 

.(
a: r .
‘ . . 15.21,.»
a «5%. ”M...
utdaaun u ﬂth
. : 2
z.

 

out

, lit}.

.V Tﬂ s.
’1

536‘.

 

 

 

1 1-
gKu \I
\u :33:

 

THESIS

8
QC; 6) * —--
I F— LIBRARY
| Michigan State .
University

 

 

This is to certify that the
dissertation entitled

Deciphering the molecular mechanisms of rootstock
induced dwarﬁng in cherries (PRUNUS SPP.)

presented by

Constantinos Prassinos

has been accepted towards fulﬁllment
of the requirements for the

PhD degree in Deg‘tment of Forestry

 

 

/

 

/ ‘ Major Professor’s Signature
ﬂ/M/o 7

Date

MSU is an afﬁnnative-action, equal-opportunity employer

on-.-.-o-u-o-n-n-- .—

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DaleDueindd-p.1

 

 

DECIPHERING THE MOLECULAR MECHANISMS OF ROOTSTOCK INDUCED
DWARF ING IN CHERRIES (Prunus spp.)

By

Constantinos Prassinos

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILISOPHY
Department of Forestry

2007

ABSTRACT

DECIPHERING THE MOLECULAR MECHANISMS OF ROOTSTOCK INDUCED
DWARF ING IN CHERRIES (Prunus spp.)

By

Constantinos Prassinos

Rootstock induced dwarﬁng has been one of the major breakthroughs in orchard
management in the twentieth century. The majority of the cherry rootstocks have been
produced in the last 30 years. Nevertheless, breeding of new rootstocks has proven
challenging, due to the lack of understanding of the dwarﬁng phenomenon. This project
explores the phenotypic and genetic differences between cherry graft combinations that
exhibit varying degrees of vigor. Growth data have indicated a consistent cessation of
shoot grth across rootstocks of varying vigor and across growing seasons. The initial
rate of shoot elongation was the same for all graft combinations tested, but dwarﬁng graft
combinations showed faster cessation of shoot growth. The same pattern was observed
for the number of nodes added during the elongation of the shoot. The average metamer
length though was not affected between grafts, indicating that cessation of growth is due
to reduced cell grth and expansion at the apical meristem. This was further conﬁrmed
by the absence of signiﬁcant difference in the size or number of cells within the metamer,
between grafts. Complementary DNA Ampliﬁed Fragment Length Polymorphism of
shoot and graft union samples revealed a high degree of co-regulation in gene expression
between the dwarﬁng ‘Bing’/Gi5 and semi-vigorous ‘Bing’/Gi6 graft combinations. Few
genes showed differential expression between the two graft combinations. F orty-three of

those genes were differentially expressed in the shoot samples and 56 in the graft union

samples. The differentially expressed genes had a variety of functions with the most
interesting being a group of genes previously involved in brassinosteroid signaling. The
analysis of gene expression also revealed the presence of the Cherry Virus A in the
dwarﬁng combination ‘Bing’/Gi5. Screening of rootstocks that confer different degrees
of vigor did not show any correlation between the presence of the virus and the vigor of
the rootstock. Also, the absence of the virus from some rootstocks is circumstantial rather
than due to resistance, which was shown by screening different scions grafted on the
same rootstock variety. The current study provides an initial cataloging of genes that may
be involved in the process of rootstock-induced dwarﬁng. No certain conclusion can be
drawn from these results and further study will be necessary to identify which of these

genes contribute signiﬁcantly to this phenomenon.

COPYRIGHT
CONSTANTINOS PRASSINOS
2007

DEDICATION

To my parents

AKNOWLEDGEMENTS

I would like to thank my committee members Dr.Kyung-Hwan Han, Dr.Gregory
Lang, Dr.Amy Iezzoni and Dr.Wayne Loescher for their continuous support and
guidance. I am grateful to Dr.Han, Dr.Lang and Dr.lezzoni for establishing the rootstock
induced dwarﬁng experiment, which addressed a question that had been in my mind since
my undergraduate years. They gave me a great opportunity to work on this complicated
phenomenon. Particularly, I would like to thank Dr.Han for giving me the opportunity to
work in his lab and perform state-of—the-art research on tree molecular biology and tissue
culture. I would also like to thank him for his generous ﬁnancial support throughout my
studies. Without this support my research would not have been possible. I would like to
thank Dr.Lang for his continuous support with orchard management, for providing
support with tree measurements and for the helpﬁil advice. I am also grateﬁil to him for
the support he provided me to attend the 5th International Cherry Symposium in Turkey. I
would like to thank Dr.lezzoni for introducing me into the world of plant breeding, which
helped me understand the system I chose to work as well as other modern agricultural
systems. Finally, I would like to thank Dr.Loescher for his helpful discussions and
advice.

Many thanks go to DrJim Olmstead and Zoltan Horvath for their help in
maintaining the orchard and taking measurements. Special thanks to Jim whose help in

the orchard has been instrumental into the continuous progress of this project.

vi

Special thanks to Dr. Jae Heung Ko for his help in the establishment of the
cDNA-AF LP experiment and for his helpﬁil advice throughout my project. I am also
grateful to other members of the HanLab, Dr. Sunchung Park, Dr. 800 Kyung Oh, Dr.
Jaemo Yang, Andrew Park and Sang Hyuck Park for their help with protocols,
experiments and discussions. Special thanks to Merilyn Ruthig, the HanLab manager, for
her precious help in ordering, making sure everything runs ﬁne and for editing my
writing.

Many thanks to Dr.Kenneth Sink and Dr.David Douches for allowing me to be a
teaching assistant in their course CSS 451 “Techniques in Plant Biotechnology”. I gained
very good experience in preparing for a lecture/laboratory, interact with students and
guide them through the experiments.

I am grateful to the RTSF staff for their helpful advice on high-throughput
sequencing and microarrays. More speciﬁcally I would like to thank Annette Thelen and
Jeff Landgraf for the helpful discussions and advices.

Here I would like to mention my undergraduate advisor at the Agricultural
University of Athens, Dr. Polydeﬂcis Hatzopoulos for his encouragement to pursue this
degree.

Without friendship life is very difﬁcult. During my 6 years of studies I have met a
signiﬁcant number of people many of whom became my friends. I would like to thank
them for being there when I needed them, for trusting me, for adding color into my life

and for teaching about life outside the lab.

vii

Finally, I would like to thank my parents George and Georgia for always trying to
help me the best they could even from so far away and my brothers Nikita and Pavlo for

being supportive and interested in my progress.

viii

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... xii
LIST OF FIGURES ...................................................................................................... xiii
LITERATURE REVIEW ............................................................................................. 1
Historical background .......................................................................................... 2
Signiﬁcance of rootstocks and Rootstock Induced Dwarﬁng (RID) .................. 3
The apple system: Hypotheses and advances ..................................................... 4
Rootstock Induced Dwarﬁng in other systems ................................................... 9
Response of Cherry Trees to Grafting ................................................................ 10
Interstocks ............................................................................................... 10
Budding height ........................................................................................ 11
Phenolics in grafted cherry trees .............................................................. 12
Hormonal control of growth ................................................................... 14
Hydraulic conductance in grafted trees ................................................... 14
Long distance transport of macromolecules ....................................................... l6
Hypothesis and objectives .................................................................................. 20
LITERATURE CITED .................................................................................................... 22

CHAPTER 1: GROWTH CHARACTERISTICS OF DWARF AND SEMI-

VIGOROUS CHERRY GRAF T COMBINATIONS .............................. 29
INTRODUCTION ........................................................................................................... 30
MATERIALS AND METHODS ..................................................................................... 33

Plant material ....................................................................................................... 33
Measuring ............................................................................................................ 34
Laser scanning confocal microscopy of pith and epidermal cells ....................... 35
Statistical analysis .............................................................................................. 36
RESULTS ........................................................................................................................ 36
Monitoring tree growth in three growing seasons .............................................. 36
Expansion of trunk girth does not differ signiﬁcantly between graft
combinations ........................................................................................................ 37
Main-shoot elongation in three growing seasons and different graft
combinations ........................................................................................................ 45
Measurements of intemode number and metamer length .................................... 54
Counting of cell size and number in metamers .................................................... 58
DISCUSSION ................................................................................................................. 6O
LITERATURE CITED .................................................................................................... 65

ix

CHAPTER 2: IDENTIFICATION OF GENES EXPRESSED AT THE CRITICAL
POINTS IN GROWTH BETWEEN DWARF AND SEMI-VIGOROUS

GRAF T COMBINATIONS ..................................................................... 68
INTRODUCTION ........................................................................................................... 69
MATERIALS AND METHODS .................................................................................... 72

Plant material ...................................................................................................... 72
Sampling ............................................................................................................. 73
RNA extraction ................................................................................................... 73
cDNA-AF LP analysis ........................................................................................ 75

i) Preparation of double stranded cDNA ....................................... 75

ii) Preparation of the primary template ........................................... 76

iii) Preparation of the secondary template ........................................ 77

iv) Selective ampliﬁcation of the secondary template ..................... 78

Excision, cloning and sequencing of differentially expressed bands ................. 81
Microarray construction ..................................................................................... 82
Microarray hybridization .................................................................................... 83
Microarray analysis ............................................................................................ 84
Northern hybridization ....................................................................................... 85
RESULTS ....................................................................................................................... 86

Complementary DNA Ampliﬁed Fragment Length Polymorphism (cDNA-AF LP)
analysis of scion main shoots in grafts showing differential cessation of

growth .................................................................................................................. 86
cDNA-AF LP analysis of the graft union area during shoot growth cessation 88
Conﬁrmation of cDNA-AF LP with the use of microarrays ............................... 92
DISCUSSION ................................................................................................................. 99
Discrepancies in phenotype between graft combinations are accompanied by
genetic changes ................................................................................................... 99
The graft union region is transcriptionaly diverse between graft combinations at
the time of differentiation in shoot growth ......................................................... 102
Parallel gene regulatory pathways between cherry and apple graft combinations
............................................................................................................................. 104
Genes implicated in brassinosteroid response .................................................... 104
Conclusions ......................................................................................................... 105
LITERATURE CITED .................................................................................................... 107

CHAPTER 3: CHERRY VIRUS A HAS NO DIRECT EFFECT ON ROOTSTOCK-

INDUCED DWARF ING OF GRAFTED CHERRY TREES ................. 111
INTRODUCTION ........................................................................................................... 1 12
MATERIALS AND METHODS .................................................................................... 116

Plant material ...................................................................................................... 116
RNA extraction ................................................................................................... 118
cDNA-AF LP analysis ........................................................................................ 118
Sequencing ........................................................................................................... 1 18

Reverse Transcription Polymerase Chain Reaction (RT-PCR) .......................... 118

Growth measurements ......................................................................................... 119
Northern hybridization analysis ........................................................................ 119
RESULTS ........................................................................................................................ 120
Initial detection of the Virus .............................................................................. 120
CVA is not associated with rootstock control of scion vigor ............................. 121
The absence of CVA is not linked to rootstock resistance ................................. 125
DISCUSSION ................................................................................................................. 129
LITERATURE CITED .................................................................................................... 131
CONCLUSIONS, SUMMARY AND IDEAS ............................................................. 133

xi

LIST OF TABLES

Table 1.1: Growth measurements of rootstocks produced in the Giessen cherry rootstock
breeding program (adapted from Seif and Gruppe, 1985). Letters in shoot length indicate
signiﬁcant similarities .............................................................................. 32

Table 1.2: Growth characteristics of cherry 'Bing' scions grafted on Gi5 and Gi6
rootstocks across three growing seasons.In 2002 and 2005 measurements were taken
from 2-year old trees, while in 2003 measurements were taken from 3-year old trees.
Letters denote signiﬁcant difference at an a=0.05. (n=9 except for Trunk diameter where
n=10) ................................................................................................. 44

Table 2.1: Genes differentially expressed at the upper shoot microarray experiment.
Positive fold change values refer to B5 up-regulation while negative values to B6 up-
regulation. 3/6: 3 June, 20/6: 20 June, 3/7: 3 July, B5: ‘Bing’/Gi5, B6: ‘Bing’/Gi6, R:
Rootstock, S: Scion, GU: Graft Union, Sh: Shoot, ns: non-singiﬁcant ............................. 95

Table 2.2: Genes differentially expressed in the graft union microarray experiment.
Positive fold change values refer to B5 up-regulation; negative values to B6 up-
regulation. R: Rootstock, GU: Graft Union, S: Scion, Sh: Shoot, B5: ‘Bing’/Gi5, B6:
‘Bing’/Gi6, ns: non-signiﬁcant .................................................................... 96

Table 3.1: cDNA-AF LP fragments with homology to CVA. Location of each fragment is
given in nucleotides on the CVA genome. Size of each fragment is given in bases. . 124

Table 3.2: End of season shoot length of ‘Hedelﬁngen’ scions grafted on 9 different
rootstocks. Values present the average shoot length on July 29 and letters denote
statistical differences at a=0.05 .................................................................. 127

xii

LIST OF FIGURES

Figure 1.1: ‘Bing’ cherry tree trunk diameter growth across three growing seasons and
across three graft combinations. A. Growing season of 2002, B. Growing season of 2003,
C. Growing season of 2005. In 2002 and 2005 measurements were taken from 2—year old
trees, while in 2003 measurements were taken from 3-year old trees. Error bars indicate
standard error ........................................................................................ 39

Figure 1.2: Trunk diameter growth rate in ‘Bing’/Gi5 and ‘Bing’/Gi6 cheery trees in 2002
(A), 2003 (B) and 2005(C). In 2003 trunk expansion rate was calculated also for
‘Bing’/Edabriz trees. Expansion rate was calculated as the ratio of weekly trunk diameter
increase divided by the number of days between measurements. In 2002 and 2005
measurements were taken from 2-year old trees, while in 2003 measurements were taken
from 3-year old trees ............................................................................... 41

Figure 1.3: (A) Trunk diameter of ungrafted Gi5 and Gi6 rootstocks measured in 2003,
(B) Trunk expansion rate of the same trees expressed as the ratio of weekly trunk
expansion divided by the number of days between measurements. Error bars indicate
standard error ........................................................................................ 43

Figure 1.4: Main shoot length of ‘Bing’/Gi5, ‘Bing’/Gi6 cherry trees taken in 2002 (A),
2003 (B) and 2005 (3). Measurements were taken from bud break to bud set.
Measurements were also taken for ‘Bing’/Edabriz trees in 2003. In 2002 and 2005
measurements were taken from 2-year old trees, while in 2003 measurements were taken
from 3-year old trees. Error bars indicate standard error. ..................................... 47

Figure 1.5: Main shoot elongation rate in ‘Bing’/Gi5 and ‘Bing’/Gi6 cherry trees in 2002
(A), 2003 (B) and 2005(C). In 2003 shoot elongation rate was calculated also for
‘Bing’/Edabriz trees. Expansion rate was calculated as the ratio of weekly shoot
elongation divided by the number of days between measurements. In 2002 and 2005
measurements were taken from 2-year old trees, while in 2003 measurements were taken
from 3-year old trees. Error bars indicate standard error ....................................... 49

Figure 1.6: Shoot growth characteristics of ungrafted rootstocks Gi5 and Gi6 in the
growing season of 2003. (A) Shoot length from bud break to bud set and (B) shoot
elongation rate expressed as the ratio of weekly shoot growth divided by the number of
days between measurements. Error bars indicate standard error .............................. 51

xiii

Figure 1.7: Cumulative bud set in ‘Bing’/Gi5 and ‘Bing’/Gi6 trees for the growing
seasons of 2002 (A), 2003 (B) and 2005 (C). Bud set was recorded for ‘Bing’/Edabriz,
Gi5 and Gi6 ungrafted rootstocks in the growing season of 2003 (B). Cumulative bud set
is the percentage of trees for which the main shoot has set bud. The week when shoot
growth cessation occurred was considered as the time of bud set (n=9-11). In 2002 and
2005 measurements were taken from 2-year old trees, while in 2003 measurements were
taken from 3-year old trees ........................................................................ 52

Figure 1.8: The number of nodes was measured in the main shoot of ‘Bing’/Gi5 and
‘Bing’/Gi6 in the growing season of 2003 (A) and 2005 (B). In 2005 measurements were
taken from 2-year old trees, while in 2003 measurements were taken from 3-year old
trees. Node number was also measured in 2-year old ‘Bing’/Edabriz trees in the growing
season of 2003 (A). Error bars indicate standard error ........................................ 56

Figure 1.9: Metamer length of ‘Bing’/Gi5 and ‘Bing’/Gi6 trees in the growing season of
2003 (A) and 2005 (B). In 2005 measurements were taken from 2-year old trees, while in
2003 measurements were taken from 3-year old trees. Metamer length was also
calculated for 2-year old ‘Bing’/Edabriz trees in the growing season of 2003 (A).
Metamer length was derived by the division of shoot length to node number for each
measurement point. Higher metamer lengths were observed in June due to higher
elongation rates. Reduction of metamer length reﬂects shorter intemodes as shoot growth
rates dropped near the time of shoot growth cessation ......................................... 57

Figure 1.10: Cell size and number in the pith and epidermis of the 6th intemode collected
on the 3rd of July and the 10th intemode collected on the 29th of July of ‘Bing’ sweet
cherr'y trees on Gisela 5 (Gi5) and Gisela 6 (Gi6) rootstocks. (A) Pith cells were counted
in radial sections of the intemode. (B) Cell ﬁles in radial sections of the intemode were
used to measure the cell number per millimeter of pith. (C) Epidermal cell ﬁles were
counted in tangential sections of the bark. (D) Cell number per millimeter of epidermis
was counted in cell ﬁles of the 6th intemode. Measurements were taken on 2-year old
trees. Error bars indicate standard error ........................................................... 59

Figure 2.1: cDNA-AFLP analysis at the main shoot. (A) Portion of a cDNA-AF LP gel
produced by four primer pairs and showing high degree of co-regulation between
samples. (B) Detail of gene expression patterns putatively related to shoot growth
cessation, identiﬁed in 8 cDNA-AF LP gels as the one shown in (A). 3/6: 3 June, 20/6: 20
June, 3/7: 3 July, B5: ‘Bing’/Gi5, B6: ‘Bing’/Gi6 .............................................. 90

xiv

Figure 2.2: cDNA-AF LP analysis at the graft union. (A) Portion of a cDNA-AF LP gel
produced by four primer pairs and showing high degree of co-regulation between
samples. (B) Detail of interesting expression patterns identiﬁed in 8 cDNA-AF LP gels as
the one shown in (A). R: Rootstock, GU: Graft Union, S: Scion, B5: ‘Bing’/Gi5, B6:
‘Bing’/Gi6 .......................................................................................... 91

Figure 3.1: Genome organization of Cherry Virus A. Shaded areas denote the position of
the proteins as described above these areas. ORF 1: Open Reading Frame], ORF2: Open
Reading F rame2, DUF: Domain of Unknown Function, RdRP: RNA depended RNA
Polymerase. ORF 1 is 2,360aa and ORF2 is 463aa .......................................... 115

Figure 3.2: Cherry Virus A detection. (A) cDNA-AF LP proﬁles across tree sections and
growing season dates for different parts of the CVA genome that was digested with
ApoI/MseI restriction enzymes, R: rootstock, GU: graft union, Sc: scion, BS:
‘Bing’/GiselaS, B6: ‘Bing’/Gisela6. (B) The position of each cDNA-AF LP fragment on
the 7,3 83bp CVA genome is indicated by arrows. Direction of the arrow is from the ApoI
to MseI restriction site. (C) Alignment of the translated cDNA-AF LP fragments with the
two open reading frames (ORF) of CVA. The alignment was obtained from the BlastX
results. Amino acid conservation is indicated as follows; black: identical, gray:
conserved, white: non-conserved. ORFl and ORF 2: open reading frames of the CVA
genome with GenBank accession numbers CAA57896 and CAA57897, respectively.
Amino acid numbering of the cDNA-AF LP fragments is based on the individual TDF
amino acid sequence, while for CVA ORF 1 and ORF2 it is based on the complete amino
acid sequence. Sorting of the sequences is based on the CVA ORFs ....................... 122

Figure 3.3: CVA detection in the sweet cherry variety ‘Hedelﬁngen’ grafted on 9
rootstocks that exert different degrees of vigor to the scion. (A) RT-PCR ampliﬁcation of
a 1532bp fragment of the CVA genome using primers CVA4-CVA5. W158: Weiroot158,
W10: Weiroot10. Rootstocks are arranged from the most dwarﬁng (GiSelA3) to the most
vigorous (Erdi V). (B) Northern blot hybridization was used to conﬁrm the RT—PCR
result. The image is aligned to the RT-PCR image in (A). rRNA denotes the RNA
loading control .................................................................................... 126

Figure 3.4: CVA detection in 5 sweet cherry varieties grafted on 3 rootstocks. (A) RT-
PCR ampliﬁcation of a 1532bp fragment of the CVA genome using primers CVA4-
CVAS. G15: GiSelAS, G16: GiSelA6. (B) Northern blot hybridization was used to
conﬁrm the RT-PCR result. The image is aligned to the RT-PCR image in (A). rRNA
denotes the RNA loading control ............................................................... 128

XV

LITERATURE REVIEW

Historical background

Early reports on the use of rootstocks came from ancient Greece and Rome
(Hedrick, 1915). Nevertheless, the interest at that time was concentrated in the
acquisition of new cherry varieties for the two cherry species that are in commercial use
up to this day, namely Prunus avium L. and Prunus cerasus L. The most extended report
of cherry varieties was performed by Pliny, who described the varieties collected by the
wealthy Roman Lucullus (DeCandole, 1890). For many centuries and through the middle
ages, cherry cultivation and varieties did not develop further (Hedrick, 1915). It was only
after the 16th and 17th centuries that cherry cultivation started to develop once again.
Nevertheless, it was only at the level of fruit varieties that progress was made. In the
beginning of the 20th century, there were two main choices for rootstocks. These were the
‘Mazzard’ (P. avium) and the ‘Mahaleb’ (P. mahaleb L.), both propagated through
seedlings. In an early comparison of rootstock performance between ‘Mazzard’ and
‘Mahaleb’, Hendrick (1915) list of the advantages of the second over the ﬁrst: a) cold
hardiness, b) dwarﬁng ability, c) precocity, d) no change in fruit size and e) better
adaptation to diverse soils and the disadvantages; a) weaker graft union, b) shorter life
cycle and c) lower yield. Hedrick (1915) discusses a number of other rootstocks that were
less popular and used in special environments that mainly had to do with low
temperatures or low water availability. These were the Russian Cherry which was used in
cold regions of the United States, the pigeon cherry Prunus pensylvam'ca L.f. (Linnaeus
ﬁlius) in cold regions with the potential to dwarf scions, the sand cherry Prunus pumila
L. in cold and dry areas and a Japanese variety of Prunus pseudocerasus L. used in Japan

for its cold hardiness and ability to root easily. Hedrick also refers to the ability of

rootstocks to dwarf trees, with ‘Morello’ (P. cerasus) producing dwarf and ‘Mahaleb’
very dwarf trees. Nevertheless, dwarfness could be accomplished only through the
appropriate pruning and working of the crown.

During most of the 20th century ‘Mahaleb’ and ‘Mazzard’ remained the
predominant cherry rootstocks (Perry, 1987). They were propagated through seed and
only the clone F12/1 of ‘Mazzard” was clonally propagated. Other rootstocks that have
been used in a smaller extent are the ‘Stockton Morello’ and ‘Colt’ (P. avium x P.
pseudocerasus). Since the 80’s cherry rootstock breeding has been accelerated through
the production of interspeciﬁc hybrids. Some of the most important rootstock breeding
programs were developed in Germany with Weiroot, Gisela and Pi-Ku series, in Italy the
CAB series and in Denmark the DAN series (Hrotko, 2005). Individual rootstocks with
considerable success were also produced such as Edabriz, Victor, Damil and Camil

(Hrotko, 2005).

Significance of rootstocks and rootstock induced dwarﬁng (RID)

One of the most important cultural advances in temperate tree fruit production has
been the development and adoption of dwarﬁng rootstocks. Trees on dwarﬁng rootstocks
can exhibit several economically important traits, including precocious ﬂowering,
increased yield, reduced tree height and disease/virus resistance (Lang et al. 1997;
Webster, 1998; Atkinson and Else, 2001). The coexistence of two organisms of the same
or different species background, in a single plant structure requires a high degree of co-

regulation in molecular, biochemical and physiological processes. The rootstock is the

source of nutrients that are transported to the scion to be metabolized, and the scion is the
source of photoassimilates, which will partially reach the rootstock for maintenance.
Also, the rootstock is the receiver of many signals from the soil environment, while the
scion is the receiver of the signals from the open air. Survival of the grafted tree depends
largely on the ability of the rootstock and the scion to communicate effectively.

Reduced tree height, also known as dwarﬁsm, conferred by the rootstock to the

scion remains a scientiﬁc mystery.

The apple system: hypotheses and advances

Signiﬁcant progress in the understanding of RID has been made through research
performed in the apple rootstocks. Apple has a longer history of rootstock breeding and
an extended array of rootstock varieties.

Early attempts to explain the dwarﬁng phenomenon were summarized in 1956 by
Beakbane, who listed ﬁve theories for the mechanism of the rootstock effect. The ﬁrst
theory suggested that competition for resources between various parts of the grafted tree
results in the restriction of scion growth. According to the author dwarﬁng rootstocks
tend to attract more resources due to higher content in living tissue. The second theory
suggested that transport of water and metabolites differs between rootstocks due to the
differences in vessel element diameter. Dwarﬁng rootstocks contain smaller vessels
compared to vigorous rootstocks thus having lower transport capacity. The third theory
suggests that the ratio of living tissue to plant surface affects the amount of oxygen that is

available for respiration. Dwarf trees with a high ratio tend to absorb less oxygen than

necessary thus having reduced growth as a result. The fourth hypothesis involves the
ability of the rootstock to transport ions, based on the percentage of live tissue. Finally, a
ﬁfth theory postulates that each rootstock variety has different capacity to form
elaborated compounds, such as phenolics. A year later Rogers and Beakbane (1957)
reduced the number of mechanisms to three. These were nutrient availability, nutrient
transport and auxin metabolism. In 1967, Tubbs proposed four mechanisms to explain the
phenomenon, with the assumption that these mechanisms are not exclusive, but each
contributes partially to the phenomenon. The mechanisms were i) nutrient availability, ii)
variation in metabolism between tissues, iii) effect of morphogenesis in resource
production and allocation and iv) grth regulators, such as grth promoting hormones
and grth inhibitors. These theories have been revised or rejected through time. More
sophisticated chemical and physiological analysis methods have lead to more detailed
and substantiated hypotheses. These are described in the following paragraphs.

The ﬁrst hypothesis proposes that auxin, which is produced in the aerial parts of
the grafted tree, is transported at different concentrations between grafts on different
rootstock genotypes, thus affecting cytokinin production in the root, which further causes
differences in shoot grth (Lockard and Schneider, 1981; Webster, 1998). Lockard and
Schneider (1981) proposed that auxin is degraded during its flow from the scion to the
rootstock. Auxin is degraded by enzymes such as 1AA oxidase, peroxidase and phenols.
Concentration of these compounds differs between graft combinations, which results in
differences in auxin concentration that reach the root. Some support for this hypothesis
has been provided in recent years for the apple rootstocks (Soumelidou et al. 1994a;

Kamboj et al. 1997). Auxin was found to be transported at lower rates in shoots of the

dwarﬁng rootstock M.9, than the vigorous MM.111 (Soumelidou et al. 1994a). During
active grth of apple shoots, auxin is taken-up much faster in the vigorous roostocks
MM.104 and MM.111 than in the dwarﬁng M9 and M.27 (Kamboj et al. 1997).
Nevertheless, measurements of endogenous auxin in the bark of rootstock stems did not
show signiﬁcant differences in concentration across rootstocks of varying vigor and
across the growing season. In contrast, ABA concentration was found to be higher in
shoots of the M9 and M.27 dwarﬁng rootstocks than in the vigorous MM. 106 and
MM.111 (Kamboj et al. 1999a). ABA injections have been shown to have higher impact
on shoot growth retardation of dwarﬁng apple trees (Robitaille and Carlson, 1971). In
another report, Jones (1986) gives cytokinin a more important role in the control of scion
growth, through rootstocks or interstocks. Cytokinin in the form of zeatin and zeatin
riboside is present at higher concentration in the xylem sap of invigorating MM. 106
rootstocks than in dwarﬁng M.27 and M9 rootstocks (Kamboj et al. 1999b). The
concentration of these two cytokinins does not change between shoots of grafted or
ungrafted rootstocks (Kamboj et al. 1999b), excluding thus an effect of the graft union in
the acropetal transport of these hormones. Nevertheless, when the variety “Fiesta” was
used as a scion it grew more on M9 and M.27 rootstocks compared to the shoots of the
ungrafted rootstocks, but grafted trees were signiﬁcantly shorter than on MM. 106
rootstocks (Kamboj et al. 1999b).

The second hypothesis assumes that phenols accumulating at the graft union
reduce tissue viability and perhaps the rate of auxin break down (Lockard and Schneider,
1981). This hypothesis has been based on observations from heterologous systems and in

few reports on apple rootstocks. A number of phenolic compounds have been found to

act synergistically or antagonistically to IAA (Lockard and Schneider, 1981). The most
abundant of these compounds is phloridzin, which accumulates in the phloem of apple
stems and acts antagonistically to IAA. It was found to comprise 9.2% and 7.7% of the
dry weight in MM.111 and M.26 rootstocks respectively (Lockard and Schneider, 1981).
During active growth of the shoots, phloridzin concentration drops and increases again at
the onset of dormancy. Even though phloridzin has been given an important role in this
hypothesis, the concentrations mentioned above show that it is present in higher
concentration in the vigorous rootstocks. Even though its seasonal concentration
correlates well with inhibition of growth it has not been proven that phloridzin has a
regulatory role in this phenomenon.

The last hypothesis postulates that reduced tree size conferred by dwarﬁng
rootstocks is caused by reduced solute transport across the graft union (Atkinson et a1.
2003; Basile et al. 2003). In support of this hypothesis, Atkinson et al., (2003)
demonstrated that hydraulic conductance in apples increased with the vigor of the
rootstocks. In peach, reduced stem extension rates were signiﬁcantly correlated with
lower stern water potential when comparing dwarﬁng and invigorating rootstock/scion
combinations (Basile et al. 2003). This difference in transport can be caused by vessel
diameter and number. It was found that vessel elements at the graft union of dwarf trees
on the M9 rootstock were initially few, but larger than those of trees grafted on the semi
dwarﬁng MM106 (Soumelidou et al. 1994b). A year later though, vessel elements
became smaller in the dwarﬁng combination, thus suggesting a more unstable auxin

supply compared to the semi-dwarﬁng trees that had normal vessel development.

Nevertheless, current knowledge is still inconclusive on which is the exact
mechanism of RID in apples. All of the above hypotheses have produced a number of
important ﬁndings on the biology of grafted trees. Undoubtedly, the ﬁndings portray the
differences between grafts of varying vigor. What remains to be explored is the series of
events; a separation of primary and secondary effects.

In recent years research has focused on the genetic aspect of the dwarﬁng
phenomenon. Map based cloning was used by Rusholme et al. (2004) to screen a
population of rootstocks generated by crossing the dwarf apple rootstock M9 and the
vigorous rootstock Robusta 5. The population segregated for the dwarﬁng trait and was
used in a bulked segregant analysis with RAPD markers. Trees were grouped to four
categories based on visual assessment of dwarfness and trunk circumference. These two
phenotypic characters though were not always co-segregating. Thus, trees considered
dwarf based on height were vigorous in terms of trunk circumference and vice versa,
indicating the inﬂuence of more than one loci. Four RAPD loci were found to be linked
to the dwarﬁng phenotype. Detailed mapping identiﬁed a 2.50M region named DwI
containing the dwarﬁng locus. Nevertheless, 15 individuals from the cross between M9
to R5 that were classiﬁed as vigorous also contained M.9 alleles from the DwI locus.
This is expected since the dwarﬁng trait probably involves more than one gene or genetic
loci as mentioned above. This is the ﬁrst attempt to directly target the genes involved in
the dwarﬁng phenomenon and it has a promising future for the identiﬁcation of dwarﬁng
related genes in apple.

Jensen et al. (2004) applied a genomics approach in apple to compare the

dwarﬁng rootstock M9 to the vigorous M.7 rootstock. The analysis identiﬁed 92 genes

that were differentially expressed between the shoots of the two rootstocks. Of those, 56
were up-regulated in M9, the most important of which were cell cycle and signaling
related genes. This approach provided information on the gene expression differences
between dwarf and vigorous trees, but more work is needed to identify which of these
genes act at the initial stages of growth control.

Even though the above hypotheses and new advances are presented individually,
they can all be interrelated and orderly placed in the process of dwarﬁng. In the effort to
answer the complex question of RID in apples and other systems it is important to reduce
the complexity as much as possible. Thus it would be wise to focus on speciﬁc group of
genotypes and later expand the knowledge obtained to a larger collection of genotypes.
The apple system is and will continue to be the pioneer system in the quest for signals

that promote RID.

Rootstock Induced Dwarfing in other systems

The advances achieved towards the understanding of the dwarﬁng phenomenon in
apples can be used as guides for explaining dwarﬁng in other ﬁ'uit tree systems.
Nevertheless, one should be cautious when relying on the apple system to study RID in
other tree species. Even within the same species the causes of dwarﬁsm may differ

between genotypes.

Response of cherry trees to grafting

Similarities and discrepancies between the apple and cherry models are discussed
below in an attempt to compare the two mechanisms. It is critical to understand how the
two systems work, to avoid transfer of knowledge that will complicate either system
more than it is in reality. Unfortunately, current research in cherry RID is focusing
heavily on knowledge obtained in apples. Thus, more data are needed to establish a

model for RID in cherries.

Interstock contribution to RID

As discussed previously, rootstock varieties that have been used as interstocks in
apple grafts respond in the same fashion as when they are used as rootstocks. In cherries
there is conﬂicting evidence about the role of interstocks in the control of scion vigor.
Some report the inability of interstocks to dwarf scions (Jones, 1986; Webster, 1995,
1998), while others report marginal to signiﬁcant changes in vigor. In a sixteen year trial
using 14 different rootstocks as interstocks, grafted on P. avium or P. mahaleb seedlings,
the TCSA and crown volume of ‘Van’ or ‘Btittner’s Red’ scions were statistically
different from control trees without interstocks (Rozpara et al. 1998). Nevertheless, the
vigor of the rootstocks without interstocks was not reported. Thus, it cannot be concluded
whether these genotypes behave in a similar fashion when used as rootstocks or
interstocks. Furthermore, yield efﬁciency of the trees does not correlate well with tree
vigor, suggesting a putative effect of the age of these trees. In another trial when sour
cherry varieties were used as interstocks on P. mahaleb seedlings, there was no change in

vigor of ‘Van’ and ‘Germersdorﬁ oriés’ scions, while a P. fruticosa interstock caused a

10

signiﬁcant reduction in vigor (Hrotko et al. 1997,1998). Perry (1987) reported several
instances in which interstocks caused subtle reduction in tree vigor, and in only a few
combinations, did the reduction reach 20-30% under certain conditions. These results
suggest that the genotype of the plant material used in interstock grafting is the critical
factor affecting vigor of the resulting tree. This effect of the interstock does not seem to
be proportional to the size controlling ability of the same genotype when used as
rootstock. It should be ﬁrrther pursued whether such a response is the result of

compatibility or the physiology of the grafted material.

Budding height effect on the control of vigor

Budding height in cherries, as in the case of interstocks, is not clearly deﬁned as a
vigor controlling technique. Webster (1998) reported that budding height has no effect
on the degree of dwarfness achieved, in contrast to apple, where increases in budding
height decrease in scion vigor (Lockard and Schneider, 1981). More recently though,
Santos et al. (2004) identiﬁed budding height as a signiﬁcant contributor to vigor control
in ﬁve different rootstocks grafted with three sweet cherry varieties. Not all rootstock to
scion combinations responded the same to differences in budding height. ‘Summit’ sweet
cherry showed the most signiﬁcant difference in TCSA between 10 and 60 cm budding.
Edabriz, GiselaS and Cabl 1E were the rootstocks with the most signiﬁcant difference in
TCSA between 10 and 60 cm budding. The effect though of the rootstocks to the control
of vigor contributed 80% of the difference, while budding height only 4%. Also, the
reduction in vigor conferred by the increase of budding height was proportional for all

rootstocks, suggesting a physiological rather than genotypic cause for this phenomenon.

11

Webster (1998), based on the inability of cherry interstocks and budding height to confer
dramatic change in scion vigor, proposed that the source of the dwarﬁng signal must
originate in the roots. In contrast, in the apple system the signal originates in the stem.
This hypothesis is supported by the signiﬁcant dwarﬁng ability of interstocks and

budding height (Webster, 1998).

Phenolics in grafted cherry trees

Extended analyses in the phenolics content of various cherry graft combinations
have been performed over many years by W. F eucht and his coworkers in Germany. This
is perhaps the most extended study on the interaction between rootstock and scion in
cherries. His group analyzed many different rootstock and scion varieties for phenolic
and protein content in relation to tree vigor and graft compatibility. Credit should be
given to other researchers as well, who have contributed to the enrichment of our
knowledge on the effect of phenolic compounds in cherry biology.

Catechin is the most abundant phenol in cherry stems and more speciﬁcally in the
phloem and the cambium (Treutter and F eucht, 1991; Usenik and Stampar, 2001).
Results, however, are conﬂicting on the concentration of catechin between graft
combinations of varying vigor. Treutter and F eucht (1991) found a higher concentration
of cathechin in the phloem in comparison to the cambium in two year old stems of
ungrafted ‘Van’ and ‘Werdersche Braume’ grafted on ‘Stockton Morello’. Catechin
concentration was higher in shoot tips of ungrafted and vigorous graft combinations and
furthermore promoted larger callus formation when applied to excised shoots compared

to non-treated explants, suggesting a role in cell division (Feucht and Nachit, 1977;

12

Treutter and F eucht, 1991). In contrast, Usenik and Stampar (2001) showed that catechin
concentration is higher in graft unions of dwarﬁng graft combinations. The effect of
catechin on cell divisions may explain the higher swelling observed in the callus of the
graft union as the dwarﬁng ability of the rootstock increases (Wagner and Gruppe, 1985)

Another phenolic compound that accumulates in the stems of cherry trees is
prunin. In contrast to catechin, prunin accumulates at higher concentrations in the
cambium region rather than the phloem. Like catechin its concentration is higher above
the graft union of dwarﬁng combinations (Treutter and F eucht, 1991; Usenik and
Stampar, 2001). Prunin has been found to inhibit growth of callus cells and prevent
xylogenesis (F eucht et al. 1988). This is in contrast to the effect of catechin that promotes
cell divisions. Prunin, when applied to shoot calluses or shoots, promotes its synthesis
and that of ﬂavan-3-ols, such as catechini(Yuri et al. 1990). Based on the previous
observations a balance between prunin and ﬂavan-3-ols determines callus or shoot
growth rates.

Other phenolic compounds that have been found in cherry tissues are
dihydrowogonin-7-glucoside (DWG) and chlorogenic acids. DWG accumulates at higher
concentrations in the phloem rather than the cambium (Treutter and F eucht, 1991). It is
also found at higher levels above the graft union of dwarﬁng graft combinations in
comparison to vigorous ones (Usenik and Stampar, 2001). Chlorogenic acids accumulate
in the phloem and their concentration is higher in the rootstock directly below the graft
union compared to above it (Feucht and Schmid, 1979). An effect of DWG and

chlorogenic acid in grth has not been studied.

13

As presented before there is no strong correlation between phenolic content and
dwarﬁsm, but rather an indication that phenolics may affect the rate of division of callus
cells at the graft union. Current research relates the accumulation of phenolic compounds
in the graft union as a response to stress produced by the joining of two different
genotypes (Treutter and Feucht, 1991; Usenik and Stampar, 2001). Further research is
necessary to test any direct effect of phenolic compounds on the grth of grafted cherry

trees.

Hormonal control of growth

Differences in growth between genotypes of the same plant species unavoidably
lead to the investigation of hormone levels. As it was mentioned before, auxin has been
implicated in the dwarﬁng phenomenon in apples, thus making it the ﬁrst candidate for
the control of RID in cherries. Leaf indole content has been found to differ signiﬁcantly
between ungrafted cherry rootstocks of varying vigor (Hrotko, 1996). Interestingly,
indole content was inversely proportional to rootstock vigor. Sweet cherry varieties ‘Van’
and ‘Germersdorﬁ drias’ had the highest indole concentration in comparison to the
ungrafted rootstocks. Nevertheless, when these two cherry varieties were grafted on
rootstocks of varying vigor there was no signiﬁcant difference in the indole content of

leaves.

Hydraulic conductance irgrafted trees

Transport rates of water and solutes between rootstock and scion have been

hypothesized to be one of the factors involved in RID. In apple and peach stem hydraulic

l4

conductance has been shown to be lower in the more dwarﬁng rootstocks compared to
the invigorating ones (Atkinson et al. 2003; Basile et al. 2003). Differences in vessel
diameter due to grafting have been implicated in this differentiation in water transport
(Soumelidou et al. 1994b). In newly established cherry grafts, vessel element size was
largely affected by the graft combination (Olmstead et al. 2006). It was more variable in
interspeciﬁc graft combinations where the dwarﬁng rootstocks produced vessels of
smaller diameter (Olmstead et al. 2006a). In 2-year old grafts, vessel diameter was
smaller in the dwarﬁng combination ‘Lapins’/Gi5 compared to the more vigorous
‘Lapins’lColt (Olmstead et al. 2006b). In the same study it was shown that translocation
of the dye Safranin 0 through the graft union was slower in dwarﬁng trees, suggesting a
slower water conductance. Nevertheless, differences in leaf water potential in 7-year-old
grafted cherry trees were attributed to ﬂow resistance within the rootstock-to-shoot
pathway rather than graft union resistances (Schmitt et al. 1989). Similar results have
been obtained for peach trees, where hydraulic resistance ﬂuctuated mainly between the
rootstock and the scion rather than the graft union, in combinations of varying vigor
(Solari et al. 2006). Dwarf trees had smaller hydraulic conductance mainly exerted by the
higher hydraulic resistance of the rootstock (Solari et al. 2006). According to the above
observations lower hydraulic conductance in dwarf trees can be attributed to the smaller
vessel diameter in the rootstock rather than the barrier of the graft union. Such lower rates
of solute transport are expected to have an impact on growth rates of the grafted trees.
According to the above mentioned responses of the cherry trees to grafting on size
controlling rootstocks or interstocks, any transfer of knowledge from the apple system to

cherry should be performed with caution. Physiological changes that occur as a result of

15

the dwarfing phenomenon may be similar between the two systems as secondary

responses, but the leading causes seem to be different.

Long distance transport of macromolecules

The plant body consists of a large number of cells that perform different
functions. Cells of the same or synergistic function are organized into tissues that support
certain physiological processes in the plant. Tissues are combined to form higher order
structures called the organs. As a result plants have developed mechanisms to support this
complex and highly organized structure. Communication between cells and even organs
is necessary for the transmission of information critical to the survival of the organism.
Very early in the history of plant biology it was identiﬁed that plants uptake nutrients
from the soil and distribute them across the plant body to promote its expansion and
reinforcement. The distribution occurs either through the symplast or the apoplast.
Symplastic transport is performed by cell-to-cell transport through plasmodesmata, while
apoplastic transport is performed through the vascular system, which is comprised of
xylem and phloem. The discovery of plant hormones added another factor in the array of
substances that can be transported through the vascular system to exert their action.
Viruses were another group of molecules that were identiﬁed to move through the
vascular system, even spreading to uninfected stock material grafted to infected stocks.
Other signals affecting plant growth and development were identiﬁed early in the history
of plant biology, such as the “ﬂorigen” responsible for the promotion of ﬂowering

(Zeevaart, 2006). Such signals were identiﬁed only through elegant experiments that

16

involved the grafting of tissues that had undergone different treatments (Zeevaart, 2006),
without determining the nature of the signal though. Only recently the detection of such
molecules has been made possible with the advent of new technologies and techniques
(Hoffman-Benning et al. 2002; Huang et al. 2005).

Viral movement through the vascular system occurs via transport of nucleic acids,
either DNA or RNA, and proteins. The transport is made possible by specialized proteins
called Movement Proteins (MPs) that bind to the viral nucleic acids (Leisner and
Turgeon, 1993; Fujiwara et al. 1993; Lee and Lucas, 2001; Itaya et al. 2002). The
nucleoprotein complex that forms can move through the phloem and enter destination
tissues by passing through plasmodesmata. This is accomplished by the increase of the
size exclusion limit (SEL) in plasmodesmata by the MPs. The ability of viral proteins and
nucleic acids to move through the vascular system triggered the initiation of research on
long distance transport of native proteins and nucleic acids. Furthermore, the use of plant
species that yield signiﬁcant amounts of phloem sap has permitted the isolation of many
proteins and RNAs. As a result in the past decade we have seen major advances in the
identiﬁcation and understanding of macromolecular trafﬁcking through the phloem
(Lough and Lucas, 2006).

Several proteins have been identiﬁed to move non-cell autonomously, by cell-to-
cell movement or through the phloem translocation stream (Lough and Lucas, 2006).
Some of the proteins belong to the translocation machinery, facilitating the movement of
other proteins or RNAs, while others function as long distance signals involved in
developmental control. Important components of this transport pathway are the

plasmodesmata, which form pores between neighboring cells of the phloem tissue for

17

intracellular communication (Zambryski and Crawford, 2000). Transport facilitators have
similar properties as the virus movement proteins, allowing them to dilate plasmodesmata
increasing their SEL and allow translocation of large proteins in the phloem translocation
stream. Some of these proteins such as CmPPl6 and CmHSC70-1 or -2 cause an increase
in the SEL of plasmodesmata and facilitate their own transport (Xoconostle-Cazares et al.
1999; Aoki et al. 2002). CmPP16 shares some conserved domains with the movement
protein of the Red Clover Necrotic Mosaic Virus (Xoconostle-Cazares et al. 1999), while
CmHSC70-1 and -2 are heat shock cognate chaperones that use a conserved C-terminal
domain to interact with the plasmodesmata and move through the phloem (Aoki et a1.
2002). But probably the most signiﬁcant effect of long distance transport is exerted by
mRN As involved in plant development. The Mouse ears (Me) leaf phenotype of tomatoes
is caused by a ﬁrsion between a KNOTTEDI-lz'ke gene LeT6 and the PYROPHOSPHA T E-
DEPENDENT PHOSPHO—FR UCTOKINASE (PFP) (Kim et al. 2001). Even though the
fused transcripts can be translocated to wild type scions after grafting, either protein
cannot perform its function, producing a visible leaf phenotype. Another gene found to be
transported through the phloem is CmNA CP, which encodes a NAC domain protein
(Ruiz-Medrano et al. 1999). The mRNA of CmNA CP was translocated through the graft
union ﬁ'om pumpkin stocks to cucumber scions and was localized in the apical meristem.
Other regulatory sequences were also found to be transported through the graft union and
accumulate in the shoot apical meristem of the interspeciﬁc scion (Ruiz-Medrano et al.
1999). One of these genes, CmGAIP is involved in gibberellin signaling and over-
expression or dominant negative mutation causes alteration in the leaf phenotype of wild

type scions in tomato (Haywood et al. 2005). Finally, the mRNA of the FLOWERING

18

LOC US T gene in Arabidopsis has been shown to promote ﬂowering by being expressed
in the companion cells of the phloem and move systemically as a protein from the leaf to
the shoot apex, where it is expressed, to the shoot apex and induce ﬂowering, probably
one of the most signiﬁcant recent discoveries in plant biology (Huang et al. 2005; Jaeger
et al. 2007; Mathieu et al. 2007). The property of these RNAs to be transported at long
distances through interspeciﬁc graft unions shows that the latter do not impede movement
through the vasculature. Nevertheless, the systems studied involve herbaceous plants and
not trees. It is expected though that a similar system would apply to trees as well. Allelic
variation between rootstocks may result in graft union transmissible RNAs that are
perceived differently by the scion meristems or tissues. A notable discovery is that
mRNA also can be transported between host and parasitic plants. Dodder (Cuscuta
pentagona Engelm.) growing on tomato plants contained mRN A from the latter, some of
which have been shown to move systemically in the phloem (Roney et al. 2007). One
such gene was LeGAI, a homolog of CmGAIP that moves systemically into the phloem as
discussed previously. Other unknown tomato sequences also were identiﬁed. Such a
discovery suggests that mRNAs may not be targeted by the silencing machinery that
exists in these plants. A similar response should be expected in the interspeciﬁc grafts as
well, excluding the possibility of incompatibility. Nevertheless, such a hypothesis has to
be tested further.

Long distance signals also move systemically in the form of peptides
(Matsubayashi and Sakagami, 2006). The ﬁrst such peptide was identiﬁed in tomato and
induced systemic acquired resistance to herbivory and was thus named systemin. It

induces resistance by activating the jasmonic acid signaling pathway (Ryan and Pearce,

19

2003). Another peptide signal is CLAVATA3 (CLV3) involved in shoot apical meristem
determination of cell fate (Matsubayashi and Sakagami, 2006). CLV3 is expressed in
layers L1 and L2 of the SAM and moves extracellularly to the central zone to activate the
membrane bound receptors CLV1/2 (Rojo et al. 2002). The role of these peptides
coincides with that of mobile RNA molecules, which act at the destination tissue or cell
type. In contrast, phloem mobile proteins, as discussed before, are acting as helper
molecules in macromolecular transport rather than signaling molecules.

In conclusion, all of these factors can affect communication between heterologous
grafts through reduced interaction afﬁnities. Identiﬁcation of such signals though is a
challenge for species that produce small amounts of phloem sap. Development of new
techniques and use of the already known homologs from model species can advance
signiﬁcantly our understanding on the function of these molecules in diverse plant

species and a putative role in RID.

Hypothesis and objectives

Previous research has shown that a signiﬁcant factor in RID is the genotype of the
rootstock and to some extent of the scion. Current advances in molecular techniques are
allowing the investigation of the genetic differences in non-model plant systems. The
hypothesis to be tested in this study is that “Long distance signaling between the
rootstock and the scion has an effect on gene expression in both genotypes. Eventually
these differences should lead to differential growth between graft combinations of

varying vigor”. Identiﬁcation of genes responsible for the differences in growth between

20

combinations of varying vigor should eventually lead to the signals that cause dwarﬁng.
The objectives of this study were: 1) To identify the most informative measures of
growth capable of showing differences in vigor within a growing season. These measures
will be used as guides for the analysis of gene expression, 2) Screen samples of dwarﬁng
and vigorous combinations to identify differentially expressed genes. Samples will
originate from tissues that have the most important contribution in the control of vigor.
The long term aim of this study is the production of genetic markers that can be
used in future rootstock breeding programs. The long generation time and the difﬁculties
in crossing due to self incompatibility make cherry a difﬁcult ﬁ'uit crop to breed. Any
marker that can easily discriminate the desirable traits is of great need to the cherry
breeder. Nevertheless, the range of rootstock vigor suggests a quantitative trait locus
rather than a single or a few genes, adding to the difﬁculty of breeding the desirable
rootstocks. Furthermore, dwarfness in the currently established rootstocks is not

necessarily linked to the same QTLs.

21

LITERATURE CITED

Aoki, K., Kragler, F ., Xoconostle-Cazares, B. and Lucas, W. (2002) A subclass of heat
shock cognate 70 chaperones carries a motif that facilitates trafﬁcking through

plasmodesmata. Proceedings of the National Academy of Sciences, USA 99: 16342-
16347

Atkinson, C.J. and Else, MA. (2001) Understanding how rootstocks dwarf fruit trees.
The Compact Fruit Tree 34: 46—49

Atkinson, C.J., Else, M.A., Taylor, L. and Dover, C.J. (2003) Root and stem hydraulic
conductivity as determinants of growth potential in grafted trees of apple (Malus
pumila Mill.) Journal of Experimental Botany 54(385): 1221-1229

Basile, B., Marsal, J., Solar, L.I., Tyree, M.T., Bryla, DR. and Dejong, T.M. (2003)
Hydraulic conductance of peach trees grafted on rootstocks with differing size-

controlling potentials. Journal of Horticultural Science and Biotechnology 78(6): 768-
774

Beakbane, AB. (1956) Possible mechanisms of rootstock effect. Annals of Applied
Biology 44: 517-521

DeCandole, A. (1890) Origin of Cultivated Plants, D.Appleton and Company, New York,
NY p.205-211

F eucht, W. and Nachit, M. ( 1977) F lavolans and growth-promoting catechins in young
shoot tips of Prunus species and hybrids. Physiologia Plantarum 40: 230-234

Feucht, W. and Schmid, P.P.S. (1979) Phenolic compounds in the phloem of Prunus
trees, section Eucerasus. Scientia Horticulturae 10: 387-394

Feucht, W., Treutter, D. and Schmid, P. (1988) Inhibition of growth and xylogenesis and
promotion of vacuolation in Prunus callus by the ﬂavanon prunin. Plant Cell Reports
7(3): 189-192

22

Fujiwara, T., Giesman-Cookrneyer, D., Ding, B., Lommel., SA. and Lucas, W.J. (1993)
Cell-to-cell trafﬁcking of macromolecules through plasmodesmata potentiated by the
Red Clover Necrotic Mosaic Virus movement protein. Plant Cell 5: 1783-1794

Haywood, V., Yu, T.S., Huang, NC. and Lucas, W.J. (2005) Phloem long-distance
trafﬁcking of GIBBERELIC ACID-INSENSIT I VE RNA regulates leaf development.
Plant Journal 42: 49-68

Hedrick, UP. (1915) The cherries of New York, J .B.Lyon Company, State Printers,
Albany, NY

Hoffmann-Benning, S., Gage, D.A., McIntosh, L., Kende, H. and Zeevaart, J .A. (2002)
Comparison of peptides in the phloem sap of ﬂowering and non-ﬂowering Perilla and
lupine plants using microbore HPLC followed by matrix-assisted laser
desorption/ionization time-of-ﬂight mass spectrometry. Planta 216(1): 140-7

Hrotké, K. (1996) Leaf indole content and vigor of cherry rootstocks and scion varieties.
Acta Holrticulturae 410: 189-195

Hrotko, K., Magyar, L., Simon, G. and Hanusz, B. (1997) Effect of rootstcks on growth
and yield of sweet cherry trees. Acta Horticulturae 451: 231-236

Hrotko, K., Magyar, L. and Simon, G. (1998) Growth and productivity of sweet cherry
interstem trees. Acta Horticulturae 468: 353-362

Hrotko, K. (2005) Progress in cherry rootstock research, breeding and evaluation.
Personal communication

Huang, T., Bohlenius, H., Eriksson, S., Parcy, F. and Nilsson, O. (2005) The mRNA of
the Arabidopsis gene FT moves from leaf to shoot apex and induces ﬂowering.
Science 309: 1964-1966

Itaya, A., Ma, F ., Qi, Y., Matsuda, Y., Zhu, Y., Liang, G. and Ding, B. (2002)
Plasmodesmata-mediated selective protein trafﬁc between “symplasmically isolated”
cells probed by a viral movement protein. Plant Cell 14: 2071-2083

23

Jaeger, K.E. and Wigge, PA. (2007) FT protein acts as a long-range signal in
Arabidopsis. Current Biology 17: 1050-1054

Jensen, P.J., Rytter, J ., Detwiler, E.A., Travis, J.W. and McNellis, T.W. (2003) Rootstock
effects on gene expression patterns in apple tree scions. Plant Molecular Biology 53:
493-511

Jones, GP. (1986) Endogenous growth regulators and rootstock/scion interactions in
apple and cherry trees. Acta Horticulturae 179: 177-184

Kamboj, J .S., Browning, G., Quinlan, P.S., Blake, RS. and Baker, DA. (1997) Polar
transport of [3H]-IAA in apical shoot segments of different apple rootstocks. Journal
of Horticultural Science 72: 773-780

Kamboj, J.S., Browning, G., Blake, P.S., Quinlan, RS. and Baker, D.A. (1999a) GC-MS-
SIM analysis of abscisic acid and indole-3-acetic acid in shoot bark of apple
rootstocks. Plant Growth Regulation 28: 21-27

Kamboj, J .S., Blake, P.S., Quinlan, PS. and Baker, D.A. (1999b) Identiﬁcation and
quantitation by GC-MS of zeatin and zeatin riboside in xylem sap from rootstock and
scion of grafted apple trees. Plant Growth Regulation 28: 199-205

Kim, M., Canio, W., Kessler, S. and Sinha, N. (2001) Developmental changes due to
long-distance movement of a homeobox fusion transcript in tomato. Science 293:
287-289

Lang, G., Howell, W., Ophardt, D. and Mink, G. (1997) Biotic and abiotic stress
responses of interspeciﬁc hybrid cherry rootstocks. Acta Horticulturae 451: 217-224

Lee, J .Y. and Lucas, W.J. (2001) Phosphorylation of viral movement proteins —
regulation of cell - to — cell trafﬁcking. Trends in Microbiology 9: 5-8

Leisner, SM. and Turgeon, R. (1993) Movement of virus and photoassimilate in the
phloem: A comparative analysis. BioEssays 15: 741-748

Lockard, R.G. and Schneider, G.W. (1981) Stock and scion grth relationships and the
dwarﬁng mechanism in apple. Horticulural Reviews 3: 315-375

24

Lough, T.J. and Lucas, W.J. (2006) Integrative plant biology: Role of phloem long-
distance macromolecular trafﬁcking. Annual Reviews in Plant Biology 57: 203-232

Mathieu, J., Warthmann, N., Kuttner, F. and Schmid, M. (2007) Export of FT protein
from phloem companion cells is sufﬁcient for ﬂoral induction in Arabidopsis.
Current Biology 17: 1055-1060

Matsubayashi, Y., Yang, H. and Sakagami, Y. (2001) Peptide signals and their receptors
in higher plants. Trends in Plant Science 6: 573-577

Olmstead, M.A., Lang, N.S., Ewers, F.W. and Owens, S.A. (2006a) Xylem vessel
anatomy of sweet cherries grafted onto dwarﬁng and nondwarﬁng rootstocks. Journal
of the American Society for Horticultural Science 131: 577-585

Olmstead, M.A., Lang, N.S., Lang, G.A., Ewers, F.W. and Owens, S.A. (2006b)
Examining the vascular pathway of sweet cherries grafted onto dwarﬁng rootstocks.
HortScience 41: 674-679

Perry, R. (1987) Cherry rootstocks, in Rootstocks for fruit crops, Ed. Rom, R. and
Carlson, R. John Wiley and Sons, New York, NY p. 217-264

Robitaille, H. and Carlson, RF. (1971) Response of dwarfed apple trees to stem
injections of gibberellic and abscisic acids. HortScience 6: 539-540

Rogers, W.S. and Beakbane, AB. (1957) Stock and scion relations. Annual Reviews of
Plant Physiology 8: 217-236

Rojo, E., Sharma, V.K., Kovaleva, V., Raikhel, NV. and Fletcher, J .C. (2002) CLV3 is
localized to the extracellular space, where it activates the Arabidopsis CLAVATA
stem cell signaling pathway. Plant Cell 14:969-977

Roney, J .K., Khatibi, RA. and Westwood, J.H. (2007) Cross-species transocation of
mRNA from host plants into the parasitic plant dodder. Plant Physiology 143: 1037-
1043

25

Rozpara, E., Grzyb, 2.8. and Zdyb, H. (1998) Growth and fruiting of two sweet cherry
cultivars with different interstems. Acta Horticulturae 468: 345-352

Ruiz-Medrano, R., Xoconostle-Cazares, B. and Lucas, W.J. (1999) Phloem long-distance
transport of CmNACP mRNA: implications for supracellular regulation in plants.
Development 126:4405-44 19

Rusholme, R.L., Gardiner, S.E., Bassett, H.C.M., Tustin, D.S., Wang, SM. and Didier,
A. (2004) Identifying genetic markers for an apple rootstock dwarﬁng gene. Acta
Horticulturae 663: 405-409

Ryan, CA. and Pearce, G. (2003) Systemins: A functionally deﬁned family of peptide
signals that regulate defensive genes in Solanaceae species. Proceedings of the
National Academy of Sciences, USA 100(Sup.2): 14577-14580

Santos, A., Ribeiro, R. and Crespi, AL. (2004) Sweet cherry (Prunus avium) growth is
mostly affected by rootstock and much less by budding height. New Zealand Journal
of Crop and Horticultural Science 32: 309-318

Schmidt, H and Gruppe, W. (1988) Breeding dwarﬁng rootstocks for sweet cherries.
HortScience 23: 112-114

Schmitt, E.R., Duhme, F. and Schmid, P.P.S. (1989) Water relation in sweet cherries
(Prunus avium L.) on sour cherry rootstocks (Prunus cerasus L.) of different
compatibility. Scientia Horticulturae 39: 189-200

Solari, L.I., Johnson, S. and Dejong, T.M. (2006) Hydraulic conductance characteristics
of peach (Prunus persica) trees on different rootstocks are related to biomass
production and distribution. Tree Physiology 26: 1343-1350

Soumelidou, K., Morris, D.A., Battey, N.H., Barnett, JR. and John, P. (1994a) Auxin
transport capacity in relation to the dwarﬁng effect of apple rootstocks. Journal of
Horticultural Science 69: 719-725

Soumelidou, K., Battey, N.H., John, P. and Barnett, J .R. (1994b) The anatomy of the
developing bud union and its relationship to dwarﬁng in apple. Annals of Botany 74:
605-611

26

Tubbs, F .R. (1967) Tree size control through dwarﬁng rootstocks. Proceedings of the
XVII International Horticultural Congress Volume II: 43-56

Treutter, D. and F eucht, W. (1991) Accumulation of phenolic compounds above the graft
union of cherry trees. Gartenbauwissenschaft 56: 134-137

Usenik, V. and Stampar, F. (2001) Different rootstocks for cherries — inﬂuence on
polyphenol content and graft incompatibility. Acta Horticulturae 557: 175-179

Usenik, V., Stampar, F. and F ajt, N. (2005) Seasonal changes in polyphenols of ‘Lapins’
sweet cherry grafted on different rootstocks. Acta Horticulturae 667: 239-246

Wagner, H. and Gruppe, W. (1985) Swelling of scion trunk above graft unions of cherry
trees. Acta Horticulturae 169: 269-274

Webster, AD. (1995) Rootstock and interstock effects on deciduous fruit tree vigor,
precocity, and yield productivity. New Zealand Journal of Crop and Horticultural
Science 23: 373-382

Webster, AD. and Schimdt, H. (1996) Rootstocks for sweet and sour cherries. In:
CherrieszCrop physiology, production and uses. Editor A.D. Webster and
N.E.Looney, CAB International, Wallingford, Oxon, UK. Pp. 127-163

Webster, AD. (1998) Strategies for controlling the size of sweet cherry trees. Acta
Horticulturae 468: 229-239

27

Xoconostle-Cazares, B., Xiang, Y., Ruiz-Medrano, R., Wang, H.L., Monzer, J ., Yoo,
B.C., McFarland, K.C., Franceschi, V.R. and Lucas, W.J. (1999) Plant paralog to
viral movement protein that potentiates transport of mRN A into the phloem. Science
283: 94-98

Yuri, A., Schmitt, E., Feucht, W. and Treutter, D. (1990) Metabolism of Prunus tissues
affected by Ca2+ -deﬁciency and addition of pruning. Journal of Plant Physiology
135: 692-697

Zambryski, P. and Crawford, K. (2000) Plasmodesmata: Gatekeepers for cell-to-cell
transport of developmental signals in plants. Annual Reviews in Cell Developmental
Biology 16: 393-421

Zeevaart, J. (2006) Florigen coming to an age after 70 years. Plant Cell 18: 1783-1789

28

CHAPTER 1

GROWTH CHARACTERISTICS OF DWARF AND SEMI-VIGOROUS CHERRY

GRAF T COMBINATION S

29

INTRODUCTION

Characterization of a rootstock according to its vigor is not strictly deﬁned. The
most common method of categorizing rootstock vigor makes use of the trunk cross
sectional area (TCSA) or trunk circumference in comparison to a reference vigorous
rootstock. Reference rootstocks are those used traditionally before the breeding of size
reducing rootstocks. In apple the reference rootstock is MM] 1 1, in peaches ‘Nemaguard’
and in cherry P. avium ‘Mazzard’ seedlings. Caution is needed though in this
categorization due to the unparallel expansion of the trunk and elongation of the stems.
Rusholme et al. (2004) in an attempt to map the dwarﬁng locus in apple, have shown that
a population of rootstocks segregating for the dwarﬁng phenotype exhibited varying
vigor, but were ordered differently whether TCSA or tree height was used. Other
measures of grth include total tree height, current year shoot elongation and spread of
canopy. It is not yet established which grth size indicator of a tree should be used to
deﬁne its vigor in relation to other rootstocks.

One of the most important cherry rootstock breeding programs was developed in
Giessen, Germany in the 1960’s through 1980’s under Professor Werner Gruppe (1985a).
The program made use of an extended collection of Prunus species that are compatible
and of smaller size than the standard sweet (P. avium) and sour cherry (P. cerasus) trees.
Several interspeciﬁc crosses were tested which produced an array of rootstocks with
varying vigor (Gruppe, 1985c). The most promising rootstocks were those whose parents
belonged to the P. cerasus, P. avium, P. canescens Bois. and P. fruticosa Pall. species.
The selection process involved testing the bred rootstocks for vigor control, precocity,

yield, suckering, graft-incompatibility and winter hardiness (Gruppe, 1985b; Strauch and

30

Gruppe, 1985). Vigor in the local conditions ranged from 80% to 15% of standard
rootstocks, when measured as shoot length (Seif and Gruppe, 1985). Sizes used for the
determination of vigor were TCSA, shoot length, tree height and spread of canopy
(Gruppe, 1985c; Franken-Bembenek, 1985; Wagner and Gruppe, 1985; Seif and Gruppe,
1985). Bud set was also measured in both grafted and ungrafted rootstocks. Shoot growth
cessation was shown to be the driving force for height control (Seif and Gruppe, 1985).
Dwarﬁng rootstocks tended to cease growing earlier than vigorous ones and as a result
ﬁnal shoot length was smaller. When the sweet cherry cultivar ‘Hedelﬁngen’ was used as
scion, shoot growth cessation occurred signiﬁcantly earlier than the ungrafted rootstocks,
even for the vigorous ones (Table 1.1).

Perhaps one of the most successful rootstocks produced in this project was named
Gisela 5 (Gi5) for Giessen Selection A 5 with an initial code 148/2. Gi5 has been tested
extensively in rootstock trials and has proven to be precocious, dwarﬁng and high
yielding. Several rootstock trials have incorporated Gi5 into their tests, which showed a
30-50% reduction in vigor compared to standard rootstocks (F acteau et al. 1996;
Franken-Bembenek, 1996; Webster and Lucas, 1997; Lichev, 2001; Santos et al. 2006).
The rootstock that most resembles Gi5, but was produced independent of the Giessen
program, is Tabel® Edabriz that belongs to the P. cerasus species. Edabriz shows the
same vigor and precocity as Gi5, but is less productive than Gi5 (Santos et al. 2006). It
would be interesting to see whether this similarity in phenotype is due to the P. cerasus

background of both rootstocks.

31

Table 1.1: Growth measurements of rootstocks produced in the Giessen cherry rootstock
breeding program (adapted from Seif and Gruppe, 1985). Letters in shoot length indicate

signiﬁcant similarities.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. . Cessation of shoot growth

Rootstock origin Clone No. Shoot length(cm) (days from bud break)

Cross Ungrafted Grafted Ungrafted Grafted
P.avium (control) F12/1 142.7 a 37.7 a 102 61
P. cerasus x Pfruticosa 154/4 62.1 g 25.7 de 82 55
154/7 54.1 g 25.9 de 81 55
P. cerasus x P. canescens 148/1(Gi6) 84.5 de 35.8 ab 81 61
148/2(Gi5) 59.8 g 29.4 bd 69 55
148/8 80.0 de 34.1 ab 81 61
148/9 97.7 ed 22.9 de 90 47
Pfruticosa x P.avium 172/3 81.6 de 23.5 de 74 47
172/9 20.0 h 11.9 54 47
Pfruticosa x P.cerasus 173/5 62.2 fg 20.3 cf 79 55
173/9 64.9 fg 22.7 de 82 40
P.canescens x P.avium 196/4 96.1 c 26.6 cde 85 55
196/13 93.3 cd 27.1 cde 95 55
P.canescens x P.cerasus 195/1 110.5 b 24.5 de 89 47
195/2 73.8 ef 29.6 bd 84 47

 

 

 

 

 

 

32

 

Another promising rootstock that was produced in the Giessen program is Gi6
(148/1). It is more vigorous than Gi5, reaching 70-80% of standard rootstocks, but it is
precocious and high yielding (Franken-Bembenek, 1996). Since both Gi5 and Gi6 have
the same genetic background (P. cerasus x P. canescens) it will be interesting to see
which locus is responsible for vigor control.

The objective pursued in this chapter is to identify the most informative measures
of growth capable of showing differences in vigor within a growing season. These
measures will be used as guides for the analysis of gene expression in the following
chapters. The rootstocks Gi5, Gi6 and Tabel® Edabriz were used in this study due to

their genetic and phenotypic similarities.

MATERIALS AND METHODS

Plant material

The trees used in this experiment were purchased from commercial nurseries. The
graft combinations were ‘Bing’/Gi5 and ‘Bing’/Gi6 in which the scion was l-year-old
and the rootstock 2-years-old when purchased. ‘Bing’ is a commercial sweet cherry
cultivar while Gi5 and Gi6 rootstocks are triploid F1 progeny from an interspeciﬁc cross
between the tetraploid sour cherry (P. cerasus L. cv. Schattenmorelle) and the diploid
greyleaf cherry (P. canescens Bois.) (Franken-Bembenek, 1996). The trees for each graft
combination were planted in the spring of 2001 in two rows of 50 trees or in spring 2004
in one row of 50 trees, each with 6 meter row spacing and a 2 meter tree spacing with a

North to South orientation at the MSU Clarksville Horticultural Experiment Station,

33

Clarksville, Michigan. Trees of the graft combination ‘Bing’/Edabriz were planted in
2002 in two rows of 50 trees at the same location and the same planting distances with
the other two combinations. Pruning was performed every spring prior to bud break so I
that only the main trunk was retained above ground. Flowers were removed before

pollination.

Measuring

Ten trees for each graft combination were selected for the measurements. The
trees were selected with the following criteria: healthy trunk before bud break devoid of
canker or freezing damage and intact apical bud. The morphology of the plot consisted of
a wet, fertile north part and a sandy, dry south part. The selected trees were equally
distributed across the plot, to include both plot conditions. For the trunk diameter
measurements, the trunk of the trees was marked initially with a tape and later with paint
10cm above the graft union. Measurements were taken with a digital caliper (VWR)
which was always placed in the marked position with the same direction to avoid
ﬂuctuations due to positioning since the trunk is not a perfect cylinder. Main shoot length
measurements were taken with a tailor’s meter. The zero point was placed in the base of
the shoot and the apical meristem was the end. Arching shoots were straightened during
measuring. Shoots that were damaged by wind or deer were removed from the analyses.
Node number was measured by counting the number of buds present on the growing
shoot. The basal buds were not considered in the measurements. Counting started from
the ﬁrst bud that would form a metamer. The node formed in the shoot apex by the ﬁrst

fully unfolded leaf was considered the last node. In September 2002 after the leaves had

34

dropped all trees were measured for shoot length and node number. Those with damaged

shoots were excluded from further analysis.

Laser scanning confocal microscopy of pith and epidermal cells

Cell number and approximate cell size within a speciﬁc intemode length were
determined from main shoot sections taken from the sixth intemode on the 3lrd of July
2002 or the tenth intemode on the 29th of July. Pith cell measurements were obtained
from radial sections of a 1 cm portion of the intemode that were obtained manually with a
razor blade. Epidermal cell measurements were obtained from a thin layer of epidermal
cells that was obtained in tangential sections from the same portion of the intemode with
a razor blade. The sections were washed in 100% ethanol for 4 hr to reduce browning and
remove the chlorophyll. Sections were then washed twice in distilled water for 30 min to
remove the ethanol. Staining was performed in 0.01% acrydine orange for 12 hr followed
by a wash in distilled water for 5 min to remove excess dye. The Zeiss Axiophot Laser
Scanning Microscope (LSM) was used for ﬂuorescent imaging. A laser line of 488 nm
was induced by an argon laser, passed through a primary dichroic mirror at 488 nm and
the produced ﬂuorescence was ﬁltered through a secondary dichroic mirror at 545 nm.
Emission was viewed with a band pass ﬁlter at 505-530 nm. LSM pictures were analyzed
using the software accompanying the microscope. Vertical lines of known size were
drawn parallel to the shoot axis in the pictures of pith and epidermal cells. The number of
cells falling within the line was counted and the average cell size was determined as
(linear length)/(cell number). Cells forming continuous ﬁles were counted. One or more

measurements were obtained for two or more sections from three independent

35

shoots/trees. Data were analyzed using ANOVA for unequal number of replications with
sub-samples at a=0.05. Epidermal cells on the 29th of July were embedded in a thick

cuticle, preventing imaging of the cells.

Statistical analysis
Statistical analysis was performed using the statistical software SAS v8.0. The

appropriate test statistics were used where applicable.

RESULTS

Monitoring tree growth in three growing seasons

The objective of this project was the identiﬁcation of critical points in growth of
dwarf and vigorous trees within the growing season. Establishment of those critical
points would allow in depth and more focused analysis of the changes in the biology
between trees of different vigor. Measurements were taken for trunk circumference, main
shoot elongation and node number. Two or three year old scions of ‘Bing’ sweet cherry
grafted on GiselaS, Gisela6 or Edabriz were used for the measurements. Vegetative
growth was reduced to a single growing point by removal of the side branches and the
ﬂowers before pollination. This action limited resources to those stored in the main stem
and gave all graft combinations a common initiation point. Any growth potential of the
pruned trees would be supported by a single stem compared to multiple stems in
unpruned trees. The removal of ﬂowers was aimed at eliminating any effect that a strong

sink such as fruit, would have on vegetative growth. ‘Bing’/Gi5 trees exhibit signiﬁcant

36

differences in shoot elongation between fruit thinned, un-thinned and trees without fruit
(Whiting and Lang, 2004). The latter showed the most vigorous and homogenous growth,
while shoots on trees with fruit exhibited reduction in the elongation rates at stage III of
fruit growth that were restored after harvest (Whiting and Lang, 2004).

Plantings occurred in 2001 and 2004 for ‘Bing’/Gi5 and ‘Bing’/Gi6 trees and in
2002 for ‘Bing’/Edabriz. Measurements were taken weekly following bud break in the
years 2002, 2003 and 2005. In the ﬁrst year (2002) measurements were taken twice a
week at the end of the growing season to have a more detailed monitoring of tree growth

cessation.

Expansion of trunk girth does not differ signiﬁcantly between graft combinations
The traditional measure of tree vigor is trunk circumference that can also be
presented as Trunk Cross Sectional Area (TCSA), if assumed that the trunk is circular. It
is measured in trees of 6 to 10 years of age, when the difference in vigor is more distinct
between graft combinations, since trunk circumference expansion is additive through the
years. In this study we further measured the progress of trunk expansion within the
growing season to detect any discrepancies between growth of dwarﬁng and vigorous
rootstocks. The trees were very young with a thin trunk, thus measuring the diameter was
more accurate than measuring the trunk circumference. Values of trunk diameter were
taken 10cm above the graft union to avoid the effect of the graft union swelling. The
trunk diameter of ‘Bing’/Gi6 trees was consistently larger than that of ‘Bing’/Gi5 and
‘Bing’/Edabriz trees, but the difference depended on the age of the trees (Figure 1.1,

Table 1.2). Two year old trees (2002, 2005) did not show signiﬁcant difference in

37

diameter, but in three year old trees (2003) ‘Bing’/Gi6 trees had larger diameter than
‘Bing’/Gi5. Expansion of the trunk for all the grafts was parallel throughout the growing
season, which is illustrated in detail by the expansion rate of the trunk (Figure 1.2). In all
years examined the expansion rate was higher in the end of July, after the cessation of
shoot growth as it is shown later. In 2003 the trunk diameter of ungrafted Gi5 and Gi6
trees was measured. Expansion in these trees is more stable throughout the growing
season (Figure 1.3) and follows a similar expansion rate as the two year old ‘Bing’ scions
in 2002 and 2005 (Figure 1.2A, C). The expansion rate of the trees in all growing seasons
is signiﬁcantly affected by the environmental conditions rather than the age of the trees.
This conclusion is supported by the highly synchronized growth rate within each year
between graft combinations but also the changes in growth rate within each year due to

changes in the environment (Figure 1.2).

38

Figure 1.1: ‘Bing’ cherry tree trunk diameter growth across three growing seasons and
across three graft combinations. A. Growing season of 2002, B. Growing season of 2003,
C. Growing season of 2005. In 2002 and 2005 measurements were taken from 2-year old
trees, while in 2003 measurements were taken from 3-year old trees. Error bars indicate
standard error.

39

—-x— 'Bing'/Gi5
+ 'Bing'/Gi6

   

_
.x NOBNE

 

_.. News
a «28%
M. 858
L4 moans
T NONE
,

ﬁ Same
«93$
«2me

.v 853

L

_ moat».

NQQm

15

+ 'Bing'/Gi5
+ 'Bing'/Gi6
—o— 'Bing'/Edabriz

4 lliJ JI I I4! a.

as... 83.55

A 85%
I. Baa

_
_
__- 883

A Sea
A

r 853
f 8&5
A 88%
.. 833
7 8:3
85?.

ﬂ momma

l+ moE Ev
5

1

'lGi6

—x— 'Bing'/Gi5
+ 'Bing

. i ii ll..l.||Jri

:55 causes

.. 885g
883$
888$
M. 8853
. 8833
_ 885:
8833
8835
885a
- 885a

. mOONBB

.. moon 2v
"m

40

Figure 1.2: Trunk diameter growth rate in ‘Bing’/Gi5 and ‘Bing’/Gi6 cheery trees in 2002
(A), 2003 (B) and 2005(C). In 2003 trunk expansion rate was calculated also for
‘Bing’/Edabriz trees. Expansion rate was calculated as the ratio of weekly trunk diameter
increase divided by the number of days between measurements. In 2002 and 2005
measurements were taken from 2-year old trees, while in 2003 measurements were taken
from 3-year old trees.

41

             

 

 

 

 

42

  

m
5 6 5 mm
mm mam mm
.m. .m. .m m .m .m. .m
m m H H “H M H
_ _.
. 83:8 . 88:8
% 88:8 L
m «283 _ .-
L. _. 85$ 85%
L .-
U «293 . 8B5 88%
. L ,
.1855 835 . 85$
.- 83: _. 83: .. 83:
a 855 - 856 N. 82%
._ 8E8 . 85¢ - 8E8
\ 83mm 8an . 83$
_
1.. 1-1.1111111 85% . . . 1,1- .1.-1.1 .-11-+ 852m 1. - .1.-1.11 . 1 85%
8642186420 36.420.36.420 354.218.54.20
1. 1. 1. 1. 0. 0. 0. 0. 1 1 1 1 1 0 0 0 0 0 1 1 1 1 0 0 0 0
3825.5 9.22598 .25.... AxooEEEv c2235 35:... 30225:... 5.2596 :55.

 

30.0 ..
25.0 -‘
E 20.0 i
E
v 15 0 ! +Gi5 ungrafted
g i +Gi6 ungrafted
E 10.0
5.0
1
0.0+ # A # f f.-.f-__a_w, ,#.Lg
(‘0 m (‘0 CO (‘0 0') (O (‘0 ('0 (O
O O O O O O O O O O
3 o‘o >- 1?) a. 8 B B B i:
" Q E Q rt Q B Q o‘: E
B in to co r~ co or
B
E
E,
g +Gi5 ungrafted
E +Gi6 ungrafted
GB
&
o
.8
r:
E
I-
0.0 -—— _, _

 

 

 

7/9/03 1

3120/03
9/3/03

9/17/03 i

1 ’ r
co co co co
9 o 9 9
v a ‘— l!)
C N S Q‘
L0 is (D (D

7/23/03
816/03

Figure 1.3: (A) Trunk diameter of ungrafted Gi5 and Gi6 rootstocks measured in 2003,
(B) Trunk expansion rate of the same trees expressed as the ratio of weekly trunk
expansion divided by the number of days between measurements. Error bars indicate
standard error.

43

 

2.5m 90.34 mm. _ m sad. 0&2 mm.: pom n3. ohm 9a.? mmﬁm have SUE—am

 

 

mwdm ﬂdm so. ﬂ m «M: 3.: «NM: 0mm mom «mm 3.9 2.3. 3.3 magma—E
moom moom Noom moo-o. meow. meow meow meow Noom meow meow NOON
955 080806 x55... NEEV 5w=0_ 0:502 09:5: 000: 35m A83 alvwc2 Hoonm 35m

 

8 _ M: 00055 080806 0:5:- com E080 ans .3.on :0 Ha 00:0h0b6 “:85ch
08:30 80:04 800.5 20 80%;“ 80¢ :83 803 8:080:58:— moom E 223 .0000“ 20 Sofa Bob :83 203 3580:6008 moon
was moom 5.3808 wEBEw 00:: 8800 0.0—03308 90 can 30 co “00¢me 0:28 .wEm. E050 mo 8558820 538$ ”N; 033-

 

44

Main-shoot elongation in three growing seasons and different graft combinations

As a measure of vegetative growth and a major contributor to tree height, we
monitored the elongation of the main shoot in the same trees as those used for measuring
trunk diameter. All the trees, irrespective of the rootstock, broke bud simultaneously in
mid April consistently for all three years. Shoot elongation followed a sigmoidal curve
for the ‘Bing’ scions and the ungrafted Gi6 rootstocks, but followed a more linear curve
for the ungrafted Gi5 rootstocks (Figure 1.4, 1.6A). Elongation was slow in the ﬁrst 6
weeks, but remained equal for all grafted trees (Figure 1.4). During the 7th or 8th week,
elongation differentiated between the semi-vigorous ‘Bing’/Gi6 trees and the dwarﬁng
‘Bing’/Gi5 or ‘Bing’/Edabriz, which is shown in Figures 4 and 5, by the separation of the
curves. In 2003 and 2005 this differentiation coincided with the highest elongation rate
(Figure 1.5B,C), but in 2002 it occurred at a stage when elongation rate was not
maximum (Figure 1.5A). Following shoot growth cessation, bud set occurred 1 to 3
weeks earlier for ‘Bing’/Gi5 and ‘Bing’/Edabriz trees compared to ‘Bing’/Gi6 trees
(Figure 1.7), with the exception of 2005 when bud set was completed the same week.
‘Bing’/G15 trees showed consistently earlier bud set compared to ‘Bing’/Gi6 in all three
years. At the end of the growing season ‘Bing’/Gi5 shoots grew at 73%, 79% and 73% of
the ‘Bing’/Gi6 shoots in 2002, 2003 and 2005 respectively (Table 1.2). It has to be noted
that ‘Bing’/Gi5 trees sustained frost damage in the winter of 2004-2005, which led to the
loss of many trees. Many of the trees that survived showed bleeding from the trunk.
‘Bing’/G16 showed much lower damage.

In contrast, ungrafted trees had a slower shoot elongation, which though lasted for

a longer duration, until the middle of September (Figure 1.6). This is 7 to 8 weeks later

45

than the grafted trees, but is largely due to 10-20% of the trees that did not set bud
(Figure 1.78). Elongation rate for Gi5 was stable across the growing season, in contrast
the elongation rate of Gi6 which reached a maximum in the beginning of July, 8 weeks
after bud break (Figure 1.6B). Nevertheless, 100% bud set occurred simultaneously for
the two rootstocks. Final shoot length was less for Gi5 compared to Gi6, which is
consistent with the results in the grafted trees.

As it was observed for trunk diameter, the weather seemed to affect the temporal
changes in elongation rate, depicted by the parallel changes in the elongation rate shown

in Figure 1.5.

46

Figure 1.4: Main shoot length of ‘Bing’/Gi5, ‘Bing’/Gi6 cherry trees taken in 2002 (A),
2003 (B) and 2005 (3). Measurements were taken from bud break to bud set.
Measurements were also taken for ‘Bing’/Edabriz trees in 2003. In 2002 and 2005
measurements were taken from 2-year old trees, while in 2003 measurements were taken
from 3-year old trees. Error bars indicate standard error.

47

Shoot elongation (cm)

Shoot length (cm)

Shoot length (cm)

701

—x— Bind/615
+ 'Bing'/616

    

 

_.,___

Aﬂ__._ﬁ __T_.__

4119/02 5/3/02 5/17/02 5131/02 6114/02 6/28/02 7/12/02 7/26/02 819/02

 

 

 

 

 

8
i I

 

 

-x— 'Bing'/Gi5
+ 'Bing'/616
+ 'Bing'/Edabriz

o 8 8 ‘8 8
__A_——_i__—-—-—_
|

I

 

I a ﬁ I 1' T_—"

4/19/03 513/03 5117/03 5131/03 6114/03 6128/03 7/12/03 7/26/03 819/03

—x— 'Bing'/615
+‘Bing'lGiG

 

 

101

4/19/05 513/05 5117/05 5131/05 6/14/05 6128/05 7112/05 7/26/05 8/9/05

48

Figure 1.5: Main shoot elongation rate in ‘Bing’/Gi5 and ‘Bing’/Gi6 cherry trees in 2002
(A), 2003 (B) and 2005(C). In 2003 shoot elongation rate was calculated also for
‘Bing’/Edabriz trees. Expansion rate was calculated as the ratio of weekly shoot
elongation divided by the number of days between measurements. In 2002 and 2005
measurements were taken from 2-year old trees, while in 2003 measurements were taken
from 3-year old trees. Error bars indicate standard error.

49

20.00 -~
18.00 i
16.0 1
14.00 1

Elongation rate (mm/day)

12.00
mo 1
0.00
6.00
4.00
2.00
0.00 1

1
+ 'Bing'/615

+ 'Bing'/Gi6

1

 

 

5/5/02 5/19/02 6/2/02 6/16/02 6130102 7/14/02 7128/02 8111/02

20.0 7
18.0 ~

1
1

Elongation rate (mm/day)

Elongation rate (mm/day)

16.0 7-
14.0 1

2.0 7
0.0 1
8.0 i
6.0 l
4.0 1
2.0 1
0.0 +—--- g-.. - 7..---__ - - - ~
5/5/03 5119/03 6/2/03 6116/03 6130/03 7/14/03 7128/03 8111/03

+ 'Bing'/G15
+ 'Bing'/616
+ 'Bing'/Edabriz

 

 

d —l d _L N
O
L—+2_L—«4.$—~___L_— 1

—I
ONAO’QON‘OQ

+ 'Bing'/615
+Bing°1616

l___84

 

M-..

 

 

r 7 T——— ' ——-—-———._~

7 g'—__' T 1’ if 7 T’

516/05 5/20/05 613105 6117/05 711/05 7115/05 7129/05 8112/05

50

66‘”
£401 ‘ +Gi$ungraﬂed
5 - .. +Gi6ungraﬂed
§3° ’
520 ’
’./

101 ’

0 ﬁ— T—‘Tﬁ T_—"' __ mm —

8 8 8 8 8 8 8 8 8 8 8

2 g E :7; s a a a as a is.

3 n a a a 1: 1: a 2: °’

B

16.0».
A1601
)5 1
314.0%
E
$120:
$10.01: -x—Gi§ungraﬂed
8 8.0 +GiSungra1ted
g 6.0
2’
2 4.0
I”

2.0
0.0 .

     

 

 

5114/03
5128/03 1
6/11/03
6125/03 “
7/9/03 1
7/23/03
816/03 ‘
8/20/03
9/3/03

Figure 1.6: Shoot growth characteristics of ungrafted rootstocks Gi5 and Gi6 in the
growing season of 2003. (A) Shoot length from bud break to bud set and (B) shoot
elongation rate expressed as the ratio of weekly shoot growth divided by the number of
days between measurements. Error bars indicate standard error.

51

Figure 1.7: Cumulative bud set in ‘Bing’/Gi5 and ‘Bing’/Gi6 trees for the growing
seasons of 2002 (A), 2003 (B) and 2005 (C). Bud set was recorded for ‘Bing’/Edabriz,
Gi5 and Gi6 ungrafted rootstocks in the growing season of 2003 (B). Cumulative bud set
is the percentage of trees for which the main shoot has set bud. The week when shoot
growth cessation occurred was considered as the time of bud set (n=9-1 1). In 2002 and
2005 measurements were taken from 2-year old trees, while in 2003 measurements were
taken from 3-year old trees

52

'1 615

+ 'Bing'/Gi6

9
.m
M

v

f «2%

f 833
C

W. N928

     

m
V 85.:
m
A

 

 

83S
83%
835
1 11- 11.- . 1- , 18:8
m m w m m 0
3c .3 .23 22....an

'IG16

—o— 'Bing'/Edabriz
—o— 615 ungrafted
+ 616 ungrafted

mm
m
.9
.m
a“

+ 'Bing

 

_.
w
.7 meta

 

1 MQVNB

. MOB—.3

, motht.

 

88 Ex.

”GENE

832m

-11 11 1.1 11-1 8:6

mmmwmo

as .8 25 2,3:an

 

m
35 .8 25 2:2:an

_)(._

'Bing'/G15
+ 'Bing'/616

.. mgtm

 

_. 833
W 85%
a 85:
. 885
. 85%

m mom—.6

 

.. . - . .35
w m m o

53

Measurements of intemode number and metamer length

Shoot elongation measurements showed that tree height is controlled by shoot
elongation and more speciﬁcally shoot grth cessation. The difference in shoot length
was constant through the years, but it wasn’t clear if grth cessation was the only
determinant of this difference. The various plant hormones have been shown to affect
shoot elongation, most notably gibberellin, which affects intemode length and cell size
(Fleet and Sun, 2005). Node number was measured in the same trees as for shoot
elongation and trunk diameter, to detect any discrepancies in the number of nodes and
length of intemodes. In 2002 nodes were measured only after shoots ceased growing, but
in 2003 and 2005 counting was performed in parallel to shoot elongation. In 2002, 50
trees were measured for each graﬁ combination and gave an average node number of 30
in ‘Bing’/G15 and 39 in ‘Bing’/616. In 2003 these numbers were 25 and 36 respectively,
while for 2005 they were 23 and 29. The curve for node number increase was similar to
the shoot elongation curves. When shoots were growing with the same rate, nodes were
also added equally between dwarf and semi-vigorous trees (Figure 1.8). When shoot
growth diverged between dwarf and semi-vigorous trees, node number followed this
change (Figure 1.8). Metamer length at each time of the growing season was calculated

as follows:

Mean Metamer length = shoot length/node number

Mean metamer length was not constant throughout the growing season (Figure 1.9). In

the ﬁrst 3 weeks of shoot growth metamers were short, since they are few in number and

54

are all elongating. During active shoot elongation, mean metamer length was maximal
and it stabilized to a smaller length during shoot growth cessation (Figure 1.9). This
decrease in metamer length at the end of the season does not indicate shrinking of the
metamers, but rather inability of the last metamers to elongate fully, reducing thus the
mean value. The length in the three growing seasons ranged between 1.7-2.0cm, but
except from 2005 there was no signiﬁcant difference in intemode length between
‘Bing’/Gi5 and ‘Bing’/Gi6 trees (Table 1.2). In 2005 metamer length contributed to 29%
of the difference in shoot length between ‘Bing’/Gi5 and ‘Bing’/Gi6 trees, with the rest

of the difference (71%) being due to timing of shoot growth cessation.

55

 

 

% —x-— 'Bing'/615
g + 'Bing'/Gi6
g —o—'Blng'/Edabrlz
O 8 ﬂ .
8 8 8 8 8 8 8 8
<2 8 v- B 5 B B Q
"' N " N R 9‘. B
is P5 B B Is B
B
40 1
1
—x— 'Bing'/615
+ ’Bing'/Gi6

 

 

5114/05
5128/05
6/11/05
6125105 .
719105
7123105
816105
8120/05 ‘

Figure 1.8: The number of nodes was measured in the main shoot of ‘Bing’/Gi5 and
‘Bing’/Gi6 in the growing season of 2003 (A) and 2005 (B). In 2005 measurements were
taken from 2-year old trees, while in 2003 measurements were taken from 3-year old
trees. Node number was also measured in 2-year old ‘Bing’/Edabriz trees in the growing
season of 2003 (A). Error bars indicate standard error.

56

 

2.5 ,
A 2.0 1
E
,9, 1
g 1.5 1 —x—'Bing'/GIS
§ +‘Bing'lGi6
g 1.0 —o—'Bing'/Edabn'z
S 0.5 71
0.01————~. ____._.--_ ___ _
{'0 C") (‘0 0‘) C") C’)
O O O O O O
:1 as : B a a
" N 2 Q r: q
E E5 co co rs
B
2.5
‘g 2.0
3
E 1.5 +‘Bingvcis
3 1.0 ‘ +‘Bing'lGiG
E
8 .
g 0.5 .
1

.0
o

4—
10
Q
In
'—
\
ID

5129/05 1
6/12/05 1
6/26/05
7/10/05 “
7124/05

Figure 1.9: Metamer length of ‘Bing’/Gi5 and ‘Bing’/Gi6 trees in the growing season of
2003 (A) and 2005 (B). In 2005 measurements were taken from 2-year old trees, while in
2003 measurements were taken from 3-year old trees. Metamer length was also
calculated for 2-year old ‘Bing’/Edabriz trees in the growing season of 2003 (A).
Metamer length was derived by the division of shoot length to node number for each
measurement point. Higher metamer lengths were observed in June due to higher
elongation rates. Reduction of metamer length reﬂects shorter intemodes as shoot growth
rates dropped near the time of shoot growth cessation.

57

Counting of cell size and number in metamers

The average metamer length was not signiﬁcantly different between dwarﬁng
(‘Bing’/Gi5) and non—dwarﬁng trees (‘Bing’/Gi6), with an average length of 1.7-2.0 cm
in two of the three growing seasons as shown before. Even though the genetic
background of the scions was the same, it was necessary to test whether there was a
difference in growth at the cellular level within an intemode. The size and number of
cells that constitute the pith and the epidermis of the sixth intemode were measured in
shoots collected on 3 and 29 July 2002. On the 3rd of July, the sixth intemode
corresponded to approximately the ﬁrst intemode produced during the initiation of
differential growth between dwarﬁng and non-dwarﬁng trees (Figure 1.4). Also, at this
stage cells had stopped dividing and reached their ﬁnal size. No signiﬁcant differences
were identiﬁed for cell number or size for the pith and epidermis from both graft
combinations (Figure 1.10). This indicates that between the two rootstocks, main shoot
intemodes have equivalent cell numbers and cell sizes; however, in vigorous trees, the

initial rate of shoot growth is maintained for a longer duration.

58

(a) Pith cell 8120 (C) Epldermal cell size

.5
o
o

   

 

 

 

 

 

 

 

 

 

 

 

8 5° 8

a 40 a 80

8 8

g 30 2 60

8 E

To 20 7a 40

E 5

E10 3 20

2' 2'

3 0 3 o

(b) Pith cell number (d) Epldermal cell number

20

O

$15 1 T

3: z:

B ‘01

= = 5

8 8
0

07:03:03 07129103 0 Bing/615 07:03:03
Sample collection I Bing/GS Sample collection

Figure 1.10: Cell size and number in the pith and epidermis of the 6th intemode collected
on the 3rd of July and the 10th intemode collected on the 29th of July of ‘Bing’ sweet
cherry trees on Gisela 5 (Gi5) and Gisela 6 (Gi6) rootstocks. (A) Pith cells were counted
in radial sections of the intemode. (B) Cell ﬁles in radial sections of the intemode were
used to measure the cell number per millimeter of pith. (C) Epidermal cell ﬁles were
counted in tangential sections of the bark. (D) Cell number per millimeter of epidermis
was counted in cell ﬁles of the 6th intemode. Measurements were taken on 2-year old
trees. Error bars indicate standard error.

59

DISCUSSION

Understanding the morphological changes that lead to dwarﬁsm in cherry is
crucial for the identiﬁcation of the genetic factors that contribute to this phenomenon.
Previous studies or rootstock trials have placed priority on the size of the trees at a
reproductive age, which for most rootstocks is achieved after the 4th year of growth.
Vigor is expressed through the trunk cross sectional area (TCSA) or the total tree height.
But these are static measures that do not provide any insight in the progression of tree
size. In the extended rootstock breeding program that produced the Gisela rootstocks,
data on size were collected throughout the growing season to demonstrate the changes in
size between the new rootstocks. In that trial, shoot elongation measurements of
ungrafted and grafted rootstocks indicated a differentiation in the time of shoot growth
cessation, as dwarf rootstocks stopped growing earlier (Seif and Gruppe, 1985; Franken-
Bembenek, 1996). The same was observed in this study in graﬁed rootstocks.

The beginning of the growing season found all graft combinations to break bud at
the same time. Shoot elongation was parallel for all the graft combinations. When shoot
elongation reached the highest rate, there was difference between dwarf and semi-
vigorous rootstocks. Dwarf rootstocks elongated more slowly and ceased growing earlier
than the semi-vigorous rootstocks leading to shorter shoots and tree height. The initial
parallel growth of all trees indicates the potential for growth exerted by the physiological
status and genetic background of the scion. This also indicates that reserves and nutrient
supply are not limiting at this stage of shoot growth. The onset of differential growth rate
signiﬁes a change in the biology of the various graft combinations. The signal for such a

differentiation is not known, neither is its source. The driving force of this change should

6O

relate to the environment, either directly or indirectly. Direct signals from the
environment can be a change in photoperiod, a change in day to night temperature or
water supply to the roots. Any effect of temperature or soil moisture can be excluded,
since the response of the trees is consistent each year. In 2002 trees were not irrigated,
while in 2003 they were, but the pattern of shoot growth did not change. Furthermore, the
summer of 2002 was warmer than 2003 in mid-June, but shoot growth differentiated with
a few days difference between the two years. When shoots of ‘Bing’/Gi5 and
‘Bing’/Edabriz trees started growing slower than ‘Bing’/Gi6 shoots, photoperiod was
reaching the summer solstice (June 22-23) at the Clarksville Horticulture Experiment
Station. When ‘Bing’/Gi6 shoots started growing slower the summer solstice was already
past, thus downgrading the importance of photoperiod in the control of this phenomenon.
In an extensive cataloguing of tree species according to their photoperiod response,
Nitsch (1957) had placed cherry in the non-photoperiod responsive species, which is in
agreement with the data presented in this chapter. Previous research on the effects of
grafting has revealed a differential hyrdraulic resistance exerted by the difference in
vessel diameter at the scion (Olmstead et al. 2006). Lower hydraulic conductance could
have an impact on the support of growth especially at times of maximum growth rates.
As shown in Figure 1.5B and C, shoot elongation rates differentiated between dwarﬁng
and semi-vigorous trees when they exceeded 10mm/day. This was not evident in 2002
though, when shoot elongation rate differentiated as the rate of grth was dropping.
Even though an effect of hydraulic conductance cannot be excluded, the consistency on

the time of shoot growth cessation suggests a larger impact of another factor.

61

In contrast to the grafted trees, the ungrafted rootstocks continued growing until
September, independent of their dwarﬁng ability. This observation increases the
importance of the scion in the control of the dwarﬁng phenomenon and furthermore its
interaction with the rootstocks. The same difference in shoot growth cessation between
grafted and ungrafted rootstocks was identiﬁed by Seif and Gruppe (1985). Rootstocks
grafted with ‘Hedelﬁngen’ had signiﬁcantly earlier cessation of shoot growth compared
to the respective ungrafted rootstocks that varied between 7 to 41 days. The genetic
background of the rootstock is the indisputable factor that controls tree vigor, but based
on the previous observation, the signal originates in the scion. It is then perceived
differently by the various rootstocks and the process of shoot growth cessation is
initiated. As indicated by the shoot elongation and node number curves, differential
perception of the signal may be triggered when a threshold is achieved that is different
between rootstocks. The nature of the response by the rootstock is not known. The
absence of signiﬁcant difference in the length of the intemodes, the number and size of
epidermal and pith cells, indicates that the difference in shoot length is due to control of
the shoot apical meristem rather than cell division and expansion outside the meristem.
Cells at the shoot apical meristem are produced in slower rates in the dwarf trees rather
than the semi-vigorous. Cell division frequency thus should be lower in the ﬁrst rather
than the latter. As a result fewer nodes are produced by the end of the growing season in
the dwarf trees (Figure 1.8).

In contrast to cherry, apples and peaches show a difference in shoot elongation
between graft combinations of varying vigor, right from the beginning of the growing

season (Weibel and DeJong, 2003; Webster, 1995). In the same systems, shoot growth

62

differences are a result of intemode length and to a lesser extent of node number. This
suggests an alteration in hormone perception or metabolism. Mutations in genes involved
in gibberellin signaling lead to alterations in intemode length and cell size (Peng et al.
1999; Boss and Thomas, 2002; Fleet and Sun, 2005). Loss of function mutants show
reduced intemode lengths, while gain of function mutations or application of gibberellins
increases intemode length (Fleet and Sun, 2005). Thus, in apples and peaches
gibberellins may play a signiﬁcant role in shoot elongation by affecting intemode length.
In cherries gibberellins do not seem to affect shoot growth, since intemode length does
not differ between grafts of different vigor.

Trunk diameter did not prove a good indicator of tree vigor within a growing
season. Different graft combinations had parallel growth rates and any difference in vigor
was not observed. Nevertheless, the diameter change is additive through the years and as
the trees grow older the difference in vigor is made obvious. It should be noted though
that trunk expansion rate was higher after shoot growth cessation, which is probably
attributed to the change of sinks. Shoots are not allocating any more carbohydrates for
their growth, but rather transfer it to the trunk and roots for storage. Thus, the trunk
expands and stores more carbohydrates in the wood tissues. At the time of shoot grth
cessation, new ﬂower meristems start forming. Thus, dwarf trees that cease growing
earlier than vigorous rootstocks can allocate more resources towards ﬂoral bud formation.
This hypothesis can explain the increased productivity of dwarf rootstocks in contrast to
vigorous ones, which allocate more photoassimilates to support shoot growth.

Similar growth patterns in cherry trees have been observed after the application of

paclobutrazol (PBZ) through the soil. This growth regulator has been proposed as an

63

alternative to dwarﬁng rootstocks for cherry varieties planted on their own roots
(Quinlan, 1985). Signiﬁcant shoot length reduction has been reported in cherry trees after
soil or foliar treatment with PBZ (Facteau and Chestnut, 1991; Snir, 1988; Asamoah and
Atkinson, 1985; Looney and McKellar, 1987; Walser and Davis, 1989). Soil application
of PBZ is comparable to the effect of the rootstock. The message for shoot grth
cessation in dwarﬁng rootstocks originates in the root, while PBZ acts through the root.
PBZ blocks the oxidation of ent—kaurene into ent-kaurenoic acid in the early steps in the
biosynthesis of gibberellins (Rademacher, 2000). As a result of this blockage shoot
intemodes expand less compared to untreated shoots and the ﬁnal shoot length is reduced
(Webster, 1998).

Shoot elongation proved the most informative measure to depict the discrepancies
in growth between dwarf and semi-vigorous graft combinations. Although the rootstock
is controlling the ﬁnal shoot length and as a result tree height, it is the scion that activates
the initiation of shoot growth cessation. The nature of the signal originating in the scion is
not yet determined. Study of the changes that occur between rootstocks at the time of
initiation of shoot grth cessation can yield signiﬁcant information on the nature of the

agents and genes involved in this process.

64

LITERATURE CITED

Asamoah, T.E.O. and Atkinson, D. (1985) The effects of (2RS, 3RS)-1-(4-chlorophenyl)-
4, 4-dimethyl-2-(1H-l,2,4 triazol-l-y1)pentan-3-ol (Paclobutrazol:PP333) and root
pruning on the growth, water use and response to drought of Colt cherry rootstocks.
Plant Growth Regulation 3: 37-45

Boss, PK. and Thomas, MR. (2002) Association of dwarﬁsm and ﬂoral induction with a
grape 'green revolution' mutation. Nature 416: 847-50

F acteau, T.J. and Chestnut, NE. (1991) Growth, ﬁuiting, ﬂowering and fruit quality of
sweet cherries treated with paclobutrazol. HortScience 26: 276-278

F acteau, T.J., Chestnut, NE. and Rowe, K.E. (1996) Tree, fruit size and yield of ‘Bing’
sweet cherry as inﬂuenced by rootstock, replant area, and training system. Scientia
Horticulturae 67: 13-26

Fleet, CM. and Sun, TR (2005) A DELLAcate balance: the role of gibberellin in plant
morphogenesis. Current Opinion in Plant Biology 8: 77-85

Franken-Bembenek, S. and Gruppe, W. (1985) Variability in vegetative growthof
different cherry hybrids (Prunus x spp). Acta Horticulturae 169: 257-262

Franken-Bembenek, S. (1996) The Giessen cherry rootstocks. Compact Fruit Tree 29: 19-
56

Gruppe, W. (1985a) An overview of the cherry rootstock breeding program at Giessen
1965-1984. Acta Horticulturae 169: 189-198

Gruppe, W. (1985b) Evaluating orchard behavior of cherry rootstocks. Acta Horticulturae
169: 199-208

Gruppe, W. (1985c) Size control in sweet cherry cultivars (Prunus avium) induced by
rootstocks from interspeciﬁc crosses and open pollinated Prunus species. Acta
Horticulturae 169: 209-218

65

Lichev, V. (2001) First results from testing cherry clonal rootstocks Gisela and Weiroot
in Bulgaria. Proceedings of the 9th International Conference of Horticulture 1: 111-
115

Looney, NE. and McKellar, J .E. (1987) Effect of foliar- and soil surface-applied
Paclobutrazol on vegetative growth and fruit quality of sweet cherries. Journal of the
American Society for Horticultural Science 112: 71-76

Nitch, J .P. (195 7) Photoperiodism in woody plants. Proceedings of the American Society
for Horticultural Science 70: 526-544

Olmstead, M.A., Lang, N.S., Lang, G.A., Ewers, F.W. and Owens, SA. (2006)
Examining the vascular pathway of sweet cherries grafted onto dwarﬁng rootstocks.
HortScience 41: 674-679

Peng, J ., Richards, D.E., Hartley, N.M., Murphy, G.P., Devos, K.M., F lintham, J .E.,
Beales, J ., Fish, L.J., Worland, A.J., Pelica, F., Sudhakar, D., Christou, P., Snape,
J .W., Gale, MD. and Harberd, NP. (1999) 'Green revolution' genes encode mutant
gibberellin response modulators. Nature 400: 256-261

Quinlan, J .D. (1985) Chemical regulation of fruit tree growth I the development of new
production systems. In: Growth regulators in Horticulture (Menhennet, R. and
Jackson, M.B.,Editors) British Plant Growth Regulator Group Monograph 13, Long
Ashton, 63-70.

Rademacher, W. (2000) GROWTH RETARDANTS: Effects on gibberellin biosynthesis
and other metabolic pathways. Annual Review of Plant Physiology and Plant
Molecular Biology 51: 501-531

Rusholme, R.L., Gardiner, S.E., Bassett, H.C.M., Tustin, D.S., Wang, SM. and Didier,
A. (2004) Identifying genetic markers for an apple rootstock dwarﬁng gene. Acta
Horticulturae 663: 405-409

Santos, A., Santos-Ribeiro, R., Cavalheiro, J ., Cordeiro, V. and Lousada, J .L. (2006)
Initial grth and fruiting of ‘Summit’ sweet cherry (Prunus avium) on ﬁve
rootstocks. New Zealand Journal of Crop and Horticultural Science 34: 269-277

66

Seif, S. and Gruppe, W. (1985) Shoot growth in sweet cherry (P.avium), cherry hybrid
rootstocks, and grafted trees. Acta Horticulturae 169: 251-256

Snir, I. (1988) Inﬂuence of Paclobutrazol on in vitro growth of sweet cherry shoots.
HortScience 23: 304-305

Strauch, H. and Gruppe, W. (1985) Results of laboratory tests for winter hardiness of P.

avium cultivars and interspeciﬁc cherry hybrids (Prunus x spp). Acta Horticulturae
169: 281-288

Wagner, H. and Gruppe, W. (1985) Swelling of scion trunk above graft unions of cherry
trees. Acta Horticulturae 169: 269-274

Walser, RH. and Davis, TD. (1989) Growth, reproductive development and dormancy
characteristics of paclobutrazol-treated tart cherry trees. Journal of Horticultural
Science 64: 435-441

Webster, A.D. ( 1995) Rootstock and interstock effects on deciduous fruit tree vigour,
precocity, and yield productivity. New Zealand Journal of Crop and Horticultural
Science 23: 373-382

Webster, AD. and Lucas, A. (1997) Sweet cherry rootstock studies: Comparisons of
Prunus cerasus L. and Prunus hybrid clones as rootstocks for Van, Merton Glory and
Merpet scions. Journal of Horticultural Science 72: 469-481

Webster, AD. (1998) Strategies for controlling the size of sweet cherry trees. Acta
Horticulturae 468: 229-239

Weibel, A., Johnson, RS. and DeJong, T.M. (2003) Comparative vegetative growth
responces of two peach cultivars grown on size-controlling versus standard
rootstocks. Journal of the American Society for Horticultural Science 128: 463-471

67

CHAPTER 2

IDENTIFICATION OF GENES EXPRESSED AT THE CRITICAL POINTS IN
GROWTH BETWEEN DWARF AND SEMI-VIGOROUS GRAFT COMBINATIONS

68

INTRODUCTION

A signiﬁcant amount of research has been performed in cherries and other fruit
trees for the identiﬁcation of the leading causes of rootstock induced dwarﬁng. Most of
the research has focused on the physiological changes between dwarf and vigorous trees,
as described in the Literature Review. Existing hypotheses involve changes in the
concentration of chemical signals, such as hormones and phenolics, changes in the
anatomy of the graft union or changes in the physiology of the trees (Literature Review).
The leading cause for the ability of some rootstocks to dwarf scions lies in on the genetic
makeup of those trees in comparison to vigorous rootstocks. Since allelic variation occurs
in more than one genetic locus, it is expected that the varying degrees of vigor are due to
many loci. This is supported by a map based cloning approach in apple aiming to identify
dwarﬁng loci (Rusholme et al. 2004). Mapping resulted in the identiﬁcation of a single
dominant locus, though this could not explain 100% of the variation occurring in the
segregating population. Many of the new cherry rootstocks are crosses between diploid
and tetraploid Prunus species and they show a wide variety of vigor even within the same
cross (Gruppe, 1985). This is a result of the numerous possible combinations of alleles
originating from the parent species. Identiﬁcation of this genetic variation would advance
our knowledge of the interaction between rootstock and scion, and would lead to more
efﬁcient and rapid breeding or genetic engineering of new rootstocks. Recently, the
importance of this genetic variation has been identiﬁed and a number of studies have
emerged, attempting to identify genes involved in this phenomenon (Rusholme et al.
2004; Jensen et al. 2004). It is still early to conclude with certainty which signaling

pathways or what enzymatic reactions control tree growth in a timely fashion.

69

The meristems responsible for the increase in plant size are the shoot apical
meristem (SAM) and the vascular cambium. The activity of SAM determines aerial part
architecture and plant size, whereas the cambium is involved in the increase of stem girth.
Regulation of SAM has been extensively studied in plants, with Arabidopsis being the
model for SAM development. Stem cell identity at the SAM is maintained by the
W USCHEL (W US) - CLA VA TA (CL V) interaction system (Williams and Fletcher, 2005).
Expression of WUS is localized in the organizing center (OC), which is formed by the
stem cells. Differentiation of stem cells occurs by the production of CLV3, a small
peptide that binds to the CLVl-CLV2 receptor complex and signals the suppression of
WUS outside the OC (Williams and Fletcher, 2005). SHOOT MERISTMLESS (ST M) is
another gene involved in the proliferation of the stem cells, acting in parallel to WUS in
the maintenance of stem cells at the SAM (Veit, 2006). This is a very simplistic
description of the regulation of stem cell maintenance that describes the major players in
this process. New genes continue to be identiﬁed as suppressors or inducers of stem cell
identity at the SAM. Cell divisions at the OC occur continuously during active grth to
promote differentiation of new organs without depleting the OC of stem cells. Activation
or deactivation of cell cycle related proteins occur at the post-translational level through
phosphorylation or protein degradation (Horvath et al. 2003; Gegas and Doonan, 2006).
During dormancy induction, cells at the apical meristem are arrested in the G1 phase of
cell divisions, before replication of the DNA (Gegas and Doonan, 2006). The control of
cell division during induction of dormancy is not well known. In poplars, shoot grth
cessation is controlled by the CO/F T regulatory pathway (Bohlenius et al. 2005). PtF T 1

down-regulation through siRNA causes faster shoot growth cessation under short days

70

compared to wild type trees, indicating the importance of PtF T 1 in this process. The
effect of PtF T 1 on cell cycle related genes has not been established. At the hormonal
level, auxin and cytokinin are the most important players in the control of cell divisions
(Shani et al. 2006; Veit, 2006). Cytokinin signaling positively regulates ST M, but is
negatively regulated by W US to create a more precise control of SAM proliferation
(Shani et al. 2006). Auxin has a role in organ development through suppression of ST M
and CUP SHAPPED COTYLEDONI (CUCI) (Shani et al. 2006; Veit, 2006). The
seasonal regulation of SAM activity has not been studied extensively. At the molecular
level, cessation of shoot growth in trees is an uncharted territory. More research is
necessary to identify the mechanisms regulating cell cycle, hormone signaling and stem
cell differentiation at the SAM on a seasonal basis.

Tree growth measurements described in Chapter 1 have identiﬁed the critical
points that differentiate dwarf from vigorous trees in cherry grafts. More speciﬁcally,
‘Bing’/Gi5 shoots consistently ceased growing earlier than ‘Bing’/Gi6 trees across three
growing seasons (Figure 1.3, Chapter 1). A similar response was identiﬁed for node
number, as nodes in ‘Bing’/Gi5 trees stopped being produced earlier than in ‘Bing’/Gi6
trees, resulting in the cessation of shoot growth (Figure 1.8, Chapter 1). It was expected
that changes in gene expression would precede shoot growth cessation at least for genes
involved in the early stages of the response, while at the same time the same genes in
‘Bing’/Gi6 trees would not be modiﬁed, but only later. During shoot growth cessation in
‘Bing’/Gi5 trees, changes in gene expression also were expected in the graft union, which

is the link between rootstock and scion. Crossing of the signal through the graft union

71

will occur at or before the time of shoot growth cessation. Thus, dwarﬁng rootstocks are
expected to show earlier changes in gene expression.

A genomics approach was taken for the identiﬁcation of genes differentially
expressed between the main shoot and the graft union of dwarf ‘Bing’/Gi5 and semi-
vigorous ‘Bing’/Gi6 trees. The main shoots of both graft combinations are of the same
genetic background, but the rootstocks are the siblings of the interspeciﬁc cross between
P. cerasus cv. ‘Schattenmorelle’ and P. canescens. As siblings, the two rootstocks are
expected to share a high genetic homology, but also allelic variation (due to self-
incompatibility driven heterozygosity). Complementary DNA Ampliﬁed Fragment
Length Polymorphism (cDNA-AF LP) was used to screen main shoot and graft union
samples and reveal a number of differentially expressed genes. Microarrays were
constructed to conﬁrm the differential expression of those genes. Ninety-nine genes were
conﬁrmed as differentially expressed, of which 43 were in the main shoot and 56 in the

graft union.

MATERIALS AND METHODS

Plant material

Trees used in this experiment were purchased from commercial nurseries. The
graft combinations were ‘Bing’/Gi5 and ‘Bing’/Gi6 in which the scion was l-year-old
and the rootstock 2-years-old when purchased. ‘Bing’ is a commercial sweet cherry
cultivar while Gi5 and Gi6 rootstocks are triploid F 1 progeny from an interspeciﬁc cross

between the tetraploid sour cherry (Prunus cerasus L. cv. Schattenmorelle) and the

72

diploid greyleaf cherry (Prunus canescens Bois.) (Franken-Bembenek, 1996). The trees
for each graft combination were planted in the spring of 2001 in two rows of 50 trees
each with 6 meter row spacing and a 2 meter tree spacing with a North to South
orientation at the MSU Clarksville Horticultural Experiment Station, Clarksville,
Michigan. Pruning was performed every spring prior to bud break so that only the main

stem was retained above ground. Flowers were removed before fruit set.

Sampling

In 2002, four trees were sampled per graft combination, ‘Bing’/Gi5 and
‘Bing’/Gi6, on 3 June, 20 June, 3 July. The ‘Bing’ shoots were sampled by removing the
main shoot, defoliating it and freezing it on dry ice. The trees were then removed from
the soil and the following tissue samples were obtained: rootstock (a 10 cm region
directly below the graft union), graft union (including the swollen tissues of the rootstock
and the scion) and scion (a 10 cm region directly above the graft union). All tissue
samples were directly ﬁ'ozen in dry ice within separately labeled bags. For long term

storage, the samples were kept in a -800 C ultralow freezer.

RNA extraction

For total RNA extraction, samples from each tree were ground separately in a
mortar and pestle for shoots, or stainless steel blender (Waring, Connecticut) and
subsequently stainless steel coffee grinder (BCGIOO, Kitchenaid, Michigan) for woody
tissues (rootstock, graft union, scion). During grinding, tissues were kept frozen with

liquid nitrogen. From the main shoot samples, the part containing the upper 10 buds was

73

used for grinding. Of that sample, the upper 3 nodes including the apical meristem (called
“shoot apex”) were ground separately from the “remaining shoot”. Equal amounts of
ground sample from three of four trees were mixed in a centrifuge tube totaling 2 ml of
ground tissue. For the main shoot, 500 u] of the “shoot apex” and 1500 [.11 of the
“remaining shoot” from three trees were mixed to produce 2 ml of tissue. This action was
taken to enrich the extracted RNA with shoot apex mRNAs. For RNA extraction the
protocol of Wang et al., (2000) was used. This protocol was designed for RNA extraction
from plant tissue samples with high content in polyphenols and polysaccharides. Brieﬂy,
the protocol is as follows:

Five volumes of homogenization buffer (0.3M LiCl, 200 mM Tris-HCl pH 8.5, 10
mM EDTA, 1.5% Sodium dodecyl sulfate, 1% (w/v) sodium deoxyholate, 1% (v/v)
NP-40, 1 mM aurintricarboxylic acid, 10 mM DTT, 5 mM thiourea, 2% (w/v)
PVPP) were mixed with one volume of ground tissue. The mixture was incubated at -800
C for 2 hours and then heated to 370 C until just thawed. Plant debris was removed by
centrifugation at 5,000xg for 20 min. The supernatant was mixed with 1/30th of the
volume 3M sodium acetate pH 5.2 and 100% ethanol to a ﬁnal volume of 10%. The
mixture was placed on ice for 10 min and then centrifuged at 5,000xg for 20 min. This
step removes carbohydrates that are found in large concentration in cherry stems. The
supernatant was mixed with 1/9th of the volume 3M sodium acetate pH 5.2 and 30% of
the ﬁnal volume iso-propanol. The mixture was placed at -200 C for 2 hours, followed by
centrifugation at 5,000xg for 30 min. The pellet was diluted in 3 m1 of TE buffer (10 mM
Tris-HCl pH 8.0, 1 mM EDTA pH 8.0) and placed on ice for 30 min. The samples were

then centrifuged at 10,000xg for 30 min and the supernatant was transferred to a fresh

74

tube, where it was mixed with 1A of the supernatant volume 10M lithium chloride. After
mixing well, the samples were placed at 40 C overnight. This step precipitates exclusively
the RNA without transfer of double stranded DNA. The next day they were centrifuged at
10,000xg for 30 min. The pellet was resuspended in 1.5 ml of TE buffer pH 8.0 and
mixed with 1.5 volumes 5M potassium acetate (pH not adjusted) and placed on ice for 3
hours. This step precipitates the RNA. Next, the samples were centrifuged for 30 min at
10,000xg. The pellet was resuspended in 1 ml of TE buffer pH 8.0 and placed on ice for
30 min, followed by centrifugation at 10,000xg for 30 min. The supernatant was mixed
with 1/10th of the volume 3M sodium acetate pH 5.2 and 2 volumes of 100% ethanol. It
was then placed at -200 C for 2 hours, followed by centrifugation at 10,000xg for 30 min.
The pellet was washed in 75% ethanol and centrifuged at 10,000xg for 10 min. The pellet
was air dried and then resuspended in 30 ul of diethyl pyrocarbonate-treated (DEPC)
water. RNA quantity was measured on a SmartSpec3000 (BIO-Rad, California)

Spectrophotometer at OD260 with OD260/230 serving as the quality control.

cDNA-AFLP analysis

The protocol was an adaptation of the protocol described by Bachem et al. (1996).
i) Preparation of double stranded cDNA

One hundred micrograms of total RNA were used for puriﬁcation of polyA RNA.
The Dynabeads® (#61006, Dynal, New York) or poly(A)PuristTM Mag (#1922, Ambion,
Texas) kits were used for the puriﬁcation. Two hundred and ﬁfty nanograms of polyA
RNA were used for the preparation of double stranded cDNA. For synthesis of the ﬁrst

strand, polyA RNA was mixed with 1 ng of oligodes primer and water up to 12 11.1 and

75

incubated at 700 C for 10min. After denaturation, 4 ul of lSt strand buffer, 2 111 of 0.1M
DTT, 0.5 ul of 20mM dNTPs and 0.5 ul of RNAse inhibitor (Promega, Wisconsin) were
added. The mixture was heated to 420 C for 2 min, then 1 ul of SuperscriptII (200U/ul,
Invitrogen, California) was added and the reaction was incubated at 420 C for 50 min
followed by heat inactivation at 700 C for 15 min. For synthesis of the second strand, 8 [.11
of 10x 2nd strand DNA polymerase buffer, 2 pl 20 mM dNTPs, 1 pl RNaseH (2U/ul,
Invitrogen, California), 4 111 DNA Polymerase I (9U/ul, Promega, Wisconsin) and 45 ul
water were added to the ﬁrst strand reaction and incubated at 160 C for 2 hours and 30
min, followed by deactivation at 680 C for 10 min. At this stage, 5 11.1 of the reaction were
analyzed on a 1% agarose gel to examine if the concentration between samples remained

equaL

ii) Preparation of the primaﬁ template

The cDNA was precipitated with 10% 3M sodium acetate pH 5.2 and 2.5 volumes
of ethanol at -800 C for 2 hours or at -200 C for 8 hours. Samples were then centrifuged at
10,000xg for 30 min. The pellet was washed with 70% ethanol and centrifuged at
10,000xg for 10 min. The pellet was air dried and resuspended in 10 [.11 of HPLC grade
water. The cDNA was digested with the MseI and ApoI restriction enzymes. Since the
two enzymes have different restriction temperatures, the digestions occur sequentially
and not at once. For the MseI digestion, 10 ul of cDNA were mixed with 4 ul 10x NEB4
buffer, 0.4 ul lOOxBSA, 111.1 of MseI (IOU/[11, NEB, MA) and 25 pl of HPLC grade
water. The reaction was incubated at 370 C for 2 hours and heat inactivated at 650 C for

30 min. For the ApoI reaction, the following components were added to the MseI

76

reaction: 1 11.1 10x NEB4 buffer, 0.1 ul lOOxBSA, 2 ul ApoI (4U/ul, NEB, MA) and 7 0]
HPLC grade water, This was incubated at 500 C for 2 hours and heat inactivated at 800 C
for 20 min. Following the digestion, ApoI and MseI speciﬁc adaptors are added to the

digestion reaction. The adaptor sequences are:

Apo-adap-top: 5' - CTC GTA GAC TGC GTA cc - 3'

Apo-adap-bot: 5’ - AAT TGG TAC GCA GTC TAC - 3 ’

Mse-adap-top: 5’ - GAC GAT GAG TCC TGA G - 3 ’

Mse-adap-bot: 5' - TAC TCA GGA CTC AT — 3’

Prior to ligation, both strands of the adaptor were heat denatured at 650 C for 10 min and
allowed to cool to room temperature.

For ligation of the adaptors to the digested cDNA, the following components were
added: 0.5 ul 10x NEB4 buffer, 1 ul ApoI adaptor (5 pmoles), 1 [1.1 MseI adaptor (50
pmoles), 0.6 111 IM DTT, 1.6 11] 1M Tris-HCl pH 7.5 and 0.3 [11 T4 DNA ligase (3U/ul,
Promega, Wisconsin). The mixture was incubated at 160 C overnight. Four microliters of
the reaction were loaded on a 1% agarose gel to inspect the quality of digestion. Bands

should appear at sizes 100-1,000bp.

iii) Preparation of the secondary template

Secondary template is the PCR ampliﬁed primary template. The ampliﬁcation

primers are designed to anneal on the adaptor sequences. Primer sequences are:

77

Apo-pre: 5' - CTC GTA GAC TGC GTA CCA ATT - 3'

Mse-pre: 5’ - GAC GAT GAG TCC TGA GTA A - 3'

The PCR reaction was as follows: 10 ul of Primary template, 5 ul of 10x Taq
Polymerase buffer, 5 111 Mng (25 mM ), 5 ul dNTPs (2 mM), 1 [.11 Apo-pre (10
pmoles/ul), 1 ul Mse-pre (10 pmoles/ul), 0.5 u] Taq polymerase (5 U/ul, Promega,
Wisconsin) and 22.5 111 of HPLC grade water. The ampliﬁcation program was as
follows: 940 C for 30 sec, 520 C for 30 sec, 720 C for 1 min in 15 cycles. A 5 [11 sample
of the reaction was loaded on a 1% agarose gel to inspect the quality and quantity of
DNA. At this point, it was very critical to have equal amounts of DNA for all the
samples that were to be used in the cDNA-AF LP analysis. Quantiﬁcation was performed
on a 1% agarose gel using the gel image analysis software ImageQuant (Molecular

Dynamics, California).

iv) Selective ampliﬁcation of the secondary template

At this point, the secondary template was subjected to selective ampliﬁcation to
reduce the number of cDNA fragments present in each sample. For that reason, primers
similar to the pre-ampliﬁcation primers, but with a 3’ extension into the sequence of the

unknown gene, promoted selective ampliﬁcation. The primers used were:

Apo—sel-CG : 5’ - GAC TGC GTA CCA ATT CG - 3’

Apo-sel-CA : 5’ - GAC TGC GTA CCA ATT CA - 3’

78

Apo—sel—CC : 5’ - GAC TGC GTA CCA ATT CC - 3’

Apo-sel—CT : 5’ - GAC TGC GTA CCA ATT CT - 3’
Apo-sel-TG : 5’ - GAC TGC GTA CCA ATT TG - 3’
Apo-sel-TA : 5’ - GAC TGC GTA CCA ATT TA - 3’
ApO-sel-TC : 5’ - GAC TGC GTA CCA ATT TC - 3’
Apo-sel-TT : 5’ - GAC TGC GTA CCA ATT TT - 3’

The Apo-sel primers had only two variable nucleotides in position 16 because the

restriction site includes either G or A.

Mse-sel-GG : 5’ - GAT GAG TCC TGA GTA AGG - 3’
Mse—sel—GA : 5’ - GAT GAG TCC TGA GTA AGA - 3'
Mse-sel-GC : 5’ - GAT GAG TCC TGA GTA AGC - 3’
Mse-sel-GT : 5’ - GAT GAG TCC TGA GTA AGT - 3'

Mse-sel-AG : 5’ - GAT GAG TCC TGA GTA AAG - 3’

Mse-sel—AA : 5’ - GAT GAG TCC TGA GTA AAA - 3’
Mse-sel-AC : 5’ - GAT GAG TCC TGA GTA AAC - 3’
Mse-sel—AT : 5’ - GAT GAG TCC TGA GTA AAT - 3’
Mse—Sel-CG : 5’ - GAT GAG TCC TGA GTA ACG - 3’
Mse-Sel-CA : 5’ - GAT GAG TCC TGA GTA ACA - 3’
Mse-sel-CC : 5’ - GAT GAG TCC TGA GTA ACC - 3’
Mse—sel-CT : 5' — GAT GAG TCC TGA GTA ACT - 3’
Mse-sel-TG : 5' — GAT GAG TCC TGA GTA ATG - 3’
Mse—sel-TA : 5’ - GAT GAG TCC TGA GTA ATA - 3’

79

Mse-Sel-TC : 5’ - GAT GAG TCC TGA GTA ATC - 3’

Mse-sel-TT : 5’ - GAT GAG TCC TGA GTA ATT - 3’

Visualization of the PCR product was accomplished by radioisotope labelling of
the 5’ nucleotide of the Apo-sel primers. Seven pmoles of labeled primer were used for
each reaction. Depending on the number of wells for which each selective primer was
going to be used its concentration was adjusted adequately for labeling. For example, if
the primer was going to be used in 100 re-ampliﬁcations, we needed 700 pmoles. The
labeling reaction was: 80 [r] (700 pmoles) of Apo-selective primer, 10 pl 10x PNK buffer
(NEB, Massachusetts), 4 11.1 [qr-33P] ATP (NEG602H , NEN, Massachusetts), 3 ul PNK
(N EB, Massachusetts) and 3 [rl HPLC grade water. The reaction was placed at 370 C for l
houn

The selective PCR reaction was: 1 ul template (1/50 dilution of secondary
template in water), 1 pl 10xTaq polymerase buffer, 1 ulMgC12(25 mM), 1 ul dNTPs (2
mM), 1 ul labeled Apo-Sel (7 pmol/ul), 1 ul Mse-Sel (7 pmol/ul), 0.5 ul Taq polymerase
(Promega, Wisconsin) and 3.5 ul HPLC grade water. The program used for the
ampliﬁcation was: 940 C for 30 sec, 650 C for 30 sec, 720 C for 1 min repeated in 10
cycles and followed by 940 C for 30 sec, 560 C for 30 sec, 720 C for l min repeated 25
times. After the PCR reaction ﬁnished, 2.5 ul of gel loading dye were added and the
reaction was denatured at 990 C for 5 min.

A 5% polyacrylamide gel was previously prepared. The PCR reaction was heat
denatured at 1000C for 5 min and then cooled on ice. For loading, 2.5 ul of the reaction
were used. A 50 bp step ladder (Promega, Wisconsin) was labeled with [7-33P] ATP in a

reaction similar to that for the Apo-sel primers and 2.5 ul were loaded on the side lanes of

80

the gel. The gel was pre-run for 1 hour at 80 Watt and after loading it ran for 3.5 hours.
The gel was then attached to a Whattman 3MM paper, covered on the other side with
Saran wrap and dried on a 583 Gel dryer (BioRad, California) for 2 hrs at 80°C. The dry
gel was placed in a cassette with Glogos® II Autorad Markers (Stratagene, California)
attached to its corners for alignment during excision. An X-ray ﬁlm (Kodak, New York)
was placed on top of the gel and exposed for 24 hours at -800 C. The ﬁlm was developed
by soaking for 2 min in developing solution, washed brieﬂy in water, ﬁxed for 2 min and

then air dried.

Excision, cloning and sequencing of differentially expressed bands

The autoradiographic ﬁlm was aligned to the dried gel and selected fragments
were excised by cutting both ﬁlm and gel using a scalpel (No l 1). These are called
Transcript Derived Fragments (TDFs). The excised TDFs were soaked in 50 pl of
distilled water overnight at 370 C to elute the DNA. 5 111 were used for re-ampliﬁcation of
the TDFs in a 20 pl PCR reaction using the same primers as for pre-ampliﬁcation (Apo-
pre 5 ’-CTCGTAGACTGCGTACCAATT-3 ’; Mse-pre, 5 ’-
GACGATGAGTCCTGAGTAA-3’). The following program was used: 30 sec at 940 C,
30 sec at 600 C and 1 min at 720 C for 30 cycles. The products were separated on a 1%
agarose gel and the double or smeared bands were excluded from further processing. The
remaining TDFs were puriﬁed by precipitation in 2 volumes of ethanol and 10% 3M
sodium acetate pH 5.2, overnight at -200 C, washed in 100% ethanol, air dried and re-

suspended in distilled water.

81

TDFs selected for sequencing were cloned into the pGEM-T Easy vector
(Promega, Wisconsin) and inserted into Subcloning efﬁciency E.coli DH5a cells
(#18265-017, Invitrogen, California) according to Dilks et al., (2003). Four colonies were
selected for each TDF which were then screened using a colonies PCR reaction with the
following primers: M13F 5’-GTTTTCCCAGTCACGACGTTG-3’ and M13R 5’-
GAGCGGATAACAATTTCACACAG-3’. The colonies were diluted in the reaction
mixture in a 96-well format and then streaked on an LB agar plate containing 100 ug/ml
ampicillin, divided in a 96-well format and incubated at 370 C overnight. The reaction
conditions were the same as for the re-ampliﬁcation reaction. Positive colonies were then
transferred from the agar plate to a 96-deep well plate containing 350 pl of LB with 100
ug/ml ampicillin and incubated at 370 C overnight with constant agitation. 150 pl of 50%
sterile glycerol were then added for storage at -800 C. 1 pl of the culture was used for
ﬁrrther conﬁrmation of the clones with the Apo-pre/Mse-pre primers.

The M13F primer was used for sequencing the cloned TDFs in an ABI Prism
3700 DNA analyzer. Sequencing was performed at the Genomics Technology Support

Facility, Michigan State University.

Microarray construction

The previously selected TDF clones were PCR ampliﬁed using 0.5 ul of the
glycerol stock cells or 1/100th dilution of PCR product for not cloned TDFs, as template.
The reaction was set to 5011.] ﬁnal volume in 96-well plates. Apo-pre and Mse-pre were
used as the ampliﬁcation primers in the following reaction: template 0.5 ul, 1x Taq

polymerase buffer (Promega, Wisconsin), 200 p.M dNTP mix, 10 pmole Apo-pre, 10

82

pmole Mse-pre, 1 Unit Taq polymerase (Promega, Wisconsin) and HPLC grade water up
to 50 111. The reaction conditions were: 940 C for 30 sec, 600 C for 30 sec, 720 C for 1 min
for 40 cycles. Four microliters of the product were analyzed on a 1% Agarose gel and the
remaining sample was precipitated with 10% 3M sodium acetate pH 5.2 and 2 volumes
of 95% ethanol overnight at -200 C. The next day the plates were centrifuged at 3,000
rpm for 40-50 min on a bench top centrifuge (HN-SI, Damon/IEC) and then washed in
75% Ethanol, followed by centrifugation at 3,000 rpm for 10 min. The pellet was air
dried until no water was visible, re-suspended in 15 ul of 3xSSC solution to
approximately 100 ng/ul and stored at 400 C for 24 hours and then transferred to -200 C.
DNA was transferred to 384 well plates to facilitate robotic printing. The Arrayit
SuperAmine glass slides (SMM, Telechem, California) were used as substrate on a
GeneMachines OmniGrid 100 robot (Genomic Solutions, Michigan) with Telechem
Chipmaker pins. Each slide contained 1040 DNA samples printed in triplicate on distant
locations of the slide to avoid position speciﬁc bias. The microarray was formed by 24

grids (each grid formed by one pin), each of which contained 132 spots or less.

Microarray hybridization

Probe was labeled with the amino-ally] method described by Hegde et al. (2000).
The RNA used for probe preparation was the same as that used for the cDNA-AF LP
analysis, in addition to RNA from the fourth tree that served as the biological replication.
Arrays were hybridized at 420 C for 16 hr and washed once in 1xSSC, 0.2% SDS at 420 C
for 5 min, 0.1xSSC, 0.2% SDS at room temperature for 5 min, and 0.1xSSC at room

temperature for 5 min.

83

The experimental design was as follows:

Upper shoot Graft union
B5 6/3<—>B5 6/20<—>B5 7/3 BSRHBS GUHBSS
1 1 1 I l 1
B6 6/3+—>B6 6/20<—>B6 7/3 B6R+—>B6 GUHB6S

 

BS: ‘Bing’/G15 tree, B6: ‘Bing’/G16 tree, 6/3: mm/dd/2002, R: rootstock, GU: Graft
union, S: Scion, two-headed arrows denote dye reversal. Dye reversal served as the
technical replication and two RNA samples from independent trees served as the

biological replication.

Microarray analysis

The slides were scanned in an Affymetrix 428 microarray scanner (Affynetrix,
CA) and data were analyzed with the GenePix Pro v3.0 software (Molecular Devices,
CA). Lowess normalization and ANOVA were performed using the R/maanova package
(Wu et al. 2002; Churchill, 2004; http://www.jax.org/staff/churchill/labsite/index.html).
The models used for the two experiments were:
a) Main shoot

y = Anay+spot+Dye+Date+R00tStock+DateIROOtStOCk'l'SammC

We were interested in the Date, Rootstock and DatezRootstock effect

b) Graft union

y = Array+Spot+Dye+P0sition+Rootstock+Position:Rootstock+Sample

84

We were interested in the Position, Rootstock and Positioanootstock effect.

Expression ratios at the log2 scale between ‘Bing’/G15 and ‘Bing’/G16 samples
within Date or Position were obtained with the SMA package
(http://www.stat.berkeley.edu/users/terry/Group/software.htrnl) in an R language
environment. Signiﬁcantly differentially expressed genes were those that showed a P-
value lower than 0.05, more than 1.5 fold-difference within Date or Position, and had
more than 3 spots not ﬂagged as bad per clone per sample. Grouping of the signiﬁcant
genes was performed manually due to the small number of genes. Genes were sorted
based on their fold-difference and then grouped based on the similarity of their

expression pattern.

Northern hybridization

Total RNA was extracted as above. A 5 pg sample of total RNA was analyzed on
a denaturing 1% agarose gel containing formaldehyde. RNA was then transferred on a
nylon membrane (Hybond N+, Amersham-Biosciences, New Jersey) according to
Sambrook et al. (1989). The probe was prepared from the cDNA-AF LP re-ampliﬁed
band using primers Apo-pre/Mse-pre as before. The probe was gel puriﬁed with the
QIAEX 11 gel puriﬁcation system (QIAGEN, California). For probe labeling, 25 ng of
puriﬁed PCR product were incorporated in the one tube reaction RedyPrime system
(Amersham-Biosciences, New Jersey) with the addition of [01-3 2P]ATP (N EN-Perkin

Elmer, Massachusetts). The hybridization reaction was performed at 42°C for 16 h and

85

the membrane was then washed (according to the manufacturer) and exposed to an X-ray

ﬁlm (Kodak, Connecticut).

RESULTS

Complementary DNA amplified fragment length polymorphism (cDNA-AFLP)
analysis of scion main shoots in grafts showing differential cessation of growth
Comparison of growth between dwarﬁng and semi-vigorous trees revealed that
‘Bing’/Gi5 shoots cease growing earlier than ‘Bing’/Gi6 shoots (Figure 1.4, Chapter 1).
Differentiation in shoot growth between these two graft combinations occurred
approximately on the 17th of June 2002. To understand the genetic changes leading to this
differential growth, a differential display cDNA-AF LP screen was performed on main
shoot samples of ‘Bing’/Gi5 and ‘Bing’/Gi6 trees, collected in 2002. Using the shoot
elongation curve as a reference, samples were collected on three different dates from each
graft combination. The ﬁrst sample was collected on 3 June, when shoots in both
combinations were elongating at the same rate. The second sample was collected on 20
June at the point of initiation of differential shoot elongation. Finally, the third sample
was collected on 3 July when both combinations had reduced elongation rates and
approached cessation of terminal meristem growth. The restriction enzymes ApoI and
MseI were used in the cDNA-AF LP analysis. ApoI can recognize 4 consensus sequences
(RAATTY, R=A or G, Y=C or T), including that of EcoRI (GAATTC), thus reducing its
hypothetical restriction band size to 1,024 bp, instead of 4,096 bp for the regular six

cutter enzymes. Such an enzyme combination can increase the representation of cDNAs

86

screened by four-fold compared to the standard cDNA-AF LP protocol. Selective primers
for MseI had two random nucleotides in the 3’ end giving rise to 16 primers, while ApoI
selective primers had a C or T in the second to last 3’ end and a random nucleotide in the
last position giving rise to 8 primers. Thus the number of all possible primer
combinations consisted of 128 primer pairs. At a rate of 100 bands per primer set after
PCR ampliﬁcation and 6 samples per primer set, the total number of transcript derived
fragments (TDFs) was 76,800 (128 x100 x 6).

Overall gene expression was similar between the two graft combinations for
samples collected at the same day. Only 111 Transcript Derived Fragments (TDFs)
showed differences in their expression level between ‘Bing’/Gi5 and ‘Bing’/Gi6 trees
(Figure 2.1). The change in expression occurred earlier in ‘Bing’/Gi5 samples, as it is
clear in Figure 2.1. Differences in expression were apparent as early as the 3rd of June
when shoots of ‘Bing’/Gi5 and ‘Bing’/Gi6 were still growing at the same rate. Some of
the most common expression patterns are presented in Figure 2.1B. Patterns 1-4 show
genes that were down-regulated ﬁrst in ‘Bing’/G15 shoots, while patterns 5-8 show genes
that were up-regulated ﬁrst in ‘Bing’/Gi5 shoots (Figure 2.1B). Patterns 9 and 10
represent genes whose expression changes at the same time for ‘Bing’/Gi5 and’Bing’/Gi6
shoots, in response to a non-deﬁned signal (Figure 2.1B). Pattern 1 shows a gene which is
down-regulated in ‘Bing’/Gi6 shoots while it is not already transcribed at the time of
sampling in ‘Bing’/Gi5. Pattern 2 shows a gene with an expression change that seems to
follow the change in pattern 1, but is consistently down-regulated ﬁrst in ‘Bing’/Gi5
trees. In patterns 3 and 4 gene expression is not completely blocked, but it is down-

regulated ﬁrst in ‘Bing’/Gi5 on June 20. Such genes are expected to be downstream of

87

genes exhibiting patterns 1 and 2 and act in the late stages of shoot growth cessation.
Since samples represent only points in time the exact date of change in expression of
those genes is not known. In contrast to patterns 1-4, pattern 5 shows a gene that is up-
regulated in ‘Bing’/Gi5 on the 3rd of June, while in ‘Bing’/Gi6 up-regulation occurs later
on the 20th of June. These genes are expected to be positive regulators of shoot growth
cessation. Pattern 6 shows a gene which is up-regulated ﬁrst in ‘Bing’/Gi5 shoots on the
20th of June and later on the 3rd of July in ‘Bing’/Gi6. Patterns 7 and 8 represent genes
that show a late up-regulation in ‘Bing’/Gi5 on the 3rd of July. The same genes have very
low expression or they are not expressed in ‘Bing’/Gi6. Such genes may not be involved
in the process of shoot growth cessation, but rather in other processes, such as formation
of ﬂoral primordia or induction of dormancy. All 111 TDFs were cloned in a plasmid
vector to obtain pure fragments. Cloning produced 277 clones and each TDF was

represented by one or more clones.

cDNA-AFLP analysis of the graft union area during shoot growth cessation

The graft union presents another point of inﬂuence in the dwarﬁng phenomenon.
It was expected that analysis of gene expression would reveal genes that are either
differentially expressed in the rootstock, graft union and scion or that are differentially
transported through the graft union. The second hypothesis is the most interesting, but
also more difﬁcult to prove. cDNA-AFLP analysis of the main shoot revealed that the
majority of the genes differentially expressed between ‘Bing’/Gi5 and ‘Bing’/Gi6 trees

occur in 20 June sample. Rootstock tissue, graft union and scion trunk above the graft

88

union were used in a second cDNA—AF LP analysis to screen the graft union region for
possible differential gene expression.

Overall, there was a high level of similarity in the gene expression proﬁles
between rootstock and scion (Figure 2.2), even though they represent different species.
The genotype of the scion is P. avium and the rootstock is a hybrid of P. cerasus and P.
canescens. The genetic similarity revealed by the cDNA-AF LP analysis can be explained
by the close phylogeny of P.avium to P.cerasus (Bortiri et al. 2001). The most variable
pattern is shown in Figure 2.2 (pattern 5), in which the TDF is only present in the graft
union of the ‘Bing’/Gi6 trees. Sequencing of TDFs exhibiting this cDNA-AF LP proﬁle
revealed genes with high similarity to peach, apricot or almond ESTs, which excludes the
possibility of contamination of the sample by other organisms. Also, alignment of TDFs
with those ESTs excluded the presence of genomic DNA that might have contaminated
the RNA. Northern hybridization and RT-PCR analysis with independent RNA samples
for both graft combinations did not show the expected difference in expression, but rather
equal level of expression (data not shown). This further complicates the reason for this
pattern, because it excludes the possibility that the ‘Bing’/Gi6 sample was contaminated.
If ‘Bing’/Gi6 samples were contaminated, we should not expect to see expression in
‘Bing’/G15 samples. Any further step to characterize this pattern was not taken, as the

cDNA-AF LP data could not be replicated.

89

 

 

 

 

(a) “9° 2A. E E 29. (b) 316 2016 3!?

“’° °° °° °° CT BEEELSEEE
1
2 v. ‘e.
3 “. a.
4 b-‘ueun-
6 Q~- m, “a.
6 ~ ”n
7 l
8 aft-w
9 ”quail. “-1-— ’~-'

. ﬁa-r:‘“*.-f" 10 G “I“ a...

 

 

 

Figure 2.1: cDNA-AFLP analysis at the main shoot. (A) Portion of a cDNA-AF LP gel
produced by four primer pairs and showing high degree of co-regulation between
samples. (B) Detail of gene expression patterns putatively related to shoot growth
cessation, identiﬁed in 8 cDNA-AFLP gels as the one shown in (A). 3/6: 3 June, 20/6: 20
June, 3/7: 3 July, BS: ‘Bing’/Gi5, B6: ‘Bing’/Gi6.

9O

 

 

 

 

 

(army. 0_0 g E (b) R G” 8
“8° AC AC AC AT E_EEEE
1 .
........ 2 any
T-~-" ‘ 3 -- “5"-
u‘w ":-—_ 4
W 5 . ‘
“-7--..” u 6 - al.—o-
”‘woa- .. . 7 can't-tum
m
_ _-~ “~01.“ ’- 8
"-
9 can“...

 

 

 

 

Figure 2.2: cDNA-AFLP analysis at the graft union. (A) Portion of a cDNA-AF LP gel
produced by four primer pairs and showing high degree of co-regulation between
samples. (B) Detail of interesting expression patterns identiﬁed in 8 cDNA-AF LP gels as
the one shown in (A). R: Rootstock, GU: Graft Union, S: Scion, B5: ‘Bing’/Gi5, B6:
‘Bing’/Gi6.

91

Genes, expressed either in the rootstock or scion, also were present in the graft
union, since this tissue combines cells from both genotypes (Figure 2.2B, patterns 1-4, 6-
10). In some cases (Figure 2.2B, patterns 9 and 10), genes exhibited a gradual reduction
in expression from the rootstock to the scion and vice-versa. This pattern may indicate
RNA transport across the graft union through the vascular system. It is not known
whether or not this pattern was due to RNA transport across the graft union through the
vascular system, since it proved difﬁcult to extract phloem sap from cherry trees.

The cDNA-AF LP analysis revealed 249 differentially expressed TDFs. Sixty-four
of the TDFs were differentially expressed in the rootstock (Figure 2.2B, patterns 1-4, 10),
49 in the graft union (Figure 2.2B, pattern 5) and 136 in the scion (Figure 2.23, patterns
6-9). Patterns 3, 4 and 6-8 were differentially expressed in the graft union, but the genes
originate in the rootstock or the scion and are categorized as such. The 249 TDFs were
cloned into a plasmid vector to obtain pure fragments. One or more clones were obtained

for each TDF resulting in 646 clones for all the TDFs.

Confirmation of cDNA-AF LP with the use of microarrays

Microarrays were used to conﬁrm the expression proﬁles of cloned TDFs. Sixty-
nine constitutively expressed TDFs from the shoot and the graft union, and 48 ESTs from
the shoot apical meristem, also were cloned and printed on the arrays. In total 1040 DNA
samples (277 clones from shoot, 646 clones from graft union, 48 ESTs from the SAM
region and 69 control sequences form the shoot and the graft union) were printed in
triplicate on the arrays. Two experiments were conducted, which tested the gene

expression changes in the main shoot and the graft union region, respectively. The

92

experimental design for both microarray experiments was in accordance to the cDNA-
AF LP experiments (see Materials and Methods). All 3120 spots printed on the
microarray were included in the data analysis for both experiments with the aim to
maximize the output of differentially expressed genes and identify genes differentially
expressed at both locations in the plant.

Of the 1040 clones printed on the microarray, only 99 showed statistically
signiﬁcant differences following microarray analysis: 43 in the shoot and 56 in the graft
union experiments (Tables 2.1 and 2.2). Of those, 6 clones were only expressed in
‘Bing’/Gi5 trees in both experiments and these belong to the Cherry Virus A (CVA)
RNA genome (Table 2.1, Chapter 3). CVA is a capillovirus that has no deﬁned
symptoms in cherry trees (J elkman, 1995). No other TDFs showed an expression pattern
similar to that of the virus, indicating reduced or no effect on the trees by the presence of
the virus. The other sequences exhibiting signiﬁcant differences in microarray
expression could serve as primary candidates for the promotion of dwarﬁng, since they
show a differential response between the two rootstocks.

The differentially expressed genes were clustered according to their pattern of
expression in 6 clusters. The majority of the genes differentially expressed in the main
shoot fell into two clusters, Clusters 1 and 3 (Table 2.1). The genes in Cluster 1 were
expressed higher in ‘Bing’/Gi5 compared to ‘Bing’/Gi6 trees on 3 July, when the rate of
growth for the ﬁrst graft combination was decreasing. These genes are similar to a
phospholipaseD (RGUS1010), catalasel (RGUS1271-72), MYB protein (RGUS1306), a
subtilisin serine protease (RGU81529-31), a touch induced protein (RGUSl636) and an

unknown sequence (RGUS1342) (Table 2.1). In Cluster 3, genes are expressed higher in

93

‘Bing’/GI6 shoots on 3 July than in ‘Bing’/Gi5. Genes in Cluster 3 include an AP2
domain containing protein (RGUS1210), an Armadillo beta-catenin domain protein
(RGUS 1367), an AtVOZl-like transcription factor (SM1045) and a zinc ﬁnger protein
(SM1026) (Table 2.1). In the same cluster, two genes represented by four clones
(RGUS1121, RGUS1124, SM1063, SM1064) are putatively involved in senescence.
Cluster 2 includes four genes that were down-regulated in ‘Bing’/Gi5 compared to

‘Bing’/Gi6 shoots.

94

Table 2.1: Genes differentially expressed at the upper shoot microarray experiment.
Positive fold change values refer to BS up-regulation while negative values to B6 up-
regulation. 3/6: 3 June, 20/6: 20 June, 3/7: 3 July, B5: ‘Bing’/Gi5, B6: ‘Bing’/Gi6, R:
Rootstock, S: Scion, GU: Graft Union, Sh: Shoot, ns: non-sirgiﬁcant.

 

 

 

 

 

 

 

 

TDF TDF TO(3) Geanfmk BSyBiZmlgggofoBigzng Annomion
. accessron u-

lD( l) srze( 2) number 6 /3 6/20 7/3 ster Product PubMed ID E-value
RGUS 1010 171 GU DV17538S ns ns 1.50 l Phospholipase D At3g15730 5E-18
RGUSlZ'Il 130 GU DV 175414 ns ns 1.56 l Catalase l Atlg20630 lE-l l
RGUSlZ72 129 GU DV175415 ns ns 1.65 l Catalase l Atlg20630 5E-12
RGUSl306 316 S DV175422 ns 1.11 1.58 l MYB transcription factor At5g49330 315-23
RGUSl342 200 S DV 175426 ns 1.29 1.60 1 No hit
RGUS 1636 296 R DV175451 ns 1.33 1.52 l Putative calcium binding protein XP_550050 lE—09
RGU51529 251 S DV175439 ns -l.l7 1.50 l Subtilisin-like protease AAQS4525 213-23
RGUS 1530 285 S DV175440 ns -1 .22 1.70 l Subtilisin-like protease AAQS4525 7E-22
RGUSlS3l 319 S DV175441 ns ns 1.58 l Subtilisin-like protease AAQS4525 713-24
RGUSISI l 213 S DV 175437 ns ns - l .59 2 Nucleic acid binding/ pancreatic ribonuclease Ath67210 0.001
RGUS1589 157 GU DV175447 ns ns -1 .57 2 UDP-D-glucuronate 4-epimerase GAEl At4g30440 3E-l7
RGUSl6l6 180 S DV175449 ns ns -l.51 2 No hit
RGUS1619 132 S DV175450 ns ns - l .54 2 Delta 8-sphingghpid desaturase AAG43277 6E-63
RGUS1084 155 R DV 175391 ns -l.27 -l.63 3 No hit
RGUSl 1 17 202 R DV 175394 ns - l .97 - l .79 3 Alcohol dehydrogenase (ADH) At l 377 l 20 313-05
RGUSl 121 251 GU DV175395 ns -2.58 -2.16 3 Putative replication factor C 36kDa subunit XP_468050 2E-22
RGUSl 124 305 GU DV 175396 ns -2.03 -2. l 3 3 Putative senescence-usociated protein AAR25995 lE-44
RGUS1127 21 l GU DV175397 ns -l.76 ns 3 Putative ATP-dependent RNA helimse A At2g47680 lE-09
RGUSlZlO 140 GU DV175401 ns -l.7l -l.42 3 AP2 domain containing protein RAP2.12 AAC24587 2E-74
RGUS 1230 202 S DV 175405 ns -1 .l l - l .69 3 Hemoglobin CAA68405 3E- l 6
RGUS 1231 202 S DV175406 ns -1 .30 - l .55 3 Hemoglobin CAA68405 3E- l 6
RGUSl321 170 R DV175424 ns -1 .31 -l .51 3 No hit
RGUS1367 177 GU DV 175429 ns -1.53 ns 3 Annadillo/beta-catenin repeat U-box protein At3g07360 5E-l6
RGUS1478 234 S DVI7S435 ns -2.24 -l.80 3 Cytochrome P450 like_TBP BAA10929 2E-28
RGUSlS63 230 GU DV175446 ns -1.87 -l.57 3 Polyprotein -related Atlg21945 3E-20
SM 1045 391 Sh DV1754S7 ns -1 .34 -l .56 3 Transcription factor AtVOZl Atl g28520 2E-21
SM 1063 297 Sh DV 175459 ns -l.45 -2.60 3 Putative senescence-associated protein BAB33421 5E-l9
SM 1064 367 Sh DV175460 ns -l.79 -2.46 3 Putative senescence-associated protein BAB33421 2E-l9
SM 1 123 587 Sh DV 175463 ns -1 .97 ns 3 GDSL-motif Iipase/hydrolase AAM64368 lE-23
SM1228 386 Sh DV175468 ns -2.22 -l.88 3 DEAD/DEAR box helicase, putative RHlS At5gl 1170 4E-38
SM 1026 337 Sh DV175454 -l.23 -l.68 -l.49 3 CZFl/ZFARl zinc ﬁnger protein AtZg40140 lE-ll
RGUS 1068 255 8 DV 175389 2.48 ns ns SPX domain-containing protein At5g20150 lE-20
RGUS l 303 252 S DV175420 1.29 -l .49 -l .94 3 RNA polymerase alpha chain ArtthOSS lE-08
SM 1 146 244 Sh DV175466 53.12 32.49 3.63 4 RNA dependent RNA polymerase [CVA] AAL60496 9E-29
SMl 124 245 Sh DV175464 50.34 35.62 5.23 4 RNA dependent RNA polymerase [CVA] AAL60496 2E-28
SM1002 204 Sh DV175453 27.90 28.82 2.46 4 RNA replicase; coat protein [CVA] CAA57896 5E-l7
RGUS l 393 153 R DV175430 13.33 22.69 2.99 4 RNA dependent RNA polymerase [CVA]

RGUS I336 273 GU DV175425 -2.40 ns ns 5 Chloroplast omega-3 desaturase AAM77643 513-35
SMl 155 213 Sh DV175467 -l.80 ns ns 5 Valencene synthase AAX16077 3E-ll
SM 1089 275 Sh DV 175461 - l .40 ns 2.26 6 Sesquiterpene cyclase CAA04773 313-27
SM 1090 275 Sh DV 175462 -1 .36 ns 2.34 6 Sesquiterpene cyclase CAA04773 3E-26

 

1.RGUS: gene obtained in the graft-union cDNA-AF LP experiment, SM: gene obtained in the upper shoot
cDNA-AF LP experiment; 2.TDF size refers to the sequence length in base pairs; 3. Tissue of origin

according to the cDNA-AF LP proﬁle.

95

Table 2.2: Genes differentially expressed in the graft union microarray experiment.
Positive fold change values refer to BS up-regulation; negative values to B6 up-
regulation. R: Rootstock, GU: Graﬁ Union, S: Scion, Sh: Shoot, B5: ‘Bing’/Gi5, B6:

‘Bing’/Gi6, ns: non-signiﬁcant.

 

 

 

 

 

 

 

 

 

TDF TDF To(3) Genebank Microarray fold changeI Annotation
. accession 85/36 B5/B6 BS/B6 C u- Puchd

ID( 1) srze(2) number R GU S ster Product ID E-value
RGU81219 302 S DV175402 ns ns 1.65 1 transferase family protein At1g31490 3E-30
RGUSl305 204 S DV175421 ns ns 1.75 1 pentatricopeptide (PPR) repeat-containing Atl g19720 313-22
RGUS 1 590 302 S DV175448 ns ns 1.60 1 transferase family protein At1g31490 313-30
SM 1057 174 Sh DV175458 ns ns 1.74 1 AP2 domain-containing transcription factor At4g39780 7E-09
RGUSIOOI 254 S DV175382 ns ns -1.51 3 CSLD2 cellulose synthase catalytic subunit-like At5g16910 213-29
RGUS 1 336 273 GU DV175425 ns ns -1.68 3 chloroplast omega-3 desaturase AAM77643 513-35
RGUSl358 171 S DV175427 ns ns -l.63 3 No hit
RGUS 1 399 135 S DV175431 ns ns -1.51 3 No hit
RGUSI449 300 S DV175434 ns ns -1.59 3 COBRA-like protein 7 precursor At4g16120 813-33
RGU81526 210 S DV175438 ns ns -1.53 3 calcium binding protein CAC43238 2E-50
RGUSlS42 96 GU DV175443 ns ns -1.68 3 expressed protein At5g54870 213-58
RGUS l 589 157 GU DV175447 ns ns -1.52 3 UDP-D-glucuronate 4-epimerase GAE] At4g30440 313-17
RGUSI 1 14 267 R DV175393 ns 1.53 ns 5 tcrpenc synthase/cyclase family At3g29190 1E-12
RGUSI 136 272 S DV175398 ns 1.75 ns 5 sesquiterpengyclase CAA04773 915-30
RGUSI 143 223 S DV175399 1.92 1.73 1.41 4 No hit
RGUSIO67 197 R DV175388 1.98 1.56 1.78 4 803 beta-1,3-g1ucanase At3g57240 513-20
RGUSlO8O 164 R DV175390 1.64 ns 1.43 4 glycosyl hydrolase family 17 At4g16260 713-07
RGU81278 270 S DV1754|7 1.59 ns 1.44 4 calreticulin 3 AAQ19995 2E-32
RGUS 1279 201 S DV 175418 1.62 ns 1.30 4 calreticulin 3 AAQ19995 213-32
RGUSl636 296 R DV175451 1.49 ns 1.81 4 putative calcium binding protein XP_550050 113-09
RGUSIS34 142 R DV175442 1.18 ns 1.69 4 CHIBI Acidic endochitinase At5g24090 0.0002
RGU81234 327 S DV175407 1.61 1.47 ns 4 alpha 1,4-g1ucan phosphorylase L isozyme AAK15695 4E-41
RGU81235 327 S DV175408 1.61 1.46 ns 4 alpha 1,4-g1ucan phosphorylase L isozyme AAK15695 413-41
RGUSlOO7 172 R DV175383 1.60 ns ns 4 subtilisin-like serine protease At3g14240 2E-07
RGUS 1009 170 GU DV175384 1.52 ns ns 4 No hit
RGU81022 289 GU DV175387 1.61 ns ns 4 zinc ﬁnger protein 291 NP_065894 2E-21
RGUSI 108 302 GU DV175392 1.50 ns ns 4 putative serine threonine kinase C1PK9 AAL85889 715-29
RGUSI 166 436 S DV175400 1.54 ns ns 4 BRll-associated receptor kinase 1 (BAKI) AAK68074 6E-26
RGU81223 140 GU DV175403 1.52 ns ns 4 C3HC4-type zinc ﬁnger protein (RING ﬁnger) Atlg69330 615-15
RGU81226 141 GU DV175404 1.51 ns ns 4 RING zinc ﬁnger protein, putative ABF94597 6E-14
RGU81239 449 S DV175409 1.72 ns ns 4 DNA-binding bromodomain-containing protein BAA97526 2E-32
RGU81252 273 GU DV175410 1.92 ns ns 4 heat shock protein hsc70-1 (hsp70-1) At5g02500 3E-40
RGU81276 426 S DV175416 1.62 ns ns 4 putative nodulin-like protein BAD34364 3E-43
RGUS 1 544 94 GU DV175444 1.60 ns ns 4 expressed protein At5g54870 515-06
SM 1039 369 Sh DV175455 1.68 ns ns 4 putative heat shock protein 90 At5g56000 913-51
SM1040 369 Sh DV 175456 1.59 ns ns 4 putative heat shock protein 90 At5g56000 9E-51
SM1126 287 Sh DV175465 1.57 ns ns 4 t-complex polypeptide 1 BAC22124 213-37
RGUS 1 362 381 R DV175428 1.81 ns -1.44 4 centromere protein At3g_22790 2E-24
RGUSIOIB 256 S DV175386 -1.42 ns -1.71 6 putative regulator of gene silencing XP_463755 315-15
RGUSI418 329 S DV175432 -1.51 ns -1.57 6 chorismate mutase. cytosolic (CM2) At5g10870 4E-31
RGUS 1419 363 S DV175433 -1.59 ns -1.62 6 putative chorismate mutase CM2 At5g10870 315-31
RGU81282 191 R DV175419 -1.61 1.07 ns 6 cinnamate 4-hydroxy1ase CYP73 AAF66065 113-12
RGUSISO6 207 R DV175436 -1.62 ns ns 6 cinnamate 4-hydroxylase CY P73 AAF66065 113-12
RGUS 1 679 138 R DV175452 -1.52 ns ns 6 Ttrans-cinnamate 4-monooxygenase (C411) At2g30490 0.009
RGUS 1 342 200 S DV175426 -1.66 -l.19 ns 6 No hit
SM1312 322 Sh DV175469 -1.36 -1.52 ns 6 xyloglucan endotransglycosylase, putative At4g03210 98-“
SM 1 348 386 Sh DV175470 -1.44 -1.64 ns 6 No hit
RGUSl316 365 GU DV175423 -1.57 ns 1.58 6 weak similarity to CTD phosphatase-like 3 At5g58000 7E-12
RGUS 1545 187 S DV175445 - 1.22 ns 1.55 6 inwardly rectifying potassium channel subunit CAGZ7094 313-05
RGUSIZ64 401 R DV17541 1 -1.98 ~2.37 —1.90 7 No hit
RGU81265 401 R DV175412 -1.72 -2.21 -1.71 7 No hit
RGU81266 402 R DV175413 -1.88 -2.28 -1.76 7 No hit
RGU81530 285 S DV175440 -1.31 -1.56 ns 7 subtilisin-like protease AAQ54525 713-22
1. RGUS: gene obtained in the graft-union cDNA-AF LP experiment, SM: gene obtained in the upper shoot

cDNA-AF LP experiment; 2. TDF size refers to the sequence length in base pairs; 3. Tissue of origin

according to the cDNA-AFLP proﬁle

96

The differentially expressed genes in the graft union experiment fell into 7
clusters (Table 2.2). Cluster 1 includes 3 genes up-regulated in the scion of ‘Bing’/Gi5
trees compared to ‘Bing’/Gi6, which encode for a transferase (RGUSl219, RGUSlS90)
and the other two for a lectin (RGUS 1 305) and an AP2-domain protein (SM1057) (Table
2.2). In contrast, Cluster 3 is formed by 10 genes that are up-regulated in the scion of
‘Bing’/Gi6 compared to that of ‘Bing’/Gi5. A COBRA-like 7 protein and a calcium
binding protein are among these genes. Cluster 4 contains the largest number of clones
(24), which represent genes expressed higher in ‘Bing’/Gi5 than ‘Bing’/Gi6 rootstocks.
Many of the genes are involved in post-translational modiﬁcation such as kinases,
molecular chaperones and proteases (Table 2.2). The most interesting of the kinases is a
BRIl-associated receptor kinasel (BAK1,RGUS1166) involved in the perception of
brassinosteroids (BRs) through interaction with the BRIl receptor (Li et al. 2002; Nam
and Li, 2002; Russinova et al. 2004). RGU81252 is also interesting since it encodes for
HSC70-1 a molecular chaperone with the ability to move non-cell autonomously within
the phloem sieve elements and may act in long distance signaling (Aoki et al. 2002).
Cluster 6 is comprised of 12 clones and represents ‘Bing’/Gi6 genes expressed at higher
levels in the rootstock than in ‘Bing’/Gi5. The majority of the genes are homologous to
cell wall formation related genes, while one gene is similar to a putative regulator of gene
silencing from rice (RGUSlOl8).

Four transcription factors were identiﬁed in the shoot microarray experiment: a
putative homolog to an AtVOZl protein from Arabidopsis (SM1045), an AP2-domain

containing protein (RGUSlZlO), a MYB-domain protein (RGUSl306) and a zinc-ﬁnger

97

containing protein (SM1026). AtVOZl is a transcription factor in Arabidopsis with only
one homolog called AtVOZZ (Mitsuda et al. 2004). AtVOZl is expressed speciﬁcally in
the phloem (Mitsuda et a1. 2004), but its function remains unknown. The AP2-domain
containing protein is similar to an ethylene responsive transcription factor (ERF/AP2)
with unknown function that belongs to the B2 family of ERF/AP2 factors (Nakano et al.
2006). The MYB-domain homolog of Arabidopsis is called MYB111 and belongs to the
R2R3-MYB subfamily, but its function remains unknown (Stracke et al. 2001). A
phylogenetically close homolog of MYBl 1 1, MYB12 is involved in the transcriptional
control of ﬂavonoid biosynthesis genes (Mehrtens et al. 2005). Interestingly, most of the
ﬂavonoid biosynthesis related genes were identiﬁed in the graft union region and not in
the shoot (Tables 2.1 and 2.2). Except for RGUSl306, the other three genes are expressed
higher in ‘Bing’/Gi6 on 20 June and 3 July, due to the reduction in ‘Bing’/Gi5
expression. In contrast to the shoot, only one transcription factor was differentially
expressed between ‘Bing’/Gi5 and ‘Bing’/Gi6, in the graft union. It is similar to an AP2-
domain protein (SM1057), but different from that identiﬁed in the scion (Tables 2.1 and
2.2). This AP2-domain protein also belongs to the ERF/AP2 family of transcription
factors and falls into subfamily A6. It is not known if these two proteins function in the
same process.

Finally, ﬁve TDFs were expressed differentially both at the shoot and the graft
union region. The most interesting of those genes is a touch-induced protein with a
calmodulin calcium-binding domain (RGUSl636), because it may function in the

brassinosteroid signaling pathway. The other four genes are a subtilisin-like serine

98

protease (RGUSlS30), a nucleotide sugar epimerase (RGUSIS89), a chloroplast omega-3

desaturase (RGUSl336) and a sequence with no homology to known genes (RGUSl342).

DISCUSSION

Differences in phenotype between graft combinations are accompanied by genetic
changes

Shoot elongation measurements revealed a mid-growing season differentiation in
shoot elongation rate between dwarﬁng and semi-vigorous trees. Such changes in grth
were expected to be the result of signaling events occurring in the tree system. These
events involve an array of proteins whose levels or activity is modiﬁed to produce the
observed changes. The level of some proteins is controlled transcriptionally or post-
transcriptionally and can be screened with high-throughput methods such as cDNA-
AF LP. The collection of samples before, during and after the divergence in shoot grth
between ‘Bing’/Gi5 and ‘Bing’/Gi6 trees enabled the identiﬁcation of genes that were
differentially modiﬁed between the two grafts and possibly contributed to the changes in
shoot growth. It was expected that cessation of growth is followed by other physiological
processes, such as flower primordia formation, reserve storage and dormancy. Thus genes
differentially expressed on 3 or 20 June, before or at the onset of shoot grth cessation,
respectively, are expected to be involved in growth cessation. Genes conforming to this
criterion are represented by patterns 1, 2, 3, 5 and 6 (Figure 2.1B).

cDNA-AFLP is a powerful molecular technique for high throughput screening of

transcriptomes across organisms (Breyne and Zabeau, 2001). Its major advantage is the

99

ability to analyze any RNA sample without previous knowledge of the genome sequence
for the respective organism. Due to its reliance on PCR ampliﬁcation, even low
abundance signals can be recovered and compared. The sensitivity of detection is high
enough to detect temporary RNA molecules, such as miRNA precursors (Prassinos et al.
2005). The proﬁles produced in each run are highly reproducible making it a method of
choice over differential display (Kuhn, 2001). Allelic variation and post-transcriptional
modiﬁcation, such as miRNA cleavage, can be detected with the use of adequate
restriction enzyme combinations. One disadvantage of the technique is the dependence on
the technical skills of the researcher that will allow reliable comparison between samples.
The enzyme combination ApoI/Msel used in this study was selected due to its higher
restriction frequency compared to other six/ four cutter combinations. ApoI has the
potential to identify four restriction sites [(A/G)AATT(T/C)] compared to the popular
restriction enzymes such as EcoRI, HindIII, PstI, BamHI that recognize only one. As a
result, the number of TDFs produced by this combination can reach up to 4 times more
fragments than the enzymes mentioned before. Each sample used in this study returned
approximately 100 TDFs. The primer combinations used were 128 and the samples used
for each primer pair were 6. Thus the total number of TDFs produced per cDNA-AF LP
experiment reached 76,800 (128 x 100 x 6).

Despite the large number of genes screened by cDNA-AF LP, comparison of gene
expression at the main shoot, between dwarﬁng and semi-vigorous trees, revealed a high
degree of co-regulation. Only 111 TDFs were selected as differentially expressed
between the two grafts. Such a low number of differentially expressed genes can be

explained by several observations such as: a) the genotype of the scion is the same

100

(‘Bing’) in both graft combinations, b) the trees were growing under the same
environmental conditions, c) apart from the difference in shoot elongation there was no
other obvious phenotype and (I) shoot length did not differ substantially between graft
combinations indicating a very similar transcriptome. Thus, it is prudent to expect that
these 111 TDFs are involved in shoot growth cessation. Nevertheless, evaluation of
differential expression was performed visually and thus it was necessary to quantify the
level of expression at each sample. Furthermore, cloning of some TDFs returned more
than one gene sequences due to contamination from neighboring fragments. The
microarray technology was selected to solve these problems and simultaneously serve as
a second independent replication of the experiment. In contrast to cDNA-AF LP,
microarrays use sequence homology to detect the level of gene expression. Thus,
microarrays cannot differentiate effectively between alleles, reﬂecting the cumulative
expression of a gene. Since the microarrays in this analysis were constructed using the
cloned TDFs as template, there could be no selection over the quality of the sequence.
Lack of sequence information did not allow ﬁltering of the TDFs on the basis of
sequence conservation. The average size of the TDFs was 200bp, thus highly conserved
genetic regions could cover its complete sequence. Even under stringent hybridization
conditions, cross-hybridization of conserved regions cannot be avoided. Statistical
analysis of the microarray data allowed the identiﬁcation of signiﬁcant differences in
gene expression. Genes with statistically signiﬁcant differences in expression between
rootstocks were subjected ﬁrrther to ﬁltering for fold change in expression levels. A fold
change of 2 proved to be very stringent, since as it is shown in Tables 3 and 4 only 17

TDFs could pass that ﬁlter. Thus genes with fold change higher than 1.5 were selected.

101

This is considered a small change in gene expression, but tagging as statistically
signiﬁcant indicates a consistent difference between samples. That difference in the
concentration of transcripts may be enough to produce the observed phenotypic change.
For these reasons only 43 gene clones were proven to be differentially expressed. Since
the microarray included genes from the shoot and the graft union regions, some of the
differentially expressed genes were originally identiﬁed in the graft union cDNA-AF LP

experiment.

The graft union region is transcriptionaly diverse between graft combinations at the
time of differentiation in shoot growth
The graft union is the bridge between the rootstock and the scion, where tissues

from both genotypes combine to produce a functional tree. Swelling at the graft union
denotes an interaction between rootstock and scion, a phenomenon inversely proportional
to rootstock vigor (Wagner and Gruppe, 1985). Identiﬁcation of the interaction effect at
the genetic level was performed by cDNA-AF LP. Analysis at the shoot apex indicated
the 20th of June as the most indicative time for differential gene expression between the
two graft combinations. This date was selected as the most appropriate for the graft
union analysis. Due to the common origin of the two rootstocks, it was expected that they
share the same genes, but at the same time they exhibit allelic variation.

cDNA-AF LP analysis at the graft union revealed the expected variation in TDF
proﬁles as shown in Figure 2.2. The majority of the differentially expressed TDFs were
identiﬁed in the scion area bordering the graft union. This is unexpected since the scion

tissue for both graft combinations used in the experiment have the same genetic

102

background (‘Bing’). Only 64 TDFs were differentially expressed in the rootstock, which
would have been expected to show a higher degree of differential expression than the
scion. Furthermore, TDFs with patterns similar to pattern 1 (Figure 2.2) were the result of
allelic variation and not due to a difference in the transcriptional levels as it was proven
by the microarray analysis. Nevertheless, allelic variation can affect the function of the
gene at the translational level. Enzymes or signaling proteins with altered sequence in one
genotype may have different afﬁnity for their substrate or other interacting proteins. This
possibility was not tested however. Gene expression at the graft union proved to be even
more complex. The graft union combines cells from both rootstock and scion that interact
at the point of union formation allowing interaction between cells and concomitantly the
rootstock and the scion. TDFs with patterns such as pattern 15 (Figure 2.2), were initially
believed to be ‘Bing’/Gi6 graft union speciﬁc. Microarray analysis proved that these
genes were expressed in both graft combinations without any sign of differential
expression. The reason for this anomaly in the cDNA-AF LP proﬁle is not known. Except
for 2 TDFs expressed in ‘Bing’/G15, all others were speciﬁcally expressed in ‘Bing’/G16
graft unions according to the cDNA-AF LP. This observation excludes the effect of allelic
variation in restriction digestion. In that case, we would expect a similar number of
differentially expressed bands between the two graft combinations. Furthermore,
sequencing of some of these TDFs excluded contamination by other organisms or
genomic DNA, due to high similarity to other Rosaceae cDNA sequences available in
public databases and the methods used to isolate and purify the RNA. Nevertheless, the
inability of these genes to pass the ﬁlters set for the microarray experiment did not allow

any further examination of their behavior in the cDNA-AFLP experiment.

103

Parallel gene regulatory pathways between cherry and apple graft combinations

An interesting gene (RGUSSl 166) encodes a protein similar to the Arabidopsis
BRASSINOSTEROID IN SENSITIVE l-associated receptor kinase 1 (BAKI/SERKB).
This gene also has been found to be up-regulated in dwarﬁng apple grafts, but at the stem
region (Jensen et al. 2003). The parallel differential expression of BAKl in these two
dwarﬁng systems implies an important role for brassinosteroids (BRs) in the control of
tree growth. Other genes found in both studies include an AP2 domain-containing protein
(RGUSlZ 10), a GDSL—motif lipase hydrolase (SM1123), a touch-induced protein
(RGUSl636) and a C3HC4 zinc-ﬁnger protein (RGUSl223), indicating an across-species

response mechanism to the dwarﬁng phenomenon.

Genes implicated in brassinosteroid response

Three independent microarray studies on the effect of BRs in Arabidopsis gene
expression revealed a signiﬁcant number of affected genes (Miissig et al. 2002; Yin et al.
2002; Goda et al. 2004). Interestingly, genes with similar annotation are between those
differentially expressed in the current study. The cherry orthologs of the Arabidopsis
genes affected by BRs are a touch-induced protein (RGUSl636), the subtilisin serine
protease (RGUSIS29-153 l) and a beta-1,3-glucanase (RGUSlO67). Together with the
homolog of BAKl (RGUSI 166), this makes a signiﬁcant number of differentially
expressed BR related genes and requires special attention. In the shoot, RGUSl636 and
RGU81529-31 are up-regulated in ‘Bing’/Gi5 on the 3rd of July, compared to ‘Bing/Gi6.
RGUSl636 also is up-regulated in the rootstock and scion of ‘Bing’/Gi5 trees, together

with RGUSlO67. RGUSI 166, the homolog of BAKl, is up-regulated in the rootstock of

104

‘Bing’/Gi5 trees. Only RGUSlS30 shows up-regulation in the rootstock and graft union
of ‘Bing’/Gi6 trees. This is contradictory to the role of BRs in the control of grth
since, in this study, genes responsive to BRs are up-regulated in the dwarf ‘Bing’/Gi5
trees rather than the semi-vigorous ‘Bing’/Gi6. It is in agreement, though, with the
ﬁndings in dwarf apple trees (Jensen et al. 2003). Recently, it was shown that the
epidermis plays a crucial role in the promotion of growth in Arabidopsis (Savaldi-
Goldstein et al. 2007). This was shown in response to local activation of BR signaling,
indicating the important role of BRs in the control of growth. A more in-depth analysis of
the response of various rootstocks to brassinolide treatment or internal concentration of

BRs is necessary to clarify their role in the dwarﬁng phenomenon.

Conclusions

The parallel analysis of gene expression at the shoot and the graft union region
allowed a more complete coverage of the changes in the biology of the two graft
combinations. We have shown that shoots of the common genotype ‘Bing’ respond
differently when grafted on two rootstocks of different vigor. The dwarﬁng rootstock Gi5
caused earlier changes in gene expression of the ‘Bing’ shoot, compared to the non-
dwarﬁng rootstock Gi6. The graft union region experiment revealed much more diverse
gene expression in the rootstock (Table 2.2), which is expected to be the cause of
differential shoot grth in the scion. Additionally, as shown in Figure 2.2B patterns 9
and 10, several TDFs exhibited a differential expression across the graft union, which
implies transport of the signals between rootstock and scion. It is not known, however,

whether these genes are transported through the vascular system or if they are expressed

105

in the trunk cells. These observations lead to the formulation of a fourth hypothesis that
“rootstock-induced dwarﬁng is caused by rootstock encoded signals able to move
through the graft union and affect scion growth”. In support of this hypothesis, recent
studies have shown that there are a signiﬁcant number of mobile macromolecules that
can move through the vascular system (Kim et a1. 2001; Mallory et al. 2003; Ding et al.
2003; Lucas and Lee, 2004). Signaling proteins or RNA can move through the graft
union and small differences in the receptors/targets of those signals can have negative or
positive effects on growth (Kim et al. 2001; Mallory et al. 2003).

Rootstock-induced dwarﬁng remains a complex and poorly understood
phenomenon. Physiological data have provided much information on the changes
occurring in the grafted trees, especially at the vicinity of the graft union. Clues provided
by these studies also point to the idea that all dwarﬁng systems do not function by the
same mechanism. The genotypic differences of rootstocks that exert varying levels of
vigor to the same scion variety prompted us to study the response of genes in both the
scion and the rootstock. We have successfully focused on the transition stage during the
growing season when gene expression differentiates between dwarﬁng and vigorous
rootstocks and studied both the biology of the shoot and the graft union area. Further
characterization of the identiﬁed genes should eventually lead to the confounding signals
that are responsible for the determination of the transitions in shoot growth. Control in
the production of these signals could eventually lead in the control of tree grth

depending upon the desirable attributes of the exploited trees.

106

LITERATURE CITED

Aoki, K., Kragler, F ., Xoconostle-Cazares, B., Lucas, W.J. (2002) A subclass of plant
heat shock cognate 70 chaperones carries a motif that facilitates trafﬁcking through

plasmodesmata. Proceedings of the National Academy of Science, USA 99: 16342-
16347

Bachem, C. W. B., Van der Hoeven, R. S., de Bruijin, S. M., Vreugdenhil, D., Zabeau,
M. and Visser, R. G. F. (1996) Visualisation of differential gene expression using a
novel method of RNA ﬁngerprinting based on AF LP: analysis of gene expression
during tuber development. Plant Journal 9: 745-753

Bortiri, E., Oh, S.-H., Jiang, J ., Baggett, S., Granger, A., Weeks, C., Buckingham, M.,
Potter, D. and Parﬁtt, DE. (2001) Phylogeny and systematics of Prunus (Rosaceae)
as determined by sequence analysis of ITS and the chloroplast trnL-trnF Spacer

DNA. Systematic Botany 26: 797-807

Breyne, P. and Zabeau, M. (2001) Genome-wide expression of plant cell cycle modulated
genes. Current Opinion in Plant Biology 4: 136-142

Churchill, GA. (2004) Using ANOVA to analyze microarray data. BioTechniques 37:
173-177

Dilks, D.W., Ring, R.H., Khawaja, X.Z., Novak, T.J., Aston, C. (2003) High-throughput
conﬁrmation of differential display PCR results using reverse Northern blotting.
Journal of Neuroscience Methods 123: 47-54

Ding, B., Itaya, A. and Qi, Y. (2003) Symplasmic protein and RNA trafﬁc: regulatory
points and regulatory factors. Current Opinion in Plant Biology 6: 596-602

F ranken-Bembenek, S. (1996) The Giessen cherry rootstocks. Compact Fruit Tree 29: 19-
56

Gruppe, W. (1985) Size control in sweet cherry cultivars (Prunus avium) induced by
rootstocks from interspeciﬁc crosses and open pollinated Prunus species. Acta
Horticulturae 169: 209-218

107

Hegde, P., Qi, R., Abernathy, K., Gay, C., Dharap, S., Gaspard, R., Hughes, J .E.,
Snesrud, E., Lee, N., Quackenbush, J. (2000). A concise guide to cDNA microarray
analysis. Biotechniques 29: 548-562

Jelkmann, W. (1995) Cherry Virus A: cDNA cloning of dsRNA, nucleotide sequence
analysis and serology reveal a new plant capillovirus in sweet cherry. Journal of.
General Virology 76: 2015-2024

Jensen, P.J., Rytter, J ., Detwiler, E.A., Travis, J .W. and McNeils, T.W. (2003) Rootstock
effects on gene expression in apple tree scions. Plant Molecular Biology 53: 493-511

Kim, M., Canio, W., Kessler, S. and Sinha, N. (2001) Developmental changes due to
long-distance movement of a homeobox fusion transcript in tomato. Science 293:
287-289

Kuhn, E. (2001) From library screening to microarray technology: strategies to determine
gene expression proﬁles and to identify differentially regulated genes in plants.
Annals of Botany 87: 139-155

Li, J ., Wen, J., Lease, K.A., Doke, J .T., Tax, F .E. and Walker, J .C. (2002) BAKl, an

Arabidopsis LRR receptor-like protein kinase, interacts with BRIl and modulates
brassinosteroid signaling. Cell 1 10: 213-222

Lucas, W.J. and Lee, J.Y. (2004) Plasmodesmata as a supracellular control network in
plants. Nature Reviews in Molecular and Cellular Biology 5: 712-726

Mallory, A.C., Mlotshwa, S., Bowman, L.H. and Vance, VB. (2003) The capacity of
transgenic tobacco to send a systemic RNA silencing signal depends on the nature of
the inducing transgene locus. Plant Journal 35: 82-92

Mehrtens, F., Kranz, H., Bednarek, P., Weisshaar, B. (2005). The Arabidopsis
transcription factor MYB12 is a ﬁavonol-speciﬁc regulator of phenylpropanoid
biosynthesis. Plant Physiology 138: 1083-1096

108

Mitsuda, N., Hisabori, T., Takeyasu, K., Sato, M.H. (2004) V02; isolation and
characterization of novel vascular plant transcription factors with a one-zinc ﬁnger
from Arabidopsis thaliana. Plant Cell Physiology 45: 845-854

Nakano, T., Suzuki, K., F ujimura, T. and Shinshi, H. (2006) Genome-wide analysis of
the ERF gene family in Arabidopsis and rice. Plant Physiology 140: 411-432

Nam, K.H. and Li, J. (2002) BRIl/BAKl, a receptor kinase pair mediating
brassinosteroid signaling. Cell 110: 203-12

Prassinos, C., Ko, J .-H., Yang, J. and Han, K., -H. (2005) Transcriptome Proﬁling of
Vertical Stem Segments Provides Insights into the Genetic Regulation of Secondary
Growth in Hybrid Aspen Trees. Plant Cell Physiology 46: 1213-1225

Russinova, E., Borst, J.W., Kwaaitaal M, Cano-Delgado, A., Yin, Y., Chory, J. and de
Vries, SC. (2004) Heterodimerization and endocytosis of Arabidopsis brassinosteroid
receptors BRIl and AtSERK3 (BAKl). Plant Cell 16: 3216-3229

Savaldi-Goldstein, S., Peto, C. and Chory, J. (2007) The epidermis both drives and
restricts plant shoot growth. Nature 446: 199-202

Stracke, R., Werber, M. and Weisshaar, B. (2001) The R2R3-MYB gene family in
Arabidopsis thaliana. Current Opinions in Plant Biology 4: 447-56

Veit, B. (2006) Stem cell signalling networks in plants. Plant Molecular Biology 60: 793-
810

Wagner, H. and Gruppe, W. (1985) Swelling of scion trunk above graft unions of cherry
trees. Acta Horticulturae 169: 269-274

Wang, S.X., Hunter, W. and Plant, A. (2000) Isolation and puriﬁcation of ﬁrnctional total
RNA from woody branches and needles of Sitka and white spruce. Biotechniques 28:
292—296

Williams, L. and Fletcher, J.C. (2005) Stem cell regulation in the Arabidopsis shoot
apical meristem. Current Opinion in Plant Biology 8: 582-586

109

 

Wu, H., Kerr, MK. and Churchill, GA. (2002) MAANOVA: a software package for the

analysis of spotted cDNA microarray experiments. The analysis of gene expression
data: methods and software. Springer, New York

110

 

CHAPTER 3

CHERRY VIRUS A HAS NO DIRECT EFFECT ON ROOTSTOCK-INDUCED

DWARF ING OF GRAF TED CHERRY TREES

lll

 

INTRODUCTION

Plant viruses can cause signiﬁcant losses in agricultural production. Their impact
is usually in the quality of the product, which is usually unacceptable for the market, but
also in the quantity of the product, which is reduced. Protection of crops from viruses is
usually performed through the distribution of certiﬁed material that is virus-free, since
there are no products in the market for virus eradication. Virus-free material can be
produced by tissue culture or propagation of uninfected stocks. For the detection of
viruses in crops many tests are commercially available and are usually based on antibody
detection, such as the Enzyme-Linked ImmunoSorbent Assay (ELISA). However, tests
are not available for all viruses. Development of a test depends on the importance and
severity of infection by the virus on various crops. Fruit crops are prone to virus
infections due to their longevity, clonal propagation and attraction of sap feeding insects
that can transmit viruses through their proboscis. Previously reported viruses in cherry
are the Prune Dwarf Virus (PDV), Prunus Necrotic Ringspot Virus (PNRSV), Little
Cherry Virus-1 (LChV-l), Little Cherry Virus-2 (LChV-2), Cherry Necrotic Rusty
Mottle Virus (CNRMV), Cherry Mottle Leaf Virus (CMLV), Cherry Rasp Leaf Virus
(CRLV), Cherry Virus A (CVA), and Cherry Green Ring Mottle Virus (CGRMV)
(Jelkmann, 1995; Lang and Howell, 2001; Isogai et al. 2005). As the name of PDV
suggests, the virus reduces the size of the infected trees, but not in cherries (VIDE web
site). Nevertheless, the ability to produce dwarf cherry trees may exist in other viruses.

Plants in contrast to mammals do not produce antibodies against viruses. They
have developed a system of suppression of the viral genomic RNA called silencing. In

this system, double stranded RNA molecules are recognized and cleaved by internal plant

112

 

enzymes that belong to the Dicer-like family (Wang and Metzlaff, 2005). Cleavage
results in small double stranded RNAs of 21 and 24 nucleotides in size (Lecelier and
Voinnet, 2004). These small dsRNA molecules are perceived by the ‘RNA induced
silencing complex’ (RISC), which will then convert them into single stranded RNAs and
use them as templates for the detection and cleavage of more viral RNAs (Lecelier and
Voinnet, 2004). Plant RNA dependent RNA polymerases (RdRPs) are responsible for the
ampliﬁcation of the silencing signal by producing more dsRNAs (Lecelier and Voinnet,
2004). Small RNAs can also move systemically into the plant vascular system and confer
virus resistance to the rest of the plant body (Yoo et al. 2004). Systemic silencing is
necessary to protect plants from systemically spreading viruses. Some plant viruses can
accomplish long distance trafﬁcking with the use of a movement protein (Nelson and
Citovsky, 2005; Lucas, 2006). Silencing though is not enough to protect plants from viral
infection. Viruses have developed mechanisms to overcome host-speciﬁc resistance. An
example is the ability of the HC-Pro protein to suppress the accumulation of virus
induced siRNAs, thus promoting viral infection (Llave et al. 2000).

Cherry Virus A (CVA) was reported ﬁrst in sweet cherry (Prunus avium L.) by
Jelkmann (1995) in a study aimed at isolating Little Cherry Disease (LCD). The virus
belongs to the genus Capillovirus, which includes the type member Apple Stem
Grooving Virus (ASGV). CVA is a single stranded RNA virus, with a 3’ attached poly-
adenylated tail. Its genome has a size of 7,3 83bp and contains two open reading frames
(ORFs, Figure 3.1). The ﬁrst ORF (ORFl) covers almost the complete sequence of the
virus and produces a sequence of 2,360 amino acids in frame 1. ORFI contains a domain

of unknown function (DUF17l. 7), a RNA dependent RNA polymerase (RdRP), a viral

113

 

 

helicase 1 and a coat protein. The second ORF (ORF 2) is located in the C-terminal region

of the virus and has a sequence of 463 amino acids in frame 3. It contains the movement

protein of the virus.

1 1-4.

 

 

A32 £553: «22. a p.20 23 $83 a Ego 888838 $2 828% «id ”.232 5883
E50523 .«o :6an ”SQ .NoEEm wﬁeaom con—O ”N20 .3895 wEeBM EEO ”EMO .38“ 805 gone
contact me 2:895 05 .«o 5:63 2: 30:3. mecca 322% .< mag 35:0 mo 5331890 08050 ”mm oeswi

 

 

 

 

 

 

i
mum—O
Ewan: “COE®>O<‘

i] 4.11 i :1 1. - 1 r l
Flmo E 1:11 I E l . r

590.5 500 amum _ owwo=mI .95 NE...

n50
_ _ _ _ _ _ _ _

oooo ocov Doom o

115

Since 1995, the virus has been detected in Germany, Canada, the United
Kingdom, France and Japan (Eastwell and Bemardy, 1998; Foissac et al. 2001; Isogai et
al. 2004; James and Jelkmann, 1998; Jelkmann, 1995; Kirby et al. 2001). No symptoms
on sweet cherry trees or fruits have been linked to the virus, which explains its relatively
new and accidental discovery (Jelkmann, 1995; Eastwell and Bemardy, 1998). CVA does
not exhibit a synergistic effect in the presence of LCD and thus does not amplify the
symptoms associated with LCD (Eastwell and Bemardy, 1998).

Here we report the discovery of CVA in sweet cherry cultivars grafted on
interspeciﬁc hybrid rootstocks in Michigan, United States. The virus was identiﬁed in a
complementary DNA-Ampliﬁed Fragment Length Polymorphism (cDNA-AF LP) screen
between dwarﬁng and non-dwarﬁng trees, aimed at identifying genes involved in the
phenomenon of rootstock induced dwarﬁng. Genes showing differential expression
between the two graft combinations were of primary interest. Nine cDNA-AF LP
fragments that were present only in the dwarf trees aligned to various regions of the 7,383
bp CVA genome, conﬁrming that they originated from CVA RNA. We tested whether

the presence of CVA is linked to dwarﬁsm induced by rootstocks of varying vigor.

MATERIALS AND METHODS

Plant material
Tissue samples were harvested for the cDNA-AFLP experiment from orchard-
grown shoots and trunk of two-year-old ‘Bing’ sweet cherry on GiSelAS and GiSelA6

(both clonal rootstocks derived from hybridization of Prunus cerasus L. x Prunus

116

 

canescens Bois.) (Horticulture Experiment Station, Clarksville, Michigan). Shoots were
collected on 3 June, 20 June and 3 July 2002, while trunk samples of the rootstock and
the scion of the same trees were collected on 20 June 2002. For the additional screening
tests, shoot samples were collected in June 2003 from: 1) shoots of ‘Hedelﬁngen’ sweet
cherry grafted on the clonal rootstocks GiSelA3 (P. cerasus x P. canescens), GiSelA5,
GiSelA6, Edabriz (P. cerasus), Gi195/20 (P. canescens x P. cerasus), Weiroot10 and
Weiroot158 (both P. cerasus), as well as on the seedling rootstocks Mazzard (P. avium),
Mahaleb (Prunus mahaleb L.) and Erdi V (P. mahaleb) (Northwest Michigan
Horticultural Experiment Station, Traverse City, Michigan); 2) shoots of ‘Bing’,
‘Hudson’, and ‘Attika’ sweet cherry grafted on GiSelA5 and GiSelA6 rootstocks
(Horticulture Experiment Station, Clarksville, Michigan); and 3) shoots of ‘Sam’/
GiSelA5, ‘Brooks’/ GiSelA5 and ‘Bing’/Edabriz trees (Horticulture Experiment Station,
Clarksville, Michigan). Samples consisted of the upper 10 cm of the shoot without the
leaves. One to three shoots were collected from each of three to four trees for every graft
combination. Samples harvested in the orchard were placed in coolers ﬁlled with dry ice,
transported to the laboratory, and ground to a ﬁne powder in liquid nitrogen. Shoot
samples were ground with a mortar and pestle while trunk samples from the rootstock
and the scion were ground in a 1 liter stainless steel commercial blender (Waring,
Torrington, Connecticut). All of the samples were maintained frozen in liquid nitrogen
during grinding (samples were frozen in liquid nitrogen, but were dry when the blender

was operated). Ground samples were stored at -800 C until RNA could be extracted.

9117

 

 

RNA extraction

A common total RNA extraction protocol was used throughout these experiments
(Wang et al. 2000). mRNA was isolated for the cDNA-AFLP experiment with
Dynabeads paramagnetic particles (Dynal Biotech, Lake Success, New York) as

described by the manufacturer.

cDNA-AFLP analysis

As described in Chapter 2.

Sequencing

The cDNA-AF LP fragments were PCR ampliﬁed using the Apo-pre and Mse-pre
primers and directly sequenced using ApoI-pre (Molecular Structure Facility, Michigan
State University) as the sequencing primer. Sequencing was performed on an ABI 7700
Sequencer (Genomics Technology Support Facility, Michigan State University). The
sequences were introduced in the BlastN and BlastX search engines of the National
Center for Biotechnology Information (NCBI) for alignment to known gene or protein

sequences, respectively.

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)

RT -PCR primers were designed based on the initially reported CVA sequence
(X82547) by Jelkmann (1995). The downstream primer was CVA4 (6372): 5’
TCCTTTGAGAATTGCACTTATC 3’, and the upstream primer CVAS (4840): 5’

CGTACAATAAAGGCGATCACC 3’. A brief description of the RT-PCR protocol is as

118

h.

‘1' ' .'

 

 

follows: Total RNA (1 pg) was reverse transcribed using an oligo(dT25)N primer and
SuperscriptII reverse transcriptase, as described by the manufacturer (Life Technologies,
Rockville, Maryland). The reaction was incubated at 42°C for one hour followed by
deactivation of the enzyme at 65°C for 15 min. Ten percent (2 11.1) of the reaction was
used for PCR ampliﬁcation in a 20 ul reaction using 10 pmol of CVA4 and CVAS
primers that produce a 1,532 bp fragment. The reaction conditions were as follows: 1)
940C for two min; 2) 40 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 70 s; 3) 72°C

for 7 min; and 4) storage at 40C. The product was analyzed on a 1% agarose gel.

Growth measurements

Shoot length measurements were taken from trees growing at the Northwest
Michigan Horticultural Research Station at Traverse City, Michigan. All trees had
‘Hedelﬁngen’ as the scion and the rootstocks were Gisela5, Edabriz, Gi195/20, Gisela 6,
Weiroot10, Weiroot158, Mahaleb, Mazzard and Erdi V. Eight trees were used for each
graft combination and three shoots were measured per tree. The model and analysis of
variance were for unequal number of replications and subsamples. The SAS statistical
package was use in the analysis and for multiple comparisons using the proc glm

function.

Northern hybridization analysis

Ten micrograms of total RNA were analyzed on a 1.2% denaturing agarose gel
and then transferred overnight to a nylon membrane Hybond N+ (Amersham
Biosciences, Piscataway, New Jersey) by capillary transfer (Sambrook et al. 1987). The

probe consisted of three cDNA-AFLP fragments (F 1, F6, F8) that represented the 5’ and

119

 

 

the 3’ region of the CVA genome and were labeled with 32P using the Rediprime 11 DNA
labeling kit (Amersham Biosciences, Piscataway, New Jersey). Hybridization was
performed in UltraHyb hybridization buffer (Ambion, Austin, Texas) at 42°C for 18 h.
The membranes were washed twice in 42°C 2xSSC, 0.1% SDS solution for 5 min and
twice in 42°C 0.1xSSC, 0.1% SDS solution for 15 min. After washing and drying,

membranes were exposed to an X-Ray ﬁlm (Kodak, X-OMAT) overnight.

RESULTS

Initial detection of the virus

A study was conducted to identify gene expression changes between dwarﬁng and
non-dwarﬁng scion-rootstock graft combinations in sweet cherry. ‘Bing’/Gi5 and
‘Bing’/Gi6 trees were used as the dwarﬁng and non-dwarﬁng plant materials,
respectively. Screening of shoot and trunk samples by cDNA-AFLP revealed various
types of gene expression differences between dwarﬁng and non-dwarﬁng trees (Chapter
2). Some of these expression patterns showed genes expressed only in ‘Bing’/Gi5 trees
(Figure 3.2A). Sequencing of the fragments representing these patterns revealed a high
degree of similarity to Cherry Virus A. Of the 13 fragments showing this pattern, 9 were
sequenced and aligned to the virus genome. The distribution was balanced across the
viral genome, spanning from the 5’ to the 3’ region of the 7,383bp sequence (Figure
3.28). The average size of the fragments was 157bp, with the smallest being 60bp and the
largest 343bp. The same exact fragments were present in the shoot, the scion trunk, the

graft union and the rootstock trunk (Figure 3.28). Alignment of the translated sequences

120

 

 

revealed a high degree of conservation between the virus identiﬁed in Germany and the
United States, with few non-conserved amino-acid changes (Figure 3.2C). Fragments F6
and F7 align to the ORF2 that codes the movement protein, while the remaining
fragments code for ORFl which contains the other viral proteins. It should be noted that
even though cDNA-AFLP was used as a gene expression screening tool, the
identiﬁcation of CVA was made possible by the polyadenylation and not the transcription
of its genomic RNA. Thus the cDNA-AF LP patterns for CVA reﬂect the concentration of

the viral genome and not viral gene expression.

CVA is not associated with rootstock control of scion vigor

The distinct presence of the virus in scion and rootstock tissues associated with
the dwarﬁng Gi5 rootstock, yet absence in the same tissues associated with the semi-
vigorous Gi6, led to the hypothesis that CVA may play a role in the ability of dwarﬁng
cherry rootstocks to reduce sweet cherry scion vigor. To test this hypothesis, shoots from
‘Hedelﬁngen’ sweet cherry grafted on nine different rootstocks of varying vigor were
screened for presence of the virus. The screening was performed using RT-PCR and
Northern hybridization. The primers designed for the RT-PCR were based on the
published sequence as another proof for the sequence conservation, producing a fragment
of 1,532 bp. Northern hybridization was used as a conﬁrmation method to avoid
problems with miss-priming due to mutations on the virus sequence. The two methods
returned consistent results, with CVA genomic RNA clearly present in the samples from

‘Hedelﬁngen’ on Gi5, but not on Gi6 (Figure 3.3). However, CVA RNA was

121

 

 

Figure 3.2: Cherry Virus A detection. (A) cDNA-AFLP proﬁles across tree sections and
growing season dates for different parts of the CVA genome that was digested with
ApoI/MseI restriction enzymes, R: rootstock, GU: graft union, Sc: scion, B5: ‘Bing’/Gi5,
B6: ‘Bing’/Gi6. (B) The position of each cDNA-AFLP fragment on the 7,383bp CVA
genome is indicated by arrows. Direction of the arrow is from the ApoI to MseI restriction
site. (C) Alignment of the translated cDNA-AF LP fragments with the two open reading
frames (ORF) of CVA. The alignment was obtained from the BlastX results. Amino acid
conservation is indicated as follows; black: identical, gray: conserved, white: non-
conserved. ORF] and ORF2: open reading frames of the CVA genome with GenBank
accession numbers CAA57896 and CAA57897, respectively. Arrrino acid numbering of
the cDNA-AF LP fragments is based on the individual TDF amino acid sequence, while
for CVA ORFl and ORF2 it is based on the complete amino acid sequence. Sorting of

the sequences is based on the CVA ORFs.

122

 

P1 :
ORP1:

P2 r
ORPlr

P3 8
ORPI:

P4 3
ORPlr

P5 r
ORPI:

P8 r
ORPlr

P9 3
ORPlr

P6 r
ORPZ:

P7 3
ORP2:

 

 

 

"r

I
.4-
I

D

 

F4
16 V S 7 1 k K
F6
1“F* UM“ ‘Eﬁ“ ]?7
“his!“ .4. r W W w '4?’ 1'8
I
~ u dune ’ m
0 2000 4000 6000
l I l I l l I AAA
-+-<— <——><h- <- -e’—> ->
I! P! ID P4 I! II P! ID ID

        

HKRSPWSFLSDAKNYIDSWIIQSPFLRRIFPVGSRAITELIRDWIANAESLKIQT
NKRSPWSFLSDAKNY DSWIIQSPFLRRIFPVGSRAITELIRDWIANAESLKTOT

1 TSKRFSGGSYS‘ CRSGLLVDSMRQNTSSSSEVFVDLFPST RP‘ 45
‘75 TSKRFSGGSYS SRMGLLVDSMRQNTS'SﬁEVFVDLFPS RP‘ 519

   

: 62
C 436

 

 

1 KLQiIEENS E KENLKTHLPISYSGL 27
1170 KLOtIEENSIES ENLKTHLPISYSGL 1196

1 PNNFDQRMYDESVSEHEEOKISHN- ”'1 26
p.

1293 DQRMYDESVSEFEEHKISHNA I 1323

69
2178

IYNEIRRGLGNYIWENMIDPRDLLHLTAKPAVEASEGVAATPAITLSENQRAVHNTIF
YNEIRRGLGNYIWENMIDPRDLLHLTAKPAVEASEGVAATPAITLSENORAVKNTTF

     

70 NYYLRIMFG I'VMGTSEQTDYPGEHLAIPRPVIENQEHLTAHLPAGMSLL 120
2179 NYYLRIMFG AVMGTSEQTDYPGEHLAIPRPVIENQE‘LTAHLPAGMSLL 2229

   
     

  
 

1 HGAVSFIYLKNPGAYFNCPAVVFDFNKGLPLTIIKIGKNANAISACNQRLFNREGKKAVFAAQGEVNL 69
2270 NHGAVSFIYLKNPGAYFNCPAVVFDFNKGLPLTIIKIGKNANAISACNQRLFNREGKKAVFAAOGEVNL 2338
1 KKPINGRIVYFDPRFLDKNDACQAGFSFQLQTGSVYYLYRPNYPMSTHDPD 51

104 KKPINGRIVYFDPRFLDKNDACQAGFSFQLQTGS‘ YLYRPNYPMSTHDPN 154

1 R” NPPTVALIDVSVDQSF 63

181 NPPTVALIDVSVDQSF 243

 

123

 

 

Table 3.1: cDNA-AFLP fragments with homology to CVA. Location of each fragment is
given in nucleotides on the CVA genome. Size of each fragment is given in bases.

 

 

 

 

 

 

 

 

 

 

 

Fragment Start Finish Size
F1 1180 1341 161 F3
F2 1475 1598 123 ;1
F3 3223 3371 148 i
F4 3508 3588 80 i
F5 3977 4037 60
F6 5718 5861 143
F7 5975 6118 143
F8 6391 6734 343
F9 6862 7067 205

 

 

 

 

 

 

 

124

only faintly present in ‘Hedelﬁngen’ on the dwarﬁng rootstock Edabriz, yet RNA
concentration was quite strong in ‘Hedelﬁngen’ on Weiroot158, which is similar in vigor
to Gi6 (Table 3.2). The rootstock GB is the most dwarﬁng rootstock among those tested,
but there was no indication of the presence of CVA. These results do not support the

hypothesis that CVA plays a role in the dwarﬁng ability of cherry rootstocks.

The absence of CVA is not linked to rootstock resistance

Based on the cDNA-AF LP, RT-PCR and Northern data, a second hypothesis was
developed to determine whether the absence of the virus from Gi6 is due to resistance. To
test this hypothesis, tissues were examined from four different sweet cherry cultivars
grafted onto Gi5 and Gi6 rootstocks. RT-PCR and Northern hybridization were used as
described above. The virus was present in all scions grafted on G15, while two scions
(‘Hudson’ and ‘Attika’) grafted on Gi6 also were infected (Figure 3.4). ‘Bing’ on Edabriz
which has the same vigor as ‘Bing’/Gi5 was also found to be infected. These data
indicate that the absence or presence of the virus in a particular rootstock or scion cultivar
is probably by chance and not due to genetic resistance or susceptibility. Inoculation

studies should be performed to clarify the level of resistance of each rootstock to CVA.

125

4 - I...
I
‘:

 

| Greek 3
Gt: ela 5
Edabriz
(31195120

. Gisela 6
W158
W10
Mazzard
Mahaleb
Errh V

In

t

2
_ ; ;.
w ‘
“It

 

Figure 3.3: CVA detection in the sweet cherry variety ‘Hedelﬁngen’ grafted on 9
rootstocks that exert different degrees of vigor to the scion. (A) RT-PCR ampliﬁcation of
a 1532bp fragment of the CVA genome using primers CVA4-CVA5. W158: Weiroot158,
W10: Weiroot10. Rootstocks are arranged from the most dwarﬁng (Gi3) to the most
vigorous (Erdi V). (B) Northern blot hybridization was used to conﬁrm the RT-PCR

result. The image is aligned to the RT-PCR image in (A). rRNA denotes the RNA
loading control.

126

 

 

.modnm um moococobze 185323 30:3 £082 98 mm 33. so Emcﬂ 80cm awake: 05
“comma 33:3 @6882: “:2on a co Bag 328 .eowcﬂowom. mo Ewan: 805 2038 mo cam ”NM 033.

 

8.3 nomdm Exam 095mm on. 5 amoNN mm. FN mu. FN mm. rm

 

.>_Ew EmNNmE 322—ms. wmﬁ>> 9.3 @3090 owﬁmEO Ntnmom m<_mm_0

 

127

'Sam’l G15
‘Hudsnn’lGIS
‘Bing'/Edabriz

‘Brooks’lGIS
‘Hndsnn 'IGI6
‘Attikn'lGIS

‘Atxikn'lGIS
‘Bing'/016

:1
a 53
4 4

 

--7::-- . m CVA

aﬂﬂuunau: rRNA

Figure 3.4: CVA detection in 5 sweet cherry varieties grafted on 3 rootstocks. (A) RT-
PCR ampliﬁcation of a 1532bp fragment of the CVA genome using primers CVA4-
CVAS. G15: GiSelA5, G16: GiSelA6. (B) Northern blot hybridization was used to
conﬁrm the RT-PCR result. The image is aligned to the RT-PCR image in (A). rRNA
denotes the RNA loading control.

 

128

DISCUSSION

Viruses of the genus Capillovirus, such as ASGV, Citrus Tatter Leaf Virus
(CTLV) and Lilac Chlorotic Leaf Spot Virus (LCLV), have not been linked previously to
tree height or vigor reduction (Biichen-Osmond, 2004). While our initial study revealed a
coincidental association of CVA with the dwarﬁng cherry rootstock Gi5, but not the more
vigorous Gi6, our subsequent screenings suggest that CVA is not responsible for the
alteration of tree vigor in dwarﬁng cherry rootstocks. CVA was present in trees on
relatively vigorous rootstocks (Weiroot158 and GiSelA6 with some scion varieties) that
show the expected growth. Comparing previous records on the virus (Eastwell and
Bemardy, 1998; James and Jelkmann, 1998; Jelkmann, 1995) and the experiments
reported in this study, the absence of the virus from certain trees or graft combinations is
circumstantial rather than due to a consistent association with rootstock vigor or
genotypic tolerance.

There have been few reports on the occurrence of CVA since its discovery by
Jelkmann (1995), likely because CVA has not been linked to any symptom that reduces
cherry tree productivity or fruit quality (Eastwell and Bemardy, 1998; James and
J elkmann, 1998). As a result, no commercial detection assay has been developed to track
virus abundance or distribution. Thus, it is possible that, in addition to its natural
occurrence in mature trees, the virus has been spread via propagation of nursery stock as
well. Indeed, that appears to be the case from the range of scions and rootstocks we
screened as young trees in this study. Additionally, even though we found that mutations
have been accumulating at the nucleotide level, they have been translated to only a few

amino acid changes.

129

 

The cDNA-AF LP proﬁles during three different dates (Figure 3.2A) show that the
concentration of the viral RNA is higher in mid-June, when the trees have reached the
peak of growth activity. This observation is important for easier detection of the virus
(when a genome detection method is used). Furthermore, total RNA was enough to detect
the virus, since the CVA genome consists of a long RNA molecule with a poly-
adenylated tail. The cDNA-AF LP data also provide another proof to Jelkmann’s report
(1995) that this virus is graft transmissible. This is important for the effective control of
the virus, especially at the nursery level.

Future research on CVA should include further characterization of its possible
symptomatology (possibly in synergy with other common viruses like Prunus Necrotic
Ringspot and Prune Dwarf Virus, which are both pollen-borne and hence can infect an

orchard at maturity); production of a fast detection assay (perhaps by immunoassay); and
since it may be spread through grafting, determination of whether CVA-free certiﬁcation

should be a consideration for nursery mother blocks to prevent its continued spread.

130

 

all

LITERATURE CITED

Bachem, C.W.B., van der Hoeven, R.S., de Bruijn, S.M., Vreugdenhil, D., Zabeau, M.
and Visser, R.G.F. (1996). Visualization of differential gene expression using a novel
method of RNA ﬁngerprinting based on AFLP: Analysis of gene expression during
potato tuber development. Plant Journal 9: 745-753

Bﬁchen-Osmond, C. (2004). ICTVdB: The Universal Virus Database of the International
Committee on Taxonomy of Viruses. Biomedical Informatics Core. Northeastern
Biodefense Center. Regional Center of Excellence in Emerging Infectious Diseases
and Biodefense. Columbia University, New York. 11 February 2004
<http://www.11cbi.nlm.nihgov/lCTde/index.htm>

 

Eastwell, KC. and Bemardy, MG. (1998) Relationship of Cherry Virus A to Little
Cherry Disease in British Columbia. Acta Horticulturae 472: 305-313.

Foissac, X., Svanella-Dumas, L., Dulucq, M.J., Candresse, T. and Gentit, P. (2001).
Polyvalent detection of fruit tree tricho, capillo and foveaviruses by nested RT-PCR
using degenerate and inosine containing primers (PDO RT-PCR). Acta Horticulturae
550: 37-43.

Isogai, M., Aoyagi, J ., Nakagawa, M., Kubodera, Y., Satoh, K., Katoh, T., Inamori, M.,
Yamashita, K. and Yoshikawa, N. (2004) Molecular detection of ﬁve cherry viruses
from sweet cherry trees in Japan. Journal of General Plant Pathology 70: 288-291

James, D. and Jelkmann, W. (1998). Detection of Cherry Virus A in Canada and
Germany. Acta Horticulturae 472: 299-303.

Jelkmann, W. (1995) Cherry Virus A: cDNA cloning of dsRNA, nucleotide sequence
analysis and serology reveal a new plant capillovirus in sweet cherry. Journal of
General Virology 76: 2015-2024

Kirby, M.J., Kirby, M.J. and Adams, A.N. (2001). Occurrence of Cherry Virus A in the
UK. Plant Pathology 50: 801

131

 

 

Lang, GA. and Howell, W. (2001) Lethal sensitivity of some new cherry rootstocks to
pollen-borne viruses. Acta Horticulturae 557: 151-154

Lecelier, C-H. and Voinnet, O. (2004) RNA silencing: no mercy for viruses?
Immunological Reviews 198: 285-303

Llave, C., Kasschau, KB. and Carrington, J .C. (2000) Virus-encoded suppressor of
posttranscriptional gene silencing targets a maintenance step in the silencing pathway.
Proceedings of the National Academy of Sciences, USA 97: 13401-13406

Lucas, W.J. (2006) Plant viral movement proteins: Agents for cell-to-cell trafﬁcking of
viral genomes. Virology 344: 169-184

Nelson, RS. and Citovsky, V. (2005) Plant viruses. Invaders of cells and pirates of
cellular pathways. Plant Physiology 138: 1809-1814

Sambrook, J., Fritsch, BF. and Maniatis, T. (1987). Molecular cloning: a laboratory
manual. Cold Spring Harbor Laboratory Press

Virus Identiﬁcation Data Exchange (VIDE) http://i111age.fs.uidaho.cdu/vidol/"(108011157.11tm

 

Vos, P. ,R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Homes, A. Frijters, J. Pot,
J. Peleman, M. Kuiper and Zabeau, M. (1995). AFLP: a new technique for DNA
ﬁngerprinting. Nucleic Acids Research 23: 4407-4414

Wang, M-B. and Metzlaff, M. (2005) RNA silencing and antiviral defense in plants.
Current Opinion in Plant Biology 8: 216-222

Wang, X.S., W. Hunter and Plant, A. (2000). Isolation and puriﬁcation of ﬁrnctional total
RNA from woody branches and needles of Sitka and White Spruce. BioTechniques
28: 292-296

Yoo, B.C., Kragler, F ., Varkonyi-Gasic, E., Haywood, V., Archer-Evans, S., Lee, Y.M.,
Lough, T.J. and Lucas, W.J. (2004) A systemic small RNA signaling system in
plants. Plant Cell 16: 1979-2000

132

 

‘I‘F.’ .. . _'

 

DISSERTATION SUMMARY, CONCLUSIONS AND IDEAS

 

133

SUMMARY CONCLUSIONS AND IDEAS

The aim of this study was the identiﬁcation of genes involved in rootstock
induced dwarﬁng. Several ideas have been put forth, some of which were attempted and
some remained in the blueprint. In this section initial ideas will be presented and
conclusions will be drawn based on the current results.

The idea for this project was formed by the need for genetic markers that will
assist the efﬁcient breeding of dwarﬁng rootstocks. Many of the current cherry rootstocks
are products of interspeciﬁc crosses that are sterile, thus making genetic mapping

impossible. Availability of genetic markers linked to RID would make breeding of

 

dwarﬁng rootstocks an easier task. A fact in RID is that the genetic background of the
rootstock is the driving force for the degree of vigor excerpted to the scion. Thus
rootstocks with the same genetic background and different dwarﬁng capacity should be
easier to compare for the identiﬁcation of genetic loci involved in this phenomenon.
Gisela5 and Gisela6, as discussed previously, are two of the most successful cherry
rootstocks that confer different degrees of vigor to the cherry variety ‘Bing’. The
mechanism by which the rootstock dwarfs the scion is not clear. Several hypotheses were

discussed in the Introduction, most of which relate to secondary effects of grafting.

 

cDNA-AF LP was selected to test the hypothesis that assumes the existence of a mobile
mRN A signal from the rootstock to the scion and its effect on gene expression. One
concern was whether this approach is going to return any meaningful result, but the data
in Chapter 2 proved that it did. It would have been a much more straight forward

approach to extract phloem sap and study its transcriptome. Cherry, however, is a

134

difﬁcult plant to obtain phloem sap from, and several attempts to do so proved
unsuccessful. Another concern was that changes in the transcriptome may not be
reﬂected in the proteome or the opposite. Such differences would not be detected by
cDNA-AF LP. Two-dimentional (2-D) protein electrophoresis would be able to provide
some clues on this question. Nevertheless, 2-D gels are very difﬁcult to produce or
reproduce, and low level signals cannot be detected. Thus, the method was excluded from
further consideration as impractical and deviating from the main target, which was
identiﬁcation of RNA signals.

During the course of the project it was shown that dwarf trees respond faster to a
signal that is consistent throughout the growing seasons. The signal is unknown, but the
consistency of the data indicates that it is a periodic environmental signal, which may
function through a signaling cascade or through changes in the physiology of the tree that
trigger downstream signaling. The dominance of this signal is not known. Is it
deprivation of some signiﬁcant grth compound that is reduced or is it a repressor of
growth that triggers dwarﬁng? To test that question two approaches were taken. In the
ﬁrst and probably the most direct, ‘Bing’/Gi5 and ‘Bing’/Gi6 trees were approach grafted
to produce trees with two rootstocks. Grafts were performed at the ‘Bing’ scions to avoid
any incompatibility issues. Measurements were taken on these trees to test their behavior
in comparison to regular ‘Bing’/Gi5 and ‘Bing’/Gi6 grth rates. Unfortunately, the
approach grafts created enormous wounds to the trees, which were not able to recover
and eventually grafts were aborted. Information from this experiment was going to be
very informative on the dominance of the dwarﬁng signal and will provide clues on the

nature of the signal. Experienced grafters may be the solution to the problem of aborted

135

 

 

 

grafts. The second approach was reciprocal grafting, which would allow testing the effect
of one rootstock on top of the other. Unfortunately, even in this case the grafting was
unsuccessful due to the advanced growth of the ungrafted trees. The success rate was 2-
5%, which is translated to 1-3 trees per combination. This number of trees is not enough

for a well designed experiment.

 

One would expect that changes in shoot length are caused by the action of a
certain hormone and especially the traditional hormones linked to stem elongation, such
as auxin and gibberellin. Such an analysis seems reasonable, but the growth
measurements showed that differences in shoot length between graft combinations are

due to cessation of grth rather than metamer length. This was an initial indication that

 

auxin or gibberellin are not involved in this phenomenon for these particular rootstocks.
Indeed, cDNA-AF LP analysis did not return any of the known auxin or gibberellin
regulated genes. Nevertheless, a group of genes was annotated as brassinosteroid
regulated genes. Brassinosteroids are also linked to plant growth and development. It
cannot be concluded though if these genes are responding to brassinosteroids, since no
such experiment has been conducted in this project.

Abscisic acid (ABA) is another hormone that has been related to growth and more
speciﬁcally with dormancy in seeds. It is reasonable to believe that cessation of shoot
grth and bud set are part of the dormancy process. Nevertheless, cDNA-AFLP analysis
did not return any ABA related genes. This may be explained by the absence of
dormancy at the time of shoot growth cessation. Even though shoots have ceased growing
and bud has set, the trunk is still active and expands until September. Leaf drop occurred

later in October, thus explaining the absence of ABA regulated genes in June.

136

From these observations it seems that shoot grth cessation is related to the
reduction in the rate of cell divisions and cell elongation. Such a process is difﬁcult to
explain since all mechanisms for SAM maintenance and cell differentiation from the
SAM are functional, but they only occur in slower rates and eventually stop. This is
supported by the absence of SAM regulatory or cell division related genes from the
analysis.

The use of rootstocks from the same cross provided the opportunity to study gene
expression at the graft union. If the rootstocks were not closely related then genetic
differences would be enhanced and thus make the analysis of gene expression more
difﬁcult. Nevertheless, in the analysis of gene expression at the upper main shoot, the
ideal comparison would be between graft combinations with signiﬁcant difference in
vigor. That selection would have made the differences in gene expression more obvious
and may have also returned many more genes.

Genes identiﬁed in this project should be analyzed further for their involvement in
the dwarﬁng phenomenon. To achieve this goal it is important to involve more graft
combinations and test the correlation of gene expression to rootstock vigor. Consistency
between gene expression and rootstock vigor should qualify these genes for markers in
the screening of promising rootstock breeding trials. The regulation and the impact of
polymorphisms (SNPs, indels) in these genes should be studied further to allow more
accurate screening in rootstock trials.

The presence of the CVA in the ‘Bing’/Gi5 combinations was initially a concern,
since viruses can have a signiﬁcant effect in tree physiology. Tests presented in Chapter 3

reduced this possibility since the virus is also present in vigorous trees. Nevertheless,

137

 

presence of the virus may have affected the expression of some genes. Expression
patterns similar to the CVA proﬁle were not observed in the cDNA-AF LP analysis or the
microarray screening providing another proof that the virus is not affecting the

physiology of the tree.

138

 

   

S

ullllﬁlljlllljljﬂljj111111111