length:width ratio at this level of statistical significance. Both significantly correlated genes were regulated in cis by the trait QTL region one gene encoded the poplar homolog of the Arabidopsis salt -inducable protein AtCP1 [204] while the other encoded an ADP ribosylation factor (ARF) GT Pase which we designated PtARF1 We also analyzed the relationship of each gene expressed above array background in leaf tissue to both lamina shape phenotypes, by correlating each phenotypic trait to the gene expression measurements and ranking these corr elations by their significance. Among 21,810 genes detected above background on the microarrays, 257 were significantly correlated with leaf width (FDR < .05, P < .00061). No genes correlated with lamina length:width ratio at an analogous level of signific ance. Among the 257 genes identified, only three underlie the trait QTL interval regulating leaf width (Table 4 6 ). Yet, among these three genes there was a clear discrepancy just one gene, ARF1 was strongly related to to the leaf width phenotype while being regulated in cis by the lamina phenotypic QTL interval. Conversely, the other two genes were not differentially regulated by the QTL interval despite their correlation with the phenotype

PAGE 116

116 (Table 4 6) in fact neither gene had a significant eQTL anywhere in the genome Furthermore, it is notable that AtCP1 ortholog regulated in cis by the QTL interval and significantly associated with the leaf width trait by the previous analysis was not significantl y associated when considered on a transcriptome -wide scale. It might be expected that among a large gene family such as the ARF type GTPases, compensatory expression by another ARF gene family member could occur in the P. deltoides background to modulate t he effect of decreased ARF1 expression. To determine whether additional members of the ARF gene family might respond in their expression to the varying alleles at ARF1 we analyzed the correlation of expression of all predicted ARF type and ARF like genes (n = 25) to ARF1 We also calculated the level of expression for each ARF, relative to the expression detected for ARF1 to determine whether a true compensatory effect could be expected. The expression of ARF1 is positively correlated to most other ARFs, rather than negatively correlated as would be expected for a compensatory expression effect ( Table 4 7). Furthermore, the two negatively correlated genes were detected in low relative abundance compared to ARF1 and many of the other ARF family genes (Table 4 7) in fact, both were below background for the microarrays as judged by the 97.5th percentile of the negative control probes [173] Finally, no significant cis -eQTL was detected for ARF1 in either root or xylem tissues that were also assayed for transcript abundance (data not shown), suggesting ARF1 may be expressed or modulated specifically in leaf tissue. Taken together, the entire body of statistical evidence, coupled with the robust tQTL/eQTL overlap ( Figure 4 3 ) and functional annotations, led us to pursue ARF1 for further characterization as a candidate gene for quantitative variation in leaf shape within our segregating population.

PAGE 117

117 Isolation of Interspecific ARF1 Coding and Promoter Polymorphisms To identify polymorphisms in ARF1 among the Family 52 124 parent trees, we utilized PCR to isolate the cDNA coding sequence corresponding to the ARF1 gene. Sequence analysis revealed minimal polymorphism between haplotypes, including only three synonymous nucleotide substitutions and no insertions or deletions that would result in translational frame sh ifts (not shown). The predicted function of ARF1 in the core vesicle trafficking machinery, indicates that the sequence is likely to be strongly conserved [205] so these results were not surprisin g. We speculated that differential expression of ARF1 would most likely be explained by interspecific polymorphisms in the 5 regulatory region adjacent to the gene. To explore this possibility, we utilized 4 steps of Genome Walker PCR to sequentially isol ate sequence 5 to the start codon of ARF1 from each parent tree. Haplotypes arising from each parent were discriminated on the basis of first-exon synonymous substitutions initially detected in the fulllength cDNA clones above. Extensive sequence heterog eneity between the parent trees made it challenging to readily amplify and isolate clones derived from the target sequence as such, only 500bp could be readily isolated from all three P. deltoides haplotypes of the parent trees, whereas >1.2Kb was isolat ed from the P. trichocarpa haplotype of the hybrid tree. Characterization of polymorphism between the haplotypes revealed extensive interspecific differences between the haplotypes in this limited region, including at least seven small insertions/deletions and 18 SNP ( Figure 4 4 ). Interestingly, however, the sequence from haplotypes within each species was highly conserved: no in/dels were identified in either species, and only two SNP w ere identified for P. trichocarpa (relative to the Nisqually 1 genome s equence) within the 500bp region of comparison. Sequence corresponding to the P. deltoides haplotypes was fully conserved relative to one another.

PAGE 118

118 Localized E xpression of ARF1 in E xpanding Leaves To determine the expression pattern of ARF1 throughout varying stages of leaf expansion, we conducted an experiment to sample localized regions of expanding leaves from P. trichocarpa genotype Nisqually1. Five cuttings of Nisqually1 were grown for 12 weeks under conditions mimicking those of the original experiment [75] From four different expanding and intact leaves of three clones, we extracted small leaf disc samples (~50mg) from eight distinct regions of the leaf blade ( Figure 4 5A ) to measure ARF1 expression by quantitative real time PCR (qRT PCR). ARF1 expression, when normalized to the geometric mean of three internal control genes, was surprisingly uniform across all sampled positions ( Figure 4 5B) within a leaf. Similarly, across the four leaves sampled, expression was relatively consistent wi thin sample positions. Slight variation was noted within some sampled positions, corresponding to generally increased expression of ARF1 near the midvein with increasing maturity ( Figure 4 5B, positions 3, 4, 6, 8). However, error variance associated with the measurements and the small sample size of the experiment minimizes the number of differences that could be deemed significant. This preliminary analysis suggests that ARF1 expression is likely to be relatively homogenous throughout the leaf lamina and that cellular position within the lamina, or leaf expansion stage, does little to alter the expression of ARF1 in the P. trichocarpa genetic background. Verification of Allele Specific Expression Effects for ARF1 The presence of a cis -eQTL for ARF1 in the segregating population dictates that expression variability for the gene exists in the population and associates with the ARF1 locus. Lower expression of ARF1 is associated with a P. deltoides -like leaf shape. Given the minimal variation noted in expressio n across the leaf blade in our previous experiment, we speculated that the low microarray signal obtained for ARF1 in progeny homozygous for the P. deltoides alleles

PAGE 119

119 might be accounted for by an overall lack of ARF1 expression in the P. deltoides background. To test this hypothesis, we propagated 60 progeny genotypes of Family 52 124, including 30 genotypes each that are either homozygous or heterozygous for P. deltoides ARF1 alleles (as judged by flanking microsatellite locus genotypes) and measured ARF1 expression in expanding leaves by qRT PCR. The results of this experiment indicated that the expected allelic effect on ARF1 expression due to substitution of a P. trichocarpa allele for a P. deltoides allele could be recapitulated in the segregat ing subpopulation using qRT -PCR technology ( Figure 4 6 ). However, significant differences were not observed between allelic classes in the context of this experiment ( P < .253, measured by Students t -test with unequal variance) and expression was readily detected from both genetic backgrounds. Furthermore, while variation in ARF1 expression at the original LG X eQTL interval was observed, this variation was not sufficient to surpass a standard significance threshold of LOD > 2.5 (LODobserved ot shown). The outcome of this experiment was likely affected by the small subpopulation size, coupled with the measurement of only one biological replicate of progeny. Nonetheless, the observation that allelic effects can be readily captured on the basis of qRT -PCR is an important outcome for future testing of the effects of ARF1 expression in independent verification populations. Furthermore, an overall lack of ARF1 transcript within P. deltoides is not the explanatory factor in leaf phenotypic variation. Discussion A major goal of plant biology is to identify genes implicated in adaptation and evolution, as they may provide the genetic tools to develop more productive genotypes that are capable of withstanding varying sources of biotic and abiotic stress. Many of these adaptive traits are complex in nature and thus are suitable to dissect using a combination of traditional quantitative genetic approaches such as QTL analysis and gene expression phenotyping in segregating

PAGE 120

120 populations [78, 89, 195] Here, we considered leaf lamina shape among an interspecific hybrid family of Populus in a genetical genomics context [78] From a series of QTL for leaf lamina shape characters, we identified a major QTL implicated in lamina morphology pleiotropically regulating both lamina width and lamina length:width ratio. Utilizing whole -genome microarrays for expression analysis o f leaf tissue, we identified a group of genes regulated in cis by the phenotypic QTL interval, and/or whose expression was statistical l y correlated with both leaf width and length:width ratio Considering the statistical relationships, genetic regulation of expression, and functional annotations, we selected an A DP ribosylation factor GTPase ARF1 as the primary candidate gene governi ng lamina shape characters in the interspecific hybrid pedigree. Both our microarray and real time PCR analyses clearly indicate that ARF1 is differentially regulated in cis by the phenotypic QTL region. It is also clear that among all genes regulated by this region, ARF1 has the strongest statistical relationship with leaf lamina width. Furthermore, considering all genes in the genome, the correlation of ARF1 is among the strongest to the phenotype regardless of whether any genetic regulation of the trans cript is considered. From the sequencing data produced for ARF1 from P. trichocarpa and P. deltoides it is unlikely that transcript stability due to mRNA sequence characteristics profoundly affects the abundance of ARF1 in vivo Rather, our promoter seque ncing data (albeit limited) suggests that cis regulatory polymorphisms in alleles of ARF1 leads to differential transcript accumulation observed in our experiments. While further promoter cloning and reporter gene testing are clearly necessary to validate this hypothesis, it is interesting that haplotype structures of the promoter region cloned thusfar were highly conserved within each species. This observation might indicate that the regulation of ARF1 while different between species, is important for pla nt or leaf development

PAGE 121

121 in Populus However, the sample size of ARF1 promoter haplotypes is in sufficient to draw meaningful conclusions about specific polymorphisms that might direct ARF1 expression. Sequencing ARF1 promoter regions from a larger compendium of P. trichocarpa and P. deltoides individuals will shed additional light on the regulatory mechanism and cis elements important for ARF1 regulation. Previous evidence [206, 207] suggests that vital cis -elements can be located at long distances from the gene in question thus the availability of resequenced poplar genomes from a collection of unrelated individuals could provide a key piece of data in the search for specific sequencees regulating this and other genes of interest. Homogene ous expression of ARF1 throughout the leaf lamina and at different stages of leaf expansion in P. trichocarpa indicates that ARF1 is likely not regulated in a regionally specific manner during leaf development or expansion. It is possible that expression is regionally controlled only in P. deltoides in response to presence or absence of a specific cis -element. Similarly, it is possible that expression could be restricted to a specific tissue layer or type within the leaf lamina in one or both species. The sampling method we utilized to measure regional expression of ARF1 would not detect tissue -specific differences, because sections were taken encompassing all l ayers of the leaf lamina (mesophyll, epidermis, and vasculature) in proportions in similar to which they would occur in an intact leaf. Finally, a more significant and functionally important ARF1 expression difference might be noted among newly forming lea ves within the apical meristem. Additional studies employing more sophisticated sampling methods, (i.e., laser capture microdissection) would be required to reconcile these possible scenarios of ARF1 regulation. Evidence recently generated in the Arabidops is system suggests the underpinnings of a hypothetical mechanism by which differential expression of ARF1 in the two Populus genetic

PAGE 122

122 backgrounds could lead to variation in leaf lamina shape. The action of AtARF1 (the ortholog of PtARF1 ) has been shown to modul ate the kinetics of endocytosis, PIN FORMED 2 (PIN2) localization, and cell polarity in Arabidopsis roots [198] Several PIN auxin efflux carrier proteins have also been shown to be post -translationally localized to apical and basal membranes of Arabidopsis epidermal cells in an endocytic process mediated in part by two Rab5 GTPase homologs [208] The c onnection between ce ll polarity, auxin, and directional growth has been intricately studied in several model plant systems [209, 210] Auxin has already been implicated in initial leaf formation [211, 212] lamina margin elaboration [188, 189] leaf vasculature patterning [213] Thus, it is likely that leaf expansion is also directly affected by auxin flux. In Populus our evidence supports a model whereby differential expres sion of PtARF1 in P. trichocarpa and P. deltoides changes the dynamics of endocytosis -mediated PIN localization in leaf cells. Increased abundance of ARF1 transcript, and hence ARF1 protein in P. trichocarpa speeds the process of PIN polarization through the endocytic pathway during development and expansion More rapid acquisition and maintenance of PIN polarization decreases lateral auxin flux and, subsequently, lamina expansion in the leaf width direction (Figure 4 7A) Conversely, in P. deltoides lower abundance of ARF1 transcript leads to a more limiting supply of ARF1 protein, which decreases the relative rate of e ndocytosis and hence, PIN polarization Slower PIN polarization throughout development increases lateral auxin flux and expansion of the lamina in the leaf width direction (Figure 4 7B). An important experiment to verify this model will be to determine the rate of endocytosis -mediated PIN localization in each species. A template for this experiment, providing a quantitative cell polarity index of fluorescently tagged PIN protein, has recently been published [208] and will provide a powerful framework in which to test the hypothesis

PAGE 123

123 Three other genes significantly associate with the lamina shape phenotypes through the two methods of statistical assoc iation used in our analysis. These genes include: (1) the poplar ortholog of AtCP1 a calcium binding calmodulinrelated protein inducible by salt stress [204] ; (2) a chaperonin protein (3) and a gene encoding a calmodulin calcium binding protein. While a role for these genes in leaf development is not implausible, the evidence developed thusfar strongly favors ARF1 as the causal regulator of lamina shape differences in the Populus segregating population Furthermore, the known function of ARF1 places it directly in welldescribed pathways affecting cell expansion and leaf development, whereas to our knowledge, no significant evidence has been developed for the other genes in leaf shape or development Collectively, our results provide another piece of compelling evidence for the role of transcriptional diversity in shaping variations in plant form. Similarly, we provide yet another way in which the key plant hormone auxin could shape diversity in plant morphology. We have demonstrated that natural variation for auxin response could play a key role in plant diversity we observe in nature, and our analysis suggests that natural variation in hormone response pathways clearly warrants additional investigation as we strive to learn more about the evolution of morpholog ical differences among plant taxa.

PAGE 124

124 Figure 4 1. Leaf lamina shape variation among parents and progeny of the P. trichocarpa X P. deltoides pseudobackcross pedigree Family 52124. The donor parent, P. trichocarpa, has a lamina length/width ratio of ~3.0, while the recurrent parent, P. deltoides has a lamina ratio of ~1.0. The trait exhibits additive variation, as the hybrid parent has a ratio of ~1.5. Segregating BC1 progeny span the spectrum of phenotypic varation from the hybrid to the recurrent parent with limited transgressive segregation observed.

PAGE 125

125 Figure 4 2. Genome -wide composite interval mapping scan for leaf lamina shape characters in Family 52 124. Least -square means of triplicate measurements from 396 progeny were utilized as the genotypic value for mapping against an SSR and SFP -based single tree genetic map produced for the hybrid parent, genotype 52225.

PAGE 126

126 Figure 4 3. Fine -scale mapping of the major lamina width QTL on LG X in the segregating pedigree. Framework SSR loci were genotyped in 396 progeny (G_ and P_ loci) while additional SSR underlying the QTL were genotyped in 96 recombinant progeny. Recombinants were identified by maternally inherited marker genotypes at locus P_2855 and G_2122. The approximate genomic location of the ARF1 gene is denoted by the downward facing arrow.

PAGE 127

127 Figure 4 4. Sequence alignment for ARF1 upstream regions isolated from P. trichocarpa and P. deltoides Sequence was isolated by sequential steps of Genome Walker PCR as described in Materials and Methods. SNPs detected within the P. trichocarpa haplotype, relative to comparison to Nisqually 1, are depicted by N. Polymorphisms between haplotypes are highlighted in yellow.

PAGE 128

128 F igure 4 5. Quantitative PCR analysis of regional ARF1 expression in differently staged expanding leaves of P. trichocarpa. (A) 25mg lyophilized leaf discs were harvested from the leaf regions depicted. Samples of the same number were pooled across the midvein for subsequent RNA extraction and analysis. The first, third, sixth, and ninth fully unfurled leaves were sampled in this manner for comparison of the effect of leaf maturity. (B) Results of qRT -PCR analysis. Bars represent mean and standard error of three measurements of PtARF1 expression normalized to the geometric mean of three internal control genes. Lack of an error bar indicates that sample was not triplicated due to experimental failure at one or more stages.

PAGE 129

129 Figure 4 6. Quantitative RT -PCR verification of genetic effects on ARF1 expression in the segregating population. Sixty progeny genotypes were grown in a randomized design, including thirty genotypes each homozygous (D) or heterozygous (H) for P. deltoides alleles at SSRs flanking the ARF1 locus. Mean and standard error for each genotypic class (Students t test P < .253) are shown.

PAGE 130

130 Figure 4 7. A hypothesized molecular mechanism for how differential ARF1 expression affects leaf lamina phenotypic variation in P. trichocarpa and P. deltoides PIN protein flux and localization in the leaf cell is depicted in orange thoughout the figure. Figure modified from the model of PIN localization presented by Dhonukshe et al. [208] (A) In P. deltoides newly synthesize d PIN, secreted to the plasma membrane in a non -polar pattern, is endocytically localized to apical/basal membranes slowly due to decreased ARF1 abundance. This results in increased auxin flux from the lateral cell membranes and increased expansion/growth in the lamina width direction. (B) In P. trichocarpa, newly synthesized PIN is also secreted in a non -polar manner. Higher ARF1 abundance increases the rate of endocytic apical/basal PIN localization, minimizing lateral auxin flux and polarizing cell growt h and expansion preferentially in the lamina length direction.

PAGE 131

131 Table 4 1. Primers utilized to amplify microsatellite loci across the major linkage group X QTL for leaf lamina characters in Family 52124. Locus Primer forward s equence Primer r everse s equence Motif Repeats Linkage g roup X p hysical l ocation (bp) PMGC_2855 GGTATCTTGTTATCCACTGCC TTTTCCTCGTTAATTAGAGTCG GA n/a 11, 122, 293 LG_X01 GCCACCAATCCAGCAAGTAA AATGGAGGTGTGGCAGTAGC TA 9 11, 872, 486 LG_X06 GAGAAGCAGCAATGCAGGAT ATGCAAACTGGTCCGGATAC CTT 6 12, 645, 373 LG_X03 ACCTGGTCCATTTGTTGAGC TGCAGGCAATCTCAAACTCA GA 5 13 378 326 LG_X04 GGGATGGCAAAATACGTTCA CATCATCGTACAACCTCACCTT TA 5 14, 126, 639 ARF1US3 GGCAAGGACAGCGCATGATG TTTCACCGAACTTCCACACTTT n/a n/a 14, 365, 406 ARF1DS2 GTGATAGCAGAGAGCCGAAA GTAAGTATGAGAGGAATGAGGGG n/a n/a 14, 369, 387 LG_X05 TGGTAGATTGCGAGCTGAGA AGGGCTTTCCGGCTATTAAA CT 5 15 159 499 GCPM_2122 TCAGCAACTATCACCATGAA GGAATGTGCAGCATATACAA GT 12 15, 661, 771

PAGE 132


PAGE 133


PAGE 134

134 Table 4 4. Summary of QTL detected for leaf lamina shape characters in P. trichocarpa X P. deltoides X P. deltoides Family 52124. Trait Clonal repeatability (Std e rror) QTL number Linkage group Flanking m arker 1 Flanking m arker 2 Origin of positive allele LOD peak Phenotypic v ariance e xplained Lamina Length .2186 (.0244) 1 IV G1809 G3847 P. deltoides 3.14 6.31% Lamina Width .2407 (.0249) 1 VI P2221 W12 P. trichocarpa 3.69 5.12% 2 X P2855 G2122 P. deltoides 4.59 5.99% Lamina Length:Width Ratio .3490 (.0276) 1 I G124 G2903 P. deltoides 4.51 6.16% 2 X P2855 G2122 P. trichocarpa 12.77 14.20% 3 XV G1245 G1424 P. deltoides 5.73 5.12% 4 XVII G3580 G641 P. trichocarpa 3.45 5.00% 5 XIX O597 G2319 P. deltoides 3.69 4.00%

PAGE 135

135 Table 4 5. Genes with significant phenotypic correlations to leaf lamina and eQTL regulated by the major lamina shape QTL locus. Significance was judged by a Bonferroni correction for 161 tests of expression -phenotypic correlation. P. trichocarpa g ene m odel Cis/ t rans eQTL Correlation l amina w idth Correlation s ignificance l amina w idth ( P value) Ath ortholog Ath o rtholog a nnotation estExt_Genewise1_v1.C_LG_X0744 cis 0.3669 0.0001 AT1G10630 ATARFA1F; GTP binding / phospholipase activator/ protein binding estExt_fgenesh4_pg.C_LG_X1324 cis 0.2864 0.0032 AT5G49480 ATCP1 (CA2+ BINDING PROTEIN 1); calcium ion binding

PAGE 136

136 Table 4 6. Gene models expressed above microarray background and encoded within the lamina shape QTL interval with statistically significant correlation s to lamina width Significance was judged at a modified FDR threshold P < 6.1E 04. P. trichocarpa g ene m odel Correlation lamina w idth Correlation P value Correlation rank l amina w idth Cis/t rans Ath ortholog Ath o rtholog a nnotation estExt_Genewise1_v1.C_LG_X0744 0.3668554 7.4854E 05 49 cis AT1G10630 ATARFA1F; GTP binding / phospholipase activator/ protein binding fgenesh4_pm.C_LG_X000536 0.35563947 1.2789E 04 69 n/a AT1G26230 chaperonin grail3.0006040101 0.3487108 1.7631E 04 90 n/a AT3G25600 calmodulin

PAGE 137

137 Table 4 7. Expression characteristics of ARF type gene family members relative to ARF1 P. trichocarpa g ene m odel Expression c orrelation to ARF1 Correlation significance ( P value) Relative a bundancea (m edian) Expressed a bove microa rray b ackground? estExt_Genewise1_v1.C_LG_I4117 0.39521569 1.764E 05 0.63332863 Y estExt_Genewise1_v1.C_LG_V4094 0.33574725 3.153E 04 0.71948814 Y estExt_Genewise1_v1.C_LG_XIX1392 0.28014634 2.901E 03 0.35397767 N estExt_Genewise1_v1.C_LG_XVIII2916 0.2098178 2.709E 02 0.47177235 N estExt_fgenesh4_kg.C_LG_XII0031 0.43947159 1.396E 06 0.61256805 Y estExt_fgenesh4_kg.C_LG_XV0033 0.43443381 1.898E 06 0.35373527 N estExt_fgenesh4_pg.C_LG_IX0748 0.37086319 6.152E 05 0.74631113 Y estExt_fgenesh4_pg.C_LG_XVIII0541 0.26424056 5.072E 03 0.56174961 N estExt_fgenesh4_pm.C_LG_VI0838 0.27363038 3.662E 03 0.44748339 N estExt_fgenesh4_pm.C_LG_VII0412 0.431432 2.274E 06 1.3114264 Y estExt_fgenesh4_pm.C_LG_VIII0406 0.30101811 1.326E 03 0.44116017 N eugene3.00130048 0.51562433 6.988E 09 9.64183283 Y eugene3.00131058 0.40448725 1.068E 05 0.9141763 Y grail3.0016004802 0.2856963 2.369E 03 7.16681348 Y grail3.0050012801 0.45146446 6.585E 07 4.38820514 Y grail3.0061009602 0.45939209 3.944E 07 3.35093299 Y gw1.I.3910.1 0.1974934 3.774E 02 0.2810372 N gw1.II.442.1 0.42146597 4.092E 06 0.99980107 Y gw1.IV.3637.1 0.32899233 4.225E 04 0.95721298 Y gw1.IX.2325.1 0.04655942 6.275E 01 0.28198022 N estExt_Genewise1_v1.C_1310224 0.34693432 1.912E 04 0.29649079 N estExt_fgenesh4_pm.C_1710005 0.02297545 8.108E 01 0.41922534 N eugene3.12450001 0.4747608 1.406E 07 0.33204579 N grail3.0131005102 0.1257081 1.886E 01 0.30736057 N gw1.3608.1.1 0.48736645 5.806E 08 0.3344929 N a. Relative abundance is the proportion of signal detected relative to ARF1

PAGE 138

138 CHAPTER 5 CONCLUSIONS The overall goal of this project was to determine the role of genetic variation in shaping gene expression diversity in genus Populus and to extend these findings to the level of morphological phenotypes to understand how transcriptional diversity affects evolutionarily conserved variation for leaf shape between the Tacamahaca and Aigeiros sections of the genus. An important component of understanding the role of genetic variation on gene expression and phenotypes is the ability to accurately and densely map the genome of the species in question. In Chapter 2 of this study, I reported the first utilization of sequence anchored, microarr ay -based markers for high density genetic map construction in an outcrossing species. The genetic map produced by this effort was important for the subsequent analysis of trait and expression QTL, while also resulting in the placement of nearly 50% of the whole -genome shotgun scaffold sequence that previously had no known location in the genome of P. trichocarpa [4]. This outcome paints a favorable picture for further utilization of high density genetic mapping to aid in draft assemblies of whole genome shotgun sequences, even from highly heterozygous spe cies. Similarly, a preliminary microarray -based genetic map for Eucalyptus comprising nearly 10,000 loci, has been produced (L.G. Neves and M. Kirst, personal communication) and is likely to play an important role in the genome assembly currently in progress by the Department of Energy [214] New approaches, includi ng microarray based [215] and in -solution sequence capture platforms [216] coupled to high throughput sequencing [217, 218] will continue to refine genotyping in forest tree species and help to provide even more saturated SNP -based genetic maps for QTL and association -based experiments.

PAGE 139

139 From genetic map data and whole genome expression phenotypes, a preliminary goal is to determine the overall ontogeny of the genetic regulation over gen e expression. Early studies in model species (both plants and other eukaryotes) characterized cis and trans acting regulation on a genome wide scale and determ ined heritability and quantitative genetic effects in gene expression phenotypes [62, 84, 91] Subsequently, it was found that eQTL studies may be useful to identify gene coexpression networks on account of key network regulators that segregate in these populations [183] In Chapter 3 of this study, I utilized the high density genetic map together with whole genome eQTL data to characterize the mechanism of tissue -specific genetic regulation of mRNA abundance and coexpression networks I n addition to generating a wealth of information regarding tissue -specific regulation of expression among individual loci and networked genes, two important and overarching conclusions could be reached. First, it is apparent that major loci controlling the abundance or turnover of large suites of transcripts are common in the Populus genetic system. Selective forces shaping the evolution of the modern poplar genome [4] may have driven these loci to accumulate on specific chromosomes ( Figure s 3 1 & 3 2 ). Secondly, major loci often control transcript level s in a tissue -specific manner, and common networks of genes that are regulated by a given locus in one tiss ue frequently remain co -regulate d in other tissues, albeit by different major loci In addition to the ability to describe and define the ontogeny of genome -wide expression, eQTL studies, when allied to quantitative phenotypic data gathered from the same population, can facilitate the a posteriori identification of candidate genes regulating complex phenotypes Co localization of regulation of expression and phenotypic trait s together with a significant statistical relationship s between phenotypic and gene expression measurements, can define a list of candidate genes for further investigation [78] Frequently this list is orders of magnitude

PAGE 140

140 smaller than a list that considers all genes in a tQTL interval to be equally likely as the phenotypic regulator [83, 89] Chapter 4 describes preliminary evidence implicating an ADP rybosylation factor GTPase, ARF1 as a candidate gene underlying a QTL for leaf width in the pseudobackcross pedigree. Among all genes whose regulation is controlled by the leaf width QTL interval, ARF1 displays the most significant statistical relationship with the trait. Furthermo re, allele -specific effects of gene expression are prominently displayed in qRT PCR experiments, and the analysis identified a number of candidate cis regulatory polymorphisms that may work independently or together to affect differential transcript abunda nce in P. trichocarpa and P. deltoides Previous data have implicated the Arabidopsis ARF1 ortholog in the endocytic pathway, where it influences the distribution of PIN auxin efflux carriers on the cell membrane surface. The predisposed role of ARF1 in this pathway supports a generalized model wherein higher expression of ARF1 from the P. trichocarpa background leads to stronger basal localization of auxin efflux carriers, driving cell expansion in the longitudinal plane of the leaf lamina In P. del toides weaker ARF1 expression leads to decreased basal localization of the auxin efflux machinery may result in increased lateral expansion of the leaf lamina. Clearly, a number of testable hypotheses remain unadd ressed in the scope of the current work. O f immediate importance is developing a more thorough understanding of the role that co expression networks play in phenotypic variation in Populus While it is apparent from the work outlined in Chapter 4 that individual genes can be implicated in vital ph enotypic differences, past precedent indicates that differential regulation of biochemical pathways or transcriptional and signaling cascades is similarly, if not more important for complex phenotypic variation [97, 219221] An immediate step towards attainment of this goal in Family 52 124 is the alliance of previously collected phenotypic data [75] with the co -expression networks identified in Chapter

PAGE 141

141 3. A preliminary connection between phenotypes and these networks can be established simply due to overlap of genomic inter vals regulating phenotypic QTL and network eQTL. These compar is ons indicate several cases of measured phenotypes exhibiting common genetic regulation with established a posteriori coexpression networks particularly for cell wall constituent and biomass phenotypes with xylem networks (data not shown) Transgenic modulation of candidate network regulators could verify the role of these networks in the phenotypes in question. Of course, it is likely that more complex coexpre ssion networks subject to mu ltiple points of genetic regulation will influence some if not most, of the measured traits. These would not have been c l early identified by the analysis described in Chapter 3, since it focused on known regions of pl ei otropic transcriptional regulation. One case of a more complex regulatory network has already been identified in Family 52 124, wherein a regulatory locus encoding a putative S adenosyl methionine synthase (SAMS) is co-expressed with several members of the lignin biosynthesis pathway. The eQ TL for the putative SAMS co localizes with a pl ei otropic locus regulating wood chemistry and biomass phenotypes and SAMS expression correlates strongly with each phenotype (E. Novaes, M. Kirst; personal communication). Thus, while the initial results descr ibed in Chapter 3 give a favorable starting point to understand the role of transcriptional networks in phenotypic diversity, additional analysis may frequently be required to understand the specific transcriptional mechanisms underlying variation in indiv idual phenotypic QTLs It will also be important to continue building an understanding of the role of genome evolution in shaping transcriptional diversity. It is clear from the results presented in Chapter 3 that the distribution of pl ei otropic expression regulators is non random and that specific chromosomes harbor these regulators more frequently than others. Equally obvious are the facts

PAGE 142

142 that sequence polymorphism s in the poplar genome are also distributed non-randomly across chromosomes (Table 1 2) an d certain chromosomes have been subjected to complex patterns of evolution in order to reach their modern structure [4] For instance, chromosome XVII was shown to harbor one of the highest normalized frequencies of eQTL for leaf, root, and xylem tissues (Figure 3 2). In the poplar genome assembly, chrom osome XVII is unique in that less than 60% of its predicted sequence has been contiguously assembled suggesting it harbors an unusually high rate of polymorphism ([4], Table 2 5 ). In agreement with this suggestion the chromosome possesses one of the highest frequencies of SNP per base pair in the Nisqu ally 1 genome (Table 1 2 ). Furthermore, chromosome XVII has a high frequency of scaffold sequence mapping to it ( Table 2 5 ) and is one chromosome predicted to possess a more complex evolutionary history [4]. Scaffold sequence, while generally less polymorphic for SNP substitutions is highly variable for insertion/deletion polymorphism both within and outside gene sequence relative to the genome as a whole ( Table 1 2 ). Thus, chromosome XVII and other similar chromosomes could provide a unique laboratory in which to study the role of chromosome structure and sequence features in shaping interspecific diversity of gene expression. Particularly powerful studies to address these questions could be developed based on the expression data contained within this study after the genome sequences of several P. tric hocarpa and P. deltoides clones become available. These sequences are currently being generated and are expected to reach the public realm within the next 2 3 years (G. Tuskan, personal communication). A shortcoming of the work presented in Chapter 4 is th e lack of conclusive functional verification of the role of ARF1 in PIN localization and leaf shape determination. Overexpression and RNAi constructs for ARF1 have been produced and plant transformation is currently

PAGE 143

143 underway. While the evidence presented h erein indicates that ARF1 should be specialized relative to other ARFs for its function in leaf development in Populus it is possible that radical perturbation of ARF1 expression may result in compensatory expression of another ARF gene in leaf tissue. Accordingly, mutant constructs that repress the function of ARF 1 interacting proteins may play a key role in the verification of function, as described recently in Arabidopsis [198] These constructs have been generated by a standard site -directed mutagenesis protocol and are also currently being used in plant transformation. Analysis of each o f these transgenic constructs should increase our understanding of whether ARF1 affects leaf shape in P. trichocarpa and P. deltoides and the mechanism by which the effect occurs. A final and overarching insight that can be reached from this study relates to the role of tissue -specific regulation in shaping phenotypic diversity. There is little doubt as to the importance of transcriptional diversity, both in the realm of evolution between species and in the evolution of variations in form within an organis m (i.e., tissues and organs). A commonly considered mechanism by which transcriptional diversity can effect its action is through expression of tissue organ or species -specific genes. However, we [35] and others [186] have noted that quite frequently, the contingent of genes expressed in a given species, organ, or tissue is not highly discriminatory of the phenotypic diversity between these levels. An alternate mechanism to drive diversification is the specialized regulation of genes or groups of genes common to all tissues, organs, or species. Indeed, observations supporting this mecha nism have been made in higher plant species [185, 186, 222] and recently also in mammalian cell lines [207, 223] The results outlined in Chapte r 3, when viewed in light of these observations, tend to support the hypothesis that diversification of organs and tissues may be driven extensively by differential points of regulatory input in common suites of genes. The analysis we describe in

PAGE 144

144 Chapter 3 found this pattern to be more prevalent for trans acting regulatory loci than for cis acting effects (mirrored by a recent outcome in mice [223] ), but the resolution associated with eQTL mapping in our study makes it difficult to determin e if the same cis acting factors are at work across all tissues. High resolution association -based eQTL mapping in human cell lines indicated that even among closely related cell types, less than 3 0% of cis acting regulatory polymorphisms could be accounte d for by a common SNP variant [207] It will be interesting to determine whether a similar pattern of transcriptional diversity is common for both cis and trans acting regulatory loci in poplar by employing association genetic populations that are currently being developed and propagated (G. Tuskan, M. Kirst; personal communication). The importance of these hypotheses and outcomes must also be considered with regards to the role of ARF1 in regulation of leaf shape. While the initial attempts to localize ARF1 expression to spe cific hotspots of activity in expanding leaves of P. trichocarpa proved generally i neffective (Figure 4 5 ), it must be noted that leaf samples obtained in this manner still consisted of no less than four distinct cell/tissue types (spongy and palisade mesophyll, epidermis, vasculature) and may have been no more enriched for differential ARF1 expression than a full leaf sample The aspects of ARF1 regulation that result in differential leaf shapes may be r estricted to only a subset of these cell types and/or may function homogenously across the plane of the leaf lamina Therefore, more specialized techniques such as laser capture microdissection or single -cell micropipette transcript isolation [224] may be required to determine the specific mechanism of ARF1 regulation and action. Similarly, phenotypes (or apparent lack thereof) in experiments undertaken with transgenic plants expressing mutant or RNAi constructs from constitutive prom oters should be interpreted with appropriate

PAGE 145

145 consideration to the tissue or cell -specific context in which ARF1 and its interac ting factors may normally be functioning in vivo

PAGE 146

146 LIST OF REFERENCES 1. FAO: Global forest resources assessment 2000 Main rep ort In: FAO Forestry paper 140. 2000. 2. FAO: Activities Related to Poplar and Willow Cultivation and Utilization, 2004 through 2007. In: Poplars, Willows, and People's Wellbeing: 23rd Session of the International Poplar Commission. Rome: Food and Agricul ture Organization of the United Nations; 2008. 3. Dickmann DI, Isebrands JG, Eckenwalder JE, Richardson JE: Poplar Culture in North America Ottawa: National Research Council of Canada; 2001. 4. Tuskan GA, Difazio S, Jansson S, Bohlmann J, Grigoriev I, Hel lsten U, Putnam N, Ralph S, Rombauts S, Salamov A et al : The genome of black cottonwood, Populus trichocarpa (Torr. & Gray) Science 2006, 313 (5793):15961604. 5. Taylor G: Populus: arabidopsis for forestry. Do we need a model tree? Ann Bot (Lond) 2002, 90(6):681689. 6. Tuskan GA, DiFazio SP, Teichmann T: Poplar genomics is getting popular: the impact of the poplar genome project on tree research Plant Biol (Stuttg) 2004, 6 (1):2 4. 7. Wullschleger SD, Jansson S, Taylor G: Genomics and forest biology: Popu lus emerges as the perennial favorite. Plant Cell 2002, 14(11):26512655. 8. Stettler R, Bradshaw HD, Heilman PE, Hinckley TM: Biology of Populus and its implications for management and conservation Ottawa, Ontario, Canada: NRC Research Press; 1996. 9. Ro ttmann WH, Meilan R, Sheppard LA, Brunner AM, Skinner JS, Ma C, Cheng S, Jouanin L, Pilate G, Strauss SH: Diverse effects of overexpression of LEAFY and PTLF, a poplar (Populus) homolog of LEAFY/FLORICAULA, in transgenic poplar and Arabidopsis Plant J 2000, 22(3):235 245. 10. Hsu CY, Liu YX, Luthe DS, Yuceer C: Poplar FT2 shortens the juvenile phase and promotes seasonal flowering Plant Cell 2006, 18(8):18461861. 11. Bohlenius H, Huang T, Charbonnel Campaa L, Brunner AM, Jansson S, Strauss SH, Nilsson O: CO/FT regulatory module controls timing of flowering and seasonal growth cessation in trees Science 2006, 312(5776):10401043. 12. Rae AM, Street NR, Rodriguez -Acosta M: Populus Trees In: Genome Mapping and Molecular Breeding in Plants: Forest Trees. Ed ited by Kole C, vol. 7. Berlin: Springer Verlag; 2007: 121.

PAGE 147

147 13. Fillatti JJ, Sellmer J, Mccown B, Haissig B, Comai L: Agrobacterium mediated transformation and regeneration of Populus Molecular & General Genetics 1987, 206:192 199. 14. Analysis of the genome sequence of the flowering plant Arabidopsis thaliana Nature 2000, 408 (6814):796815. 15. Goff SA, Ricke D, Lan TH, Presting G, Wang RL, Dunn M, Glazebrook J, Sessions A, Oeller P, Varma H et al : A draft sequence of the rice genome (Oryza sativa L. ssp japonica) Science 2002, 296 (5565):92100. 16. Yu J, Hu SN, Wang J, Wong GKS, Li SG, Liu B, Deng YJ, Dai L, Zhou Y, Zhang XQ et al : A draft sequence of the rice genome (Oryza sativa L. ssp indica) Science 2002, 296(5565):7992. 17. Bhalerao R, Keskita lo J, Sterky F, Erlandsson R, Bjorkbacka H, Birve SJ, Karlsson J, Gardestrom P, Gustafsson P, Lundeberg J et al: Gene expression in autumn leaves Plant Physiol 2003, 131(2):430 442. 18. Dejardin A, Leple JC, Lesage Descauses MC, Costa G, Pilate G: Express ed sequence tags from poplar wood tissues -a comparative analysis from multiple libraries Plant Biol (Stuttg) 2004, 6 (1):55 64. 19. Kohler A, Delaruelle C, Martin D, Encelot N, Martin F: The poplar root transcriptome: analysis of 7000 expressed sequence t ags Febs Letters 2003, 542(1 3):3741. 20. Lee H, Lee JS, Noh EW, Bae EK, Choi YI, Han MS: Generation and analysis of expressed sequence tags from poplar (Populus alba x P -tremula var. glandulosa) suspension cells Plant Science 2005, 169(6):11181124. 21. Nanjo T, Sakurai T, Totoki Y, Toyoda A, Nishiguchi M, Kado T, Igasaki T, Futamura N, Seki M, Sakaki Y et al : Functional annotation of 19,841 Populus nigra full -length enriched cDNA clones Bmc Genomics 2007, 8 22. Ralph S, Oddy C, Cooper D, Yueh H, Ja ncsik S, Kolosova N, Philippe RN, Aeschliman D, White R, Huber D et al : Genomics of hybrid poplar (Populus trichocarpax deltoides) interacting with forest tent caterpillars (Malacosoma disstria): normalized and full -length cDNA libraries, expressed sequenc e tags, and a cDNA microarray for the study of insect -induced defences in poplar Mol Ecol 2006, 15(5):12751297. 23. Ranjan P, Kao YY, Jiang HY, Joshi CP, Harding SA, Tsai CJ: Suppression subtractive hybridization -mediated transcriptome analysis from mult iple tissues of aspen (Populus tremuloides) altered in phenylpropanoid metabolism Planta 2004, 219(4):694 704.

PAGE 148

148 24. Sterky F, Bhalerao RR, Unneberg P, Segerman B, Nilsson P, Brunner AM, Charbonnel Campaa L, Lindvall JJ, Tandre K, Strauss SH et al : A Populus EST resource for plant functional genomics Proc Natl Acad Sci U S A 2004, 101(38):1395113956. 25. Sterky F, Regan S, Karlsson J, Hertzberg M, Rohde A, Holmberg A, Amini B, Bhalerao R, Larsson M, Villarroel R et al: Gene discovery in the wood -forming ti ssues of poplar: analysis of 5, 692 expressed sequence tags Proc Natl Acad Sci U S A 1998, 95(22):1333013335. 26. Major IT, Constabel CP: Molecular analysis of poplar defense against herbivory: comparison of wound and insect elicitor -induced gene expres sion New Phytol 2006, 172(4):617 635. 27. Miranda M, Ralph SG, Mellway R, White R, Heath MC, Bohlmann J, Constabel CP: The transcriptional response of hybrid poplar (Populus trichocarpa x P. deltoides) to infection by Melampsora medusae leaf rust involves induction of flavonoid pathway genes leading to the accumulation of proanthocyanidins Mol Plant Microbe Interact 2007, 20(7):816831. 28. Rinaldi C, Kohler A, Frey P, Duchaussoy F, Ningre N, Couloux A, Wincker P, Le Thiec D, Fluch S, Martin F et al : Tran script profiling of poplar leaves upon infection with compatible and incompatible strains of the foliar rust Melampsora larici -populina Plant Physiol 2007, 144(1):347 366. 29. Zhang Y, Zhang X, Chen Y, Wang Q, Wang M, Huang M: Function and chromosomal loc alization of differentially expressed genes induced by Marssonina brunnea f. sp multigermtubi in Populus deltoides Journal of Genetics and Genomics 2007, 34(7):641 648. 30. Bogeat Triboulot MB, Brosche M, Renaut J, Jouve L, Le Thiec D, Fayyaz P, Vinocur B Witters E, Laukens K, Teichmann T et al: Gradual soil water depletion results in reversible changes of gene expression, protein profiles, ecophysiology, and growth performance in Populus euphratica, a poplar growing in arid regions Plant Physiol 2007, 1 43 (2):876 892. 31. Brosche M, Vinocur B, Alatalo ER, Lamminmaki A, Teichmann T, Ottow EA, Djilianov D, Afif D, Bogeat Triboulot MB, Altman A et al : Gene expression and metabolite profiling of Populus euphratica growing in the Negev desert Genome Biol 2005, 6 (12):R101. 32. Street NR, Skogstrom O, Sjodin A, Tucker J, Rodriguez -Acosta M, Nilsson P, Jansson S, Taylor G: The genetics and genomics of the drought response in Populus Plant J 2006, 48(3):321341. 33. Taylor G, Street NR, Tricker PJ, Sjodin A, Graham L, Skogstrom O, Calfapietra C, Scarascia -Mugnozza G, Jansson S: The transcriptome of Populus in elevated CO2 New Phytol 2005, 167(1):143 154.

PAGE 149

149 34. Andersson Gunneras S, Mellerowicz EJ, Love J, Segerman B, O hmiya Y, Coutinho PM, Nilsson P, Henrissat B, Moritz T, Sundberg B: Biosynthesis of cellulose -enriched tension wood in Populus: global analysis of transcripts and metabolites identifies biochemical and developmental regulators in secondary wall biosynthesi s Plant J 2006, 45(2):144165. 35. Quesada T, Li Z, Dervinis C, Li Y, Bocock PN, Tuskan GA, Casella G, Davis JM, Kirst M: Comparative analysis of the transcriptomes of Populus trichocarpa and Arabidopsis thaliana suggests extensive evolution of gene expre ssion regulation in angiosperms New Phytol 2008, 180(2):408420. 36. Levine M, Tjian R: Transcription regulation and animal diversity Nature 2003, 424(6945):147151. 37. Keim P, Paige KN, Whitham TG, Lark KG: Genetic analysis of an interspecific hybrid s warm of Populus: occurrence of unidirectional introgression Genetics 1989, 123(3):557 565. 38. Yin T, Huang M, Wang M, Zhu LH, Zeng ZB, Wu R: Preliminary interspecific genetic maps of the populus genome constructed from RAPD markers Genome 2001, 44(4):60 2 609. 39. Yin T, Zhang X, Huang M, Wang M, Zhuge Q, Tu S, Zhu LH, Wu R: Molecular linkage maps of the Populus genome Genome 2002, 45(3):541 555. 40. Yin TM, DiFazio SP, Gunter LE, Riemenschneider D, Tuskan GA: Large -scale heterospecific segregation disto rtion in Populus revealed by a dense genetic map Theor Appl Genet 2004, 109 (3):451463. 41. Gupta PK, Rustgi S: Molecular markers from the transcribed/expressed region of the genome in higher plants Funct Integr Genomics 2004, 4 (3):139162. 42. Kelleher CT, Chiu R, Shin H, Bosdet IE, Krzywinski MI, Fjell CD, Wilkin J, Yin T, DiFazio SP, Ali J et al: A physical map of the highly heterozygous Populus genome: integration with the genome sequence and genetic map and analysis of haplotype variation Plant J 2007, 50(6):10631078. 43. Tuskan GA, Gunter LE, Yang ZMK, Yin TM, Sewell MM, DiFazio SP: Characterization of microsatellites revealed by genomic sequencing of Populus trichocarpa Canadian Journal of Forest Research 2004, 34(1):8593. 44. Dayanandan S, Rajora OP, Bawa KS: Isolation and characterization of microsatellites in trembling aspen ( Populus tremuloides ). Theor Appl Genet 1998, 96:950956. 45. Rahman MH, Dayanandan S, Rajora OP: Microsatellite DNA markers in Populus tremuloides Genome 2000, 43(2):293 297.

PAGE 150

150 46. Smulders MJM, Van Der Schoot J, Arens P, Vosman B: Trinucleotide repeat microsatellite markers for black poplar ( Populus nigra L.) Molec Ecol Notes 2001, 1 (3):188190. 47. van der Schoot J, Pospiskova M, Vosman B, Smulders MJM: Development and characterization of microsatellite markers in black poplar (Populus nigra L.) Theoretical and Applied Genetics 2000, 101(1 2):317322. 48. Rahman MH, Rajora OP: Microsatellite DNA fingerprinting, differentiation, and genetic relationships of clones, culti vars, and varieties of six poplar species from three sections of the genus Populus Genome 2002, 45(6):10831094. 49. Rajora P, Rahman H: Microsatellite DNA and RAPD fingerprinting, identification and genetic relationships of hybrid poplar (Populus x canadensis) cultivars Theor Appl Genet 2003, 106 (3):470477. 50. Unneberg P, Stromberg M, Lundeberg J, Jansson S, Sterky F: Analysis of 70,000 EST sequences to study divergence between two closely related Populus species Tree Genetics & Genomes 2005, 1 (3):109115. 51. Gilchrist EJ, Haughn GW, Ying CC, Otto SP, Zhuang J, Cheung D, Hamberger B, Aboutorabi F, Kalynyak T, Johnson L et al: Use of Ecotilling as an efficient SNP discovery tool to survey genetic variation in wild populations of Populus trichocarpa. Mo lecular Ecology 2006, 15:1367 1378. 52. Hall D, Luquez V, Garcia VM, St Onge KR, Jansson S, Ingvarsson PK: Adaptive population differentiation in phenology across a latitudinal gradient in European aspen (Populus tremula, L.): a comparison of neutral marke rs, candidate genes and phenotypic traits Evolution 2007, 61(12):28492860. 53. Ingvarsson PK: Nucleotide polymorphism and linkage disequilibrium within and among natural populations of European aspen (Populus tremula L., Salicaceae) Genetics 2005, 169 (2 ):945 953. 54. Ingvarsson PK: Multilocus patterns of nucleotide polymorphism and the demographic history of Populus tremula Genetics 2008, 180(1):329 340. 55. Ingvarsson PK, Garcia MV, Hall D, Luquez V, Jansson S: Clinal variation in phyB2, a candidate ge ne for day -length-induced growth cessation and bud set, across a latitudinal gradient in European aspen (Populus tremula) Genetics 2006, 172(3):18451853. 56. Heuertz M, De Paoli E, Kallman T, Larsson H, Jurman I, Morgante M, Lascoux M, Gyllenstrand N: Mu ltilocus patterns of nucleotide diversity, linkage disequilibrium and demographic history of Norway spruce [Picea abies (L.) Karst] Genetics 2006, 174(4):20952105.

PAGE 151

151 57. Krutovsky KV, Neale DB: Nucleotide diversity and linkage disequilibrium in cold hardinessand wood quality -related candidate genes in Douglas fir. Genetics 2005, 171(4):20292041. 58. Ralph SG, Chun HJ, Cooper D, Kirkpatrick R, Kolosova N, Gunter L, Tuskan GA, Douglas CJ, Holt RA, Jones SJ et al : Analysis of 4,664 high quality sequence -fin ished poplar full -length cDNA clones and their utility for the discovery of genes responding to insect feeding BMC Genomics 2008, 9 :57. 59. Novaes E, Drost DR, Farmerie WG, Pappas GJ, Jr., Grattapaglia D, Sederoff RR, Kirst M: High -throughput gene and SNP discovery in Eucalyptus grandis, an uncharacterized genome BMC Genomics 2008, 9 :312. 60. Borevitz JO, Liang D, Plouffe D, Chang HS, Zhu T, Weigel D, Berry CC, Winzeler E, Chory J: Large -scale identification of single -feature polymorphisms in complex geno mes Genome Research 2003, 13(3):513 523. 61. Winzeler EA, Richards DR, Conway AR, Goldstein AL, Kalman S, McCullough MJ, McCusker JH, Stevens DA, Wodicka L, Lockhart DJ et al : Direct allelic variation scanning of the yeast genome Science 1998, 281(5380):11941197. 62. Brem RB, Yvert G, Clinton R, Kruglyak L: Genetic dissection of transcriptional regulation in budding yeast Science 2002, 296(5568):752755. 63. West MAL, van Leeuwen H, Kozik A, Kliebenstein DJ, Doerge RW, St Clair DA, Michelmore RW: High density haplotyping with microarray based expression and single feature polymorphism markers in Arabidopsis Genome Research 2006, 16(6):787 795. 64. Singer T, Fan Y, Chang HS, Zhu T, Hazen SP, Briggs SP: A high-resolution map of Arabidopsis recombinant inbred lines by whole -genome exon array hybridization PLoS Genet 2006, 2 (9):e144. 65. Bradshaw HD, Villar M, Watson BD, Otto KG, Stewart S, Stettler RF: Molecular Genetics of Growth and Development in Populus .3. a Genetic -Linkage Map of a Hybrid Popl ar Composed of Rflp, Sts, and Rapd Markers Theoretical and Applied Genetics 1994, 89(2 3):167178. 66. Zeng ZB: Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc Natl Acad Sci U S A 1993, 90(23):10972 10976. 67. Zeng ZB: Precision mapping of quantitative trait loci. Genetics 1994, 136 (4):14571468. 68. Bradshaw HD, Stettler RF: Molecular Genetics of Growth and Development in Populus .4. Mapping Qtls with Large Effects on Growth, Form, and Phenology Traits in a Forest Tree. Genetics 1995, 139(2):963 973.

PAGE 152

152 69. Rae AM, Ferris R, Tallis MJ, Taylor G: Elucidating genomic regions determining enhanced leaf growth and delayed senescence in elevated CO2 Plant Cell Environ 2006, 29(9):17301741. 70. Wu R, Brad shaw HD, Stettler RF: Molecular genetics of growth and development in Populus (Salicaceae) .5. Mapping quantitative trait loci affecting leaf variation American Journal of Botany 1997, 84(2):143153. 71. Wu R, Stettler RF: Quantitative Genetics of Growth and Development in Populus .1. a 3 Generation Comparison of Tree Architecture During the First 2 Years of Growth Theoretical and Applied Genetics 1994, 89(7 8):10461054. 72. Wu RL: Genetic mapping of QTLs affecting tree growth and architecture in Populus : implication for ideotype breeding Theoretical and Applied Genetics 1998, 96(3 4):447457. 73. Grattapaglia D, Sederoff R: Genetic linkage maps of Eucalyptus grandis and Eucalyptus urophylla using a pseudo -testcross: mapping strategy and RAPD markers Ge netics 1994, 137(4):11211137. 74. Myburg AA, Griffin AR, Sederoff RR, Whetten RW: Comparative genetic linkage maps of Eucalyptus grandis, Eucalyptus globulus and their F -1 hybrid based on a double pseudo -backcross mapping approach Theoretical and Applied Genetics 2003, 107(6):10281042. 75. Novaes E, Osorio L, Drost DR, Miles BL, Boaventura Novaes CR, Benedict C, Dervinis C, Yu Q, Sykes R, Davis M et al: Quantitative genetic analysis of biomass and wood chemistry of Populus under different nitrogen levels New Phytol 2009. 76. Wullschleger S, Yin TM, DiFazio SP, Tschaplinski TJ, Gunter LE, Davis MF, Tuskan GA: Phenotypic variation in growth and biomass distribution for two advancedgeneration pedigrees of hybrid poplar Can J Forest Res 2005, 35(8):17791789. 77. Yin TM, DiFazio SP, Gunter LE, Jawdy SS, Boerjan W, Tuskan GA: Genetic and physical mapping of Melampsora rust resistance genes in Populus and characterization of linkage disequilibrium and flanking genomic sequence New Phytologist 2004, 164(1):95 105. 78. Jansen RC, Nap JP: Genetical genomics: the added value from segregation Trends Genet 2001, 17(7):388391. 79. King MC, Wilson AC: Evolution at 2 levels in humans and chimpanzees Science 1975, 188 (4184):107116. 80. Damerval C, Maurice A, Josse JM, de Vienne D: Quantitative trait loci underlying gene product variation: a novel perspective for analyzing regulation of genome expression Genetics 1994, 137(1):289 301.

PAGE 153

153 81. Schena M, Shalon D, Davis RW, Brown PO: Quantitative monitoring of gene expres sion patterns with a complementary DNA microarray Science 1995, 270(5235):467470. 82. Hughes TR, Mao M, Jones AR, Burchard J, Marton MJ, Shannon KW, Lefkowitz SM, Ziman M, Schelter JM, Meyer MR et al : Expression profiling using microarrays fabricated by an ink -jet oligonucleotide synthesizer. Nature Biotechnology 2001, 19(4):342 347. 83. Schadt EE, Monks SA, Drake TA, Lusis AJ, Che N, Colinayo V, Ruff TG, Milligan SB, Lamb JR, Cavet G et al : Genetics of gene expression surveyed in maize, mouse and man Na ture 2003, 422(6929):297302. 84. Kirst M, Basten CJ, Myburg AA, Zeng ZB, Sederoff RR: Genetic architecture of transcript -level variation in differentiating xylem of a eucalyptus hybrid Genetics 2005, 169 (4):22952303. 85. Kliebenstein DJ, West MA, van Le euwen H, Kim K, Doerge RW, Michelmore RW, St Clair DA: Genomic survey of gene expression diversity in Arabidopsis thaliana Genetics 2006, 172 (2):11791189. 86. Monks SA, Leonardson A, Zhu H, Cundiff P, Pietrusiak P, Edwards S, Phillips JW, Sachs A, Schadt EE: Genetic inheritance of gene expression in human cell lines American Journal of Human Genetics 2004, 75(6):10941105. 87. Morley M, Molony CM, Weber TM, Devlin JL, Ewens KG, Spielman RS, Cheung VG: Genetic analysis of genome wide variation in human ge ne expression Nature 2004, 430(7001):743747. 88. Rockman MV, Kruglyak L: Genetics of global gene expression Nature Reviews Genetics 2006, 7 (11):862872. 89. Wayne ML, McIntyre LM: Combining mapping and arraying: An approach to candidate gene identificat ion Proc Natl Acad Sci USA 2002, 99(23):1490314906. 90. Potokina E, Druka A, Luo Z, Wise R, Waugh R, Kearsey M: Gene expression quantitative trait locus analysis of 16 000 barley genes reveals a complex pattern of genome wide transcriptional regulation Plant J 2008, 53(1):90 101. 91. West MA, Kim K, Kliebenstein DJ, van Leeuwen H, Michelmore RW, Doerge RW, St Clair DA: Global eQTL mapping reveals the complex genetic architecture of transcript -level variation in Arabidopsis Genetics 2007, 175(3):1441145 0. 92. Brem RB, Kruglyak L: The landscape of genetic complexity across 5,700 gene expression traits in yeast Proc Natl Acad Sci U S A 2005, 102(5):15721577. 93. Keurentjes JJ, Fu J, Terpstra IR, Garcia JM, van den Ackerveken G, Snoek LB, Peeters AJ, Vreu gdenhil D, Koornneef M, Jansen RC: Regulatory network construction in

PAGE 154

154 Arabidopsis by using genome wide gene expression quantitative trait loci Proc Natl Acad Sci U S A 2007, 104(5):17081713. 94. Schadt EE, Monks SA, Friend SH: A new paradigm for drug dis covery: integrating clinical, genetic, genomic and molecular phenotype data to identify drug targets Biochemical Society Transactions 2003, 31:437443. 95. Eaves IA, Wicker LS, Ghandour G, Lyons PA, Peterson LB, Todd JA, Glynne RJ: Combining mouse congenic strains and microarray gene expression analyses to study a complex trait: the NOD model of type 1 diabetes Genome Res 2002, 12(2):232 243. 96. Karp CL, Grupe A, Schadt E, Ewart SL, Keane -Moore M, Cuomo PJ, Kohl J, Wahl L, Kuperman D, Germer S et al : Identification of complement factor 5 as a susceptibility locus for experimental allergic asthma Nature Immunol 2000, 1 (3):221226. 97. Wentzell AM, Rowe HC, Hansen BG, Ticconi C, Halkier BA, Kliebenstein DJ: Linking metabolic QTLs with network and ci s -eQTLs controlling biosynthetic pathways PLoS Genet 2007, 3 (9):16871701. 98. Yvert G, Brem RB, Whittle J, Akey JM, Foss E, Smith EN, Mackelprang R, Kruglyak L: Trans acting regulatory variation in Saccharomyces cerevisiae and the role of transcription f actors Nat Genet 2003, 35(1):57 64. 99. DeRisi JL, Iyer VR, Brown PO: Exploring the metabolic and genetic control of gene expression on a genomic scale Science 1997, 278 (5338):680686. 100. Kirst M, Myburg AA, De Leon JP, Kirst ME, Scott J, Sederoff R: C oordinated genetic regulation of growth and lignin revealed by quantitative trait locus analysis of cDNA microarray data in an interspecific backcross of eucalyptus Plant Physiol 2004, 135 (4):23682378. 101. Wentzell AM, Boeye I, Zhang Z, Kliebenstein DJ: Genetic networks controlling structural outcome of glucosinolate activation across development PLoS Genet 2008, 4 (10):e1000234. 102. Zhang B, Horvath S: A general framework for weighted gene co -expression network analysis Statistical Applications in Gen etics and Molecular Biology 2005, 4 103. Lee TI, Rinaldi NJ, Robert F, Odom DT, Bar Joseph Z, Gerber GK, Hannett NM, Harbison CT, Thompson CM, Simon I et al : Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 2002, 298(5594):799804. 104. Ideker T, Thorsson V, Ranish JA, Christmas R, Buhler J, Eng JK, Bumgarner R, Goodlett DR, Aebersold R, Hood L: Integrated genomic and proteomic analyses of a systematically perturbed metabolic network Science 2001, 292(5518):929934.

PAGE 155

155 105. Bing N, Hoeschele I: Genetical genomics analysis of a yeast segregant population for transcription network inference Genetics 2005, 170(2):533542. 106. Wu C, Delano DL, Mitro N, Su SV, Janes J, McClurg P, Batalov S, Welch GL, Zhang J, Orth AP et al : Gene set enrichment in eQTL data identifies novel annotations and pathway regulators PLoS Genet 2008, 4 (5):e1000070. 107. Zhu J, Lum PY, Lamb J, GuhaThakurta D, Edwards SW, Thieringer R, Berger JP, Wu MS, Thompson J, Sachs AB et al : An integrative genomics ap proach to the reconstruction of gene networks in segregating populations Cytogenet Genome Res 2004, 105 (2 4):363374. 108. Chaibub Neto E, Ferrara CT, Attie AD, Yandell BS: Inferring causal phenotype networks from segregating populations Genetics 2008, 1 79(2):10891100. 109. Dobrin R, Zhu J, Molony C, Argman C, Parrish ML, Carlson S, Allan MF, Pomp D, Schadt EE: Multi-tissue coexpression networks reveal unexpected subnetworks associated with disease Genome Biol 2009, 10(5):R55. 110. Emilsson V, Thorleifs son G, Zhang B, Leonardson AS, Zink F, Zhu J, Carlson S, Helgason A, Walters GB, Gunnarsdottir S et al : Genetics of gene expression and its effect on disease Nature 2008, 452(7186):423428. 111. Hood L, Heath JR, Phelps ME, Lin B: Systems biology and new technologies enable predictive and preventative medicine Science 2004, 306(5696):640643. 112. Mehrabian M, Allayee H, Stockton J, Lum PY, Drake TA, Castellani LW, Suh M, Armour C, Edwards S, Lamb J et al : Integrating genotypic and expression data in a se gregating mouse population to identify 5-lipoxygenase as a susceptibility gene for obesity and bone traits Nat Genet 2005, 37(11):12241233. 113. Shlomi T, Cabili MN, Herrgard MJ, Palsson BO, Ruppin E: Network based prediction of human tissue -specific met abolism Nat Biotechnol 2008, 26(9):10031010. 114. Fu J, Keurentjes JJ, Bouwmeester H, America T, Verstappen FW, Ward JL, Beale MH, de Vos RC, Dijkstra M, Scheltema RA et al: System wide molecular evidence for phenotypic buffering in Arabidopsis Nat Gene t 2009, 41(2):166167. 115. Zhu J, Zhang B, Smith EN, Drees B, Brem RB, Kruglyak L, Bumgarner RE, Schadt EE: Integrating large -scale functional genomic data to dissect the complexity of yeast regulatory networks Nat Genet 2008, 40(7):854861. 116. Eckenwa lder JE: Systematics and evolution of Populus In: Biology of Populus and its implications for management and conservation. Edited by Stettler RF, Bradshaw HD, Heilman PE, Hinckley TM. Ottawa, Canada: NRC Research Press; 1996: 7 -32. 117. Cronk QC: Plant ec o -devo: the potential of poplar as a model organism New Phytol 2005, 166 (1):39 48.

PAGE 156

156 118. Collinson ME: The early fossil history of Salicaceae: a brief review Proc Roy Soc Edinburgh 1992, 98B :155 167. 119. Harrington CA, Radwan MA, DeBell DS: Leaf characteristics reflect growth rates of 2 -year -old Populus trees Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 1997, 27(8):13211325. 120. Marron N, Villar M, Dreyer E, Delay D, Boudouresque E, Petit JM, Delmotte FM, Guehl J M, Brignolas F: Diversity of leaf traits related to productivity in 31 Populus deltoides x Populus nigra clones Tree Physiol 2005, 25(4):425435. 121. Marron N, Ceulemans R: Genetic variation of leaf traits related to productivity in a Populus deltoides x Populus nigra family Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 2006, 36(2):390 400. 122. Rae AM, Tricker PJ, Bunn SM, Taylor G: Adaptation of tree growth to elevated CO2: quantitative trait loci for biomass in Populus New Phytol 2007, 175(1):59 69. 123. Street NR, Sjodin A, Bylesjo M, Gustafsson P, Trygg J, Jansson S: A cross species transcriptomics approach to identify genes involved in leaf development BMC Genomics 2008, 9 :589. 124. Ferris R, Long L, Bunn SM, Robinson KM, Bradshaw HD, Rae AM, Taylor G: Leaf stomatal and epidermal cell development: identification of putative quantitative trait loci in relation to elevated carbon dioxide concentration in poplar Tree Physiology 2002, 22(9):633 640. 125. Yin T, Difazio S P, Gunter LE, Zhang X, Sewell MM, Woolbright SA, Allan GJ, Kelleher CT, Douglas CJ, Wang M et al : Genome structure and emerging evidence of an incipient sex chromosome in Populus Genome Res 2008, 18(3):422 430. 126. Morreel K, Goeminne G, Storme V, Sterck L, Ralph J, Coppieters W, Breyne P, Steenackers M, Georges M, Messens E et al : Genetical metabolomics of flavonoid biosynthesis in Populus: a case study Plant J 2006, 47(2):224 237. 127. Rae AM, Pinel MPC, Bastien C, Sabatti M, Street NR, Tucker J, Dixon C, Marron N, Dillen SY, Taylor G: QTL for yield in bioenergy Populus: identifying GxE interactions from growth at three contrasting sites Tree Genetics & Genomes 2008, 4 (1):97 112. 128. Mackay TF: The genetic architecture of quantitative traits Annu Rev Genet 2001, 35:303339. 129. Tsukaya H: Mechanism of leaf -shape determination. Annu Rev Plant Biol 2006, 57:477496. 130. Donnelly PM, Bonetta D, Tsukaya H, Dengler RE, Dengler NG: Cell cycling and cell enlargement in developing leaves of Arabidopsis Dev Biol 1999, 215(2):407419.

PAGE 157

157 131. Tsuge T, Tsukaya H, Uchimiya H: Two independent and polarized processes of cell elongation regulate leaf blade expansion in Arabidopsis thaliana (L.) Heynh Development 1996, 122 (5):15891600. 132. Kim GT, Tsukaya H, Uchimi ya H: The ROTUNDIFOLIA3 gene of Arabidopsis thaliana encodes a new member of the cytochrome P -450 family that is required for the regulated polar elongation of leaf cells Genes Dev 1998, 12(15):2381 2391. 133. Narita NN, Moore S, Horiguchi G, Kubo M, Demu ra T, Fukuda H, Goodrich J, Tsukaya H: Overexpression of a novel small peptide ROTUNDIFOLIA4 decreases cell proliferation and alters leaf shape in Arabidopsis thaliana Plant J 2004, 38(4):699713. 134. Kim GT, Shoda K, Tsuge T, Cho KH, Uchimiya H, Yokoyama R, Nishitani K, Tsukaya H: The ANGUSTIFOLIA gene of Arabidopsis, a plant CtBP gene, regulates leaf -cell expansion, the arrangement of cortical microtubules in leaf cells and expression of a gene involved in cell wall formation Embo J 2002, 21(6):12671279. 135. Corda D, Colanzi A, Luini A: The multiple activities of CtBP/BARS proteins: the Golgi view Trends Cell Biol 2006, 16(3):167173. 136. Chinnadurai G: CtBP, an unconventional transcriptional core pressor in development and oncogenesis Mol Cell 2002, 9 (2):213224. 137. Stern MD, Aihara H, Cho KH, Kim GT, Horiguchi G, Roccaro GA, Guevara E, Sun HH, Negeri D, Tsukaya H et al : Structurally related Arabidopsis ANGUSTIFOLIA is functionally distinct from the transcriptional corepressor CtBP Dev Genes Evol 2007, 217 (11 12):759769. 138. Horiguchi G, Kim GT, Tsukaya H: The transcription factor AtGRF5 and the transcription coactivator AN3 regulate cell proliferation in leaf primordia of Arabidopsis thaliana Plant J 2005, 43(1):68 78. 139. Kumar R, Qiu J, Joshi T, Valliyodan B, Xu D, Nguyen HT: Single feature polymorphism discovery in rice PLoS ONE 2007, 2 (3):e284. 140. Ronald J, Akey JM, Whittle J, Smith EN, Yvert G, Kruglyak L: Simultaneous genotyping, ge ne -expression measurement, and detection of allele -specific expression with oligonucleotide arrays Genome Res 2005, 15(2):284291. 141. Coram TE, Settles ML, Wang M, Chen X: Surveying expression level polymorphism and single -feature polymorphism in near -i sogenic wheat lines differing for the Yr5 stripe rust resistance locus Theor Appl Genet 2008, 117(3):401 411. 142. Cui X, Xu J, Asghar R, Condamine P, Svensson JT, Wanamaker S, Stein N, Roose M, Close TJ: Detecting single -feature polymorphisms using oligonucleotide arrays and robustified projection pursuit Bioinformatics 2005, 21(20):38523858.

PAGE 158

1 58 143. Luo ZW, Potokina E, Druka A, Wise R, Waugh R, Kearsey MJ: SFP genotyping from affymetrix arrays is robust but largely detects cis acting expression regulators Genetics 2007, 176 (2):789 800. 144. Rostoks N, Borevitz JO, Hedley PE, Russell J, Mudie S, Morris J, Cardle L, Marshall DF, Waugh R: Single -feature polymorphism discovery in the barley transcriptome Genome Biol 2005, 6 (6). 145. West MA, van Leeuwen H, K ozik A, Kliebenstein DJ, Doerge RW, St Clair DA, Michelmore RW: High density haplotyping with microarray based expression and single feature polymorphism markers in Arabidopsis Genome Res 2006, 16(6):787795. 146. Kirst M, Caldo R, Casati P, Tanimoto G, W albot V, Wise RP, Buckler ES: Genetic diversity contribution to errors in short oligonucleotide microarray analysis Plant Biotechnol J 2006, 4 (5):489498. 147. Cooke JEK, Brown KA, Wu R, Davis JM: Gene expression associated with N -induced shifts in resour ce allocation in poplar Plant Cell and Environment 2003, 26(5):757 770. 148. Chang S, Puryear J, Cairney J: A simple and efficient method for isolating RNA from pine trees Plant Mol Biol Rep 1993, 11:117121. 149. Tuskan GA, Gunter LE, Yang ZMK, Yin TM, Sewell MM, DiFazio SP: Characterization of microsatellites revealed by genomic sequencing of Populus trichocarpa Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 2004, 34(1):85 93. 150. L ander ES, Green P, Abrahamson J, Barlow A, Daly MJ, Lincoln SE, Newburg L: MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations Genomics 1987, 1 (2):174181. 151. Ma CX, Yu Q, Berg A, Drost D, Novaes E, Fu G, Yap JS, Tan A, Kirst M, Cui Y et al: A statistical model for testing the pleiotropic control of phenotypic plasticity for a count trait Genetics 2008, 179(1):627 636. 152. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Bas ic local alignment search tool J Mol Biol 1990, 215(3):403410. 153. Bassam BJ, Caetanoanolles G, Gresshoff PM: Fast and Sensitive Silver Staining of DNA in Polyacrylamide Gels Analytical Biochemistry 1991, 196(1):8083. 154. Bolstad BM, Irizarry RA, Ast rand M, Speed TP: A comparison of normalization methods for high density oligonucleotide array data based on variance and bias Bioinformatics 2003, 19(2):185 193.

PAGE 159

159 155. Storey JD, Tibshirani R: Statistical significance for genomewide studies Proc Natl Aca d Sci U S A 2003, 100(16):94409445. 156. Haldane JB, Waddington CH: Inbreeding and Linkage Genetics 1931, 16(4):357374. 157. Zeileis A, F. Leish, K. Hornik, and C. Kleiber: Strucchange: An R package for testing structural change in linear regression mod els J Statistical Software 2002, 7 :1 38. 158. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R et al : Clustal W and Clustal X version 2.0 Bioinformatics 2007, 23(21):29472948. 159. Cerver a MT, Storme V, Ivens B, Gusmao J, Liu BH, Hostyn V, Van Slycken J, Van Montagu M, Boerjan W: Dense genetic linkage maps of three Populus species (Populus deltoides, P. nigra and P. trichocarpa) based on AFLP and microsatellite markers Genetics 2001, 158(2):787809. 160. Das S, Bhat PR, Sudhakar C, Ehlers JD, Wanamaker S, Roberts PA, Cui X, Close TJ: Detection and validation of single feature polymorphisms in cowpea (Vigna unguiculata L. Walp) using a soybean genome array BMC Genomics 2008, 9 :107. 161. Re nnie C, Noyes HA, Kemp SJ, Hulme H, Brass A, Hoyle DC: Strong position dependent effects of sequence mismatches on signal ratios measured using long oligonucleotide microarrays BMC Genomics 2008, 9 (1):317. 162. Ching A, Caldwell KS, Jung M, Dolan M, Smith OS, Tingey S, Morgante M, Rafalski AJ: SNP frequency, haplotype structure and linkage disequilibrium in elite maize inbred lines BMC Genet 2002, 3 :19. 163. Kolkman JM, Berry ST, Leon AJ, Slabaugh MB, Tang S, Gao W, Shintani DK, Burke JM, Knapp SJ: Single nucleotide polymorphisms and linkage disequilibrium in sunflower Genetics 2007, 177(1):457468. 164. Tenaillon MI, Sawkins MC, Anderson LK, Stack SM, Doebley J, Gaut BS: Patterns of diversity and recombination along chromosome 1 of maize (Zea mays ssp. mays L.) Genetics 2002, 162 (3):14011413. 165. Voorrips RE: MapChart: software for the graphical presentation of linkage maps and QTLs J Hered 2002, 93(1):77 78. 166. Vuylsteke M, van Eeuwijk F, Van Hummelen P, Kuiper M, Zabeau M: Genetic analysis of vari ation in gene expression in Arabidopsis thaliana Genetics 2005, 171(3):12671275. 167. Breitling R, Li Y, Tesson BM, Fu J, Wu C, Wiltshire T, Gerrits A, Bystrykh LV, de Haan G, Su AI et al: Genetical genomics: spotlight on QTL hotspots PLoS Genet 2008, 4 (10):e1000232.

PAGE 160

160 168. Sonderby IE, Hansen BG, Bjarnholt N, Ticconi C, Halkier BA, Kliebenstein DJ: A systems biology approach identifies a R2R3 MYB gene subfamily with distinct and overlapping functions in regulation of aliphatic glucosinolates PLoS ONE 2007, 2 (12):e1322. 169. Kliebenstein DJ, West MA, van Leeuwen H, Loudet O, Doerge RW, St Clair DA: Identification of QTLs controlling gene expression networks defined a priori BMC Bioinformatics 2006, 7 :308. 170. DeCook R, Lall S, Nettleton D, Howell SH: Gen etic regulation of gene expression during shoot development in Arabidopsis Genetics 2006, 172(2):11551164. 171. Chesler EJ, Lu L, Shou S, Qu Y, Gu J, Wang J, Hsu HC, Mountz JD, Baldwin NE, Langston MA et al : Complex trait analysis of gene expression unco vers polygenic and pleiotropic networks that modulate nervous system function. Nat Genet 2005, 37(3):233 242. 172. Petretto E, Mangion J, Dickens NJ, Cook SA, Kumaran MK, Lu H, Fischer J, Maatz H, Kren V, Pravenec M et al : Heritability and tissue specificity of expression quantitative trait loci. PLoS Genet 2006, 2 (10):e172. 173. Drost DR, Novaes E, Boaventura Novaes C, Benedict CI, Brown RS, Yin T, Tuskan GA, Kirst M: A microarray -based genotyping and genetic mapping approach for highly heterozygous outcro ssing species enables localization of a large fraction of the unassembled Populus trichocarpa genome sequence Plant J 2009. 174. Wang S. CJB, and Z. -B. Zeng Windows QTL Cartographer 2.5 In Raleigh, NC: Department of Statistics, North Carolina State Univ ersity; 2007: http://statgen.ncsu.edu/qtlcart/WQTLCart.htm 175. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT et al: Gene ontology: tool for the unification of biology. The Gene Ontology Consortium Nat Genet 2000, 25(1):25 29. 176. Higo K, Ugawa Y, Iwamoto M, Korenaga T: Plant cis acting regulatory DNA elements (PLACE) database: 1999 Nucleic Acids Res 1999, 27(1):297300. 177. Yan T, Yoo D, Berardini TZ, Mueller LA, Weems DC, Weng S, Cherry JM, Rhee SY: PatMatch: a program for finding patterns in peptide and nucleotide sequences Nucleic Acids Res 2005, 33(Web Server issue):W262 266. 178. McAndrew RS, Froehlich JE, Vitha S, Stok es KD, Osteryoung KW: Colocalization of plastid division proteins in the chloroplast stromal compartment establishes a new functional relationship between FtsZ1 and FtsZ2 in higher plants Plant Physiol 2001, 127 (4):16561666.

PAGE 161

161 179. Stokes KD, McAndrew RS, Figueroa R, Vitha S, Osteryoung KW: Chloroplast division and morphology are differentially affected by overexpression of FtsZ1 and FtsZ2 genes in Arabidopsis Plant Physiol 2000, 124(4):16681677. 180. Sun C, Palmqvist S, Olsson H, Boren M, Ahlandsberg S, Jansson C: A novel WRKY transcription factor, SUSIBA2, participates in sugar signaling in barley by binding to the sugar -responsive elements of the iso1 promoter Plant Cell 2003, 15(9):20762092. 181. Muller R, Morant M, Jarmer H, Nilsson L, Nielsen TH: G enome wide analysis of the Arabidopsis leaf transcriptome reveals interaction of phosphate and sugar metabolism Plant Physiol 2007, 143(1):156171. 182. Koch KE: Carbohydrate -Modulated Gene Expression in Plants Annu Rev Plant Physiol Plant Mol Biol 1996, 47:509540. 183. Jansen RC, Tesson BM, Fu J, Yang Y, McIntyre LM: Defining gene and QTL networks Curr Opin Plant Biol 2009. 184. Blanc G, Wolfe KH: Functional divergence of duplicated genes formed by polyploidy during Arabidopsis evolution. Plant Cell 2004, 16(7):16791691. 185. Potokina E, Druka A, Luo Z, Moscou M, Wise R, Waugh R, Kearsey M: Tissue dependent limited pleiotropy affects gene expression in barley Plant J 2008, 56(2):287 296. 186. Schmid M, Davison TS, Henz SR, Pape UJ, Demar M, Vingron M, Scholkopf B, Weigel D, Lohmann JU: A gene expression map of Arabidopsis thaliana development Nat Genet 2005, 37(5):501506. 187. Byrne ME: Networks in leaf development Curr Opin Plant Biol 2005, 8 (1):59 66. 188. Barkoulas M, Hay A, Kougioumoutzi E, Tsia ntis M: A developmental framework for dissected leaf formation in the Arabidopsis relative Cardamine hirsuta Nat Genet 2008, 40(9):11361141. 189. Hay A, Tsiantis M: The genetic basis for differences in leaf form between Arabidopsis thaliana and its wild relative Cardamine hirsuta Nat Genet 2006, 38(8):942 947. 190. Holtan HE, Hake S: Quantitative trait locus analysis of leaf dissection in tomato using Lycopersicon pennellii segmental introgression lines Genetics 2003, 165(3):15411550. 191. Gailing O: Q TL analysis of leaf morphological characters in a Quercus robur full -sib family (Q. robur x Q. robur ssp. slavonica) Plant Biol (Stuttg) 2008, 10(5):624 634.

PAGE 162

162 192. Mickelson SM, Stuber CS, Senior L, Kaeppler SM: Quantitative trait loci controlling leaf and tassel traits in a B73 x MO17 population of maize Crop Science 2002, 42(6):19021909. 193. Juenger T, Perez -Perez JM, Bernal S, Micol JL: Quantitative trait loci mapping of floral and leaf morphology traits in Arabidopsis thaliana: evidence for modular g enetic architecture. Evol Dev 2005, 7 (3):259271. 194. Kearsey M: QTL analysis: problems and (possible) solutions In: Quantitative Genetics, Genomics and Plant Breeding. Edited by Kang MS. Wallingford, U.K.: CABI Publishing; 2002: 45 58. 195. Hansen BG, Halkier BA, Kliebenstein DJ: Identifying the molecular basis of QTLs: eQTLs add a new dimension. Trends Plant Sci 2008, 13(2):72 77. 196. Tuskan GA, DiFazio, S.P., Teichmann, T.: Poplar Genomics is Getting Popular: The Impact of the Poplar Genome Project o n Tree Research Plant Biology 2004, 6 :2 4. 197. Marron N, Villar M, Dreyer E, Delay D, Boudouresque E, Petit JM, Delmotte FM, Guehl JM, Brignolas F: Diversity of leaf traits related to productivity in 31 Populus deltoides x Populus nigra clones Tree Phys iology 2005, 25(4):425 435. 198. Xu J, Scheres B: Dissection of Arabidopsis ADP -RIBOSYLATION FACTOR 1 function in epidermal cell polarity Plant Cell 2005, 17(2):525536. 199. Wang SC, Basten CJ, Zeng ZB: Windows QTL Cartographer 2.5. Raleigh, NC.: Depart ment of Statistics, North Carolina State University.; 2007. 200. Dascher C, Balch WE: Dominant inhibitory mutants of ARF1 block endoplasmic reticulum to Golgi transport and trigger disassembly of the Golgi apparatus J Biol Chem 1994, 269 (2):14371448. 201. Pepperkok R, Whitney JA, Gomez M, Kreis TE: COPI vesicles accumulating in the presence of a GTP restricted arf1 mutant are depleted of anterograde and retrograde cargo J Cell Sci 2000, 113 ( Pt 1) :135144. 202. Peirson SN, Butler JN, Foster RG: Experime ntal validation of novel and conventional approaches to quantitative real -time PCR data analysis Nucleic Acids Res 2003, 3 (e73). 203. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van R, De Paepe A, Speleman F: Accurate normalization of real -time quantitative RT -PCR data by geometric averaging of multiple internal control genes Genome Biol 2002, 3 :0034. 204. Jang HJ, Pih KT, Kang SG, Lim JH, Jin JB, Piao HL, Hwang I: Molecular cloning of a novel Ca2+ -binding protein that is induced by NaCl stress Plant Mol Biol 1998, 37(5):839 847.

PAGE 163

163 205. Click ES, Stearns T, Botstein D: Systematic structure -function analysis of the small GTPase Arf1 in yeast Mol Biol Cell 2002, 13(5):16521664. 206. Clark RM, Wagler TN, Quijada P, Doebley J: A distant upstream enh ancer at the maize domestication gene tb1 has pleiotropic effects on plant and inflorescent architecture Nat Genet 2006, 38(5):594597. 207. Dimas AS, Deutsch S, Stranger BE, Montgomery SB, Borel C, Attar -Cohen H, Ingle C, Beazley C, Gutierrez Arcelus M, Sekowska M et al : Common regulatory variation impacts gene expression in a cell type -dependent manner Science 2009, 325(5945):12461250. 208. Dhonukshe P, Tanaka H, Goh T, Ebine K, Mahonen AP, Prasad K, Blilou I, Geldner N, Xu J, Uemura T et al : Generation of cell polarity in plants links endocytosis, auxin distribution and cell fate decisions Nature 2008, 456(7224):962966. 209. Boutte Y, Ikeda Y, Grebe M: Mechanisms of auxin -dependent cell and tissue polarity Curr Opin Plant Biol 2007, 10(6):6 16 623. 210. Grebe M, Friml J, Swarup R, Ljung K, Sandberg G, Terlou M, Palme K, Bennett MJ, Scheres B: Cell polarity signaling in Arabidopsis involves a BFA -sensitive auxin influx pathway Curr Biol 2002, 12(4):329 334. 211. Reinhardt D, Pesce ER, Stieger P, Mandel T, Baltensperger K, Bennett M, Traas J, Friml J, Kuhlemeier C: Regulation of phyllotaxis by polar auxin transport Nature 2003, 426(6964):255260. 212. Scanlon MJ: The polar auxin transport inhibitor N -1 -naphthylphthalamic acid disrupts leaf ini tiation, KNOX protein regulation, and formation of leaf margins in maize Plant Physiol 2003, 133(2):597 605. 213. Scarpella E, Marcos D, Friml J, Berleth T: Control of leaf vascular patterning by polar auxin transport Genes Dev 2006, 20(8):10151027. 214. Myburg A, Grattapaglia D, Tuskan G, Schmutz J, Barry K, Bristow JD, Network TEG: Sequencing the Eucalyptus Genome: Genomic Resources for Renewable Energy and Fiber Production; W195 In: International Plant and Animal Genome Conference XVI: 2008; San Diego, CA ; 2008. 215. Okou DT, Steinberg KM, Middle C, Cutler DJ, Albert TJ, Zwick ME: Microarray based genomic selection for high -throughput resequencing Nature Methods 2007, 4 (11):907909. 216. Gnirke A, Melnikov A, Maguire J, Rogov P, LeProust EM, Brockman W, Fennell T, Giannoukos G, Fisher S, Russ C et al: Solution hybrid selection with ultra -long oligonucleotides for massively parallel targeted sequencing Nat Biotechnol 2009, 27(2):182 189.

PAGE 164

164 217. Eid J, Fehr A, Gray J, Luong K, Lyle J, Otto G, Peluso P, R ank D, Baybayan P, Bettman B et al: Real Time DNA Sequencing from Single Polymerase Molecules Science 2009, 323 (5910):133138. 218. Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, Berka J, Braverman MS, Chen YJ, Chen Z et al : Genome seque ncing in microfabricated high density picolitre reactors Nature 2005, 437(7057):376380. 219. Chesler EJ, Lu L, Shou SM, Qu YH, Gu J, Wang JT, Hsu HC, Mountz JD, Baldwin NE, Langston MA et al : Complex trait analysis of gene expression uncovers polygenic a nd pleiotropic networks that modulate nervous system function. Nature Genetics 2005, 37(3):233 242. 220. de Haan G, Bystrykh LV, Weersing E, Sutton S, Manly K, Chesler E, Jansen RC, Vellenga E, Williams RW, Cooke MP: Genetical genomics to identify gene net works in hematopoietic stem cells Experimental Hematology 2005, 33(7):60 60. 221. de Haan G, Weersing E, Dontje B, Sutton S, Vellenga E, Manly KF, Williams RW, Cooke M, Bystrykh LV: Identification of transcriptional networks affecting hematopoietic stem cell function using "genetical genomics" Blood 2003, 102(11):343A 343A. 222. Segerman B, Jansson S, Karlsson J: Characterization of genes with tissue specific differential expression patterns in Populus Tree Genetics & Genomes 2007, 3 (4):351362. 223. Ge rrits A, Li Y, Tesson BM, Bystrykh LV, Weersing E, Ausema A, Dontje B, Wang X, Breitling R, Jansen RC et al: Expression quantitative trait l oci are highly sensitive to cellular differentiation state PLoS Genet 2009, 5 (10):e1000692. 224. Lee JY, Levesque M Benfey PN: High -throughput RNA isolation technologies. New tools for high -resolution gene expression profiling in plant systems Plant Physiol 2005, 138 (2):585 590.

PAGE 165

165 BIOGRAPHICAL SKETCH Derek R. Drost was born in Madison, Wisconsin in 1983. Upon graduating from high school in 2001, he attended the University of Nebraska Lincoln, where he obtained a Bachelor of Science degree in a gricultural s c iences and majored in a gronomy. During that time, Derek conducted research to determine the molecular gene tic s of resistance to acetolactate synthase inhibitor herbicides among invasive Sorghum bicolor populations. In 2005, Derek joined the Plant Molecular and Cellular Biology graduate program at the University of Florida, where he researched the quantitative genetics of transcription in Populus under the direction of Drs. Matias Kirst and Gary Peter Derek completed his Ph.D. in December of 2009 and is currently employed by the Monsanto Corporation, where he contribute s to a research program aimed at discover ing genes influencing biotic stress tolerance in maize.