<%BANNER%>

Regulation of Sink Strength in Developing Maize Florets

Permanent Link: http://ufdc.ufl.edu/UFE0022061/00001

Material Information

Title: Regulation of Sink Strength in Developing Maize Florets Implications for Seed Set and Grain Yield
Physical Description: 1 online resource (149 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: 454, drought, inflorescence, maize, maizeflower, maizeovary, microarray, profiling, transcript
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

Notes

Abstract: The pre- and early post-pollination phases of maize (Zea mays L.) reproductive development are critical for seed set and subsequent grain yield. During this time, the plant is especially sensitive to abiotic stresses, such as drought, which can reduce pollination efficiency or lead to kernel abortion. A key determinant of reproductive success is carbohydrate allocation and use in the developing female inflorescence, which is often disrupted by stress. Understanding the mechanisms that underlie regulation of sink strength during normal progression of maize floral development is thus central to improving seed set under adverse environmental conditions. In this work, we tested the hypothesis that expression of genes related to specific metabolic or regulatory pathways would change in association with carbohydrate allocation and use during silk exsertion and pollination in maize. Individual stages of pre- and early post-pollination maize female florets were characterized based on physical characteristics and expression of a molecular marker for development. Subsequent analyses revealed a shift in sink strength, as approximated by dry and fresh weights, during the pollination period from rapidly expanding silks and subtending floral structures (lemma, palea, and glumes) to the developing ovary and pedicel. This shift coincided with isoform-specific expression of sucrose metabolizing invertases, which provide hexose substrates essential for turgor-based expansion prior to pollination and also for post-pollination growth of symplastically-isolated filial tissues. In addition, accumulation of sucrose and hexoses in the pedicel and ovary, respectively, indicated that invertases could contribute to spatial regulation of cell expansion and differentiation during development. We used a genome-wide transcript profiling approach to determine whether co-expressed genes were related to specific functional processes or associated with relevant metabolic pathways during pre- and early post-pollination ovary development. A gene-specific, sequence-based, 3'-UTR profiling strategy was developed and tested in parallel to microarray analyses. We resolved co-expression profiles for key genes related to nitrogen and amino acid metabolism, carbohydrate metabolism, lignin biosynthesis, and cell growth during ovary development. Transcript profiles were combined with sugar and metabolite analyses and fresh/dry weight quantifications to further support relevance of key sets of co-expressed genes during ovary sink establishment in maize. Results from this study provide evidence for testable roles of such genes in kernel set.
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.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Koch, Karen E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022061:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022061/00001

Material Information

Title: Regulation of Sink Strength in Developing Maize Florets Implications for Seed Set and Grain Yield
Physical Description: 1 online resource (149 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: 454, drought, inflorescence, maize, maizeflower, maizeovary, microarray, profiling, transcript
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

Notes

Abstract: The pre- and early post-pollination phases of maize (Zea mays L.) reproductive development are critical for seed set and subsequent grain yield. During this time, the plant is especially sensitive to abiotic stresses, such as drought, which can reduce pollination efficiency or lead to kernel abortion. A key determinant of reproductive success is carbohydrate allocation and use in the developing female inflorescence, which is often disrupted by stress. Understanding the mechanisms that underlie regulation of sink strength during normal progression of maize floral development is thus central to improving seed set under adverse environmental conditions. In this work, we tested the hypothesis that expression of genes related to specific metabolic or regulatory pathways would change in association with carbohydrate allocation and use during silk exsertion and pollination in maize. Individual stages of pre- and early post-pollination maize female florets were characterized based on physical characteristics and expression of a molecular marker for development. Subsequent analyses revealed a shift in sink strength, as approximated by dry and fresh weights, during the pollination period from rapidly expanding silks and subtending floral structures (lemma, palea, and glumes) to the developing ovary and pedicel. This shift coincided with isoform-specific expression of sucrose metabolizing invertases, which provide hexose substrates essential for turgor-based expansion prior to pollination and also for post-pollination growth of symplastically-isolated filial tissues. In addition, accumulation of sucrose and hexoses in the pedicel and ovary, respectively, indicated that invertases could contribute to spatial regulation of cell expansion and differentiation during development. We used a genome-wide transcript profiling approach to determine whether co-expressed genes were related to specific functional processes or associated with relevant metabolic pathways during pre- and early post-pollination ovary development. A gene-specific, sequence-based, 3'-UTR profiling strategy was developed and tested in parallel to microarray analyses. We resolved co-expression profiles for key genes related to nitrogen and amino acid metabolism, carbohydrate metabolism, lignin biosynthesis, and cell growth during ovary development. Transcript profiles were combined with sugar and metabolite analyses and fresh/dry weight quantifications to further support relevance of key sets of co-expressed genes during ovary sink establishment in maize. Results from this study provide evidence for testable roles of such genes in kernel set.
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.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Koch, Karen E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022061:00001


This item has the following downloads:


Full Text





REGULATION OF SINK STRENGTH IN DEVELOPING MAIZE FLORETS:
IMPLICATIONS FOR SEED SET AND GRAIN YIELD


















By

ANDREA LEE EVELAND


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

UNIVERSITY OF FLORIDA

2008


































2008 Andrea L. Eveland




























To Joshua Shome, in loving memory.









ACKNOWLEDGMENTS

Many thanks go to the members of my committee, Donald McCarty, John Davis, Robert

Ferl, and Edward Braun, for all of their advice and critical insight to specific areas of my

research. I also whole-heartedly thank the Koch lab group, Sylvia De Sousa, Chip Hunter, Brent

O'Brien, and especially Wayne Avigne, for their friendship and support every step of the way.

Special thanks go to Li-Fen Huang, whose work ethic and precision inspired me as a young

graduate student. I thank Jie Yang, Matias Kirst, and Lauren McIntyre for their time and advice

in the statistical analyses for this work. Also, thanks go to Marina Telonis-Scott and Mick Popp

for allowing me to use their microarray facility and providing support during optimization of

array hybridizations. I thank Eric Schmelz and Denise Tieman for their assistance with GC/MS

and HPLC experiments, respectively. Many thanks go to Tom Davenport and Jonathan Crane at

T.R.E.C. for their kind support during my stay at their research facility. I thank the camaraderie

of friends and fellow colleagues, Stefanie Maruhnich, Maggie Kellogg, Michele Auldridge,

Travis Baughman, and Matt Reyes for insightful conversations over the years. Very special

thanks go to my parents for their love and support throughout my life. They have given so much

and I owe to them my determination and achievements.

Finally, I thank Karen Koch for her support and guidance during this very important time

in my life. She has been not only an advisor and colleague, but an extraordinary role model for

me throughout the years. Her passion for science and her constructive criticisms have cultivated

my development into an aspiring young investigator.









TABLE OF CONTENTS

page

A CK N O W LED G M EN T S ................................................................. ........... ............. .....

LIST OF FIGURES .................................. .. ..... ..... ................. .8

L IST O F A B B R E V IA T IO N S .................................................................. .............................. 10

A B S T R A C T ................................ ............................................................ 1 1

CHAPTER

1 L IT E R A TU R E R E V IE W ............................................................................... .................. 13

In tro d u ctio n ............... .. ........... .......... .......... ........................ ................ 13
Source/Sink Relations and Sucrose Metabolizing Enzymes ...............................................14
Sugar Sensing and Signaling ............................................................. .... ........................... 16
Horm one and Sugar Signaling N networks ........................................ .......................... 19
M aize R productive D evelopm ent .............................................................. .....................2 1

2 TRANSCRIPT PROFILING BY 3'-UTR SEQUENCING RESOLVES EXPRESSION
O F G EN E FA M IL IE S ..................................................... ........................ ...............23

Intro du action ................... .......................................................... ................ 2 3
R esu lts...................................................................................................... 2 7
Construction of a 3'-cD N A Library ..............................................................................27
D ata A ssem bly and A nalysis..................................................................... ............... 28
Analysis of 3'-UTR Profile Reveals a Dynamic Range of Expression ...........................29
Distinguishing Gene Family M embers....................................................................... 30
Evaluation of Differential Expression between Multiplexed Sub-libraries ..................31
Resolution of Near-Identical Transcripts by Polymorphisms ......................................32
Validation of SNPs and Homopolymer-Based Polymophisms................. .............. 33
D iscu ssion .. ............. ...........................................................................................34
Future Prospects .............................................. 40
C onclu sions.......... .........................................................4 1
M materials and M methods ...................................... .. .......... ....... ...... 41
Plant M materials .............................. ............................................41
Sub-library Preparation and Sequencing ................................................. .............. 42
D ata A naly sis................................... ............... ........... ...............43
Real-time RT-PCR Analysis for Validation of 454 Data ........... ...........44









3 EXPRESSION PROFILING OF DEVELOPING MAIZE OVARIES USING
MICROARRAYS AND SEQUENCING OF 3'-UTRS .............. ............ .....................57

In tro d u c tio n ....................................................................................................................... 5 7
R results and D iscu ssion ................. ................................................................ ...... ..... ........... .. 6 1
M icroarray Data Analysis and Clustering .............. ....... ....... ............. ............... 61
Identifying Clusters of Co-Expressed Genes .................................... ............... 62
Co-Expression of Genes Related to Common Metabolic Pathways ............................63
Sequence-Based Analyses by 3'-UTR Profiling................. ............................................65
Comparison of Microarray and Quantitative 3'-UTR Profiles .............. ...................66
C onclusions.....................................................................68
M materials and M methods ................................... ... .. .......... ....... ...... 68
Plant M material ............................................68
RNA Extraction and Target Labeling................................................. 69
Microarray Slide Preparation and Target Hybridization ............ ...............................69
M icroarray D ata A nalysis....................... ................................................. ............... 71
Cluster Analysis using Orthogonal Polynomials.................................. ............... 72
Construction of a 454 3'-U TR library...................................... ...................... ............ 72

4 C ALLOCATION AND USE IN DEVELOPING MAIZE FEMALE FLORETS ................90

In tro d u ctio n ............. ...... ......... ....................................................................................... 9 0
Results ......... ...... .. .... .......... ..............94
Staging of Pre-Pollination Floral Developm ent ......................................... .......... .....94
Pre-Pollination Carbon Allocation and Sucrose Use in Individual Floral Organs..........95
Temporal and Spatial, IVR2-Based Sucrose Use in Developing Female Florets............96
Post-Pollination C Accumulation in Pedicels............... ...... ..............98
Co-Expression of Genes Related to C Sink Development .......... .........................98
Post-Pollination Lignin Biosynthesis in Pedicels............................................... 99
D iscu ssio n ................... ...................9...................9..........
Conclusions..........................................102
M materials an d M eth od s ................................................................................ ................... 103
Plant M material and Sam pling ........................................................................... 103
Quantification of mRNAs by Real-Time RT-PCR .............. .................................. 103
Soluble Sugar Extraction and Quantification.............................. ......... ............. 104
Assay for Soluble Invertase Activity............... ........... ............ ........... ......... 105
GC/M S Quantification of M etabolites ........................................ ....... ............... 105
Phloroglucinol-HCL Staining ......... .............. .. .................................. 106

5 SU M M A R Y ......... ................................................................................... ............. 119

APPENDIX

PERTURBATION BY DROUGHT AND DISRUPTED VP1-BASED ABA SENSING..122

B IO G R A PH IC A L SK E T C H ......................................................................... ... ..................... 149




6









LIST OF TABLES
Table page

2-1 Summary statistics of a two-sample multiplex 3'-UTR library. ......................................46

2-2 CesA family 3'-anchored consensus sequences. ....................................................47

2-3 Best-match cDNAs and associated annotations (BLASTN) for consensus sequences
showing highly significant differences in transcript abundance between wild-type
and vpl mutant drought-stressed ovary sub-libraries ............................. ...................48

2-4 Polymorphisms detected by W22 3'-anchored 454 sequence reads. ................................49

3-1 Number of genes associated with each co-expression cluster and percentage having
significant differences in expression or large-fold changes during development..............74

3-2 Annotated genes from clusters 1 and 5 that showed significant differences in
expression (FDR 20%) or large fold-change (>2) during development ..........................75

3-3 Annotated genes from cluster 7 that showed significant differences in expression
(FDR 20%) or large fold-changes during development...............................................76









LIST OF FIGURES


Figure page

2-1 Comparison of shotgun versus 3'-anchored approaches for using 454-sequences as a
platform for transcript profiling. ................................. ..................................... 50

2-2 Strategy for 3'UTR profiling by 454-sequencing ................................... .................51

2-3 Distribution of tag lengths for a simulated Mspl digest of 70,000 3'-oriented ESTs
from B73 maize (maize full length cDNA project [www.maizecdna.org]) ...................52

2-4 A graphic presentation of the quantitative 3'UTR profile representing 11,559
consensus sequences that matched cDNAs in the 2-sample multiplexed library ..............53

2-5 Distribution of transcripts (log-log scale) in selected functional classes from Figure
4B and read count for each. Resolution of individual gene family members was
enabled by the specificity of the 3'U TR ..................................... ..................... ............. 54

2-6 Validation of technical accuracy for determining differences in transcript abundance
based on read number .................... ......................... .......... 55

2-7 A 3' sequence polymorphism resolved nearly-identical Auxin Repressed Dormancy
Associated paralogs with differences in mRNA abundance............................................56

3-1 Experimental design used to compare transcriptional profiles during maize ovary
develop ent ................................................................................77

3-2 A heat map generated by a two-way, hierarchical clustering of expression profiles for
856 genes that showed significant changes in transcript abundance during
develop ent (FD R 10% ) ........................................ .................. ... ..................... 78

3-3 Annotated genes from hierachical clusters 1 and 2 (Figure 3-2) that showed either
decreased or increased expression during maize ovary development were classified
into functional categories based on Gene Ontologies.......................................................79

3-4 Co-expression clusters of genes were analyzed after fitting expression patterns
during development to orthogonal polynomials ..................................... .................80

3-5 Annotated genes in clusters 1, 5, and 7 were grouped by functional processes using
G en e O ntolog ies..................................................... ................ 82

3-6 Experimental design and sequencing results for a 12-sample, multiplexed, 3'-UTR
lib ra ry ................... ............................................................ ................ 8 3

3-7 The 3'-UTR consensus tags were compared to the 70-mer microarray probe
sequences to determine extent of overlap. .............................................. ............... 84









3-8 Selected genes with high transcript abundances in the 3'-UTR library that were used
to validate m icroarray expression data ........................................ ......................... 85

3-9 Quantitative 3'-UTR profiles for cell-wall-related genes with matches to array probes
in clusters 1, 2, 5, and 7 (see Figure 3-4)...................................................... ............... 86

3-10 Transcript profiles of four Phenylalanine ammonia lyase (PAL) gene family
members were compared by microarray- and sequence-based methods.........................87

3-11 Resolution of quantitative transcript profiles by the 3'-UTR sequencing approach for
near-identical mRNAs that matched a single array probe ...........................................88

4-1 Definition of developmental stages for pre-pollinated maize female florets by
physical growth parameters, anatomical features, and expression of a ZAG2
m olecu lar m ark er ......... ............... ............................................................. 107

4-2 Partitioning of carbon and water to dry weight and fresh weight accumulation among
individual organs of developing maize female florets............. .... ...............108

4-3 Spatial and temporal expression of soluble acid invertases and changes in sugar
composition in maize florets just prior to pollination ............. ..... .. ............... 109

4-4 Activity of the vacuolar invertase, IVR2, is associated with turgor-based expansion
in rapidly elongating silks ............................................... .. ......... ........ .... 110

4-5 Diurnal changes in sucrose and hexose levels in rapidly expanding silks on the day
of their em ergence from husks....................................................................... .............111

4-6 Relative abundance of mRNAs for vacuolar and cell-wall inverase isoforms in maize
female florets during the pre- to- post-pollination period ...............................................112

4-7 Carbon deposition and relative water content in maize ovaries and pedicels during
post-pollination develop ent........................................................................... ..... .... 113

4-8 Sugar composition of maize ovaries and pedicels during early post-pollination
d ev elo p m en t ...................................... .................................................. 1 14

4-9 Co-expression profiles for genes related to C and N metabolism during ovary and
pedicel development from genome-wide microarray analyses ............. ... .................115

4-10 Co-expression of genes related to lipid metabolism and abundance of lipid-based
m etabolites in developing m aize ovaries ...................................................................... 116

4-11 Expression profiles for genes involved in phenylpropanoid biosynthesis in the
developing m aize ovary-plus-pedicel. ................................................................... ..... 117

4-12 Accumulation of lignin precursors and pedicel-localized staining of lignin in the
developing m aize ovary ...................................................... .... .. ............ 118









LIST OF ABBREVIATIONS


UTR

PCR

EST

SAGE

MPSS

SNP

QTL

MAGI

DBP

DAP

ASI

ANOVA

FDR

BIC

LSMEANS

PTT

C

N


Untranslated region

Polymerase chain reaction

Expressed sequence tag

Serial analysis of gene expression

Massively parallel signature sequencing

Single nucleotide polymorphism

Quantitative trait loci

Maize Assembled Genomic Islands

Days before pollination

Days after pollination

Anthesis-Silking Interval

Analysis of Variance

False Discovery Rate

Bayesian Information Criterion

Least square means

Parts per ten thousand

Carbon

Nitrogen









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

REGULATION OF SINK STRENGTH IN DEVELOPING MAIZE FLORETS:
IMPLICATIONS FOR SEED SET AND GRAIN YIELD

By

Andrea L. Eveland

May 2008

Chair: Karen E. Koch
Major: Plant Molecular and Cellular Biology

The pre- and early post-pollination phases of maize (Zea mays L.) reproductive

development are critical for seed set and subsequent grain yield. During this time, the plant is

especially sensitive to abiotic stresses, such as drought, which can reduce pollination efficiency

or lead to kernel abortion. A key determinant of reproductive success is carbohydrate allocation

and use in the developing female inflorescence, which is often disrupted by stress.

Understanding the mechanisms that underlie regulation of sink strength during normal

progression of maize floral development is thus central to improving seed set under adverse

environmental conditions.

In this work, we tested the hypothesis that expression of genes related to specific metabolic

or regulatory pathways would change in association with carbohydrate allocation and use during

silk exsertion and pollination in maize. Individual stages of pre- and early post-pollination maize

female florets were characterized based on physical characteristics and expression of a molecular

marker for development. Subsequent analyses revealed a shift in sink strength, as approximated

by dry and fresh weights, during the pollination period from rapidly expanding silks and

subtending floral structures lemmaa, palea, and glumes) to the developing ovary and pedicel.

This shift coincided with isoform-specific expression of sucrose metabolizing invertases, which









provide hexose substrates essential for turgor-based expansion prior to pollination and also for

post-pollination growth of symplastically-isolated filial tissues. In addition, accumulation of

sucrose and hexoses in the pedicel and ovary, respectively, indicated that invertases could

contribute to spatial regulation of cell expansion and differentiation during development.

We used a genome-wide transcript profiling approach to determine whether co-expressed

genes were related to specific functional processes or associated with relevant metabolic

pathways during pre- and early post-pollination ovary development. A gene-specific, sequence-

based, 3'-UTR profiling strategy was developed and tested in parallel to microarray analyses.

We resolved co-expression profiles for key genes related to nitrogen and amino acid metabolism,

carbohydrate metabolism, lignin biosynthesis, and cell growth during ovary development.

Transcript profiles were combined with sugar and metabolite analyses and fresh/dry weight

quantifications to further support relevance of key sets of co-expressed genes during ovary sink

establishment in maize. Results from this study provide evidence for testable roles of such genes

in kernel set.









CHAPTER 1
LITERATURE REVIEW

Introduction

In plants, maintaining a delicate balance between photosynthesis, the energy demands of

growing sink organs, and storage of essential reserves is central to whole plant function and

developmental progression. Reproduction is a costly process requiring large amounts of

photosynthetic resources for floral organ expansion (Woodson and Wang, 1987; Makela et al.,

2005; Borras et al., 2007), respiration (Bustan and Goldschmidt 1998), and/or grain filling (Cruz-

Aguado et al., 1999; Maitz et al., 2000; Weschke et al., 2003). Therefore, careful regulation of

carbon assimilation and allocation is necessary for timing and maintenance of reproductive

growth. Sucrose import into developing sink structures involves sensing of carbohydrate status

in both source and sink cells (Koch, 1996; Chiou and Bush, 1998; Stitt et al., 2007) as well as

intercellular signaling based on steady-state pools and flux of sugars through membrane

transporters (Lalonde et al., 1999; Barker et al., 2000; Vaughn et al., 2002; Sauer, 2007). Such

signals can influence sucrose export from source organs or regulate sink strength at the site of

phloem unloading. Perturbation at either end alters the source/sink balance (Paul and Foyer,

2001; Borras et al., 2003; McCormick et al., 2008).

Adjustment of source-to-sink resource allocation has a substantial impact on fruit set and

grain yield and thus is centrally important to crop improvement programs. In maize (Zea mays

L.), the pre- and early post-pollination phases of female reproductive development are critical for

sufficient pollination and seed set (Zinselmeier et al., 1995; McLaughlin and Boyer 2004a).

During this time, female reproductive tissues are most sensitive to abiotic stresses such as

drought (Westgate and Boyer, 1985; Andersen et al., 2002; McLaughlin and Boyer 2004b) and

shade (Setter et al., 2001; Zinselmeier et al., 2002). Such stresses negatively affect









photosynthesis and thus alter whole plant carbohydrate status (Roitsch, 1999; Trouverie and

Prioul, 2006). Resulting losses in grain yield are devastating on a world-wide scale (Barnabas et

al., 2008), yet the processes underlying the pre- and early post-pollination phases of maize

reproductive development have received comparatively little attention at the molecular/metabolic

level.

Source/Sink Relations and Sucrose Metabolizing Enzymes

Sucrose is synthesized in source-leaf mesophyll cells and in most plant systems transported

through the phloem to developing sink organs by "mass flow." This is a function of a turgor

pressure gradient due to sucrose concentration at both source- and sink-ends of the phloem

(Lalonde et al., 2004; Carpaneto et al., 2005). An organ's 'sink strength' is dependent upon a

number of factors related to this turgor gradient, including localized osmotic status and the

activity of sucrose-cleaving enzymes (Sturm and Tang, 1999; Roitsch et al., 2000). Recent work

has shown that specific quantitative trait loci (QTL) for growth in maize are associated

predominantly with carbohydrate metabolic enzymes involved in source/sink relations rather

than photosynthetic genes (Pelleschi et al., 2006). Invertases, which catalyze the irreversible

cleavage of disaccharide sucrose into glucose and fructose, are implicated as major determinants

of sink strength in the early development of importing structures (Koch, 1996, 2004; Strum

1999; Roitsch, 1999; Weschke et al., 2003; McLaughlin and Boyer, 2004b; Schaarschmidt et al.,

2007). The conversion of one molecule of sucrose to two hexoses regulates osmotically-active

solutes in sink tissues and can affect the descending turgor gradient for sucrose in the phloem.

Carbon assimilates can move into a developing sink either symplastically through

interconnecting plasmodesmata or via apoplastic phloem unloading into the cell wall space.

Invertase is active at both of these interfaces with isoform-specific compartmentalization

(Tymowska-Lalanne and Kreis, 1998; Sturm, 1999; Godt and Roitsch, 2006). Inside the cell,









there are two types of soluble invertases that can cleave symplastically-delivered sucrose. Of

these, the acidic vacuolar invertases are typically most active. However, recent findings support

specific roles for the elusive cytoplasmic, alkaline invertases (Flemetakis et al., 2006; Lou et al.,

2007; Vargas et al, 2007). Apoplast-localized cell wall invertases, also optimally active at acidic

pH, catalyze the cleavage of sucrose unloaded from the phloem into hexoses which are

transported into symplastically-isolated sinks such as the developing embryonic tissues (Cheng

et al., 1996; Sherson et al., 2003).

Invertases are encoded by small gene families, members of which are spatially and

temporally regulated (Tymowska-Lalanne and Kreis, 1998; Fridman and Zamir, 2003; Cho et al.,

2005; Huang, 2006). Only two vacuolar isoforms have been described in Arabidopis

(Tymowska-Lalanne and Kreis, 1998), tomato (Fridman and Zamir, 2003), poplar (Bocock et al.,

2008), and maize (Xu et al., 1996), suggesting evolutionary conservation among diverse species.

Vacuolar and cell wall-bound invertases are implicated not only in the establishment of sink

strength in developing organs, but also in the generation of hexose-based signals, providing cues

for developmental and metabolic processes (Koch, 1996, 2004; Sturm and Tang, 1999; Rolland

et al., 2006). Hexose-based signaling is discussed further in the sugar sensing/signaling section

of this review.

Sink development is mediated by the temporal expression patterns of sucrose-cleaving

enzymes. During initial phases of sink establishment, vacuolar invertases generate essential

hexoses for turgor-driven cell expansion (Sturm and Tang, 1999; Koch, 2004), whereas

apoplastic invertase activity becomes the predominant source of hexose-based signals during

expansion through the cell differentiation phase. The reversible activity of sucrose synthase,

which cleaves sucrose into fructose and UDP-glucose, becomes important during the maturation









phases of sink development. In general, a high level of hexoses relative to sucrose induces cell

division and expansion, whereas sucrose accumulation is associated with differentiation and

maturation (Zinselmeier et al., 1995; Patrick, 1997; Wobus and Weber, 1999; Winter and Huber,

2000). Consequently, acid invertases, and in particular vacuolar invertases, have been identified

as central to turgor-mediated growth in expansion sinks such as floral organs (Woodson and

Wang, 1987; Xu et al., 1996) and roots (Sergeeva et al., 2006; Huang, 2006).

The essential contribution of sucrose metabolizing enzymes to sink establishment and

growth has been described for reproductive development and is discussed in the following

section. Similar modes of action for such enzymes have been revealed in plant-microorganism

interactions. Symbiotic relationships such as mycorrhizal arbuscules and root nodules (Blee and

Anderson, 2002) as well as parasitic associations of nematode induced syncytia (Hofmann et al.,

2007) and Agrobacterium infection (Wachter et al, 2003), induce carbohydrate demands that

mimic reproductive structures. Whole plant source/sink adjustments result accordingly.

Sugar Sensing and Signaling

Central to photosynthate allocation is nutrient status and the mechanisms by which cells

sense this status. Sucrose, and to a larger extent the hexoses generated through its cleavage, can

act as signaling molecules in a variety of developmental processes (Koch, 1996, 2004;

Smeekens, 2000; Rolland et al., 2006). Therefore, sugar status can modulate the expression of

specific genes based on photosynthate availability and/or environmental cues. Sucrose and

hexose signals can reportedly be generated either by concentration or by flux through sugar-

specific sensors and/or transporters. Long-distance sucrose transport depends on temporal and

spatial regulation of a sucrose transporter (SUT) gene family (Sauer, 2007), members of which

have been identified as pivotal to long-distance transport in cereal crops such as maize (Aoki et

al., 1999) and rice (Scofield et al., 2007). Certain SUT isoforms have been shown to regulate









phloem loading and unloading in response to changes in sucrose concentrations (Lalonde et al.,

1999, 2004; Barker et al., 2000; Vaughn et al., 2002).

In yeast, membrane proteins with homology to hexose transporters have been found to

sense glucose and relay the signal to hexose-responsive genes (Ozcan et al., 1998; Rolland et al.,

2002). In plants, monosaccharide/H+ symporters catalyze the uptake of hexoses from the

apoplast (Sherson et al, 2003; Weschke et al., 2003; Buttner, 2007) or into the vacuole (Wormit

et al., 2006) and may be involved in the transduction of hexose-specific cellular signals.

However, an actual sensor function for these transporters in plants has not been reported to date.

Hexoses may be sensed as substrates for hexokinase, where the phosphorylation event

leads to a signal cascade and subsequent changes in gene expression. The hexokinase-dependent

mechanism couples signaling with hexose-phosphate flux through glycolysis (Harrington and

Bush, 2003; Moore et al, 2003). Recent work has identified a set of interacting proteins that

form a complex with hexokinase in the nucleus (Cho et al., 2006), and the association of

hexokinase with actin filaments (Balasubramanian et al., 2007), however the mechanisms for

signaling are still unknown. Alternately, sensing via a hexokinase-independent pathway is

possible (Xiao et al., 2000). Recent work has described the role of certain G-protein coupled

receptors in direct sensing and transduction of the sugar signal in yeast (Lemaire et al., 2004) and

in plants (Huang et al., 2006). In addition, sucrose has also been shown to induce

phosphorylation (Nittylae et al., 2007) and thus stimulate signal cascades. These sensing

mechanisms enable intercellular signaling that can modulate changes in gene expression

according to the amount of photosynthate available.

Sucrose-cleaving enzymes are among such sugar-regulated gene products. In maize, both

invertases and sucrose synthases show isoform-specific induction (Xu et al., 1996) in a









"feast/famine" association. The two vacuolar invertase isoforms, IVR1 and IVR2, are up-

regulated in a reciprocal manner by sucrose availability and sucrose starvation, respectively.

Similarly, sucrose synthase isoforms also display concentration-dependent induction by sugars

(Xu et al., 1996). This reciprocal mode of regulation has been identified in pairs of vacuolar

invertase isoforms in Arabidopsis (Huang, 2006), poplar (Bocock et al., 2008), tomato (Godt and

Roitsch, 1997) and rice (Huang, 2006). Whether this dual sub-functionalization is evolutionarily

conserved or an example of convergent evolution is currently unknown. However, such

differential responses to sugar availability or starvation based on isoform specificity enable fine

adjustment of photosynthate in response to environmental perturbation or developmental cues

(Osuna et al., 2007; Smith and Stitt, 2007).

Signals based on carbohydrate status can regulate plant growth in response to

environmental or developmental stimuli. Sugar availability can influence cell division,

expansion, or proliferation (Nicolai et al., 2006; Rook et al., 2006; Kojima et al., 2007; Smith

and Stitt, 2007). Glucose and/or sucrose can induce specific D cyclins, thus coupling cell cycle

regulation to carbohydrate status (Riou-Khamlichi et al., 2000; Menges et al., 2006). Recent

work has also linked sugar sensing and signaling to regulation of cell wall expansion (Lee et al.,

2007; Li et al, 2007). Carbohydrate availability can also regulate gene expression based on its

relative abundance relative to nitrogen (Coruzzi and Bush, 2001; Cooke et al., 2003; Fritz et al.,

2006) or to phosphate (Karthikeyan et al., 2007). Sugar signals based on starch metabolism and

sugar starvation also affect gene expression, aspects of which have been supported by global

assessment of diurnal regulation (Gibon et al., 2004; Blasing et al., 2005; Smith and Stitt, 2007).

In addition, evidence for signaling by sugars such as trehalose has revealed developmental and

regulatory roles. Trehalose metabolism has been shown to influence embryonic development









(Eastmond, 2002), reproductive architecture (Satoh-Nagasawa et al., 2006), and ABA-mediated

responses (Avonce et al., 2004; Ramon et al., 2007), possibly through sensing and signaling of

trehalose itself or of its phosphorylated form.

Hormone and Sugar Signaling Networks

The regulation of gene expression by sugars can be viewed as analogous to phytohormone-

based interactions. Therefore, it is not surprising that extensive crosstalk between sugar and

hormonal pathways has been described for many developmental and metabolic processes

(Gazzarrini and McCourt, 2001; Leon and Sheen, 2003; Gibson, 2004; Rolland et al., 2006).

The biosynthesis of phytohormones critical to certain developmental cues can be dependent on

carbohydrate status (Cheng et al., 2002). Alternatively, differential responses of hormone-

modulated genes may rely on sugar availability (Huang, 2006) and can thus be influenced by

whole plant source/sink relations. Studies in Arabidopsis have revealed an extensive interface

between ABA and sugar regulatory networks (Finkelstein and Gibson, 2002; Li et al., 2006).

Most of these interactions were identified by mutant loci redundant for sugar- and ABA- sensing

and signaling phenotypes (Arenas-Huertero et al., 2000; Huijser et al., 2000; Laby et al., 2000;

Brocard et al., 2002; Arroyo et al., 2003; Brocard-Gifford et al., 2004). In addition, examples of

converging sugar- and ABA-based signals in the regulation of specific transcription factors have

been described in maize (Niu et al., 2002) and in grape (Cakir et al., 2003).

Much of the work to date in sugar sensing and signaling has used Arabidopsis thaliana as a

model system due to the ease of phenotypic screening on sugars or other growth regulators.

Although these studies have focused primarily on seed germination, some sugar- and hormone-

based interactions have been described in whole plant source/sink relations. In particular, the

ABA has been implicated in regulation of leaf senescence (Pourtau et al., 2004), modulation of









phloem unloading ((Peng et al., 2003; Oliver et al., 2007), and transitory starch biosynthesis

(Rook et al., 2001).

Regulation of invertase-mediated carbon allocation includes ABA-based signals as well as

those of various other phytohormones. Studies have shown that acid invertase responds to ABA

(Setter et al., 2001; Kim et al., 2000a; Pan et al., 2005; Huang, 2006), ethylene (Linden et al.,

1996; Huang, 2006), cytokinins (Setter et al., 2001; Lara et al., 2004), and auxin (Yun et al.,

2002; Long et al., 2002). Since sucrose metabolism and use are essential aspects of

developmental growth, it is not surprising that multiple, developmentally-induced hormone

signals are involved. Invertase activities may also be mediated by hormone ratios and

antagonisms. For example, ethylene and ABA are antagonistic, yet a high level of ABA tends to

promote ethylene production (Gazzarrini and McCourt, 2001; Leon and Sheen 2003).

Hexose-based signals generated by invertase activity can affect the extent of hormonal

responses either through biosynthesis or signal transduction. Glucose has been found to promote

ABA biosynthesis via up-regulation of genes encoding key enzymes in this process (Cheng et al.,

2002). This suggests potential feedback regulation of invertase expression that could operate on

ABA levels. In Arabidopsis, sucrose-based induction of AtvaclNV], the putative functional

ortholog of ZmlVR2, is enhanced by ABA (Huang, 2006). Alternatively, AtvaclNV]

transcriptional induction by ABA is dampened by sucrose. In addition, isoform-specific

regulation of invertases extend beyond sugar and hormonal signals (Roitsch and Gonzalez, 2004;

Huang et al., 2007). Mechanisms of regulation include mRNA stability (Yun et al., 2002),

turgor-based sensing via cell wall associated kinases (Kohorn et al., 2006), and sequestration in

pre-vacuolar secretary vesicles (Rojo et al., 2003; Huang et al., 2007).









Maize Reproductive Development

The activities of sucrose-cleaving enzymes during reproductive development are central to

promoting optimal pollination and subsequent seed set in cereal crops such as maize. Both

vacuolar (Xu et al., 1996; Andersen et al., 2002; McLaughlin and Boyer 2004b; Qin et al., 2004)

and cell wall invertases (Cheng et al., 1996; Taliercio et al., 1999; Kim et al., 2000b; Chourey et

al., 2006) have been implicated in early maize floral development, and especially ovary

expansion. The developing maize ovary is composed predominantly of maternal tissue and

receives sucrose primarily via symplastic transfer from the phloem ending in the pedicel region

(Fisher and Wang, 1999).

During the pre- and early post-pollination phases of reproductive development, insoluble

invertase is localized to the area beneath the nucellus in the upper pedicel tissues (Cheng et al.,

1996; Kladnik et al., 2004; McLaughlin and Boyer 2004b). Soluble invertase is expressed

throughout the nucellus and pericarp tissues (Andersen et al., 2002; McLaughlin and Boyer

2004b). Upon embryonic growth, a cell wall invertase, INCW2, becomes the dominant isoform,

providing sucrose across the apoplastic barrier between maternal tissue and the developing

embryo (Cheng et al., 1996; Kladnik et al., 2004). Together, activities of specific invertase

isoforms are responsible for maintaining an effective sucrose gradient to promote ovary growth

and seed set. Accordingly, a mutation in the INCW2 gene resulted in a small kernel, or

miniature, phenotype (Cheng et al., 1996) and drought-induced repression of vacuolar invertases,

IVR1 and IVR2, caused inefficient pollination and/or kernel abortion (Andersen et al., 2002;

Boyer and Westgate, 2004; McLaughlin and Boyer, 2004b).

Maize is a monoecious plant with male tassels borne terminally from the differentiated

shoot apical meristem and female inflorescences, or ears, laterally in the axils of leaves. In the

female florets, the ovary, composed of three fused carpels, is subtended by a palea and lemma









(outer whorl structures with similarities to prophylls and bracts, respectively) and an additional

set of bracts called glumes (Clifford, 1987; Ambrose et al., 2000; Whipple et al., 2004). An

inner whorl of petal-like structures, or lodicules, is present only in the male florets. Outgrowths

from two of the three fused carpel units form the silk, a rapidly elongating stigmatic structure

which is unique to the maize flower. Silk expansion is critical to pollination efficiency and is

directly correlated with biomass accumulation (Borras et al., 2007). Previous work has identified

soluble invertase (Xu et al., 1996) and expansion (D Cosgrove, personal communication)

activities in growing silks and suggests their roles in turgor-mediated expansion. However, few

studies have investigated the regulation of carbohydrate allocation and use in silks.

A central aspect of the pollination process and ovary receptivity depends on coordination

of silk emergence from husk leaves with anther dehiscence, also known as the Anthesis Silking

Interval (ASI) (Bolanos and Edmeades, 1996; Borras et al., 2007). Initial studies by Westgate

and Boyer (1985) showed that low water potential due to soil drying had the most dramatic effect

on silk turgor and growth as compared to turgor in leaf, stem, and root. These findings indicate

that osmotic adjustment in the silk, based on whole plant source/sink balance, has a significant

impact on reproductive success in maize.









CHAPTER 2
TRANSCRIPT PROFILING BY 3'-UTR SEQUENCING RESOLVES EXPRESSION OF
GENE FAMILIES

Introduction

Functional analysis of plant genomes requires methods for resolving differential

expression of closely-related genes. The ability to distinguish between paralogs (e.g. gene

family members) and alleles on a genome-wide scale is key to understanding the genetic basis of

quantitative traits in diverse plant populations. Genes with extensive sequence similarity may

comprise a significant portion of a given transcriptome. Among maize inbreds, for example,

90% of the alleles are polymorphic (Wright et al., 2005) and approximately one-third of maize

genes are tandemly duplicated (Messing et al., 2004). The extent of these sequence similarities

in maize and other complex genomes pose a clear challenge to delineation of gene-specific

function.

Differential expression of related, duplicated genes has been linked to functional diversity

within species (Gu et al., 2004), and sub-functionalization can provide a basis for genome

evolution (Moore and Purugganan, 2005; Emrich et al., 2007b). The impetus for resolving

expression among paralogs is further motivated by the extent of polyploidy, which is estimated

to affect 50-80% of angiosperm species, including maize, wheat, cotton, and other important

crops (Osborn et al., 2003; Blanc and Wolfe, 2004). Moreover, allele-specific differences in

gene expression that contribute to variations in phenotype are widespread in both animal

(Cowles et al., 2002; Yan et al., 2002) and plant species (Cong et al., 2002; Guo et al., 2004).

Accordingly, transcript profiling has been adapted as a means of appraising quantitative traits

(Schadt et al., 2003; Borevitz and Chory, 2004). Association mapping using eQTLs has

identified candidate genes for important traits in tomato (Baxter et al., 2005) and poplar (Street et

al., 2006). In addition, comparing allele-specific expression among inbred parental varieties and









F hybrids of reciprocal crosses has revealed deviations from dosage dependency and enabled

analysis of imprinting in maize endosperm (Guo et al., 2003). The ability to resolve individual

transcripts with similar sequences and quantitatively compare their expression is thus central to

addressing questions in functional genomics and defining genetic contributions to hybrid vigor

(Birchler et al., 2003; Swanson-Wagner et al., 2006; Springer and Stupar, 2007).

Despite rapid advances in expression profiling techniques, capacity to distinguish among

closely-related transcripts on a genome-wide scale remains a challenge. In microarray analyses,

cross-hybridization of similar transcripts to a given oligonucleotide probe may confound

expression of individual genes. With sequencing of several genomes complete (e.g.

Arabidopsis, rice, and poplar), whole-genome tiling arrays have allowed unbiased interrogation

of the transcriptome (Yamada et al., 2003; Mockler and Ecker, 2004). These platforms have

successfully uncovered discrete polymorphisms and alternate splice sites, but depend on fully-

sequenced genomes and/or are limited to sequence variants present on the array (typically

derived from a single reference strain). In addition, quantitative measures of gene expression are

limited by probe-specific hybridization efficiencies.

The emergence of high volume, short-read sequencing technologies has increased

resolution for quantitative transcriptome analysis in organisms for which complete genomic

sequence is available. Advances in serial analysis of gene expression (SAGE) have opened

transcript profiling to unbiased sampling and quantitative analysis of gene expression (Saha et

al., 2002; Bao et al., 2005). Although limited in throughput, the sequencing of novel cDNAs

following 3' extension and amplification of short SAGE tags has been successfully utilized for

gene discovery (Chen et al., 2004).









Alternatively, genome-wide profiling by massively parallel sequencing, MPSS (Meyers et

al., 2004; Jongeneel et al., 2005) and the more recent Solexa 1-G technology (Barski et al.,

2007), has facilitated detection of rare transcripts and comprehensive cataloging of non-coding

RNAs (Lu et al., 2005; Nobuta et al., 2007). The capacity of these massively parallel approaches

to generate millions of short sequence tags can enable reliable, cost-effective coverage of the

transcriptome. However, in genomes where sequence information is fragmentary, the short

length of these reads (17 to 36b) provides a limited capability for unambiguous gene assignment.

Likewise, near-identical transcripts are difficult to discern with short-sequence reads, even in a

fully-sequenced genome. Estimates from rice and Arabidopsis MPSS libraries indicate that

approximately 11% of signature sequences matched multiple target sites in the genome (Nobuta

et al., 2007).

Longer-read lengths are achieved with 454-based pyrosequencing, initially described by

Marguiles et al. (2005), and more recently implemented as a platform for transcript profiling

(Emrich et al., 2007a; Weber et al., 2007). A key advantage of the longer reads generated by this

technology is greater capability for gene annotation and discovery in both sequenced and non-

sequenced genomes. A recent upgrade of the Genome Sequencer FLX system (Harkins and

Jarvie, 2007) has extended average read lengths to >200 bases. A tradeoff for obtaining more

informative read lengths is a lower depth of sequencing achieved with 454 compared to short-

read technologies (e.g. Solexa 1-G). Therefore, one method for improving the efficiency of 454-

based transcript profiling is to anchor 454-reads to unique sites near the 3' ends of expressed

sequences in order to 1) reduce the number of reads necessary to identify individual mRNAs and

2) maximize recovery of gene-specific polymorphisms. The 3'-untranslated region (UTR) is rich

in single-feature polymorphisms that distinguish closely-related transcripts (Bhattramakki et al.,









2002; Vroh Bi et al., 2006). The specificity of 3'UTR sequence reads thus allows effective

annotation of individual mRNAs without assembly of complete cDNAs (Figure 2-1).

Here we present a strategy that harnesses the specificity and information content of the

3'UTR in a long-read, 454-based sequencing approach to transcript profiling. A key to this

method is the use of 3'-anchored sequence reads long enough for unambiguous identification of

closely-related transcripts. By targeting the 3'UTR of mRNAs, an unprecedented resolution is

achieved for gene- and allele-specific transcripts, even for genomes that are only partially-

sequenced or lack extensive expressed sequence tag (EST) coverage. In addition, detection of

haplotypes containing multiple polymorphisms is facilitated by the longer-read length. These

components of the transcriptome are thus opened to quantitative analysis beyond that currently

accessible with short-tag sequencing technologies. In the present work, maize provides an ideal

system to assess our 3'-anchored strategy, because the genome is rich in genetic complexity from

extensive gene duplication (Messing et al., 2004; Messing and Dooner, 2006) and currently is

not fully sequenced.

In this study we introduce a 3'UTR profiling method that allows quantitative analysis of

gene-specific expression on a genome-wide level, here using mutant and wild-type maize

ovaries. Concurrent sequencing of multiple mRNA samples was enabled by use of a

multiplexing strategy. Results provided quantitative expression profiles with read output evenly

distributed between samples. The frequency of 3'-anchored sequence reads aligning to a given

cDNA was used to quantify mRNA abundance and to measure differential gene expression. The

long read lengths, combined with the specificity of the 3'UTR, were sufficient to distinguish

individual members of a previously-characterized gene family as well as provide quantitative

comparisons of closely-related transcripts that matched unique maize expressed sequence tags









(ESTs) or assembled cDNAs. In addition, insertion deletion (indel) polymorphisms were readily

detectable by this method and resolved nearly-identical paralogous gene products.

Results

Construction of a 3'-cDNA Library

We synthesized 3'-anchored cDNA template libraries to generate gene-specific sequence

reads by 454 using the protocol shown in Figure 2-2. Concurrent sequencing of up to 16

individual sub-libraries is enabled by incorporation of a three-base multiplex key in the A-

adaptor (Figure 2-2A). By using a subset of 16 three-base keys, we could detect single-base

errors in the multiplex key. Addition of a fourth base to the multiplex key would enable up to 64

unique combinations with error detection, thus enhancing the number of concurrently sequenced

sub-libraries. Each 3'UTR sub-library was constructed from total RNA (Figure 2-2B) using a

modified, biotinylated 454-B adaptor that incorporates an oligo(dT) tail for priming cDNA

synthesis from poly(A) RNA. Following second-strand cDNA synthesis, biotinylated cDNAs

were bound to steptavidin beads, purified by magnetic pull down, and digested with MspI to

generate 3'-cDNA fragments with 2-base (CG) overhangs. Specific multiplex A-adaptors were

then ligated to the purified 3' fragments. A detailed description is provided in Materials and

Methods. Mspl was selected based on simulated digests of 70,000 3'-orientated ESTs of maize

(Figure 2-3) from the maize full-length cDNA project (www.maizecdna.org).

Predicted tag lengths were used to assess the proportion of 3'-enrichment in comparison to

rice 3'UTRs. While the expected size distribution ofMspl-digested cDNA fragments is optimal

for the GS20 read length (-100 b), the longer reads generated by the 454-FLX instrument (-250

b average) would likely extend through the 3'UTR of many transcripts. If so, the number of FLX

reaction cycles could be configured to optimize average read length (Harkins and Jarvie, 2007).

Although a single Mspl digest was used in this study, potential increases in coverage of 3'-ends









could be achieved by combining digests made with compatible restriction enzymes (e.g. TaqI,

Maell). This would further improve coverage when used in conjunction with the longer read

lengths and enhanced read output achieved by 454-FLX technology.

To test the 3'UTR profiling strategy, we sequenced 3'-anchored cDNA sub-libraries

prepared from immature ovaries of isogenic viviparous-1 mutant and wild-type maize plants in a

W22 inbred genetic background. Prior to RNA sampling, plants were subjected to a drought

stress treatment (Materials and Methods). VP1 is a transcription factor that mediates a subset of

responses to the plant stress hormone, abscisic acid (ABA), including maturation and onset of

seed desiccation tolerance. The classic vpl phenotype is that of precocious germination due to

reduced ABA sensitivity (McCarty et al., 1991). More recently, however, VP1 has been

implicated in stress responses of non-seed tissues (Cao et al., 2007). In addition, preliminary

evidence suggests that VP1 may be involved in modulating female reproductive quiescence in

maize under drought stress (Eveland, unpublished). To resolve differences in expression profiles

that would help define roles of the VP1 gene, cDNA sub-libraries, each tagged with a unique

multiplex key, were prepared from wild-type and vp]-mutant maize ovaries.

Data Assembly and Analysis

A sequencing reaction on the Genome Sequencer 20 instrument (Marguiles et al, 2005)

yielded 228,595 high-quality-reads with an average, trimmed length of 95 bases. Of these, 93%

were identified as correctly oriented, 3'-anchored cDNAs with equal representation of wild-type

and mutant sub-libraries (Table 2-1). The 7% of reads that were excluded from further analysis

contained errors in the multiplex key (1.5%) or invalid ligation junctions (5.5%). Assembly of

validated, trimmed, high-quality-reads using CAP3 (Huang and Madan, 1999) revealed 14,822

non-redundant 3'-anchored consensus sequences, each represented by 2 to 2,500 reads, and

32,477 singlets.









The capacity of these consensus sequences to identify individual genes based on specificity

of their 3'UTRs was tested by aligning these reads to available maize cDNA databases (TIGR

Zea mays Gene Index [ZmGI] and Industry UniGene [IUC]) using BLASTN (cutoff: E < 10-7).

At least 87% of the consensus tags matched cDNAs and 66% aligned with a gene-enriched

maize genomic assembly (MAGI) (Fu et al., 2005). In addition, BLASTN searches of the TIGR

maize repeat database (http://maize.tigr.org/repeat db.shtml) indicated that only 1.9% of 3'-

anchored consensus sequences contained retrotransposons or other repetitive sequences whereas

another 1.2% were identified as organellar or cytosolic rRNA contaminants. The latter were

most likely due to rare mispriming by the oligo(dT)-B adaptor during cDNA synthesis.

Analysis of 3'-UTR Profile Reveals a Dynamic Range of Expression

Based on the set of unique consensus sequences obtained from the two-sample library, we

developed a graphic display for the quantitative transcriptome profile (Figure 2-4A and 2-4B).

We quantified gene expression for each of 11,559 consensus sequences that matched unique

cDNAs using read frequencies. The results are plotted on a logarithmic scale to capture the full

range of expression. The 11,559 3'-sequences profiled were also analyzed based on Gene

Ontology functional classifications determined by PFam searches derived from ZmGI and IUC

databases. Analysis of respective maize cDNAs revealed 5,202 (45%) that were unclassified and

lacked annotation based on sequence similarity. An additional 578 (5%) of consensus sequences

matched genes having conserved domains of unknown function.

The relationship between abundance of each mRNA and its rank (ordered from least- to

most- prevalent) in the whole dataset approximated a Zipf-power law (ranked slope near -1 on a

log-log scale). This distribution was evident among transcripts overall (Figure 2-4), and within

individual functional classes (Figure 2-4B, 2-5A). Zipf's power law relationships are observed

in a wide range of natural phenomena including the distribution of gene expression in a variety









of organisms (Kuznetsov et al., 2002; Furusawa and Kaneko, 2003). Accordingly, it has been

used as a tool for normalization in some SAGE and microarray analyses. Although our results

were consistent with this distribution, an interesting exception is shown for the chromatin-related

functional class. As shown in Figures 2-4B and 2-5B, distribution of expression was skewed for

this group of mRNAs by a disproportionate number of highly abundant transcripts.

Distinguishing Gene Family Members

In order to evaluate 3'UTR profiling for resolution of individual gene family members, we

analyzed the Cellulose Synthase (CesA) gene family (Figure 2-5A). The assembled 3'-anchored

sequences distinguished 12 unique transcripts representing nine annotated CesA gene family

members (Table 2-2) that were previously characterized in maize (Holland et al., 2000;

Appenzeller et al., 2004). The full-length CesA cDNAs (ZmGI) share up to 94% sequence

identity. In some cases, extensive sequence similarity between CesA genes and their proximal

mapped locations to each other in the genome are suggestive of paired duplications (e.g. CesA1

and CesA2 on chromosomes 6 and 8 and CesA4 and CesA9 on chromosome 7). The cDNA

sequences for CesA4 and CesA9 differ almost exclusively in their 3'UTRs, thus complicating

resolution of these two genes in previous expression studies (Holland et al., 2000). Here, the

corresponding 3'-anchored 454 reads for these closely-related gene family members aligned with

gene-specific regions in the 3'UTR (www.plantphysiol.org/cgi/data/pp. 107.108597/DC1/1).

Polymorphic variants for CesA4 and CesA6 were also identified. Alignments of consensus tags

to a CesA4 cDNA (TC287832) indicated that a novel transcript variant (CesA4c) contained an

MspI restriction site polymorphism as well as 35 bp of an unspliced intron (Table 2-2).

Although no other ESTs having these features were detected in maize databases, nine reads in

our maize-ovary dataset aligned with the CesA4c variant. Consistent with the possibility that

these sequences identify a second CesA4 gene, CesA4 has been mapped to two locations (2.06









and 7.01; Holland et al., 2000) corresponding to duplicated chromosome segments (Helentjaris et

al., 1988).

In addition, we analyzed a group of closely-related Histonel (H/)-like transcripts (Figure

2-5B). These transcripts matched a unique, non-redundant set of ESTs from various maize

cDNA libraries and were annotated based on sequence similarities in other species. Although

these HI genes have not been individually characterized in maize, BLASTN results provided

insight for eventual functional analysis. For example, a very highly expressed HI-like transcript

(TC292133a) matched a drought- and ABA-induced gene that had been characterized in tomato

(Bray et al., 1999). These results indicate that unbiased profiling of closely-related transcripts

can facilitate studies of functional genomics with or without a fully annotated EST dataset or a

completely sequenced genome.

Evaluation of Differential Expression between Multiplexed Sub-libraries

The use of 3'UTR profiling as an effective strategy for detecting quantitative differences in

transcript abundance between samples was evaluated based on read frequencies generated from

individual sub-libraries. Read frequencies representing each expressed gene were determined for

wild-type and mutant sub-libraries by parsing the CAP3 ace file output. We analyzed 4,147

consensus sequences that were represented by a total of 10 or more reads using a chi-square

statistic. Of these, 202 showed significant differences (p < 0.0015) in frequency between the two

samples, indicating putative differences in transcript levels. A subset of these consensus

sequences with highly significant differences between libraries was annotated by BLASTN to

identify best-match ESTs (Table 2-3). Of the 30 sequences listed, three matched to unannotated

cDNAs which appeared to be maize specific (TC286704, TC300122 [ZmGI], 2569799 [NCBI

UniGene]) and, based on searches of public databases, 10 were found exclusively or highly

represented in cDNA libraries from reproductive tissues (2568974/TC285721, 514900, 2566963,









2568212/TC286030, 507904/TC301902 [NCBI UniGene/ZmGI]) or from drought-stressed

plants (2564044/TC285867, 2714857/TC286791, 508486/TC29233, 2561245/TC299973,

2567165 [NCBI UniGene/ZmGI]).

Quantitative differences in levels of specific mRNAs were confirmed for a subset of genes

by real-time RT-PCR analyses of the wild-type and vpl mutant samples (Figure 2-6). Results

showed that differences in transcript abundance between wild-type and mutant RNA samples

used in 3'-cDNA sub-library construction paralleled the 454-based expression profiles.

Resolution of Near-Identical Transcripts by Polymorphisms

Analyses of the maize genome have revealed a high frequency of nearly-identical paralogs

with > 98% identity (Emrich et al., 2006b). In most instances, both gene copies are expressed.

Identification of single feature polymorphisms in the 3' sequences can effectively distinguish a

subset of such paralogs. At least one example where 3'UTR profiling effectively resolved near-

identical paralogs was evident for closely-related, but differentially expressed Auxin Repressed

Dormancy Associated transcripts (we designated these genes as ARDA1 and ARDA2). The

ARDA1 and ARDA2 sequences share > 98 % identity (99% in the coding region and 97% in their

3'UTRs). Two distinct 3'UTRARDA consensus sequences detected an 18-bp indel

polymorphism that distinguished these two paralogs. Read frequencies showed reciprocal

responses in the mutant background by ARDA1 (p < 10-53) and ARDA2 (p < 10-12). Differential

profiles for these genes were confirmed by amplifying the region in or around the indel using

real-time RT-PCR (Figure 2-7A). These reciprocal expression profiles could not be resolved

when regions outside of the indel sequence were amplified due to confounding effects of the

nearly-identical sequences.

Earlier work identified ARDA1 as a potentially important contributor to stress-tolerance in

hybrid maize (Guo et al., 2004). The previously-undetected ARDA2, resolved in the mono-









allelic W22 inbred, matched a unique maize EST. Alignment of the two consensus sequences to

a region within assembled genomic sequence (MAGI4_156527) verified presence of two

paralogous gene products (Figure 2-7B). Both ARDA paralogs appeared to be drought-

responsive in preliminary analyses. Currently, there is little information for putative roles of

Auxin Repressed Dormancy Associated genes. Studies in pea (Pisum sativum) characterized

similar genes as markers for dormancy in axillary buds (Stafstrom et al., 1998).

Validation of SNPs and Homopolymer-Based Polymophisms

We conducted a detailed analysis of polymorphisms detected by a preliminary dataset

comprised of 1,263 W22 consensus sequences using BLASTN alignment to MAGI4 B73

genomic sequences. We expected that some portion of apparent polymorphisms in consensus

sequences ranging from 2 to 75 reads (56.6%, Table 2-4) was due to sequence errors. To

estimate the contribution of sequence errors in the 454 data, we evaluated polymorphisms

detected by a subset of 107 cDNA consensus sequences (7 to 75 reads) with respect to B73

MAGI assemblies by independent BLASTN searches of IUC cDNA and public EST databases.

We confirmed 93.8% of 146 sequence polymorphisms detected within 52 W22 alleles by

identical cDNA matches indicating that most identify independently-documented maize alleles.

Because the pyrosequencing method used by 454 is prone to errors in estimating lengths of

long homopolymer runs (Margulies et al., 2005), we investigated the effect this may have on

SNP detection in maize sequences. Overall, 29% of the 1,263 W22 consensus sequences

analyzed above contained one or more homopolymer tracts of 5-bp or longer. In order to assess

the impact of homopolymer read errors on SNP detection by 454, we analyzed the

polymorphisms detected by a set of 211 W22 consensus sequences (5 to 75 reads) in best

alignments to the MAGI4 (B73) dataset. Of the total 257 polymorphisms detected (counting

indels as one), at least 89.9% were independently confirmed by identical cDNA matches. In









addition, only 60 (23%) were potentially attributed to simple or compound (e.g. CCTTT -*

CCCTT) homopolymer base-calling errors. Moreover, the 60 homopolymer-based

polymorphisms were distributed randomly (p > 0.9) between alignments that included long

homopolymer tracts of > 5 bp (21% of consensus sequences analyzed) and alignments that

lacked them. Finally, all but 7 of 60 homopolymer length polymorphisms were supported by

independent EST sequences from W22 or other sources. Hence, these homopolymer-based

polymorphisms were not appreciably less reliable (88.3% confirmed) than other substitution and

indel polymorphisms (93.8% confirmed by independent cDNA sequences).

Discussion

Our results demonstrate that 3'UTR profiling is an effective strategy for high-resolution

global analysis of gene expression that does not require a complete genome sequence. Using this

approach, we were able to identify over 14,000 gene-specific mRNAs and quantify expression

based on read frequencies occurring in 3'-anchored consensus sequences. Analysis of the

quantitative 3'UTR profile revealed a dynamic range of gene expression spanning greater than

three orders of magnitude.

Our strategy of using long-read, 454 sequencing to target gene-specific 3'UTRs offers

several advantages over previous tag-based approaches to global expression profiling. First,

depth of sequencing is enhanced by anchoring the 454 reads to unique sites proximal to the 3'

ends of transcripts. This eliminates redundancy associated with shotgun sequencing of cDNA

fragments, thus providing more reads per unique transcript and reducing the potential for highly-

expressed mRNAs to saturate the library (Weber et al., 2007). In this study, the two-sub-library

analysis using the 454 Genome Sequencer 20 instrument identified 47,299 distinct mRNAs

(including 14,822 consensus sequences represented by 2 to 2500 reads). In comparison,

sequencing of nebulized Arabidopsis cDNAs yielded approximately 17,500 unique transcripts









after two GS 20 sequencing reactions (Weber et al., 2007). Although we cannot discount the

possibility that a portion of the singlets identified in our dataset are due to sequencing errors,

deeper sampling with the upgraded FLX technology will provide enhanced statistical support for

rare transcripts.

Second, the specificity of these long, 3'UTR-based sequence reads facilitates unambiguous

gene assignment. Our analyses indicated that individual gene family members can be resolved

by unique, gene-specific 3'-anchored tags and the corresponding closely-related ESTs

characterized. Finally, enrichment of 3'UTR sequences provides a useful source of polymorphic

information for studies of natural variation. Identification and analysis of nearly-identical

paralogous genes is improved on a genome-wide scale by enrichment for polymorphisms in the

3' sequences. Even in cases where genomic information is very limited, high-throughput

sequencing of 3'UTRs from species' variants allows direct comparison of polymorphic loci.

This approach thus provides a tool for genotyping and assessing genetic diversity contributing to

quantitative traits without the need for a sequenced genome or extensive EST collections.

Approximately 22% of the unique mRNAs identified in this study by > 2 reads did not

match ESTs in either ZmGI or IUC databases. A similar percentage of novel sequences (30%)

were also observed for a transcript profile from maize shoot apical meristem (SAM) using 454-

based shotgun sequencing of sheared cDNAs (Emrich et al., 2007a). The 8% difference may

reflect an increased specificity of our BLASTN results using the IUC cDNA collection and/or

more novel transcripts identified in the non-differentiated SAM tissue. Our data also showed

that among distinct mRNAs matching cDNAs, approximately 50% either contained domains of

unknown function and/or were unclassified based on lack of homology to annotated genes in









other species. This percentage demonstrates the potential for gene discovery with unbiased

sampling and sequencing of gene-specific 3'UTRs.

Furthermore, quantitative analyses of closely-related transcripts can extend studies of

functional genomics to species without completely sequenced genomes and where gene families

are largely uncharacterized. We addressed this possibility with an analysis ofHistone-1 (H1)

transcripts in maize ovaries. Although the individual genes have not been characterized in

maize, identification of the corresponding HI ESTs indicated that these unique, non-redundant

transcripts are indeed expressed. One highly represented Histone-1 mRNA in maize,

TC292133a, was annotated as a drought- and ABA-inducible HI gene based on sequence

similarities in tomato (Bray et al., 1999). This annotation is consistent with a function for this

highly expressed HI in drought-stressed maize.

For organisms that have limiting cDNA resources, 3'-cDNA tags will be less likely to align

with upstream coding sequences, thus constraining functional annotation. Nonetheless, 3'UTR

sequences enable resolution of unique mRNAs and distinguish among closely-related transcripts.

Quantitative data on transcript abundance is also provided, as well as an open, unbiased sampling

of the transcriptome. Where additional cDNA information is available, the 3'-cDNA sequences

can be extended by BLASTN alignments. Alternatively, the sequence tags can be used to design

primers or probes for screening of cDNA libraries. While the divergence of 3'UTR sequences

facilitates resolution of genes within a genome, it may limit the effectiveness of cross-species

comparisons for annotation of transcripts. For example, alignment of maize ovary 3'-cDNA

consensus tags to the complete set of rice genes (OsGI) using BLASTN produced matches

(expectation score








Based on our analysis of SNPs identified within consensus sequences and comparisons

with B73 MAGI genomic assemblies (Table 2-4), we confirmed at least 89.9% of

polymorphisms independently by identical cDNA matches. These data are consistent with a

recent study by Barbazuk et al. (2007) in which 88% of SNPs sampled by 2 or more 454 reads

were validated by Sanger sequencing. Removal of the unconfirmed SNPs from our analysis

reduced the estimated polymorphisms in W22 relative to B73 to 43.9%. That estimate is

comparable to the 44% polymorphism reported for B73 and Mol7 alleles (Vroh Bai et al., 2006).

Due to incomplete coverage of the B73 genome, it is likely that some W22 consensus sequences

were aligned to closely-related, paralogous MAGI4 sequences rather than alleles (e.g. Auxin

Repressed Dormancy Associated paralogs).

In addition, our preliminary results indicate that homopolymer base-calling errors will

have a minor impact on ability to analyze polymorphisms in maize cDNAs. Importantly, even

where errors of this type occur, the consistency of base calling in reads derived from independent

454-libraries suggests that non-identical alleles may still be distinguished if they give rise to

different consensus sequences. This level of specificity in gene expression analysis is invaluable

to uncovering novel variation in polyploid or paleopolyploid genomes (Osborn et al., 2003).

Evidence of ancient tetraploidization in the maize genome can be observed for roughly 60% of

genes in duplicated regions (Messing et al., 2004). Conservative estimates indicate that

extensive amplification oftandemly duplicated genes may represent approximately one-third

(35%) of maize genes (Messing et al., 2004).

Our analysis of expressed CesA gene family members demonstrates the capacity of the

approach described to provide quantitative resolution of closely-related transcripts. This is

achieved by specificity of the 3'UTRs for individual cDNAs. Cross-hybridization of near-









identical transcripts often complicates identification of individual gene family members in array-

based experiments. Consistent with this, the resolution of three CesA4 transcripts, including a

putative splice variant, denotes the complexity within the CesA gene family in maize. Even with

the most stringent probe designs, cross-hybridization with unknown family members remains a

challenge in non-sequenced genomes. With unbiased sampling and sequencing, resolution of

tissue and/or temporal-specific transcripts and polymorphic variants will provide functional clues

in complex genomes such as maize (Ma et al., 2006). Furthermore, quantitative assessment of

transcription among individual members of a gene family can facilitate analyses of functional

genomics and address key questions in evolution. Studies in Arabidopsis have identified

instances of functional diversification among duplicated genes either in parallel biochemical

pathways (Blanc and Wolfe, 2004) or within specific developmental and metabolic processes

(Schmid et al, 2005).

Results from a quantitative 3'UTR expression profile showed that the Zipf power

distribution of gene expression in the entire dataset overall, but was not conserved within the

chromatin-related functional class. This group of mRNAs showed a skewed distribution of

abundance due mainly to a large number of distinct highly-expressed, Histone-3 (H3) transcripts.

Among these we identified 67 mRNAs having H3 functional domains and 39% of the consensus

sequences were represented by 100 to 1,000 reads. Results may be due to transcriptional

responses to the stress treatment or be specific to the reproductive tissues examined.

Validation of differences in transcript abundance for a subset of genes by real-time RT-

PCR in RNA samples used for sub-library construction supports 3'UTR profiling as a platform

for quantitative expression profiling between samples. Furthermore, construction of the 3'-

cDNA libraries by this method yielded sequences with very low retrotransposon content and









nominal rRNA contamination. In addition, read distribution between multiplexed samples was

well balanced. Thus, a multiplexing strategy can be used to concurrently profile multiple

samples for increased cost effectiveness. Incorporation of a four-base error detecting key

enables up to 64 unique combinations for individual sample recognition.

Preliminary data generated with the recently upgraded FLX 454 technology (Harkins and

Jarvie, 2007) identified approximately 22,920 unique consensus sequences with a much higher

depth of sequencing for 12 multiplexed samples in a single reaction (Eveland, unpublished). The

enhanced sequencing capacity of FLX will therefore provide improved statistical analyses while

increasing the number of multiplexed cDNA libraries (e.g. biological replicates and treatments).

In the present study, a single GS20 run enabled detection of unique transcripts (2 or more reads)

with a sensitivity of approximately 1 in 100,000 mRNA molecules. A similar run on the 454-

FLX instrument is expected to increase sensitivity by at least 2-fold. This will be directly

applicable to identifying rare transcripts and resolving complex gene families.

Also, the range of gene expression quantified by the 454-based 3'UTR profile provides

higher resolution in global transcript profiling analyses compared to array-based hybridization

experiments. Accordingly, our results include detection of many rare mRNAs as well as

quantification of highly abundant transcripts. In contrast, this level of resolution was not

observed in initial microarray analyses of the same tissues (Eveland, unpublished) due to

threshold levels of detection and saturation. Likewise, with array-based interpretation of fold-

changes, subtle variations in gene expression are often not detected, but can have a significant

impact on physiology. A quantitative appraisal of all expressed sequences is thus invaluable to

studies of quantitative traits such as heterozygosity (Birchler et al., 2003; Stupar and Springer,

2006).









Future Prospects

With 454-based, long-read sequencing of 3'UTRs, quantitative profiles for allele-specific

inheritance patterns can be generated in the absence of apriori data on polymorphisms. Allelic

variants are frequently distinguished by single-feature polymorphisms such as those that marked

nearly-identical paralogs in this study. Identifying allele-specific differences in gene expression

and quantifying parental contributions to complex traits in Fl hybrids is key to understanding

genetic mechanisms such as imprinting (Guo et al., 2003) and heterosis (Bircher et al., 2003;

Springer and Stupar, 2007). In addition, strategies for eQTL analyses (Schadt et al., 2003;

Borevitz and Chory, 2004) and genome-wide linkage studies (Cheung et al., 2005) are improved

by a high-resolution, non-biased approach to quantifying allelic imbalances in gene expression

(including those resulting from imprinting or X-chromosome inactivation). Furthermore,

analysis of natural variation is enhanced by recovery of haplotypes in species where genomic

information is limited.

Natural variation can also be assessed with array-based probe sets generated from 3'-

anchored sequence reads (Borevitz et al, 2003). For species in which comprehensive microarray

platforms are not available, the 3'UTR sequence reads can serve as blueprints for chip

construction with highly-specific probe sets representing an unbiased sample of expressed

sequences. Alternatively, for genomes with limited EST support, this method can enhance

efficiency of cDNA sequencing by pre-screening libraries to eliminate redundancy. Fine

mapping and marker-assisted breeding can also be facilitated by utilizing insertion-deletion

polymorphisms (indels) in the 3'UTRs as molecular markers (Bhattramakki et al., 2002; Vroh Bi

et al., 2006). In addition, anchoring the 454 sequences proximal to the 3' ends of transcripts

enables resolution of 3'-RNA processing variants. Instances of differential polyadenylated

transcripts were readily detected in this study (data not shown). Currently, identification and









characterization of alternate poly(A) sites is fragmentary for most genes, since the required

length of 3'UTR sequence has been largely outside the range of short-read technologies

(Jongeneel et al., 2005).

Conclusions

By combining the specificity of 3'UTRs with long-read, high-throughput sequencing, we

are able to distinguish expression of newly-identified genes and closely-related transcripts on a

genome-wide scale. This can also be accomplished without reference to a completely sequenced

genome. The approach provides an efficient avenue for gene discovery and elucidation of

variations in expression that underlie natural variation and contribute to complex genetics of

heterosis and imprinting. In addition, 3'UTR profiling advances studies of comparative and

functional genomics by quantitatively resolving expression of gene families and identifying

unknown gene family members.

Materials and Methods

Plant Materials

Maize (Zea mays) plants were grown in 14"/ 7 gallon pots under greenhouse conditions

(Sept. to Nov. in Gainesville, FL) at 12-h-light/12-h-dark cycles. Sibling wild-type and vpl

mutant plants were in a W22 inbred background were derived from a self-pollinated vpl/+

heterozygous ear. A drought-stress treatment was initiated by gradually withholding water

beginning two weeks prior to tassel emergence. Soil was covered to restrict water loss by

evaporation and pots were weighed at the end of each day to determine water loss to

transpiration. Water lost to transpiration was added back. One week after ears first appeared,

water was withheld completely. Ears were collected right before silk emergence from wild-type

and vpl mutant plants. Immature ovaries (with pedicels) were hand-dissected from equivalent









sections of each ear (base-to-mid section), weighed to 50mg FW (15 ovaries per ear), and frozen

in liquid N2.

Sub-library Preparation and Sequencing

Tissue was homogenized in TRIzol reagent (Invitrogen) using a FastPrep lysis system (Q-

BIOgene). RNA was extracted using standard methods based on protocols from the University

of Arizona (www.maizearray.org). Total RNA (5ug) from wild-type and vpl mutant ovaries was

used for cDNA synthesis (MessageAmp II, Ambion) and primed with 6 pmol biotinylated (T12)

B-adaptor (modified from Margulies et al. [2005]) oligo: Biotin-

CCTATCCCCTGTGTGCCTTGCCTATCCCTGTTGCGTGTCTCAGTTTTTTTTTTTT[AGC].

Purified cDNA (DNA clear, Ambion) was bound to M-270 Streptavidin beads (Dynal),

immobilized on a Magnabot 96 (Promega), and digested with Mspl (Promega) to create 2-base

CG overhangs for adaptor ligation. A-adaptor oligos (modified from Margulies et al. [2005])

included 3-base muliplex keys (wild-type sub-library top strand, 5'-

CCATCTCATCCCTGCGTGTCCCATCTGTTCCCTCCCTGTCTCAGCAT-3', wild-type sub-

library bottom strand, 5'-

CGATGCTGAGACAGGGAGGGAACAGATGGGACACGCAGGGATGA-3'; vpl mutant

sub-library top strand, 5'-

CCATCTCATCCCTGCGTGTCCCATCTGTTCCCTCCCTGTCTCAGACT-3', vpl mutant

sub-library bottom strand, 5'-

CGAGTCTGAGACAGGGAGGGAACAGATGGGACACGCAGGGATGA-3').

Adapter pairs were combined and concentrated to 1 pmol/uL in salt buffer (10 mM Tris, 1

mM EDTA, 50 mM NaCl [pH 8]) and annealed by incremental, -1 degree/min decreases (95 -

4C, with a 30-min hold at 72 71C). Adaptors (5 pmol) with multiplex keys CAT and AGT

were ligated to digested wild-type and mutant cDNA samples, respectively. The 3-base key









sequences enabled detection of single-base errors in the multiplex key. Un-ligated adaptors were

removed by washing beads twice with 1X B&W buffer (2.5 mM Tris-HCL [pH 7.5], 0.25 mM

EDTA, .5 M NaC1) and twice with ddH20. The desired 5'-A-cDNA-B-3' template strand was

eluted with 100 mM NaOH, neutralized and concentrated on a Qiagen column (Margulies et al.,

2005). Sequencing was conducted as per Margulies et al. (2005) using a 454 GS-20 instrument.

The expected yield of -3x109 template molecules for the combined libraries was confirmed

by a SYBR Green Q-PCR strategy (MyiQ, Bio-Rad). Molecules iL-1 of amplified product were

calculated from an in-vitro transcribed (MAXIscript, Ambion) alpha-tubulin (Z. mays) standard:

alpha-tubulin forward, 5'-TTGTGCCTGGTGGCGACCTGG-3' and alpha-tubulin reverse: 5'-

ACCGACCTCCTCGTAGTCCT-3'.

Data Analysis

Quality-trimmed 454-sequences (FASTA format) were filtered for valid key and ligation

junction (CGG) sequences at 5'-ends and poly-A tails were trimmed using custom programs

written in java. Validated, trimmed sequences (93% of total reads) were assembled using CAP3

(http://genome.cs.mtu.edu/sas.html). The non-redundant set of consensus cDNA sequences

represented by 2 or more reads (14,822 total assemblies) were annotated by BLASTN searches

of cDNA databases for maize. These included TIGR Maize Gene Index Assemblies (ZmGI) and

Industry UniGene (IUC), a collection of cDNAs provided by an industry consortium via a user's

agreement (http://www.maizeseq.org). Functional classifications of cDNA matches were based

on Gene Ontology terms associated with PFam (http://www.sanger.ac.uk/Software/Pfam/)

assignments in IUC. In addition, consensus sequences were aligned by BLASTN to a maize

genome sequence assembly (MAGI, version 4.0 [http://magi.plantgenomics.iastate.edu/]).

Trace files for 454 sequences were deposited in NCBI.









Real-time RT-PCR Analysis for Validation of 454 Data

Real-time RT-PCR was carried out to validate technical replicates of RNA samples used in

sub-library construction. For real-time PCR analysis, cDNA was synthesized from DNaseI-

treated (Ambion) total RNA using an oligo(dT) primer (TaqMan Reverse Transcription

Reagents, ABI). Real-time PCR was monitored using the MyiQ Single Color Real-Time PCR

Detection system (Bio-Rad). Each reaction contained 10 uL of 2x iQ SYBR Green Supermix

(Bio-Rad), 1.0 uL of cDNA sample, and and 200 nM gene-specific primer in a final volume of

20 uL. All reactions were performed in triplicate. The relative abundance of transcripts was

normalized with 18S ribosomal RNA control values using Taqman (Ribosomal RNA Control

Reagents, ABI) and to the constitutive expression of an alpha-tubulin mRNA using SYBR Green

on cDNA templates (MyiQ, Bio-Rad). SYBR Green was used to amplify a subset of transcripts

with gene-specific primers. Primer pairs were designed using the 454-read and adjacent

sequences in best-match ESTs identified by BLASTN as templates.

CBS Chloride channel ([2562879] Cbs forward, 5'-ATGGATGCTGCTGTTCTCATGCTC-3'

and Cbs reverse, 5'- ATGGAGTCTCCTGGCGTGCTAC-3'), Thaumatin/osmotin ([1321765]

Osmotin forward, 5'-TACCGCAGCAGCTGAACAACG-3' and Osmotin reverse, 5'-

ATGTTCCGTCGCAGTCGCTAGG-3'), Senescence-associated/tetraspannin ([TC299489] Sa

forward, 5'-AACGACGAGGACGACCTCTGC-3' and Sa reverse, 5'-

AGTTTGATTAAGCGTCACCGCCTCG-3'), Chlorophyll a-b binding protein ([TC299127]

Cab forward, 5'-TGTACCCTGGCGGCAGCTTC-3' and Cab reverse, 5'-

ATCCACGTACGTACACCCTCTCC-3'), Cu transport ATPase/heavy-metal-associated

([2562278] Hma forward, 5'-AGCCAAAGCTGACGCCTGATC-3' and Hma reverse, 5'-

TCCTGCAAGGGATGTGTTGTTC-3'), Glycine-rich protein ([2923887], Grp forward, 5'-

ATCAGGTGAAGGATACGGACAAGGTG-3'and Grp reverse, 5'-









ACAGGACAAATTACAAGCCTTGCGGTG-3'), Dehydrin DHN1/RAB-17 ([TC286791]

dehydrin forward, 5'-ACAGCACTGAGCGGCGCCTATAC-3' and dehydrin reverse, 5'-

ACGTAGCAGCATAAACAGTACACGGACC-3').

Relative expression levels of ARDA1 and ARDA2 were compared by real-time RT-PCR

using SYBR Green and gene-specific primers within and around the 18-bp indel sequence.

(Ardal forward, 5'-TACAAGCGGGCGCAGTCG-3', Ardal reverse, 5'-

AGCAAACATGGCCTCTTCACTG; Arda2 forward, 5'-TACAAGCGGGCGCAGTCG-3',

Arda2 reverse 5'-TGGCCTGACAGAGACACCG-3').









Table 2-1. Summary statistics of a two-sample multiplex 3'-UTR library.
Total high-quality reads 228,595
Wild-type sub-library reads 105,289
Mutant sub-library reads 109,958
Error-detected reads 13,348
Errors in multiplex key 1.5%
Invalid ligation junctions 5.5%
Total unique sequences 49,822
Singletsc 32,477
Consensus sequences (> 2 reads) 14,822
EST matches 11,559
Genome matches 9,740
aAbundance of reads after filtering for Mspl-ligation junctions. bNonredundant set of sequences
representing unique transcripts. cSingle-copy transcripts. dAbundance of consensus sequences
matching cDNAs in Industry Unigene or ZmGI databases. eAbundance of consensus sequences
matching available maize genomic sequence assemblies (MAGI4).









Table 2-2. CesA family 3'-anchored consensus sequences.
Read
3'-anchored consensus sequence Ra


CGGAGGCTGCGGCAACCTTGTGCAGTTCGGCCACGAATATACTAGGGAAGATCGC
CesA6a
GACCAATCAATCAATCGATGACCGAGTTCAATTGTTCAAAG
CesA6ba CGGATCGACCCTTTCCTTGCGAAGGATGATGGTCCCCTGTTGGAGGAGTGTGGTCT
GGATTGCAACTAGGAGGTCAGCACGTGGACCTTCCCGTNAGTGTGTGG
CGGGATCTCGAACGCGATCAACAACGGGTACGAGTCGTGGGGCCCCCTGTTCGGG
CesA8 AAGCTCTTCTTCGCCTTCTGGGTGATCGTCCACCTGTACCCGTTCCTCAAGGGTCTG
GGTGGGG
CGGACGCCCACCATCGTCGTGCTCTGGTCCATCCTCCTCGCCTCCATCTTCTCGCTC
CesA10
GTCTGGGTCAGGATCGACCCGTTTATCCCGAAGGCCAAGGGCCCCAT
CGGATACCCAGACGTGTGGCATCAACTGCTAGGGAAGTGGAAGGTTTGTACTTTGT
CesA4a
AGAAACGGAGGAATACCACGTGCCATCTGTTGTCTGTTAAGTTATATAT
CGGATACCCAGACGTGTGGCATCAACTGCTAGGGAIGTGGAAGGTTTGTACTTTG
CesA4bb AGAAAC
GGAGGAATACCACGTGCCATCTGTTGTCT GTTAAGTTATATA
CGGGGTGTTGGTGATATTGATGCATCAACTGATTACAACATGGAAGATGCCTTATT


CesA4c

CesA9

CesA5

CesA2

CesAl

CesA3d


CGGATACTCGAACGTGTGGCATCAACTGCTAGGGAGGTGGAAGGTTTGTAGAACA
GAGAGATACCACGAATGTGCCGCTGCCACAAATTGTCTGTTAAG
CGGGTCACTGGCCCTGATATCGCGAAATGTGGCATCAACTGCTAGGATGAGCTGA
ATATAGTTAAAGAGTGGAACTAGACGCATTGTGG
CGGTGCTGCTGCAGACAATCATGGAGCCTTTCTACCTTGCTTGTAGTGCTGGCCAG
CAGCGTAAATTGTGAATTCTGCTTATTTTTTTAG
CGGTGCTGCTGCGGACTAAGAATCACGGAGCCTTTCTACCTTCCATGTAGCGCCAG
CCAGCAGCGTAAGATGTGTAATTTTGTAAGTTTTGTTATGTC
CGGCACAATCATGATCTACCCCTTCGTGTAAATACCAGAGGTTAGGCAAGACTTTT
CTTGGTAGGTGGCGAAGATGTGTCGTTTAAGTTCACTCTACTGCTAGTTTGGGGG


26 100 (93)

9 97 (104)


8 100(113)


2 100 (104)

5 100 (106)


2 95 (107)


2 100 (57)

10 100 (98)

27 96 (88)

14 100 (78)

18 97(96)

17 100 (100)


aCesA6b reads are located upstream of the CesA6a consensus tag, consistent with a transcript containing a restriction site
polymorphism. However, absence of the downstream Mspl site was not independently confirmed. bCesA4b contains 3 indel
polymorphisms in comparison to CesA4a and all 3 are represented in both reads. cCesA4c aligns with a different genomic assembly.


% Match











Table 2-3. Best-match cDNAs and associated annotations (BLASTN) for consensus sequences
showing highly significant differences in transcript abundance between wild-type and
vpl mutant drought-stressed ovary sub-libraries.


cDNA ID
TC285867b
TC285721a
TC286704
TC305930
2569891
TC285789
TC292711
514' I i
TC286791b
TC292358
507881
TC286485
2566963a
TC286030a
TC310545
TC294233b
2708354
TC298173
TC294(51)5
TC301902a
TC305186
1572511
TC299973b
654573
TC280589
2567165 b
2569891
TC310820
2569799
TC300122


% match
(length)
100 (87)
100 (98)
98 (94)
98 (80)
100 (52)
100 (96)
100 (100)
97(116)
100 (86)
100 (96)
100 (95)
100 (87)
99 (104)
100(111)
100 (93)
100 (104)
100 (97)
95 (93)
100 (95)
100 (87)
100 (78)
100 (49)
95 (84)
100 (45)
100 (98)
90(110)
100 (52)
100 (91)
100 (74)
99 (103)


Read frequency differences for consensus sequences were analyzed using a Chi square statistic.
Consensus sequences were aligned to best-match cDNAs using BLASTN in ZMGI and Industry
UniGene databases. NCBI UniGene IDs are listed for sequences where TC numbers were not
available.
acDNAs highly represented in Genbank libraries from maize reproductive tissues. bcDNAs
highly represented in Genbank libraries from drought-stressed maize plants.


BLASTn annotation (species)
Auxin-repressed dormancy (R. pseudoacacia)
GRP (0. sativa)
Unannotated
farnesylated protein 3 (H. vulgare)
XET (H. vulgare)
Auxin-repressed dormancy (P. sativum)
nodulin MtN3 family (A. thaliana)
Atlg74950 (A. thaliana)
dehydrin RAB-17 protein (Z. mays)
Thr rich extension (Z. mays)
unnamed protein product (0. sativa)
histone H2A (Z. mays)
unknown protein (0. sativa)
harpin induced gene 1 (0. sativa)
histone Hi (Z. mays)
putative cystatin cc3 (S. ottic i/lirn'/I
unannotated (0. sativa)
histone 3 (0. sativa)
EF-hand Ca2+-binding CCD1 (T. aestivum)
AP2 domain, EREBP (0. sativa)
subtilisin-like proteinase (0. sativa)
hypothetical protein (0. sativa)
glycogenin-like (0. sativa)
Ca+-binding EF hand family (A. thaliana)
phosphate-induced protein 1-like (P. ciliare)
heavy-metal-associated (0. sativa)
xyloglucan endo-1,4-beta-D-glucanase (H. vulgare)
CCCH-type zinc finger protein-like (0. sativa)
Unannotated
Unannotated


read frequency
wt vpl
238 709
57 244
25 167
14 119
234 81
53 1
0 52
127 37
8 69
1070 1414
93 21
575 396
74 12
106 31
259 135
65 158
75 16
151 61
102 32
66 13
113 41
31 0
31 0
53 10
146 67
141 64
43 6
15 61
59 14
127 56


P Value
10-53
10-27
10-24
10-20
10-18
10-12
10-13
10-12
10-12
10-12
10-11
10-11
10-11
10-10
10-10
10-10
10-10
10-10
10-9
10-9
10-9
10-8
10-8
10-8
10-8
10-8
10-7
10-7
10-7
10-7









Table 2-4. Polymorphisms detected by W22 3'-anchored 454 sequence reads.
Consensus tags with 2 to 75 reads (n = 1263)
Total polymorphic W22 alleles 56.6%
B73 MAGI Matches1 92.4%
Consensus tags with 7 to 75 reads (n=107)
Total polymorphic W22 alleles 52
# confirmed 47 (43.9%)1
Total polymorphisms 146*
# confirmed 137 (93.8%)o
Consensus tags with 5 to 75 reads (n=211)
Total polymorphisms 257*
# confirmed 231 (89.9%)0
Homopolymer-type polymorphisms 60*
# confirmed 53 (88.8%)0
Contiguous BLASTN alignment of >50 bp or > 90% of query length. Percent of matches
having one or more SNP or indel polymorphism within the best aligned segment. Allele
sequences supported by identity to independent EST sequences.












5' 3'
EST

I =


B
5' 3'
EST





Figure 2-1. Comparison of shotgun versus 3'-anchored approaches for using 454-sequences as a
platform for transcript profiling. A) Sequencing of randomly sheared cDNA
fragments, followed by contig assembly can provide coverage of full-length cDNAs.
However the redundancy of sequence-reads for a given transcript limits information
returned per number of reads and thus statistical analyses of expression. B) Our
3'UTR profiling method identifies unique ESTs by anchoring each sequence read to a
gene-specific region of cDNAs. This increases depth of sequencing and facilitates
assembly and analysis.













multiplex adaptor
S 3'sequence read ( 100bp)


A AAAA -
sequence primer for derived
454 Mspl (5'overhang) bead-linker
4-base calibration key
3-base multiplex error-detecting key
(16 combinations)

B
AAAAAAA
t biotin labeled 3' anchor oligo
AAAAAAA
^I TT [>
t 454 bead anchor copied
I ITT >
t streptavidin bead pull down
I =TT ff-O
t Mspl 5' overhang
AS- O
t multiplex adapter ligation


t elute
IAAAD


Figure 2-2. Strategy for 3'UTR profiling by 454-sequencing. A) A 3'-anchored cDNA library is
restriction-digested and tagged via ligation to a multiplex adaptor. Unique
combinations of a 4-base multiplex error-detecting key enables sample identification
during concurrent sequencing of up to 16 sub-libraries. B) Sub-library construction
from total RNA.














O)



0.1


LL .
U -


0
<25 100 200 300 400 500 600 700 800
Tag length


Figure 2-3. Distribution of tag lengths for a simulated Mspl digest of 70,000 3'-oriented ESTs
from B73 maize (maize full length cDNA project [www.maizecdna.org]). A
comparison of simulated tag lengths to the distribution of annotated rice 3'UTRs
indicated enrichment of 3'UTR sequence with an average tag length of 100-200 bases.
Proportion of short tags is reduced due to the low frequency of Mspl sites proximal to
the poly(A) tail.













1 *

**
10
I m



1000

IOK -

I 10 100 100] liK
TrnrinFt wMhrdhn ka scidali
B
blolicdaban siress
Call aide
ielwal





rrrent ssragem
dher
p-o l lshesis








m classrled
dIrrii s























frequency was used as a quantitative measure of mRNA abundance. A) Transcript
abundance is plotted on a log-log scale for respective genes in rank order from least-
to most-highly expressed. B) Transcripts are grouped into functional classifications
based on Gene Ontologies and plotted linearly along the y-axis. Read count for each
transcript is plotted on the x-axis (log scale). Color scale (dark-to-light) denotes the
dynamic range of mRNA abundance.












j* 1 *

0)
o i
10 CesA5
o CesA6a
e CesA1
CesA3
100 CesA3
CesA9
CesA6b
CesA4c CesA10 CesA8 CesA6b
CesA4c A ,, H CesAb
CesA4b CesA4a

0 1 10 100
Transcript abundance (log scale)
B


1 *

S 10 --TC292133a
0
10 f---TC292749
100 gi:14203885
TC294259
TC292749 TC292931
1000 TC292133b gi:149102785
TC292750
0 1 10 100
Transcript abundance (log scale)


Figure 2-5. Distribution of transcripts (log-log scale) in selected functional classes from Figure
4B and read count for each. Resolution of individual gene family members was
enabled by the specificity of the 3'UTR. A) Quantitative measure of mRNAs for all
transcripts classified with cell-wall-related functions (Fig. 4B) and resolution of 12
unique mRNAs representing 9 previously-characterized members of the CesA gene
family (including transcript variants for CesA4 and CesA6) in maize ovaries. Read
frequencies for CesA gene family members range from 2 to 27 reads. B) Quantitative
measure of mRNAs for all transcripts with chromatin-related functions (Fig. 4B) and
identification of Histone (H/)-like transcripts that matched unique, but
uncharacterized maize cDNAs (indicated by TC or GenBank IDs). Read frequencies
for HI gene family members range from 2 to 684 reads.














100-


^ Sa Dehydrin: Grp: alpha-tubulin:
Q- 454 Q-PCR 454 Q-PCR 454 Q-PCR 454
S100

<
z
E 50





Swild-type vp


Figure 2-6. Validation of technical accuracy for determining differences in transcript abundance
based on read number. Parallel real-time RT-PCR (SYBR green, MyIQ, BioRad)
analyses (technical error based on 3 independent determinations) on identical RNA
samples as used in 3'UTR sub-library construction validated the significant
differences in read frequency (p < 0.0015 except for alpha-tubulin control) for a
subset of genes. Gene-specific primers were designed using the 454-sequence-reads
and associated EST matches as templates.












VP I
Arda2


Ardal wt

0 20 40 60 80 100
Relative abundance (% mRNA)

B
GGCC 5'

3'
ARDA2


3' GGCC 5'
ARDA1 intron


5' CCGGCCGCAGCCCAACTCCCCCACCGTCTACG
ACTGGCTCTACAGCGACGAGACCAGGAGCAAC
CACCGCTAGATCGGGCGAGATGGACAGTGAAG
AGGCCCGGTGTCTCTGTCApGCCATGTTTGCT
GGTCCTTGTGGTTATCTAAAACGCATGCATGCT
CTATTATAGCTAGTCATCACTATATATATAC 3'


Figure 2-7. A 3' sequence polymorphism resolved nearly-identical Auxin Repressed Dormancy
Associated paralogs with differences in mRNA abundance. A) Q-PCR analyses
confirmed the reciprocal responses of near-identical Ardal and Arda2 (> 98%
identity) on wild-type and mutant samples used for sub-library construction (technical
error based on 3 independent determinations). B) The near-identical ARDA1
(EE188942) and ARDA2 (EE679809) ESTs differ by an 18-bp insertion within
ARDA2, which is not present in MAGI4_156527 (region of available maize genomic
sequence aligning with the ESTs). The 454-sequence-reads representing ARDA1
(blue) and ARDA2 (brown) are highlighted within the 3' end of the ARDA2 EST
sequence (top) and within a schematic diagram of the two ESTs (3' ends) aligned to
maize genomic sequence (bottom).









CHAPTER 3
EXPRESSION PROFILING OF DEVELOPING MAIZE OVARIES USING MICROARRAYS
AND SEQUENCING OF 3'-UTRS

Introduction

Improving agronomic traits for grain yield and quality is facilitated by increasing

availability of relevant genomic sequence information and functional data. Cereal crop

productivity represents approximately one-third of the world's food supply (faostat.fao.org), with

maize (Zea mays L.) producing the largest yield in total bushels. Although grain yield in the

USA has increased over the past 80 years with improved, commercially-available hybrids,

variation in harvestable yield has also increased (Bruce et al., 2002). Much of this variability is

attributable to drought stress (Boyer, 1982; Boyer and Westgate, 2004), one of the major causes

for yield reductions in most areas of the world.

The pre- and early post-pollination phases of maize reproductive development are critical

for seed set and subsequent grain yield (Zinselmeier et al., 1995, 2002; Andersen et al., 2002;

McLaughlin and Boyer 2004b). During this time, the plant is especially sensitive to abiotic

stresses, which can decrease pollination efficiency and/or promote kernel abortion, thus leading

to devastating losses in grain yield worldwide. Selective breeding for traits related to female

inflorescence and ear growth have improved pollination efficiency and seed set (Bolanos and

Edmeades, 1996; Bruce et al., 2002; Campos et al., 2004). However, relatively little is known

about key metabolic processes that influence maize female floret growth and development during

the silking and pollination periods.

Overall productivity in maize depends largely on photosynthetic capacity and the

subsequent allocation of carbohydrates to the developing female inflorescence (Boyer and

Westgate, 2004; Borras et al., 2007). In addition, nitrogen-based assimilates are essential for

kernel growth and have been implicated in maintaining the carbon: nitrogen balance in









developing ears (Seebauer et al., 2004; Martin et al., 2006). Both carbon and nitrogen

metabolites can provide a source of signals that regulate gene expression based on nutrient status

in sink cells (Koch, 1996, 2004; Seebauer et al., 2004; Smith and Stitt, 2007). Sink strength in

the female inflorescence is dependent on such signals to ensure sufficient availability of

resources for silk exsertion, successful fertilization, and to carry the developing kernel to

maturity.

Unfavorable conditions such as drought and shade stress negatively affect photosynthesis

in source leaves and typically reduce C available for import by developing ears (Westgate and

Boyer, 1985; Zinselmeier et al., 1995; 2002; Setter et al., 2001; Borras et al., 2007). Repression

of genes for sucrose metabolizing enzymes at low water potentials can also contribute to this

reduction in sink strength and resulting inhibition of growth (Andersen et al., 2002; McLaughlin

and Boyer, 2004b). In contrast, tassel development appears less affected by reduced

photosynthate availability. Therefore, stress can result in asynchronous timing of pollen shed

and silk emergence, evident as an enhanced Anthesis Silking Interval (ASI) (Bolanos and

Edmeades, 1996; Borras et al., 2007). Studies using reciprocal crosses determined that the

female reproductive tissues also controlled early, post-pollination kernel abortion (Westgate and

Boyer, 1986).

Recent expression profiling studies in maize have focused on effects of drought (Zhuang et

al., 2007) and shade stress (Zinselmeier et al., 2002) on immature reproductive structures.

However, the metabolic processes underlying normal ear development during silking and

pollination have not been widely investigated. Studies in animals (Newport and Kirschner,

1982; Newman-Smith and Rothman, 1998; Pelegri, 2003) and in plants (Vielle-Calzada et al.,

2000; Grimanelli et al., 2005) have suggested that maternal-based regulation dominates during









the initial phases of embryonic development. In maize, this was evident as a shift in maternal-to-

zygotic transition of transcripts several days after fertilization (Grimanelli et al., 2005). Also,

expression studies in barley showed co-regulation of genes related to maternal-specific metabolic

processes during early post-pollination development (Sreenivasulu et al., 2004).

Understanding the mechanisms underlying seed set in maize, and extent to which they are

altered by abiotic stress, can be valuable to maximizing yield potential (Bruce et al., 2002;

Tuberosa and Salvi, 2006; Vij and Tyagi, 2007). With release of the maize draft genome

sequence in 2008 and recent advances in sequence-based technologies for expression profiling

(Meyers et al., 2004; Nobuta et al., 2007; Eveland et al., 2008), the challenges facing researchers

are increasingly focused on genome annotation and functional prediction. Annotation of maize

genes is not trivial, however, due to extensive genome duplication and approximately one-third

of genes in tandemly arrayed gene families (Messing et al., 2004; Messing and Dooner, 2006).

In addition, analyses of the maize genome have revealed a high frequency of genes that share

>98% identity to a near-identical paralog and in many cases both are expressed (Emrich et al.,

2007b). Such sequence similarities complicate resolution of gene-specific expression profiles

due to potentially confounding effects from cross-hybridization of closely-related transcripts.

Functional genomics and expression profiles in complex genomes such as maize depend

on gene-specificity and the capacity to distinguish gene family members from other closely-

related transcripts with specialized functions. In addition, expression data from individual

organs or specific cell types (Ma et al., 2006; Zhang et al., 2007) are essential for relating gene

expression to putative function, especially if a range of developmental stages are included. For

example, work by Zhuang et al. (2007) showed that drought stress had very different effects on

gene expression in the immature tassel verses comparable tissues in ears. Also, tissue-specific









profiles from shoot apical meristems in maize revealed novel transcripts and cell-specific gene

expression (Zhang et al., 2007). Global transcript profiles compiled from tissue- and stage-

specific analyses are available for Arabidopsis and have contributed to the understanding of gene

function at the whole-plant level (Zimmermann et al., 2004). Such profiles for important crop

species would provide a framework for testing effects of environmental or genetic perturbation

on gene expression.

Advances in expression profiling technologies have also enabled resolution of natural anti-

sense transcripts (Ma et al., 2006; Zhang et al., 2007), alternate splice variants (Ner-Gaon et al.,

2007), and non-coding RNAs (Lu et al., 2005; Nobuta et al., 2007). Emerging sequencing

technologies (Meyers et al., 2004; Jorgoneel et al., 2005; Margulies et al., 2005), and their

applications for gene expression profiling (Nobuta et al., 2007, Weber et al., 2007; Eveland et al.,

2008), provide quantitative, high-resolution methods for gene-specific analysis. Studies

comparing array-based profiling methods with short-tag sequencing platforms such as SAGE

(van Ruissen et al., 2005) and MPSS (Liu et al., 2007), have shown overall agreement between

strategies. Inherent limitations with both array- and sequence-based profiles, however, indicated

that a complimentary approach was most effective.

In a previous study, we described a 454-based sequencing approach to transcript profiling

that used the specificity of 3'UTRs to distinguish between gene family members and other

closely-related transcripts on a genome-wide scale (Eveland et al., 2008). Here, we evaluated the

potential of 3'UTR profiling by concurrently sequencing a 12-sample library that included four

stages of maize ovary development during pollination. Resulting data were compared to parallel

microarray analyses. The significance of this work is thus two-fold. 1) A gene-specific,

sequence-based strategy is evaluated for its use in analysis of the maize transcriptome. 2) Sets of









co-expressed genes related to specific functional processes or key metabolic pathways during

maize ovary development are identified We concluded that the 3'-UTR sequence-based strategy

provided an effective complement to microarray analyses for gene-specific expression profiling

during maize ovary development.

Results and Discussion

To profile global transcriptional changes in developing maize ovaries, we sampled ears

from field-grown plants at two-day intervals over four stages of pre- and early post-pollination

growth (Figure 3-1A). Six individual ears were sampled each time point, which included: pre-

silking (4 DBP), silk emergence (2 DBP), time of pollination (OT), and two days post-pollination

(2 DAP). A detailed description of sampling procedures appears in Materials and Methods.

Ovary-plus-pedicel samples were hand-dissected from florets positioned equatorially around the

mid-basal region of each ear. Expression profiles were evaluated from each sample in all

possible pair-wise comparisons to samples at other developmental stages using maize long-

oligonucleotide arrays (University of Arizona). Dye swaps were included in the experimental

design (Figure 3-1B).

Microarray Data Analysis and Clustering

We first used a hierarchical clustering method to test whether overall trends in gene

expression coincided with development. Median signal intensities were extracted, background

subtracted, and log-transformed before fitting an ANOVA model. We identified 856 genes that

showed significant changes in expression during development with a False Discovery Rate

(FDR) of 10%. The 70-mer sequences corresponding to all 856 probes were annotated by

BLASTN alignments to cDNAs in Zea mays Gene Index (ZmGI) [tigr.org] and Industry Unigene

(IUC) databases. Of these, 606 were classified into functional processes based on Gene









Ontologies and 250 matched available cDNA sequences that were either unannotated or

contained domains of unknown function.

The two-way cluster hierarchy showed that genes clustered into three major groups based

on their expression pattern over time. These were 1) expression decreasing during development,

2) increasing during this time, or 3) showing combinations of both (Figure 3-2). Interestingly,

only a small proportion of genes responded specifically to pollination (evident as differences

between OT and 2 DAP). We compared annotations of genes that showed 1) decreases and 2)

increases in expression during development to determine whether functional processes were

differentially represented between these two groups (Figure 3-3). In general, expression shifted

during development from genes related to signaling, protein turnover/modification, and nucleic

acid metabolism to pathways involved in secondary metabolism and defense.

Identifying Clusters of Co-Expressed Genes

One approach we used to identify groups co-expressed genes was a non-linear clustering

analysis described in Qu and Xu (2006). This method is based on fitting a polynomial model for

expression of change and has been adapted to analyses of time course datasets (Fung et al.,

2008). Here, genes were clustered based on the pattern of mean differences in expression over

the four developmental stages. Background-corrected mean signal intensities were compared to

a spotted set of negative controls to determine percentage of genes that showed positive signal

intensities. The criteria used for probe elimination in the present work was based on a cut-off for

positive signal intensities and poor spot frequencies, and are described in the Materials and

Methods. A calculated, weighted Kappa value ranged from 69-91% across all genes with

positive signal intensities and indicated close agreement among biological replicates from a

given developmental stage.









In total, 40,686 genes showed positive signal intensities and a full vector of least square

means estimates after ANOVA. These 40,686 genes were fit to cubic polynomials describing

their patterns of change, and a minimum Bayesian Information Criterion (BIC) (grouping that

best fit the data) was used to determine an optimal number of clusters (Qu and Xu, 2006). We

identified a total of 7 co-expression clusters (Figure 3-4A) and highlighted those genes which

showed significant changes in expression during development at an FDR of 20% (Figure 3-4B).

Total per-cluster gene numbers and percentages of differentially expressed genes in each cluster

are shown in Table 3-1. In addition, 3,671 of the 40,686 genes (9%) showed greater than 2-fold

change during development. A quality check for background uniformity revealed areas of high

background for some slides. High background is common with the maize long-oligonucleotide

platform possibly due to interference from complex carbohydrates, enhanced cross-

hybridization, or inefficient blocking procedures (J. Gardiner, personal communication).

Because of such inconsistencies in background levels, analyses of genes that showed low signal

intensities across development were omitted from analyses done here. Instead, we focused

primarily on those genes with large fold-change or significant differences in expression during

development.

Co-Expression of Genes Related to Common Metabolic Pathways

The total set of 40,686 genes that showed a positive signal was annotated based on their

alignment to cDNAs in Zea mays Gene Indices (ZmGI). Annotated genes were grouped

according to putative functional processes using Gene Ontologies and, where appropriate, KEGG

pathway assignments (Kanehisa et al., 2007). Approximately 95% of the genes with positive

signals were associated with cluster 2, in which relatively little change in expression was

observed during development. Although 1,280 genes in cluster 2 were differentially expressed

during development (FDR 20%) and/or had large fold-changes, these represented only 3.3%









percent of the total in this group (Table 3-1). We focused on those clusters which had the

highest proportion of genes with significant changes in expression or high fold change during

development. These were clusters 1, 5, and 7 (Table 3-1).

Clusters 1 and 5 were similar both in their expression profiles (tended to decrease after

pollination) and in functional composition of annotated genes. Of the 1,567 genes that grouped

with Clusters 1, 5, and 7, we found that 672 were not annotated or were unclassified. We

compared functional classes of genes combined from clusters 1 and 5 to those from cluster 7

(Figure 3-5). Genes that showed either significant differences in expression (FDR 20%) or large

fold changes during development from clusters 1, 5, and 7 are listed with their associated

functions in Tables 3-2 and 3-3.

In general, expression profiles for genes in clusters 1 and 5 tended to show mRNA levels

that fluctuated during development and typically either decreased or showed very little change

after pollination. We compared profiles for these genes in clusters 1 and 5 to those associated

with cluster 7, which tended to show increasing expression during development, and often with

largest fold changes between OT and 2 DAP (Figure 3-4B). Comparison of functional processes

associated with the different clusters revealed an increase in genes related to nitrogen and amino

acid metabolism after pollination relative to those involved in carbohydrate metabolism (Figure

3-5). In clusters 1 & 5, N- and C-metabolism are equally represented. This balance shifts to a 6-

fold higher proportion of genes related to N metabolism that are expressed later in development,

as shown in cluster 7.

Also, genes for protein synthesis were heavily represented in cluster 7 (24% of the total)

while genes related to proteolysis were equally represented in clusters 1, 5, and 7. Interestingly,

a number of genes annotated as protease inhibitors in clusters 1 and 5 showed significant









differences in expression or large fold changes during development (Table 3-2). Increased

expression of genes for protein synthesis in maternal tissues during the later stages of ovary

development could be central to growth and development of filial tissues. Regulation of protein

synthesis and turnover has also been linked to C and N availability and signaling (Palenchar et

al., 2004).

Consistent with possible increases in C availability, genes related to phenylpropanoid

biosynthesis were highly represented in cluster 7. Many of these genes also showed significant

differences in expression or large fold changes during development (Table 3-3). Previous work

has shown that a number of key genes in lignin biosynthesis are sugar-responsive (Rogers et al.,

2005). We also noted a shift from prominent expression of genes related to cell growth, to those

related to differentiation during development (Figure 3-5).

Sequence-Based Analyses by 3'-UTR Profiling

A gene-specific, sequence-based 3'-UTR profiling strategy was previously described by

Eveland et al. (2008). To evaluate the potential of this approach in a transcriptome-wide

assessment of developing maize ovaries, we compared a parallel, multiplexed 3'-UTR

sequencing profile to the microarray dataset. Sub-libraries were constructed, as described in

Chapter 2 (Eveland et al., 2008), for three of the six replicate ovary samples chosen at random

from each of the four developmental stages. The unique key codes in the multiplex adaptors

enabled sample recognition during concurrent sequencing of all 12 samples (Figure 3-6). The

12-sample, multiplex library was sequenced using both the 454 Genome Sequencer 20

instrument (Margulies et al., 2005) and the recently upgraded FLX system (Harkins and Jarvie,

2007). Sequences from both instruments were analyzed collectively.

In total, we obtained 578,088 good-quality, 3'-anchored sequences that represented 22,920

unique consensus mRNAs, each with two or more reads. Although distribution of reads among









multiplexed biological replicate tended to vary (Figure 3-6), overall differences in read

frequency between the four time points were marginal (+ 9,382 reads). Of the 22,920 unique

consensus sequences identified, 21,273 aligned to cDNA sequences using BLASTN in both

ZmGI and IUC databases. Read count for each unique mRNA was normalized to reads per ten

thousand (PTT) and used to profile quantitative changes in transcript abundance during

development.

To determine the extent to which the 70-mer array probes and unique 3'-UTR sequences

matched identical cDNAs, we aligned both sets of sequences to IUC cDNAs using BLASTN. In

total, 17,882 of the 3'-UTR consensus sequences associated with array probes that showed

positive signal intensities (Figure 3-7). There were 5,038 consensus sequences that did not

correspond to an array probe based on our analyses. In addition, 1,344 of the 3'-UTR sequences

associated with genes that were eliminated from the microarray analysis due to undetectable

levels of expression or poor spot flags. Therefore, our 3'-UTR profiling strategy provided

resolution of expression changes that were not detected by microarray analyses. Since sampling

of the transcriptome by microarray hybridization was limited to spotted oligos, this could explain

a portion of the transcripts resolved only by 454-based sequencing.

Comparison of Microarray and Quantitative 3'-UTR Profiles

We compared expression profiles of genes detected by both microarray- and sequence-

based methods. Quantitative data from the 3'-UTR profiles were in accordance with gene

expression trends from the array analyses (Figure 3-8). Sequence-based profiles showed ranges

of expression that spanned three orders of magnitude and enabled quantitative comparisons of

transcript levels between genes (Figure 3-9). Our ability to reliably quantify rarer transcripts was

limited by the depth of sequencing for 12 multiplexed samples. In addition, only 3 biological

replicates were analyzed by 3'-UTR profiling and therefore greater variation was encountered









than with the 6 replicates used for microarray analyses (indicated by standard errors in Figure 3-

8). However, in addition to quantitative, gene-specific profiles, the 454 data also provided

validation for microarray-based expression, especially for those genes with high read frequencies

(Figure 3-8).

To determine whether co-expressed genes of related function from microarray analyses

showed similar expression profiles when examined using the 454-based method, we analyzed

clusters of cell-wall-related genes. The quantitative 3'-UTR profiles for genes involved in cell

wall biosynthesis and modification from clusters 1, 5, and 7 are shown in Figure 4-9. In

addition, we presented in Figure 3-10 an example of four Phenylalanine Ammonia Lyase (PAL)

gene family members from the microarray dataset and compared these to their respective read

counts from the 3'-UTR profile. In both instances, the sequence-based method provided

quantitative resolution of expression for genes with related functional processes as well as

individual gene family members.

Resolution of Near-Identical Transcripts

Comparison of 3'-UTR tags to the 70-mer oligo sequences on the array identified 1,394

array probes which matched two or more consensus sequences. A number of these

disproportionate matches could be explained by misalignment of a given oligo probe and 3'-UTR

sequence due to incomplete genome sequence or proximity of MspI sites to the 3' end. Also, we

cannot rule out the possibility that a small proportion of digests were incomplete or that

premature termination of first-strand cDNA synthesis could have occurred for certain transcripts.

We investigated these possibilities by aligning the corresponding 3'-UTR sequences with cDNAs

in ZmGI and IUC databases and with maize genomic assemblies [MAGI (Fu et al., 2005)]. In

several cases, we resolved expression profiles for two or more near-identical transcripts that









matched a common oligo sequence. Some of these closely-related mRNAs were identified by

small feature polymorphisms such as a SNP or indel in the 3'-454-sequence read (Figure 3-11A).

Also, putative 3'-RNA processing variants were resolved for genes represented by a single array

element (Figure 3-11B). Differential profiles for these near-identical transcripts indicated that

cross-hybridization could lead to confounding results with some microarray analyses.

Conclusions

In this work we determined that a sequence-based 3'-UTR profiling strategy can be used to

analyze global transcriptional changes for twelve multiplexed samples. In addition we showed

that this sequence-based method is an effective compliment to microarray analyses and enabled

gene-specific resolution of expression among complex gene families and near-identical

transcripts. Clustering analyses showed shifts in expression of genes related to specific

functional processes during maize ovary development. Among these were changes in expression

of genes for C and N metabolism, cell growth and differentiation, and for protein synthesis.

These findings were consistent with enhanced C-import by these developing sink tissues and

could provide a basis for understanding regulation of maternal-based adjustment of sink strength

at the transcriptional level.

Materials and Methods

Plant Material

Maize (Zea mays L.), W22 inbred plants were field-grown at the University of Florida

Research farm in Citra, Florida from April to June, 2005. Ears were sampled at four stages of

development: 1) Pre-silk emergence (4 days before pollination [4 DBP]), 2) First silk emergence

(2 days before pollination [2 DBP]), 3) Time of pollination (OT), and 4) Two days after

pollination (2 DAP). Ears sampled at 4 DBP included only those in which longest silks reached

approximately 2 cm below the tip of the husk. Ears sampled at 2 DBP had been protected from









prior pollination by inverting small, paper, ear-capping bags over them. Silks at this stage had

typically elongated about 2 cm beyond the husk and emerged from mid-basal florets. Ears

sampled at OT were collected 1 hour after pollination. All four stages were sampled on each of

seven consecutive days. Florets were from sampled from 3-cm equatorial sections starting

approximately 2 cm from the base of each ear. Ovary-plus-pedicel samples were hand-dissected,

pooled to 100 mg FW, and immediately frozen in liquid N2. Ovary-plus-pedicel samples were

collected from 6 individual ears at each of the 4 developmental stages for subsequent analyses.

RNA Extraction and Target Labeling

Frozen ovary samples (Ig FW) were pulverized in TRIzol (Invitrogen) using a FastPrep

lysis system (Q-BIOgene). Total RNA was extracted as described at www.maizearray.org,

treated with DNase (Ambion), and quantified by NanoDrop (ND-1000, NanoDrop

Technologies). A MessageAmp II kit (Ambion) was used to synthesize cDNA from 1.5 ug total

RNA using an oligo(dT) primer and subsequent cRNA by in vitro transcription and amplification

in the presence of aminoallyl UTP. Each purified, aminoallyl-labeled cRNA sample was

quantified by NanoDrop (ND-1000, NanoDrop Technologies), 6 ug transferred to an RNase-free

eppendorf tube, and dried using a SpeedVac (Vacufuge Concentrator 5301, Eppendorf) at room

temperature. Indirect coupling of Cy Dye esters (Cy3 and Cy5 Mono-Reactive Dye, GE-

Healthcare) to aminoallyl-labeled cRNA and removal of unincorporated dye was conducted

according to protocols at www.maizearray.org.

Microarray Slide Preparation and Target Hybridization

A total of 12 arrays were used (University of Arizona), each represented by a two-slide set

("A" and "B" slides). Information on probe sequences and accession numbers are available at

www.maizearray.org. The 70-mer oligonucleotide probes were cross-linked to glass slides via a

Stratalinker (Stratagene) at 180 mJ. All "A" slides were prepared and hybridized concurrently









on day 1 and "B" slides were prepared and hybridized concurrently on day 2. Slides were

washed, with stirring, in filtered 1% SDS for 5 min at room temperature. To remove SDS, slides

were dipped 10 times in Milli-Q water, 5 times in 100% EtOH, then incubated, while shaking,

for 3 min in 100% EtOH. Each slide was then placed in an individual 50-mL Falcon tube, dried

in a centrifuge at 200 x g for four minutes, and transferred to a new 50-mL Falcon tube.

The following pre-hybridization and hybridization procedures were conducted

concurrently for three sets of four slides. For each slide, 2 of the 4 developmental stages (4

DBP, 2 DBP, OT, 2 DAP) were compared. Four 50-mL Falcon tubes, each containing a single

slide, were filled with pre-warmed (42 C), filtered, pre-hybridization buffer (4g BSA, 100 mL

20X SSC, 4 mL 10% SDS, 296 mL Milli-Q water), incubated with gentle agitation for 2 min,

and then in an oven at 42 C for 45 min. Slides were immediately transferred to Milli-Q water,

submerged 5 times, then moved to a fresh volume of Milli-Q for another 5 submersions.

Isopropanol was used for the final 5 emersion rinses, followed by drying with compressed

nitrogen gas. The labeled cRNA samples were quantified by NanoDrop (ND-1000, NanoDrop

Technologies) and lug of both Cy3- and Cy5-labeled targets were combined for co-hybridization

to arrays based on the experimental design in Figure 3-1B. Each Cy3/Cy5 target mixture was

dried in a SpeedVac (Vacufuge Concentrator 5301, Eppendorf) and then resuspended in 60 uL

1X hybridization buffer (0.55 uL 10% filtered SDS, 13.75 uL 20X filtered SSC, 5.51 uL 50X

Denhards, 16.51 uL Formamide, 18.15 uL HPLC grade water). Hybridizations were carried out

in a darkened room. The hybridization mixture was then denatured by heating to 95 C for 3 min,

transferred to ice for 30 sec, and then centrifuged at 14,000 x g for 2 min to pellet particulates.

Each of four MAUI mixers (model SC, BioMicro) were attached to the glass slides via an

Assembly/Disassembly apparatus (BioMicro) and firmly sealed. The labeled target mixture (50









uL) was injected slowly into the fill port using a positive displacement Eppendorf Combitip

syringe (Brinkmann, Cat No. 022-26-595-4). Fill and vent ports were lightly dried with a

Kimwipe, sealed with adhesive discs, and 4 slides were concurrently hybridized in a single

MAUI hybridization chamber at 42 C on mixing level "D". A total of three, 4-bay MAUI

hybridization chambers were used for concurrent hybridization of all 12 arrays. The slides were

removed from the MAUI after 14 hours of hybridization and submerged in pre-warmed (42 C)

Wash-1 buffer (IX SSC, 0.2% SDS, filtered). Mixers were then gently removed, and slides were

transferred to a glass frame in stirred Wash-1 buffer for a 4-min incubation. The frame was then

transferred to stirred Wash-2 buffer (0.1X SSC, 0.2% SDS, filtered) for 2 min, then to a series of

three, 2-min incubations, each in a fresh, stirred reservoir of Wash-3 buffer (0.1% SSC, filtered).

Slides were then gently removed and dried with compressed nitrogen gas. All washing was done

in a darkened room.

Microarray Data Analysis

Slides were scanned using an Agilent scanner setting of 80 PMT. Scanned images (TIFF

format) were imported, along with the maize array annotation GAL file (version 10.1) available

on www.maizearray.org, into ImaGene Version 6.0. Custom grids were fit to each slide for spot

finding and signal mean, median, and background intensities were extracted. Based on mean

signal differences to negative control spots present on the slide, 16,274 out of 57,452 genes

showed no detectable signal corresponding to a 90% negative control quantile, which was used

as the cut-off value. For the analysis, a gene at a given developmental stage was defined to be

"off" if four of the six biological replicates showed signals at or below background. In addition,

expression of a gene was determined to be below detectable levels if it was defined as "off" at all

four developmental stages.









The "marray" package in Bioconductor software was used to generate background

intensity plots in R for a quality check of the array data. For each combination of slide and dye,

genes having no detectable expression, based on the 90% negative control quantile, were deleted

before fitting an ANOVA model. Poor spots flagged for contamination, saturation, or showed

negative signal intensities after background subtraction, were also eliminated from the analysis.

The signal values were defined to be signal mean-background mean. Data were log-

transformed to meet the normal assumption for fitting an ANOVA model: log(signal mean)

=dye+ treatment+ error. Gene-specific ANOVA models were fitted for each probe across all

developmental stages. The above model was fitted for log (signal median) in hierarchical

clustering analyses using JMP Genomics.

Cluster Analysis using Orthogonal Polynomials

The SAS code for the clustering analysis was adapted from Qu and Xu (2006). Our dataset

included four sequential developmental stages. Therefore, we could only fit cubic polynomials

and the order value was set at 3. The polynomial parameters (developmental stages here) were

re-scaled before constructing the orthogonal polynomials so that only the pattern of the

expression profile was used. Therefore, whether we let development equal (4 DBP, 2 DBP, OT,

2 DAP) or (1, 2, 3, 4), the same set of orthogonal polynomials was generated. A minimum

Bayesian Information Criterion (BIC) was used to determine the optimal number of clusters to

best fit the data (Qu and Xu, 2006). Clustering analysis of least square means (LSMEANS) were

run using log scale and original scale values. Clusters in Figure 3-4 are reported in the original

scale to show maximum differences in gene-specific patterns.

Construction of a 454 3'-UTR library

Three of the six biological replicate samples from each of the four developmental stages

were chosen at random for parallel, 454-based, 3'UTR profiling analyses. Total RNA (5 ug)









from each of the 12 samples was used as template for first-strand cDNA synthesis (MessageAmp

II [cDNA synthesis only], Ambion) and primed with 6 pmol biotinylated (T12) B-adapter

(modified from Margulies et al. [2005]) oligonucleotide. Sub-libraries were constructed as

described in Eveland et al. (2008). Each sample was tagged with one of 12 unique, four-base

multiplex keys in the ligated A-adaptor oligonucleotide. Sub-libraries were then pooled. The

desired 5'-A-cDNA-B-3' template strand was eluted with 100 mM NaOH, neutralized, and

concentrated on a Qiagen column (Margulies et al., 2005). Sequencing was conducted as per

Margulies et al. (2005) using a 454 GS-20 instrument for one sequencing reaction and the 454

FLX system for a second sequencing reaction. Data were combined and reads were filtered for

correct ligation junctions and validation of the error detection key (Eveland et al., 2008).









Table 3-1. Number of genes associated with each co-expression cluster and percentage having
significant differences in expression or large fold-changes during development.
l St g s Percent Genes with Percent large
Total Significant genes
Cluster a significant large fold- fold-change
genes (FDR 20%) (%)b change (%)c


1 407 26 6.39 44 10.81
2 38,350 1,280 3.34 3450 9
3 420 11 2.62 6 1.43
4 102 4 3.92 5 4.90
5 639 30 4.69 62 9.70
6 247 19 7.69 6 2.43
7 521 48 9.21 86 16.51


a40,686 genes were defined as "on" and used in clustering analyses. bPercentage of genes per
cluster that showed significant differences in expression during development using a FDR of
20%. "Percentage of genes per cluster that showed expression fold changes > 2 during
development.











Table 3-2. Annotated genes from clusters 1 and 5 that showed significant differences in
expression (FDR 20%) or large fold changes (> 2) during development.


MZ
Assignment
Cluster 1
MZ00003791
MZ00027467
MZ00017851
MZ00042055
MZ00004477
MZ00018463
MZ00021835
MZ00036098
MZ00012431
MZ00004658
MZ00030532
MZ00044388
MZ00057049
MZ00014170
MZ00040508
MZ00030872
MZ00043399
MZ00024089
MZ00003640
MZ00024317
MZ00044854
MZ00056428
MZ00018500
MZ00035817
MZ00040858
MZ00041925
MZ00030106
Cluster 5
MZ00016522
MZ00024039
MZ00024043
MZ00024066
MZ00025588
MZ00025431
MZ00037253
MZ00026127
MZ00012051
MZ00016305
MZ00024782
MZ00036659
MZ00013279
MZ00027021
MZ00027846
MZ00028568
MZ00028196
MZ00004010
MZ00013531
MZ00013851
MZ00025380


*Fold-
P-Value
change


0.003
0.013
0.020
0.003
0.010
0.023
0.022
0.005
0.045
0.0007
0.006
0.035
0.035
0.397
0.002
0.034
0.001
0.0003
0.0008
0.122
0.002
0.001
0.079
0.197
0.135
0.236
0.005

0.004
0.004
0.005
0.001
0.006
0.002
0.015
0.005
0.0006
0.002
0.004
0.003
0.012
0.0002
0.001
0.0008
0.004
0.016
0.145
0.856
0.008


1.595
0.873
0.854
0.755
1.070
1.635
1.516
1.150
0.948
1.935
1.999
1.933
1.758
1.988
0.800
1.600
0.482
0.266
0.384
1.126
1.537
1.383
0.733
1.058
1.028
1.376
0.270

1.292
0.856
0.589
0.551
0.751
0.434
1.578
0.899
0.587
0.301
0.336
0.208
0.714
0.582
0.611
0.287
0.456
1.154
0.884
1.747
0.747


Putative annotation [Species]


ATP-binding-cassette transporter [0. sativa]
Putative phosphate translocator [0. sativa]
O-methyltransferase ZRP4 [Z. mays]
4-coumarate--CoA ligase 4CL2 [L. perenne]
N-hydroxycinnamoyl/benzoyltransferase [A. thaliana]
Peroxidase [0. sativa]
N-hydroxycinnamoyl/benzoyltransferase [A. thaliana]
Metallothionein-like protein 1 [Z. mays]
Putative laccase [0. sativa]
Laccase [0. sativa]
Protease inhibitor [0. sativa]
Protease inhibitor [0. sativa]
Putative protease inhibitor [0. sativa]
Cystatin [T. aestivum]
Cytochrome f [Z. mays]
Uclacyanin 3-like protein [0. sativa]
Citrate lyase [A. thaliana]
Putative DNA topoisomerase II [0. sativa]
Phospholipase D 2 [0. sativa]
Gipl-like protein [Petunia x hybrida]
Early nodulin 75 precursor-like protein [0. sativa]
Beta-expansin [0. sativa]
Xyloglucan endo-1,4-beta-D-glucanase [H. vulgare]
Proline-rich protein [0. sativa]
Proline-rich protein-like [0. sativa]
Proline-rich protein-like [0. sativa]
Beta-glucuronidase [Z. mays]

Early nodulin 75 precursor-like protein [0. sativa]
Photosystem-I PSI-F chain precursor [H. vulgare]
Chlorophyll a/b-binding protein precursor [Z. mays]
Oxygen-evolving enhancer protein precursor [Z. mays]
Proteinase inhibitor [G. max]
Bowman-Birk type trypsin inhibitor [H. vulgare]
Subtilisin/chymotrypsin inhibitor [Z. mays]
Development regulation OsNAC4 [0. sativa]
Putative bHLH transcription protein [0. sativa]
Putative CTP synthase [0. sativa]
Starch branching enzyme I [Z. mays]
Putative trehalose-6-phosphate synthase [0. sativa]
Putative triosephosphate isomerase [0. sativa]
Putative 6-phosphogluconolactonase [0. sativa]
Auxin efflux carrier protein [M. truncatula]
Viviparous-14 protein [Z. mays]
Syntaxin [G. max]
Putative serine/threonine kinase [0. sativa]
Light regulated protein precursor [0. sativa]
Superoxide dismutase [Cu-Zn] [Z. mays]
Putative fiddlehead-like protein [0. sativa]


Function


Transport
Transport
2 metabolism
2 metabolism
2 metabolism
2 metabolism
2 metabolism
Redox
Redox
Redox
Protease inhibitor
Protease inhibitor
Protease inhibitor
Protease inhibitor
Photosynthesis
Photosynthesis
TCA cycle
Nucleic acid
Lipid metabolism
Gibberellin
Nodulin-related
Cell wall
Cell wall
Cell wall
Cell wall
Cell wall
Cell wall

Nodulin-related
Photosynthesis
Photosynthesis
Photosynthesis
Protease inhibitor
Protease inhibitor
Protease inhibitor
Transcription
Transcription
Nucleic acid
C-metabolism
C metabolism
Glycolysis
OPP
Auxin
ABA
Secretory trafficking
Signaling
Signaling
Redox
Lipid metabolism


Large fold change = ln(2) = 0.7











Table 3-3. Annotated genes from cluster 7 that showed significant differences in expression
(FDR 20%) or large fold changes during development.


MZ
Assignment
MZ00023389
MZ00054418
MZ00019305
MZ00023972
MZ00024296
MZ00027767
MZ00026125
MZ00037083
MZ00013363
MZ00035394
MZ00020674
MZ00035052
MZ00036117
MZ00027578
MZ00041508
MZ00014474
MZ00014516
MZ00015368
MZ00016005
MZ00016987
MZ00025089
MZ00041601
MZ00057269
MZ00015899
MZ00017982
MZ00018613
MZ00029356
MZ00043035
MZ00041278
MZ00018945
MZ00026029
MZ00038502
MZ00041611
MZ00041613
MZ00019713
MZ00041203
MZ00033558
MZ00015327
MZ00018436
MZ00024269
MZ00036403
MZ00032935
MZ00035277
MZ00043252
MZ00043254
MZ00014807
MZ00025294


p-Value
0.001
0.0001
0.015
0.014
0.001
3.61E-05
0.0009
1.70E-05
0.001
0.008
0.0001
0.019
0.051
0.003
0.003
0.461
0.578
0.0007
0.0001
3.47E-0
5.45E-05
0.002
0.005
0.008
0.0003
0.0002
1.33E-06
0.0001
0.013
0.004
0.001
0.0001
0.002
3.75E-05
0.076
0.465
0.001
0.01
0.032
9.64E-06
9.78E-05
0.017
0.015
1.21E-05
3.44E-07
0.002
0.035


*Fold-
change
0.979
2.128
0.706
0.939
2.606
3.575
0.607
1.732
1.355
0.764
3.085
0.827
0.834
0.716
0.402
1.893
2.818
0.768
1.788
2.513
1.113
1.227
0.572
0.806
2.034
0.973
1.792
1.463
2.648
0.923
1.229
1.258
1.041
1.825
0.779
0.854
0.960
0.945
1.136
1.739
1.932
0.713
1.240
1.884
2.119
0.585
6.653


Large fold change = ln(2) =0.7


Putative annotation [Species]
ABC transporter protein [0. sativa]
WRKY transcription factor 45 [0. sativa]
Homeodomain protein [0. sativa]
OsNAC3 protein [0. sativa]
2-oxoglutarate-dependent oxygenase [Z. mays]
Naringenin 3-dioxygenase [Z. mays]
pyridoxal kinase-like protein SOS4 [A. thaliana]
Metallothionein-like MT-1 [Z. mays]
Metallothionein-like protein 1 [Z. mays]
Cytochrome P450 monooxygenase [Z. mays]
Elicitor-inducible cytochrome P450
Pathogenesis related protein-5 [Z. mays]
Pathogenesis related protein-5 [Z. mays]
Peptidase family Al [0. sativa]
Ubiquitin-conjugating enzyme [0. sativa]
26S proteasome regulatory particle [0. sativa]
Putative ubiquitin-protein ligase [0. sativa]
Gaiacol peroxidase [G. hirsutum]
Peroxidase [Z. mays]
Polyphenol oxidase [T. aestivum]
Inducible PAL[T. aestivum]
B-glucosidase aggregating factor precursor [Z. mays]
Caffeoyl CoA 3-O-methyltransferase [Z. mays]
Cinnamoyl CoA reductase [Z. mays]
Phytosulfokine peptide precursor [0. sativa]
Putative MtN3 [0. sativa]
Putative MtN3 [0. sativa]
Chitinase PRm 3 [Z. mays]
Chitinase A [Z. mays]
Storage/lipid transfer protein (LTP) [0. sativa]
Putative lipid transfer protein [0. sativa]
Nonspec lipid-transfer protein precursor [Z. mays]
Phospholipid transfer protein [Z. mays]
Phospholipid transfer 9C2 precursor [Z. mays]
GDSL-motif lipase/hydrolase protein [0. sativa]
Lipid transfer protein [S. italica]
Gibberellin-stimulated transcript 1 like [O sativa]
Gibberellin-stimulated transcript 1 like [O sativa]
ACC oxidase [0. sativa]
Proline-rich protein [Z. mays]
Proline-rich protein [Z. mays]
Xyloglucan fucosyltransferase [A. thaliana]
Glycosyltransferase QUASIMODO 1 [A. thaliana]
Glutamine synthetase [Z. mays]
Glutamine synthetase [Z. mays]
Nitrate transporter NRT1-5 [0. sativa]
Putative NAD svnthetase [0. sativa]


Function

Transport
Transcription
Transcription
Transcription
2 metabolism
2 metabolism
2 metabolism
Redox
Redox
Redox
Redox
Defense
Defense
Proteolysis
Proteolysis
Proteolysis
Proteolysis
Phenypropanoid
Phenylpropanoid
Phenylpropanoid
Phenylpropanoid
Phenylpropanoid
Phenylpropanoid
Phenylpropanoid
Differentiation
Nodulin-related
Nodulin-related
Nodulin-related
Nodulin-related
Lipid metabolism
Lipid metabolism
Lipid metabolism
Lipid metabolism
Lipid metabolism
Lipid metabolism
Lipid metabolism
Gibberellin
Gibberellin
Ethylene
Cell wall
Cell wall
Cell wall
Cell wall
N metabolism
N metabolism
N metabolism
N metabolism









4DBP 2DBP OT 2DAP





Developmental progression
4DBP: pre-silk emergence
2DBP: silk emergence
OT: time of pollination
2DAP: two days post-pollination

B
4DBP 2DBP


3 4


11 F2 11


I- -------O
OT 2DAP
cy3 -
cy5

Figure 3-1. Experimental design used to compare transcriptional profiles during maize ovary
development. A) Stages of development sampled at two-day intervals included pre-
silking at 4 days before pollination (4DBP), silk emergence at 2 days before
pollination (2DBP), time of pollination (OT), and 2 days after pollination (2DAP)/
post-fertilization. B) Transcript profiles for each developmental stage were analyzed
by all possible pair-wise comparisons and included dye swaps.
















































IL QI I
mm <(
^CM0 CM


Figure 3-2. A heat map generated by a two-way, hierarchical clustering of expression profiles
for 856 genes that showed significant changes in transcript abundance during
development (FDR 10%). Based on this hierarchy, genes clustered into three main
groups: 1) mRNAs that decreased during development, 2) increased during this time,
or 3) showed combinations of both.













(5%)
Secondary metabolism (10%)
(3.9%)

Defense/ stress-related (12.5%)

Signaling (16.6%)

(12.2%)



Proteolysis (8.8%)
Nucleic acid-related (7.2%) (5%)
(2.5%)
Transcription (8.8%) (5%)







Ovary development

Figure 3-3. Annotated genes from hierachical clusters 1 and 2 (Figure 3-2) that showed either
decreased or increased expression during maize ovary development were classified
into functional categories based on Gene Ontologies. The most prominent differences
in functional representation between the two clusters are noted.














Cluster 1
n=407


1.0 15 20 25 30 35 40
Time

Cluster 5
n=639


Cluster 2
n=38350


10 15 20 25 30 35 40
Time

Cluster 6
n=247


Cluster 3
n=420


II
10 15 20 25 30 35 40
Time


Cluster 4
n=102


10 15 20 25 30 35 40
Time


Cluster 7
n=521


10 15 20 25 30 35 40
Time


10 15 20 25 30 35 40
Time


10 15 20 25 30 35 40
Time


Figure 3-4. Co-expression clusters of genes were analyzed after fitting expression patterns
during development to orthogonal polynomials. Data shown are based on original
signal intensity values after removing the gene-wise mean (y-axis). Cooridnates for
the polynomials (1, 2, 3, and 4) along the x-axis correspond to 4DBP, 2DBP, OT, and
2DAP, respectively. A) Cluster assignments for the total set of 40,686 genes that
showed positive signal intensities on the array. Numbers of genes associated with a
given cluster are noted. B) Genes that showed significant changes in expression
during development (FDR 20%) are highlighted in red.














B




8
5 s




S-


Cluster 1

n=407


10 15 2.0 25 30 35 40
Time

Cluster 5
n=639
8












10 15 20 25 30 35 40
Time



Figure 3-4 continued.
Figure 3-4 continued.


Cluster 2

n=38350


UJ c
a C-


10 15 20 25 3.0 35 40
Time

Cluster 6
n=247


10 15 20 25 30 35 40
Time


Cluster 3

n=420


8
8-




o


Cluster 4

n=102


8




uJ cr
I C-


10 15 20 25 30 35 40
Time

Cluster 7
n=521


10 15 20 25 30 35 40


10 15 20 25 30 35 40
Time


0

00

o8-

1 o


8



8
ru rr
E
Y o-


OSO%










Clusters 1 & 5


Signaling (16%)





Redox-related (7%)

Proteolysis (8%)


Nitrogen and amino acid metabolism (3%)
/ Carbohydrate metabolism (3%)
S--Cell-wall-related (5%)
/ Chromatin-related (3%)
SCell growth (6%)


Energy metabolism (7%)


\ Phenylpropanold biosynthesis (1%)
Protein synthesis (5%) Photosynthesis (3%)


Cluster 7


Signaling (3%)

Redox-related (3%)

Proteolysis (8%)




Protein synthesis (24%)


Nitrogen and amino acid metabolism (7%)
Carbohydrate metabolism (1%)
^/N .-Cell-wall-related (5%)
S Chromatin-related (5%)
-Cell growth (2%)


Energy metabolism (7%)


Phenylpropanold blosynthesis (3%)


Figure 3-5. Annotated genes in clusters 1, 5, and 7 were grouped by functional processes using
Gene Ontologies. Results from clusters 1 and 5 (gene expression fluctuates during
development) were combined and compared to those from cluster 7 (gene expression
increases during development). Prominent differences in functional processes during
development are noted as well as examples of those that remained constant.










# reads in consensus # consensus
sequences (2 2 reads) sequences
Jm AACA: 41,585
4 DBP 137,986 ] ACGG: 52,497 16,147
I AGGT: 43,904
SIAAGC: 32,968
2 DBP 130,799 M l ACTT: 45,929 16,112
1 AGTA: 51,902
SMI AATG: 43,354
OT 141,347 ]UAGAC: 31,699 17,042
S ATAG: 66,294
mACAA: 70,200
4 DAP 167,956 Jl AGCG: 32,659 17,264
m ATGA: 65,097
Total: 578,088 22,920

Figure 3-6. Experimental design and sequencing results for a 12-sample, multiplexed, 3'-UTR
library. Unique key codes in the multiplex adaptor were used to distinguish three
biological replicates at each of the four stages of maize ovary development. Total
reads of consensus sequences (unique mRNAs represented by 2 or more reads) are
shown for each of the tags as well as the number of unique transcripts identified at
each developmental stage.










454 consensus tags
total: 22,920


array probes
total: 40,686


Figure 3-7. The 3'-UTR consensus tags (unique mRNAs represented by 2 or more reads) were
compared to the 70-mer microarray probe sequences to determine extent of overlap.

























o oE Cluster 5
Light-regulated protein precursor (5) Early nodulin precursor (5)
6-
2-

4-






Auxin-independent growth promoter protein (2) Inositol polyphosphate 5-phosphatase (2)
3-


2
> 06- Cluster 2
0




4- Proline-rich protein (7) 4. Galacol peroxidase (7)
S30-
U
s) 33

2-


10 1

.-- ------- _--- "___
02
Glbberellin-stimulated transcript (7) 6- Beta-glucosidase aggregating factor precursor (7)
12

4-



4 2-



Glutamine synthetase (7) 14 Glycosyltransferase-1 (QUASIMODO) (7)

3-





1 -
0s












genes that matched array probes associated with clusters 2,5, and 7.
/ 1 I












genes that matched array probes associated with clusters 2,5, and 7.
















10- .-



.1


**** r
**


.01 -


- Clusters 1 &5
MZ00024544 Glycine-rich protein 2 [AI sylveshis]
MZ0008500 Xylaglucan endo-1,4-beta-D-glucanase [H. vulgare]
MZ00042900 Osr40g3 glycine-rich cell wall protein [0. sativa]
MZ0016735 Cellulose synthase-1 [Z. meys]
MZ00015880 Callosesynthase-like [0. sativaI
MZ003t80g Xyloglucan endo-1,4-beta-D-glucanase [4. vulgarej
MZ00041809 Proline-rich protein Z. maya]
MZ00O30105 Beta-glucuronldase [0. satlva
- Cluster 2
MZ00027309a Beta-galactosidase precursor/lactase [A. officinalis]
MZ00027309b Bet -galactosidase precursor/ lactase [A. offcinalias
MZ00027418 Putative beta-1,3-glucanase [0. etiva]
MZ00021494 Putative xyloglucan enda-1,4-beta-D-glucanase [0. satva]
MZ00030125 Etensln-IIlke protein [. seUva]


S l.. Cluster7
4DBP 2DBP OT 2DAP MZ00036403 Proline-rich protein [Z maya]
S MZ00035277 Glycosyltransferase QUASIMODO1 [A. theliene]
Ovary development MZ00043552 Arabinoxylan arabinofuranohydrolase isoenzyme AXAH-II [H. vulgare]
MZ00043429 Proline-rich-ike protein [A. offcinkals]
MZ00043513 Putative arabinogalactan-like protein [0. saefvae
MZ00024711 Xyloglucan endoransglucosylase/hydrolase [T. aestivum]
MZ00032935 Xyloglucan fucosyltranmferas [A. thaianal


Figure 3-9. Quantitative 3'-UTR profiles for cell-wall-related genes with matches to array
probes in clusters 1, 2, 5, and 7 (see Figure 3-4). Transcript abundance was
quantified by read frequency and was plotted on a log scale to compare ranges of
expression during maize ovary development.










Phenylalanlne Ammonia Lyase (PAL)


array data PAL2
PAL4
PAL3
PAL1


Ovary development


PAL1



PAL2
PAL4

PAL3


I
2DAP


Ovary development

Figure 3-10. Transcript profiles of four Phenylalanine ammonia lyase (PAL) gene family
members were compared by microarray- and sequence-based methods. The 3'-UTR
profiles showed quantitative comparisons among gene family members during
development. Microarray-based profiles are depicted as heat maps (light-to-dark
denotes low-to-high gene expression).


9






e 3
ac
'-i


a,


I
4DBP


I I
2DBP OT











454


2 Chil


S Chi2 ..
0

15

10 m vosin-like1

5
mvosin-like 2


6-

3-
Ao2
0 2DBP OT 2DAP
4DBP 2DBP OT 2DAP


array
Chalcone Isomerase






mvosin-like (indel)


Amino oxidase (SNP)



Ovary development


Ovary development

Figure 3-11. Resolution of quantitative transcript profiles by the 3'-UTR sequencing approach
for near-identical mRNAs that matched a single array probe. Microarray-based
profiles are depicted as heat maps (light-to-dark denotes low-to-high gene
expression). A) Examples of near-identical paralogs identified by small feature
polymorphisms in the 3'-UTR. Type of polymorphism is noted in parentheses. B)
Resolution of differential transcript profiles for 3'-RNA processing variants.












454 data
15

10
XET-1
5
XET-2


2 H13-1



S H13-2
1"3 ^ ______ ~ '


4DBP 2DBP OT
Ovary development


array data
Xyloglucan B-
Endotransferase (XET)


Minor hito


Minor histocompatibility
antigen (H13)



Ovary development


2DAP


Figure 3-11 continued.


m









CHAPTER 4
C ALLOCATION AND USE IN DEVELOPING MAIZE FEMALE FLORETS

Introduction

The allocation and use of photosynthate is essential to whole plant function, particularly

during reproductive development. Modulation of carbohydrate synthesis in source leaves and C

transport to growing sinks is maintained by a system of nutrient-based checks and balances.

Reproduction is a costly process and necessitates abundant photosynthate for sink establishment

and growth (Sturm and Tang, 1999; Paul and Foyer, 2001; Koch, 2004), including respiration

(Bustin and Goldschmidt, 1999) and post-pollination fruit development or grain-fill (Cruz-

Aguado et al., 1999; Maitz et al., 2000; Borras et al., 2003; Weschke et al., 2003). Reproductive-

based food stuffs such as cereal grains, fruits, and nuts are essential agricultural commodities in

all areas of the world. Therefore, regulation of carbohydrate allocation and use in reproductive

sinks has been a central aspect of breeding and crop improvement programs for centuries (Boyer

and Westgate, 2004).

In maize, the pre- and early post-pollination phases of reproductive development are

critical for seed set and subsequent grain yield. Sufficient import of C resources is required by

the developing female inflorescence for both silk (stigma) exsertion and early grain

establishment (Zinselmeier et al., 1995; McLaughlin and Boyer, 2004b; Borras et al., 2007).

Optimal source-to-sink C transport is dependent on water available and therefore maize is most

sensitive to drought stress during the pre- and early post-pollination period (Westgate and Boyer,

1985; Zinselmeier et al., 2002). Although efforts to enhance yield potential have focused on ear-

specific traits during the silking and pollination period, relatively little is known about the

mechanisms underlying female floret growth and development at the molecular/metabolic level.









In maize, dry weight accumulation in the female inflorescence, or ear, was directly linked

to whole plant growth rate under conditions analyzed by Borras et al. (2007). Based, on their

findings, rate of silk exsertion is strongly influenced by photosynthetic capacity and allocation to

the developing ear. In addition, silk elongation is compromised under water-limiting conditions

by a decrease in turgor potential (Westgate and Boyer, 1985). Carbon allocation from source to

sink is regulated, in part, by acid invertase activities which catalyze the irreversible cleavage of

sucrose to glucose and fructose. Isoform-specific localization of these sucrose-metabolizing

enzymes determines their activity at either symplastic or apoplastic sites of phloem unloading

(Tymowska-Lalanne and Kreis, 1998; Sturm, 1999; Godt and Roitsch, 2006). Invertases are

typically characterized by gene families, members of which are differentially regulated based on

temporal-, spatial-, and/or isoform-specificity (Xu et al., 1996; Godt and Roitsch, 2006; Huang et

al., 2007).

The conversion of one molecule of sucrose to two hexoses provides substrates for osmotic

adjustment and turgor-based expansion. Invertase activities immediately outside the transport

path for assimilates entering sink tissues can thus generate an enhanced turgor gradient within

the phloem from source to sink (Lalonde et al., 2004; Carpaneto et al., 2005). In addition, both

sucrose and its hexose products can act as signaling molecules that regulate gene expression

based on the carbohydrate status of the cell (Koch, 1996, 2004; Smeekins, 2000; Rolland et al.,

2006). Certain invertase isoforms are sugar-responsive at the transcriptional level and regulated

by sucrose availability and/or accumulation of their own hexose products (Xu et al., 1996;

Huang, 2006). Studies in maize, Arabidopsis, poplar, tomato, and rice have each revealed a pair

of vacuolar invertase isoforms with reciprocal responses to sugar (Xu et al., 1996; Tymowska-

Lalanne and Kreis, 1998; Fridman and Zamir, 2003; Cho et al, 2005; Huang, 2006; Bocock et al.,









2008). Whether conserved or independently acquired in distinct species, such reciprocal

regulation provides a mechanism for fine adjustment of gene expression to in response to "feast"

or "famine" conditions (Koch, 1996; Smith and Stitt, 2007).

Vacuole-localized, soluble invertase activity is central to maintaining an osmotic gradient

for symplastic phloem unloading during the initial phases of sink cell expansion (Sturm, 1999;

Koch, 1996; 2004; Godt and Roitsch, 2006). Post-fertilization, the activities of cell wall-bound

invertases provide essential hexoses to the developing embryo across the apoplastic

maternal/filial barrier. Both vacuolar and apoplastic invertases have been implicated as key

enzymes during the pre- and early post-pollination phases of maize female reproductive

development. A mutation in a maize apoplastic invertase, ZmlNCW2, resulted in a small kernel,

or miniature phenotype (Cheng et al., 1996). Drought-induced repression of vacuolar invertases

limited carbon influx to the developing ovary and thus led to growth inhibition and/or abortion

(Zinselmeier et al., 1995; Andersen et al., 2002; McLaughlin and Boyer, 2004b).

Expression of invertases and other sugar-responsive genes is modulated, in part, by

carbohydrate availability during sink establishment and growth. In the maize ovary, transient

starch reserves are remobilized to promote growth and development during periods of low C

availability, or "famine". Extended periods of "famine", such as under drought stress, can

deplete essential C reserves and ultimately lead to ovary abortion (Zinselmeier et al., 1995;

Westgate and Boyer, 1986; McLaughlin and Boyer, 2004a). Sucrose metabolizing activity is

also associated with developmental progression in sink organs. For example, while a greater

hexose-to-sucrose ratio favors growth and expansion, higher levels of sucrose promote

differentiation and storage (Koch, 1996; 2004; Winter and Huber, 2000). Accordingly, sucrose

synthase activity predominates during sink maturation, generating a one-to-one, sucrose-to-









hexose conversion, with UDPG as an additional product that can serve as a substrate for cell wall

biosynthesis (Koch et al., 2000; Koch, 2004).

Carbohydrate sensing is also mediated by nitrogen availability and metabolism (Koch,

1997; Coruzzi and Bush, 2001; Cooke et al., 2003; Palenchar et al., 2004). As such, both C and

N metabolite signals, and their relative ratios, can affect whole plant source/sink relations.

Nitrate, and its assimilated products such as glutamine, can act as signals for N availability

(Coruzzi and Bush, 2001; Foyer et al., 2003). Gln has been shown to feedback-inhibit nitrogen

uptake and reduction, while C metabolites promote up-regulation of genes involved in N

acquisition and metabolism (Coruzzi and Bush, 2001). Therefore, in growing C sinks, genes for

N assimilation are up-regulated, as are those involved in proteolysis, lipid synthesis, and storage

(Koch, 1997; Foyer et al., 2003). In addition, source-sink adjustment of resources may include

endogenous hormonal cues and/or responses to environmental stimuli. Accordingly, regulation

of acid invertases at the transcriptional level by sucrose (Xu et al., 1996; Koch, 1996; 2004), has

also been shown to respond to hormones such as abscisic acid (Trouverie et al., 2003), ethylene

acid (Linden et al., 1996), cytokinin (Lara et al., 2004), and auxin (Long et al., 2002).

Understanding nutrient-based modulation of sink strength and associated links to

developmental progression will ultimately enhance yield potential in maize (Borras et al., 2003;

Boyer and Westgate, 2004; Barnabas et al., 2007). Much of the work to date investigating sugar-

based effects on developmental- and hormonal-based regulatory networks has focused on

Arabidopsis as a model system. Extending key insights from these studies to crop systems will

be important in determining areas of functional significance for agricultural applications.

Mutations in key regulators that link carbohydrate metabolism and development have revealed

phenotypes of agricultural relevance, e.g. RAMOSA3 (encoding a trehalose-phosphate









phosphatase) in maize (Satoh-Nagasawa et al., 2006). In addition, genotype differences in rate

of exsertion and yield performance under source-limiting conditions have been identified (Bruce

et al 2002), however the underlying mechanisms remain obscure.

In this work, we compared C-allocation (as determined by dry weight deposition and

metabolism) among individual floral organs during pre- and early post-pollination maize female

reproductive development. Spatial and temporal regulation of sink strength and sucrose-

metabolizing activity of a specific vacuolar invertase were tested under normal growth

conditions. An apparent shift in sink strength to the developing ovary during the pollination

period coincided with co-expression of genes related to phenylpropanoid biosynthesis and N

metabolism from profiling experiments in Chapter 3. Data presented here provide evidence for

changes in gene expression that associate with C allocation and related expansion or

differentiation processes in developing maize ovaries.

Results

Staging of Pre-Pollination Floral Development

To establish a foundation for molecular and metabolic analyses of developing maize

female florets, we characterized a series of individual growth stages that occur prior to

pollination. Immature ears were sampled from field-grown, W22 inbred maize plants at two-day

intervals from time of whole-ear silk expansion to two days post silk emergence. Florets from

the mid-to-basal equatorial region of a given ear (Figure 4-1A) were hand-dissected into ovary-

plus-pedicel, with and without subtending floral organs lemmaa, palea, and glumes), and silk.

Physical characteristics and expression of a developmental marker were used to assign

dissected florets to a stage from 1 to 7, each of which corresponded to sequential points of ear

development from 12 days before pollination (DBP) to the time of pollination (Figure 4-1B).

Silks of stage 1 florets were approximately 6 cm in length and elongated about 1.7 centimeters









per day during pre-pollination development. Ovary fresh weight increased exponentially, but

very slowly, and in proportion to linear increases in silk length (Figure 4-1C).

In addition, expression of ZAG2, a floral homeotic gene, was used as a molecular marker

for development. Quantification of Zag2 mRNAs was used to validate the assignment of floret

samples to a given stage (Figure 4-ID). ZAG2 is expressed specifically in carpels and shares

homology to the APETALA floral identity "C" gene in Arabidopsis (Schmidt et al., 1993). Based

on these combined physical and molecular markers, 6 individual ears at each of the 7

developmental stages were used in subsequent analyses.

Pre-Pollination Carbon Allocation and Sucrose Use in Individual Floral Organs

One goal of this work was to determine partitioning of carbon and water as evident in dry

and fresh weight accumulation among individual floral organs prior to pollination. We

quantified dry and fresh weights for 1) ovary-plus-pedicel, 2) ovary-plus-pedicel including

subtending floral parts, and 3) silk over a developmental time course (Figure 4-2A). Because

individual organs could not be effectively separated at stages 1 and 2, we excluded them from

this analysis. Per-floret weights were based on measurements of six individual equatorial florets

from each of six separate ears at a given developmental stage. Dry and fresh weights for

subtending floral parts were determined by subtracting respective ovary-plus-pedicel weights.

Deposition of dry weight in the silk was nearly linear and increased a total of three-fold

during pre-pollination growth. Fresh weight accumulation was also linear and paralleled

increases in silk length during development as shown in Figure 4-1C. A three-fold gain in fresh

weight was observed in the subtending floral organs immediately prior to pollination, however

dry weight accumulated to a lesser degree. The resulting increase in ratio of fresh/dry weight

was linear during the period 4 DBP to OT (Figure 4-2B) and is consistent with an osmotically-

driven expansion of subtending floral organs. Also compatible with turgor-based silk elongation









was the maintenance of a substantial fresh/dry weight ratio in silks during pre-pollination growth

(Figure 4-2B). The prominence of dry and fresh weight gain indicated substantial sink strength

in silks prior to pollination. Although C accumulated to a lesser degree in the ovary-plus-pedicel

during pre-pollination growth, rate of fresh weight increase was maintained proportional to that

of silk elongation

Temporal and Spatial, IVR2-Based Sucrose Use in Developing Female Florets

Based on the linear increase in fresh-to-dry weight ratio in subtending floral parts, we

hypothesized that hexoses generated via soluble invertase activity could provide substrates for

osmotic-based expansion in these tissues. We quantified the expression of two vacuolar

invertase isoforms, IVR1 and IVR2, in ovary-plus-pedicel samples, with and without subtending

floral parts, during this pre-pollination period of expansion (Figure 4-3A). Levels of Ivr2

mRNAs were significantly higher in subtending floral parts than in samples of ovary-plus-

pedicel alone, whereas levels of vr] transcripts were similar between fractions. Consistent with

expression oflvr2, hexose-to-sucrose ratios were at least 2-fold higher in whole florets compared

to ovary-plus-pedicel samples at each stage (Figure 4-3B). These results suggest that soluble

invertase activity in subtending floral parts is predominantly IVR2-specific.

Since rapidly growing silks also maintained high fresh/dry weight prior to pollination, we

hypothesized that IVR2 might provide essential hexoses for turgor-based expansion in silks. We

observed a linear increase in cell size measured at the base of the silk during pre-pollination

growth (Figure 4-4A). These data are consistent with work of Westgate and Boyer (1985) who

suggested that expansion, rather than cell division, was the mechanism behind silk elongation.

We quantified Ivr2 mRNAs in whole silk tissues sampled during development in three separate

field seasons (Citra, Florida fall plantings 2003, 2004, and spring planting 2004) (Figure 4-4B).

All silks were collected at 9AM to exclude possible variation due to diurnal effects (Figure 4-5).









During post-emergence stages, only silks that had exserted from the husk were collected. Levels

oflvr2 mRNAs tended to increase during early stages of development and had dropped to nearly

undetectable limits by two days after pollination. Previous work had also demonstrated a

coincident cessation of silk elongation after pollination (Westgate and Boyer, 1985).

Interestingly, decreased levels oflvr2 mRNAs were apparent at silk emergence from the husk,

but prior to pollination. This finding indicated that pollination was not the primary signal for

repression of IVR2. In addition, decreases in soluble acid invertase activity paralleled those of

Ivr2 transcript abundance (Figure 4-4C). A decrease in the hexose/sucrose ratio by

approximately 50% was shown between time of pollination and two days post-pollination in

silks (Figure 4-4D). Although the latter was observed post-pollination, it may well have resulted

from pre-pollination events and been more dependent on these than previously recognized.

Previous studies have provided evidence for diurnal regulation of invertase-mediated

expansion (Gonzalez et al., 2005). Also, Westgate and Boyer (1985) showed that rate of silk

elongation was highest pre-dawn. To evaluate the sugar composition of the expanding silk tissue

over a diurnal time course, we collected newly-emerged silks over a continuous 24 hours

including time points at 10 AM, 4 PM, 10 PM, 2 AM, and 6 AM. Sucrose and hexose

concentrations were quantified and results showed that hexoses were most abundant in silks at

6AM, or pre-dawn (Figure 4-5).

We also compared expression profiles oflvr2 and a cell-wall bound invertase, Incw2

during pre- and early post-pollination floret development. While Ivr2 mRNAs were maintained

at relatively constant levels in ovaries, Incw2 expression is significantly enhanced post-

pollination and is specific to the developing ovary-plus-pedicel (Figure 4-6). The observed

spatial and temporal expression of Incw2 is consistent with its proposed function in providing









hexoses to the symplastically-isolated, developing filial tissues (Cheng et al., 1996; Kladuik et

al., 2004).

Post-Pollination C Accumulation in Pedicels

To approximate the extent of change that might be occurring during C deposition in

developing ovaries and pedicels after pollination, we quantified dry and fresh weights for 1)

ovary-plus-pedicel, 2) ovary only, and 3) pedicel only. Ovary-plus-pedicel samples were left

intact or further dissected into ovary and pedicel fractions with a razor blade. Pedicel samples

included the transfer region. Increases in dry and fresh weights were evident in ovary-plus-

pedicel during early post-pollination development, however dry weight in the pedicel alone

almost doubled between 2 and 4 days after pollination (DAP) (Figure 4-7). Sucrose and hexoses

were also quantified in the above samples (Figure 4-8). While hexose-to-sucrose ratio was

maintained in ovary tissues (developing kernel post-fertilization), sucrose accumulated to high

levels in the pedicel. The resulting decrease in the hexose-to-sucrose ratio in the pedicel would

theoretically favor differentiation as opposed to expansion (Winter and Huber, 2000; Koch,

2004).

Co-Expression of Genes Related to C Sink Development

Genome-wide expression data from developing maize ovaries were analyzed from parallel

microarray- and sequence-based profiles in Chapter 3. We identified co-expression clusters of

genes involved in C and N metabolism that showed high fold changes and/or significant

expression differences at a FDR of 20% during development. Genes annotated as having

functions in sucrose and starch metabolism tended to associate with a wider range of co-

expression clusters than those related to N metabolic processes (Figure 4-9). In general, C-

related metabolic genes also showed lower fold-changes over development than those associated

with N metabolism. Genes for N metabolism were highly represented in cluster 7, which









included those genes up-regulated during later stages of ovary development (Figure 4-9B). Since

lipid metabolism is, in part, regulated by C: N status (Koch, 1997; Foyer et al., 2003), we

evaluated co-expression profiles of genes involved in lipid metabolism. The developmental

trends observed from the expression data were indicative of a shift from lipid degradation to lipid

transfer, synthesis of lipid-based metabolites, and storage (Figure 4-10A). These data were

further supported by metabolite profiles for linoleic acid (18:02), linolenic acid (18:03), and

oxophytodienoic acid (OPDA) using GC/MS-based quantifications (Figure 4-10B).

Post-Pollination Lignin Biosynthesis in Pedicels

In addition to enhanced dry weight and sucrose accumulation, the rigidity of pedicels

increased rapidly after pollination. Microarray- and sequence-based expression profiles in

Chapter 3 identified co-expressed clusters of genes related to phenylpropanoid biosynthesis.

Expression profiles associated with cluster 7 tended to increase over development in ovary-plus-

pedicel samples and included a number of genes annotated as having functions related to

phenylpropanoid biosynthesis (Table 3-3). The developmental expression profiles for select

genes related to phenylpropanoid biosynthesis and their co-expression cluster assignments are

shown in Figure 4-11. In addition, quantification of cinnamic acid using GC/MS showed a linear

increase of this lignin precursor in the ovary-plus-pedicel fractions, while levels of salicylic acid

remained unchanged (Figure 4-12A). Lignin accumulation in the pedicel was visualized by

phloroglucinol-HCL staining of fresh, longitudinal sections of developing florets during pre- and

early post-pollination (Figure 4-12B).

Discussion

Our findings demonstrated that the shift from pre- to early post-pollination maize

reproductive growth coincided with a concomitant shift in resource allocation among individual

floral organs. We observed a temporal and spatial regulation of sink strength as determined by









accumulation of dry and fresh weights. Such regulation was consistent with expansion and/or

differentiation in specific floral organs during female floret development. The substantial

deposition of dry weight observed in silks before pollination is not surprising given the

reproductive advantage of concurrent anther and silk exsertion (translating to a short Anthesis-

Silking Interval) (Bolanos and Edmeades, 1996; Borras et al., 2007). Also, accumulation of

fresh weight in subtending floral parts during the pollination period was analogous to petal

expansion. Although a functional role for the subtending floral structures lemmaa, palea, and

glumes) has not been described, post-pollination remobilization of C assimilates from these sinks

is one possibility. The proportionate relationship maintained between ovary-plus-pedicel fresh

weight and silk length over the course of pre-pollination growth suggested spatial regulation of

photoassimilate import during development.

The elevated hexose-to-sucrose ratio and turgor-based expansion in silks and subtending

floral parts coincided with expression of a vacuolar invertase, IVR2. The IVR2 gene in maize is

up-regulated by C availability at the transcriptional level (Andersen et al., 2002; Gonzalez et al.,

2005) and has been implicated as a key enzyme for establishing and maintaining sink strength in

developing ovaries (Zinselmeier et al., 1995; Andersen et al., 2002; McLaughlin and Boyer

2004a). Previous findings showed that maximum rates of silk expansion occurred pre-dawn

(Westgate and Boyer, 1995). Our data were consistent with these findings and indicated that

hexoses were most abundant at a 6 AM sampling time. One possible explanation for a rapid

burst of expansion at the end of the dark period is remobilization of transitory starch in the silks.

Preliminary evidence for starch accumulation in silks included high levels of a leaf-specific

ADP-G Pyrophosphorylase mRNA and localization of iodine-stained starch to chloroplasts along

the length of the silk (A. Eveland, unpublished). Previous work has shown that starch reserves in









ovary tissue were also remobilized during times of carbohydrate starvation, presumably aiding

maintenance of normal growth and development (McLaughlin and Boyer, 2004b). Theoretically,

remobilization of this transitory starch could provide substrates for IVR2 activity in silks.

An increase in dry weight deposition by developing ovary-plus-pedicels coincided with the

time of pollination. The observed accumulation of sucrose in pedicels could by key to promoting

vascular differentiation. Also, evidence indicates that lignin biosynthesis is a strong sink for C

assimilates in the pedicel immediately after pollination. First, genome-wide expression data

from analyses in Chapter 3 identified co-expressed clusters of lignin biosynthetic genes that

showed high-fold change during post-pollination development. Second, linear increases of

cinnamic acid in the ovary-plus-pedicel suggested flux of C-assimilates to phenylpropanoid

biosynthesis rather than to salicylic acid conversions. Finally, phloroglucinol staining of 4-0-

linked hydroxycinnamyl aldehydes provided a visual appraisal of pedicel-specific lignin

accumulation over time. Lignin biosynthesis is also irreversible and energy-intensive, thus an

irretrievable, high-cost investment in plant growth and development. Therefore, regulation of C

flux to lignin biosynthesis is probably modulated on a number of levels including nutrient status,

hormone signals, and developmental cues. Genes involved in lignin biosynthesis, for example,

respond to sugar signals at the transcriptional level (Rogers et al., 2005). Such signals could be

central to maintaining a spatial and temporal balance between cell expansion and differentiation

during sink development. Further analyses of lignins and their composition using GC/MS will

provide quantitative evidence for the degree to which C is utilized in lignin biosynthesis during

pedicel development.

Recent work has shown that an altered C: N balance can modulate the extent of

phenylpropanoid biosynthesis (Fritz et al., 2006). Analysis of genome-wide transcript profiles









during ovary-plus-pedicel development (Chapter 3) also showed that genes involved in N

assimilation were co-expressed with those for lignin biosynthesis. These expression data are

consistent with metabolic trends observed in growing C-sinks (Koch, 1996, 1997; Cooke et al.,

2003; Foyer et al., 2003). Shifting the C: N balance in favor of C assimilates could result in up-

regulation of nitrogen assimilation via sugar-responsive expression of glutamate synthase and

glutamine synthetase (Koch, 1996, 1997; Fritz et al., 2006; Martin et al., 2006). Studies have

shown that key factors for kernel set in maize include N assimilation during the early post-

pollination period, particularly an up-regulation of glutamine synthetase isoforms (Seebauer et

al., 2004; Martin et al., 2006).

During the pre-pollination period of expansion by silks and subtending floral organs, C

deposition is comparatively less in the ovary and pedicel. Remobilization of C resources from

senescing floral organs to the developing ovary-plus-pedicel could promote C-based signals

during early post-pollination growth. Increasing C relative to N would favor up-regulation of

genes involved in nitrogen assimilation, lipid synthesis and storage, and phenylpropanoid

biosynthesis. Genes associated with such processes tended to show co-expression during

development while those involved in starch and sucrose metabolism showed a wider range of

expression profiles and smaller fold-changes.

Conclusions

Individual stages of pre- and early post-pollination maize female floret development were

characterized together with estimations of C-allocation among individual floral organs. Prior to

pollination, the major sinks appeared to be the silks and the floral structures subtending the

ovary. During early post-pollination development, dry weight accumulated in the pedicel and C

was allocated predominantly to the ovary-plus-pedicel. This shift in sink strength was consistent

with differential expression of a vacuolar and cell-wall invertase, respectively. In addition, there









was an apparent pre-to-post-pollination shift from accumulation of soluble sugars to lignin in the

pedicel. These data indicated that a shift in balance occurs between expansion and

differentiation during maize ovary development. Analyses of changes in co-expressed genes

associated with this shift suggest possible contributions by C- and N-based sensing mechanisms.

Materials and Methods

Plant Material and Sampling

W22 inbred maize plants grown in various field seasons and under greenhouse conditions.

Data are from pre-pollination, developmental analyses of material grown under field conditions

at the Tropical Research and Education Center (TREC), Homestead, Florida, winter planting.

Three ears were collected daily at 9 AM over the course of two weeks in March, 2005. Ear

development was staged based on anatomical characteristics which included ear length, silk

length, and floret fresh weight. Florets were dissected from the mid-to-basal equatorial region of

each ear, separated from silk, and both immediately frozen in liquid nitrogen for further analyses.

Ovary-plus-pedicel samples were hand-dissected and collected separately from ears 6 DBP to OT

(stages 5 to 7). A total of six ears that showed greatest uniformity in growth characteristics and

expression of the ZAG2 molecular marker were used in subsequent analyses. Post-pollination

analyses were repeated in material from both a summer field (2005) grown in Citra, FL and from

a winter greenhouse (2008) in Gainesville, FL. The latter were grown in 14" pots with daylight

extended to 12 hrs.

Quantification of mRNAs by Real-Time RT-PCR

Quantitative analyses of mRNAs by real-time RT-PCR used either an ABI 2600

Instrument or the ABI StepOne Plus (Applied Biosystems). Freshly dissected silk and floret

samples were immediately frozen in liquid nitrogen and homogenized in 1 mL TriZol

(Invitrogen) using a FastPrep lysis system (Q-Biogene). Total RNA was extracted as described









in Materials and Methods (Chapter 3). Gene-specific primers and Taqman probes were designed

for IVR1, IVR2, and INCW2 using PrimerExpress software (ABI): Ivrl FP: 5'-

CGGCAGCCTCCAAACTTTC-3', Ivrl RP: 5'-CCCGTATACTCTCTTAACCAGATCGT-3',

Ivrl probe: 5'-TCTGCCAAGACGAGGTCAGGGCA-3'; Ivr2 FP: 5'-

TGGCTACTACTTATCTTCCAGCACTAGT-3', Ivr2 RP: 5'-TGCATGATGCGGTGCTACA-

3', Ivr2 probe: 5'-CATGTACAACTAGAGGCTACACGTCTTCCCACTG-3'; Incw2 FP; 5'-

GGTCACTCTCAGGAACAGGGTAA-3', Incw2 RP: 5'-AGCCTGTGCCGTTTGTATCC-3',

Incw2 probe: 5'-CACCTCGACGTGATGTCCTGCCTTG -3'. For each Taqman reaction, 100

ng of total RNA was used as template and run in triplicate. Data were analyzed using a AA CT

method by comparing all values to a single sample and calculating RQ or relative percent. A

SYBR green method was used to quantify Zag2 transcripts using gene-specific primers: Zag2

FP: 5'-TTGGCTTCCATGACCTTGCT-3'. Zag2 RP: 5'-GCACAAGGAGAATCACACACAAA-

3'. For SYBR green analyses, 1 ug total RNA was converted to cDNA using a High-Capacity

cDNA RT reaction (ABI) and random primers. Parallel amplification of a VIC-labeled 18S

rRNA control (ABI) was used to normalize all mRNA levels

Soluble Sugar Extraction and Quantification

Concentrations of sucrose, glucose, and fructose were determined using either enzymatic

analysis or chemical separation with High Performance Liquid Chromatography (HPLC).

Soluble sugars were extracted from 200 ng whole silk tissue after homogenization in liquid

nitrogen and incubation in 1 mL extraction buffer (200 mM KOH, .08% Triton X). A

Sucrose/D-Glucose/D-Fructose UV method (R-Biopharm, Roche) was used to quantify absolute

amounts of sucrose, glucose, and fructose. For HPLC analyses, a Carbohydrate Analysis

Column (AMINEX Carbohydrate HPX-87C, BioRad) was used to separate soluble sugars. Pre-

and post-pollination floret, ovary, and pedicel samples were weighed and homogenized in 2 mL









tubes with 500 uL boiling 80% ethanol using a FastPrep lysis system (Q-Biogene). Prior to

homogenization, samples were spiked with a xylose control. Soluble sugars were separated by

centrifugation for 10 minutes and the ethanol extraction was repeated three times. The 1.5 mL

supernatant sample was lyophilized and resuspended in 900 uL HPLC grade ddH20. Filitered

samples were then analyzed by HPLC and peak areas were calculated on the basis of a standard

curve.

Assay for Soluble Invertase Activity

Whole silk tissue (200 mg) was homogenized in liquid nitrogen and extracted on ice in 1

mL extraction buffer (50 mM MOPs-NaOH [pH 7.5], 5mM MgC1, ImM EDTA, 0.05% w/v

Triton K-100, 2.5 mM DTT, 0.1 mM DMSF, 1% PVPP). Soluble protein samples were

separated by centrifugation and transferred to dialysis bags (MWCO 50 kD). Dialysis was

carried out overnight in three changes of 1XPBS buffer to remove low-molecular-weight

invertase inhibitors. Total protein concentration was determined by Bradford (BioRad) using a

BSA standard curve. Soluble acid invertase activity was assayed in a 1:3 dilution of total soluble

protein: 17mg/mL sucrose in 50mM sodium citrate buffer (pH 5) at 37 C for 5 minutes. Total

glucose was quantified in relation to a glucose standard (GO Assay Kit GAGO20-1KT, Sigma).

GC/MS Quantification of Metabolites

Cinnamic acid and salicylic acid were quantified in pre-pollination whole floret and ovary-

plus-pedicel samples. Samples were frozen immediately after dissection in liquid nitrogen and

extracted in 2 mL tubes using a FastPrep lysis system (Q-Biogene). Metabolite extraction,

methylation, and isobutene-chemical ionization gas chromatography/ mass spectroscopy

(GC/MS) analysis were carried out according to Schmelz et al. (2003) at the USDA-ARS Center

for Medical and Veterinary Entomology, Chemistry Unit, Gainesville, FL.









Phloroglucinol-HCL Staining

Fresh longitudinal sections of florets at 2 DBP, OT, 2 DAP, and 4 DAP were stained with

phloroglucinol/ethanol (96%)/HCL (37%) for five minutes, washed with ddH20 and

photographed under a dissecting scope (Leica MZ 12.5).


























silk



ovary
subtending floral
] pedicel


stage 1 2 3 4 5 6 7
DBP 12 10 8 6 4 2 0


A
C,
0)
C,
E
W,
(D
(D
I)


A
C
0
(U
C
c
Q-


silk length (cm)


3 Zag2

Z 2
E
a

0

stage 1 2 3 4 5 6 7
DBP 12 10 8 6 4 2 0



Figure 4-1. Definition of developmental stages for pre-pollinated maize female florets by
physical growth parameters, anatomical features, and expression of a ZAG2 molecular
marker. A) Florets were sampled and hand-dissected from the mid-to-base region
(boxed in red) of a given ear and B) assigned to growth stages 1 to 7 (also
corresponding in this study to days before pollination [DBP] shown immediately
below the stages indicated). C) Stages were determined by fresh weight of ovary-
plus-pedicel in relation to silk length and D) Relative abundance ofZag2 mRNAs as a
molecular marker for development.


P













0.045 -




0.023 -


stage


ovary + pedicel

3 4 5 6 7


DBP 8 6 4 2 0

12
silk

8-
subtending floral

ovary + pedicel


Os-t
stage


3 4 5 6 7
3 4 5 6 7


DBP 8 6 4 2 0


Figure 4-2. Partitioning of carbon and water to dry weight and fresh weight accumulation among
individual organs of developing maize female florets. A) Dry and fresh weights were
quantified on a per-floret basis. Note the prominence of dry weight gain (and
approximate sink strength) of the rapidly elongating silk. B) Silks maintained a high
fresh-to-dry weight ratio over the course of pre-pollination elongation. Also, fresh-
to-dry weight increased linearly during rapid expansion of subtending floral parts just
prior to pollination.


dry weight


u --

















0
-o

CU 0
z 5- Ivr
E

.E2.5



0
stage 5 6 7
DBP 4 2 0
I ovary + pedicel whole floret


01
stage
DBP


5
4
- sucrose


6
2
LWWW


7
0
hexoses


Figure 4-3. Spatial and temporal expression of soluble acid invertases and changes in sugar
composition in maize florets just prior to pollination. A) Relative mRNA levels for
two vacuolar invertase isoforms, Ivr2 and Ivrl in ovary-plus-pedicel (white bars) and
whole florets (grey bars). B) Sucrose (dk bars) and hexoses (It bars) were quantified
in ovary-plus-pedicel and whole florets prior to pollination.











A 09 C
1.2
.06-
E 0.8-
N 0
0.4-. .03-
0.4-
0 0
E
0- 0-
100- B --Fall 2003 12- D
-*-Spring 2004
-m Fall 2004 U -)

0 8-
0.

0 0 04 101
stage 3 4 5 6 7 4 DBP 2 DBP OT 2 DAP
DBP 8 6 4 2 0 +2

Figure 4-4. Activity of the vacuolar invertase, IVR2, is associated with turgor-based expansion in
rapidly elongating silks. A) Epidermal cell size at the base of rapidly elongating silks
increases linearly over a pre-pollination time course. B) Relative levels of vr2
mRNAs during three different field seasons. C) soluble acid invertase activity and D)
Hexose-to-sucrose ratio in whole silks pre- and early post-pollination.











hexoses
3- -




0) sucrose



0
4pm 10pm lam 4am 6am 10am



Figure 4-5. Diurnal changes in sucrose and hexose levels in rapidly expanding silks on the day
of their emergence from husks.
















50-





S100- Incw2

E


50




-6 -4 -2 0 2 4
days before pollination days after pollination

Figure 4-6. Relative abundance of mRNAs for vacuolar and cell-wall inverase isoforms in maize
female florets during the pre- to- post-pollination period. While Ivr2 mRNAs
accumulated in the subtending floral parts, Incw2 expression was specific to the
ovary-plus-pedicel fraction and increased after pollination.













P w: pedicel


.004
Dry weight
ovary + pedicel
.003-


.002- _
ovary
001
.001- pedicel

E
M 0-
,, Fresh weight
.03


.02 -ovary


.01

pedicel
0-
OT 2DAP 4DAP


Figure 4-7. Carbon deposition and relative water content in maize ovaries and pedicels during
post-pollination development. Ovary tissue (or developing kernel post-fertilization)
was separated immediately above the pedicel as shown in the diagram above. Dry
and fresh weights were quantified for ovary-plus-pedicel, ovary, and pedicel samples
on a per-floret basis.











1800ovary/young kernel

hexoses
1200-


sucrose
S600-

LL
0)
E 0
0 2500- pedicel

hexoses


1500-


sucrose
500

0
OT 2DAP 4DAP


Figure 4-8. Sugar composition of maize ovaries and pedicels during early post-pollination
development. Hexoses and sucrose were quantified in separated ovary tissue (or
developing kernel if post-fertilization) and pedicels during early post-pollination
development.












A C metabolism-related
'I-------------


S10-


*-

0-
SB-
E 7-


4 DBP 2 DBP OT 2DAP










N metabolism-related
110



.............,.,,


9.


B"


7-


6-___________________________


..... Cluster 1
UDP-glucose pyrophosphorylase [A. fruicosa]
ADP-glucose pyrophosphorylase small subunit [Z. mays]
Susl [Z. mays]
UDP-glucuronic acid decarboxylase [0. safva]
Putative UDP-glucose dehydrogenase [0. seuva]
Putative amylase [0. satave]
Cluster 7
Putative alcohol dehydrogenase [O. satval
Enolase I [Z. mays]
Shrunken-I [Z. mays]
UDP-glucose pyrophosphorylase [B. oldhami/]
Similarto alpha-glucosidase [Z. maysI
***** Cluster 6
ADP glucose pyrophosphorylase large subunit [0. sarva]
Putative sugar transporter protein [0. sativa]
- Cluster 5
Putative beta-amylase [0. saeva]
Putative sus3 [Z. mays]
Sucrose transporter [Z. maya]
Putative sucrose phosphate synthase [0. sativa]
Alkalinelneutral invertase [0. sativa]
Starch branching enzyme I [Z. mays]
- Cluster 2
Putative glucosyltransferase [0. saeUva]
Trehalose-6-phosphate synthase [0. safiva]

-*-** Cluster 1
Glutamine synthetase [Z. mays]
Glycine hydroxymethyltransferase [A. thalianra]
Alanine transaminase [P. miiaceum]
Gamma-amninobutyrate transaminase precursor [0. saeval
Cysteine synthase [Z. maya]
Arginine bansamidinase [0. safeal
Cluster7
Putative nitrate transporter NRTt-5 [0. sat6val
Putative NAD synthetase [0. aefve]
Glutamine synthetase [Z. maya]

Glutamine synthetase [Z mays]
Glutamate-5-semialdehydelehydrogenase
Glutamine-fructose-6-phosphate transaminase 2 [A. thaliana]
Putative gamma-lvase r0. saeval


4 DBP 2 DBP OT 2DAP Homocystine S-methyltransferase-3 [Z. maya]
S-adenosylmethionine decarboxylase [Z. maysI
Methionine synthase [Z. mays]
Cluster 2
Nitrate reductase [Z. mays]
Asparaginase [H. vulgare]
Glutamate decarboxylase [0. sativa]
Glutamate-ammonia ligase [Z. mays]


Figure 4-9. Co-expression profiles for genes related to C and N metabolism during ovary and
pedicel development from genome-wide microarray analyses. A) Genes annotated as
having functions in sucrose and starch metabolism tended to associate with a wide
range of co-expression clusters. B) Genes involved in N metabolism and assimilation
tended to show larger fold-changes during development and associate predominantly
with cluster 7 (expression increases after pollination).


*- _

-. a - -
ar













CT


2 1
_o







a-
C

E

























52
tD


20-


30

S1803

20-


10

0


2






3-
D-



4--


12 10 8 6 4 2 0

days before pollination


Figure 4-10. Co-expression of genes related to lipid metabolism and abundance of lipid-based
metabolites in developing maize ovaries. A) Select genes annotated as having
functions related to lipid metabolism in co-expression clusters 1, 2, and 7 from
microarray analyses (Chapter 3). Relative expression profiles in the developing
ovary-plus-pedicel are depicted and select genes are listed. B) Quantitative
metabolite profiles are shown for linoleic acid (18:02), linolenic acid (18:03), and
oxophytodienoic acid (OPDA) in whole female florets (12 DBP to OT) and ovary-
plus-pedicel fractions (4DBP to OT) during pre-pollination development.


***** Cluster 1
......... Phaspholipase D alpha 2 [0. safvaI
s.. m.enFa Fatty acid elongase [Z maya]
Fatty acid desaturase FAD7 [Z. mays]
Omega-3 fatty acid desaturase [Z. mays]
Putative enoyl-ACP reductase [0. saUva]
Acyl-CoA-binding protein [0. sateva]
S- Putative 12-oxophytodienoate reductase [0. saiva]
-- -- Cluster 7
Putative lipid transfer protein [O. saffva]
Putative lipoxygenase [O. aatival
Allene oxide cyclase [Z maysl
Phospholipid transfer protein 9C2 precursor [Z. mays]
Long chain acyl-CoA synthetase 7 [A. thaliana]
Lipoxygenase [Z mays]
Storage!lipid transfer protein (LPT) [O. sativa]
4 DBP 2 DBP OT 2DAP Cluster2
AE9 stearoyl-ACP desaturase [0. sabrva]
Lipoxygenase [0. saiva]
Cis-12-oxophytodienoic acid reductase [0. seava]
18t02 Putative lipase [0. aaevea
Lipase-like protein [0. saUva]

Putative phospholipase A2 [0. sa#va]
Fatty acid desaturase [Z. mays]
SGDSL-motif lipasefhydrolase protein [0. sativa]


60


40











12. *m** Clusteri
4-coumarate--CoA ligase 4CL2 [L. perenne]
**------- N-hydroxycinnamoylibenzoyltransferase-like protein [A. thaiana]
S10. ..= 'I Cinnamyl alcohol dehydrogenase [L pernne]
10- Putative peroxidase precursor [0. save]
Cluster'
S""'"Gaiacol peroxidase [G. hirsutum]
C Peroxidase [Z. mays]
Polyphenol oxidase [T. aestivum]
6 Inducible PAL [T. aesfivun]
Beta-glucosidase aggregating factor precursor [Z. maysI
S" Caffecyl CoA 3-0-methyltransferase [Z. mays]
E 4. Cinnamoyl CoA reductase [Z. mays]
4 iCaffeoyl CoA 3-0-methyltransferase [Z. mays]
4-coumarate-CoA ligase 4CL2 [L. perenne]
2 Cluster 2
4 DBP 2 DBP OT 2DAP Cinnamyl alcohol dehydrogenase [L perenne]
Peroxidase-like protein [0. satival
Putative peroxidase [0. sa'va]
Cinnamoyl-CoA reductase [Z. maysI
Peroxidase [Z. maya]
Putative 4-coumarate-CoA ligase 1 [O. satival
Putative cinnamoyl-CoA reductase [0. satia]


Figure 4-11. Expression profiles for genes involved in phenylpropanoid biosynthesis in the
developing maize ovary-plus-pedicel.












150-

100-

50-

0-
300-

200-

100-

0-
staqe


cinnamic acid ovary + pedicel
whole florets

".3





salicylic acid





I I I I I I I
1 2 3 4 5 6 7


DBP 12 10 8 6 4 2 0


2 DBP OT 2 DAP 4 DAP


Figure 4-12. Accumulation of lignin precursors and pedicel-localized staining of lignin in the
developing maize ovary. A) Cinnamic acid and salicylic acid were quantified in
whole female florets and ovary-plus-pedicel during pre-polliantion development. B)
A phloroglucinol stain for cinnamylaldehyde residues was used as a qualitative test
for lignin accumulation and was localized to vascular tissue in the pedicel during the
early post-pollination period.









CHAPTER 5
SUMMARY

The overall hypothesis tested here was that genes related to specific functional processes or

metabolic pathways would be co-expressed in association with carbohydrate allocation and use

during silk exsertion and pollination in maize. Further, the approaches used allowed such genes

to be identified, clustered with other sets of co-regulated genes, and their expression quantified.

The broader significance of this work was two-fold. First, maize now represents both the largest

crop yield on earth in total bushels, as well as an emerging model for grain yield and biomass

production. Second, the reproductive phases examined here are pivotal for successful maize

pollination and establishment of kernel number, a primary determinant of yield.

With the recent release of the maize draft genome sequence, focus is now directed to

building a resource infrastructure specific to maize. Results from the work presented here will

contribute to a foundation for functional, maize-based research. Currently, major limitations to

determining specific functional roles of maize genes include insufficient genome annotation,

biased array platforms that limit gene discovery, and large-scale confounding effects from

expression of paralogous genes.

To achieve quantitative resolution of gene family members and other closely-related

paralogs, we developed a sequence-based profiling method that uses the gene-specificity of the

3'-UTR (Chapter 2). We tested this method parallel with microarray analyses in a genome-wide

transcript profiling approach to identify co-expressed genes with related functions during maize

ovary development (Chapter 3). Resulting analyses revealed clusters of genes involved in import

and use of C- and N-resources that were co-expressed during development. In Chapter 4, we

determined spatial and temporal changes in C-allocation and turgor-based expansion, as

evidenced by dry and fresh weights, among individual floral organs. In addition, the









contribution of specific invertase isoforms was tested in relation to these changes. Key

conclusions based on the combined analyses are as follows:

* A 3'-UTR profiling strategy was developed that resolved quantitative expression of gene
family members and other closely-related transcripts in maize ovaries and provided a
useful compliment to microarray analyses.

* A combined microarray- and sequence-based approach was found most effective for
genome-wide transcript profiling in developing maize ovaries. Advantages and limitations
of both methods were addressed.

* Genes related to specific functional processes were co-expressed during maize ovary
development. Key differences in functional classes of genes were evident at specific
stages of development.

* Prior to pollination, increases in dry and fresh weights in silks and subtending floral organs
coincided with expression of a sugar-responsive vacuolar invertase, IVR2. Prominent dry
weight gain indicated that silks were the major sink in pre-pollination maize florets.

* Immediately after pollination, allocation of C-resources (as evident by dry weight) shifted
from silks and subtending floral organs to the developing ovary and pedicel.

* Co-expression analyses indicated that genes related to C-sink development were up-
regulated in concert with the increased accumulation of dry weight and sucrose in the
pedicel of the ovary.

* Genes related to N metabolism and lignin biosyntheses were co-expressed during
development of maize ovaries and pedicels and tended to increase post-pollination.

* Evidence indicated that lignin biosynthesis in the pedicels of maize ovaries was a
prominent sink for C resources during early post-pollination development.

* Global expression profiles and invertase activity data indicated that regulation of sink
strength in maize female floral organs was more strongly related to developmental changes
in maternal tissues than to direct effects of pollination or fertilization.

Central contributions to functional genomics in maize can be made by organ-specific

expression profiles, metabolite analyses, and biochemical testing. The maize female

inflorescence is characterized by unique anatomical features and therefore functional

comparisons to Arabidopsis, or even to rice, can be insufficient. For many years, corn breeders

have focused on genetic traits specific to the female inflorescence to improve kernel set and









grain yield. Although genome-wide expression analyses and functional annotation are becoming

more accessible for maize, the pre- and early post-pollination phases of maize reproductive

development have received comparatively little attention at the molecular/metabolic level.

Results from the present work and from other studies suggested that maternal-based regulation of

reproductive development in cereals included C-allocation and use (Westgate and Boyer, 1986;

Zinselmeier et al., 2003; McLaughlin and Boyer, 2004a; Sreenivasulu et al., 2004). Co-

expressed genes related to these processes have been identified here, providing evidence for

testable roles in kernel set. Future work can also use these developmental analyses as a baseline

for environmental or genetic perturbation of the system.









APPENDIX
PERTURBATION BY DROUGHT AND DISRUPTED VP1-BASED ABA SENSING

Hypothesis

The hypothesis tested here was that C allocation and use in developing female

inflorescences of maize would be disrupted by drought stress and could involve interactions of

ABA and sugar signals. We analyzed developing female florets (as described in Chapter 4) from

maize plants that were subjected to a drought stress treatment and from an ABA insensitive

mutant to test affects of environmental and genetic perturbation, respectively.

Background and Significance

Drought stress during flowering can negatively impact yield in maize either by disruption

of meiotic, pollination, and/or fertilization processes or by inhibiting grain fill (Westgate and

Boyer, 1985; Andersen et al., 2002; Boyer and Westgate, 2004; Barnabas et al., 2007). Westgate

and Boyer (1985) showed that drought stress during pre-pollination growth significantly reduced

solute concentration in the maize female inflorescence and did so to a greater extent than

observed in leaves, stems, or roots. Resulting decreases in turgor were associated with inhibition

of ear growth and failure of silk exsertion.

Recent work by Borras et al. (2007) defined a threshold level of ear growth, as measured

by dry weight accumulation, which was indispensable for timely silk emergence. When C

allocation to the developing ear fell below this threshold, a larger Anthesis Silking Interval (ASI)

resulted. Since tassel development showed little response to reduction in photosynthate

availability, timing of pollen-shed was used as a marker to compare ASI in stressed plants

(Borras et al., 2007). A short ASI is essential for optimal silk receptivity, pollination, and

subsequent grain yield. Efforts to decrease the ASI through breeding and QTL analyses have









been successful, however the mechanisms underlying regulation remain elusive (Bruce et al.,

2002).

A key regulator of drought-induced gene expression is the phytohormone, abscisic acid

(ABA). Recent work has demonstrated that local biosynthesis of ABA in leaves is dependent on

a hydraulic signal from the roots (Christmann et al., 2007). Accumulation of ABA in response to

soil drying causes stomatal closure, presumed to be a mechanism of acclimation that reduces

evapotranspiration. Photosynthesis is also repressed under these conditions, and thus disruption

of the source-to-sink balance requires whole-plant adjustment. Evidence for ABA production in

maternal reproductive tissues has also been demonstrated (Myers et al., 1992; Setter et al., 2001),

and drought stress can result in transient increases of this hormone during early development of

female florets (Andersen et al., 2002). ABA-based signaling networks overlap with those of

sugar-based gene regulation on a number of levels (Finkelstein and Gibson, 2002; Li et al.,

2006). Many of these connections have been demonstrated with redundant mutant loci in

Arabidopsis that can result in ABA- and sugar-responsive phenotypes (Arenas-Huertero et al.,

2000; Huijser et al., 2000; Laby et al., 2000; Brocard et al., 2002; Arroyo et al., 2003; Brocard-

Gifford et al., 2004). However, very few studies have extended such analyses to maize (Niu et

al., 2002).

Specific invertase isoforms respond to ABA in both maize (Kim et al., 2000b; Trouverie et

al., 2004) and Arabidopsis (Huang, 2006). In maize, a vacuolar invertase, ZmlVR2, responds

differentially to drought and associated rises in ABA levels in leaves and developing ears. While

ZmlVR2 is up-regulated in leaves during drought stress, the same isoform is repressed in florets

(Andersen et al., 2002; Trouverie et al., 2003). This spatial-specific mode of regulation suggests

an association between the phytohormone ABA and whole-plant source/sink balance. In









addition, different levels of ABA can have contrasting effects on organ growth. Low levels of

ABA tend to promote growth while higher levels are growth-inhibiting (Cheng et al., 2002; Peng

et al., 2003). Work by Suzuki et al. (2003) showed that AtvaclNV1, the putative functional

ortholog of ZmlVR2, was ABA up-regulated, but repressed by VP 1 (a component of one ABA

sensing system). Although VP1 has classically been implicated as seed-specific in maize

(McCarty et al., 1991; Hoecker et al., 1995), evidence from its Arabidopsis counterpart, ABI3,

indicates putative functions in vegetative and maternal reproductive tissues (Parcy et al., 1994;

Rohde et al., 1999; Rohde et al., 2002). A recent study by Cao et al. (2006) showed that VP1

was induced by desiccation stress in maize and localized to the phloem in leaves and pedicels of

developing female florets. In addition, studies in Arabidopsis have shown that ABI3 is involved

in both ABA and auxin-based signaling pathways (Suzuki et al., 2001; Brady et al., 2003).

Results

In this study, we compared mRNA levels for two vacuolar invertases, Ivrl and Ivr2, in

developing maize ovaries before pollination. Staging of pre-pollination development was as

described in Chapter 4. We also quantified endogenous levels of ABA and Vpl mRNAs in the

same material. Transcript profiles for both vacuolar invertases were quantified by Taqman Q-

PCR analyses and showed similar trends during pre-pollination ovary development (Figure A-

1A). A significant peak was evident in Ivrl mRNA levels early in ovary development (stage 2 or

10 days prior to pollination according to our scale). Transient accumulation of ABA (quantified

by GC/MS analysis) coincided with elevated Ivrl mRNA levels at stage 2 (10 DBP) (Figure A-

1B). Andersen et al. (2002) showed comparable levels of endogenous ABA in young maize

ovaries. They also showed that this transient peak in ABA was enhanced 3-fold under drought

stress. In work here, we show a concurrent, transient rise in levels of Vpl mRNAs in young

ovaries (stage 4 [6 DBP]). This peak in Vpl mRNAs was preceded by ABA and Ivrl mRNAs









(Figure A-1C). These findings showed that endogenous levels of Vp] mRNAs and ABA were

present in maternal tissues under normal, well-watered growth conditions.

One goal of this work was to test the extent of drought stress effects on the Anthesis

Silking Interval (ASI) in maize plants lacking a VP1-based ABA-sensing system. To do this, we

compared W22 wild-type and vpl mutant maize in well-watered and drought-stressed conditions.

ASI was measured in days from first pollen shed to first silk emergence (Table A-i). The vpl

mutants consistently showed a reduced ASI compared to the wild-type plants under drought-

stress. Although the ASI was significantly increased in drought-stressed vpl mutants relative to

well-watered controls, all vpl ears analyzed (10/10) had exserted silks during drought treatments.

In contrast, only 40% (4/10) of the wild-type ears exserted silks. Rewatering after 7 days of

withholding water resulted in silk exsertion by 100% of the wild-type ears. Wild-type and vpl

mutants showed no significant difference in ASI or in ear length when well-watered.

Immature ears were sampled from wild-type and vpl mutant maize plants grown under

well-watered and drought conditions. All ears were sampled 2 days after first pollen shed.

Whole florets were collected from comparable sections at the base, middle, and tip of each ear,

frozen in liquid nitrogen, and pooled for subsequent analyses. Levels of ABA and IAA were

quantified in the wild-type and vpl ears by GC/MS and used to compare the extent of treatment

effects (Figure A-2A). Although total ABA levels were significantly lower in vpl mutants,

drought stress resulted in proportional increases to ABA levels for ears of both mutant and wild-

type plants. We also observed that IAA levels were elevated in drought-stressed, wild-type ears

compared to ears from well-watered controls. Interestingly, IAA levels did not change in

response to drought in vpl mutant ears. These data showed that drought stress altered the

ABA/IAA ratio in the absence of a VPl-based ABA sensing system.









Sucrose and hexoses were also quantified in the same material using HPLC. Results

showed a decreased hexose/sucrose ratio in wild-type and vpl mutant ears in response to

drought. This decrease in hexose/sucrose was dampened in the vpl mutants (Figure A-2B) by

enhanced accumulation of sucrose. Together, data from this work and elsewhere indicates a

possible association between sugar signals and ABA/auxin balance during drought-induced

inhibition of maize ear growth. Preliminary evidence presented here provides a basis for more

in-depth analyses of the hypothesized link between sugars and the ABA/auxin balance and its

role in drought-stressed maize ears.









Table Al-1. ASI for wild-type and vpl mutant maize in response to drought.
Treatment ASI (day) Ear length (cm)
Well-watered
wild-type 1.6 0.95 13.4 + 1.2
vpl mutant 1.1 0.53 13.6 0.8
Drought stressed
wild-type 5.8 + 1.2 7.2 + 1.4
vpl mutant 2.3 1.2 8.4 2.1
Only measured for ears with exserted silks (4/10).













3.5 wMY


o 2.5

1.5

m
c .5


Z





1/


stage 1 2 3 4 5 6 7
DBP 12 10 8 6 4 2 0
B
8o0- ---whole Ilorels
oveary-p -sp dicel



S400


a----I


stage 1 2 3 4 5 0 7
DBP 12 10 8 6 4 2 0

C
100-







E
0

stage 2 3 4 5 6
DBP 10 8 6 4 2


Figure Al-1. Quantification of soluble invertase mRNAs, endogenous ABA levels, and Vpl
transcripts during pre-pollination maize ovary development. A) Expression profiles
for Ivrl and Ivr2 quantified by Taqman Q-PCR. B) GC/MS analyses of endogenous
ABA levels. C) Maternal-based Vpl mRNAs quantified by Taqman Q-PCR.















3000


2000


1000


S0

C 30 IAA


20


10




B

3.5
() hexose/sucrose
U)
0
0 2.5
U)
(D-
U) 1.5
0
x

.5


0
wild-type vp


S drought-stressed well-watered


Figure A1-2. Sugar and hormone levels in drought-stressed, immature ears of wild-type and vpl
mutant maize. A) ABA and IAA levels quantified by GC/MS. B) Hexose/ sucrose
ratios determined by HPLC analysis.









LIST OF REFERENCES


Ambrose BA, Lerner DR, Ciceri P, Padilla CM, Yanofsky MF, Schmidt RJ (2000)
Molecular and genetic analyses of the silky 1 gene reveal conservation in floral organ
specification between eudicots and monocots. Molecular Cell 5: 569-579

Andersen MN, Asch F, Wu Y, Jensen CR, Naested H, Mogensen VO, Koch KE (2002)
Soluble invertase expression is an early target of drought stress during the critical,
abortion-sensitive phase of young ovary development in maize. Plant Physiology 130:
591-604

Anderson SR, Lauer MJ, Schoper JB, Shibles RM (2003) Pollination timing effects on kernel
set and silk receptivity in four maize hybrids. Crop Science 44: 464-473

Aoki N, Hirose T, Takahashi S, Ono K, Ishimaru K, Ohsugi R (1999) Molecular cloning and
expression analysis of a gene for a sucrose transporter maize (Zea mays L.). Plant and
Cell Physiology 40: 1072-1078

Appenzeller L, Doblin M, Barreiro R, Wang H, Niu X, Kollipara K, Carrigan L, Tomes D,
Chapman M, Dhugga KS (2004) Cellulose sunthesis in maize: isolation and expression
analysis of the cellulose synthase (CesA) gene family. Cellulose 11: 287-299

Arenas-Huertero F, Arroyo A, Zhou L, Sheen J, Leon P (2000) Analysis of Arabidopsis
glucose insensitive mutants, gin5 and gin6, reveals a central role of the plant hormone
ABA in the regulation of plant vegetative development by sugar. Genes & Development
14: 2085-2096

Arroyo A, Bossi F, Finkelstein RR, Leon P (2003) Three genes that affect sugar sensing
(Abscisic Acid Insensitive 4, Abscisic Acid Insensitive 5, and Constitutive Triple
Response 1) are differentially regulated by glucose in arabidopsis. Plant Physiology 133:
231-242

Avonce N, Leyman B, Mascorro-Gallardo JO, van Dijck P, Thevelein JM, Iturriaga G
(2004) The arabidopsis trehalose-6-P synthase AtTPS 1 gene is a regulator of glucose,
abscisic acid, and stress signaling. Plant Physiol 136: 3649-3659

Balasubramanian R, Karve A, Kandasamy M, Meagher RB, Moore BD (2007) A role for F-
Actin in hexokinase-mediated glucose signaling. Plant Physiology 145: 1423-1434

Bao JY, Lee S, Chen C, Zhang XQ, Zhang Y, Liu SQ, Clark T, Wang J, Cao ML, Yang
HM, Wang SM, Yu J (2005) Serial analysis of gene expression study of a hybrid rice
strain (LYP9) and its parental cultivars. Plant Physiol 138:1216-1231

Barbazuk WB, Emrich SJ, Chen HD, Li L, Schnable PS (2007) SNP discovery via 454
transcriptome sequencing. Plant J 51: 910-918









Barker L, Kuhn C, Weise A, Schulz A, Gebhardt C, Hirner B, Hellmann H, Schulze W,
Ward JM, Frommer WB (2000) SUT2, a putative sucrose sensor in sieve elements.
Plant Cell 12: 1153-1164

Barnabas B, Jager K, Feher A (2008) The effect of drought and heat stress on reproductive
processes in cereals. Plant Cell and Environment 31: 11-38

Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, Wei G, Chepelev I, Zhao K
(2007) High-resolution profiling of histone methylations in the human genome. Cell 129:
823-837

Baxter CJ, Sabar M, Quick WP, Sweetlove LJ (2005) Comparison of changes in fruit gene
expression in tomato introgression lines provides evidence of genome-wide
transcriptional changes and reveals links to mapped QTLs and described traits. J Exp Bot
56: 1699-1709

Bhattramakki D, Dolan M, Hanafey M, Wineland R, Vaske D, Register JC III, Tingey SV,
Rafalski A (2002) Insertion-deletion polymorphisms in 3' regions of maize genes occur
frequently and can be used as highly informative genetic markers. Plant Mol Biol 48:
539-547

Bate NJ, Niu XP, Wang YW, Reimann KS, Helentjaris TG (2004) An invertase inhibitor
from maize localizes to the embryo surrounding region during early kernel development.
Plant Physiology 134: 246-254

Blanc G, Wolfe KH (2004) Functional divergence of duplicated genes formed by polyploidy
during Arabidopsis evolution. Plant Cell 16: 1679-1691

Biasing OE, Gibon Y, Gunther M, Hohne M, Morcuende R, Osuna D, Thimm O, Usadel B,
Scheible WR, Stitt M (2005) Sugars and circadian regulation make major contributions
to the global regulation of diurnal gene expression in Arabidopsis. Plant Cell 17: 3257-
3281

Blee KA, Anderson AJ (2002) Transcripts for genes encoding soluble acid invertase and
sucrose synthase accumulate in root tip and cortical cells containing mycorrhizal
arbuscules. Plant Molecular Biology 50: 197-211

Birchler JA, Augar DL, Riddle NC (2003) In search of the molecular basis of heterosis. Plant
Cell 15: 2236-2239

Bocock PN, Morse AM, Dervinis C, Davis JM (2008) Evolution and diversity of invertase
genes in Populus trichocarpa. Planta 227: 565-576

Bolanos J, Edmeades GO (1996) The importance of the anthesis-silking interval in breeding for
drought tolerance in tropical maize. Field Crops Research 48: 65-80









Borevitz JO, Liang D, Plouffe D, Chang H-S, Zhu T, Weigel D, Berry CC, Winzeler E,
Chory J (2003) Large-scale identification of single-feature polymorphisms in complex
genomes. Genome Res 13: 513-523

Borevitz JO, Chory J (2004) Genomics tools for QTL analysis and gene discovery. Curr Opin
Plant Biol 7: 132-136

Borisjuk L, Rolletschek H, Radchuk R, Weschke W, Wobus U, Weber H (2004) Seed
development and differentiation: A role for metabolic regulation. Plant Biology 6: 375-
386

Borras L, Westgate ME, Otegui ME (2003) Control of kernel weight and kernel water
relations by post-flowering source-sink ratio in maize. Annals of Botany 91: 857-867

Borras L, Westgate ME, Astini JP, Echarte L (2007) Coupling time to silking with plant
growth rate in maize. Field Crops Research 102: 73-85

Boyer JS (1982) Plant productivity and environment. Science 218: 443-448

Boyer JS, Westgate ME (2004) Grain yields with limited water. Journal of Experimental
Botany 55: 2385-2394

Brady SM, Sarkar SF, Bonetta D, McCourt P (2003) The ABSCISIC ACID INSENSITIVE
(ABI3) gene is modulated by famesylation and is involved in auxin signaling and lateral
root development in Arabidopsis. Plant J 34: 67-75

Bray EA, Shih T-Y, Moses MS, Cohen A, Imai R, Plant AL (1999) Water-deficit induction of
a tomato H1 histone requires abscisic acid. Plant Growth Reg 29: 35-46

Brocard IM, Lynch TJ, Finkelstein RR (2002) Regulation and role of the Arabidopsis
Abscisic Acid-Insensitive 5 gene in abscisic acid, sugar, and stress response. Plant
Physiology 129: 1533-1543

Brocard-Gifford I, Lynch TJ, Garcia ME, Malhotra B, Finkelstein RR (2004) The
Arabidopsis thaliana ABSCISIC ACID-INSENSITIVE8 locus encodes a novel protein
mediating abscisic acid and sugar responses essential for growth. Plant Cell 16: 406-421

Bruce WB, Edmeades GO, Barker TC (2002) Molecular and physiological approaches to
maize improvement for drought tolerance. Journal of Experimental Botany 53: 13-25

Bustan A, Goldschmidt EE (1998) Estimating the cost of flowering in a grapefruit tree. Plant
Cell and Environment 21: 217-224

Buttner M (2007) The monosaccharide transporter (-like) gene family in Arabidopsis. Febs
Letters 581: 2318-2324









Cakir B, Agasse A, Gaillard C, Saumonneau A, Delrot S, Atanassova R (2003) A grape ASR
protein involved in sugar and abscisic acid signaling. Plant Cell 15: 2165-2180

Campos H, Cooper A, Habben JE, Edmeades GO, Schussler JR (2004) Improving drought
tolerance in maize: a view from industry. Field Crops Research 90: 19-34

Cao X, Costa LM, Biderre-Petit C, Kbhaya B, Dey N, Perez P, McCarty DR, Gutierrez-
Marcos JF, Becraft PW (2007) Abscisic acid and stress signals induce Viviparous-1
(Vpl) expression in seed and vegetative tissues of maize. Plant Physiol 143: 720-731

Carlson SJ, Chourey PS (1999) A re-evaluation of the relative roles of two invertases, INCW2
and IVR1, in developing maize kernels and other tissues. Plant Physiology 121: 1025-
1035

Carpaneto A, Geiger D, Bamberg E, Sauer N, Fromm J, Hedrich R (2005) Phloem-
localized, proton-coupled sucrose carrier ZmSUT1 mediates sucrose efflux under the
control of the sucrose gradient and the proton motive force. Journal of Biological
Chemistry 280: 21437-21443

Cazetta JO, Seebauer JR, Below FE (1999) Sucrose and nitrogen supplies regulate growth of
maize kernels. Annals of Botany 84: 747-754

Chen J, Sun M, Lee S, Zhou G, Rowley JD, Wang SM (2002) Identifying novel transcripts
and novel genes in the human genome by using novel SAGE tags. Proc Natl Acad Sci
USA 99: 12257-12262

Cheng WH, Endo A, Zhou L, Penney J, Chen HC, Arroyo A, Leon P, Nambara E, Asami
T, Seo M, Koshiba T, Sheen J (2002) A unique short-chain dehydrogenase/reductase in
Arabidopsis glucose signaling and abscisic acid biosynthesis and functions. Plant Cell 14:
2723-2743

Cheng WH, Taliercio EW, Chourey PS (1996) The Miniature 1 seed locus of maize encodes a
cell wall invertase required for normal development of endosperm and maternal cells in
the pedicel. Plant Cell 8: 971-983

Cheung VG, Spielman RS, Ewens KG, Weber TM, Morley M, Burdick JT (2005) Mapping
determinants of human gene expression by regional and genome-wide association. Nature
437: 1365-1369

Chiou TJ, Bush DR (1998) Sucrose is a signal molecule in assimilate partitioning. Proceedings
of the National Academy of Sciences of the United States of America 95: 4784-478

Cho JI, Lee SK, Ko SH, Kim HK, Jun SH, Lee YH, Bhoo SH, Lee KW, An GH, Hahn TR,
Jeon JS (2005) Molecular cloning and expression analysis of the cell-wall invertase gene
family in rice (Oryza sativa L.). Plant Cell Reports 24: 225-236









Cho YH, Yoo SD, Sheen J (2006) Regulatory functions of nuclear hexokinasel complex in
glucose signaling. Cell 127: 579-589

Chourey P, Jain M, Li QB, Carlson S (2006) Genetic control of cell wall invertases in
developing endosperm of maize. Planta 223: 159-167

Christmann A, Weiler EW, Steudle E, Grill E (2007) A hydraulic signal in root-to-shoot
signalling of water shortage. Plant Journal 52: 167-174

Clifford HT (1987). Spikelet and floral morphology. In Grass Systematics and Evolution, T.R.
Soderstrom, K.W. Hilu, C.S. Campbell, and M.E. Barkworth, eds (Washington, DC:
Smithsonian Institute) pp 21-30.

Cong B, Liu J, Tanksley SD (2002) Natural alleles at a tomato fruit size quantitative trait locus
differ by heterochronic regulatory mutations. Proc Natl Acad Sci USA 99: 13606-13611

Cooke JEK, Brown KA, Wu R, Davis JM (2003) Gene expression associated with N-induced
shifts in resource allocation in poplar. Plant, Cell, and Environment 26: 757-770

Coruzzi G, Bush DR (2001) Nitrogen and carbon nutrient and metabolite signaling in plants.
Plant Physiology 125: 61-64

Cowles CR, Hirschhorm JN, Altshuler D, Lander ES (2002) Detection of regulatory variation
in mouse genes. Nat Genetics 32: 432-437

Cruz-Aguado JA, Reyes F, Rodes R, Peres I, Dorado M (1999) Effect of source-to-sink ratio
on partitioning of dry matter and C-14-photoassimilates in wheat during grain filling.
Annals of Botany 83: 655-665

Eastmond PJ, van Dijken AJ, Spielman M, Kerr A, Tissier AF, Dickinson HG, Jones JD,
Smeekens SC, Graham IA (2002) Trehalose-6-phosphate synthase 1, which catalyzes
the first step in trehalose synthesis, is essential for Arabidopsis embryo maturation. Plant
J 29: 225-235

Ehness R, Ecker M, Godt DE, Roitsch T (1997) Glucose and stress independently regulate
source and sink metabolism and defense mechanisms via signal transduction pathways
involving protein phosphorylation. Plant Cell 9: 1825-1841

Emrich SJ, Barbazuk B, Schnable PS (2007a) Gene discovery and annotation using LCM-454
transcriptome sequencing. Genome Res 17: 69-73

Emrich SJ, Li L, Wen T-J, Yandeau-Nelson MD, Fu Y, Guo L, Chou H-H, Aluru S,
Ashlock DA, Schnable PS (2007b) Nearly identical paralogs: Implications for Maize
(Zea mays L.) genome evolution. Genetics 175: 429-439









Eveland AL, McCarty DR, Koch KE (2008) Transcript profiling by 3'-untranslated region
sequencing resolves expression of gene families. Plant Physiology 146: 32-44

Finkelstein RR, Gibson SI (2002) ABA and sugar interactions regulating development: cross-
talk or voices in a crowd? Current Opinion in Plant Biology 5: 26-32

Fisher DB, Wang N (1995) Sucrose concentration gradients along the post-phloem transport
pathway in the maternal tissues of developing wheat grains. Plant Physiology 109: 587-
592

Flemetakis E, Efrose RC, Ott T, Stedel C, Aivalakis G, Udvardi MK, Katinakis P (2006)
Spatial and temporal organization of sucrose metabolism in Lotus japonicus nitrogen-
fixing nodules suggests a role for the elusive alkaline/neutral invertase. Plant Molecular
Biology 62: 53-69

Fridman E, Zamir D (2003) Functional divergence of a syntenic invertase gene family in
tomato, potato, and Arabidopsis. Plant Physiology 131: 603-609

Fritz C, Palacios-Rojas N, Feil R, Stitt M (2006) Regulation of secondary metabolism by the
carbon-nitrogen status in tobacco: nitrate inhibits large sectors of phenylpropanoid
metabolism. Plant Journal 46: 533-548

Fu Y, Emrich SJ, Guo L, Wen T-J, Ashlock DA, Aluru S, Schnable PS (2005) Quality
assessment of maize assembled genomic islands (MAGIs) and large-scale experimental
verification of predicted genes. Proc Natl Acad Sci USA 102: 12282-12287

Fung RWM, Gonzalo M, Fekete C, Kovacs LG, He Y, Marsh E, McIntyre LM,
Schachtman DP, Qiu W (2008) Powdery mildew induces defense-oriented
reprogramming of the transcriptome in a susceptible but not in a resistant grapevine.
Plant Physiol 146: 236-249

Furusawa C, Kaneko K (2003) Zipf's law in gene expression. Phys Rev Lett 90: 88-102

Furutani I, Sukegawa S, Kyozuka J (2006) Genome-wide analysis of spatial and temporal
gene expression in rice panicle development. Plant Journal 46: 503-511

Gazzarrini S, McCourt P (2001) Genetic interactions between ABA, ethylene and sugar
signaling pathways. Current Opinion in Plant Biology 4: 387-391

Gibon Y, Biasing OE, Palacios-Rojas N, Pankovic D, Hendriks JHM, Fisahn J, Hohne M,
Gunther M, Stitt M (2004) Adjustment of diurnal starch turnover to short days:
depletion of sugar during the night leads to a temporary inhibition of carbohydrate
utilization, accumulation of sugars and post-translational activation of ADP-glucose
pyrophosphorylase in the following light period. Plant Journal 39: 847-862









Gibson SI (2004) Sugar and phytohormone response pathways: navigating a signalling network.
J Exp Bot: 55: 253-264

Godt D, Roitsch T (2006) The developmental and organ specific expression of sucrose cleaving
enzymes in sugar beet suggests a transition between apoplasmic and symplasmic phloem
unloading in the tap roots. Plant Physiology and Biochemistry 44: 656-665

Gonzalez MC, Roitsch T, Cejudo FJ (2005) Circadian and developmental regulation of
vacuolar invertase expression in petioles of sugar beet plants. Planta 222: 386-395

Grimanelli D, Perotti E, Ramirez J, Leblanc O (2005) Timing of the maternal-to-zygotic
transition during early seed development in maize. Plant Cell 17: 1061-1072

Guo M, Rupe MA, Danilevskaya ON, Yang X, Hu Z (2003) Genome-wide mRNA profiling
reveals heterochronic allelic variation and a new imprinted gene in hybrid maize
endosperm. Plant J 36: 30-44

Guo M, Rupe MA, Zinselmeier C, Habben J, Bowen BA, Smith OS (2004) Allelic variation
of gene expression in maize hybrids. Plant Cell 16: 1707-1716

Gu Z, Rifkin SA, White KP, Li W-H (2004) Duplicate genes increase gene expression
diversity within and between species. Nat Genetics 36: 577-578

Harkins T, Jarvie T (2007) Megogenomics analysis using the Genome Sequencer FLX system.
Nat Methods 4: application notes

Harrington GN, Bush DR (2003) The bifunctional role of hexokinase in metabolism and
glucose signaling. Plant Cell 15: 2493-2496

Helentjaris T, Weber D, Wright S (1988) Identification of the genomic locations of duplicate
nucleotide sequences in maize by analysis of restriction fragment length polymorphisms.
Genetics 118: 353-363

Hoecker U, Vasil IK, McCarty DR (1995) Integrated control of seed maturation and
germination programs by activator and repressor functions of Viviparous-] of maize.
Genes & Development 9: 2459-2469

Hofmann J, Wieczorek K, Blochl A, Grundler FMW (2007) Sucrose supply to nematode-
induced syncytia depends on the apoplasmic and symplasmic pathways. Journal of
Experimental Botany 58: 1591-1601

Holland N, Holland D, Helentjaris T, Dhugga KS, Xoconostle-Cazares B, Delmer DP (2000)
A comparative analysis of the plant Cellulose Synthase (CesA) gene family. Plant Physiol
123: 1313-1323









Hothorn M, Wolf S, Aloy P, Greiner S, Scheffzek K (2004) Structural insights into the target
specificity of plant invertase and pectin methylesterase inhibitory proteins. Plant Cell 16:
3437-3447

Huang J, Taylor JP, Chen J-G, Uhrig JF, Schnell DJ, Nakagawa T, Korth KL, Jones AM
(2006) The plastid protein THYLAKOID FORMATION1 and the plasma membrane G-
protein GPA1 interact in a novel sugar-signaling mechanism in Arabidopsis. Plant Cell
18: 1226-1238

Huang LF (2006) Molecular analysis of an invertase gene family in Arabidopsis. Thesis (PhD)
University of Florida

Huang LF, Bocock PN, Davis JM, Koch KE (2007) Regulation of invertase: a suite of
transcriptional and post-transcriptional mechanisms. Functional Plant Biology 34: 499-
507

Huang X, Madan A (1999) CAP3: A DNA sequence assembly program. Genome Res 9: 868-
877

Huijser C, Kortstee A, Pego J, Weisbeek P, Wisman E, Smeekens S (2000) The Arabidopsis
SUCROSE UNCOUPLED-6 gene is identical to ABSCISIC ACID INSENSITIVE-4:
involvement of abscisic acid in sugar responses. Plant Journal 23: 577-585

Ji XM, Raveendran M, Oane R, Ismail A, Lafitte R, Bruskiewich R, Cheng SH, Bennett J
(2005) Tissue-specific expression and drought responsiveness of cell-wall invertase
genes of rice at flowering. Plant Molecular Biology 59: 945-964

Jongeneel CV, Delorenzi M, Iseli C, Zhou D, Haudenchild CD, Khrebtukova I, Kuznetsov
D, Stevenson BJ, Strausberg RL, Simpson AJG, et al (2005) An atlas of human gene
expression from massively parallel signature sequences (MPSS) Genome Res 15: 1007-
1014

Karthikeyan AS, Varadarajan DK, Jain A, Held MA, Carpita NC, Raghothama KG (2007)
Phosphate starvation responses are mediated by sugar signaling in Arabidopsis. Planta
225: 907-918

Kim JY, Mahe A, Brangeon J, Prioul JL (2000a) A maize vacuolar invertase, IVR2, is
induced by water stress. Plant Physiology 124: 71-84

Kim JY, Mahe A, Guy S, Brangeon J, Roche O, Chourey PS, Prioul JL (2000b)
Characterization of two members of the maize gene family, Incw3 and Incw4, encoding
cell-wall invertases. Gene 245: 89-102

Kladnik A, Chamusco K, Dermastia M, Chourey P (2004) Evidence of programmed cell
death in post-phloem transport cells of the maternal pedicel tissue in developing
caryopsis of maize. Plant Physiology 136: 3572-3581









Koch KE (1996) Carbohydrate-modulated gene expression in plants. Annual Review of Plant
Physiology and Plant Molecular Biology 47: 509-540

Koch KE (2004) Sucrose metabolism: regulatory mechanisms and pivotal roles in sugar sensing
and plant development. Current Opinion in Plant Biology 7: 235-246

Koch KE, Ying Z, Wu Y, Avigne WT (2000) Multiple paths of sugar-sensing and a
sugar/oxygen overlap for genes of sucrose and ethanol metabolism. Journal of
Experimental Botany 51: 417-427

Kohorn BD, Kobayashi M, Johansen S, Riese J, Huang LF, Koch K, Fu S, Dotson A, Byers
N (2006) An Arabidopsis cell wall-associated kinase required for invertase activity and
cell growth. Plant Journal 46: 307-316

Kojima H, Suzuki T, Kato T, Enomoto K, Sato S, Kato T, Tabata S, Saez-Vasquez J,
Manuel E, Nakagawa T, Ishiguro S, Nakamura K (2007) Sugar-inducible expression
of the nucleolin-1 gene of Arabidopsis thaliana and its role in ribosome synthesis, growth
and development. Plant Journal 49: 1053-1063

Kuznetsov VA, Knott GD, Bonner RF (2002) General statistics of stochastic process of gene
expression in eukaryotic cells. Genetics 161: 1321-1332

Laby RJ, Kincaid MS, Kim DG, Gibson SI (2000) The Arabidopsis sugar-insensitive mutants
sis4 and sis5 are defective in abscisic acid synthesis and response. Plant Journal 23: 587-
596

Lalonde S, Boles E, Hellmann H, Barker L, Patrick JW, Frommer WB, Ward JM (1999)
The dual function of sugar carriers: Transport and sugar sensing. Plant Cell 11: 707-726

Lalonde S, Wipf D, Frommer WB (2004) Transport mechanisms for organic forms of carbon
and nitrogen between source and sink. Annual Review of Plant Biology 55: 341-372

Lara MEB, Garcia MCG, Fatima T, Ehness R, Lee TK, Proels R, Tanner W, Roitsch T
(2004) Extracellular invertase is an essential component of cytokinin-mediated delay of
senescence. Plant Cell 16: 1276-1287

Lee EJ, Matsumura Y, Soga K, Hoson T, Koizumi N (2007) Glycosyl hydrolases of cell wall
are induced by sugar starvation in Arabidopsis. Plant and Cell Physiology 48: 405-413

Leon P, Sheen J (2003) Sugar and hormone connections. Trends in Plant Science 8: 110-116

Li YH, Lee KK, Walsh S, Smith C, Hadingham S, Sorefan K, Cawley G, Bevan MW (2006)
Establishing glucose- and ABA-regulated transcription networks in Arabidopsis by
microarray analysis and promoter classification using a Relevance Vector Machine.
Genome Research 16: 414-427









Li YH, Smith C, Corke F, Zheng LY, Merali Z, Ryden P, Derbyshire P, Waldron K, Bevan
MW (2007) Signaling from an altered cell wall to the nucleus mediates sugar-responsive
growth and development in Arabidopsis thaliana. Plant Cell 19: 2500-2515

Linden, JC, Ehness, R, and Roitsch, T (1996) Regulation by ethylene of apoplastic invertase
expression in Chenopodium rubrum tissue culture cells. Plant Growth Regul 19: 219-222

Liu F, Jenssen T-K, Trimarchi J, Punzo C, Cepko CL, Ohno-Machado L, Hovig E, Kuo
WP (2007) Comparison of hybridization-based and sequencing-based gene expression
technologies on biological replicates. BMC Genomics 8: 153

Long JC, Zhao W, Rashotte AM, Muday GK, Huber SC (2002) Gravity-stimulated changes
in auxin and invertase gene expression in maize pulvinal cells. Plant Physiology 128:
591-602

Lou Y, Gou J-Y, Xue H-W (2007) PIP5K9, an Arabidopsis phosphatidylinositol
monophosphate kinase, interacts with a cytosolic invertase to negatively regulate sugar-
mediated root growth. Plant Cell 19: 163-181

Lu C, Tej SS, Luo S, Haudenchild CD, Meyers BC, Green PJ (2005) Elucidation of the small
RNA component of the transcriptome. Science 309: 1567-1569

Ma J, Morrow DJ, Fernandes J, Walbot V (2006) Comparative profiling of sense and
antisense transcriptome of maize lines. Genome Biology 7: R22

Maitz M, Santandrea G, Zhang ZY, Lal S, Hannah LC, Salamini F, Thompson RD (2000)
rgfl, a mutation reducing grain filling in maize through effects on basal endosperm and
pedicel development. Plant Journal 23: 29-42

Makela P, McLaughlin JE, Boyer JS (2005) Imaging and quantifying carbohydrate transport to
the developing ovaries of maize. Annals of Botany 96: 939-949

Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, Berka J,
Braverman MS, Chen Y-J, Chen Z, Dewell, et al. (2005) Genome sequencing in
microfabricated high-density picolitre reactors. Nature 437: 376-380

Martin A, Lee J, Kichey T, Gerentes D, Zivy M, Tatout C, Dubois F, Balliau T, Valot B,
Davanture M, et al (2006) Two glutamine synthetase isoforms of maize are specifically
involved in the control of grain production. Plant Cell 18: 3252-3274

McCarty DR, Hattori T, Carson CB, Vasil V, Lazar M, Vasil IK (1991) The Viviparous-1
developmental gene of maize encodes a novel transcriptional activator. Cell 66: 895-905

McCormick AJ, Cramer MD, Watt DA (2008) Changes in photosynthetic rates and gene
expression of leaves during a source-sink perturbation in sugarcane. Annals of Botany
101: 89-102









McLaughlin JE, Boyer JS (2004a) Glucose localization in maize ovaries when kernel number
decreases at low water potential and sucrose is fed to the stems. Annals of Botany 94: 75-
86

McLaughlin JE, Boyer JS (2004b) Sugar-responsive gene expression, invertase activity, and
senescence in aborting maize ovaries at low water potentials. Annals of Botany 94: 675-
689

Menges M, Samland AK, Planchais S, Murray JAH (2006) The D-type cyclin CYCD3:1 is
limiting for the Gl-to-S-phase transition in Arabidopsis. Plant Cell 18: 893-906

Messing J, Bharti AK, Karlowski WM, Gundlach H, Kim HR, Yu Y, Wei F, Fuks G,
Soderlund CA, Mayer KFX, et al. (2004) Sequence composition and genome
organization of maize. Proc Natl Acad Sci USA 101: 14349-14354

Messing J, Dooner HK (2006) Organization and variability of the maize genome. Curr Opin
Plant Biol 9: 157-163

Meyers BC, Tej SS, Vu TH, Haudenschild CD, Agrawal V, Edberg SB, Ghazal H, Decola S
(2004) The use of MPSS for whole-genome transcriptional analysis in Arabidopsis.
Genome Res 14: 1641-1653

Mockler TC, Ecker JR (2005) Applications of DNA tiling arrays for whole-genome analysis.
Genomics 85: 1-15

Moore B, Zhou L, Rolland F, Hall Q, Cheng WH, Liu YX, Hwang I, Jones T, Sheen J
(2003) Role of the Arabidopsis glucose sensor HXK1 in nutrient, light, and hormonal
signaling. Science 300: 332-336

Moore RC, Purugganan MD (2005) The evolutionary dynamics of plant duplicate genes. Curr
Opin Plant Biol 8: 122-128

Myers PN, Setter TL, Madison JT, Thompson JF (1992) Endosperm cell division in maize
kernels cultured at 3 levles of water potential. Plant Physiology 99: 1051-1056

Ner-Gaon H, Leviatan N, Rubin E, Fluhr R (2007) Comparative cross-species alternative
splicing in plants. Plant Physiol 144: 1632-1641

Newman-Smith ED, Rothman JH (1998) The maternal-to-zygotic transition in embryonic
patterning of Caenorhabditis elegans. Curr Opin Genet Dev 8: 472-480

Newport J, Kirschner M (1982) A major developmental transition in early Xenopus embryos:
I. Characterization and timing of cellular changes at the midblastula stage. Cell 30: 675-
686.









Nicolai M, Roncato MA, Canoy AS, Rouquie D, Sarda X, Freyssinet G, Robaglia C (2006)
Large-scale analysis of mRNA translation states during sucrose starvation in Arabidopsis
cells identifies cell proliferation and chromatin structure as targets of translational
control. Plant Physiology 141: 663-673

Niittylae T, Fuglsang AT, Palmgren MG, Frommer WB, Schulze WX (2007) Temporal
analysis of sucrose-induced phosphorylation changes in plasma membrane proteins of
Arabidopsis. Molecular & Cellular Proteomics 6: 1711-1726

Niu XP, Helentjaris T, Bate NJ (2002) Maize AB14 binds coupling element in abscisic acid
and sugar response genes. Plant Cell 14: 2565-2575

Nobuta K, Venu RC, Lu C, Belo A, Vemaraju K, Kulkarni K, Wang W, Pillay M, Green
PJ, Wang G-l, Meyers BC (2007) An expression atlas of rice mRNAs and small RNAs.
Nat Biotechnol 25: 473-477

Oliver SN, Dennis ES, Dolferus R (2007) ABA regulates apoplastic sugar transport and is a
potential signal for cold-induced pollen sterility in rice. Plant and Cell Physiology 48:
1319-1330

Osborn TC, Pires JC, Birchler JA, Auger DL, Chen ZJ, Lee H-S, Comai L, Madlung A,
Doerge RW, Colot V, et al. (2003) Understanding mechanisms of novel gene
expression in polyploids. Trends Genet 19: 141-147

Osuna D, Usadel B, Morcuende R, Gibon Y, Biasing OE, Hohne M, Gunter M, Kamlage B,
Trethewey R, Scheible WR, Stitt M (2007) Temporal responses of transcripts, enzyme
activities and metabolites after adding sucrose to carbon-deprived Arabidopsis seedlings.
Plant Journal 49: 463-491

Ozcan S, Dover J, Johnston M (1998) Glucose sensing and signaling by two glucose receptors
in the yeast Saccharomyces cerevisiae. EMBO 17: 2566-2573

Palenchar PM, Kouranov A, Lejay LV, Coruzzi GM (2004) Genome-wide patterns of carbon
and nitrogen regulation of gene expression validate the carbon and nitrogen (CN)-
signaling hypothesis. Genome Biol 5: R91

Pan QH, Li MJ, Peng CC, Zhang N, Zou X, Zou KQ, Wang XL, Yu XC, Wang XF, Zhang
DP (2005) Abscisic acid activates acid invertases in developing grape berry. Physiologia
Plantarum 125: 157-170

Patrick JW (1997) Phloem unloading: Sieve element unloading and post-sieve element
transport. Annual Review of Plant Physiology and Plant Molecular Biology 48: 191-222

Paul MJ, Foyer CH (2001) Sink regulation of photosynthesis. Journal of Experimental Botany
52: 1383-1400









Pelegri F (2003). Maternal factors in zebrafish development. Dev Dyn 228: 535-554

Pelleschi S, Leonardi A, Rocher JP, Comic G, de Vienne D, Thevenot C, Prioul JL (2006)
Analysis of the relationships between growth, photosynthesis and carbohydrate
metabolism using quantitative trait loci (QTLs) in young maize plants subjected to water
deprivation. Molecular Breeding 17: 21-39

Peng YB, Lu YF, Zhang DP (2003) Abscisic acid activates ATPase in developing apple fruit
especially in fruit phloem cells. Plant Cell and Environment 26: 1329-1342

Price J, Laxmi A, St Martin SK, Jang JC (2004) Global transcription profiling reveals
multiple sugar signal transduction mechanisms in Arabidopsis. Plant Cell 16: 2128-2150

Pourtau N, Mares M, Purdy S, Quentin N, Ruel A, Wingler A. (2004) Interactions of abscisic
acid and sugar signalling in the regulation of leaf senescence. Planta 219: 765-772

Proels RK, Hause B, Berger S, Roitsch T (2003) Novel mode of hormone induction of tandem
tomato invertase genes in floral tissues. Plant Molecular Biology 52: 191-201

Qin LX, Trouverie J, Chateau-Joubert S, Simond-Cote E, Thevenot C, Prioul JL (2004)
Involvement of the Ivr2-invertase in the perianth during maize kernel development under
water stress. Plant Science 166: 371-379

Qu Y, Xu S (2006) Quantitative trait associated microarray gene expression data analysis.
Molecular Biology and Evolution 23: 1558-1573

Ramon M, Rolland F, Thevelein JM, Van Dijck P, Leyman B (2007) ABI4 mediates the
effects of exogenous trehalose on Arabidopsis growth and starch breakdown. Plant
Molecular Biology 63: 195-206

Riou-Khamlichi C, Menges M, Healy JMS, Murray JAH (2000) Sugar control of the plant
cell cycle: Differential regulation of Arabidopsis D-type cyclin gene expression.
Molecular and Cellular Biology 20: 4513-4521

Ritter H, Schulz GE (2004) Structural basis for the entrance into the phenylpropanoid
metabolism catalyzed by phenylalanine ammonia-lyase. Plant Cell 16: 3426-3436

Rogers LA, Dubos C, Cullis IF, Surman C, Poole M, Willment J, Mansfield SD, Campbell
MM (2005) Light, the circadian clock, and sugar perception in the control of lignin
biosynthesis. Journal of Experimental Botany 56: 1651-1663

Rohde A, Prinsen E, De Rycke R, Engler G, Van Montagu M, Boerjan W (2002) PtABI3
impinges on the growth and differentiation of embryonic leaves during bud set in poplar.
Plant Cell 14: 1885-1901









Roitsch T (1999) Source-sink regulation by sugar and stress. Current Opinion in Plant Biology
2: 198-206

Roitsch T, Balibrea ME, Hofmann M, Proels R, Sinha AK (2003) Extracellular invertase: key
metabolic enzyme and PR protein. Journal of Experimental Botany 54: 513-524

Roitsch T, Ehness R, Goetz M, Hause B, Hofmann M, Sinha AK (2000) Regulation and
function of extracellular invertase from higher plants in relation to assimilate partitioning,
stress responses and sugar signalling. Australian Journal of Plant Physiology 27: 815-825

Roitsch T, Gonzalez MC (2004) Function and regulation of plant invertases: sweet sensations.
Trends in Plant Science 9: 606-613

Rojo E, Zouhar J, Carter C, Kovaleva V, Raikhel NV (2003) A unique mechanism for protein
processing and degradation in Arabidopsis thaliana. Proceedings of the National
Academy of Sciences of the United States of America 100: 7389-7394

Rolland F, Baena-Gonzalez E, Sheen J (2006) Sugar sensing and signaling in plants:
Conserved and novel mechanisms. Annual Review of Plant Biology 57: 675-709

Rolland F, Winderickx J, Thevelein JM (2002) Glucose-sensing and signaling mechanisms in
yeast. FEMS Yeast Res 2: 183-201

Rook F, Corke F, Baier M, Holman R, May AG, Bevan MW (2006) Impaired sucrose
inductionI encodes a conserved plant-specific protein that couples carbohydrate
availability to gene expression and plant growth. Plant Journal 46: 1045-1058

Rook F, Corke F, Card R, Munz G, Smith C, Bevan MW (2001) Impaired sucrose-induction
mutants reveal the modulation of sugar-induced starch biosynthetic gene expression by
abscisic acid signalling. Plant Journal 26: 421-433

Rook F, Hadingham SA, Li Y, Bevan MW (2006) Sugar and ABA response pathways and the
control of gene expression. Plant Cell and Environment 29: 426-434

Saha S, Sparks AB, Rago C, Akmaev V, Wang CJ, Vogelstein B, Kinzler KW, Velculescu
VE (2002) Using the transcriptome to annotate the genome. Nat Biotechnol 20: 508-512

Satoh-Nagasawa N, Nagasawa N, Malcomber S, Sakai H, Jackson D (2006) A trehalose
metabolic enzyme controls inflorescence architecture in maize. Nature 441: 227-230

Sauer N (2007) Molecular physiology of higher plant sucrose transporters. Febs Letters 581:
2309-2317

Schaarschmidt S, Kopka J, Ludwig-Muller J, Hause B (2007) Regulation of arbuscular
mycorrhization by apoplastic invertases: enhanced invertase activity in the leaf apoplast
affects the symbiotic interaction. Plant Journal 51: 390-405









Schadt EE, Monks SA, Drake TA, Lusis AJ, Che N, Collnayo V, Ruff TG, Milligan SB,
Lamb JR, Cavet G, et al (2003) Genetics of gene expression surveyed in maize, mouse,
and man. Nature 422: 297-302

Schmidt RJ, Veit B, Mandel MA, Mena M, Hake S, Yanofsky MF (1993) Identification and
molecular characterization of ZAG1, the maize homolog of the Arabidopsis floral
homeotic gene AGAMOUS. Plant Cell 5: 729-737

Scofield GN, Hirose T, Aoki N, Furbank RT (2007) Involvement of the sucrose transporter,
OsSUT1, in the long-distance pathway for assimilate transport in rice. Journal of
Experimental Botany 58: 3155-3169

Seebauer JR, Moose SP, Fabbri BJ, Crossland LD, Below FE (2004) Amino acid metabolism
in maize earshoots. Implications for assimilate preconditioning and nitrogen signaling.
Plant Physiology 136: 4326-4334

Sergeeva LI, Keurentjes JJB, Bentsink L, Vonk J, van der Plas LHW, Koornneef M,
Vreugdenhil D (2006) Vacuolar invertase regulates elongation of Arabidopsis thaliana
roots as revealed by QTL and mutant analysis. Proceedings of the National Academy of
Sciences of the United States of America 103: 2994-2999

Setter TL, Flannigan BA, Melkonian J (2001) Loss of kernel set due to water deficit and shade
in maize: Carbohydrate supplies, abscisic acid, and cytokinins. Crop Science 41: 1530-
1540

Sherson SM, Alford HL, Forbes SM, Wallace G, Smith SM (2003) Roles of cell-wall
invertases and monosaccharide transporters in the growth and development of
Arabidopsis. Journal of Experimental Botany 54: 525-531

Smeekens S (2000) Sugar-induced signal transduction in plants. Annual Review of Plant
Physiology and Plant Molecular Biology 51: 49-81

Smith AM, Stitt M (2007) Coordination of carbon supply and plant growth. Plant Cell and
Environment 30: 1126-1149

Soderman EM, Brocard IM, Lynch TJ, Finkelstein RR (2000) Regulation and function of the
arabidopsis ABA-insensitive4 gene in seed and abscisic acid response signaling
networks. Plant Physiology 124: 1752-1765

Springer NM, Stupar RM (2007) Allele-specific expression patterns reveal biases and embryo-
specific parent-of-origin effects in hybrid maize. Plant Cell 19: 2391-2402

Sreenivasulu N, Altschmied L, Radchuk V, Gubatz S, Wobus U, Weschke W (2004)
Transcript profiles and deduced changes of metabolic pathways in maternal and filial
tissues of developing barley grains. Plant Journal 37: 539-553









Stafstrom JP, Ripley BD, Devitt ML, Drake B (1998) Dormancy-associated gene expression
in pea axillary buds. Planta 205: 547-552

Stitt M, Gibon Y, Lunn JE, Piques M (2007) Multilevel genomics analysis of carbon
signalling during low carbon availability: coordinating the supply and utilisation of
carbon in a fluctuating environment. Functional Plant Biology 34: 526-549

Street NR, Skogstrom O, Sjodin A, Ticker J, Rodriguez-Acosta M, Nilsson P, Jansson S,
Taylor G (2006) The genetics and genomics of drought response in Populus. Plant J 48:
321-341

Stupar RM and Springer NM (2006) Cis-transcriptional variation in maize inbred lines B73
and Mol7 leads to additive expression patterns in the Fl Hybrid. Genetics 173: 2199-
2210

Sturm A (1999) Invertases. Primary structures, functions, and roles in plant development and
sucrose partitioning. Plant Physiology 121: 1-7

Sturm A, Tang GQ (1999) The sucrose-cleaving enzymes of plants are crucial for development,
growth and carbon partitioning. Trends in Plant Science 4: 401-407

Suzuki M, Kao CY, Cocciolone S, McCarty DR (2001) Maize VP1 complements Arabidopsis
abi3 and confers a novel ABA/auxin interaction in roots. Plant Journal 28: 409-418

Suzuki M, Ketterling MG, Li QB, McCarty DR (2003) Viviparous alters global gene
expression patterns through regulation of abscisic acid signaling. Plant Physiology 132:
1664-1677

Swanson-Wagner RA, Joa Y, DeCook R, Borsuk LA, Nettleton D, Schnable PS (2006) All
possible modes of gene action are observed in a global comparison of gene expression in
a maize Fl hybrid and its inbred parents. Proc Natl Acad Sci USA 103: 6805-6810

Taliercio EW, Kim JY, Mahe A, Shanker S, Choi J, Cheng WH, Prioul JL, Chourey PS
(1999) Isolation, characterization and expression analyses of two cell wall invertase
genes in maize. Journal of Plant Physiology 155: 197-204

Trouverie J, Chateau-Joubert S, Thevenot C, Jacquemot MP, Prioul JL (2004) Regulation
of vacuolar invertase by abscisic acid or glucose in leaves and roots from maize plantlets.
Planta 219: 894-905

Trouverie J, Prioul JL (2006) Increasing leaf export and grain import capacities in maize plants
under water stress. Functional Plant Biology 33: 209-218

Tuberosa R, Salvi S (2006) Genomics-based approaches to improve drought tolerance of crops.
Trends in Plant Science 11: 405-412









Tymowska-Lalanne Z, Kreis M (1998) Expression of the Arabidopsis thaliana invertase gene
family. Planta 207: 259-265

van Ruissen F, Ruijter JM, Schaaf GJ, Asgharnegad L, Zwijnenburg DA, Kool M, Baas F
(2005) Evaluation of the similarity of gene expression data estimated with SAGE and
Affymetrix GeneChips. BMC Genomics 6: 91

Vargas WA, Pontis HG, Salerno GL (2007) Differential expression of alkaline and neutral
invertases in response to environmental stresses: characterization of an alkaline isoform
as a stress-response enzyme in wheat leaves. Planta 226: 1535-1545

Vaughn MW, Harrington GN, Bush DR (2002) Sucrose-mediated transcriptional regulation of
sucrose symporter activity in the phloem. Proceedings of the National Academy of
Sciences of the United States of America 99: 10876-10880

Vielle-Calzada JP, Baskar R, Grossniklaus U (2000) Delayed activation of the paternal
genome during seed development. Nature 404: 91-94

Vij S, Tyagi AK (2007) Emerging trends in the functional genomics of the abiotic stress
response in crop plants. Plant Biotechnology Journal 5: 361-380

Vroh Bi I, McMullen MD, Sanchez-Villeda H, Schroeder S, Gardiner J, Polacco M,
Soderlund C, Wing R, Fang Z, Coe Jr. EH (2006) Single nucleotide polymorphisms
and insertion-deletions for genetic markers and anchoring the maize fingerprint contig
physical map. Crop Sci. 46: 12-21

Wachter R, Langhans M, Aloni R, Gotz S, Weilmunster A, Koops A, Temguia L, Mistrik I,
Pavlovkin J, Rascher U, Schwalm K, Koch KE, Ullrich CI (2003) Vascularization,
high-volume solution flow, and localized roles for enzymes of sucrose metabolism during
tumorigenesis by Agrobacterium tumefaciens. Plant Physiology 133: 1024-1037

Weber APM, Weber KL, Carr K, Wilkerson C, Ohlrogge JB (2007) Sampling the
Arabidopsis transcriptome with massively parallel pyrosequencing. Plant Physiol 144:
32-42

Welcker C, Boussuge B, Bencivenni C, Ribaut JM, Tardieu F (2007) Are source and sink
strengths genetically linked in maize plants subjected to water deficit? A QTL study of
the responses of leaf growth and of Anthesis-Silking Interval to water deficit. Journal of
Experimental Botany 58: 339-349

Weschke W, Panitz R, Gubatz S, Wang Q, Radchuk R, Weber H, Wobus U (2003) The role
of invertases and hexose transporters in controlling sugar ratios in maternal and filial
tissues of barley caryopses during early development. Plant Journal 33: 395-411

Westgate ME, Boyer JS (1985) Osmotic adjustment and the inhibition of leaf, root, stem and
silk growth at low water potentials in maize. Planta 164: 540-549









Westgate ME, Boyer JS (1986) Reproduction at low silk and pollen water potentials in maize.
Crop Science 26: 951-956

Whipple CJ, Ciceri P, Padilla CM, Ambrose BA, Bandong SL, Schmidt RJ (2004)
Conservation of B-class floral homeotic gene function between maize and Arabidopsis.
Development 131: 6083-6091

Wiese A, Elzinga N, Wobbes B, Smeekens S (2004) A conserved upstream open sucrose-
induced repression reading frame mediates of translation. Plant Cell 16: 1717-1729

Winter H, Huber SC (2000) Regulation of sucrose metabolism in higher plants: Localization
and regulation of activity of key enzymes. Critical Reviews in Biochemistry and
Molecular Biology 35: 253-289

Wobus U, Weber H (1999) Seed maturation: genetic programmes and control signals. Curr
Opin Plant Biol 2: 33-38

Woodson WR and H Wang (1987) Invertases of carnation petals. Partial purification,
characterization, and changes in activity during petal growth. Physiologia Plantarum 71:
224-228

Wormit A, Trentmann O, Feifer I, Lohr C, Tjaden J, Meyer S, Schmidt U, Martinoia E,
Neuhaus HE (2006) Molecular identification and physiological characterization of a
novel monosaccharide transporter from Arabidopsis involved in vacuolar sugar transport.
Plant Cell 18: 3476-3490

Wright SI, Bi IV, Schroeder SG, Yamasaki M, Doebley JF, McMullen MD, Gaut BS (2005)
The effects of artificial selection on the maize genome. Science 308: 1310-1314

Xiao W-Y, Sheen J, Jang J-C (2000) The role of hexokinase in plant sugar signal transduction
and growth and development. Plant Mol Biol 44: 451-461

Xu J, Avigne WT, McCarty DR, Koch KE (1996) A similar dichotomy of sugar modulation
and developmental expression affects both paths of sucrose metabolism: Evidence from a
maize invertase gene family. Plant Cell 8: 1209-1220

Yamada K, Lim J, Dale JM, Chen H, Shinn P, Palm CJ, Southwick AM, Wu HC, Kim C,
Nguyen M, et al. (2003) Empirical analysis of transcriptional activity in the Arabidopsis
genome. Science 302: 842-846

Yan H, Yuan W, Velculescu VE, Vogelstein B, Kinzler KW (2002) Allelic variation in human
gene expression. Science 297: 1143

Yun HS, Yoon IS, Kang BG (2002) Rapid repression of vacuolar invertase in mungbean
hypocotyl segments and regulation by sucrose, auxin and light. Plant Growth Regulation
38: 181-189









Zhang X, Madi S, Borsuk L, Nettleton D, Elshire RJ, Buckner B, Janick-Buckner D, Beck
J, Timmermans M, Schnable PS, Scanlon MJ (2007) Laser microdissection of narrow
sheath mutant maize uncovers novel gene expression in the shoot apical meristem. PLoS
Biology el01: 1040-1052

Zhuang Y, Ren G, Yue G, Li Z, Qu X, Hou G, Zhu Y, Zhang J (2007) Effects of water deficit
stress on the transcriptomes of developing immature ear and tassel in maize. Plant Cell
Rep 26: 2137-2147

Zimmermann P, Hirsch-Hoffmann M, Hennig L, Gruissem W (2004) GENEVESTIGATOR.
Arabidopsis Microarray Database and Analysis Toolbox. Plant Physiol 136: 2621-2632

Zinselmeier C, Jeong BR, Boyer JS (1999) Starch and the control of kernel number in maize at
low water potentials. Plant Physiology 121: 25-35

Zinselmeier C, Sun YJ, Helentjaris T, Beatty M, Yang S, Smith H, Habben J (2002) The use
of gene expression profiling to dissect the stress sensitivity of reproductive development
in maize. Field Crops Research 75: 111-121

Zinselmeier C, Westgate ME, Schussler JR, Jones RJ (1995) Low water potential disrupts
carbohydrate metabolism in maize ovaries. Plant Physiology 107: 385-391









BIOGRAPHICAL SKETCH

Andrea Lee Eveland was born on May 2nd, 1977 to Frank and Joan Eveland on Long Island

of New York. Andrea first developed a deep appreciation for biology through the guidance Dr.

Alan Katz, her biology teacher at Commack High School. She became fascinated by plants in

her botany studies at Binghamton University (SUNY) where, as an undergraduate, she served as

TA for a botany lab for 2 years. Andrea earned her Bachelors of Science degree for a dual major

in biology and environmental science in May, 2000. At graduation, she was awarded the James

D. Grierson foundation award for Excellence in Botany. During her time at Binghamton

University, Andrea studied tropical ecology in Costa Rica and was involved in sustainable

agriculture and forest restoration projects. She also held positions at various nurseries and

greenhouses as manager and consultant.

In the fall of 2000, Andrea moved to San Diego, CA, and joined a plant pathology lab

group at Torry Mesa Research Institute, Syngenta Inc, as a research assistant where she learned

molecular-based techniques. In June 2002 she entered the Plant Molecular and Cell Biology

(PMCB) program at the University of Florida as a Ph.D. student and in spring 2003 began

research in Karen Koch's lab. Here she was able to combine new-found interests in molecular

biology with her love for plant physiology. During her time as a Ph.D. student, Andrea received

a number of awards including the ASPB/Pioneer Hi-Bred International Graduate Student Prize,

the IFAS Scholarship for Women in Agriculture, awards for talks at the annual PMCB retreat,

and several travel grants. She also had the opportunity to present her work at a number of

international meetings and meet with many influential people in her field.





PAGE 1

1 REGULATION OF SINK STRENGTH IN DEVELOPING MAIZE FLORETS: IMPLICATIONS FOR SEED SET AND GRAIN YIELD By ANDREA LEE EVELAND A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Andrea L. Eveland

PAGE 3

3 To Joshua Shome, in loving memory.

PAGE 4

4 ACKNOWLEDGMENTS Many thanks go to the m embers of my comm ittee, Donald McCarty, John Davis, Robert Ferl, and Edward Braun, for all of their advice and critical insight to specific areas of my research. I also whole-heartedl y thank the Koch lab group, Sylvia De Sousa, Chip Hunter, Brent OBrien, and especially Wayne Avigne, for their friendship and support every step of the way. Special thanks go to Li-Fen Huang, whose wo rk ethic and precision inspired me as a young graduate student. I thank Jie Yang, Matias Kirst, and Lauren Mc Intyre for their time and advice in the statistical analyses for this work. Also, thanks go to Marina Te lonis-Scott and Mick Popp for allowing me to use their microarray facility and providing support during optimization of array hybridizations. I thank Eric Schmelz and De nise Tieman for their assistance with GC/MS and HPLC experiments, respectively. Many thanks go to Tom Davenport and Jonathan Crane at T.R.E.C. for their kind support durin g my stay at their research f acility. I thank the camaraderie of friends and fellow colleagues, Stefanie Maruhnich, Maggie Kell ogg, Michele Auldridge, Travis Baughman, and Matt Reyes for insightful conversations over the years. Very special thanks go to my parents for their love and suppor t throughout my life. They have given so much and I owe to them my determ ination and achievements. Finally, I thank Karen Koch for her support a nd guidance during this very important time in my life. She has been not only an advisor an d colleague, but an extraordinary role model for me throughout the years. Her passion for science and her constructive criticisms have cultivated my development into an aspiring young investigator.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF FIGURES.........................................................................................................................8 LIST OF ABBREVIATIONS........................................................................................................ 10 ABSTRACT...................................................................................................................................11 CHAP TER 1 LITERATURE REVIEW.......................................................................................................13 Introduction................................................................................................................... ..........13 Source/Sink Relations and Su crose Metabolizing Enzym es.................................................. 14 Sugar Sensing and Signaling.................................................................................................. 16 Hormone and Sugar Signaling Networks............................................................................... 19 Maize Reproductive Development.........................................................................................21 2 TRANSCRIPT PROFILING BY 3'-UTR SEQUENCING RESOLVES EXPRESSION OF GENE FAMILIES ............................................................................................................23 Introduction................................................................................................................... ..........23 Results.....................................................................................................................................27 Construction of a 3'-cDNA Library.................................................................................27 Data Assembly and Analysis........................................................................................... 28 Analysis of 3'-UTR Profile Reveals a Dynamic Range of Expression...........................29 Distinguishing Gene Family Members............................................................................ 30 Evaluation of Differential Expression between Multiplexed Sub-libraries.................... 31 Resolution of Near-Identical Transcripts by Polymorphisms......................................... 32 Validation of SNPs and Hom opolymer-Based Polym ophisms....................................... 33 Discussion...............................................................................................................................34 Future Prospects......................................................................................................................40 Conclusions.............................................................................................................................41 Materials and Methods...........................................................................................................41 Plant Materials.................................................................................................................41 Sub-library Preparation and Sequencing......................................................................... 42 Data Analysis...................................................................................................................43 Real-time RT-PCR Analysis for Validation of 454 Data................................................ 44

PAGE 6

6 3 EXPRESSION PROFILING OF DEV ELOPING MAIZE OVARIES USING MICROARRAYS AND SEQUENCING OF 3'UTRS......................................................... 57 Introduction................................................................................................................... ..........57 Results and Discussion......................................................................................................... ..61 Microarray Data Analysis and Clustering....................................................................... 61 Identifying Clusters of Co-Expressed Genes.................................................................. 62 Co-Expression of Genes Related to Common Metabolic Pathways ...............................63 Sequence-Based Analyses by 3'-UTR Profiling.............................................................. 65 Comparison of Microarray and Quantitative 3'-UTR Profiles........................................ 66 Conclusions.............................................................................................................................68 Materials and Methods...........................................................................................................68 Plant Material..................................................................................................................68 RNA Extraction and Target Labeling..............................................................................69 Microarray Slide Preparation and Target Hybridization................................................. 69 Microarray Data Analysis................................................................................................ 71 Cluster Analysis using Orthogonal Polynomials............................................................. 72 Construction of a 454 3'-UTR library..............................................................................72 4 C ALLOCATION AND USE IN DEV ELOPING MAIZE FEMALE FLORETS................ 90 Introduction................................................................................................................... ..........90 Results.....................................................................................................................................94 Staging of Pre-Pollination Floral Development.............................................................. 94 Pre-Pollination Carbon Allocation and Sucr ose Use in Individua l Floral Organs.......... 95 Temporal and Spatial, IVR2 -Based Sucrose Use in Developing Fe male Florets............ 96 Post-Pollination C Accumulation in Pedicels.................................................................. 98 Co-Expression of Genes Related to C Sink Development.............................................. 98 Post-Pollination Lignin Bi osynthesis in Pedicels............................................................ 99 Discussion...............................................................................................................................99 Conclusions...........................................................................................................................102 Materials and Methods.........................................................................................................103 Plant Material and Sampling......................................................................................... 103 Quantification of mRNA s by Real-Time RT-PCR ....................................................... 103 Soluble Sugar Extraction and Quantification................................................................ 104 Assay for Soluble Invertase Activity............................................................................. 105 GC/MS Quantification of Metabolites.......................................................................... 105 Phloroglucinol-HCL Staining........................................................................................106 5 SUMMARY..........................................................................................................................119 APPENDIX PERTURBATION BY DROUGHT AND DISRUPTED VP1-BASED ABA SENSING.. 122 BIOGRAPHICAL SKETCH.......................................................................................................149

PAGE 7

7 LIST OF TABLES Table page 2-1 Summary statistics of a two-sa m ple multiplex 3'-UTR library......................................... 46 2-2 CesA family 3'-anchored consensus sequences................................................................. 47 2-3 Best-match cDNAs and associated annot ations (BLAST N) for consensus sequences showing highly significant differences in transcript abundance between wild-type and vp1 mutant drought-stressed ovary sub-libraries........................................................48 2-4 Polymorphisms detected by W22 3'-anchored 454 sequence reads.................................. 49 3-1 Number of genes associated with each co-expression cluster and percentage having significant differences in e xpression or large-fold ch anges during developm ent..............74 3-2 Annotated genes from clusters 1 and 5 that showed signifi cant differences in expression (FDR 20%) or large fold -change (>2) during developm ent............................ 75 3-3 Annotated genes from clus ter 7 that showed significan t differences in expression (FDR 20%) or large fold-c hanges during developm ent..................................................... 76

PAGE 8

8 LIST OF FIGURES Figure page 2-1 Comparison of shotgun versus 3'-anchored approaches for using 454-sequences as a platform for transcript profiling. ...................................................................................... 50 2-2 Strategy for 3'UTR profiling by 454-sequencing.............................................................. 51 2-3 Distribution of tag lengths for a simula ted Msp1 digest of 70,000 3' -oriented ESTs from B73 maize (maize full length cDNA project [www.maizecdna.org])...................... 52 2-4 A graphic presentation of the quant itative 3' UTR profile representing 11,559 consensus sequences that matched cDNAs in the 2-sample multiplexed library.............. 53 2-5 Distribution of transcripts (log-log scale) in selected functional classes from Figure 4B and read count for each. Resolution of individual gene fam ily members was enabled by the specificity of the 3'UTR............................................................................. 54 2-6 Validation of technical a ccuracy for determ ining differe nces in transcript abundance based on read number........................................................................................................55 2-7 A 3' sequence polymorphism resolved n early-identical Auxin Repressed Dor mancy Associated paralogs with di fferences in mRNA abundance.............................................. 56 3-1 Experimental design used to compare tr anscriptional profiles during m aize ovary development.................................................................................................................... ...77 3-2 A heat map generated by a two-way, hierar chical clustering of e xpression profiles for 856 genes that showed significant cha nges in transcript abundance during developm ent (FDR 10%)................................................................................................... 78 3-3 Annotated genes from hierach ical clusters 1 and 2 (Figur e 3-2) that showed either decreased or increased expression during m a ize ovary development were classified into functional categories based on Gene Ontologies........................................................ 79 3-4 Co-expression clusters of genes were analyzed after fitting ex pression patterns during development to orthogonal polynomials................................................................ 80 3-5 Annotated genes in clusters 1, 5, and 7 were grouped by functional processes using Gene Ontologies.................................................................................................................82 3-6 Experimental design and sequencing resu lts f or a 12-sample, multiplexed, 3'-UTR library........................................................................................................................ .........83 3-7 The 3'-UTR consensus tags were co mpared to the 70-m er microarray probe sequences to determine extent of overlap.......................................................................... 84

PAGE 9

9 3-8 Selected genes with high transcript abunda nces in the 3' -UTR library that were used to validate microarray expression data.............................................................................. 85 3-9 Quantitative 3'-UTR profiles for cell-wall-related genes with matches to array probes in clusters 1, 2, 5, and 7 (see Figure 3-4) ...........................................................................86 3-10 Transcript profiles of four Phenylal anine amm onia lyase (PAL) gene family members were compared by microarra yand sequence-based methods...........................87 3-11 Resolution of quantitative transcript profiles by the 3'-UTR sequencing approach for near-identical m RNAs that matched a single array probe................................................. 88 4-1 Definition of developmental stages fo r pre-po llinated maize female florets by physical growth parameters, anatomi cal features, and expression of a ZAG2 molecular marker............................................................................................................. 107 4-2 Partitioning of carbon and water to dry weight an d fresh weight accumulation among individual organs of devel oping maize female florets..................................................... 108 4-3 Spatial and temporal expression of sol uble acid invertases and changes in sugar com position in maize florets just prior to pollination......................................................109 4-4 Activity of the vacuolar i nvertase, IVR2, is associated with turgor-based expansion in rapidly elongating silks. ...............................................................................................110 4-5 Diurnal changes in sucrose and hexose levels in rapidly expanding silks on the day of their em ergence from husks......................................................................................... 111 4-6 Relative abundance of mRNAs for vacuolar and cell-wall inverase isoforms in m aize female florets during the pretopost-pollination period................................................112 4-7 Carbon deposition and relative water conten t in maize ovaries and pedicels during post-pollination development...........................................................................................113 4-8 Sugar composition of maize ovaries a nd pedicels during ea rly post-pollination developm ent.................................................................................................................... .114 4-9 Co-expression profiles for genes relate d to C and N m etabolism during ovary and pedicel development from genome-wide microarray analyses........................................ 115 4-10 Co-expression of genes related to lipid m etabolism and abundance of lipid-based metabolites in develo ping maize ovaries......................................................................... 116 4-11 Expression profiles for genes involved in phenylpropanoid biosynthesis in the developing m aize ovary-plus-pedicel..............................................................................117 4-12 Accumulation of lignin precursors and pe dicel-localized staining of lignin in the developing m aize ovary...................................................................................................118

PAGE 10

10 LIST OF ABBREVIATIONS UTR Untranslated region PCR Polymerase chain reaction EST Expressed sequence tag SAGE Serial analysis of gene expression MPSS Massively parallel signature sequencing SNP Single nucleotide polymorphism QTL Quantitative trait loci MAGI Maize Assembled Genomic Islands DBP Days before pollination DAP Days after pollination ASI Anthesis-Silking Interval ANOVA Analysis of Variance FDR False Discovery Rate BIC Bayesian Information Criterion LSMEANS Least square means PTT Parts per ten thousand C Carbon N Nitrogen

PAGE 11

11 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy REGULATION OF SINK STRENGTH IN DEVELOPING MAIZE FLORETS: IMPLICATIONS FOR SEED SET AND GRAIN YIELD By Andrea L. Eveland May 2008 Chair: Karen E. Koch Major: Plant Molecular and Cellular Biology The preand early post-pollination phases of maize ( Zea mays L.) reproductive development are critical for seed set and subseque nt grain yield. During this time, the plant is especially sensitive to abiotic st resses, such as drought, which can reduce pollination efficiency or lead to kernel abortion. A key determin ant of reproductive succe ss is carbohydrate allocation and use in the developing female inflores cence, which is often disrupted by stress. Understanding the mechanisms that underlie regulation of sink strength during normal progression of maize floral development is t hus central to improving seed set under adverse environmental conditions. In this work, we tested the hypot hesis that expression of genes related to specific metabolic or regulatory pathways would change in asso ciation with carbohydrate al location and use during silk exsertion and pollination in ma ize. Individual stages of preand early post-pollination maize female florets were characterized based on physical characteristics and ex pression of a molecular marker for development. Subsequent analyses revealed a shift in sink strength, as approximated by dry and fresh weights, during the pollination period from rapidly expanding silks and subtending floral structures (lemma, palea, and glumes) to the developing ovary and pedicel. This shift coincided with isof orm-specific expression of sucros e metabolizing invertases, which

PAGE 12

12 provide hexose substrates essential for turgor-b ased expansion prior to pollination and also for post-pollination growth of symplastically-isolate d filial tissues. In addition, accumulation of sucrose and hexoses in the pedicel and ovary, respectively, indicated th at invertases could contribute to spatial regulati on of cell expansion and differe ntiation during development. We used a genome-wide transcript profiling approach to determine whether co-expressed genes were related to specific functional processes or associated with relevant metabolic pathways during preand early post-pollination ovary development. A gene-specific, sequencebased, 3'-UTR profiling strategy was developed and tested in parallel to microarray analyses. We resolved co-expression profiles for key genes related to nitrogen and amino acid metabolism, carbohydrate metabolism, lignin biosynthesis, a nd cell growth during ovary development. Transcript profiles were combined with sugar and metabolite analyses and fresh/dry weight quantifications to further support relevance of key sets of co-expressed genes during ovary sink establishment in maize. Results from this study provide evidence for testable roles of such genes in kernel set.

PAGE 13

13 CHAPTER 1 LITERATURE REVIEW Introduction In plants, maintaining a delic ate balance between photosynthe sis, the energy dem ands of growing sink organs, and storage of essential reserves is central to whole plant function and developmental progression. Reproduction is a costly process requi ring large amounts of photosynthetic resources for flor al organ expansion (Woodson and Wang, 1987; Makela et al., 2005; Borras et al., 2007), respiration (Bustan and Goldschmidt 1998), and/or grain filling (CruzAguado et al., 1999; Maitz et al., 2000; Weschke et al., 2003). Therefore, careful regulation of carbon assimilation and allocation is necessary for timing and maintenance of reproductive growth. Sucrose import into developing sink stru ctures involves sensing of carbohydrate status in both source and sink cells (Koch, 1996; Chiou and Bush, 1998; Stitt et al., 2007) as well as intercellular signaling based on steady-state pools and flux of sugars through membrane transporters (Lalonde et al., 1999; Barker et al., 2000; Vaughn et al., 2002; Sauer, 2007). Such signals can influence sucrose export from source or gans or regulate sink stre ngth at the site of phloem unloading. Perturbation at either end alters the source/sink balance (Paul and Foyer, 2001; Borras et al., 2003; McCormick et al., 2008). Adjustment of source-to-sink resource allocatio n has a substantial impact on fruit set and grain yield and thus is centrally important to crop improvement programs. In maize (Zea mays L.), the preand early post-pollination phases of female reproduc tive development are critical for sufficient pollination and seed set (Zinselmei er et al., 1995; McLaughlin and Boyer 2004a). During this time, female reproductive tissues are most sensitive to abiotic stresses such as drought (Westgate and Boyer, 1985; Andersen et al., 2002; McLaughlin and Boyer 2004b) and shade (Setter et al., 2001; Zinselmeier et al ., 2002). Such stresses negatively affect

PAGE 14

14 photosynthesis and thus alter whole plant carbohydrate status (Roits ch, 1999; Trouverie and Prioul, 2006). Resulting losses in grain yield ar e devastating on a world-wi de scale (Barnabas et al., 2008), yet the processes underlying the preand early post-pollination phases of maize reproductive development have received comparativel y little attention at the molecular/metabolic level. Source/Sink Relations and Sucrose Metabolizing Enz ymes Sucrose is synthesized in source-leaf mesophyll cells and in most plant systems transported through the phloem to developing si nk organs by mass flow. This is a function of a turgor pressure gradient due to sucrose concentrati on at both sourceand sink-ends of the phloem (Lalonde et al., 2004; Carpaneto et al., 2005). An organs sin k strength is dependent upon a number of factors related to th is turgor gradient, including lo calized osmotic status and the activity of sucrose-cleaving enzymes (Sturm a nd Tang, 1999; Roitsch et al., 2000). Recent work has shown that specific quantit ative trait loci (QTL) for grow th in maize are associated predominantly with carbohydrate metabolic enzymes involved in source/sink relations rather than photosynthetic genes (Pelleschi et al., 2006). Invertases, which catalyze the irreversible cleavage of disaccharide sucrose into glucose and fructose, are implicated as major determinants of sink strength in the early development of importing structures (Koch, 1996, 2004; Strum 1999; Roitsch, 1999; Weschke et al., 2003; McLaughlin and Boyer, 2004b; Schaarschmidt et al., 2007). The conversion of one molecule of sucros e to two hexoses regula tes osmotically-active solutes in sink tissues and can aff ect the descending turgor gradient for sucrose in the phloem. Carbon assimilates can move into a developing sink either sy mplastically through interconnecting plasmodesmata or via apoplastic phloem unloading into the cell wall space. Invertase is active at both of these interf aces with isoform-specific compartmentalization (Tymowska-Lalanne and Kreis, 1998; Sturm, 1999; Godt and Roitsch, 2006). Inside the cell,

PAGE 15

15 there are two types of soluble i nvertases that can cleave symplast ically-delivered sucrose. Of these, the acidic vacuolar inve rtases are typically most active. However, recent findings support specific roles for the elusive cytoplasmic, alkalin e invertases (Flemetakis et al., 2006; Lou et al., 2007; Vargas et al, 2007). Apoplast-localized cell wall invertases, also optim ally active at acidic pH, catalyze the cleavage of sucrose unloade d from the phloem into hexoses which are transported into symplastically-isolated sinks such as the developing embryonic tissues (Cheng et al., 1996; Sherson et al., 2003). Invertases are encoded by small gene families, members of which are spatially and temporally regulated (Tymowska-Lalanne and Kr eis, 1998; Fridman and Zamir, 2003; Cho et al., 2005; Huang, 2006). Only two vacuolar isofor ms have been described in Arabidopis (Tymowska-Lalanne and Kreis, 1998), tomato (Fridman and Zamir, 2003), poplar (Bocock et al., 2008), and maize (Xu et al., 1996), suggesting evolutionary conser vation among diverse species. Vacuolar and cell wall-bound invert ases are implicated not only in the establishment of sink strength in developing organs, but also in the generation of hexose-base d signals, providing cues for developmental and metabolic processe s (Koch, 1996, 2004; Sturm and Tang, 1999; Rolland et al., 2006). Hexose-based signalin g is discussed further in the sugar sensing/signaling section of this review. Sink development is mediated by the tempor al expression patterns of sucrose-cleaving enzymes. During initial phases of sink establishment, vacuolar invertases generate essential hexoses for turgor-driven cell expansion (Sturm and Tang, 1999; Koch, 2004), whereas apoplastic invertas e activity becomes the predominant s ource of hexose-ba sed signals during expansion through the cell different iation phase. The reversible activity of sucrose synthase, which cleaves sucrose into fructose and UDP-g lucose, becomes important during the maturation

PAGE 16

16 phases of sink development. In general, a high level of hexoses relative to sucrose induces cell division and expansion, whereas sucrose accumulation is associ ated with differentiation and maturation (Zinselmeier et al., 1995; Patrick, 199 7; Wobus and Weber, 1999; Winter and Huber, 2000). Consequently, acid invertases, and in partic ular vacuolar invertases have been identified as central to turgor-mediated growth in expans ion sinks such as floral organs (Woodson and Wang, 1987; Xu et al., 1996) and root s (Sergeeva et al., 2006; Huang, 2006). The essential contribution of sucrose metabolizing enzymes to sink establishment and growth has been described for reproductive de velopment and is discussed in the following section. Similar modes of action for such enzymes have been revealed in plant-microorganism interactions. Symbiotic relations hips such as mycorrhizal arbusc ules and root nodules (Blee and Anderson, 2002) as well as parasitic associations of nematode induced syncytia (Hofmann et al., 2007) and Agrobacterium infection (Wachter et al, 2003), induce carbohydrate demands that mimic reproductive structures. Whole plant source/sink adjustments result accordingly. Sugar Sensing and Signaling Central to photosynthate allo cation is nutrient status and the m echanisms by which cells sense this status. Sucrose, and to a larger extent the hexoses generated through its cleavage, can act as signaling molecules in a variety of developmental processes (Koch, 1996, 2004; Smeekens, 2000; Rolland et al., 2006). Therefore, sugar status can modulate the expression of specific genes based on photosyntha te availability and/or environmental cues. Sucrose and hexose signals can reportedly be generated ei ther by concentration or by flux through sugarspecific sensors and/or transpor ters. Long-distance su crose transport depends on temporal and spatial regulation of a sucrose transporter (SUT) gene family (Sauer, 2007), members of which have been identified as pivotal to long-distance tr ansport in cereal crops such as maize (Aoki et al., 1999) and rice (Scofield et al ., 2007). Certain SUT isoforms have been shown to regulate

PAGE 17

17 phloem loading and unloading in response to change s in sucrose concentra tions (Lalonde et al., 1999, 2004; Barker et al., 2000; Vaughn et al., 2002). In yeast, membrane proteins with homology to hexose transporters have been found to sense glucose and relay the signal to hexose-resp onsive genes (Ozcan et al., 1998; Rolland et al., 2002). In plants, monosaccharide/H+ symporters catalyze the uptake of hexoses from the apoplast (Sherson et al, 2003; Weschke et al., 20 03; Buttner, 2007) or into the vacuole (Wormit et al., 2006) and may be involved in the transd uction of hexose-specific cellular signals. However, an actual sensor function fo r these transporters in plants has not been reported to date. Hexoses may be sensed as substrates fo r hexokinase, where the phosphorylation event leads to a signal cascade and subs equent changes in gene expression. The hexokinase-dependent mechanism couples signaling with hexose-phosph ate flux through glycolysis (Harrington and Bush, 2003; Moore et al, 2003). Recent work has identified a set of interacting proteins that form a complex with hexokinase in the nucle us (Cho et al., 2006), a nd the association of hexokinase with actin filaments (Balasubramanian et al., 2007), however the mechanisms for signaling are still unknown. Alternately, sensin g via a hexokinase-independent pathway is possible (Xiao et al., 2000). Recent work has de scribed the role of certain G-protein coupled receptors in direct sensing and transduction of th e sugar signal in yeast (Lemaire et al., 2004) and in plants (Huang et al., 2006). In addition, sucrose has also been shown to induce phosphorylation (Nittylae et al., 2007) and thus stimulate signal cascades. These sensing mechanisms enable intercellu lar signaling that can modulat e changes in gene expression according to the amount of photosynthate available. Sucrose-cleaving enzymes are among such suga r-regulated gene products. In maize, both invertases and sucrose synthases show isoform-specific induction (Xu et al., 1996) in a

PAGE 18

18 feast/famine association. The tw o vacuolar invertase isoforms, IVR1 and IVR2, are upregulated in a reciprocal manne r by sucrose availability and su crose starvation, respectively. Similarly, sucrose synthase isoforms also disp lay concentration-dependent induction by sugars (Xu et al., 1996). This reciprocal mode of regul ation has been identified in pairs of vacuolar invertase isoforms in Arabidops is (Huang, 2006), poplar (Bocock et al., 2008), tomato (Godt and Roitsch, 1997) and rice (Huang, 2006). Whether this dual sub-functionalizati on is evolutionarily conserved or an example of convergent evol ution is currently unknown. However, such differential responses to sugar availability or st arvation based on isoform sp ecificity enable fine adjustment of photosynthate in response to en vironmental perturbation or developmental cues (Osuna et al., 2007; Smith and Stitt, 2007). Signals based on carbohydrate status can regulate plant growth in response to environmental or developmental stimuli. Suga r availability can in fluence cell division, expansion, or proliferation (Nicolai et al., 2006; Rook et al., 2006; Kojima et al., 2007; Smith and Stitt, 2007). Glucose and/or sucrose can induce specific D cy clins, thus coupling cell cycle regulation to carbohydrate status (Riou-Khamlic hi et al., 2000; Menges et al., 2006). Recent work has also linked sugar sensing and signaling to regulation of cell wall expansion (Lee et al., 2007; Li et al, 2007). Carbohydrate availability can also regulat e gene expression based on its relative abundance relative to n itrogen (Coruzzi and Bush, 2001; Cooke et al., 2003; Fritz et al., 2006) or to phosphate (Karthikeyan et al., 2007). Sugar signals based on starch metabolism and sugar starvation also affect ge ne expression, aspects of whic h have been supported by global assessment of diurnal regulation (Gibon et al., 2 004; Blasing et al., 2005; Smith and Stitt, 2007). In addition, evidence for signaling by sugars such as trehalose has revealed developmental and regulatory roles. Trehalose metabolism has b een shown to influence embryonic development

PAGE 19

19 (Eastmond, 2002), reproductive arch itecture (Satoh-Nagasawa et al., 2006), and ABA-mediated responses (Avonce et al., 2004; Ramon et al., 2007), possibly through sensing and signaling of trehalose itself or of its phosphorylated form. Hormone and Sugar Signaling Networks The regulation of gene expression by sugars can be viewed as analogous to phytohormonebased interactions. Therefore, it is n ot surprising that extens ive crosstalk between sugar and hormonal pathways has been described for many developmental and metabolic processes (Gazzarrini and McCourt, 2001; Leon and Sheen, 2003; Gibson, 2004; Rolland et al., 2006). The biosynthesis of phytohormones critical to ce rtain developmental cues can be dependent on carbohydrate status (Cheng et al., 2002). Alternatively, differential responses of hormonemodulated genes may rely on sugar availability (Huang, 2006) and can thus be influenced by whole plant source/sink relations. Studies in Arabidopsis have revealed an extensive interface between ABA and sugar regulatory networks (Fin kelstein and Gibson, 2002; Li et al., 2006). Most of these interactions were identified by mu tant loci redundant for sugarand ABAsensing and signaling phenotypes (Arenas-Hu ertero et al., 2000; Huijser et al., 2000; La by et al., 2000; Brocard et al., 2002; Arroyo et al ., 2003; Brocard-Gifford et al., 2004). In addition, examples of converging sugarand ABA-based signals in the regulation of specific tran scription factors have been described in maize (Niu et al., 2002) and in grape (Cakir et al., 2003). Much of the work to date in s ugar sensing and signaling has used Arabidopsis thaliana as a model system due to the ease of phenotypic screen ing on sugars or other growth regulators. Although these studies have focused primarily on seed germination, some sugarand hormonebased interactions have been described in whole plant source/sink relations In particular, the ABA has been implicated in regulation of leaf senescence (Pourtau et al., 2004), modulation of

PAGE 20

20 phloem unloading ((Peng et al., 2003; Oliver et al., 2007), and transitory starch biosynthesis (Rook et al., 2001). Regulation of invertase-mediated carbon alloca tion includes ABA-based signals as well as those of various other phytohormones. Studies ha ve shown that acid inve rtase responds to ABA (Setter et al., 2001; Kim et al ., 2000a; Pan et al., 2005; Huang, 2006), ethylene (Linden et al., 1996; Huang, 2006), cytokinins (Set ter et al., 2001; Lara et al ., 2004), and auxin (Yun et al., 2002; Long et al., 2002). Since sucrose meta bolism and use are essential aspects of developmental growth, it is not surprising that multiple, developmentally-induced hormone signals are involved. Invertas e activities may also be mediated by hormone ratios and antagonisms. For example, ethylene and ABA are antagonistic, yet a high level of ABA tends to promote ethylene production (Gazzarrini a nd McCourt, 2001; Leon and Sheen 2003). Hexose-based signals generated by invertase ac tivity can affect the extent of hormonal responses either through biosynt hesis or signal transduction. Gl ucose has been found to promote ABA biosynthesis via up-regulation of genes encoding key enzymes in this process (Cheng et al., 2002). This suggests potential feed back regulation of invertase e xpression that could operate on ABA levels. In Arabidopsis, sucrose-based induction of AtvacINV1 the putative functional ortholog of ZmIVR2 is enhanced by ABA (Huang, 2006). Alternatively, AtvacINV1 transcriptional induction by ABA is dampened by sucrose. In addition, isoform-specific regulation of invertases extend beyond sugar and hormonal signals (Roitsch and Gonzalez, 2004; Huang et al., 2007). Mechanisms of regulati on include mRNA stabil ity (Yun et al., 2002), turgor-based sensing via cell wall associated kinases (Kohorn et al., 2006), and sequestration in pre-vacuolar secretory vesicles (Roj o et al., 2003; Huang et al., 2007).

PAGE 21

21 Maize Reproductive Development The activities of sucrose-cleav ing enzym es during reproductive development are central to promoting optimal pollination and subsequent seed set in cereal crops such as maize. Both vacuolar (Xu et al., 1996; Andersen et al., 2002; McLaughlin and Boye r 2004b; Qin et al., 2004) and cell wall invertases (Cheng et al., 1996; Taliercio et al., 1 999; Kim et al., 2000b; Chourey et al., 2006) have been implicated in early mai ze floral development, and especially ovary expansion. The developing maize ovary is com posed predominantly of maternal tissue and receives sucrose primarily via symplastic transf er from the phloem ending in the pedicel region (Fisher and Wang, 1999). During the preand early post-pollination phases of reproductive development, insoluble invertase is localized to the area beneath the nu cellus in the upper pedice l tissues (Cheng et al., 1996; Kladnik et al., 2004; McLaughlin and B oyer 2004b). Soluble invertase is expressed throughout the nucellus and pericarp tissues (Andersen et al., 2002; McLaughlin and Boyer 2004b). Upon embryonic growth, a cell wall invertase, INCW2, becomes the dominant isoform, providing sucrose across the apoplastic barrier between maternal tissu e and the developing embryo (Cheng et al., 1996; Kladnik et al., 2004). Together, activities of specific invertase isoforms are responsible for maintaining an eff ective sucrose gradient to promote ovary growth and seed set. Accordingly, a mutation in the INCW2 gene resulted in a small kernel, or miniature, phenotype (Cheng et al., 1996) and drought-indu ced repression of vacuolar invertases, IVR1 and IVR2 caused inefficient pollination and/or kernel abortion (Andersen et al., 2002; Boyer and Westgate, 2004; McLaughlin and Boyer, 2004b). Maize is a monoecious plant wi th male tassels borne terminally from the differentiated shoot apical meristem and female inflorescences, or ears, laterally in the axils of leaves. In the female florets, the ovary, composed of three fused carpels, is subtende d by a palea and lemma

PAGE 22

22 (outer whorl structures with similarities to pr ophylls and bracts, respectively) and an additional set of bracts called glumes (Clifford, 1987; Ambrose et al., 2000; Whipple et al., 2004). An inner whorl of petal-like structures or lodicules, is present only in the male florets. Outgrowths from two of the three fused carpe l units form the silk, a rapidl y elongating stigmatic structure which is unique to the maize flower. Silk expans ion is critical to pollination efficiency and is directly correlated with bioma ss accumulation (Borras et al., 2007). Previous work has identified soluble invertase (Xu et al ., 1996) and expansin (D Cosgrove, personal communication) activities in growing silk s and suggests their roles in turgo r-mediated expansion. However, few studies have investigated th e regulation of carbohydrate allocation and use in silks. A central aspect of the pollination process and ovary receptivity depends on coordination of silk emergence from husk leaves with anther dehiscence, also known as the Anthesis Silking Interval (ASI) (Bolanos and Edmeades, 1996; Borra s et al., 2007). Initia l studies by Westgate and Boyer (1985) showed that low water potential due to soil drying had the most dramatic effect on silk turgor and growth as compared to turgor in leaf, stem, and root. These findings indicate that osmotic adjustment in the silk, based on whole plant source/sink balance, has a significant impact on reproductive success in maize.

PAGE 23

23 CHAPTER 2 TRANSCRIPT PROFILING BY 3'-UTR SEQUENCING RESOLVES EXPRESSION OF GENE F AMILIES Introduction Functional analysis of plant genom es requires methods for resolving differential expression of closely-related genes. The ability to distinguish between paralogs (e.g. gene family members) and alleles on a genome-wide scal e is key to understanding the genetic basis of quantitative traits in diverse pl ant populations. Gene s with extensive sequence similarity may comprise a significant portion of a given transc riptome. Among maize inbreds, for example, 90% of the alleles are polymorphic (Wright et al ., 2005) and approximately one-third of maize genes are tandemly duplicated (Messing et al., 2004). The extent of these sequence similarities in maize and other complex genomes pose a clear challenge to delinea tion of gene-specific function. Differential expression of related, duplicated genes has been linked to functional diversity within species (Gu et al., 2004), and sub-func tionalization can provide a basis for genome evolution (Moore and Purugganan, 2005; Emrich et al., 2007b). The impetus for resolving expression among paralogs is further motivated by the extent of polyploidy, which is estimated to affect 50-80% of angiosperm species, incl uding maize, wheat, cott on, and other important crops (Osborn et al., 2003; Blanc and Wolfe, 200 4). Moreover, allele-specific differences in gene expression that contribute to variations in phenotype are wide spread in both animal (Cowles et al., 2002; Yan et al., 2002) and plan t species (Cong et al., 20 02; Guo et al., 2004). Accordingly, transcript profiling has been adapte d as a means of appraising quantitative traits (Schadt et al., 2003; Borevitz and Chory, 2004). Association mapping using eQTLs has identified candidate genes for important traits in tomato (Baxter et al., 200 5) and poplar (Street et al., 2006). In addition, comparing allele-specific expression among i nbred parental varieties and

PAGE 24

24 F1 hybrids of reciprocal crosse s has revealed deviations from dosage dependency and enabled analysis of imprinting in maize endosperm (Guo et al., 2003). The ability to resolve individual transcripts with similar sequences and quantitatively compare their expression is thus central to addressing questions in functiona l genomics and defining genetic contributions to hybrid vigor (Birchler et al., 2003; Swanson-Wagner et al., 2006; Springer and Stupar, 2007). Despite rapid advances in expression prof iling techniques, capacity to distinguish among closely-related transcripts on a genome-wide scale remains a challenge. In microarray analyses, cross-hybridization of similar transcripts to a given oli gonucleotide probe may confound expression of individual genes. With sequencing of several genomes complete (e.g. Arabidopsis, rice, and poplar), whole-genome tiling arrays have allowed unbiased interrogation of the transcriptome (Yamada et al., 2003; Mock ler and Ecker, 2004). These platforms have successfully uncovered discrete po lymorphisms and alternate splice sites, but depend on fullysequenced genomes and/or are limited to sequenc e variants present on the array (typically derived from a single reference strain). In ad dition, quantitative measures of gene expression are limited by probe-specific hybridization efficiencies. The emergence of high volume, short-read sequencing technologies has increased resolution for quantitative transcriptome analysis in organisms for which complete genomic sequence is available. Advances in serial analysis of gene expression (SAGE) have opened transcript profiling to unbiased sampling and quantitative analysis of gene expression (Saha et al., 2002; Bao et al., 2005). Although limited in throughput, the sequencing of novel cDNAs following 3' extension and amplifi cation of short SAGE tags has b een successfully utilized for gene discovery (Chen et al., 2004).

PAGE 25

25 Alternatively, genome-wide profiling by massive ly parallel sequencing, MPSS (Meyers et al., 2004; Jongeneel et al., 2005) and the more re cent Solexa 1-G technology (Barski et al., 2007), has facilitated detection of rare transcripts and compre hensive cataloging of non-coding RNAs (Lu et al., 2005; Nobuta et al., 2007). The cap acity of these massively parallel approaches to generate millions of short sequence tags can enable reliable, cost-effective coverage of the transcriptome. However, in genomes where sequence information is fragmentary, the short length of these reads (17 to 36b) provides a lim ited capability for unambiguous gene assignment. Likewise, near-identical transcri pts are difficult to discern with short-sequence reads, even in a fully-sequenced genome. Estimates from rice and Arabidopsis MPSS libra ries indicate that approximately 11% of signature sequences matched multiple target sites in the genome (Nobuta et al., 2007). Longer-read lengths are achieved with 454-base d pyrosequencing, initially described by Marguiles et al. (2005), and more recently impl emented as a platform for transcript profiling (Emrich et al., 2007a; Weber et al., 2007). A key advantage of the longer reads generated by this technology is greater capability for gene annotation and discov ery in both sequenced and nonsequenced genomes. A recent upgrade of the Genome Sequencer FLX system (Harkins and Jarvie, 2007) has extended averag e read lengths to >200 bases. A tradeoff for obtaining more informative read lengths is a lower depth of sequencing achieved with 454 compared to shortread technologies (e.g. Solexa 1-G). Therefore, one method for improving the efficiency of 454based transcript profiling is to anchor 454-reads to unique sites near the 3' ends of expressed sequences in order to 1) reduce the number of reads necessary to identify individual mRNAs and 2) maximize recovery of gene-specific polymorphism s. The 3'-untranslated region (UTR) is rich in single-feature polymorphisms that distinguish closely-related transcript s (Bhattramakki et al.,

PAGE 26

26 2002; Vroh Bi et al., 2006). The specificity of 3'UTR sequence reads thus allows effective annotation of individual mRNAs without asse mbly of complete cDNAs (Figure 2-1). Here we present a strategy that harnesses th e specificity and information content of the 3'UTR in a long-read, 454-based sequencing appro ach to transcript profiling. A key to this method is the use of 3'-anchored sequence read s long enough for unambiguous identification of closely-related transcripts. By targeting th e 3'UTR of mRNAs, an unpr ecedented resolution is achieved for geneand allele-specific transcripts, even for genomes that are only partiallysequenced or lack extensive expressed sequence tag (EST) coverage. In addition, detection of haplotypes containing multiple polymorphisms is facilitated by the longe r-read length. These components of the transcriptome are thus opened to quantitative analysis beyond that currently accessible with short-tag sequencing technologies. In the present work, maize provides an ideal system to assess our 3'-anchored strategy, because the genome is rich in genetic complexity from extensive gene duplication (Mes sing et al., 2004; Messing and Dooner, 2006) and currently is not fully sequenced. In this study we introduce a 3'UTR profiling me thod that allows quantitative analysis of gene-specific expression on a genome-wide leve l, here using mutant and wild-type maize ovaries. Concurrent sequencing of multiple mRNA samples was enabled by use of a multiplexing strategy. Results provided quantitativ e expression profiles with read output evenly distributed between samples. The frequency of 3'-anchored sequence reads aligning to a given cDNA was used to quantify mRNA abundance and to measure differential gene expression. The long read lengths, combined with the specificity of the 3'UTR, we re sufficient to distinguish individual members of a previ ously-characterized gene family as well as provide quantitative comparisons of closely-related transcripts that matched unique maize expressed sequence tags

PAGE 27

27 (ESTs) or assembled cDNAs. In addition, insert ion deletion (indel) polymorphisms were readily detectable by this method and resolved n early-identical paralogous gene products. Results Construction of a 3'-cDNA Library We synthesized 3'-anchored cDNA template li braries to generate gene-specific sequence reads by 454 using the protocol show n in Figure 2-2. Concurrent sequencing of up to 16 individual sub-libraries is enab led by incorporation of a thre e-base multiplex key in the Aadaptor (Figure 2-2A). By usi ng a subset of 16 three-base keys, we could detect single-base errors in the multiplex key. Addition of a fourth base to the multiplex key would enable up to 64 unique combinations with error detection, thus enhancing the number of concurrently sequenced sub-libraries. Each 3'UTR sub-library was cons tructed from total RNA (Figure 2-2B) using a modified, biotinylated 454-B adaptor that in corporates an oligo(dT ) tail for priming cDNA synthesis from poly(A) RNA. Following sec ond-strand cDNA synthesis, biotinylated cDNAs were bound to steptavidin beads, purified by magnetic pull down, and digested with MspI to generate 3'-cDNA fragments with 2-base (CG) overhangs. Specific multiplex A-adaptors were then ligated to the purified 3' fragments. A detailed description is provided in Materials and Methods. MspI was selected based on simulated digest s of 70,000 3'-orientated ESTs of maize (Figure 2-3) from the maize full-len gth cDNA project (www. maizecdna.org). Predicted tag lengths were used to assess the proportion of 3'-enrichment in comparison to rice 3'UTRs. While the expect ed size distribution of MspI -digested cDNA fragments is optimal for the GS20 read length (~100 b), the longer re ads generated by the 454-FLX instrument (~250 b average) would likely extend through the 3'UTR of many transcripts. If so, the number of FLX reaction cycles could be configured to optimi ze average read length (Harkins and Jarvie, 2007) Although a single MspI digest was used in this study, potenti al increases in coverage of 3'-ends

PAGE 28

28 could be achieved by combining digests made with compatible restriction enzymes (e.g. TaqI MaeII ). This would further improve coverage when used in conjunction with the longer read lengths and enhanced read out put achieved by 454-FLX technology. To test the 3'UTR profiling strategy, we sequenced 3'-a nchored cDNA sub-libraries prepared from immature ovaries of isogenic viviparous-1 mutant and wild-type maize plants in a W22 inbred genetic background. Prior to RNA sampling, plants were subjected to a drought stress treatment (Materials and Methods). VP1 is a transcription factor that mediates a subset of responses to the plant stress hor mone, abscisic acid (ABA), in cluding maturation and onset of seed desiccation tolerance. The classic vp1 phenotype is that of pr ecocious germination due to reduced ABA sensitivity (McCarty et al., 1991) More recently, however, VP1 has been implicated in stress responses of non-seed tissu es (Cao et al., 2007). In addition, preliminary evidence suggests that VP1 may be involved in modulating fema le reproductive quiescence in maize under drought stress (Eveland, unpublished). To resolve differences in expression profiles that would help define roles of the VP1 ge ne, cDNA sub-libraries, each tagged with a unique multiplex key, were prepared from wild-type and vp1 -mutant maize ovaries. Data Assembly and Analysis A sequencing reaction on the Genom e Sequencer 20 instrument (Marguiles et al, 2005) yielded 228,595 high-quality-r eads with an average, trimmed lengt h of 95 bases. Of these, 93% were identified as correctly oriented, 3'-anchored cDNAs with e qual representation of wild-type and mutant sub-libraries (Table 21). The 7% of reads that were excluded from further analysis contained errors in the multiplex key (1.5%) or in valid ligation junctions (5.5%). Assembly of validated, trimmed, high-quality-reads using CAP3 (Huang and Madan, 1999) revealed 14,822 non-redundant 3'-anchored consensus sequences, each represented by 2 to 2,500 reads, and 32,477 singlets.

PAGE 29

29 The capacity of these consensus sequences to identify individual genes based on specificity of their 3'UTRs was tested by aligning these reads to available maize cDNA databases (TIGR Zea mays Gene Index [ZmGI] and Industry UniGen e [IUC]) using BLASTN (cutoff: E < 10-7). At least 87% of the consensus tags matched cDNAs and 66% aligned with a gene-enriched maize genomic assembly (MAGI) (Fu et al., 2005). In addition, BLASTN searches of the TIGR maize repeat database (http://maize.tigr.org/ repeat_db.shtml) indicated that only 1.9% of 3'anchored consensus sequences contained retrotransposons or other repetitive sequences whereas another 1.2% were identified as organellar or cytosolic rRNA contaminants. The latter were most likely due to rare mispriming by the oligo(dT)-B adaptor during cDNA synthesis. Analysis of 3'-UTR Profile Reveal s a Dynamic Range of Expression Based on the set of unique consensus sequences obtained from the two-sam ple library, we developed a graphic display for the quantitative transcriptome profile (Fig ure 2-4A and 2-4B). We quantified gene expression for each of 11,559 consensus sequences that matched unique cDNAs using read frequencies. The results are plotted on a logarithmic scale to capture the full range of expression. The 11,559 3'-sequences pr ofiled were also analyzed based on Gene Ontology functional classifications determined by PFam searches derived from ZmGI and IUC databases. Analysis of respective maize cDNAs revealed 5,202 (45%) that were unclassified and lacked annotation based on sequence similarity. An additional 578 (5%) of consensus sequences matched genes having conserved domains of unknown function. The relationship between abundance of each mR NA and its rank (ordered from leastto mostprevalent) in the whole dataset approximate d a Zipf-power law (ranked slope near -1 on a log-log scale). This distribution was evident among transcripts overall (F igure 2-4), and within individual functional classes (Fi gure 2-4B, 2-5A). Zipfs power law relationships are observed in a wide range of natural phenomena including th e distribution of gene expression in a variety

PAGE 30

30 of organisms (Kuznetsov et al., 2002; Furusa wa and Kaneko, 2003). Accordingly, it has been used as a tool for normalization in some SAGE and microarray analyses. Although our results were consistent with this distri bution, an interesting exception is shown for the chromatin-related functional class. As shown in Figures 2-4B a nd 2-5B, distribution of e xpression was skewed for this group of mRNAs by a disproportionate number of highly abundant transcripts. Distinguishing Gene Family Members In order to e valuate 3'UTR profiling for resolution of individual gene family members, we analyzed the Cellulose Synthase ( CesA ) gene family (Figure 2-5A). The assembled 3'-anchored sequences distinguished 12 unique tran scripts representing nine annotated CesA gene family members (Table 2-2) that were previously ch aracterized in maize (Holland et al., 2000; Appenzeller et al., 2004). The full-length CesA cDNAs (ZmGI) share up to 94% sequence identity. In some cases, extens ive sequence similarity between CesA genes and their proximal mapped locations to each other in the genome are suggestive of paired duplications (e.g. CesA1 and CesA2 on chromosomes 6 and 8 and CesA4 and CesA9 on chromosome 7). The cDNA sequences for CesA4 and CesA9 differ almost exclusively in their 3'UTRs, thus complicating resolution of these two genes in previous expr ession studies (Holland et al., 2000). Here, the corresponding 3'-anchored 454 reads for these closel y-related gene family members aligned with gene-specific regions in the 3'UTR (www .plantphysiol.org/cgi/d ata/pp.107.108597/DC1/1). Polymorphic variants for CesA4 and CesA6 were also identified. Ali gnments of consensus tags to a CesA4 cDNA (TC287832) indicated that a novel transcript variant ( CesA4c ) contained an MspI restriction site polymorphism as well as 35 bp of an unspliced intron (Table 2-2). Although no other ESTs having these features were detected in ma ize databases, nine reads in our maize-ovary dataset aligned with the CesA4c variant. Consistent with the possibility that these sequences identify a second CesA4 gene, CesA4 has been mapped to two locations (2.06

PAGE 31

31 and 7.01; Holland et al., 2000) corresponding to duplicated chromosome segments (Helentjaris et al., 1988). In addition, we analyzed a group of closely-related Histone1 ( H1 )-like transcripts (Figure 2-5B). These transcripts matched a unique, non-redundant set of ESTs from various maize cDNA libraries and were annotated based on sequ ence similarities in other species. Although these H1 genes have not been individually characte rized in maize, BLASTN results provided insight for eventual functi onal analysis. For example, a very highly expressed H1 -like transcript (TC292133a) matched a droughtand ABA-induced gene that had been characterized in tomato (Bray et al., 1999). These results indicate that unbiased profiling of closely-related transcripts can facilitate studies of functi onal genomics with or without a fully annotated EST dataset or a completely sequenced genome. Evaluation of Differential Expressi on betw een Multiplexed Sub-libraries The use of 3'UTR profiling as an effective stra tegy for detecting quantitative differences in transcript abundance between samples was evaluate d based on read frequencies generated from individual sub-libraries. Read frequencies representing each expressed gene were determined for wild-type and mutant sub-libraries by parsi ng the CAP3 ace file output. We analyzed 4,147 consensus sequences that were represented by a total of 10 or more reads using a chi-square statistic. Of these, 202 showed significant diffe rences (p < 0.0015) in fr equency between the two samples, indicating putative differences in tran script levels. A subs et of these consensus sequences with highly significant differences between libraries was annotated by BLASTN to identify best-match ESTs (Table 2-3). Of th e 30 sequences listed, thr ee matched to unannotated cDNAs which appeared to be maize specific (TC286704, TC300122 [ZmGI], 2569799 [NCBI UniGene]) and, based on searches of public da tabases, 10 were found exclusively or highly represented in cDNA libraries from reproductive tissues (2568974/TC285721, 514900, 2566963,

PAGE 32

32 2568212/TC286030, 507904/TC301902 [NCBI UniGene/ZmGI ]) or from drought-stressed plants (2564044/TC285867, 2714857/TC286791, 508486/TC29233, 2561245/TC299973, 2567165 [NCBI UniGene/ZmGI]). Quantitative differences in levels of specific mRNAs were confirmed for a subset of genes by real-time RT-PCR analyses of the wild-type and vp1 mutant samples (Figure 2-6). Results showed that differences in tr anscript abundance between wild -type and mutant RNA samples used in 3'-cDNA sub-library construction paralleled the 454-based expression profiles. Resolution of Near-Identical Transcripts by Polymorphisms Analyses of the m aize genome have revealed a high frequency of nearly-identical paralogs with 98% identity (Emrich et al., 2006b). In most instances, both gene copies are expressed. Identification of single feature polymorphisms in the 3' sequences can effectively distinguish a subset of such paralogs. At least one example where 3'UTR prof iling effectively resolved nearidentical paralogs was evident for closel y-related, but diffe rentially expressed Auxin Repressed Dormancy Associated transcripts (we designated these genes as ARDA1 and ARDA2). The ARDA1 and ARDA2 sequences share 98 % identity (99% in the coding region and 97% in their 3'UTRs). Two distinct 3'UTR ARDA consensus sequences detected an 18-bp indel polymorphism that distinguished these two paralogs. Read frequencies showed reciprocal responses in the mutant background by ARDA1 (p < 10-53) and ARDA2 (p < 10-12). Differential profiles for these genes were confirmed by amp lifying the region in or around the indel using real-time RT-PCR (Figure 2-7A). These reciprocal expression profiles could not be resolved when regions outside of the indel sequence we re amplified due to confounding effects of the nearly-identical sequences. Earlier work identified ARDA1 as a potentially important cont ributor to stress-tolerance in hybrid maize (Guo et al., 2004). The previously-undetected ARDA2, resolved in the mono-

PAGE 33

33 allelic W22 inbred, matched a unique maize EST. Alignment of the two consensus sequences to a region within assembled genomic sequen ce (MAGI4_156527) verified presence of two paralogous gene products (Figure 2-7B). Both ARDA paralogs appeared to be droughtresponsive in preliminary analyses. Currently, th ere is little informati on for putative roles of Auxin Repressed Dormancy Associated genes. Studies in pea (Pisum sativum) characterized similar genes as markers for dormancy in axillary buds (Stafstrom et al., 1998). Validation of SNPs and Homopolymer-Based Polymophisms We conducted a detailed analysis of polym or phisms detected by a preliminary dataset comprised of 1,263 W22 consensus sequences using BLASTN alignment to MAGI4 B73 genomic sequences. We expected that some portion of apparent polymorphisms in consensus sequences ranging from 2 to 75 reads (56.6%, Ta ble 2-4) was due to sequence errors. To estimate the contribution of sequence errors in the 454 data, we evaluated polymorphisms detected by a subset of 107 cDNA consensus sequ ences (7 to 75 reads) with respect to B73 MAGI assemblies by independent BLASTN searches of IUC cDNA and public EST databases. We confirmed 93.8% of 146 sequence polymorphi sms detected within 52 W22 alleles by identical cDNA matches indicating that most identify independently -documented maize alleles. Because the pyrosequencing method used by 454 is prone to errors in estimating lengths of long homopolymer runs (Margulies et al., 2005), we investigated the effect this may have on SNP detection in maize sequences. Overa ll, 29% of the 1,263 W22 consensus sequences analyzed above contained one or more homopolymer tracts of 5-bp or longer. In order to assess the impact of homopolymer read errors on SNP detection by 454, we analyzed the polymorphisms detected by a set of 211 W22 consen sus sequences (5 to 75 reads) in best alignments to the MAGI4 (B73) dataset. Of the total 257 polymorphisms detected (counting indels as one), at l east 89.9% were independently confir med by identical cDNA matches. In

PAGE 34

34 addition, only 60 (23%) were potentially at tributed to simple or compound (e.g. CCT TT CCC TT) homopolymer base-calling errors. Moreover, the 60 homopolymer-based polymorphisms were distributed randomly (p > 0.9) between alignments that included long homopolymer tracts of 5 bp (21% of consensus sequences analyzed) and alignments that lacked them. Finally, all but 7 of 60 homopolymer length po lymorphisms were supported by independent EST sequences from W22 or othe r sources. Hence, th ese homopolymer-based polymorphisms were not appreciabl y less reliable (88.3% confirmed) than other substitution and indel polymorphisms (93.8% confirmed by independent cDNA sequences). Discussion Our results dem onstrate that 3'UTR profiling is an effective strategy for high-resolution global analysis of gene expressi on that does not require a complete genome sequence. Using this approach, we were able to identify over 14,000 gene-specific mRNAs and quantify expression based on read frequencies occurring in 3'-anc hored consensus sequences. Analysis of the quantitative 3'UTR profile revealed a dynamic ra nge of gene expression spanning greater than three orders of magnitude. Our strategy of using long-read, 454 sequenci ng to target gene-spe cific 3'UTRs offers several advantages over previous tag-based approaches to global expression profiling. First, depth of sequencing is enhanced by anchoring the 454 reads to uni que sites proximal to the 3' ends of transcripts. This eliminates redunda ncy associated with shotgun sequencing of cDNA fragments, thus providing more reads per unique transcript a nd reducing the potential for highlyexpressed mRNAs to saturate the library (Weber et al., 2007). In this st udy, the two-sub-library analysis using the 454 Genome Sequencer 20 instrument identified 47,299 distinct mRNAs (including 14,822 consensus sequences represen ted by 2 to 2500 reads). In comparison, sequencing of nebulized Arabidopsis cDNAs yielded approximately 17,500 unique transcripts

PAGE 35

35 after two GS 20 sequencing reactions (Weber et al., 2007). Although we cannot discount the possibility that a portion of the singlets identifie d in our dataset are due to sequencing errors, deeper sampling with the upgraded FLX technology will provide enhanced statistical support for rare transcripts. Second, the specificity of these long, 3'UTR-based sequence reads facilitates unambiguous gene assignment. Our analyses indicated that indivi dual gene family members can be resolved by unique, gene-specific 3'-anchored tags and the corresponding closely-related ESTs characterized. Finally, enrichment of 3'UTR se quences provides a useful source of polymorphic information for studies of natural variation. Id entification and analysis of nearly-identical paralogous genes is improved on a genome-wide scale by enrichment for polymorphisms in the 3' sequences. Even in cases where genomic information is very limited, high-throughput sequencing of 3'UTRs from species variants a llows direct comparison of polymorphic loci. This approach thus provides a tool for genotyping and assessing genetic diversity contributing to quantitative traits without the need for a se quenced genome or extensive EST collections. Approximately 22% of the unique mRNAs identified in this study by 2 reads did not match ESTs in either ZmGI or IUC databases. A similar percentage of novel sequences (30%) were also observed for a transc ript profile from maize shoot apical meristem (SAM) using 454based shotgun sequencing of sheared cDNAs (E mrich et al., 2007a). The 8% difference may reflect an increased sp ecificity of our BLASTN results us ing the IUC cDNA collection and/or more novel transcripts identified in the non-differentiated SAM tissu e. Our data also showed that among distinct mRNAs matching cDNAs, appr oximately 50% either contained domains of unknown function and/or were uncla ssified based on lack of homology to annotated genes in

PAGE 36

36 other species. This percentage demonstrates the potential for gene discovery with unbiased sampling and sequencing of gene-specific 3'UTRs. Furthermore, quantitative analyses of closel y-related transcripts can extend studies of functional genomics to species without completely sequenced genomes and where gene families are largely uncharacterized. We addressed this possibility with an analysis of Histone-1 ( H1 ) transcripts in maize ovaries. Although the individual genes have not been characterized in maize, identification of the corresponding H1 ESTs indicated that these unique, non-redundant transcripts are indeed expre ssed. One highly represented Histone-1 mRNA in maize, TC292133a, was annotated as a droughtand ABA-inducible H1 gene based on sequence similarities in tomato (Bray et al., 1999). This annotation is consistent with a function for this highly expressed H1 in drought-stressed maize. For organisms that have limiting cDNA resources, 3'-cDNA tags will be less likely to align with upstream coding sequences, thus constraini ng functional annotation. Nonetheless, 3'UTR sequences enable resolution of unique mRNAs and distinguish among closely-related transcripts. Quantitative data on transcript abundance is also provided, as well as an open, unbiased sampling of the transcriptome. Where additional cDNA information is av ailable, the 3'-cDNA sequences can be extended by BLASTN alignments. Alterna tively, the sequence tags can be used to design primers or probes for screening of cDNA libraries While the divergence of 3'UTR sequences facilitates resolution of genes within a genome, it may limit the effectiveness of cross-species comparisons for annotation of transcripts. For example, alignment of maize ovary 3'-cDNA consensus tags to the complete set of ri ce genes (OsGI) using BL ASTN produced matches (expectation score <1e-5) for only 20.6% of the transcripts.

PAGE 37

37 Based on our analysis of SNPs identified within consensus sequences and comparisons with B73 MAGI genomic assemblies (Table 2-4), we confirmed at least 89.9% of polymorphisms independently by identical cDNA matc hes. These data are consistent with a recent study by Barbazuk et al. (2007) in which 88 % of SNPs sampled by 2 or more 454 reads were validated by Sanger sequencing. Removal of the unconfirmed SNPs from our analysis reduced the estimated polymorphisms in W22 re lative to B73 to 43.9%. That estimate is comparable to the 44% polymorphism reported fo r B73 and Mo17 alleles (V roh Bai et al., 2006). Due to incomplete coverage of the B73 genome, it is likely that some W22 consensus sequences were aligned to closely-rela ted, paralogous MAGI4 sequences rather than alleles (e.g. Auxin Repressed Dormancy Associated paralogs). In addition, our preliminary re sults indicate that homopolymer base-calling errors will have a minor impact on ability to analyze polymorphisms in ma ize cDNAs. Importantly, even where errors of this type occur, the consistency of base calling in reads derived from independent 454-libraries suggests that non-identi cal alleles may still be distingui shed if they give rise to different consensus sequences. This level of specifi city in gene expression analysis is invaluable to uncovering novel variation in polyploid or paleopolyploid genomes (Osborn et al., 2003). Evidence of ancient tetraploidi zation in the maize genome can be observed for roughly 60% of genes in duplicated regions (Messing et al., 2004 ). Conservative estimates indicate that extensive amplification of tandemly duplicated genes may represent approximately one-third (35%) of maize genes (Messing et al., 2004). Our analysis of expressed CesA gene family members demonstrates the capacity of the approach described to provide qua ntitative resolution of closely-re lated transcripts. This is achieved by specificity of the 3'UTRs for indi vidual cDNAs. Cross-hybridization of near-

PAGE 38

38 identical transcripts often complicates identification of individual gene family members in arraybased experiments. Consistent w ith this, the resolution of three CesA4 transcripts, including a putative splice variant, denotes the complexity within the CesA gene family in maize. Even with the most stringent probe designs, cross-hybridiz ation with unknown family members remains a challenge in non-sequenced genomes. With unbi ased sampling and sequencing, resolution of tissue and/or temporal-specific tr anscripts and polymorphi c variants will provide functional clues in complex genomes such as maize (Ma et al., 2006). Furthermore, quantitative assessment of transcription among individual members of a gene family can facilitate analyses of functional genomics and address key questions in evoluti on. Studies in Arabidopsis have identified instances of functional diversif ication among duplicated genes either in parallel biochemical pathways (Blanc and Wolfe, 2004) or within specific developmental and metabolic processes (Schmid et al, 2005). Results from a quantitative 3'UTR expression profile showed that the Zipf power distribution of gene expression in the entire dataset overall, but was not conserved within the chromatin-related functional cla ss. This group of mRNAs show ed a skewed distribution of abundance due mainly to a large num ber of distinct highly-expressed, Histone-3 ( H3 ) transcripts. Among these we identified 67 mRNAs having H3 functional domains and 39% of the consensus sequences were represented by 100 to 1,000 reads. Results may be due to transcriptional responses to the stress treatme nt or be specific to the re productive tissues examined. Validation of differences in transcript abundance for a subset of genes by real-time RTPCR in RNA samples used for sub-library construction supports 3'UTR pr ofiling as a platform for quantitative expression profili ng between samples. Furtherm ore, construction of the 3'cDNA libraries by this method yielded sequences with very low retr otransposon content and

PAGE 39

39 nominal rRNA contamination. In addition, read distribution between multiplexed samples was well balanced. Thus, a multiplexing strategy can be used to concurrently profile multiple samples for increased cost effectiveness. In corporation of a four-b ase error detecting key enables up to 64 unique combinations fo r individual sample recognition. Preliminary data generated with the recen tly upgraded FLX 454 technology (Harkins and Jarvie, 2007) identified approximately 22,920 unique consensus sequences with a much higher depth of sequencing for 12 multiplexed samples in a single reaction (Eveland, unpublished). The enhanced sequencing capacity of FLX will therefore provide improved statistical analyses while increasing the number of multiplexed cDNA libraries (e.g. biological replicat es and treatments). In the present study, a single GS20 run enabled detec tion of unique transcripts (2 or more reads) with a sensitivity of approximately 1 in 100,000 mRNA molecules. A similar run on the 454FLX instrument is expected to incr ease sensitivity by at least 2-fold. This will be directly applicable to identifying rare transcri pts and resolving complex gene families. Also, the range of gene expression quantif ied by the 454-based 3'UTR profile provides higher resolution in global tran script profiling analyses compar ed to array-based hybridization experiments. Accordingly, our results includ e detection of many rare mRNAs as well as quantification of highly abundant transcripts. In contrast, th is level of resolution was not observed in initial microarray analyses of th e same tissues (Eveland, unpublished) due to threshold levels of detection and saturation. Li kewise, with array-based interpretation of foldchanges, subtle variations in gene expression are often not det ected, but can have a significant impact on physiology. A quantitative appraisal of a ll expressed sequences is thus invaluable to studies of quantitative traits such as heterozygosity (Birchler et al., 2003; Stupar and Springer, 2006).

PAGE 40

40 Future Prospects W ith 454-based, long-read sequencing of 3'UTRs, quantitative profiles for allele-specific inheritance patterns can be ge nerated in the absence of a priori data on polymorphisms. Allelic variants are frequently distinguished by single-feature polymorphism s such as those that marked nearly-identical paralogs in this study. Identifying allele-specifi c differences in gene expression and quantifying parental contributions to comple x traits in F1 hybrids is key to understanding genetic mechanisms such as imprinting (Guo et al., 2003) and heterosis (Bircher et al., 2003; Springer and Stupar, 2007). In addition, strategies for eQTL analyses (Schadt et al., 2003; Borevitz and Chory, 2004) and genome-wide linka ge studies (Cheung et al., 2005) are improved by a high-resolution, non-biased ap proach to quantifying allelic im balances in gene expression (including those resultin g from imprinting or X-chromosome inactivation). Furthermore, analysis of natural variation is enhanced by recovery of haplotypes in species where genomic information is limited. Natural variation can also be assessed with array-based probe sets generated from 3'anchored sequence reads (Borevitz et al, 2003). For species in which comprehensive microarray platforms are not available, the 3'UTR sequence reads can serve as blueprints for chip construction with highly-specific probe sets representing an unbiased sample of expressed sequences. Alternatively, for genomes with limited EST support, this method can enhance efficiency of cDNA sequencing by pre-screeni ng libraries to elimin ate redundancy. Fine mapping and marker-assisted breeding can also be facilitated by utilizing insertion-deletion polymorphisms (indels) in the 3'UTRs as molecula r markers (Bhattramakki et al., 2002; Vroh Bi et al., 2006). In addition, anchoring the 454 sequences proximal to the 3' ends of transcripts enables resolution of 3'-RNA processing variants Instances of differential polyadenylated transcripts were readily detected in this study (data not shown) Currently, identification and

PAGE 41

41 characterization of alternate poly( A) sites is fragmentary for mo st genes, since the required length of 3'UTR sequence has been largely out side the range of short-read technologies (Jongeneel et al., 2005). Conclusions By com bining the specificity of 3'UTRs w ith long-read, high-throughput sequencing, we are able to distinguish expressi on of newly-identified genes and closely-related transcripts on a genome-wide scale. This can also be accomplishe d without reference to a completely sequenced genome. The approach provides an efficient av enue for gene discovery and elucidation of variations in expression that unde rlie natural variation and contri bute to complex genetics of heterosis and imprinting. In addition, 3'UTR profiling advances studies of comparative and functional genomics by quantitatively resolving expression of gene families and identifying unknown gene family members. Materials and Methods Plant Materials Maize (Zea mays) plants were grown in 14/ 7 ga llon pots under gree nhouse conditions (Sept. to Nov. in Gainesville, FL) at 12-h-li ght/12-h-dark cycles. Sibling wild-type and vp1 mutant plants were in a W22 inbred bac kground were derived from a self-pollinated vp1 /+ heterozygous ear. A drought-str ess treatment was initiated by gradually withholding water beginning two weeks prior to tassel emergence. Soil was covered to restrict water loss by evaporation and pots were weighed at the e nd of each day to determine water loss to transpiration. Water lost to transpiration was added back. One week after ears first appeared, water was withheld completely. Ears were collec ted right before silk emergence from wild-type and vp1 mutant plants. Immature ovaries (with pe dicels) were hand-dissec ted from equivalent

PAGE 42

42 sections of each ear (base-to-mid section), weighed to 50mg FW ( 15 ovaries per ear), and frozen in liquid N2. Sub-library Preparation and Sequencing Tissue was hom ogenized in TRIz ol reagent (Invitrogen) using a FastPrep lysis system (QBIOgene). RNA was extracted using standard me thods based on protocols from the University of Arizona ( www.maizearray.org ). Total RNA (5ug) from wild-type and vp1 mutant ovaries was used for cDNA synthesis (MessageAmp II, Ambion) and primed with 6 pmol biotinylated (T12) B-adaptor (modified from Margulie s et al. [2005]) oligo: BiotinCCTATCCCCTGTGTGCCTTG CCTATCCCTGTTGCGTGTC TCAGTTTTTTTTTTTT[AGC]. Purified cDNA (DNA clear, Ambion) was bound to M-270 Streptavidin beads (Dynal), immobilized on a Magnabot 96 (Promega), and digested with Msp1 (Promega) to create 2-base CG overhangs for adaptor ligation. A-adaptor ol igos (modified from Margulies et al. [2005]) included 3-base muliplex keys (wild -type sub-library top strand, 5'CCATCTCATCCCTGCGTGTCCCATCTGTTCCCTCCCTGTCTCAGCAT -3', wild-type sublibrary bottom strand, 5'CGATG CTGAGACAGGGAGGGA ACAGATGGGACACGCAGGGA TGA-3'; vp1 mutant sub-library top strand, 5'CCATCTCATCCCTGCGTGTCCCATCTGTTCCCTCCCTGTCTCAGACT -3', vp1 mutant sub-library bottom strand, 5'CGAGT CTGAGACAGGGAGGGAAC AGATGGGACACGCAGGGATGA-3'). Adapter pairs were combined and concentrated to 1 pmol/uL in salt buffer (10 mM Tris, 1 mM EDTA, 50 mM NaCl [pH 8]) an d annealed by incremental, -1 degree/min decreases (95 4C, with a 30-min hold at 72 71C). Adaptors (5 pmol) with multiplex keys CAT and AGT were ligated to digested wild -type and mutant cDNA samples, respectively. The 3-base key

PAGE 43

43 sequences enabled detection of singl e-base errors in the multiplex key. Un-ligated adaptors were removed by washing beads twice with 1X B& W buffer (2.5 mM Tris -HCL [pH 7.5], 0.25 mM EDTA, .5 M NaCl) and twice with ddH20. The desired 5'-A-cDNA-B-3' template strand was eluted with 100 mM NaOH, neutralized and concen trated on a Qiagen co lumn (Margulies et al., 2005). Sequencing was conducted as per Margulies et al. (2005) using a 454 GS-20 instrument. The expected yield of ~3x109 template molecules for the combined libraries was confirmed by a SYBR Green Q-PCR strategy (Myi Q, Bio-Rad). Molecules L-1 of amplified product were calculated from an in-vitro transcribe d (MAXIscript, Ambion) alpha-tubulin ( Z. mays) standard: alpha-tubulin forward, 5'-TTG TGCCTGGTGGCGACCTGG-3 and alph a-tubulin reverse: 5'ACCGACCTCCTCGTAGTCCT-3'. Data Analysis Quality -trimmed 454-sequences (FASTA format) were filtered for valid key and ligation junction (CGG) sequences at 5'-ends and polyA tails were trimmed using custom programs written in java. Validated, trimmed sequences (9 3% of total reads) were assembled using CAP3 (http://genome.cs.mtu.edu/sas.html). The non-redundant set of consensus cDNA sequences represented by 2 or more reads (14,822 total assemblies) were annotated by BLASTN searches of cDNA databases for maize. These included TIGR Maize Gene Index Assemblies (ZmGI) and Industry UniGene (IUC), a collection of cDNAs pr ovided by an industry consortium via a users agreement ( http://www.maizeseq.org ). Functional classifications of cDNA m atches were based on Gene Ontology terms associated with PFam ( http://www.sanger.ac.uk/Software/Pfam/) assignm ents in IUC. In addition, consensus sequences were aligned by BLASTN to a maize genome sequence assembly (MAGI, version 4.0 [ http://magi.plantgenomics.iastate.edu/ ]). Trace files for 454 sequences were d eposited in NCBI.

PAGE 44

44 Real-time RT-PCR Analysis for Validation of 454 Data Real-tim e RT-PCR was carried out to validate technical replicates of RNA samples used in sub-library construction. For real-time PCR analysis, cDNA wa s synthesized from DNaseItreated (Ambion) total RNA using an oligo(dT) primer (TaqMan Reverse Transcription Reagents, ABI). Real-time PCR was monitore d using the MyiQ Single Color Real-Time PCR Detection system (Bio-Rad). Each reaction c ontained 10 uL of 2x iQ SYBR Green Supermix (Bio-Rad), 1.0 uL of cDNA sample, and and 200 nM gene-specific primer in a final volume of 20 uL. All reactions were performed in tripli cate. The relative abundance of transcripts was normalized with 18S ribosomal RNA control values using Taqman (Ribosomal RNA Control Reagents, ABI) and to the constitutive expression of an alpha-tubulin mRNA using SYBR Green on cDNA templates (MyiQ, Bio-Rad) SYBR Green was used to amplify a subset of transcripts with gene-specific primers. Primer pairs were designed using the 454-read and adjacent sequences in best-match ESTs iden tified by BLASTN as templates. CBS Chloride channel ([2562879] Cbs forward, 5-ATGGATG CTGCTGTTCTCATGCTC-3 and Cbs reverse, 5ATGGAGTCTCCTGGCGTGCT AC-3), Thaumatin/osmotin ([1321765] Osmotin forward, 5-TACCGCAGCAGCTGAACAACG-3 and Osmotin reverse, 5ATGTTCCGTCGCAGTCGCTAGG-3), Senescen ce-associated/tetraspannin ([TC299489] Sa forward, 5-AACGACGAGGACGACCTCTGC-3 and Sa reverse, 5AGTTTGATTAAGCGTCACCGCCTCG-3), Chlo rophyll a-b binding protein ([TC299127] Cab forward, 5-TGTACCCTGGCGGCAGCTTC-3 and Cab reverse, 5ATCCACGTACGTACACCCTCTCC-3), Cu transport ATPase/heavy-metal-associated ([2562278] Hma forward, 5-AGCCAAAGCTGACGCCTGATC-3 and Hma reverse, 5TCCTGCAAGGGATGTGTTGTT C-3), Glycine-rich protein ([2923887], Grp forward, 5ATCAGGTGAAGGATACG GACAAGGTG-3and Grp reverse, 5-

PAGE 45

45 ACAGGACAAATTACAAGCCTTGCGGTG-3), Dehydrin DHN1/RAB-17 ([TC286791] dehydrin forward, 5-ACAGCACTGAGCGGCGCCTATAC-3 and dehydrin reverse, 5ACGTAGCAGCATAAACAGTACACGGACC-3'). Relative expression levels of ARDA1 and ARDA2 were compared by real-time RT-PCR using SYBR Green and gene-specific primers within and around the 18-bp indel sequence. ( Arda1 forward, 5'-TACAAGCGGGCGCAGTCG-3', Arda1 reverse, 5'AGCAAACATGGCCTCTTCACTG; Arda2 forward, 5'-TACAAGCGGGCGCAGTCG-3', Arda2 reverse 5'-TGGCCTGACAGAGACACCG-3').

PAGE 46

46 Table 2-1. Summary statistics of a two-sample multiplex 3'-UTR library. Total high-quality readsa 228,595 Wild-type sub-library reads 105,289 Mutant sub-library reads 109,958 Error-detected reads 13,348 Errors in multiplex key 1.5% Invalid ligation junctions 5.5% Total unique sequencesb 49,822 Singletsc 32,477 Consensus sequences ( 2 reads) 14,822 EST matches 11,559 Genome matchesd 9,740 aAbundance of reads after filtering for Msp1ligation junctions. bNonredundant set of sequences representing unique transcripts. cSingle-copy transcripts. dAbundance of consensus sequences matching cDNAs in Industry Unig ene or ZmGI databases. eAbundance of consensus sequences matching available maize genomic sequence assemblies (MAGI4).

PAGE 47

47Table 2-2. CesA family 3'-anchored consensus sequences. 3'-anchored consensus sequence Read # % Match CesA6a CGGAGGCTGCGGCAACCTTGTGCAGTTCGGCCACGAATATACTAGGGAAGATCGC GACCAATCAATCAATCGATGACCGAGTTCAATTGTTCAAAG 26 100 (93) CesA6ba CGGATCGACCCTTTCCTTGCGAAGGATGA TGGTCCCCTGTTGGAGGAGTGTGGTCT GGATTGCAACTAGGAGGTCAGCACGTGGACCTTCCCGTNAGTGTGTGG 9 97 (104) CesA8 CGGGATCTCGAACGCGATCAACAACGGGTACGAGTCGTGGGGCCCCCTGTTCGGG AAGCTCTTCTTCGCCTTCTGGGTGATCGT CCACCTGTACCCGTT CCTCAAGGGTCTG GGTGGGG 8 100 (113) CesA10 CGGACGCCCACCATCGTCGTGCTCTGGT CCATCCTCCTCGCCTCCATCTTCTCGCTC GTCTGGGTCAGGATCGA CCCGTTTATCCCGAAGGC CAAGGGCCCCAT 2 100 (104) CesA4a CGGATACCCAGACGTGTGGCATCAACT GCTAGGGAAGTGGAAGGTTTGTACTTTGT AGAAACGGAGGAATACCA CGTGCCATCTGTTGTCTGTTAAGTTATATAT 5 100 (106) CesA4bb CGGATACCCAGACGTGTGGCATCAACTGCTAGGGA GTGGAAGGTTTGTACTTTGAGAAAC GGAGGAATACCACGTGCCATCTGTTGTCT A GTTAAGTTATATA 2 95 (107) CesA4cc CGGGGTGTTGGTGATATTGATGCATCAACTGATTACAACATGGAAGATGCCTTATT G TGAGTTCCAACACCCTCCTT CAGCTCACTCATTTG 2 100 (57) CesA9 CGGATACTCGAACGTGTGGCATCAACTG CTAGGGAGGTGGAAGGTTTGTAGAACA GAGAGATACCACGAATGTGCCGC TGCCACAAATTGTCTGTTAAG 10 100 (98) CesA5 CGGGTCACTGGCCCTGATATCGCGAAATGT GGCATCAACTGCTAGGATGAGCTGA ATATAGTTAAAGAGTGGAACTAGACGCATTGTGG 27 96 (88) CesA2 CGGTGCTGCTGCAGACAATCATGGAGCC TTTCTACCTTGCTTGTAGTGCTGGCCAG CAGCGTAAATTGTGAAT TCTGCTTATTTTTTTAG 14 100 (78) CesA1 CGGTGCTGCTGCGGACTAAGAATCACGGAGC CTTTCTACCTTCCATGTAGCGCCAG CCAGCAGCGTAAGATGTGTAA TTTTGTAAGTTTTGTTATGTC 18 97 (96) CesA3d CGGCACAATCATGATCTACCCCTTCGTGT AAATACCAGAGGTTAGGCAAGACTTTT CTTGGTAGGTGGCGAAGATGTGTCGTTT AAGTTCACTCTACTGCTAGTTTGGGGG 17 100 (100) aCesA6b reads are located upstream of the CesA6a consensus tag, consistent with a tran script containing a restriction site polymorphism. However, absence of the downstream Msp1 site was not inde pendently confirmed. bCesA4b contains 3 indel polymorphisms in comparison to CesA4a and all 3 are represented in both reads. cCesA4c aligns with a different genomic assembly.

PAGE 48

48 Table 2-3. Best-match cDNAs a nd associated annotations (BLA STN) for consensus sequences showing highly significant differences in transcript abundance between wild-type and vp1 mutant drought-stressed ovary sub-libraries. read frequency cDNA ID % match (length) BLASTn annotation (species) wt vp1 P Value TC285867b 100 (87) Auxin-repressed dormancy ( R. pseudoacacia ) 238 709 10-53 TC285721a 100 (98) GRP (O. sativa ) 57 244 10-27 TC286704 98 (94) Unannotated 25 167 10-24 TC305930 98 (80) farnesylated protein 3 ( H. vulgare ) 14 119 10-20 2569891 100 (52) XET ( H. vulgare ) 234 81 10-18 TC285789 100 (96) Auxin-repressed dormancy ( P. sativum ) 53 1 10-12 TC292711 100 (100) nodulin MtN3 family ( A. thaliana ) 0 52 10-13 514900a 97 (116) At1g74950 ( A. thaliana ) 127 37 10-12 TC286791b 100 (86) dehydrin RAB-17 protein ( Z. mays ) 8 69 10-12 TC292358 100 (96) Thr rich extensin ( Z. mays ) 1070 1414 10-12 507881 100 (95) unnamed protein product ( O. sativa ) 93 21 10-11 TC286485 100 (87) histone H2A ( Z. mays ) 575 396 10-11 2566963a 99 (104) unknown protein ( O. sativa ) 74 12 10-11 TC286030a 100 (111) harpin induced gene 1 ( O. sativa ) 106 31 10-10 TC310545 100 (93) histone H1 ( Z. mays ) 259 135 10-10 TC294233b 100 (104) putative cystatin cc3 ( S. officinarum ) 65 158 10-10 2708354 100 (97) unannotated ( O. sativa) 75 16 10-10 TC298173 95 (93) histone 3 ( O. sativa ) 151 61 10-10 TC294050 100 (95) EF-hand Ca2+-binding CCD1 ( T. aestivum ) 102 32 10-9 TC301902a 100 (87) AP2 domain, EREBP ( O. sativa ) 66 13 10-9 TC305186 100 (78) subtilisin-like proteinase ( O. sativa ) 113 41 10-9 1572511 100 (49) hypothetical protein ( O. sativa ) 31 0 10-8 TC299973b 95 (84) glycogenin-like ( O. sativa ) 31 0 10-8 654573 100 (45) Ca+-binding EF hand family ( A. thaliana ) 53 10 10-8 TC280589 100 (98) phosphate-induced protein 1-like ( P. ciliare ) 146 67 10-8 2567165b 90 (110) heavy-metal-associated ( O. sativa ) 141 64 10-8 2569891 100 (52) xyloglucan endo-1,4-beta-D-glucanase ( H. vulgare ) 43 6 10-7 TC310820 100 (91) CCCH-type zinc finger protein-like ( O. sativa ) 15 61 10-7 2569799 100 (74) Unannotated 59 14 10-7 TC300122 99 (103) Unannotated 127 56 10-7 Read frequency differences for consensus sequences were analyzed using a Chi square statistic. Consensus sequences were aligned to best-mat ch cDNAs using BLASTN in ZMGI and Industry UniGene databases. NCBI UniGene IDs are list ed for sequences where TC numbers were not available. acDNAs highly represented in Genbank libra ries from maize reproductive tissues. bcDNAs highly represented in Genbank libraries from drought-stressed maize plants.

PAGE 49

49 Table 2-4. Polymorphisms detected by W22 3'-anchored 454 sequence reads. Consensus tags with 2 to 75 reads (n = 1263) Total polymorphic W22 alleles 56.6% B73 MAGI Matches1 92.4% Consensus tags with 7 to 75 reads (n=107) Total polymorphic W22 alleles 52 # confirmed 47 (43.9%) Total polymorphisms 146* # confirmed 137 (93.8%) Consensus tags with 5 to 75 reads (n=211) Total polymorphisms 257* # confirmed 231 (89.9%) Homopolymer-type polymorphisms 60* # confirmed 53 (88.8%) 1 Contiguous BLASTN alignment of >50 bp or > 90% of query length. Percent of matches having one or more SNP or indel polymorphi sm within the best aligned segment. Allele sequences supported by identity to independent EST sequences.

PAGE 50

50 Figure 2-1. Comparison of shotgun versus 3'-anchored approaches for using 454-sequences as a platform for transcript profiling. A) Sequencing of randomly sheared cDNA fragments, followed by contig assembly can provide coverage of full-length cDNAs. However the redundancy of sequence-reads fo r a given transcript limits information returned per number of reads and thus st atistical analyses of expression. B) Our 3'UTR profiling method identifies unique ESTs by anchoring each sequence read to a gene-specific region of cDNAs. This increases depth of sequencing and facilitates assembly and analysis. A B 5 3 EST 5 3 EST

PAGE 51

51 Figure 2-2. Strategy for 3'UTR profiling by 454-se quencing. A) A 3'-anchored cDNA library is restriction-digested and tagged via liga tion to a multiplex adaptor. Unique combinations of a 4-base multiplex error-detecting key enables sample identification during concurrent sequencing of up to 16 s ub-libraries. B) Sublibrary construction from total RNA. derived bead-linker multiplex adaptor 3-base multiplex error-detecting key (16 combinations) 4-base calibration key Msp1 (5overhang) sequence primer for 454 3sequence read ( ~100bp ) AAAA A AAAAAAA AAAAAAA TT TT TT TT AAA AAA AAA biotin labeled 3 anchor oligo streptavidin bead pull down Msp1 5 overhang TT AAA multiplex adapter ligation elute 454 bead anchor copied B

PAGE 52

52 Figure 2-3. Distribution of tag lengths for a simulated Msp1 digest of 70,000 3'-oriented ESTs from B73 maize (maize full length c DNA project [www.maizecdna.org]). A comparison of simulated tag lengths to th e distribution of annotated rice 3'UTRs indicated enrichment of 3'UTR sequence with an average tag lengt h of 100-200 bases. Proportion of short tags is reduced due to the low frequency of MspI sites proximal to the poly(A) tail. 0 0.1 0.2 0.3 <25100200300400500600700800Fraction of total tags Rice 3'UTRs B73 tags Tag length

PAGE 53

53 Figure 2-4. A graphic presentation of the quantitative 3'UTR profile representing 11,559 consensus sequences that matched cDNAs in the 2-sample multiple xed library. Read frequency was used as a quantitative measure of mRNA abundance. A) Transcript abundance is plotted on a log-log scale for re spective genes in rank order from leastto most-highly expressed. B) Transcripts are grouped into functional classifications based on Gene Ontologies and plotted linearl y along the y-axis. Read count for each transcript is plotted on the xaxis (log scale). Color scale (dark-to-light) denotes the dynamic range of mRNA abundance.

PAGE 54

54 Figure 2-5. Distribution of transc ripts (log-log scale) in selected functiona l classes from Figure 4B and read count for each. Resolution of individual gene family members was enabled by the specificity of the 3'UTR. A) Quantitative measure of mRNAs for all transcripts classified with cell-wall-relat ed functions (Fig. 4B) and resolution of 12 unique mRNAs representing 9 previously-characterized members of the CesA gene family (including transcript variants for CesA4 and CesA6) in maize ovaries. Read frequencies for CesA gene family members range from 2 to 27 reads. B) Quantitative measure of mRNAs for all tr anscripts with chromatin-rela ted functions (Fig. 4B) and identification of Histone ( H1 )-like transcripts that matched unique, but uncharacterized maize cDNAs (indicated by TC or GenBank IDs). Read frequencies for H1 gene family members range from 2 to 684 reads. Transcript abundance (log scale) B 0 1 10 100 TC292133a TC292749 gi:14203885 TC294259 TC292931 TC292749 TC292750 gi:149102785 TC292133b 1 100 10 1000 Transcript rank order (log scale) A Transcript rank order (log scale) CesA5 CesA6a CesA1 CesA3 CesA9 CesA6b CesA8 CesA4a CesA10 0 1 10 100CesA4b CesA4c 1 100 10 CesA2 Transcript abundance (log scale)

PAGE 55

55 Figure 2-6. Validation of technical accuracy for determining differences in transcript abundance based on read number. Pa rallel real-time RT-PCR (SYBR green, MyIQ, BioRad) analyses (technical error based on 3 inde pendent determinations) on identical RNA samples as used in 3'UTR sub-library construction validated the significant differences in read frequency (p < 0.0015 except for alpha-tubulin control) for a subset of genes. Gene-specific primer s were designed using the 454-sequence-reads and associated EST matches as templates. wild-type vp1 mRNA abundance (relative % ) 0 50 Q-PCR 454 Q-PCR 454 Q-PCR 454 Sa alpha-tubulin: Dehydrin: Grp: Q-454 100 50 0 100 Cab Osmotin Cbs: Hma Q-PCR 454 Q-PCR 454 Q-PCR 454 Q-PCR 454

PAGE 56

56 Figure 2-7. A 3' sequence polymor phism resolved nearly-identic al Auxin Repressed Dormancy Associated paralogs with differences in mRNA abundance. A) Q-PCR analyses confirmed the reciprocal re sponses of near-identical Arda1 and Arda2 ( 98% identity) on wild-type and mutant samples us ed for sub-library construction (technical error based on 3 independent determin ations). B) The near-identical ARDA1 (EE188942) and ARDA2 (EE679809) ESTs differ by an 18-bp insertion within ARDA2, which is not present in MAGI4_156527 (region of available maize genomic sequence aligning with the ESTs). The 454-sequence-reads representing ARDA1 (blue) and ARDA2 (brown) are highlighted within the 3' end of the ARDA2 EST sequence (top) and within a schematic diagra m of the two ESTs (3' ends) aligned to maize genomic sequence (bottom). 5' GGCC 3' 3' intron 5' ARDA1 ARDA2 GGCC Arda1 Arda2 vp1 wt 20 40 60 80 100 0 Relative abundance (% mRNA) vp1 wt A B 5' 3' CCGG CCGCAGCCCAACTCCCCCACCGTCTACG ACT GG CTCTACAGCGACGAGACCAGGAGCAAC CACCGCTAGATCGGGCGAGATGGACAGTG AAG AGGC CCGG TGTCTCTGTCAGGCCATGTTTGCT GGTCCTTGTGGTTATCTAAAACGCATGCATGCT CTATTATAGCTAGTCATCACTATATA TATAC

PAGE 57

57 CHAPTER 3 EXPRESSION PROFILING OF DEVELOPING MAIZE OVARIES USING MICROARRAYS AND SEQUENCING OF 3'-UT RS Introduction Im proving agronomic traits for grain yield and quality is facilitated by increasing availability of relevant genomic sequence in formation and functional data. Cereal crop productivity represents approximate ly one-third of the worlds f ood supply (faostat.fao.org), with maize (Zea mays L .) producing the largest yield in total bushels. Although grain yield in the USA has increased over the past 80 years with improved, commercially-available hybrids, variation in harvestable yield has also increased (Bruce et al., 2002). Much of this variability is attributable to drought stress (Boyer, 1982; Boyer and Westgate 2004), one of the major causes for yield reductions in mo st areas of the world. The preand early post-pollination phases of maize reproductive development are critical for seed set and subsequent grain yield (Zinselmeier et al., 1995, 2002; Andersen et al., 2002; McLaughlin and Boyer 2004b). During this time, th e plant is especially sensitive to abiotic stresses, which can decrease pollin ation efficiency and/or promot e kernel abortion, thus leading to devastating losses in grain yield worldwide. Selective breeding for traits related to female inflorescence and ear growth have improved pollin ation efficiency and s eed set (Bolanos and Edmeades, 1996; Bruce et al., 2002; Campos et al., 2004). However, relatively little is known about key metabolic processes that influence maize female floret growth and development during the silking and pollination periods. Overall productivity in maize depends la rgely on photosynthetic capacity and the subsequent allocation of carbohydrates to the developing female infl orescence (Boyer and Westgate, 2004; Borras et al., 2007). In addition, nitrogen-based assimilates are essential for kernel growth and have been implicated in maintaining the carbon: nitrogen balance in

PAGE 58

58 developing ears (Seebauer et al., 2004; Martin et al., 2006) Both carbon and nitrogen metabolites can provide a source of signals that regulate gene expression based on nutrient status in sink cells (Koch, 1996, 2004; Seebauer et al., 2 004; Smith and Stitt, 2007). Sink strength in the female inflorescence is dependent on such si gnals to ensure sufficient availability of resources for silk exsertion, successful fertiliz ation, and to carry the developing kernel to maturity. Unfavorable conditions such as drought and shade stress negatively affect photosynthesis in source leaves and typically reduce C availa ble for import by developing ears (Westgate and Boyer, 1985; Zinselmeier et al., 1995; 2002; Setter et al., 2001; Borras et al., 2007). Repression of genes for sucrose metabolizing enzymes at low water potentials can also contribute to this reduction in sink strength and resulting inhibitio n of growth (Andersen et al., 2002; McLaughlin and Boyer, 2004b). In contrast, tassel development appears less affected by reduced photosynthate availability. Ther efore, stresse can result in asynchronous timing of pollen shed and silk emergence, evident as an enhanced Anthesis Silking Interval (ASI) (Bolanos and Edmeades, 1996; Borras et al., 2007). Studies us ing reciprocal crosses determined that the female reproductive tissues also controlled earl y, post-pollination kernel abortion (Westgate and Boyer, 1986). Recent expression profiling studies in maize ha ve focused on effects of drought (Zhuang et al., 2007) and shade stress (Zinselmeier et al ., 2002) on immature reproductive structures. However, the metabolic processes underlyi ng normal ear development during silking and pollination have not been widely investigated. Studies in animals (Newport and Kirschner, 1982; Newman-Smith and Rothman, 1998; Pelegri, 2003) and in plants (V ielle-Calzada et al., 2000; Grimanelli et al., 2005) ha ve suggested that maternal-bas ed regulation dominates during

PAGE 59

59 the initial phases of embryonic deve lopment. In maize, this was ev ident as a shift in maternal-tozygotic transition of transcripts several days after fertilization (Grimanelli et al., 2005). Also, expression studies in barley show ed co-regulation of genes relate d to maternal-specific metabolic processes during early post -pollination development (S reenivasulu et al., 2004). Understanding the mechanisms underlying seed set in maize, and extent to which they are altered by abiotic stress, can be valuable to maximizing yield potential (Bruce et al., 2002; Tuberosa and Salvi, 2006; Vij and Tyagi, 2007) With release of th e maize draft genome sequence in 2008 and recent advances in sequenc e-based technologies for expression profiling (Meyers et al., 2004; Nobuta et al ., 2007; Eveland et al., 2008), the challenges facing researchers are increasingly focused on genome annotation an d functional prediction. Annotation of maize genes is not trivial, however, due to extensive genome duplica tion and approximately one-third of genes in tandemly arrayed gene families (M essing et al., 2004; Messing and Dooner, 2006). In addition, analyses of the maize genome have revealed a high frequency of genes that share 98% identity to a near-identic al paralog and in many cases bot h are expressed (Emrich et al., 2007b). Such sequence similarities complicate re solution of gene-speci fic expression profiles due to potentially confounding effect s from cross-hybridization of cl osely-related transcripts. Functional genomics and expression profiles in complex genomes such as maize depend on gene-specificity and the capacity to disti nguish gene family members from other closelyrelated transcripts with specialized functions. In addition, expression data from individual organs or specific cell types (Ma et al., 2006; Zhang et al., 2007) are essential for relating gene expression to putative function, especially if a ra nge of developmental stages are included. For example, work by Zhuang et al. (2007) showed that drought stress had ve ry different effects on gene expression in the immature tassel verses comp arable tissues in ears. Also, tissue-specific

PAGE 60

60 profiles from shoot apical meristems in maize re vealed novel transcripts and cell-specific gene expression (Zhang et al., 2007). Global transcri pt profiles compiled from tissueand stagespecific analyses are available for Arabidopsis an d have contributed to th e understanding of gene function at the whole-plant leve l (Zimmermann et al., 2004). Such profiles for important crop species would provide a framework for testing effects of environmental or genetic perturbation on gene expression. Advances in expression profiling technologies have also enabled resolution of natural antisense transcripts (Ma et al., 2006; Zhang et al., 2007), alternate sp lice variants (Ner-Gaon et al., 2007), and non-coding RNAs (Lu et al., 2005; Nobuta et al., 2007). Emerging sequencing technologies (Meyers et al., 2004; Jorgoneel et al., 2005; Mar gulies et al., 2005), and their applications for gene expression profiling (Nobuta et al., 2007, Webe r et al., 2007; Eveland et al., 2008), provide quantitative, high-resolution methods for gene-specific analysis. Studies comparing array-based profiling methods with sh ort-tag sequencing platforms such as SAGE (van Ruissen et al., 2005) and MPSS (Liu et al., 2007), have shown ove rall agreement between strategies. Inherent limitations with both arrayand sequence-ba sed profiles, however, indicated that a complimentary approach was most effective. In a previous study, we described a 454-based sequencing approach to transcript profiling that used the specificity of 3'UTRs to dis tinguish between gene family members and other closely-related transcripts on a genome-wide scale (Eveland et al., 2008). Here, we evaluated the potential of 3'UTR profiling by concurrently se quencing a 12-sample library that included four stages of maize ovary developmen t during pollination. Resulting da ta were compared to parallel microarray analyses. The significance of this wo rk is thus two-fold. 1) A gene-specific, sequence-based strategy is evaluated for its use in analysis of the ma ize transcriptome. 2) Sets of

PAGE 61

61 co-expressed genes related to sp ecific functional pro cesses or key metabolic pathways during maize ovary development are identified We co ncluded that the 3'-UTR sequence-based strategy provided an effective complement to microarray analyses for gene-speci fic expression profiling during maize ovary development. Results and Discussion To profile global transcripti onal changes in developing m a ize ovaries, we sampled ears from field-grown plants at two-day intervals ove r four stages of preand early post-pollination growth (Figure 3-1A). Six individual ears were sampled each time point, which included: presilking (4 DBP), silk emergence (2 DBP), time of pollination (0T), and two days post-pollination (2 DAP). A detailed description of sampling procedures appears in Materials and Methods. Ovary-plus-pedicel samples were hand-dissected from florets positioned equatorially around the mid-basal region of each ear. Expression profiles were evaluated from each sample in all possible pair-wise comparisons to samples at other developmental stages using maize longoligonucleotide arrays (University of Arizona). Dye swaps were included in the experimental design (Figure 3-1B). Microarray Data Analysis and Clustering We first used a hierarchical clustering m et hod to test whether overa ll trends in gene expression coincided with development. Medi an signal intensities we re extracted, background subtracted, and log-transformed before fitting an ANO VA model. We identified 856 genes that showed significant changes in expression during development with a False Discovery Rate (FDR) of 10%. The 70-mer sequences corres ponding to all 856 probe s were annotated by BLASTN alignments to cDNAs in Zea mays Gene Index (ZmGI) [tigr.org] and Industry Unigene (IUC) databases. Of these, 606 were classi fied into functional processes based on Gene

PAGE 62

62 Ontologies and 250 matched available cDNA se quences that were either unannotated or contained domains of unknown function. The two-way cluster hierarchy showed that genes clustered into three major groups based on their expression pattern over time. These were 1) expression decreasing during development, 2) increasing during this time, or 3) showing combinations of both (Figure 3-2). Interestingly, only a small proportion of genes responded specifi cally to pollination (evident as differences between 0T and 2 DAP). We compared annotations of genes that showed 1) decreases and 2) increases in expression during development to determine whether functional processes were differentially represented between these two groups (Figure 3-3). In general, expression shifted during development from genes related to signa ling, protein turnover/mo dification, and nucleic acid metabolism to pathways involved in secondary metabolism and defense. Identifying Clusters of Co-Expressed Genes One approach we used to identify groups co-expressed genes was a non-linear clustering analys is described in Qu and Xu (2006). This method is based on fitting a polynomial model for expression of change and has been adapted to analyses of time course datasets (Fung et al., 2008). Here, genes were clustered based on the pa ttern of mean differences in expression over the four developmental stages. Background-correct ed mean signal intensities were compared to a spotted set of negative controls to determine percentage of genes that showed positive signal intensities. The criteria used for probe eliminat ion in the present work was based on a cut-off for positive signal intensities and poor spot frequenc ies, and are described in the Materials and Methods. A calculated, weighted Kappa value ranged from 69-91% across all genes with positive signal intensities and indicated close agreement among biological replicates from a given developmental stage.

PAGE 63

63 In total, 40,686 genes showed pos itive signal intensities and a full vector of least square means estimates after ANOVA. These 40,686 genes were fit to cubic polynomials describing their patterns of change, and a minimum Bayesian Information Criterion (BIC) (grouping that best fit the data) was used to determine an op timal number of clusters (Qu and Xu, 2006). We identified a total of 7 co-expression clusters (Figure 3-4A) and highlighted those genes which showed significant changes in expression during development at an FDR of 20% (Figure 3-4B). Total per-cluster gene numbers a nd percentages of differentially ex pressed genes in each cluster are shown in Table 3-1. In addition, 3,671 of the 40,686 genes (9%) showed greater than 2-fold change during development. A quality check fo r background uniformity revealed areas of high background for some slides. High background is common with the maize long-oligonucleotide platform possibly due to interference fr om complex carbohydrates, enhanced crosshybridization, or inefficient blocking proce dures (J. Gardiner, pe rsonal communication). Because of such inconsistencies in background levels, analyses of genes that showed low signal intensities across development were omitted from analyses done here. Instead, we focused primarily on those genes with large fold-change or significant differences in expression during development. Co-Expression of Genes Related to Common Metabolic Pathw ays The total set of 40,686 genes that showed a pos itive signal was annotated based on their alignment to cDNAs in Zea mays Gene Indices (ZmGI). An notated genes were grouped according to putative functional processes using Gene Ontologi es and, where appropriate, KEGG pathway assignments (Kanehisa et al., 2007). A pproximately 95% of the genes with positive signals were associated with cl uster 2, in which relatively li ttle change in expression was observed during development. Although 1,280 genes in cluster 2 were differentially expressed during development (FDR 20%) and/or had larg e fold-changes, these represented only 3.3%

PAGE 64

64 percent of the total in this group (Table 3-1). We focused on those clusters which had the highest proportion of genes with significant changes in expression or high fold change during development. These were clusters 1, 5, and 7 (Table 3-1). Clusters 1 and 5 were similar both in their e xpression profiles (tended to decrease after pollination) and in functional composition of a nnotated genes. Of the 1,567 genes that grouped with Clusters 1, 5, and 7, we found that 672 were not annotated or were unclassified. We compared functional classes of genes combined fr om clusters 1 and 5 to those from cluster 7 (Figure 3-5). Genes that showed either significant differences in expression (FDR 20%) or large fold changes during development from clusters 1, 5, and 7 are listed w ith their associated functions in Tables 3-2 and 3-3. In general, expression profiles for genes in clusters 1 and 5 tended to show mRNA levels that fluctuated during developmen t and typically either decreased or showed very little change after pollination. We compared profiles for these genes in clusters 1 and 5 to those associated with cluster 7, which tended to show increasing expression during development, and often with largest fold changes between 0T and 2 DAP (Fi gure 3-4B). Comparison of functional processes associated with the different clusters revealed an increase in genes related to nitrogen and amino acid metabolism after pollination relative to t hose involved in carbohydrate metabolism (Figure 3-5). In clusters 1 & 5, Nand C-metabolism are equally represente d. This balance shifts to a 6fold higher proportion of genes related to N metabo lism that are expressed later in development, as shown in cluster 7. Also, genes for protein synthesi s were heavily represented in cluster 7 (24% of the total) while genes related to proteolysis were equally re presented in clusters 1, 5, and 7. Interestingly, a number of genes annotated as protease inhib itors in clusters 1 and 5 showed significant

PAGE 65

65 differences in expression or large fold cha nges during development (Table 3-2). Increased expression of genes for protein synthesis in mate rnal tissues during the later stages of ovary development could be central to growth and deve lopment of filial tissues. Regulation of protein synthesis and turnover has also be en linked to C and N availabili ty and signaling (Palenchar et al., 2004). Consistent with possible increases in C av ailability, genes rela ted to phenylpropanoid biosynthesis were highly represente d in cluster 7. Many of these genes also showed significant differences in expression or large fold changes during development (Table 3-3). Previous work has shown that a number of key genes in lignin biosynthesis are sugar-res ponsive (Rogers et al., 2005). We also noted a shift from prominent expr ession of genes related to cell growth, to those related to differentiation duri ng development (Figure 3-5). Sequence-Based Analyses by 3'-UTR Profiling A gene-spec ific, sequence-based 3'-UTR prof iling strategy was previously described by Eveland et al. (2008). To evaluate the potential of this appr oach in a transcriptome-wide assessment of developing maize ovaries, we compared a parallel, multiplexed 3'-UTR sequencing profile to the microarra y dataset. Sub-libraries were constructed, as described in Chapter 2 (Eveland et al., 2008), for three of th e six replicate ovary samples chosen at random from each of the four developmental stages. The unique key codes in the multiplex adaptors enabled sample recognition during concurrent se quencing of all 12 samples (Figure 3-6). The 12-sample, multiplex library was sequenced using both the 454 Ge nome Sequencer 20 instrument (Margulies et al., 2005) and the recen tly upgraded FLX system (Harkins and Jarvie, 2007). Sequences from both instrument s were analyzed collectively. In total, we obtained 578,088 good-quality, 3'-a nchored sequences that represented 22,920 unique consensus mRNAs, each with two or more reads. Although distribution of reads among

PAGE 66

66 multiplexed biological replicate tended to vary (Figure 3-6), overall differences in read frequency between the four time points were marginal ( 9,382 reads). Of the 22,920 unique consensus sequences identified, 21,273 aligned to cDNA sequences using BLASTN in both ZmGI and IUC databases. Read count for each unique mRNA was normalized to reads per ten thousand (PTT) and used to profile quantitat ive changes in transc ript abundance during development. To determine the extent to which the 70-me r array probes and unique 3'-UTR sequences matched identical cDNAs, we aligned both sets of sequences to IUC cDNAs using BLASTN. In total, 17,882 of the 3'-UTR consensus sequences associated with array probes that showed positive signal intensities (Figure 3-7). There were 5,038 consensus sequences that did not correspond to an array probe based on our analys es. In addition, 1,344 of the 3'-UTR sequences associated with genes that were eliminated fr om the microarray analysis due to undetectable levels of expression or poor spot flags. Th erefore, our 3'-UTR profiling strategy provided resolution of expression changes that were not de tected by microarray analyses. Since sampling of the transcriptome by microarra y hybridization was limited to spot ted oligos, this could explain a portion of the transcripts reso lved only by 454-based sequencing. Comparison of Microarray and Qu antitative 3'-UTR Profiles We com pared expression profiles of gene s detected by both microarrayand sequencebased methods. Quantitative data from the 3' -UTR profiles were in accordance with gene expression trends from the array analyses (Fi gure 3-8). Sequence-base d profiles showed ranges of expression that spanned three orders of ma gnitude and enabled quantitative comparisons of transcript levels between genes (Figure 3-9). Our ability to reliably quantify rarer transcripts was limited by the depth of sequencing for 12 multiplexed samples. In addition, only 3 biological replicates were analyzed by 3' -UTR profiling and ther efore greater variation was encountered

PAGE 67

67 than with the 6 replicates used for microarray analyses (indicated by standard errors in Figure 38). However, in addition to quantitative, gene-specific profiles, the 454 data also provided validation for microarray-based ex pression, especially for those ge nes with high read frequencies (Figure 3-8). To determine whether co-expressed genes of related function from microarray analyses showed similar expression profiles when exam ined using the 454-based method, we analyzed clusters of cell-wall-related genes. The quantitative 3'-UTR profiles for genes involved in cell wall biosynthesis and modification from cluste rs 1, 5, and 7 are shown in Figure 4-9. In addition, we presented in Figure 3-10 an exampl e of four Phenylalanine Ammonia Lyase (PAL) gene family members from the microarray dataset and compared these to their respective read counts from the 3'-UTR profile. In both in stances, the sequencebased method provided quantitative resolution of expression for genes with related functional pr ocesses as well as individual gene family members. Resolution of Near-Identical Transcripts Comparison of 3'-UTR tags to the 70-mer oligo sequences on the array identified 1,394 array probes which matched two or more c onsensus sequences. A number of these disproportionate matches could be explained by mi salignment of a given oligo probe and 3'-UTR sequence due to incomplete genome sequence or proximity of MspI sites to the 3' end. Also, we cannot rule out the possibility that a small proportion of digest s were incomplete or that premature termination of first-st rand cDNA synthesis could have occurred for certain transcripts. We investigated these possibilities by aligning the corresponding 3'-UTR sequences with cDNAs in ZmGI and IUC databases and with maize ge nomic assemblies [MAGI (F u et al., 2005)]. In several cases, we resolved expression profiles fo r two or more near-identical transcripts that

PAGE 68

68 matched a common oligo sequence. Some of th ese closely-related mRNAs were identified by small feature polymorphisms such as a SNP or inde l in the 3'-454-sequence read (Figure 3-11A). Also, putative 3'-RNA processing variants were resolved for genes represented by a single array element (Figure 3-11B). Differen tial profiles for these near-identical transcripts indicated that cross-hybridization could lead to confounding results with some microarray analyses. Conclusions In this work we determ ined that a sequence-ba sed 3'-UTR profiling strategy can be used to analyze global transcriptional changes for twelve multiplexed samples. In addition we showed that this sequence-based method is an effective compliment to microarray analyses and enabled gene-specific resolution of expression among complex gene families and near-identical transcripts. Clustering analyses showed shifts in expression of ge nes related to specific functional processes during maize ovary development. Among these were changes in expression of genes for C and N metabolism, cell growth a nd differentiation, and for protein synthesis. These findings were consistent with enhanced C-import by these developing sink tissues and could provide a basis for understand ing regulation of mate rnal-based adjustment of sink strength at the transcriptional level. Materials and Methods Plant Material Maize (Zea mays L.), W22 inbred plants were fieldgrown at the University of Florida Research farm in Citra, F lorida from April to June, 2005. Ears were sampled at four stages of development: 1) Pre-silk emergence (4 days befo re pollination [4 DBP]), 2) First silk emergence (2 days before pollination [2 DBP]), 3) Time of pollination (0T), and 4) Two days after pollination (2 DAP). Ears sampled at 4 DBP included only those in which longest silks reached approximately 2 cm below the tip of the husk. Ears sampled at 2 DBP had been protected from

PAGE 69

69 prior pollination by inverting small, paper, ear-ca pping bags over them. Silks at this stage had typically elongated about 2 cm beyond the husk and emerged from mid-basal florets. Ears sampled at 0T were collected 1 hour after pollina tion. All four stages were sampled on each of seven consecutive days. Florets were from sa mpled from 3-cm equatorial sections starting approximately 2 cm from the base of each ear. Ovary-plus-pedicel samples were hand-dissected, pooled to 100 mg FW, and imme diately frozen in liquid N2. Ovary-plus-pedicel samples were collected from 6 individual ears at each of the 4 developmental st ages for subsequent analyses. RNA Extraction and Target Labeling Frozen ovary sam ples (1g FW) were pulverized in TRIzol (Invitrogen) using a FastPrep lysis system (Q-BIOgene). Total RNA was extracted as described at www.maizearray.org treated with DNase (Ambion), and quantified by NanoDrop (ND-1000, NanoDrop Technologies). A MessageAm p II kit (Ambion) wa s used to synthesize cDNA from 1.5 ug total RNA using an oligo(dT) primer and subsequent cRNA by in vitro transcription and amplification in the presence of aminoallyl UTP. Each purified, aminoallyl-labeled cRNA sample was quantified by NanoDrop (ND-1000, NanoDrop Technologi es), 6 ug transferred to an RNase-free eppendorf tube, and dried using a SpeedVac (Vacufuge Concentrator 5301, Eppendorf) at room temperature. Indirect coupling of Cy Dy e esters (Cy3 and Cy5 Mono-Reactive Dye, GEHealthcare) to aminoallyl-labeled cRNA and removal of uni ncorporated dye was conducted according to protocols at www.maizearray.org Microarray Slide Preparation and Target Hybridization A total of 12 arrays were used (University of Arizona), each r epresented by a two-slide set (A and B slides). Information on probe se quences and accession numbers are available at www.maizearray.org T he 70-mer oligonucleotide probes were cross-linked to glass slides via a Stratalinker (Stratagene) at 180 mJ All A slides we re prepared and hybridized concurrently

PAGE 70

70 on day 1 and B slides were prepared and hybr idized concurrently on day 2. Slides were washed, with stirring, in filtered 1% SDS for 5 min at room temperature. To remove SDS, slides were dipped 10 times in Milli-Q water, 5 time s in 100% EtOH, then incubated, while shaking, for 3 min in 100% EtOH. Each slide was then placed in an individual 50-mL Falcon tube, dried in a centrifuge at 200 x g for f our minutes, and transferred to a new 50-mL Falcon tube. The following pre-hybridization and hybrid ization procedures were conducted concurrently for three sets of four slides. For each slide, 2 of the 4 developmental stages (4 DBP, 2 DBP, 0T, 2 DAP) were compared. Four 50-mL Falcon tubes, each containing a single slide, were filled with pre-warmed (42 C), filtered, pre-hybridization buffer (4g BSA, 100 mL 20X SSC, 4 mL 10% SDS, 296 mL Milli-Q water), incubated with gentle agitation for 2 min, and then in an oven at 42 C for 45 min. Slides were immediately transferred to Milli-Q water, submerged 5 times, then moved to a fresh volume of Milli-Q for another 5 submersions. Isopropanol was used for the final 5 emersion rinses, followed by drying with compressed nitrogen gas. The labeled cRNA samples were quantified by Na noDrop (ND-1000, NanoDrop Technologies) and 1ug of both Cy3and Cy5-labele d targets were combined for co-hybridization to arrays based on the experimental design in Fi gure 3-1B. Each Cy3/Cy5 target mixture was dried in a SpeedVac (Vacufuge Concentrator 5301, Eppendorf) and then resuspended in 60 uL 1X hybridization buffer (0.55 uL 10% filtered SDS, 13.75 uL 20X filtered SSC, 5.51 uL 50X Denhards, 16.51 uL Formamide, 18.15 uL HPLC grad e water). Hybridizations were carried out in a darkened room. The hybridization mixture was then denature d by heating to 95 C for 3 min, transferred to ice for 30 sec, a nd then centrifuged at 14,000 x g for 2 min to pellet particulates. Each of four MAUI mixers (model SC, BioMicro ) were attached to the glass slides via an Assembly/Disassembly apparatus (BioMicro) and fi rmly sealed. The labeled target mixture (50

PAGE 71

71 uL) was injected slowly into the fill port using a positive displacement Eppendorf Combitip syringe (Brinkmann, Cat No. 022-26595-4). Fill and vent ports were lightly dried with a Kimwipe, sealed with adhesive discs, and 4 s lides were concurrently hybridized in a single MAUI hybridization chamber at 42 C on mixing le vel D. A total of three, 4-bay MAUI hybridization chambers were used for concurrent hybridization of all 12 arra ys. The slides were removed from the MAUI after 14 hours of hybrid ization and submerged in pre-warmed (42 C) Wash-1 buffer (1X SSC, 0.2% SDS, filtered). Mixe rs were then gently re moved, and slides were transferred to a glass frame in stirred Wash-1 buffer for a 4-min incubation. The frame was then transferred to stirred Wash-2 buffer (0.1X SSC, 0.2% SDS, filtered) for 2 min, then to a series of three, 2-min incubations, each in a fresh, stirred re servoir of Wash-3 buffer (0.1% SSC, filtered). Slides were then gently removed and dried with compressed nitrogen gas. All washing was done in a darkened room. Microarray Data Analysis Slides were scanned using an Agilent scanne r setting of 80 PMT. Scanned im ages (TIFF format) were imported, along with the maize arra y annotation GAL file (version 10.1) available on www.maizearray.org into Im aGene Version 6.0. Custom gr ids were fit to each slide for spot finding and signal mean, median, and background intensities were extracted. Based on mean signal differences to negative control spot s present on the slide, 16,274 out of 57,452 genes showed no detectable signal corresponding to a 90% negative control qua ntile, which was used as the cut-off value. For the analysis, a gene at a given developmental stage was defined to be off if four of the six bi ological replicates showed signals at or below background. In addition, expression of a gene was determined to be belo w detectable levels if it was defined as off at all four developmental stages.

PAGE 72

72 The marray package in Bioconductor software was used to generate background intensity plots in R for a quality check of the a rray data. For each combination of slide and dye, genes having no detectable expression, based on the 90% negative control quantile, were deleted before fitting an ANOVA model. Poor spots fl agged for contamination, saturation, or showed negative signal intensities after background subtraction, were also e liminated from the analysis. The signal values were defined to be signal_mean-background_mean. Data were logtransformed to meet the normal assumption for fitting an ANOVA model: log(signal_mean) = dye+treatment+error. Gene-specific ANOVA models were fitted for each probe across all developmental stages. The above model was fitted for log (signal_median) in hierarchical clustering analyses using JMP Genomics. Cluster Analysis using Orthogonal Polynomials The SAS code for the clustering analysis was adapted from Qu and Xu (2006). Our dataset included four sequential developmental stages. Therefore, we could only fit cubic polynomials and the order value was set at 3. The polynomial parameters (developmental stages here) were re-scaled before constructing the orthogonal pol ynomials so that only the pattern of the expression profile was used. Therefore, whether we let development equal (4 DBP, 2 DBP, 0T, 2 DAP) or (1, 2, 3, 4), the same set of orthogonal polynomials was generated. A minimum Bayesian Information Criterion (BIC) was used to determine the optimal number of clusters to best fit the data (Qu and Xu, 2006) Clustering analysis of least square means (LSMEANS) were run using log scale and original sc ale values. Clusters in Figure 3-4 are reported in the original scale to show maximum differen ces in gene-spe cific patterns. Construction of a 454 3'-UTR library Three of the six biolo gical replicate samples from each of the four developmental stages were chosen at random for parallel, 454-based, 3'UTR profiling analyses. Total RNA (5 ug)

PAGE 73

73 from each of the 12 samples was used as templa te for first-strand cDNA synthesis (MessageAmp II [cDNA synthesis only], Ambion) and primed with 6 pmol biotinylated (T12) B-adapter (modified from Margulies et al. [2005]) oligonucle otide. Sub-libraries were constructed as described in Eveland et al. ( 2008). Each sample was tagged with one of 12 unique, four-base multiplex keys in the ligated A-adaptor oligonucle otide. Sub-libraries were then pooled. The desired 5'-A-cDNA-B-3' template strand was eluted with 100 mM Na OH, neutralized, and concentrated on a Qiagen column (Margulies et al., 2005). Sequencing was conducted as per Margulies et al. (2005) using a 454 GS-20 inst rument for one sequencing reaction and the 454 FLX system for a second sequencing reaction. Da ta were combined and reads were filtered for correct ligation junctions and va lidation of the error detection key (Eveland et al., 2008).

PAGE 74

74 Table 3-1. Number of genes a ssociated with each co-expressi on cluster and pe rcentage having significant differences in expression or large fold-c hanges during development. Cluster Total genesa Significant genes (FDR 20%) Percent significant (%)b Genes with large foldchange Percent large fold-change (%)c 1 407 26 6.39 44 10.81 2 38,350 1,280 3.34 3450 9 3 420 11 2.62 6 1.43 4 102 4 3.92 5 4.90 5 639 30 4.69 62 9.70 6 247 19 7.69 6 2.43 7 521 48 9.21 86 16.51 a40,686 genes were defined as on a nd used in clustering analyses. bPercentage of genes per cluster that showed significant differences in expression during development using a FDR of 20%. cPercentage of genes per cluster that showed expression fold changes 2 during development.

PAGE 75

75 Table 3-2. Annotated genes from clusters 1 and 5 that showed signi ficant differences in expression (FDR 20%) or large fold changes ( 2) during development. MZ Assignment P-Value *Foldchange Putative annotation [Species] Function Cluster 1 MZ00003791 0.003 1.595 ATP-binding-cassette transporter [ O. sativa ] Transport MZ00027467 0.013 0.873 Putative phosphate translocator [ O. sativa ] Transport MZ00017851 0.020 0.854 O-methyltransferase ZRP4 [ Z. mays ] 2 metabolism MZ00042055 0.003 0.755 4-coumarate--CoA ligase 4CL2 [ L. perenne ] 2 metabolism MZ00004477 0.010 1.070 N-hydroxycinnamoyl/benzoyltransferase [ A. thaliana ] 2 metabolism MZ00018463 0.023 1.635 Peroxidase [ O. sativa] 2 metabolism MZ00021835 0.022 1.516 N-hydroxycinnamoyl/benzoyltransferase [ A. thaliana ] 2 metabolism MZ00036098 0.005 1.150 Metallothionein-like protein 1 [ Z. mays ] Redox MZ00012431 0.045 0.94 8 Putative laccase [ O. sativa ] Redox MZ00004658 0.0007 1.935 Laccase [ O. sativa ] Redox MZ00030532 0.006 1.999 Protease inhibitor [ O. sativa ] Protease inhibitor MZ00044388 0.035 1.933 Protease inhibitor [ O. sativa ] Protease inhibitor MZ00057049 0.035 1.758 Putative protease inhibitor [ O. sativa] Protease inhibitor MZ00014170 0.397 1.988 Cystatin [ T. aestivum ] Protease inhibitor MZ00040508 0.002 0.800 Cytochrome f [ Z. mays ] Photosynthesis MZ00030872 0.034 1.600 Uclacyanin 3-like protein [ O. sativa ] Photosynthesis MZ00043399 0.001 0.482 Citrate lyase [ A. thaliana ] TCA cycle MZ00024089 0.0003 0.266 Putative DNA topoisomerase II [ O. sativa ] Nucleic acid MZ00003640 0.0008 0.384 Phospholipase D 2 [ O. sativa ] Lipid metabolism MZ00024317 0.122 1.126 Gip1-like protein [ Petunia x hybrida ] Gibberellin MZ00044854 0.002 1.537 Early nodulin 75 precursor-like protein [ O. sativa ] Nodulin-related MZ00056428 0.001 1.383 Beta-expansin [ O. sativa ] Cell wall MZ00018500 0.079 0.733 Xyloglucan endo-1,4-beta-D-glucanase [ H. vulgare ] Cell wall MZ00035817 0.197 1.058 Proline-rich protein [ O. sativa ] Cell wall MZ00040858 0.135 1.028 Proline-rich protein-like [ O. sativa ] Cell wall MZ00041925 0.236 1.376 Proline-rich protein-like [ O. sativa ] Cell wall MZ00030106 0.005 0.270 Beta-glucuronidase [ Z. mays ] Cell wall Cluster 5 MZ00016522 0.004 1.292 Early nodulin 75 precursor-like protein [ O. sativa ] Nodulin-related MZ00024039 0.004 0.856 Photosystem-I PSI-F chain precursor [ H. vulgare ] Photosynthesis MZ00024043 0.005 0.589 Chlorophyll a/b-binding protein precursor [ Z. mays ] Photosynthesis MZ00024066 0.001 0.551 Oxygen-evolving enhancer protein precursor [ Z. mays ] Photosynthesis MZ00025588 0.006 0.751 Proteinase inhibitor [ G. max ] Protease inhibitor MZ00025431 0.002 0.434 Bowman-Birk type trypsin inhibitor [ H. vulgare ] Protease inhibitor MZ00037253 0.015 1.578 Subtilisin/chymotrypsin inhibitor [ Z. mays ] Protease inhibitor MZ00026127 0.005 0.899 Development regulation OsNAC4 [ O. sativa] Transcription MZ00012051 0.0006 0.587 Putative bHLH transcription protein [ O. sativa ] Transcription MZ00016305 0.002 0.301 Putative CTP synthase [ O. sativa ] Nucleic acid MZ00024782 0.004 0.336 Starch branching enzyme I [ Z. mays ] C-metabolism MZ00036659 0.003 0.208 Putative trehalose-6-phosphate synthase [ O. sativa ] C metabolism MZ00013279 0.012 0.714 Putative triosephosphate isomerase [ O. sativa ] Glycolysis MZ00027021 0.0002 0.582 Putative 6-phosphogluconolactonase [ O. sativa ] OPP MZ00027846 0.001 0.611 Auxin efflux carrier protein [ M. truncatula ] Auxin MZ00028568 0.0008 0.287 Viviparous-14 protein [ Z. mays ] ABA MZ00028196 0.004 0.456 Syntaxin [ G. max ] Secretory traffiking MZ00004010 0.016 1.154 Putative serine/threonine kinase [ O. sativa ] Signaling MZ00013531 0.145 0.884 Light regulated protein precursor [ O. sativa ] Signaling MZ00013851 0.856 1.747 Superoxide dismutase [Cu-Zn] [ Z. mays ] Redox MZ00025380 0.008 0.747 Putative fiddlehead-like protein [ O. sativa] Lipid metabolism *Large fold change = ln(2) = 0.7

PAGE 76

76 Table 3-3. Annotated genes from cluster 7 that showed signifi cant differences in expression (FDR 20%) or large fo ld changes during development. MZ Assignment p-Value *Foldchange Putative annotation [Species] Function MZ00023389 0.001 0.979 ABC transporter protein [ O. sativa ] Transport MZ00054418 0.0001 2.128 WRKY transcription factor 45 [O. sativa ] Transcription MZ00019305 0.015 0.706 Homeodomain protein [ O. sativa ] Transcription MZ00023972 0.014 0.939 OsNAC3 protein [ O. sativa] Transcription MZ00024296 0.001 2.606 2-oxoglutarate-dependent oxygenase [ Z. mays ] 2 metabolism MZ00027767 3.61E-05 3.575 Naringenin 3-dioxygenase [ Z. mays ] 2 metabolism MZ00026125 0.0009 0.607 pyridoxal kinase-like protein SOS4 [ A. thaliana ] 2 metabolism MZ00037083 1.70E-05 1.732 Metallothionein-like MT-1 [ Z. mays ] Redox MZ00013363 0.001 1.355 Metallothionein-like protein 1 [ Z. mays ] Redox MZ00035394 0.008 0.764 Cytochrome P450 monooxygenase [ Z. mays ] Redox MZ00020674 0.0001 3.085 Elicitor-inducible cytochrome P450 Redox MZ00035052 0.019 0.827 Pathogenesis related protein-5 [ Z. mays ] Defense MZ00036117 0.051 0.834 Pathogenesis related protein-5 [ Z. mays ] Defense MZ00027578 0.003 0.716 Peptidase family A1 [ O. sativa ] Proteolysis MZ00041508 0.003 0.402 Ubiquitin-conjugating enzyme [ O. sativa ] Proteolysis MZ00014474 0.461 1.893 26S proteasome regulatory particle [ O. sativa ] Proteolysis MZ00014516 0.578 2.818 Putative ubiquitin-protein ligase [ O. sativa ] Proteolysis MZ00015368 0.0007 0.768 Gaiacol peroxidase [ G. hirsutum ] Phenypropanoid MZ00016005 0.0001 1.788 Peroxidase [ Z. mays ] Phenylpropanoid MZ00016987 3.47E-07 2.513 Polyphenol oxidase [ T. aestivum ] Phenylpropanoid MZ00025089 5.45E-05 1.113 Inducible PAL[T. aestivum ] Phenylpropanoid MZ00041601 0.002 1.227 B-glucosidase aggregating factor precursor [ Z. mays ] Phenylpropanoid MZ00057269 0.005 0.572 Caffeoyl CoA 3-O-methyltransferase [ Z. mays ] Phenylpropanoid MZ00015899 0.008 0.806 Cinnamoyl CoA reductase [ Z. mays ] Phenylpropanoid MZ00017982 0.0003 2.034 Phytosulfokine peptide precursor [ O. sativa ] Differentiation MZ00018613 0.0002 0.973 Putative MtN3 [O. sativa ] Nodulin-related MZ00029356 1.33E-06 1.792 Putative MtN3 [O. sativa ] Nodulin-related MZ00043035 0.0001 1.463 Chitinase PRm 3 [ Z. mays ] Nodulin-related MZ00041278 0.013 2.648 Chitinase A [ Z. mays ] Nodulin-related MZ00018945 0.004 0.923 Storage/lipid transfer protein (LTP) [ O. sativa ] Lipid metabolism MZ00026029 0.001 1.229 Putative lipid transfer protein [O. sativa ] Lipid metabolism MZ00038502 0.0001 1.258 Nonspec lipid-transfer protein precursor [ Z. mays ] Lipid metabolism MZ00041611 0.002 1.041 Phospholipid transfer protein [ Z. mays ] Lipid metabolism MZ00041613 3.75E-05 1.825 Phospholipid transfer 9C2 precursor [ Z. mays] Lipid metabolism MZ00019713 0.076 0.779 GDSL-motif lipase/hydrolase protein [ O. sativa ] Lipid metabolism MZ00041203 0.465 0.854 Lipid transfer protein [S. italica ] Lipid metabolism MZ00033558 0.001 0.960 Gibberellin-stimulated transcript 1 like [ O sativa ] Gibberellin MZ00015327 0.01 0.945 Gibberellin-stimulated transcript 1 like [ O sativa ] Gibberellin MZ00018436 0.032 1.136 ACC oxidase [ O. sativa ] Ethylene MZ00024269 9.64E-06 1.739 Proline-rich protein [ Z. mays ] Cell wall MZ00036403 9.78E-05 1.932 Proline-rich protein [ Z. mays ] Cell wall MZ00032935 0.017 0.713 Xyloglucan fucosyltransferase [ A. thaliana ] Cell wall MZ00035277 0.015 1.240 Glycosyltransferase QUASIMODO1 [ A. thaliana ] Cell wall MZ00043252 1.21E-05 1.884 Glutamine synthetase [ Z. mays] N metabolism MZ00043254 3.44E-07 2.119 Glutamine synthetase [ Z. mays] N metabolism MZ00014807 0.002 0.585 Nitrate transporter NRT1-5 [ O. sativa ] N metabolism MZ00025294 0.035 6.653 Putative NAD synthetase [ O. sativa ] N metabolism *Large fold change = ln(2) = 0.7

PAGE 77

77 A Developmental progression 4DBP 2DBP 0T 2DAP 4DBP: pre-silk emergence 2DBP: silk emergence 0T: time of pollination 2DAP: two days post-pollination cy3 cy5 B4DBP 2DBP 2DAP 0TA Developmental progression 4DBP 2DBP 0T 2DAP 4DBP: pre-silk emergence 2DBP: silk emergence 0T: time of pollination 2DAP: two days post-pollination A Developmental progression 4DBP 2DBP 0T 2DAP Developmental progression 4DBP 2DBP 0T 2DAP 4DBP: pre-silk emergence 2DBP: silk emergence 0T: time of pollination 2DAP: two days post-pollination cy3 cy5 B4DBP 2DBP 2DAP 0T cy3 cy5 cy3 cy5 B4DBP 2DBP 2DAP 0T Figure 3-1. Experimental design used to compar e transcriptional profiles during maize ovary development. A) Stages of development sampled at two-day intervals included presilking at 4 days before pollination (4DB P), silk emergence at 2 days before pollination (2DBP), time of pollination (0T) and 2 days after pollination (2DAP)/ post-fertilization. B) Transcript profiles for each developmental stage were analyzed by all possible pair-wise comparis ons and included dye swaps.

PAGE 78

78 Figure 3-2. A heat map generated by a two-way, hierarchical clustering of expression profiles for 856 genes that showed significant ch anges in transcript abundance during development (FDR 10%). Based on this hi erarchy, genes clustered into three main groups: 1) mRNAs that decreased during develo pment, 2) increased during this time, or 3) showed combinations of both.

PAGE 79

79 Figure 3-3. Annotated genes from hierachical clusters 1 and 2 (Fi gure 3-2) that showed either decreased or increased expression during ma ize ovary development were classified into functional categories base d on Gene Ontologies. The most prominent differences in functional representation betw een the two clusters are noted.

PAGE 80

80 Figure 3-4. Co-expression clusters of genes were analyzed af ter fitting expression patterns during development to orthogonal polynomials. Data shown are based on original signal intensity values after removing the ge ne-wise mean (y-axis). Cooridnates for the polynomials (1, 2, 3, and 4) along the x-axis correspond to 4DBP, 2DBP, 0T, and 2DAP, respectively. A) Clus ter assignments for the to tal set of 40,686 genes that showed positive signal intensi ties on the array. Numbers of genes associated with a given cluster are noted. B) Genes that showed significant changes in expression during development (FDR 20%) are highlighted in red. Cluster 1 Cluster 2 Cluster 3 Cluster 4 A Cluster 5 Cluster 6 Cluster 7

PAGE 81

81 Figure 3-4 continued. Cluster 1 Cluster 2 Cluster 3 Cluster 4 B Cluster 5 Cluster 6 Cluster 7

PAGE 82

82 Figure 3-5. Annotated genes in clusters 1, 5, a nd 7 were grouped by functional processes using Gene Ontologies. Results from clusters 1 and 5 (gene expression fluctuates during development) were combined and compared to those from cluster 7 (gene expression increases during development). Prominent di fferences in functiona l processes during development are noted as well as exampl es of those that remained constant.

PAGE 83

83 Figure 3-6. Experimental desi gn and sequencing results for a 12-sample, multiplexed, 3'-UTR library. Unique key codes in the multiplex adaptor were used to distinguish three biological replicates at each of the four stages of maize ovary development. Total reads of consensus sequences (unique mRNAs represented by 2 or more reads) are shown for each of the tags as well as the number of unique transcripts identified at each developmental stage.

PAGE 84

84 Figure 3-7. The 3'-UTR consensus tags (unique mRNAs represented by 2 or more reads) were compared to the 70-mer microarray probe seque nces to determine extent of overlap.

PAGE 85

85 Figure 3-8. Selected genes with hi gh transcript abundances in the 3'-UTR library that were used to validate microarray expression data. Qu antitative 3'-UTR profiles shown represent genes that matched array probes asso ciated with clusters 2,5, and 7. 0 3 2 1 0 Glutamine synthetase (7) 12 4 Gibberellin-stimulated transcript (7) 30 20 10 4 3 2 1 0 Gaiacol peroxidase (7) 6 4 2 Beta-glucosidase aggregating factor precursor (7) Glycosyltransferase-1 (QUASIMODO) (7) 14 0 Trehalose-6-P synthase (5) Putative triosephosphate isomerase (5) Light-regulated protein precursor (5). Early nodulin precursor (5) Inositol polyphosphate 5-phosphatase (2) Auxin-independent growth promoter protein (2) 4 3 2 1 0 8 Proline-rich protein (7) 0 7 0. 1 1 2 1.6 1.2 .8 .4 6 4 2 0 0.6 0 3 2 1 4DBP 2DBP 0T 2DAP Read fre q uenc y 4DBP 2DBP 0T 2DAP Cluster 5 Cluster 7 Cluster 2 0 0 0

PAGE 86

86 Figure 3-9. Quantitative 3'-UTR profiles for ce ll-wall-related genes with matches to array probes in clusters 1, 2, 5, and 7 (see Figure 3-4). Transcript abundance was quantified by read frequency and was plot ted on a log scale to compare ranges of expression during maize ovary development.

PAGE 87

87 Figure 3-10. Transcript profile s of four Phenylalanine ammonia lyase (PAL) gene family members were compared by microarraya nd sequence-based methods. The 3'-UTR profiles showed quantitative comparisons among gene family members during development. Microarray-based profiles ar e depicted as heat maps (light-to-dark denotes low-to-high gene expression).

PAGE 88

88 Figure 3-11. Resolution of quantitative transcript profiles by the 3'-UTR sequencing approach for near-identical mRNAs that matched a single array probe. Microarray-based profiles are depicted as heat maps (l ight-to-dark denotes low-to-high gene expression). A) Examples of near-identic al paralogs identified by small feature polymorphisms in the 3'-UTR. Type of polym orphism is noted in parentheses. B) Resolution of differential transcript profiles for 3'-RNA processing variants. arra y read fre q uenc y (p tt ) 454 A mino oxidase ( SNP ) 4DBP 2DBP 0T 2DAP Ovar y develo p ment 0 3 6 Ao1 Ao 2 10 m y osin-like ( indel ) m y osin-like1 m y osin-like 2 1 2 3 Chi1 Chi 2 Chalcone Isomerase 0 Ovary development 15 5 0 A

PAGE 89

89 Figure 3-11 continued. .

PAGE 90

90 CHAPTER 4 C ALLOCATION AND USE IN DEV ELOPING MAIZE FEMALE FLORETS Introduction The allocation and use of photosynthate is esse ntial to whole plant f unction, particularly during reproductive developm ent. Modulation of carbohydrate synthesis in source leaves and C transport to growing sinks is maintained by a system of nutrient-based checks and balances. Reproduction is a costly process and necessitates abundant photos ynthate for sink establishment and growth (Sturm and Tang, 1999; Paul and Foyer, 2001; Koch, 2004), including respiration (Bustin and Goldschmidt, 1999) and post-pollinat ion fruit development or grain-fill (CruzAguado et al., 1999; Maitz et al., 2000; Borras et al., 2003; Weschke et al ., 2003). Reproductivebased food stuffs such as cereal grains, fruits, and nuts are essential agri cultural commodities in all areas of the world. Therefore, regulati on of carbohydrate allocation and use in reproductive sinks has been a central aspect of breeding a nd crop improvement programs for centuries (Boyer and Westgate, 2004). In maize, the preand early post-pollina tion phases of reproductive development are critical for seed set and subse quent grain yield. Sufficient impor t of C resources is required by the developing female inflorescence for bot h silk (stigma) exsertion and early grain establishment (Zinselmeier et al., 1995; McLa ughlin and Boyer, 2004b; Borras et al., 2007). Optimal source-to-sink C transport is dependent on water availably and therefore maize is most sensitive to drought stress during th e preand early post-pollinati on period (Westgate and Boyer, 1985; Zinselmeier et al., 2002). Although efforts to enhance yield potential have focused on earspecific traits during the silking and pollinati on period, relatively little is known about the mechanisms underlying female floret growth and de velopment at the molecular/metabolic level.

PAGE 91

91 In maize, dry weight accumulation in the fema le inflorescence, or ear, was directly linked to whole plant growth rate unde r conditions analyzed by Borras et al. (2007). Based, on their findings, rate of silk exsertion is strongly infl uenced by photosynthetic cap acity and allocation to the developing ear. In addition, silk elongation is compromised under water-limiting conditions by a decrease in turgor potential (Westgate and Boyer, 1985). Carbon allocation from source to sink is regulated, in part, by acid invertase activities which catalyze the irreversible cleavage of sucrose to glucose and fructose. Isoform-speci fic localization of these sucrose-metabolizing enzymes determines their activity at either symp lastic or apoplastic sites of phloem unloading (Tymowska-Lalanne and Kreis, 1998; Sturm, 1999; Godt and Roitsch, 2006). Invertases are typically characterized by gene families, members of which are differentially regulated based on temporal-, spatial-, and/or isoform-specificity (X u et al., 1996; Godt and Roitsch, 2006; Huang et al., 2007). The conversion of one molecule of sucrose to two hexoses provides substrates for osmotic adjustment and turgor-based expansion. Invert ase activities immediately outside the transport path for assimilates entering sink tissues can thus generate an e nhanced turgor gradient within the phloem from source to sink (Lalonde et al., 2 004; Carpaneto et al., 2005). In addition, both sucrose and its hexose products can act as signaling molecules that regulate gene expression based on the carbohydrate status of the cell (K och, 1996, 2004; Smeekins, 2000; Rolland et al., 2006). Certain invertase isoforms are sugar-responsive at the transcriptional level and regulated by sucrose availability and/or accumulation of their own hexose products (Xu et al., 1996; Huang, 2006). Studies in maize, Arabidopsis, poplar tomato, and rice have each revealed a pair of vacuolar invertase isoforms with reciprocal responses to sugar (Xu et al., 1996; TymowskaLalanne and Kreis, 1998; Fridman and Zamir, 2003; Cho et al, 2005; Huang, 2006; Bocock et al.,

PAGE 92

92 2008). Whether conserved or independently acquir ed in distinct specie s, such reciprocal regulation provides a mechanism for fine adjustment of gene expression to in response to feast or famine conditions (Koch, 1996; Smith and Stitt, 2007). Vacuole-localized, soluble invertase activity is central to maintaining an osmotic gradient for symplastic phloem unloading during the initia l phases of sink cell expansion (Sturm, 1999; Koch, 1996; 2004; Godt and Roitsch, 2006). Post-f ertilization, the activ ities of cell wall-bound invertases provide essential hexoses to the developing embryo across the apoplastic maternal/filial barrier. Both vacuolar and apopl astic invertases have been implicated as key enzymes during the preand early post-pollin ation phases of maize female reproductive development. A mutation in a maize apoplastic invertase, ZmINCW2 resulted in a small kernel, or miniature phenotype (Cheng et al., 1996). Drought-indu ced repression of vacuolar invertases limited carbon influx to the devel oping ovary and thus led to grow th inhibition and/or abortion (Zinselmeier et al., 1995; Andersen et al., 2002; McLaughlin and Boyer, 2004b). Expression of invertases and other sugar-responsive gene s is modulated, in part, by carbohydrate availability during sink establishmen t and growth. In the maize ovary, transient starch reserves are remobilized to promote gr owth and development during periods of low C availability, or famine. Extended periods of famine, such as under drought stress, can deplete essential C reserves a nd ultimately lead to ovary ab ortion (Zinselmeier et al., 1995; Westgate and Boyer, 1986; McLaughlin and Boye r, 2004a). Sucrose metabolizing activity is also associated with developmental progression in sink organs. For example, while a greater hexose-to-sucrose ratio favors growth and e xpansion, higher levels of sucrose promote differentiation and storage (Koch, 1996; 2004; Wint er and Huber, 2000). Accordingly, sucrose synthase activity predominates during sink maturation, generating a one-to-one, sucrose-to-

PAGE 93

93 hexose conversion, with UDPG as an additional product that can serv e as a substrate for cell wall biosynthesis (Koch et al., 2000; Koch, 2004). Carbohydrate sensing is also mediated by nitrogen availability and metabolism (Koch, 1997; Coruzzi and Bush, 2001; Cooke et al., 2003; Palenchar et al., 2004). As such, both C and N metabolite signals, and their relative ratios, can affect whole plant source/sink relations. Nitrate, and its assimilated products such as gl utamine, can act as signals for N availability (Coruzzi and Bush, 2001; Foyer et al., 2003). Gln has been shown to feedback-inhibit nitrogen uptake and reduction, while C metabolites prom ote up-regulation of genes involved in N acquisition and metabolism (Coruzzi and Bush, 2001). Therefore, in growing C sinks, genes for N assimilation are up-regulated, as are those involved in proteolysi s, lipid synthesi s, and storage (Koch, 1997; Foyer et al., 2003). In addition, s ource-sink adjustment of resources may include endogenous hormonal cues and/or responses to en vironmental stimuli. Accordingly, regulation of acid invertases at the transcriptional level by sucrose (Xu et al., 1996; Koch, 1996; 2004), has also been shown to respond to hor mones such as abscisic acid (Trouverie et al., 2003), ethylene acid (Linden et al., 1996), cytokinin (Lara et al., 2004), and auxin (Long et al., 2002). Understanding nutrient-based modulation of si nk strength and associated links to developmental progression will ultimately enhance yield potential in maize (Borras et al., 2003; Boyer and Westgate, 2004; Barnabas et al., 2007). Much of the work to date investigating sugarbased effects on developmentaland hormona l-based regulatory networks has focused on Arabidopsis as a model system. Extending key in sights from these studies to crop systems will be important in determining areas of functiona l significance for agricult ural applications. Mutations in key regulators that link carbohydrat e metabolism and development have revealed phenotypes of agricultur al relevance, e.g. RAMOSA3 (encoding a trehalose-phosphate

PAGE 94

94 phosphatase) in maize (Satoh-Nagasawa et al., 2006) In addition, genotype differences in rate of exsertion and yield performa nce under source-limiting conditions have been identified (Bruce et al 2002), however the underlying mechanisms remain obscure. In this work, we compared C-allocation (as determined by dry weight deposition and metabolism) among individual floral organs duri ng preand early post-pollination maize female reproductive development. Spatial and tempor al regulation of sink strength and sucrosemetabolizing activity of a speci fic vacuolar invertase were tested under normal growth conditions. An apparent shift in sink strength to the developing ovary during the pollination period coincided with co-expres sion of genes related to pheny lpropanoid biosynthesis and N metabolism from profiling experiments in Chapter 3. Data presented here provide evidence for changes in gene expression that associate with C allocation and related expansion or differentiation processes in developing maize ovaries. Results Staging of Pre-Pollinat ion Floral Development To establish a foundation for m olecular and metabolic analyses of developing maize female florets, we characterized a series of individual growth stages that occur prior to pollination. Immature ears were sampled from field-grown, W22 inbred maize plants at two-day intervals from time of whole-ear silk expansion to two days post silk emergence. Florets from the mid-to-basal equatorial region of a given ear (Figure 4-1A) were hand-dissected into ovaryplus-pedicel, with and without su btending floral organs (lemma, palea, and glumes), and silk. Physical characteristics and expression of a developmental marker were used to assign dissected florets to a stage from 1 to 7, each of which corresponded to sequential points of ear development from 12 days before pollination (DBP ) to the time of pollination (Figure 4-1B). Silks of stage 1 florets were approximately 6 cm in length and elongated about 1.7 centimeters

PAGE 95

95 per day during pre-pollination development. Ov ary fresh weight increased exponentially, but very slowly, and in proportion to linear in creases in silk length (Figure 4-1C). In addition, expression of ZAG2 a floral homeotic gene, was used as a molecular marker for development. Quantification of Zag2 mRNAs was used to validate the assignment of floret samples to a given stage (Figure 4-1D). ZAG2 is expressed specifically in carpels and shares homology to the APETALA floral identity C gene in Arabidopsis (Schmidt et al., 1993). Based on these combined physical and molecular mark ers, 6 individual ears at each of the 7 developmental stages were used in subsequent analyses. Pre-Pollination Carbon Allocation and Sucr ose Use in Individual Floral Organs One goal of this work was to determine parti tioning of carbon and wate r as evident in dry and fresh weight accumulation among individual floral organs prio r to pollination. We quantified dry and fresh weights for 1) ovary-p lus-pedicel, 2) ovary-plus-pedicel including subtending floral parts, and 3) silk over a de velopmental time course (Figure 4-2A). Because individual organs could not be effectively separated at stages 1 and 2, we excluded them from this analysis. Per-floret weights were based on measurements of six individual equatorial florets from each of six separate ears at a given deve lopmental stage. Dry and fresh weights for subtending floral parts were determined by subtra cting respective ovary-plus-pedicel weights. Deposition of dry weight in the silk was nearly linear and in creased a total of three-fold during pre-pollination growth. Fresh weight accumulation was also linear and paralleled increases in silk length during de velopment as shown in Figure 4-1C A three-fold gain in fresh weight was observed in the subt ending floral organs immediatel y prior to pollination, however dry weight accumulated to a lesser degree. The resulting increase in rati o of fresh/dry weight was linear during the period 4 DBP to 0T (Figure 4-2B) and is consistent with an osmoticallydriven expansion of subtending floral organs. Also compatible with turg or-based silk elongation

PAGE 96

96 was the maintenance of a substantial fresh/dry wei ght ratio in silks during pre-pollination growth (Figure 4-2B). The prominence of dry and fresh weight gain indicated s ubstantial sink strength in silks prior to pollination. Although C accumulate d to a lesser degree in the ovary-plus-pedicel during pre-pollination growth, rate of fresh weight increase was maintained proportional to that of silk elongation Temporal and Spatial, IVR2 -Based Sucrose Use in Developing Female Florets Based on th e linear increase in fresh-to-dry we ight ratio in subtendi ng floral parts, we hypothesized that hexoses generate d via soluble invertase activity could provide substrates for osmotic-based expansion in these tissues. We quantified the expres sion of two vacuolar invertase isoforms, IVR1 and IVR2, in ovary-plus-pedicel samples, with and without subtending floral parts, during this prepollination period of expansion (Figure 4-3A). Levels of Ivr2 mRNAs were significantly higher in subtending floral parts than in samples of ovary-pluspedicel alone, whereas levels of Ivr1 transcripts were similar between fractions. Consistent with expression of Ivr2 hexose-to-sucrose ratios were at least 2-fold higher in whole florets compared to ovary-plus-pedicel samples at each stage (Fig ure 4-3B). These result s suggest that soluble invertase activity in subtending floral parts is predominantly IVR2-specific. Since rapidly growing silks also maintained hi gh fresh/dry weight prior to pollination, we hypothesized that IVR2 might provide essential hexoses for tu rgor-based expansion in silks. We observed a linear increase in ce ll size measured at the base of the silk during pre-pollination growth (Figure 4-4A ) These data are consistent with wo rk of Westgate a nd Boyer (1985) who suggested that expansion, rather than cell division, was the me chanism behind silk elongation. We quantified Ivr2 mRNAs in whole silk tissues sampled during development in three separate field seasons (Citra, Florida fall plantings 2003, 2004, and spring planting 2004) (Figure 4-4B). All silks were collected at 9AM to exclude possible variation due to diurnal effects (Figure 4-5).

PAGE 97

97 During post-emergence stages, only silks that had exserted from the husk were collected. Levels of Ivr2 mRNAs tended to increase during early stages of development and had dropped to nearly undetectable limits by two days after pollination. Previous work had also demonstrated a coincident cessation of silk elongation af ter pollination (Westgate and Boyer, 1985). Interestingly, decreased levels of Ivr2 mRNAs were apparent at silk emergence from the husk, but prior to pollination. This finding indicate d that pollination was not the primary signal for repression of IVR2. In addition, decreases in soluble acid invertase act ivity paralleled those of Ivr2 transcript abundance (Figure 4-4C). A decrease in the hexose/sucrose ratio by approximately 50% was shown between time of pollination and two days post-pollination in silks (Figure 4-4D). Although the latter was obs erved post-pollination, it may well have resulted from pre-pollination events and been more depe ndent on these than previously recognized. Previous studies have provided evidence for diurnal regulation of invertase-mediated expansion (Gonzalez et al., 2005). Also, Westgate and Boyer (1985) showed that rate of silk elongation was highest pre-dawn. To evaluate the sugar compositi on of the expanding silk tissue over a diurnal time course, we collected newly-emerged silks over a continuous 24 hours including time points at 10 AM, 4 PM, 10 PM, 2 AM, and 6 AM. Sucrose and hexose concentrations were quantified a nd results showed that hexoses we re most abundant in silks at 6AM, or pre-dawn (Figure 4-5). We also compared expression profiles of Ivr2 and a cell-wall bound invertase, Incw2 during preand early post-pollina tion floret development. While Ivr2 mRNAs were maintained at relatively constant levels in ovaries, Incw2 expression is significantly enhanced postpollination and is specific to the developing ovary-plus-pedicel (Figure 4-6). The observed spatial and temporal expression of Incw2 is consistent with its pr oposed function in providing

PAGE 98

98 hexoses to the symplastically-isolated, developi ng filial tissues (Cheng et al., 1996; Kladuik et al., 2004). Post-Pollination C Accumulation in Pedicels To approxim ate the extent of change that might be occurring during C deposition in developing ovaries and pedicels after pollination, we quantified dry and fresh weights for 1) ovary-plus-pedicel, 2) ovary only, and 3) pedice l only. Ovary-plus-pedi cel samples were left intact or further dissected into ovary and pedice l fractions with a razor blade. Pedicel samples included the transfer region. Increases in dry and fresh weights were evident in ovary-pluspedicel during early post-pollina tion development, however dry weight in the pedicel alone almost doubled between 2 and 4 days after pollina tion (DAP) (Figure 4-7). Sucrose and hexoses were also quantified in the a bove samples (Figure 4-8). While hexose-to-sucrose ratio was maintained in ovary tissues (developing kernel post-fertilization), sucrose accumulated to high levels in the pedicel. The resu lting decrease in the hexose-to-sucrose ratio in the pedicel would theoretically favor differentia tion as opposed to expansion (Winter and Huber, 2000; Koch, 2004). Co-Expression of Genes Related to C Sink Development Genom e-wide expression data fr om developing maize ovaries were analyzed from parallel microarrayand sequence-based pr ofiles in Chapter 3. We identif ied co-expression clusters of genes involved in C and N metabolism that s howed high fold changes and/or significant expression differences at a FDR of 20% duri ng development. Genes annotated as having functions in sucrose and starch metabolism te nded to associate with a wider range of coexpression clusters than those related to N meta bolic processes (Figure 4-9). In general, Crelated metabolic genes also showed lower fold -changes over development than those associated with N metabolism. Genes for N metabolism we re highly represented in cluster 7, which

PAGE 99

99 included those genes up-regulated during later stages of ovary development (Figure 4-9B). Since lipid metabolism is, in part, regulated by C: N status (Koc h, 1997; Foyer et al., 2003), we evaluated co-expression pr ofiles of genes involved in lipid metabolism. The developmental trends observed from the expression data were indicative of a shift from lipid degradation to lipid transfer, synthesis of lipid-based metabolites, and storage (Figure 4-10A ). These data were further supported by metabolite profiles for linol eic acid (18:02), linol enic acid (18:03), and oxophytodienoic acid (OPDA) using GC/MS-ba sed quantifications (Figure 4-10B). Post-Pollination Lignin Biosynthesis in Pedicels In addition to enhanced dry weight and su crose accum ulation, the rigidity of pedicels increased rapidly after pollina tion. Microarrayand sequencebased expression profiles in Chapter 3 identified co-expressed clusters of genes related to phenylpropanoid biosynthesis. Expression profiles associat ed with cluster 7 tended to incr ease over development in ovary-pluspedicel samples and included a number of gene s annotated as having functions related to phenylpropanoid biosynthesis (Tab le 3-3). The developmental expression profiles for select genes related to phenylpropanoid biosynthesis an d their co-expression cluster assignments are shown in Figure 4-11. In addition, quantificatio n of cinnamic acid using GC/MS showed a linear increase of this lignin precursor in the ovary-plus-pedicel fractions, while levels of salicylic acid remained unchanged (Figure 4-12A). Lignin acc umulation in the pedi cel was visualized by phloroglucinol-HCL staining of fr esh, longitudinal sections of de veloping florets during preand early post-pollination (Figure 4-12B). Discussion Our findings dem onstrated that the shift fr om preto early po st-pollination maize reproductive growth coincided with a concomitant shift in resource a llocation among individual floral organs. We observed a temporal and spa tial regulation of sink stre ngth as determined by

PAGE 100

100 accumulation of dry and fresh weights. Such re gulation was consistent with expansion and/or differentiation in specific floral organs during female floret development. The substantial deposition of dry weight observed in silks be fore pollination is not surprising given the reproductive advantage of concurrent anther and silk exsertion (t ranslating to a short AnthesisSilking Interval) (Bolanos and Edmeades, 1996; Borras et al., 2007). Also, accumulation of fresh weight in subtending floral parts duri ng the pollination period was analogous to petal expansion. Although a functional role for the subt ending floral structures (lemma, palea, and glumes) has not been described, post-pollination re mobilization of C assimilates from these sinks is one possibility. The proportionate relationshi p maintained between ovary-plus-pedicel fresh weight and silk length over the course of pre-pollination growth suggested spatial regulation of photoassimilate import during development. The elevated hexose-to-sucrose ratio and turgor-based expa nsion in silks and subtending floral parts coincided with expres sion of a vacuolar invertase, IVR2. The IVR2 gene in maize is up-regulated by C availability at the transcriptional level (Anderse n et al., 2002; Gonzalez et al., 2005) and has been implicated as a key enzyme fo r establishing and maintaining sink strength in developing ovaries (Zinselmeier et al., 1995; Andersen et al., 2002; McLaughlin and Boyer 2004a). Previous findings showed that maximum rates of silk expansion occurred pre-dawn (Westgate and Boyer, 1995). Our data were cons istent with these findings and indicated that hexoses were most abundant at a 6 AM sampli ng time. One possible explanation for a rapid burst of expansion at the end of the dark period is remobilization of transitory starch in the silks. Preliminary evidence for starch accumulation in s ilks included high levels of a leaf-specific ADP-G Pyrophosphorylase mRNA and localization of iodine-sta ined starch to chloroplasts along the length of the silk (A. Evela nd, unpublished). Previous work has shown that starch reserves in

PAGE 101

101 ovary tissue were also remobilized during ti mes of carbohydrate starvation, presumably aiding maintenance of normal growth and development (McLaughlin and Boyer, 2004b). Theoretically, remobilization of this transitory starch could provide substrates for IVR2 activity in silks. An increase in dry weight deposition by devel oping ovary-plus-pedicels coincided with the time of pollination. The observed accumulation of sucrose in pedicels could by key to promoting vascular differentiation. Also, evidence indicates that lignin biosynthesis is a strong sink for C assimilates in the pedicel immediately after pol lination. First, genome-wide expression data from analyses in Chapter 3 identified co-expres sed clusters of lignin biosynthetic genes that showed high-fold change during post-pollinati on development. Second, linear increases of cinnamic acid in the ovary-plus-pedicel suggest ed flux of C-assimilates to phenylpropanoid biosynthesis rather than to salicylic acid conve rsions. Finally, phloroglucinol staining of 4-Olinked hydroxycinnamyl aldehydes provided a visu al appraisal of pedicel-specific lignin accumulation over time. Lignin biosynthesis is also irreversible and energy-intensive, thus an irretrievable, high-cost investment in plant grow th and development. Therefore, regulation of C flux to lignin biosynthesis is pr obably modulated on a number of le vels including nutrient status, hormone signals, and developmental cues. Genes involved in lignin biosynthesis, for example, respond to sugar signals at the tr anscriptional level (R ogers et al., 2005). Su ch signals could be central to maintaining a spatial and temporal balance between cell expansion and differentiation during sink development. Further analyses of lignins and their composition using GC/MS will provide quantitative evidence for th e degree to which C is utilized in lignin biosynthesis during pedicel development. Recent work has shown that an altered C: N balance can modulate the extent of phenylpropanoid biosynthesis (Fritz et al., 2006). Analysis of genome-wide transcript profiles

PAGE 102

102 during ovary-plus-pedicel development (Chapter 3) also showed that genes involved in N assimilation were co-expressed with those for lig nin biosynthesis. These expression data are consistent with metabolic trends observed in growing C-sinks (Koch, 1996, 1997; Cooke et al., 2003; Foyer et al., 2003). Shifting the C: N balance in favor of C assimilates could result in upregulation of nitrogen assimilation via sugar-re sponsive expression of glutamate synthase and glutamine synthetase (Koch, 1996, 1997; Fritz et al ., 2006; Martin et al., 2006). Studies have shown that key factors for kernel set in mai ze include N assimilation during the early postpollination period, particularly an up-regulation of glutamine synthetase isoforms (Seebauer et al., 2004; Martin et al., 2006). During the pre-pollination peri od of expansion by silks and subtending floral organs, C deposition is comparatively less in the ovary and pedicel. Remobilization of C resources from senescing floral organs to the developing ova ry-plus-pedicel could pr omote C-based signals during early post-pollination grow th. Increasing C relative to N would favor up-regulation of genes involved in nitrogen a ssimilation, lipid synthesis and storage, and phenylpropanoid biosynthesis. Genes associated with such processes tended to s how co-expression during development while those involved in starch and sucrose metabo lism showed a wider range of expression profiles and smaller fold-changes. Conclusions Individual stages of prea nd early post-pollinatio n m aize female floret development were characterized together with estimations of C-allo cation among individual fl oral organs. Prior to pollination, the major sinks appeared to be the si lks and the floral structures subtending the ovary. During early post-pollination development, dry weight accumulated in the pedicel and C was allocated predominantly to the ovary-plus-pedi cel. This shift in sink strength was consistent with differential expression of a vacuolar and cell-wall invertase, respectively. In addition, there

PAGE 103

103 was an apparent pre-to-post-pollin ation shift from accumulation of so luble sugars to lignin in the pedicel. These data indicated that a shif t in balance occurs between expansion and differentiation during maize ovary development. Analyses of changes in co-expressed genes associated with this shift suggest possible cont ributions by Cand N-base d sensing mechanisms. Materials and Methods Plant Material and Sampling W22 inbred m aize plants grown in various fi eld seasons and under greenhouse conditions. Data are from pre-pollination, developmental anal yses of material grown under field conditions at the Tropical Research and Education Center (TREC), Homestead, Florida, winter planting. Three ears were collected daily at 9 AM over the course of two weeks in March, 2005. Ear development was staged based on anatomical char acteristics which included ear length, silk length, and floret fresh weight. Fl orets were dissected from the mi d-to-basal equatorial region of each ear, separated from silk, and both immediately frozen in liquid nitrogen for further analyses. Ovary-plus-pedicel samples were hand-dissected and collected sepa rately from ears 6 DBP to 0T (stages 5 to 7). A total of six ears that showed greatest uniformity in growth characteristics and expression of the ZAG2 molecular marker were used in s ubsequent analyses. Post-pollination analyses were repeated in material from both a summer field (2005) grown in Citra, FL and from a winter greenhouse (2008) in Gainesville, FL. Th e latter were grown in 14 pots with daylight extended to 12 hrs. Quantification of mRNAs by Real-Time RT-PCR Quantitative analyses of mRNAs by real -tim e RT-PCR used either an ABI 2600 Instrument or the ABI StepOne Plus (Applied Bios ystems). Freshly dissected silk and floret samples were immediately frozen in liquid nitrogen and homogenized in 1 mL TriZol (Invitrogen) using a FastPrep lysi s system (Q-Biogene). Total RNA was extracted as described

PAGE 104

104 in Materials and Methods (Chapter 3). Gene-specific primers and Taqman probes were designed for IVR1, IVR2, and INCW2 using PrimerExpress software (ABI): Ivr1 FP: 5'CGGCAGCCTCCAAACTTTC-3', Ivr1 RP: 5'-CCCGTATACTCTCTTAACCAGATCGT-3', Ivr1 probe: 5'-TCTGCCAAGACGAGGTCAGGGCA-3'; Ivr2 FP: 5'TGGCTACTACTTATCTTCCAGCACTAGT-3', Ivr2 RP: 5'-TGCATGATGCGGTGCTACA3', Ivr2 probe: 5'-CATGTACAACTAGAGGCTACACGTCTTCCCACTG-3'; Incw2 FP; 5'GGTCACTCTCAGGAACAGGGTAA-3', Incw2 RP: 5'-AGCCTGTGCCGTTTGTATCC-3', Incw2 probe: 5'-CACCTCGACGTGA TGTCCTGCCTTG -3'. For each Taqman reaction, 100 ng of total RNA was used as template and run in triplicate. Data were analyzed using a CT method by comparing all values to a single sample and calculating RQ or relative percent. A SYBR green method was used to quantify Zag2 transcripts using gene-specific primers: Zag2 FP: 5'-TTGGCTTCCATGACCTTGCT-3'. Zag2 RP: 5'-GCACAAGGAGAATCACACACAAA3'. For SYBR green analyses, 1 ug total RNA was converted to cDNA using a High-Capacity cDNA RT reaction (ABI) and random primers. Pa rallel amplification of a VIC-labeled 18S rRNA control (ABI) was used to normalize all mRNA levels Soluble Sugar Extraction and Quantification Concentrations of sucrose, glucose, and fruc tose were determined using either enzymatic analysis or chemical separation with High Performance Liquid Chromatography (HPLC). Soluble sugars were extracted from 200 ng whol e silk tissue after homogenization in liquid nitrogen and incubation in 1 mL extrac tion buffer (200 mM KOH, .08% Triton X). A Sucrose/D-Glucose/D-Fructose UV method (R-Bi opharm, Roche) was used to quantify absolute amounts of sucrose, glucose, and fructose. For HPLC analyses, a Carbohydrate Analysis Column (AMINEX Carbohydrate HPX-87C, BioRad) was used to separate soluble sugars. Preand post-pollination floret, ovar y, and pedicel samples were wei ghed and homogenized in 2 mL

PAGE 105

105 tubes with 500 uL boiling 80% ethanol using a FastPrep lysis system (Q-Biogene). Prior to homogenization, samples were spiked with a xylos e control. Soluble sugars were separated by centrifugation for 10 minutes and the ethanol extraction was rep eated three times. The 1.5 mL supernatant sample was lyophilized and resusp ended in 900 uL HPLC grade ddH20. Filitered samples were then analyzed by HPLC and peak ar eas were calculated on the basis of a standard curve. Assay for Soluble Invertase Activity Whole silk tissue (200 m g) was homogenized in liquid nitrogen and extracted on ice in 1 mL extraction buffer (50 mM MOPs-NaOH [pH 7.5], 5mM MgCl, 1mM EDTA, 0.05% w/v Triton K-100, 2.5 mM DTT, 0.1 mM DMSF, 1% PVPP). Soluble protein samples were separated by centrifugation and tr ansferred to dialysis bags (MWCO 50 kD). Dialysis was carried out overnight in three changes of 1XPBS buffer to remove low-molecular-weight invertase inhibitors. Total prot ein concentration was determined by Bradford (BioRad) using a BSA standard curve. Soluble acid invertase activ ity was assayed in a 1:3 dilution of total soluble protein: 17mg/mL sucrose in 50mM sodium citrate buffer (pH 5) at 37 C for 5 minutes. Total glucose was quantified in relati on to a glucose standard (GO A ssay Kit GAGO20-1KT, Sigma). GC/MS Quantification of Metabolites Cinnam ic acid and salicylic acid were quantifie d in pre-pollination w hole floret and ovaryplus-pedicel samples. Samples were frozen i mmediately after dissection in liquid nitrogen and extracted in 2 mL tubes using a FastPrep lysi s system (Q-Biogene). Metabolite extraction, methylation, and isobutene-chemical ioniza tion gas chromatography/ mass spectroscopy (GC/MS) analysis were carried out according to Schmelz et al (2003) at the USDA-ARS Center for Medical and Veterinary Entomology, Chemistry Unit, Gainesville, FL.

PAGE 106

106 Phloroglucinol-HCL Staining Fresh longitudinal sections of florets at 2 DBP, 0T, 2 DAP, a nd 4 DAP were stained with phloroglucinol/ethanol (96%)/ HCL (37%) for five minutes washed with ddH20 and photographed under a dissecting scope (Leica MZ 12.5).

PAGE 107

107 Figure 4-1. Definition of devel opmental stages for pre-pollinated maize female florets by physical growth parameters, anatomi cal features, and expression of a ZAG2 molecular marker. A) Florets were sampled and hand-dissected from the mid-to-base region (boxed in red) of a given ear and B) assigned to growth stages 1 to 7 (also corresponding in this study to days befo re pollination [DBP] shown immediately below the stages indicated). C) Stages were determined by fresh weight of ovaryplus-pedicel in relation to silk length and D) Relative abundance of Zag2 mRNAs as a molecular marker for development. B A silks emerge pollination stage DBP 1 2 3 8 4 6 4 2 0 7 5 6 10 12 0 10 20 30 40 0.00 0.01 0.02silk length (cm) ovary FW (g) C D stage DBP 1 2 3 8 4 6 4 2 0 7 5 6 10 12 0 Zag2 1 2 3 relative mRNAs (RQ) silk pedicel ovary subtending floral

PAGE 108

108 Figure 4-2. Partitioning of carbon and water to dry weight and fresh weight accumulation among individual organs of developing maize female florets. A) Dry and fresh weights were quantified on a per-floret basis. Note the prominence of dry weight gain (and approximate sink strength) of the rapidly el ongating silk. B) Silks maintained a high fresh-to-dry weight ratio over the course of pre-pollination elongation. Also, freshto-dry weight increased linea rly during rapid expansion of subtending floral parts just prior to pollination. A g rams 0 0.045 0.023 dry weight silk subtending floral ovary + pedicel 0.4 fresh weight 0.2 silk subtending floral ovary + pedicel B DBP 6 stage 7 6 4 5 3 8 4 2 0 12 8 4 0 fresh/dr y wei g ht DBP 6 stage 7 6 4 5 3 8 4 2 0 ovary + pedicel subtending floral silk 0

PAGE 109

109 Figure 4-3. Spatial and temporal expression of soluble acid invertases and changes in sugar composition in maize florets just prior to pollination. A) Relative mRNA levels for two vacuolar invertase isoforms, Ivr2 and Ivr1 in ovary-plus-pedicel (white bars) and whole florets (grey bars). B) Sucrose (dk bars) and hexoses (lt bars) were quantified in ovary-plus-pedicel and whole florets prior to pollination. A 5 6 7 0 2 4 5 10 Ivr 0 2.5 5 Ivr relative mRNA abundance (RQ) stage DBP 0 ovary + pedicel whole floret ovar y + p edicel whole florets4 2 0 5 6 7 stage DBP 0 10 0 10 20 30 30 20 g m g FW-1 B sucrose hexoses

PAGE 110

110 Figure 4-4. Activity of the vacuolar invertase, IVR2, is associated with turgor-based expansion in rapidly elongating silks. A) Epidermal cell size at the base of rapidly elongating silks increases linearly over a pre-pollination time course. B) Relative levels of Ivr2 mRNAs duing three different field seasons. C) soluble acid invertase activity and D) Hexose-to-sucrose ratio in whole s ilks preand early post-pollination. B 8 6 4 2 0 +2 3 4 5 6 7 stage DBP A cell size (mm) hexose/sucrose D 0 4 8 124 DBP 2 DBP 0T 2 DAP 0 nmole glc ug protein1 hr1 .03 .06 .09 C Fall 2003 Spring 2004 Fall 2004 mRNAs (relative %) 0 0.4 0.8 1.2 100 50 0

PAGE 111

111 Figure 4-5. Diurnal changes in sucrose and hexose levels in rapidly expanding silks on the day of their emergence from husks. g 100g1 10am 6am 4am 1am 10pm 4pm sucrose 3 2 1 0 hexoses

PAGE 112

112 Figure 4-6. Relative abundance of mRNAs for vacuolar and cell-wa ll inverase isoforms in maize female florets during the pretopost-pollination period. While Ivr2 mRNAs accumulated in the subtending floral parts, Incw2 expression was specific to the ovary-plus-pedicel fraction and increased after pollination. 100 100 50 0 0 50 days before pollination days after pollination -6 -4 -2 0 2 4 Ivr2 Incw2 mRNAs (relative %) whole florets ovary + pedicel

PAGE 113

113 Figure 4-7. Carbon deposition and re lative water content in maize ovaries and pedicels during post-pollination development. Ovary tissue (or developing kernel post-fertilization) was separated immediately above the pedice l as shown in the diagram above. Dry and fresh weights were quantified for ovary-plus-pedicel, ovary, and pedicel samples on a per-floret basis. g ram.004 0 ovary + pedicel ovary pedicel .003 .002 .001 Dry weight 0T2DAP4DAP .01 .02 .03 ovary + pedicel ovary pedicel Fresh weight 0 ovary pedicel

PAGE 114

114 Figure 4-8. Sugar composition of maize ovaries and pedicels during early post-pollination development. Hexoses and sucrose were quantified in separa ted ovary tissue (or developing kernel if post-fe rtilization) and pedicels during early post-pollination development. 2500 1500 1200 1800 500 0T 2DAP 4DAP sucrose hexoses pedicel 0 600 sucrose hexoses ovary/young kernel g mg FW-1 0

PAGE 115

115 Figure 4-9. Co-expression profiles for genes re lated to C and N metabolism during ovary and pedicel development from genome-wide micr oarray analyses. A) Genes annotated as having functions in sucrose and starch me tabolism tended to associate with a wide range of co-expression cluste rs. B) Genes involved in N metabolism and assimilation tended to show larger fold -changes during development and associate predominantly with cluster 7 (expression increases after pollination).

PAGE 116

116 Figure 4-10. Co-expression of genes related to lipid metabolism and abundance of lipid-based metabolites in developing maize ovaries. A) Select genes annotated as having functions related to lipid metabolism in co-expression clusters 1, 2, and 7 from microarray analyses (Chapter 3). Relati ve expression profiles in the developing ovary-plus-pedicel are depicted and select genes are listed. B) Quantitative metabolite profiles are shown for linoleic acid (18:02), linolenic acid (18:03), and oxophytodienoic acid (OPDA) in whole fema le florets (12 DBP to 0T) and ovaryplus-pedicel fractions (4DBP to 0T) during pre-pollination development. A

PAGE 117

117 Figure 4-11. Expression prof iles for genes involved in phe nylpropanoid biosynthesis in the developing maize ovary-plus-pedicel.

PAGE 118

118 Figure 4-12. Accumulation of li gnin precursors and pedicel-locali zed staining of lignin in the developing maize ovary. A) Cinnamic acid and salicylic acid were quantified in whole female florets and ovary-plus-pedice l during pre-polliantion development. B) A phloroglucinol stain for cinnamylaldehyde residues was used as a qualitative test for lignin accumulation and was localized to vascular tissue in the pedicel during the early post-pollination period. B A ovary + pedicel whole florets 50 100 150 0 cinnamic acid ng/g FW 100 200 300 salicylic acid 12 10 8 6 4 2 0 1 2 3 4 5 6 7 stage DBP 0

PAGE 119

119 CHAPTER 5 SUMMARY The overall hypothesis tested here w as that genes related to sp ecific functional processes or metabolic pathways would be co-expressed in association with carbohydra te allocation and use during silk exsertion and pollination in maize. Fu rther, the approaches used allowed such genes to be identified, clustered with ot her sets of co-regulated genes, and their expression quantified. The broader significance of this work was two-fold. First, maize now represents both the largest crop yield on earth in total bushels, as well as an emerging model for grain yield and biomass production. Second, the reproductive phases examined here are pivotal for successful maize pollination and establishment of kernel number, a primary determinant of yield. With the recent release of the maize draft genome sequence, focus is now directed to building a resource infrastructure specific to maiz e. Results from the work presented here will contribute to a foundation for functional, maize-based researc h. Currently, major limitations to determining specific functional roles of maize genes include insufficient genome annotation, biased array platforms that limit gene disc overy, and large-scale confounding effects from expression of paralogous genes. To achieve quantitative resolution of gene family members and other closely-related paralogs, we developed a sequencebased profiling method that uses the gene-specificity of the 3'-UTR (Chapter 2). We tested this method para llel with microarray anal yses in a genome-wide transcript profiling approach to identify co-expressed genes with related f unctions during maize ovary development (Chapter 3). Resulting analyses revealed cluste rs of genes involved in import and use of Cand N-resources that were co-expressed during development. In Chapter 4, we determined spatial and temporal changes in C-allocation and turgor-based expansion, as evidenced by dry and fresh weights, among individual floral organs In addition, the

PAGE 120

120 contribution of specific invertase isoforms was tested in relation to these changes. Key conclusions based on the combined analyses are as follows: A 3'-UTR profiling strategy was developed that resolved quantitative expression of gene family members and other clos ely-related transcripts in maize ovaries and provided a useful compliment to microarray analyses. A combined microarrayand sequence-based approach was found most effective for genome-wide transcript profili ng in developing maize ovaries. Advantages and limitations of both methods were addressed. Genes related to specific functional proce sses were co-expressed during maize ovary development. Key differences in functional classes of genes were evident at specific stages of development. Prior to pollination, increases in dry and fresh weights in silks and subtending floral organs coincided with expression of a s ugar-responsive vacuolar invertase, IVR2. Prominent dry weight gain indicated that silks were the major sink in pre-pollination maize florets. Immediately after pollination, allocation of C-resources (as ev ident by dry weight) shifted from silks and subtending floral organs to the developing ovary and pedicel. Co-expression analyses indicated that genes related to C-sink development were upregulated in concert with the increased accu mulation of dry weight and sucrose in the pedicel of the ovary. Genes related to N metabolism and lignin biosyntheses were co-expressed during development of maize ovaries and pedicels and tended to increase post-pollination. Evidence indicated that lignin biosynthesis in the pedicels of maize ovaries was a prominent sink for C resources during early post-pollination development. Global expression profiles and invertase activit y data indicated that regulation of sink strength in maize female floral organs was more strongly related to developmental changes in maternal tissues than to direct effects of pollination or fertilization. Central contributions to functional genomics in maize can be made by organ-specific expression profiles, metabolite analyses, and biochemical testing. The maize female inflorescence is characterized by unique an atomical features and therefore functional comparisons to Arabidopsis, or even to rice, can be insufficient. For many years, corn breeders have focused on genetic traits specific to the fe male inflorescence to improve kernel set and

PAGE 121

121 grain yield. Although genome-wid e expression analyses and functional annotation are becoming more accessible for maize, the preand early post-pollination phases of maize reproductive development have received comparatively little attention at the molecular/metabolic level. Results from the present work and from other studi es suggested that maternal-based regulation of reproductive development in cereals included Callocation and use (Westgate and Boyer, 1986; Zinselmeier et al., 2003; McLa ughlin and Boyer, 2004a; Sreenivasulu et al., 2004). Coexpressed genes related to these processes have been identified here, providing evidence for testable roles in kernel set. Future work can al so use these developmental analyses as a baseline for environmental or genetic perturbation of the system. .

PAGE 122

122 APPENDIX PERTURBATION BY DROUGHT AND DISR UPTED VP1-BASED ABA SE NSING Hypothesis The hypothesis tested here was that C allocation and use in developing fema le inflorescences of maize would be disrupted by drought stress and could in volve interactions of ABA and sugar signals. We analyz ed developing female florets (a s described in Chapter 4) from maize plants that were subjected to a drought stress treatment and from an ABA insensitive mutant to test affects of environmenta l and genetic perturbation, respectively. Background and Significance Drought stress during flowering can negatively impact yield in ma ize either by disruption of meiotic, pollination, and/or fertilization processes or by inhibiting grain fill (Westgate and Boyer, 1985; Andersen et al., 2002; Boyer and West gate, 2004; Barnabas et al., 2007). Westgate and Boyer (1985) showed that drought stress during pre-pollination growth significantly reduced solute concentration in the maize female infl orescence and did so to a greater extent than observed in leaves, stems, or root s. Resulting decreases in turgor were associated with inhibition of ear growth and failure of silk exsertion. Recent work by Borras et al. (2007) defined a th reshold level of ear growth, as measured by dry weight accumulation, which was indispensible for timely silk emergence. When C allocation to the developing ear fell below this th reshold, a larger Anthesis Silking Interval (ASI) resulted. Since tassel development showed little response to reduction in photosynthate availability, timing of pollen-shed was used as a marker to compare ASI in stressed plants (Borras et al., 2007). A short ASI is essentia l for optimal silk receptivity, pollination, and subsequent grain yield. Effort s to decrease the ASI through breeding and QTL analyses have

PAGE 123

123 been successful, however the mechanisms underl ying regulation remain elusive (Bruce et al., 2002). A key regulator of drought-indu ced gene expression is the phytohormone, abscisic acid (ABA). Recent work has demonstrated that loca l biosynthesis of ABA in leaves is dependent on a hydraulic signal from the roots (Christmann et al., 2007). Accumulation of ABA in response to soil drying causes stomatal closure, presumed to be a mechanism of acclimation that reduces evapotranspiration. Photosynthesis is also repr essed under these conditions, and thus disruption of the source-to-sink balance requires whole-plant adjustment. Evidence for ABA production in maternal reproductive tissues has also been demons trated (Myers et al., 1 992; Setter et al., 2001), and drought stress can result in tr ansient increases of this horm one during early development of female florets (Andersen et al., 2002). ABA-ba sed signaling networks ov erlap with those of sugar-based gene regulation on a number of leve ls (Finkelstein and Gibson, 2002; Li et al., 2006). Many of these connections have been demonstrated with redundant mutant loci in Arabidopsis that can result in ABAand suga r-responsive phenotypes (Are nas-Huertero et al., 2000; Huijser et al., 2000; Laby et al., 2000; Brocard et al., 2002; Arroyo et al., 2003; BrocardGifford et al., 2004). However, very few studies have extended su ch analyses to maize (Niu et al., 2002). Specific invertase isoforms respond to ABA in both maize (Kim et al., 2000b; Trouverie et al., 2004) and Arabidopsis (Huang, 2006). In maize, a vacuolar invertase, ZmIVR2 responds differentially to drought and associ ated rises in ABA levels in l eaves and developing ears. While ZmIVR2 is up-regulated in leaves during drought stress, the same isoform is repressed in florets (Andersen et al., 2002; Trouverie et al., 2003). This spatial-specifi c mode of regulation suggests an association between the phytohormone ABA a nd whole-plant source/sink balance. In

PAGE 124

124 addition, different levels of ABA can have contrasting effects on organ growth. Low levels of ABA tend to promote growth while higher levels are growth-inhibiting (C heng et al., 2002; Peng et al., 2003). Work by Suzuki et al. (2003) showed that AtvacINV1 the putative functional ortholog of ZmIVR2 was ABA up-regulated, but repressed by VP1 (a component of one ABA sensing system). Although VP1 has classically been im plicated as seed-specific in maize (McCarty et al., 1991; Hoecker et al., 1995), evidence from its Arabidopsis counterpart, ABI3, indicates putative functions in vegetative and ma ternal reproductive tissues (Parcy et al., 1994; Rohde et al., 1999; Rohde et al., 2002). A recen t study by Cao et al. (2006) showed that VP1 was induced by desiccation stress in maize and local ized to the phloem in l eaves and pedicels of developing female florets. In addition, studies in Arabidopsis have shown that ABI3 is involved in both ABA and auxin-based signaling pathways (Suzuki et al., 2001; Brady et al., 2003). Results In this study, we com pared mRNA le vels for two vacuolar invertases, Ivr1 and Ivr2 in developing maize ovaries before pollination. St aging of pre-pollination development was as described in Chapter 4. We also qua ntified endogenous levels of ABA and Vp1 mRNAs in the same material. Transcript profiles for both v acuolar invertases were quantified by Taqman QPCR analyses and showed similar trends duri ng pre-pollination ovary development (Figure A1A). A significant peak was evident in Ivr1 mRNA levels early in ovary development (stage 2 or 10 days prior to pollination according to our scal e). Transient accumulation of ABA (quantified by GC/MS analysis) coincided with elevated Ivr1 mRNA levels at stage 2 (10 DBP) (Figure A1B). Andersen et al. (2002) showed comparable levels of endogenous ABA in young maize ovaries. They also showed that this transien t peak in ABA was enhanced 3-fold under drought stress. In work here, we show a conc urrent, transient rise in levels of Vp1 mRNAs in young ovaries (stage 4 [6 DBP]). This peak in Vp1 mRNAs was preceded by ABA and Ivr1 mRNAs

PAGE 125

125 (Figure A-1C). These findings s howed that endogenous levels of Vp1 mRNAs and ABA were present in maternal tissues under norma l, well-watered growth conditions. One goal of this work was to test the exte nt of drought stress effects on the Anthesis Silking Interval (ASI) in maize plants lacking a VP1-based ABA-sensing system. To do this, we compared W22 wild-type and vp1 mutant maize in well-watered and drought-stressed conditions. ASI was measured in days from first pollen shed to first silk emergence (Table A-1). The vp1 mutants consistently showed a reduced ASI co mpared to the wild-type plants under droughtstress. Although the ASI was significan tly increased in drought-stressed vp1 mutants relative to well-watered controls, all vp1 ears analyzed (10/10) had exsert ed silks during drought treatments. In contrast, only 40% (4/10) of the wild-type ears exserted silk s. Rewatering after 7 days of withholding water resulted in silk exsertion by 100% of the wild-type ears. Wild-type and vp1 mutants showed no significant difference in ASI or in ear length when well-watered. Immature ears were sampled from wild-type and vp1 mutant maize plants grown under well-watered and drought conditions All ears were sampled 2 days after first pollen shed. Whole florets were collected from comparable sec tions at the base, middle, and tip of each ear, frozen in liquid nitrogen, and pooled for subse quent analyses. Levels of ABA and IAA were quantified in the wild-type and vp1 ears by GC/MS and used to compare the extent of treatment effects (Figure A-2A). Although total ABA levels were significantly lower in vp1 mutants, drought stress resulted in proportional increases to ABA levels for ears of both mutant and wildtype plants. We also observed that IAA levels were elevated in drought-stressed, wild-type ears compared to ears from well-watered controls. Interestingly, IAA leve ls did not change in response to drought in vp1 mutant ears. These data showed that drought stress altered the ABA/IAA ratio in the absence of a VP1-based ABA sensing system.

PAGE 126

126 Sucrose and hexoses were also quantified in the same material using HPLC. Results showed a decreased hexose/suc rose ratio in wild-type and vp1 mutant ears in response to drought. This decrease in hexose/sucrose was dampened in the vp1 mutants (Figure A-2B) by enhanced accumulation of sucrose. Together, da ta from this work and elsewhere indicates a possible association between sugar signals and ABA/auxin balance during drought-induced inhibition of maize ear growth. Preliminary evid ence presented here provides a basis for more in-depth analyses of the hypothe sized link between sugars and the ABA/auxin balance and its role in drought-stressed maize ears.

PAGE 127

127 Table A1-1. ASI for wild-type and vp1 mutant maize in response to drought. Treatment ASI (day) Ear length (cm) Well-watered wild-type 1.6 0.95 13.4 1.2 vp1 mutant 1.1 0.53 13.6 0.8 Drought stressed wild-type *5.8 1.2 7.2 1.4 vp1 mutant 2.3 1.2 8.4 2.1 *Only measured for ears with exserted silks (4/10).

PAGE 128

128 Figure A1-1. Quantification of soluble i nvertase mRNAs, endogenous ABA levels, and Vp1 transcripts during pre-pollination maize ovary development. A) Expression profiles for Ivr1 and Ivr2 quantified by Taqman Q-PCR. B) GC/MS analyses of endogenous ABA levels. C) Maternal-based Vp1 mRNAs quantified by Taqman Q-PCR.

PAGE 129

129 Figure A1-2. Sugar and hormone levels in dr ought-stressed, immature ears of wild-type and vp1 mutant maize. A) ABA and IAA levels quantified by GC/MS. B) Hexose/ sucrose ratios determined by HPLC analysis. ABA IAA 0 10 20 30 0 1000 2000 3000 wild-type vp1 hexose/sucrose 0 drought-stressed well-watered .5 1.5 2.5 B A 3.5 n g g FW-1 hexose/sucrose

PAGE 130

130 LIST OF REFERENCES Ambrose BA, Lerner DR, Ciceri P, Padilla CM, Yanofsky MF, Schmidt RJ (2000) Molecular and genetic analyses of the silky1 gene reveal conservation in floral organ specification between eudicots and monocots. Molecular Cell 5: 569-579 Andersen MN, Asch F, Wu Y, Je nsen CR, Naested H, Mogensen VO, Koch KE (2002) Soluble invertase expression is an early target of drought stress during the critical, abortion-sensitive phase of young ovary deve lopment in maize. Plant Physiology 130: 591-604 Anderson SR, Lauer MJ, Schoper JB, Shibles RM (2003) Pollination timing effects on kernel set and silk receptivity in f our maize hybrids. Crop Science 44: 464-473 Aoki N, Hirose T, Takahashi S, Ono K, Ishimaru K, Ohsugi R (1999) Molecular cloning and expression analysis of a gene for a sucrose transporter maize (Zea mays L.). Plant and Cell Physiology 40: 1072-1078 Appenzeller L, Doblin M, Barreiro R, Wang H, Niu X, Kollipara K, Carrigan L, Tomes D, Chapman M, Dhugga KS (2004) Cellulose sunthesis in maize: isolation and expression analysis of the cellulose synthase (CesA) gene family. Cellulose 11: 287-299 Arenas-Huertero F, Arroyo A, Zhou L, Sheen J, Leon P (2000) Analysis of Arabidopsis glucose insensitive mutants, gin5 and gin6, reve als a central role of the plant hormone ABA in the regulation of plant vegetative development by sugar. Genes & Development 14: 2085-2096 Arroyo A, Bossi F, Finkelstein RR, Leon P (2003) Three genes that affect sugar sensing (Abscisic Acid Insensitive 4, Abscisic Acid Insensitive 5, and Constitutive Triple Response 1) are differentially regulated by glucose in arabidopsis. Plant Physiology 133: 231-242 Avonce N, Leyman B, Mascorro-Gallardo JO van Dijck P, Thevelein JM, Iturriaga G (2004) The arabidopsis trehalos e-6-P synthase AtTPS1 gene is a regulator of glucose, abscisic acid, and stress signaling. Plant Physiol 136: 3649-3659 Balasubramanian R, Karve A, Kandasamy M, Meagher RB, Moore BD (2007) A role for FActin in hexokinase-mediated gl ucose signaling. Plant Physiology 145: 1423-1434 Bao JY, Lee S, Chen C, Zhang XQ, Zhang Y, Liu SQ, Clark T, Wang J, Cao ML, Yang HM, Wang SM, Yu J (2005) Serial analysis of gene expression study of a hybrid rice strain (LYP9) and its parental cultivars. Plant Physiol 138: 1216-1231 Barbazuk WB, Emrich SJ, Ch en HD, Li L, Schnable PS (2007) SNP discovery via 454 transcriptome sequencing. Plant J 51 : 910-918

PAGE 131

131 Barker L, Kuhn C, Weise A, Schulz A, Gebhar dt C, Hirner B, Hellmann H, Schulze W, Ward JM, Frommer WB (2000) SUT2, a putative sucros e sensor in sieve elements. Plant Cell 12: 1153-1164 Barnabas B, Jager K, Feher A (2008) The effect of drought and heat stress on reproductive processes in cereals. Plant Cell and Environment 31: 11-38 Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, Wei G, Chepelev I, Zhao K (2007) High-resolution profiling of histone methylations in the human genome. Cell 129: 823-837 Baxter CJ, Sabar M, Quick WP, Sweetlove LJ (2005) Comparison of changes in fruit gene expression in tomato introgression lines provides evidence of genome-wide transcriptional changes and reve als links to mapped QTLs a nd described traits. J Exp Bot 56: 1699-1709 Bhattramakki D, Dolan M, Hanafey M, Winela nd R, Vaske D, Register JC III, Tingey SV, Rafalski A (2002) Insertion-deletion polymorphisms in 3 regions of maize genes occur frequently and can be used as highly informative genetic markers. Plant Mol Biol 48 : 539-547 Bate NJ, Niu XP, Wang YW, Reimann KS, Helentjaris TG (2004) An invertase inhibitor from maize localizes to the embryo surroundi ng region during early kernel development. Plant Physiology 134: 246-254 Blanc G, Wolfe KH (2004) Functional divergence of duplicated genes formed by polyploidy during Arabidopsis evolution. Plant Cell 16: 1679-1691 Blasing OE, Gibon Y, Gunther M, Hohne M, Morcuende R, Osuna D, Thimm O, Usadel B, Scheible WR, Stitt M (2005) Sugars and circadian regul ation make major contributions to the global regulation of diurnal gene expression in Arabidopsis. Plant Cell 17: 32573281 Blee KA, Anderson AJ (2002) Transcripts for genes enco ding soluble acid invertase and sucrose synthase accumulate in root tip and cortical cells containing mycorrhizal arbuscules. Plant Molecular Biology 50: 197-211 Birchler JA, Augar DL, Riddle NC (2003) In search of the molecu lar basis of heterosis. Plant Cell 15: 2236-2239 Bocock PN, Morse AM, Dervinis C, Davis JM (2008) Evolution and diversity of invertase genes in Populus trichocarpa. Planta 227: 565-576 Bolanos J, Edmeades GO (1996) The importance of the anthes is-silking interval in breeding for drought tolerance in tropical maize. Field Crops Research 48: 65-80

PAGE 132

132 Borevitz JO, Liang D, Plouffe D, Chang H-S, Zhu T, Weigel D, Berry CC, Winzeler E, Chory J (2003) Large-scale identification of si ngle-feature polymorphisms in complex genomes. Genome Res 13: 513-523 Borevitz JO, Chory J (2004) Genomics tools for QTL anal ysis and gene discovery. Curr Opin Plant Biol 7: 132-136 Borisjuk L, Rolletschek H, Radchuk R, Weschke W, Wobus U, Weber H (2004) Seed development and differentiation: A role for metabolic regulation. Plant Biology 6: 375386 Borras L, Westgate ME, Otegui ME (2003) Control of kernel weight and kernel water relations by post-flowering source-sink ratio in maize. Annals of Botany 91: 857-867 Borras L, Westgate ME, Astini JP, Echarte L (2007) Coupling time to silking with plant growth rate in maize. Field Crops Research 102: 73-85 Boyer JS (1982) Plant productivity and environment. Science 218: 443-448 Boyer JS, Westgate ME (2004) Grain yields with limited water. Journal of Experimental Botany 55: 2385-2394 Brady SM, Sarkar SF, Bonetta D, McCourt P (2003) The ABSCISIC ACID INSENSITIVE3 (ABI3) gene is modulated by farnesylation a nd is involved in auxin signaling and lateral root development in Arabidopsis. Plant J 34: 67-75 Bray EA, Shih T-Y, Moses MS, Cohen A, Imai R, Plant AL (1999) Water-deficit induction of a tomato H1 histone requires abscisic acid. Plant Growth Reg 29: 35-46 Brocard IM, Lynch TJ, Finkelstein RR (2002) Regulation and role of the Arabidopsis Abscisic Acid-Insensitive 5 gene in abscisic acid, sugar, and stress response. Plant Physiology 129: 1533-1543 Brocard-Gifford I, Lynch TJ, Garcia ME, Malhotra B, Finkelstein RR (2004) The Arabidopsis thaliana ABSCISIC ACID-INSE NSITIVE8 locus encodes a novel protein mediating abscisic acid and sugar res ponses essential for growth. Plant Cell 16: 406-421 Bruce WB, Edmeades GO, Barker TC (2002) Molecular and physio logical approaches to maize improvement for drought tolerance. Journal of Experimental Botany 53: 13-25 Bustan A, Goldschmidt EE (1998) Estimating the cost of flowering in a grapefruit tree. Plant Cell and Environment 21: 217-224 Buttner M (2007) The monosaccharide transporter (-li ke) gene family in Arabidopsis. Febs Letters 581: 2318-2324

PAGE 133

133 Cakir B, Agasse A, Gaillard C, Saumonneau A, Delrot S, Atanassova R (2003) A grape ASR protein involved in sugar and abscisic acid signaling. Plant Cell 15: 2165-2180 Campos H, Cooper A, Habben JE, Edmeades GO, Schussler JR (2004) Improving drought tolerance in maize: a view from industry. Field Crops Research 90: 19-34 Cao X, Costa LM, Biderre-Petit C, Kbhaya B, Dey N, Perez P, McCarty DR, GutierrezMarcos JF, Becraft PW (2007) Abscisic acid and stress signals induce Viviparous-1 ( Vp1) expression in seed and vegetative tissues of maize. Plant Physiol 143: 720-731 Carlson SJ, Chourey PS (1999) A re-evaluation of the relati ve roles of two invertases, INCW2 and IVR1, in developing maize kernel s and other tissues. Plant Physiology 121: 10251035 Carpaneto A, Geiger D, Bamberg E, Sauer N, Fromm J, Hedrich R (2005) Phloemlocalized, proton-coupled sucrose carrier ZmSUT1 mediates sucrose efflux under the control of the sucrose gradient and the proton motive force. Journal of Biological Chemistry 280: 21437-21443 Cazetta JO, Seebauer JR, Below FE (1999) Sucrose and nitrogen supplies regulate growth of maize kernels. Annals of Botany 84: 747-754 Chen J, Sun M, Lee S, Zhou G, Rowley JD, Wang SM (2002) Identifying novel transcripts and novel genes in the human genome by usi ng novel SAGE tags. Proc Natl Acad Sci USA 99: 12257-12262 Cheng WH, Endo A, Zhou L, Penney J, Chen HC, Arroyo A, Leon P, Nambara E, Asami T, Seo M, Koshiba T, Sheen J (2002) A unique short-chain dehydrogenase/reductase in Arabidopsis glucose signaling and abscisic acid biosynthesi s and functions. Plant Cell 14: 2723-2743 Cheng WH, Taliercio EW, Chourey PS (1996) The Miniature1 seed locus of maize encodes a cell wall invertase required fo r normal development of endosperm and maternal cells in the pedicel. Plant Cell 8: 971-983 Cheung VG, Spielman RS, Ewens KG, Weber TM, Morley M, Burdick JT (2005) Mapping determinants of human gene expression by re gional and genome-wide association. Nature 437: 1365-1369 Chiou TJ, Bush DR (1998) Sucrose is a signal molecule in assimilate partitioning. Proceedings of the National Academy of Sciences of the United States of America 95: 4784-478 Cho JI, Lee SK, Ko SH, Kim HK, Jun SH, Lee YH, Bhoo SH, Lee KW, An GH, Hahn TR, Jeon JS (2005) Molecular cloning and expression analysis of the cell-wa ll invertase gene family in rice (Oryza sativa L.). Plant Cell Reports 24: 225-236

PAGE 134

134 Cho YH, Yoo SD, Sheen J (2006) Regulatory functions of nuclear hexokinase1 complex in glucose signaling. Cell 127: 579-589 Chourey P, Jain M, Li QB, Carlson S (2006) Genetic control of cell wall invertases in developing endosperm of maize. Planta 223: 159-167 Christmann A, Weiler EW, Steudle E, Grill E (2007) A hydraulic sign al in root-to-shoot signalling of water shortage. Plant Journal 52: 167-174 Clifford HT (1987). Spikelet and floral morphology. In Grass Systematics and Evolution, T.R. Soderstrom, K.W. Hilu, C.S. Campbell, and M.E. Barkworth, eds (Washington, DC: Smithsonian Institute) pp 21. Cong B, Liu J, Tanksley SD (2002) Natural alleles at a tomato fruit size quantit ative trait locus differ by heterochronic regulatory mu tations. Proc Natl Acad Sci USA 99: 13606-13611 Cooke JEK, Brown KA, Wu R, Davis JM (2003) Gene expression as sociated with N-induced shifts in resource allocation in poplar. Plant, Cell, and Environment 26: 757-770 Coruzzi G, Bush DR (2001) Nitrogen and carbon nutrient and metabolite signaling in plants. Plant Physiology 125: 61-64 Cowles CR, Hirschhorm JN, Altshuler D, Lander ES (2002) Detection of regulatory variation in mouse genes. Nat Genetics 32: 432-437 Cruz-Aguado JA, Reyes F, Rodes R, Peres I, Dorado M (1999) Effect of source-to-sink ratio on partitioning of dry matter and C-14-photoass imilates in wheat during grain filling. Annals of Botany 83: 655-665 Eastmond PJ, van Dijken AJ, Spielman M, K err A, Tissier AF, Dickinson HG, Jones JD, Smeekens SC, Graham IA (2002) Trehalose-6-phosphate synthase 1, which catalyzes the first step in trehalose s ynthesis, is essential for Arab idopsis embryo maturation. Plant J 29: 225-235 Ehness R, Ecker M, Godt DE, Roitsch T (1997) Glucose and stress independently regulate source and sink metabolism and defense mechanisms via signal transduction pathways involving protein phosphor ylation. Plant Cell 9: 1825-1841 Emrich SJ, Barbazuk B, Schnable PS (2007a) Gene discovery and annotation using LCM-454 transcriptome sequenc ing. Genome Res 17: 69-73 Emrich SJ, Li L, Wen T-J, Yandeau-Nelson MD, Fu Y, Guo L, Chou H-H, Aluru S, Ashlock DA, Schnable PS (2007b) Nearly identical para logs: Implications for Maize ( Zea mays L.) genome evolution. Genetics 175: 429-439

PAGE 135

135 Eveland AL, McCarty DR, Koch KE (2008) Transcript profili ng by 3'-untranslated region sequencing resolves expression of gene families. Plant Physiology 146: 32-44 Finkelstein RR, Gibson SI (2002) ABA and sugar interactions regulating development: crosstalk or voices in a crowd? Current Opinion in Plant Biology 5: 26-32 Fisher DB, Wang N (1995) Sucrose concentration grad ients along the post-phloem transport pathway in the maternal tissues of de veloping wheat grains. Plant Physiology 109: 587592 Flemetakis E, Efrose RC, Ott T, Stedel C, Aivalakis G, Udvardi MK, Katinakis P (2006) Spatial and temporal organization of sucros e metabolism in Lotus japonicus nitrogenfixing nodules suggests a role fo r the elusive alkaline/neutral invertase. Plant Molecular Biology 62: 53-69 Fridman E, Zamir D (2003) Functional divergence of a s yntenic invertase gene family in tomato, potato, and Arabidopsis. Plant Physiology 131: 603-609 Fritz C, Palacios-Rojas N, Feil R, Stitt M (2006) Regulation of secondary metabolism by the carbon-nitrogen status in tobacco: nitrate inhibits large sectors of phenylpropanoid metabolism. Plant Journal 46: 533-548 Fu Y, Emrich SJ, Guo L, Wen TJ, Ashlock DA, Aluru S, Schnable PS (2005) Quality assessment of maize assembled genomic isla nds (MAGIs) and large-scale experimental verification of predicted gene s. Proc Natl Acad Sci USA 102: 12282-12287 Fung RWM, Gonzalo M, Fekete C, K ovacs LG, He Y, Marsh E, McIntyre LM, Schachtman DP, Qiu W (2008) Powdery mildew induces defense-oriented reprogramming of the transcriptome in a sus ceptible but not in a resistant grapevine. Plant Physiol 146: 236-249 Furusawa C, Kaneko K (2003) Zipfs law in gene expression. Phys Rev Lett 90: 88-102 Furutani I, Sukegawa S, Kyozuka J (2006) Genome-wide analysis of spatial and temporal gene expression in rice panicl e development. Plant Journal 46: 503-511 Gazzarrini S, McCourt P (2001) Genetic interactions between ABA, ethylene and sugar signaling pathways. Current Opinion in Plant Biology 4: 387-391 Gibon Y, Blasing OE, Palacios-Rojas N, Pa nkovic D, Hendriks JHM, Fisahn J, Hohne M, Gunther M, Stitt M (2004) Adjustment of diurnal starch turnover to short days: depletion of sugar during the night leads to a temporary inhi bition of carbohydrate utilization, accumulation of sugars and pos t-translational activa tion of ADP-glucose pyrophosphorylase in the following light period. Plant Journal 39: 847-862

PAGE 136

136 Gibson SI (2004) Sugar and phytohormone response pathways: navigating a signalling network. J Exp Bot: 55: 253-264 Godt D, Roitsch T (2006) The developmental and organ sp ecific expression of sucrose cleaving enzymes in sugar beet suggests a transiti on between apoplasmic and symplasmic phloem unloading in the tap roots. Plan t Physiology and Biochemistry 44: 656-665 Gonzalez MC, Roitsch T, Cejudo FJ (2005) Circadian and deve lopmental regulation of vacuolar invertase expression in peti oles of sugar beet plants. Planta 222: 386-395 Grimanelli D, Perotti E, Ramirez J, Leblanc O (2005) Timing of the maternal-to-zygotic transition during early seed development in maize. Plant Cell 17: 1061-1072 Guo M, Rupe MA, Danilevs kaya ON, Yang X, Hu Z (2003) Genome-wide mRNA profiling reveals heterochronic allelic variation a nd a new imprinted gene in hybrid maize endosperm. Plant J 36: 30-44 Guo M, Rupe MA, Zinselmeier C, Habben J, Bowen BA, Smith OS (2004) Allelic variation of gene expression in maize hybrids. Plant Cell 16: 1707-1716 Gu Z, Rifkin SA, White KP, Li W-H (2004) Duplicate genes increase gene expression diversity within and between species. Nat Genetics 36: 577-578 Harkins T, Jarvie T (2007) Megogenomics analysis using the Genome Sequencer FLX system. Nat Methods 4: application notes Harrington GN, Bush DR (2003) The bifunctional role of hexokinase in metabolism and glucose signaling. Plant Cell 15: 2493-2496 Helentjaris T, Weber D, Wright S (1988) Identification of the genomic locations of duplicate nucleotide sequences in maize by analysis of restriction fragment length polymorphisms. Genetics 118 : 353-363 Hoecker U, Vasil IK, McCarty DR (1995) Integrated contro l of seed maturation and germination programs by activator and repressor functions of Viviparous-1 of maize. Genes & Development 9: 2459-2469 Hofmann J, Wieczorek K, Blochl A, Grundler FMW (2007) Sucrose supply to nematodeinduced syncytia depends on the apoplasmic and symplasmic pathways. Journal of Experimental Botany 58: 1591-1601 Holland N, Holland D, Helentjaris T, Dhu gga KS, Xoconostle-Cazares B, Delmer DP (2000) A comparative analysis of th e plant Cellulose Synthase ( CesA ) gene family. Plant Physiol 123: 1313-1323

PAGE 137

137 Hothorn M, Wolf S, Aloy P, Greiner S, Scheffzek K (2004) Structural insi ghts into the target specificity of plant invertase and pectin met hylesterase inhibitory proteins. Plant Cell 16: 3437-3447 Huang J, Taylor JP, Chen J-G, Uhrig JF, Schnell DJ, Nakagawa T, Korth KL, Jones AM (2006) The plastid protein THYLAKOID FORMATION1 and the plasma membrane Gprotein GPA1 interact in a nove l sugar-signaling mechanism in Arabidopsis Plant Cell 18: 1226-1238 Huang LF (2006) Molecular analysis of an invertase gene family in Arabidopsis. Thesis (PhD) University of Florida Huang LF, Bocock PN, Davis JM, Koch KE (2007) Regulation of i nvertase: a suite of transcriptional and post-transcriptional mechanisms. Functional Plant Biology 34: 499507 Huang X, Madan A (1999) CAP3: A DNA sequence assembly program. Genome Res 9: 868877 Huijser C, Kortstee A, Pego J, Weisbeek P, Wisman E, Smeekens S (2000) The Arabidopsis SUCROSE UNCOUPLED-6 gene is identical to ABSCISIC ACID INSENSITIVE-4: involvement of abscisic acid in sugar responses. Plant Journal 23: 577-585 Ji XM, Raveendran M, Oane R, Ismail A, Lafitte R, Bruskiewich R, Cheng SH, Bennett J (2005) Tissue-specific expre ssion and drought responsivene ss of cell-wall invertase genes of rice at flowering. Plant Molecular Biology 59: 945-964 Jongeneel CV, Delorenzi M, Iseli C, Zhou D, Haudenchild CD, Khrebtukova I, Kuznetsov D, Stevenson BJ, Strausberg RL, Simpson AJG, et al (2005) An atlas of human gene expression from massively parallel signa ture sequences (MPSS) Genome Res 15: 10071014 Karthikeyan AS, Varadarajan DK, Jain A, Held MA, Carpita NC, Raghothama KG (2007) Phosphate starvation responses are mediated by sugar signaling in Arabidopsis. Planta 225: 907-918 Kim JY, Mahe A, Brangeon J, Prioul JL (2000a) A maize vacuolar invertase, IVR2, is induced by water stress. Plant Physiology 124: 71-84 Kim JY, Mahe A, Guy S, Brangeon J, Roche O, Chourey PS, Prioul JL (2000b) Characterization of two members of the ma ize gene family, Incw3 and Incw4, encoding cell-wall invertases. Gene 245: 89-102 Kladnik A, Chamusco K, Dermastia M, Chourey P (2004) Evidence of programmed cell death in post-phloem trans port cells of the maternal pedicel tissue in developing caryopsis of maize. Plant Physiology 136: 3572-3581

PAGE 138

138 Koch KE (1996) Carbohydrate-modulate d gene expression in plants Annual Review of Plant Physiology and Plant Molecular Biology 47: 509-540 Koch KE (2004) Sucrose metabolism: re gulatory mechanisms and pivo tal roles in sugar sensing and plant development. Current Opinion in Plant Biology 7: 235-246 Koch KE, Ying Z, Wu Y, Avigne WT (2000) Multiple paths of sugar-sensing and a sugar/oxygen overlap for genes of sucros e and ethanol metabolism. Journal of Experimental Botany 51: 417-427 Kohorn BD, Kobayashi M, Johansen S, Riese J, Huang LF, Koch K, Fu S, Dotson A, Byers N (2006) An Arabidopsis cell wall-associated kinase required for invertase activity and cell growth. Plant Journal 46: 307-316 Kojima H, Suzuki T, Kato T, Enomoto K, Sa to S, Kato T, Tabata S, Saez-Vasquez J, Manuel E, Nakagawa T, Ishiguro S, Nakamura K (2007) Sugar-inducible expression of the nucleolin-1 gene of Arabidopsis thaliana and its role in ribos ome synthesis, growth and development. Plant Journal 49: 1053-1063 Kuznetsov VA, Knott GD, Bonner RF (2002) General statistics of stochastic process of gene expression in eukaryotic cells. Genetics 161: 1321-1332 Laby RJ, Kincaid MS, Kim DG, Gibson SI (2000) The Arabidopsis sugar-insensitive mutants sis4 and sis5 are defective in abscisic acid synthesis and response. Plant Journal 23: 587596 Lalonde S, Boles E, Hellmann H, Ba rker L, Patrick JW, Frommer WB, Ward JM (1999) The dual function of sugar carriers: Tr ansport and sugar sensing. Plant Cell 11: 707-726 Lalonde S, Wipf D, Frommer WB (2004) Transport mechanisms for organic forms of carbon and nitrogen between source and sin k. Annual Review of Plant Biology 55: 341-372 Lara MEB, Garcia MCG, Fatima T, Ehness R, Lee TK, Proels R, Tanner W, Roitsch T (2004) Extracellular inve rtase is an essential component of cytokinin-mediated delay of senescence. Plant Cell 16: 1276-1287 Lee EJ, Matsumura Y, Soga K, Hoson T, Koizumi N (2007) Glycosyl hydr olases of cell wall are induced by sugar starvation in Ar abidopsis. Plant and Cell Physiology 48: 405-413 Leon P, Sheen J (2003) Sugar and hormone connec tions. Trends in Plant Science 8: 110-116 Li YH, Lee KK, Walsh S, Smith C, Hadingham S, Sorefan K, Cawley G, Bevan MW (2006) Establishing glucoseand ABA-regulated transcription networks in Arabidopsis by microarray analysis and prom oter classification using a Relevance Vector Machine. Genome Research 16: 414-427

PAGE 139

139 Li YH, Smith C, Corke F, Zheng LY, Merali Z, Ryden P, Derbyshire P, Waldron K, Bevan MW (2007) Signaling from an alte red cell wall to the nucleus mediates sugar-responsive growth and development in Arab idopsis thaliana. Plant Cell 19: 2500-2515 Linden, JC, Ehness, R, and Roitsch, T (1996) Regulation by ethylene of apoplastic invertase expression in Chenopodium rubrum tissue culture cells. Plant Growth Regul 19: 219 Liu F, Jenssen T-K, Trimarchi J, Punzo C, Cepko CL, Ohno-Machado L, Hovig E, Kuo WP (2007) Comparison of hybrid ization-based and sequenci ng-based gene expression technologies on biological replicates. BMC Genomics 8: 153 Long JC, Zhao W, Rashotte AM, Muday GK, Huber SC (2002) Gravity-stimulated changes in auxin and invertase gene expression in maize pulvinal cells. Plant Physiology 128: 591-602 Lou Y, Gou J-Y, Xue H-W (2007) PIP5K9, an Arabidopsis phosphatidylinositol monophosphate kinase, interacts with a cytoso lic invertase to nega tively regulate sugarmediated root growth. Plant Cell 19: 163-181 Lu C, Tej SS, Luo S, Haudenchild CD, Meyers BC, Green PJ (2005) Elucidation of the small RNA component of the transcriptome. Science 309: 1567-1569 Ma J, Morrow DJ, Fernandes J, Walbot V (2006) Comparative profiling of sense and antisense transcriptome of maize lines. Genome Biology 7: R22 Maitz M, Santandrea G, Zhang ZY, Lal S, Hannah LC, Salamini F, Thompson RD (2000) rgf1, a mutation reducing grai n filling in maize through e ffects on basal endosperm and pedicel development. Plant Journal 23: 29-42 Makela P, McLaughlin JE, Boyer JS (2005) Imaging and quantif ying carbohydrate transport to the developing ovaries of maize. Annals of Botany 96: 939-949 Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, Berka J, Braverman MS, Chen Y-J, Chen Z, Dewell, et al. (2005) Genome sequencing in microfabricated high-density picolitre reactors. Nature 437: 376-380 Martin A, Lee J, Kichey T, Gerentes D, Zivy M, Tatout C, Dubois F, Balliau T, Valot B, Davanture M, et al (2006) Two glutamine synthetase is oforms of maize are specifically involved in the control of grain production. Plant Cell 18: 3252 McCarty DR, Hattori T, Carson CB, Vasil V, Lazar M, Vasil IK (1991) The Viviparous-1 developmental gene of maize encodes a novel transcriptio nal activator. Cell 66: 895 McCormick AJ, Cramer MD, Watt DA (2008) Changes in photos ynthetic rates and gene expression of leaves during a source-sink pe rturbation in sugarcane. Annals of Botany 101: 89-102

PAGE 140

140 McLaughlin JE, Boyer JS (2004a) Glucose localization in ma ize ovaries when kernel number decreases at low water potential and sucros e is fed to the stems. Annals of Botany 94: 7586 McLaughlin JE, Boyer JS (2004b) Sugar-responsive gene e xpression, invertase activity, and senescence in aborting maize ovaries at low water poten tials. Annals of Botany 94: 675689 Menges M, Samland AK, Planchais S, Murray JAH (2006) The D-type cyclin CYCD3:1 is limiting for the G1-to-S-phase tran sition in Arabidopsis. Plant Cell 18: 893-906 Messing J, Bharti AK, Karlow ski WM, Gundlach H, Kim HR, Yu Y, Wei F, Fuks G, Soderlund CA, Mayer KFX, et al. (2004) Sequence composition and genome organization of maize. Proc Natl Acad Sci USA 101 : 14349-14354 Messing J, Dooner HK (2006) Organization and variability of the maize genome. Curr Opin Plant Biol 9 : 157-163 Meyers BC, Tej SS, Vu TH, Haudenschild CD, Ag rawal V, Edberg SB, Ghazal H, Decola S (2004) The use of MPSS for whole-genome tr anscriptional analysis in Arabidopsis. Genome Res 14: 1641-1653 Mockler TC, Ecker JR (2005) Applications of DNA tiling arrays for whole-genome analysis. Genomics 85: 1-15 Moore B, Zhou L, Rolland F, Hall Q, Cheng WH, Liu YX, Hwang I, Jones T, Sheen J (2003) Role of the Arabidopsis glucose sens or HXK1 in nutrient, light, and hormonal signaling. Science 300: 332-336 Moore RC, Purugganan MD (2005) The evolutionary dynamics of plant duplicate genes. Curr Opin Plant Biol 8: 122-128 Myers PN, Setter TL, Madison JT, Thompson JF (1992) Endosperm cell division in maize kernels cultured at 3 levles of water potential. Plant Physiology 99: 1051-1056 Ner-Gaon H, Leviatan N, Rubin E, Fluhr R (2007) Comparative cross-species alternative splicing in plants. Plant Physiol 144: 1632-1641 Newman-Smith ED, Rothman JH (1998) The maternal-to-zygotic transition in embryonic patterning of Caenorhabditis elegans Curr Opin Genet Dev 8: 472 Newport J, Kirschner M (1982) A major developmental transition in early Xenopus embryos: I. Characterization and timing of cellular changes at the midblastula stage. Cell 30: 675 686.

PAGE 141

141 Nicolai M, Roncato MA, Canoy AS, Rouquie D, Sarda X, Freyssinet G, Robaglia C (2006) Large-scale analysis of mRNA translation stat es during sucrose starva tion in Arabidopsis cells identifies cell proliferation and chroma tin structure as targ ets of translational control. Plant Physiology 141: 663-673 Niittylae T, Fuglsang AT, Palm gren MG, Frommer WB, Schulze WX (2007) Temporal analysis of sucrose-induced phosphorylation changes in plas ma membrane proteins of Arabidopsis. Molecular & Cellular Proteomics 6: 1711-1726 Niu XP, Helentjaris T, Bate NJ (2002) Maize AB14 binds coupling element1 in abscisic acid and sugar response genes. Plant Cell 14: 2565-2575 Nobuta K, Venu RC, Lu C, Belo A, Vemaraju K, Kulkarni K, Wang W, Pillay M, Green PJ, Wang G-l, Meyers BC (2007) An expression atlas of rice mRNAs and small RNAs. Nat Biotechnol 25: 473-477 Oliver SN, Dennis ES, Dolferus R (2007) ABA regulates apoplasti c sugar transport and is a potential signal for cold-induced pollen st erility in rice. Plant and Cell Physiology 48: 1319-1330 Osborn TC, Pires JC, Birchler JA, Auger DL, Chen ZJ, Lee H-S, Comai L, Madlung A, Doerge RW, Colot V, et al. (2003) Understanding mechanisms of novel gene expression in polyploids. Trends Genet 19: 141-147 Osuna D, Usadel B, Morcuende R, Gibon Y, Blasing OE, Hohne M, Gunter M, Kamlage B, Trethewey R, Scheible WR, Stitt M (2007) Temporal responses of transcripts, enzyme activities and metabolites afte r adding sucrose to carbon-depr ived Arabidopsis seedlings. Plant Journal 49: 463-491 Ozcan S, Dover J, Johnston M (1998) Glucose sensing and si gnaling by two glucose receptors in the yeast Saccharomyces cerevisiae. EMBO 17: 2566-2573 Palenchar PM, Kouranov A, Lejay LV, Coruzzi GM (2004) Genome-wide patterns of carbon and nitrogen regulation of ge ne expression validate the carbon and nitrogen (CN)signaling hypothesis. Genome Biol 5: R91 Pan QH, Li MJ, Peng CC, Zhang N, Zou X, Zou KQ, Wang XL, Yu XC, Wang XF, Zhang DP (2005) Abscisic acid activates acid invert ases in developing grape berry. Physiologia Plantarum 125: 157-170 Patrick JW (1997) Phloem unloading: Sieve elem ent unloading and post-sieve element transport. Annual Review of Plant Physiology and Plant Molecular Biology 48: 191-222 Paul MJ, Foyer CH (2001) Sink regulation of photosynthesis. Journal of Experimental Botany 52: 1383-1400

PAGE 142

142 Pelegri F (2003). Maternal factors in zebrafish development. Dev Dyn 228: 535 Pelleschi S, Leonardi A, Rocher JP, Cornic G, de Vienne D, Thevenot C, Prioul JL (2006) Analysis of the relationships betwee n growth, photosynthesis and carbohydrate metabolism using quantitative trait loci (QTLs ) in young maize plants subjected to water deprivation. Molecular Breeding 17: 21-39 Peng YB, Lu YF, Zhang DP (2003) Abscisic acid activates ATPase in developing apple fruit especially in fruit phloem ce lls. Plant Cell and Environment 26: 1329-1342 Price J, Laxmi A, St Martin SK, Jang JC (2004) Global transcri ption profiling reveals multiple sugar signal transduction mechanisms in Arabidopsis. Plant Cell 16: 2128-2150 Pourtau N, Mars M, Purdy S, Quentin N, Rul A, Wingler A. (2004) Interactions of abscisic acid and sugar signalling in the regul ation of leaf senescence. Planta 219: 765 Proels RK, Hause B, Berger S, Roitsch T (2003) Novel mode of hormone induction of tandem tomato invertase genes in floral tissues. Plant Molecular Biology 52: 191-201 Qin LX, Trouverie J, Chateau-Joubert S, Simond-Cote E, Thevenot C, Prioul JL (2004) Involvement of the Ivr2-invertase in the pe rianth during maize kernel development under water stress. Plant Science 166: 371-379 Qu Y, Xu S (2006) Quantitative trait a ssociated microarray gene expression data analysis. Molecular Biology and Evolution 23: 1558-1573 Ramon M, Rolland F, Thevelein JM, Van Dijck P, Leyman B (2007) ABI4 mediates the effects of exogenous trehalose on Arabidops is growth and starch breakdown. Plant Molecular Biology 63: 195-206 Riou-Khamlichi C, Menges M, Healy JMS, Murray JAH (2000) Sugar control of the plant cell cycle: Differential regulation of Arab idopsis D-type cycl in gene expression. Molecular and Cellular Biology 20: 4513-4521 Ritter H, Schulz GE (2004) Structural basis for the entrance into the phenylpropanoid metabolism catalyzed by phenylalanine ammonia-lyase. Plant Cell 16: 3426-3436 Rogers LA, Dubos C, Cullis IF, Surman C, Poole M, Willment J, Mansfield SD, Campbell MM (2005) Light, the circadian clock, and sugar perception in the control of lignin biosynthesis. Journal of Experimental Botany 56: 1651-1663 Rohde A, Prinsen E, De Rycke R, Engler G, Van Montagu M, Boerjan W (2002) PtABI3 impinges on the growth and differentiation of embryonic leaves during bud set in poplar. Plant Cell 14: 1885-1901

PAGE 143

143 Roitsch T (1999) Source-sink regulation by sugar and stress. Current Opinion in Plant Biology 2: 198-206 Roitsch T, Balibrea ME, Hofmann M, Proels R, Sinha AK (2003) Extracellula r invertase: key metabolic enzyme and PR protein. Journal of Experimental Botany 54: 513-524 Roitsch T, Ehness R, Goetz M, Hause B, Hofmann M, Sinha AK (2000) Regulation and function of extracellula r invertase from higher plants in relation to assimilate partitioning, stress responses and sugar signalling. Australian Journal of Plant Physiology 27: 815-825 Roitsch T, Gonzalez MC (2004) Function and regul ation of plant invertas es: sweet sensations. Trends in Plant Science 9: 606-613 Rojo E, Zouhar J, Carter C, Kovaleva V, Raikhel NV (2003) A unique mechanism for protein processing and degradation in Arabidopsis thaliana. Proceedings of the National Academy of Sciences of the United States of America 100: 7389-7394 Rolland F, Baena-Gonzalez E, Sheen J (2006) Sugar sensing a nd signaling in plants: Conserved and novel mechanisms. Annual Review of Plant Biology 57: 675-709 Rolland F, Winderickx J, Thevelein JM (2002) Glucose-sensing and signaling mechanisms in yeast. FEMS Yeast Res 2: 183-201 Rook F, Corke F, Baier M, Holman R, May AG, Bevan MW (2006) Impaired sucrose induction1 encodes a conserved plant-specif ic protein that co uples carbohydrate availability to gene expression and plant growth. Plant Journal 46: 1045-1058 Rook F, Corke F, Card R, Munz G, Smith C, Bevan MW (2001) Impaired sucrose-induction mutants reveal the modulation of sugar-induced starch biosyn thetic gene expression by abscisic acid signalling. Plant Journal 26: 421-433 Rook F, Hadingham SA, Li Y, Bevan MW (2006) Sugar and ABA response pathways and the control of gene expressio n. Plant Cell and Environment 29: 426-434 Saha S, Sparks AB, Rago C, Akmaev V, Wang CJ, Vogelstein B, Kinzler KW, Velculescu VE (2002) Using the transcriptome to annotate the genome. Nat Biotechnol 20: 508-512 Satoh-Nagasawa N, Nagasawa N, Ma lcomber S, Sakai H, Jackson D (2006) A trehalose metabolic enzyme controls inflorescence architecture in maize. Nature 441: 227-230 Sauer N (2007) Molecular physiology of higher pl ant sucrose transporters. Febs Letters 581: 2309-2317 Schaarschmidt S, Kopka J, Ludwig-Muller J, Hause B (2007) Regulation of arbuscular mycorrhization by apoplastic invertases: enhanc ed invertase activity in the leaf apoplast affects the symbiotic interaction. Plant Journal 51: 390-405

PAGE 144

144 Schadt EE, Monks SA, Drake TA, Lusis AJ, Che N, Collnayo V, Ruff TG, Milligan SB, Lamb JR, Cavet G, et al (2003) Genetics of gene expressi on surveyed in maize, mouse, and man. Nature 422: 297-302 Schmidt RJ, Veit B, Mandel MA, Mena M, Hake S, Yanofsky MF (1993) Identification and molecular characterization of ZAG1, the ma ize homolog of the Arabidopsis floral homeotic gene AGAMOUS. Plant Cell 5: 729-737 Scofield GN, Hirose T, Aoki N, Furbank RT (2007) Involvement of the sucrose transporter, OsSUT1, in the long-distance pathway for assimilate transport in rice. Journal of Experimental Botany 58: 3155-3169 Seebauer JR, Moose SP, Fabbri BJ, Crossland LD, Below FE (2004) Amino acid metabolism in maize earshoots. Implications for assim ilate preconditioning and nitrogen signaling. Plant Physiology 136: 4326-4334 Sergeeva LI, Keurentjes JJB, Bentsink L, Vonk J, van der Plas LHW, Koornneef M, Vreugdenhil D (2006) Vacuolar invert ase regulates elongation of Arabidopsis thaliana roots as revealed by QTL and mutant analysis. Proceedings of the National Academy of Sciences of the United States of America 103: 2994-2999 Setter TL, Flannigan BA, Melkonian J (2001) Loss of kernel set due to water deficit and shade in maize: Carbohydrate supplies, abscis ic acid, and cytokinins. Crop Science 41: 15301540 Sherson SM, Alford HL, Forbes SM, Wallace G, Smith SM (2003) Roles of cell-wall invertases and monosaccharide transporters in the growth and development of Arabidopsis. Journal of Experimental Botany 54: 525-531 Smeekens S (2000) Sugar-induced signal transduction in plants. Annual Review of Plant Physiology and Plant Molecular Biology 51: 49-81 Smith AM, Stitt M (2007) Coordination of carbon suppl y and plant growth. Plant Cell and Environment 30: 1126-1149 Soderman EM, Brocard IM, Lynch TJ, Finkelstein RR (2000) Regulation and function of the arabidopsis ABA-insensitive4 gene in seed and abscisic acid response signaling networks. Plant Physiology 124: 1752-1765 Springer NM, Stupar RM (2007) Allele-specific expression patterns reveal biases and embryospecific parent-of-origin effects in hybrid maize. Plant Cell 19: 2391-2402 Sreenivasulu N, Altschmied L, Radchuk V, Gubatz S, Wobus U, Weschke W (2004) Transcript profiles and deduced changes of me tabolic pathways in maternal and filial tissues of developing barley grains. Plant Journal 37: 539-553

PAGE 145

145 Stafstrom JP, Ripley BD, Devitt ML, Drake B (1998) Dormancy-associated gene expression in pea axillary buds. Planta 205: 547-552 Stitt M, Gibon Y, Lunn JE, Piques M (2007) Multilevel genomics analysis of carbon signalling during low carbon availability: co ordinating the supply and utilisation of carbon in a fluctuating environm ent. Functional Plant Biology 34: 526-549 Street NR, Skogstrom O, Sjodin A, Ticker J, Rodriguez-Acosta M, Nilsson P, Jansson S, Taylor G (2006) The genetics and genom ics of drought response in Populu s. Plant J 48: 321-341 Stupar RM and Springer NM (2006) Cis -transcriptional variation in maize inbred lines B73 and Mo17 leads to additive expression patterns in the F1 Hybrid. Genetics 173: 21992210 Sturm A (1999) Invertases. Primary structures, func tions, and roles in plant development and sucrose partitioning. Plant Physiology 121: 1-7 Sturm A, Tang GQ (1999) The sucrose-cleaving enzymes of plants are crucial for development, growth and carbon partitioning. Trends in Plant Science 4: 401-407 Suzuki M, Kao CY, Cocciolone S, McCarty DR (2001) Maize VP1 complements Arabidopsis abi3 and confers a novel ABA/auxin in teraction in roots. Plant Journal 28: 409-418 Suzuki M, Ketterling MG, Li QB, McCarty DR (2003) Viviparous1 alters global gene expression patterns through regulation of abscisic acid signaling. Plant Physiology 132: 1664-1677 Swanson-Wagner RA, Joa Y, DeCook R, Borsuk LA, Nettleton D, Schnable PS (2006) All possible modes of gene action are observed in a global comparison of gene expression in a maize F1 hybrid and its inbred parents. Proc Natl Acad Sci USA 103: 6805-6810 Taliercio EW, Kim JY, Mahe A, Shanker S, Choi J, Cheng WH, Prioul JL, Chourey PS (1999) Isolation, characterizati on and expression analyses of two cell wa ll invertase genes in maize. Journal of Plant Physiology 155: 197-204 Trouverie J, Chateau-Joubert S, Th evenot C, Jacquemot MP, Prioul JL (2004) Regulation of vacuolar invertase by abscisic acid or glucose in leaves and roots from maize plantlets. Planta 219: 894-905 Trouverie J, Prioul JL (2006) Increasing leaf export and grai n import capacities in maize plants under water stress. Functional Plant Biology 33: 209-218 Tuberosa R, Salvi S (2006) Genomics-based approaches to improve drought tolerance of crops. Trends in Plant Science 11: 405-412

PAGE 146

146 Tymowska-Lalanne Z, Kreis M (1998) Expression of the Arabi dopsis thaliana invertase gene family. Planta 207: 259-265 van Ruissen F, Ruijter JM, Schaaf GJ, Asgh arnegad L, Zwijnenburg DA, Kool M, Baas F (2005) Evaluation of the similarity of gene expression data estimated with SAGE and Affymetrix GeneChips. BMC Genomics 6: 9I Vargas WA, Pontis HG, Salerno GL (2007) Differential expressi on of alkaline and neutral invertases in response to environmental stre sses: characterization of an alkaline isoform as a stress-response enzyme in wheat leaves. Planta 226: 1535-1545 Vaughn MW, Harrington GN, Bush DR (2002) Sucrose-mediated tr anscriptional regulation of sucrose symporter activity in the phloem. Proceedings of the National Academy of Sciences of the United States of America 99: 10876-10880 Vielle-Calzada JP, Baskar R, Grossniklaus U (2000) Delayed activat ion of the paternal genome during seed development. Nature 404: 91 Vij S, Tyagi AK (2007) Emerging trends in the functional genomics of the abiotic stress response in crop plants. Plant Biotechnology Journal 5: 361-380 Vroh Bi I, McMullen MD, Sanchez-Villeda H, Schroeder S, Gardiner J, Polacco M, Soderlund C, Wing R, Fang Z, Coe Jr. EH (2006) Single nucleotide polymorphisms and insertion-deletions for genetic markers and anchoring the mai ze fingerprint contig physical map. Crop Sci. 46: 12-21 Wachter R, Langhans M, Aloni R, Gotz S, We ilmunster A, Koops A, Temguia L, Mistrik I, Pavlovkin J, Rascher U, Schwalm K, Koch KE, Ullrich CI (2003) Vascularization, high-volume solution flow, and localized roles for enzymes of sucrose metabolism during tumorigenesis by Agrobacterium tumefaciens. Plant Physiology 133: 1024-1037 Weber APM, Weber KL, Carr K, Wilkerson C, Ohlrogge JB (2007) Sampling the Arabidopsis transcriptome with massively parallel pyrosequencing. Plant Physiol 144: 32-42 Welcker C, Boussuge B, Benciv enni C, Ribaut JM, Tardieu F (2007) Are source and sink strengths genetically lin ked in maize plants subjected to water deficit? A QTL study of the responses of leaf growth and of Anthesis-Silking Interval to water deficit. Journal of Experimental Botany 58: 339-349 Weschke W, Panitz R, Gubatz S, Wan g Q, Radchuk R, Weber H, Wobus U (2003) The role of invertases and hexose transp orters in controlling sugar ratios in maternal and filial tissues of barley caryopses during early development. Plant Journal 33: 395-411 Westgate ME, Boyer JS (1985) Osmotic adjustment and the inhibition of leaf, root, stem and silk growth at low water potentials in maize. Planta 164: 540-549

PAGE 147

147 Westgate ME, Boyer JS (1986) Reproduction at low silk and pollen water potentials in maize. Crop Science 26: 951-956 Whipple CJ, Ciceri P, Padilla CM, Ambrose BA, Bandong SL, Schmidt RJ (2004) Conservation of B-class floral homeotic gene function between maize and Arabidopsis. Development 131: 6083-6091 Wiese A, Elzinga N, Wobbes B, Smeekens S (2004) A conserved upstream open sucroseinduced repression reading frame me diates of translation. Plant Cell 16: 1717-1729 Winter H, Huber SC (2000) Regulation of sucrose metabol ism in higher plants: Localization and regulation of activity of key enzymes. Critical Reviews in Biochemistry and Molecular Biology 35: 253-289 Wobus U, Weber H (1999) Seed maturation: genetic programmes and control signals. Curr Opin Plant Biol 2: 33-38 Woodson WR and H Wang (1987) Invertases of carnation petals. Partial purification, characterization, and changes in activity during petal grow th. Physiologia Plantarum 71: 224-228 Wormit A, Trentmann O, Feifer I, Lohr C, Tj aden J, Meyer S, Schmidt U, Martinoia E, Neuhaus HE (2006) Molecular identification and physiological characterization of a novel monosaccharide transporter from Arabidopsis involved in vacuolar sugar transport. Plant Cell 18: 3476-3490 Wright SI, Bi IV, Schroeder SG, Yamasaki M, Doebley JF, McMullen MD, Gaut BS (2005) The effects of artificial selection on the maize genome. Science 308: 1310-1314 Xiao W-Y, Sheen J, Jang J-C (2000) The role of hexokinase in plant sugar signal transduction and growth and development. Plant Mol Biol 44: 451-461 Xu J, Avigne WT, McCarty DR, Koch KE (1996) A similar dichotomy of sugar modulation and developmental expression affects both path s of sucrose metabolism: Evidence from a maize invertase gene family. Plant Cell 8: 1209-1220 Yamada K, Lim J, Dale JM, Chen H, Shinn P, Palm CJ, Southwick AM, Wu HC, Kim C, Nguyen M, et al. (2003) Empirical analysis of transc riptional activity in the Arabidopsis genome. Science 302: 842-846 Yan H, Yuan W, Velculescu VE, Vogelstein B, Kinzler KW (2002) Allelic variation in human gene expression. Science 297: 1143 Yun HS, Yoon IS, Kang BG (2002) Rapid repression of va cuolar invertase in mungbean hypocotyl segments and regulation by sucrose, auxin and light. Plant Growth Regulation 38: 181-189

PAGE 148

148 Zhang X, Madi S, Borsuk L, Nettleton D, El shire RJ, Buckner B, Janick-Buckner D, Beck J, Timmermans M, Schnable PS, Scanlon MJ (2007) Laser microdissection of narrow sheath mutant maize uncovers novel gene expres sion in the shoot apic al meristem. PLoS Biology e101: 1040-1052 Zhuang Y, Ren G, Yue G, Li Z, Qu X, Hou G, Zhu Y, Zhang J (2007) Effects of water deficit stress on the transcriptomes of developing im mature ear and tassel in maize. Plant Cell Rep 26: 2137-2147 Zimmermann P, Hirsch-Hoffmann M, Hennig L, Gruissem W (2004) GENEVESTIGATOR. Arabidopsis Microarray Database and Analysis Toolbox. Plant Physiol 136: 2621 Zinselmeier C, Jeong BR, Boyer JS (1999) Starch and the control of kernel number in maize at low water potentials. Plant Physiology 121: 25-35 Zinselmeier C, Sun YJ, Hele ntjaris T, Beatty M, Ya ng S, Smith H, Habben J (2002) The use of gene expression profiling to dissect the stress sensitivity of reproductive development in maize. Field Crops Research 75: 111-121 Zinselmeier C, Westgate ME, Schussler JR, Jones RJ (1995) Low water potential disrupts carbohydrate metabolism in maize ovaries. Plant Physiology 107: 385-391

PAGE 149

149 BIOGRAPHICAL SKETCH Andrea Lee Eveland was born on May 2nd, 1977 to Frank and Joan Eveland on Long Island of New York. Andrea first developed a deep appreciation for biology through the guidance Dr. Alan Katz, her biology teacher at Commack High School. She became fascinated by plants in her botany studies at Binghamton University (SUNY) where, as an undergraduate, she served as TA for a botany lab for 2 years. Andrea earned her Bachelors of Science degree for a dual major in biology and environmental science in May, 20 00. At graduation, she was awarded the James D. Grierson foundation award for Excellence in Botany. During her time at Binghamton University, Andrea studied tropi cal ecology in Costa Rica and was involved in sustainable agriculture and forest restorati on projects. She also held posit ions at various nurseries and greenhouses as manager and consultant. In the fall of 2000, Andrea moved to San Diego, CA, and joined a plant pathology lab group at Torry Mesa Research Institute, Syngenta In c, as a research assistant where she learned molecular-based techniques. In June 2002 sh e entered the Plant Mo lecular and Cell Biology (PMCB) program at the University of Florid a as a Ph.D. student and in spring 2003 began research in Karen Kochs lab. Here she was able to combine new-found interests in molecular biology with her love for plant physiology. During her time as a Ph.D. student, Andrea received a number of awards including the ASPB/Pioneer Hi-Bred International Graduate Student Prize, the IFAS Scholarship for Women in Agriculture, awards for talks at the annual PMCB retreat, and several travel grants. She also had the opportunity to present her work at a number of international meetings and meet with many influential people in her field.