:x. at?
z. i :

..:r r.l,
.315 o.
g... 1..

.... $v e
11v 14““5'

\

19-h
DELTA!)
Embryflnnfifinnh. 2%.". ..
5.... .x.‘ J afl 1.33.".

E

H.
9
«n

«gl-

avg!"
1‘ . . .

MCKWV.IW
. a

on

k; ‘,
v.43“,
... ..

1
I

0.5).». x

5.3!... u.-

{Irfiil-Z‘tr
$25,...
(.2!!!

 

v .. fir . .§.r. .ll.
vVQ)I.c ‘9' (5‘ L..l .llv!
$5.31.... .s..v .1.....\ t

 

. . . . , . . . . xv ....o.....l.ah......
. . . . . . . . . .11. Haley! who.) .5 1‘
.. .. if. .2... . 5%... . . . . . 12351.», ...«.,...uyi.
I? , f a». . 11...... . . . ... . :

.. Fuluwfiwzf. : . . . . . . .

 

9““ J\l\ll\llll\lll\lll\llllll\lll§l\t

Z 3 1293 01391 819

This is to certify that the
dissertation entitled
The Consequences of Polyploidy on
Inbreeding Depression in
Vaccinium (Blueberry) Species
presented by

Karen E. Hokanson

has been accepted towards fulfillment
of the requirements for

PhD degree in HortiCUIture

 

 

Ecology and Evolutionary
Biology Program

Major professor

Date l1/2i [as

MS U it an Affirmative Action/Equal Opportunity Institution 042771

 

LIBRARY
Michigan State
University

 

 

 

PLACE ll RETURN BOX to remove thle checkout from your record.

TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU to An Affirmative Action/Equal Oppommlty Induction
W

M‘

THE CONSEQUENCES OF POLYPLOIDY ON INBREEDING DEPRESSION
IN VACCINIUIVI (BLUEBERRY) SPECIES

By

Karen E. Hokanson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Horticulture
Ecology and Evolutionary Biology Program

1995

ABSTRACT

THE CONSEQUENCES OF POLYPLOIDY ON INBREEDING DEPRESSION
IN VACCINIUM (BLUEBERRY) SPECIES

By

Karen E. Hokanson

The approach to homozygosity upon selfing is theoretically slower in an
autotetraploid, due to tetrasomic inheritance, than in a diploid. As a result, less
inbreeding depression is predicted in a tetraploid because fewer deleterious recessive
traits should be expressed. This research examined how well two autotetraploid
Vaccinium species, V. corymbosum and V. angustifolium, and one diploid, V.
myrtilloides, fit this prediction.

There was a significant reduction in fruit set and seed set following self
pollination compared to outcross pollination in all of these species. The average
relative survivorship (the percent mature seed per self pollination relative to outcross
pollination) was 0.255 in V. corymbosum, 0.079 in V. angustifolium, and 0.020 in V.
mym'lloides. Rarely have such drastic reductions in fertility following selfing been
attributed to inbreeding depression. However, in this study, support for early-acting
inbreeding depression, rather than pre- or post—zygotic self incompatibility, included:
1) a higher proportion of aborted seed in selfed fruit than in outcrossed fruit;

2) aborted seed that ranged in size and shape; 3) a positive correlation between self
and outcross seed set per pollination; 4) results of a pollen chase experiment which

provided evidence that fertilization by self pollen does occur. Inbreeding depression

was more severe in diploid V. myrtilloides and tetraploid V. angustzfolium than in
tetraploid V. corymbosum.

The level of heterozygosity and number of alleles per locus, averaged over
seven polymorphic isozyme loci, were lower in the diploid, V. myrtilloides (21.7%;
2.9), than in the tetraploids, V. angustifolium (57.1%; 3.4) and V. corymbosum (75.6%;
3.6). The proportion of fertilized ovules that developed into mature seed (percent
mature seed) in open-pollinated (OP) fruit was highest in V. corymbosum (44.6%),
intermediate in V. myrrilloides (38.3%), and lowest in V. angustifolium (27.7%).
Based on the levels of inbreeding depression in these species, the lower percent mature
seed in GP fruit from V. angustifolium was unexpected. Geitonogamous self
pollination was considered as an explanation for the lower level of OP seed set in V.
angustrfolium, and for the lower level of heterozygosity in V. angustzfolium than in V.
corymbosum. The theoretical effects of tetrasomic inheritance and the level of

heterozygosity on the expression of genetic load are discussed.

ACKNOWLEDGEMENTS

This dissertation never could have been completed without the willing
assistance of many people, beginning with those who assisted with field collections,
various greenhouse chores, occassional lab work, or other technical tasks: Stan
Hokanson, Pete Callow, Jim Hancock, LuPing Qu, Dan Prince, Muzaffar Sakin, Jean
Baker, Gary Schott, and Roger May. I would also like to thank these people, and
many others, for their friendship and support during my time at MSU.

I would like to acknowledge my thesis committee: Amy Iezzoni, Kay Gross,
and Susan Kalisz. Their input in the design and completion of my dissertation is
appreciated. Although I chose these individuals primarily for their expertise, I also
sought mentors in a field where there are few women. Perhaps without even being
conscious of their duty in this respect, they each have fulfilled their roles. I am
inspired by their achievements. I would also like to thank Eric Hanson for graciously
agreeing to take Susan‘s place during my defense, and for his helpful comments on my
dissertation. Steve Krebs must also be acknowledged for building the launching pad
in his own dissertation for a large part of my research, and for his particularly useful
input on my dissertation. .

Many people choose to acknowledge their parents at this time, probably for
providing them with all of the opportunities that brought them here and for the love

and support over the years. For these things, I am truly thankful to my own parents.

iv

But my parents also provided something more tangible that, I can surely say, made it
possible for me to finish this degree. They cared for Kate during the days that Stan
and I were most frantically trying to get our dissertations to our committees. I don't
know what we would have done without them.

It has been challenging at times to be both a mom and a graduate student.
There are two people I would like to thank for making this possible for me: Kate and
Stan. If Kate were not the happy, cooperative little girl that she is, my job would
have been much harder. She always gave me something to look forward to at the end
of the day. I also never could have come this far without the love and support of my
husband, Stan. His companionship, partnership, and gentle prodding when the going
got tough, made the whole ordeal of getting this degree bearable at the times when it
otherwise would not have been. I know that he and I are an inseparable team.

Finally, I would like to thank my advisor, Jim Hancock. I admire Jim for his
flexible mind and his ability to successfully juggle so many research interests, which
allowed me to pursue my own interests. He has been incredibly supportive of me in
all of my efforts, both personally and professionally. I realize that a good advisor can

make or break the graduate experience. I was fortunate to have one of the best.

TABLE OF CONTENTS

Page

LIST OF TABLES .............................................. vii

LIST OF FIGURES ............................................. ix

INTRODUCTION ............................................... 1

CHAPTER 1: EARLY-ACTING INBREEDING DEPRESSION AND

ESTIMATES OF EMBRYONIC GENETIC LOAD IN TWO TETRAPLOID

AND ONE DIPLOID VACCINIUM (ERICACEAE) SPECIES ............... 11
Abstract ................................................ 12
Introduction ............................................. 13
Materials and Methods ...................................... 16
Results ................................................. 24
Discussion .............................................. 34
Conclusion .............................................. 48
Literature Cited ........................................... 50

CHAPTER 2: GENETIC VARIATION AND THE EXPRESSION OF

GENETIC LOAD IN NATURAL POPULATIONS OF TWO TETRAPLOID

AND ONE DIPLOID VACCINI UM SPECIES ........................... 57
Abstract ................................................ 58
Introduction ............................................. 59
Materials and Methods ...................................... 61
Results ................................................. 68
Discussion .............................................. 74
Conclusion .............................................. 80
Literature Cited ........................................... 82

Table

LIST OF TABLES

CHAPTER 1

Names, codes, and locations (county, township range, section) of V.
corymbosum, V. angustifolium, and V. myrtilloides populations in the

lower and upper peninsulas of Michigan .......................

The average fruit and seed set after self and outcross pollinations on
fifteen plants from each of three Vaccim‘um species, and the results

of pairwise t-test comparisons (contrasts) between the species ........

Results of pollen chase experiments showing percent mature seed per
pollination for each control and treatment in two plants of each

Vaccinium species. C1 = Control 1; C2 = Control 2; C3 = Control 3;
T1 = Treatment 1; T2 = Treatment 2. Pairwise t-test comparisons of

the control and treatment means for each plant are also shown .......

The percent mature seed per self and outcross pollination, the
relative survivorship and the number of lethal equivalents in
self-fertile V. combosum, V. angustifolium, and V. myrtilloides

plants ................................................

CHAPTER 2

Allele frequencies at two enzyme loci in Michigan populations of

three Vaccinium species ...................................

Equations used to calculate expected heterozygosity at diploid and
tetraploid loci in populations with two alleles (p,q), three alleles (p,q,r),

and four alleles (p,q,r,s) ...................................

Variation at seven enzyme loci in three Vaccinium species. Values are

averaged across individuals from three p0pulations of each species .....

vii

Page

...18

...25

..31

..33

..66

..67

..69

Table

LIST OF TABLES (continued)
Page

Mean seed set in three Vaccim'um species. Values are averaged across
individuals from three populations of each species. Lower case letters
represent mean separations between species (LSDOS) ................. 71

Seed set and isozyme variation averaged across individuals within

populations of three Vaccim'um species. Observed and expected
heterozygosities are shown for each population. Lower case letters

represent mean separations between populations within species (LSDOS) . . . 72

viii

Figure

LIST OF FIGURES

Page
CHAPTER 1
Approximate locations of the Vaccinium populations on a county map
of the state of Michigan ..................................... 20
Correlation between fruit set and the % mature seed per fruit in the
three Vaccinium species: A. V. corymbosum; B. V. angustr‘folium;
C. V. mym'lloides ......................................... 28
Correlation between self and outcross % mature seed per pollination
in the three Vaccim‘um species: A. V. corymbosum; B. V. angustr‘folr'um;
C. V. myrtilloides ......................................... 30
The predicted relationship of diploid and autotetraploid relative
survivorship to the number of lethal equivalents .................... 43

ix

Figure

LIST OF FIGURES (Continued)
Page
CHAPTER 2

Figures 1-6. Five enzyme loci in diploid and tetraploid Vaccim'um,
illustrating the banding patterns of different genotypes. Loci are
designated by the numbers on the left, with superscript letters indicating
alleles. Figure 1. Monomeric SKDH in a diploid. Lane A is
homozygous for allele a (aa), with a "shadow" band below it; Lane B is
a heterozygote with the genotype ab, and with "shadow" bands below
both allelic bands. Figure 2. Monomeric SKDH in a tetraploid. Lane
A is homozygous for allele b (bbbb), with a "shadow” band below it;
Lanes B and C are di-allelic heterozygotes with the genotypes aabb and
aaab, respectively. "Shadow” bands are below both allelic bands.
Figure 3. Monomeric Pgm-Z in a tetraploid. Lane A is homozygous
for allele b (bbbb); Lanes B and C are di-allelic heterozygotes with the
genotypes bbcc and abbb; Lane D is a tri-allelic heterozygote with the
genotype aabc. Figure 4. Dimeric 6pgdh—2 in a tetraploid. Lane A

is homozygous for allele b (bbbb); Lanes B, C, and D are di-allelic
heterozygotes with the genotypes aabb, aaab, and abbb, respectively.
Heterodimeric bands are in between the allelic bands. Figure 5.
Dimeric Mdh-l and Mdh-2 in a diploid. Mdh-l: Lanes A and B are
homozygous for allele a (aaaa) and b (bbbb), respectively; Lane C is a
heterozygote with the genotype ab, and with a heterodimeric band in
between the two allelic bands. Mdh-Z: All lanes are homozygous for
allele c (cc). Figure 6. Dimeric Mdh-l and Mdh-2 in a tetraploid.
Mdh-l: Lane B is homozygous for allele a (aaaa); Lanes A,C,D, and E
are di-allelic heterozygotes with the genotype aaab, with heterodimeric
bands in between the allelic bands. Mdh-Z: Lane A is homozygous

for allele c (cccc); Lanes B,C, and D are di-allelic heterozygotes with
the genotypes aacc, accc, and bccc, respectively, with heterodimeric
bands in between the two allelic bands; Lane D is a tri-allelic
heterozygote with the genotype abcc, with a heterodimeric band in
between each pair of allelic bands .............................. 64

INTRODUCTION

Nearly half of the extant flowering plant species, or more by some estimates,
are polyploid (Stebbins, 1971; Grant, 1981; Hancock, 1992). Because polyploidy is so
prevalent, it is considered fundamentally important in angiosperm evolution. Early
evolutionary studies of polyploidy were primarily concerned with the origins of
polyploids and their cytology. At one time, most polyploids in nature were assumed
to be allopolyploid (Stebbins, 1950). However, examples of autopolyploids, supported
by the use of molecular techniques in inheritance studies, are becoming increasingly
common (Thompson and Lumaret, 1992). Because autopolyploids are more common
in nature than once thought, an understanding of the evolutionary consequences of
autOpolyploidy is essential to the understanding of angiosperm evolution.

The main difference between auto— and allo-tetraploids is their mode of
inheritance. Traits are inherited disomically in an allotetraploid, as in a diploid, which
means the alleles on only two chromosomes are associated with each other during
meiosis. In an autotetraploid, four chromosomes are randonly associated with each
other and inheritance is, therefore, tetrasomic. There are potential evolutionary
advantages shared by both types of polyploids, but there are also unique advantages

associated with the individual modes of inheritance.

2

An increased level of heterozygosity is anticipated in both types of polyploids
compared to a diploid. A tetraploid is potentially heterozygous at more loci and can
carry more alleles at a single locus than a diploid. A diploid can only be heterozygous
with one copy of two alleles at a locus (ab). A tetraploid can be heterozygous for two
alleles with one, two, or three copies of each allele at a locus (aaab, aabb, abbb), or it
can have more than two alleles at a locus (i.e., aabc. abcd). An allopolyploid can also
have ”fixed” heterozygosity, when homeologous loci are homozygous for different
alleles. Both types of polyploids presumably will be larger and more vigorous or
more widely adapted than their diploid progenitors because of the heterotic effect of
more heterozygous loci and more alleles at a locus (Stebbins, 1971; Grant, 1981;
Hancock, 1992).

Along with increased heterozygosity, autotetraploids have the potential
advantage of a slower approach to homozygosity following inbreeding, due to
tetrasomic inheritance. Inbreeding depression, or the reduced fitness of offspring upon
inbreeding, is primarily attributed to the expression of deleterious recessives in the
homozygous condition. Assuming normal Mendelian inheritance, 1/4 of the offspring
will be homozygous recessive following self fertilization at a heterozygous locus in a
diploid. At an silo. or auto-tetraploid locus with genotype abbb, 1/4 of the offspring
will also be homozygous recessive (bbbb), and none of the offspring will be
homozygous recessive at a locus with genotype aaab (assuming no double reduction).
However, at a tetraploid locus with the genotype aabb, 1/16 of the offspring will be

homozygous recessive following selfing when inheritance is disomic (assuming no

3

fixed heterozygosity), but only 1/36 of the offspring will be homozygous recessive if
inheritance is tetrasomic.

Therefore, in the early stages of inbreeding, diploids and allotetraploids may be
more prone to inbreeding depression than autotetraploids because homozygosity
increases rapidly upon selfing. At the same time, tetrasomic inheritance may allow
more deleterious recessives to accumulate than does disomic inheritance. By a
phenomenon known as purging, deleterious recessives should be selected against as
they become homozygous (Lande and Schemske, 1985; Charlesworth and
Charlesworth, 1987). In an autotetraploid, homozygosity accrues more slowly and
segregation at loci with deleterious recessives continues over many more generations.
Because of this, autOpolyploids may suffer greater inbreeding depression in later
generations of self fertilization than do diploids or allotetraploids.

Inbreeding depression in the initial generations of selfing, however, does not
always meet theoretical expectations in autotetraploids based solely on the approach to
homozygosity. The classic example is the severe inbreeding depression observed after
one generation of selfing in autotetraploid lines of Medicago sativa (alfalfa). The
unexpected severity of inbreeding depression in alfalfa has most commonly been
attributed to the loss of multiple allelic interactions at a single locus (overdominance
or segregational load) (Busbice and Wilsie, 1966; Dunbier and Bingharn, 1975; Groose
et al., 1989). For example, a genotype with four alleles at a locus (abcd) when selfed
would only produce 16.6 percent offspring with four alleles at that locus and none of

its offspring would be homozygous. Thus, assuming that more alleles at a locus is

4

associated with higher fitness, the fitness of the population would decrease
significantly with no increase in homozygosity.

Theoretical changes in the frequency of multiple allelic loci also explain the
progressive heterosis observed over generations of crosses in alfalfa and
autotetraploids, and early autopolyploid theory favored the single gene model of
multiple allelic interactions because it could explain both progressive heterosis and
inbreeding depression. Additive effects among alleles at different loci had also been
proposed as an explanation for progressive heterosis in autotetraploids, but only
recently was it theoretically shown that additive loci could also explain the unexpected
severity of inbreeding depression (Bingham et al., 1994). For example, self
fertilization of a genotype with four linked loci with a favorable dominant allele at
each locus produces only 16.6 percent offspring with a favorable allele at each locus.
Recent work with molecular markers in autopolyploids indicates that heterosis is
associated with the additive effects of favorable alleles at different loci in linkage
blocks, rather than interactions between alleles at a single locus (Bonierbale, Plaisted,
and Tanksley, 1993; Kidwell et al., 1994). The individual favorable alleles with
additive effects may also contribute to non-additive complimentary gene action, or
epistasis, by masking deleterious recessives at other loci (Bingham et al., 1994).
Therefore, additive effects and interactions among loci should be given greater
consideration in studies on inbreeding depression in autotetraploids.

Clearly, there has been a great deal of theoretical debate about the response of
autotetraploids to inbreeding, but very little empirical data has been gathered in natural

populations. The Vaccim'um (blueberry) polyploid complex is an ideal system to study

5

the consequences of autopolyploidy on genetic variation and inbreeding depression.
The complex consists of an array of species with ploidy levels from 2x to 4x to 6x
(Hancock, 1995). Although the precise origins of the polyploid species are not known,
tetrasomic inheritance has been demonstrated in at least one of the tetraploids, the
cultivated highbush species V. corymbosum (Krebs and Hancock, 1989), and in a
tetraploid hybrid between V. corymbosum and the diploid V. darrowr' (On and
Hancock, in press). Because the diploid Vaccim'um species are all interfertile,
indicating very little genomic divergence between them, and all of the diploid species
produce unreduced gametes (Ortiz, et al., 1992; Hancock, 1995), it is likely that all of
the tetraploid Vaccim'um species have tetrasomic inheritance.

It has also been demonstrated in V. corymbosum that self fertility is controlled
by post-zygotic seed abortion due to early-acting inbreeding depression rather than pre-
zygotic or post-zygotic self incompatibility (Krebs and Hancok, 1988; 1990; 1991).
Among the evidence for inbreeding depression, there is a range among genotypes
within V. corymbosum from zero to high levels of self fertility, although self fertility is
generally low. Variation in the levels of self fertility can be accounted for by
variation in the levels of genetic load being expressed as inbreeding depression. In a
self incompatible species there should be zero or near zero seed set in all individuals
following self fertilization. Variation in the levels of self fertility among Vaccim'um
species may be due to variation in the amount of inbreeding depression among species
as well.

The three most common Vaccim'um species in Michigan are included in this

study: Vaccim'um corymbosum, the common highbush blueberry; Vaccinium

6

angustrfolium, the common lowbush blueberry; and Vaccinium myrtr'llor’des, the velvet-
leaf blueberry. These species make an interesting comparison for several reasons. V.
corymbosum and V. angusnfolium are both tetraploid species (4x=2n=48), and V.
myrtillor'des is a diploid (2x=2n=24). V. angusrr‘folr‘um and V. myrrilloides are both
rhizomatous, lowbush Species, but V. corymbosum is a crown-forming, highbush
species. In addition, although moderate levels of self fertility have been reported in
individuals of each of these species, the level of self fertility is generally higher in V.
corymbosum than in V. angusrrfolr'um or V. myrrilloides. Therefore, these species
provide an opportunity to compare a diploid and a tetraploid with similar growth
habits and levels of self fertility, and two tetraploids with different growth habits and
levels of self fertility.

The present study addresses the evolutionary consequences of autopolyploidy
with regard to inbreeding depression in natural populations of these three Vaccinium
species. In the first chapter, expectations for inbreeding depression verses self
incompatibility were investigated as the control mechanisms for the low levels of self
fertility in these species. The responses to inbreeding and the levels of embryonic
genetic load, estimated from the relative survivorship of selfed progeny to outcross
progeny, were compared between the species. The second chapter explores the
possible reasons for differences between the species in levels of inbreeding depression.
Both ecological differences between the species and the theoretical expectations for

inbreeding depression in autopolyploids were given consideration.

7

Literature Cited

Bingham, E.T., R.W. Groose, D.R. Woodfield, and K.K. Kidwell. 1994.
Complementary gene interactions in Alfalfa are greater in autotetraploids than

diploids. Crop Science 34(4):823-829.

Bonierbale, M.W., R.L. Plaisted, and SD. Tanksley. 1993. A test of the maximum
heterozygosity hypothesis using molecular markers in tetraploid potatoes.

Theoretical and Applied Genetics 86:481-491.

Busbice, TH. and CF. Wilsie. 1966. Inbreeding depression and heterosis in

autotetraploids with application to Medicago sativa L. Euphytica 15:52-67.

Charlesworth, D. and B. Charlesworth. 1987. Inbreeding depression and its
evolutionary consequences. Annual Review of Ecology and Systematics

18:237-268.

Dunbier, M.W. and ET. Bingham. 1975. Maximum heterozygosity in alfalfa: results

using haploid-derived autotetraploids. Crop Science 15:527-531.

Grant, V. 1981. Plant Speciation, 2nd Ed. Columbia University Press, NY.

8
Groose, R.W., L.E. Talbert, W.P. Kojis, and ET. Bingham. 1989. Progressive

heterosis in autotetraploid alfalfa: studies using two types of inbreds. Crop

Science 29: 1 173-1 177.

Hancock, JR 1992. Plant Evolution and the Origin of Crop Species. Prentice Hall,

Inc. Englewood Cliffs, NJ.

Hancock, JP. 1995. Blueberries, Cranberries, etc. In: Smartt, J. and NW. Simmonds,
Eds. Evolution of Crop Plants, 2nd Ed. p121-123. Longman Scientific and

Technical, London.

Kidwell, K.K., E.T. Bingham, D.R. Woodfield, and TC. Osborne. 1994. Relationships
among genetic distance, forage yield, and heterozygosity in isogenic diploid

and tetraploid alfalfa poulations. Theoretical and Applied Genetics 89:323-328.

Krebs, S.L. and IF. Hancock. 1988. The consequences of inbreeding on fertility in

Vaccinr'um corymbosum L. Journal of the American Society for Horticultural

Science. 113(6):914-9l8.

Krebs, S.L. and IF Hancock. 1989. Tetrasomic inheritance of isoenzyme markers in

the highbush blueberry, Vaccim'um corymbosum L. Heredity 63:11-18.

9

Krebs, S.L. and J.F. Hancock. 1990. Early-acting inbreeding depression and
reproductive success in the highbush blueberry, Vaccinium corymbosum L.

Theoretical and Applied Genetics 79:825-832.

Krebs, S.L. and J.F. Hancock. 1991. Embryonic genetic load in the highbush blueberry

Vaccim'um corymbosum (Ericaceae). American Journal of Botany 78(19):]427-

1437.

Lande, R. and D.W. Schemske. 1985. The evolution of self-fertilization and inbreeding

depression in plants. 1. Genetic models. Evolution 39(1):24-40.

Ortiz, R, N. Vorsa, L.P. Bruederle, and T. Laverty. 1992. Occurence of unreduced
pollen in diploid blueberry species, Vaccim'um sect. Cyanococcus. Theoretical

and Applied Genetics 85:55-60.

Qu, L. and J.F. Hancock. In press. Nature of Zn gamete formation and mode of
inheritance in interspecific hybrids of diploid Vaccim‘um darrowi and tetraploid

V. corymbosum. Theoretical and Applied Genetics.

Stebbins, G.L. 1950. Plant Variation and Evolution. Columbia University Press, NY.

Stebbins, G.L. 1971. Chromosomal Evolution in Higher Plants. Addison-Wesley,

Reading MA, USA.

10

Thompson, JD. and R. Lumaret. 1992. The evolutionary dynamics of polyploid plants:
origins, establishment and persistence. Trends in Research in Ecology and

Evolution 7(9):302-307.

CHAPTER 1

EARLY-ACTING INBREEDING DEPRESSION
AND ESTIMATES OF EMBRYONIC GENETIC LOAD
IN TWO TETRAPLOID AND ONE DIPLOID

VACCINIUM (ERICACEAE) SPECIES.

11

12

Abstract

There is a significant reduction in fruit set and seed set following self
pollination in two tetraploid species, Vaccim'um corymbosum and V. angustr’folium, and
one diploid, V. mym'lloides. The results of self and outcross pollinations indicate that
self fertility in these species, although it is extremely low, is controlled by early-acting
inbreeding depression and not by a pre-zygotic or post-zygotic self incompatibility
system. The evidence includes: 1) self fruit set and seed set in some plants; 2) a
higher proportion of aborted seed in selfed fruit than in outcross fruit; 3) aborted seed
that ranges in size and shape; and 4) a positive correlation between self and outcross
seed set. Pollen chase experiments also demonstrated that self pollen does fertilize the
ovules, although the self pollen is less vigorous than outcross pollen. The low levels
of self fertility in these species are probably due to a combination of fewer ovules
being fertilized because of the reduced vigor of self pollen and subsequent abortion of
a high proportion of the ovules that are fertilized by self pollen. Relative survivorship,
or the percent mature seed per self pollination relative to outcross, was lower in
diploid V. myrtillor’des (.020) and tetraploid V. angusnfolr'um (.079) than in tetraploid
V. corymbosum (.255). The number of lethal equivalents, estimated from the relative
survivorship, was nearly the same in diploid V. myrn'lloides (15.7) and tetraploid V.

corymbosum (18.6), but was considerably higher in tetraploid V. angustrfolr'um (37.7).

13

Introduction

Self infertility in plants is most commonly attributed to pre-zygotic or late-
acting self incompatibility, but early-acting inbreeding depression can also be a cause.
The mechanisms of these three systems are different. Pre—zygotic self incompatibility
is the failure of self pollen tube development in the stigma or style, before reaching
the ovule. Genetic control of this system is usually at one or a few specific loci
(Lewis, 1979). Late-acting self incompatibility is described as the failure of self
pollen tube development in the ovule, although the mechanism and genetic control of
this system is not well understood (Seavey and Bawa, 1986). Early-acting inbreeding
depression is the failure of self-fertilized ovules to develop into mature seed due to the
expression of embryonic genetic load. Genetic load can be either mutational or
segregational. Mutational load is deleterious or lethal, partially or fully recessive
allelic mutations that are expressed in the homozygous condition. Segregational load
is alleles which are more favorable in combination at a locus (heterozygous) than
when alone (homozygous). It is mutational load that is thought to be maintained at
high levels in outcrossing species and have a significant effect early in plant
develOpment (Wiens et al., 1987; Charlesworth, 1989).

Pre-zygotic self incompatibility is well documented in many plant species, but
evidence of self seed abortion clearly indicates that the loss of fertility following self
pollination is post-zygotic in other species. Because late-acting self incompatibility
and early-acting inbreeding depression are both post-zygotic events, it is more difficult
to distinguish between these. While only a few studies report self pollen tube failure in

the ovule (Kenrick, Kaul, and Williams, 1986; Waser and Price, 1991; Broyles and

14

Wyatt, 1993), evidence of early-acting inbreeding depression, particularly in long-
lived, outcrossing species, is becoming more and more common in the literature
(Weins et al, 1987; Wiens et a1, 1989; Krebs and Hancock, 1991; Burbidge and
James, 1991; Weller and Omduff, 1991; Seavey and Carter, 1994). In at least one
species, Epilobium obcordarum, recent work supports early-acting inbreeding
depression (Seavey and Carter, 1994), even though late-acting self incompatibility had
originally been proposed (Seavey and Bawa, 1986).

In a series of studies of both cultivated and wild tetraploid Vaccim‘um
corymbosum (the highbush blueberry), Krebs and Hancock (1988; 1990; 1991)
concluded that the reduction in fertility following self pollination was due to early-
acting inbreeding depression and not pre-zygotic or late-acting self incompatibility.
There were aborted seed in the fruit, and probable fertilization of ovules by self pollen
was demonstrated both cytologically and through a pollen chase experiment. Although
a few plants were self-infertile, most had at least a low level of self fertility. Also,
there was a significant positive correlation between self and outcross fertility and
between the percent aborted ovules and the inbreeding coefficient, while seed number
was inversely correlated to levels of inbreeding. Work by Vander Kloet (1991) also
supports early-acting inbreeding depression in V. corymbosum.

In two other Vaccinium species, the tetraploid V. angusnfolium and the diploid
V. myrtr'lloides, the reduction in fertility following self pollination is more severe than
in V. corymbosum. Although a small percentage of self pollinations do set fruit on
some plants of V. angustrfolium and V. myrrilloides, most plants of these two species

are considered self-infertile (Aalders and Hall, 1961; Wood, 1968; Hall et al., 1979;

15

Vander Kloet and Hall, 1981). Such a drastic reduction in fertility following selfing
has rarely been attributed to early-acting inbreeding depression. In most species where
the reduction in self fertility relative to outcross fertility is attributed to genetic load,
only a small number of completely self-infertile individuals have been found.
However, in Epilobr'um obcordatum (Seavey and Carter, 1994) and Psuedotsuga
menziesir' (Sorensen, 1969) the average relative survivorship, or the self fertility
relative to the outcross fertility, among those plants that were self-fertile was very low
(.11 in both species) and individual plants had relative survivorships as low as .01 and
.001 in these species, respectively. Also, Wiens et al. (1989) attribute a nearly
complete loss of both self and outcross fertility in Dedeclrera eurekensr‘s to genetic
load, although it is segregational rather than mutational load that accounts for the
overall reduction in fertility in this species.

Based on the evidence in the present study, we propose that the extremely low
levels of self fertility in V. angustrfolium and V. myrtr'lloides are due to extremely high
levels of embryonic genetic load and not to pre- or post-zygotic self-incompatibilty.
We compare the level of genetic load between the two tetraploid species, V.
carymbosum and V. angustrfolium, and we compare the level of genetic load in the
two tetraploids to the diploid, V. mynilloides. The estimates of embryonic genetic
load (the number of lethal equivalents per zygote) used to make these comparisons are
based on the theory that the rate of the approach to homozygosity following selfing in
an autotetraploid is one-third that of a diploid because of tetrasomic inheritance in the
tetraploid. Therefore, an autotetraploid may maintain a higher level of genetic load

than a diploid, but be less effected by inbreeding (Haldane, 1930; Bennett, 1976).

16

Few studies have compared levels of inbreeding depression in closely related
diploid and tetraploid species. Vander Kloet (1991) compared self pollinations to
outcross pollinations in diploid, tetraploid, and hexaploid V. corynrbosum clones and
found no difference in self fertility or outcross fertility between the ploidy levels. A
recent paper by Parker, Nakamura, and Schemske (1995) reported less inbreeding
depression at all stages of the life cycle, including seed set, in the diploid Epilobium
ciliatum thzm in a tetraploid sympatric congener Epilobr’um angustrfolr’um. However,
more inbreeding depression is reported in the diploid form of Epilobium angustr‘folium
than in the tetraploid form (Husband and Schemske, 1995; Parker, Nakamura, and
Schemske, 1995). The effect of ploidy on inbreeding depression and genetic load is

discussed in the present study.

Materials and Methods

53919.5

Vaccim'um spp. are long-lived, predominantly outcrossing, woody perennials
that are North American in origin (Vander Kloet, 1988). The three most prevalent
blueberry Species in Michigan were included in this study. Vaccim'um corymbosum
(2n=4x=48) is a tall (2-3m), crown-forming shrub. It is most frequently found in Open
swamps and bogs, and ranges from Florida north to Quebec and west along the Great
Lakes of Michigan. The tetraploid cytotype of V. corymbosum is an autotetraploid
(Krebs and Hancock, 1989). Vaccim'um angustr'folium (2n=4x=48) is a small (20-
50cm), densely rhizomatous shrub found in disturbed sites in boreal forests, bogs,

barrens, and mountain meadows. Its range extends from the Maritime Provinces of

17

Canada and Maine, west to Manitoba and Minnesota, and south to the mormtains of
Virginia. There is circumstantial evidence, primarily complete inter-fertility with the
autotetraploid V. corymbosum, which suggests that V. angustifolr’um also is an
autotetraploid (Hokanson and Hancock, 1993). Also, tetrasomic inheritance has been
observed in V. angustrfolium at one MDH locus (Hokanson and Hancock, unpublished
data). Vaccim'um myrtillor'des (2n=2x=24) is similar to V. angustrfolium both in
habitat and stature, although it is typically less dense. It ranges from the Northwest
territories of Canada to eastern and central US.
mggllegtions

Each species was collected from three different sites in October 1992 (Table l
and Figure 1). At most locations, only the predominant species was collected.
However, at two sites (Blueberry Ridge and Beaver Lake), both V. myrtilloides and V.
angustrfolium were prevalent and both were collected. Ten individual plants 15-30
meters apart along a transect were dug at each site and transferred in large plastic bags
to the greenhouse at Michigan State University, where they were potted in a peat/sand
soil mix amended with sulfur (approx. 1.25g/L) in 20 liter pots.
SelfagrdQutgrosngllinations

To compare self and outcross fertility within and between species and to
determine whether early-acting inbreeding depression is effecting self-fertility, self and
outcross pollinations were made on V. myflilloides and V. angustrfolium in the spring
of 1994 and on V. corymbosum in the spring of 1995. After vemalization from
November to February in the open greenhouse (average daytime temp. 0-4°C), plants

were allowed to flower by warming the greenhouse. Greenhouse daytime temperature

18

Table 1. Site names, codes, and locations (county, township range, section) of V.

corymbosum, V. angustifolium, and V. myrtilloides populations in the lower and

upper peninsulas of Michigan.

 

 

Species Site Code County Township Range Section
V. corymbosum Otis Lake COL Barry T3N-R9W 31
V. corymbosum Ely Lake CEL Allegan TZN-RISW 26
V. corymbosum Herbert Hoover CHH Oakland T2N-R7E 4
V. angustifolium Beaver Lake ABL Alger T48N-R16W 18
V. angustifolium Blueberry Ridge ABR Marquette T47N-R25W 22
V. angustifolium Adam's Trail AAT Alger T48N-R16W 32
V. myrtilloides Beaver Lake MBL Alger T48N-R16W 18
V. myrtilloides Blueberry Ridge MBR Marquette T47N-R25W 22
V. myrtilloides Poe Reef MPR Cheboygan T38N-R1W 23

 

19

Figure 1. Approximate locations of the Vaccim'um populations on a county map of the

state of Michigan.

20

 

 

Figure 1.

‘—r

 

V. corymbosum - 4x
V. angustifolium - 4x
V. myrtilloides - 2x

9*.
fl 25 50
.——_—._
0 40 fl
Km

MICHIGAN COUNTIES

 

m-
mmm orseao um we

mm Wimt Wlmm m >
/wu«srrr woman MW comb“ mam nosey

UKEOSCEWCIMEGWWIN

 

 

 

 

 

 

 

 

 

0cm necosn SABEUJ man W
mm

 

 

 

 

 

 

 

earsxtr; WW sacrum

 

0 A T IONIA 9114
WA '
CUNTW A555!

 

 

 

 

 

 

l
A a wow mu» lumsmu
I

moo cunouu Morse" (WWIW

WWW 1......

 

 

 

 

 

 

If't

21

was maintained between 20 and 25°C from the time of pollinations until fruit harvest.
All the flowers used were emasculated after the style had elongated but before the
corolla had opened. Pollen was collected on a 24x55mm glass coverslip and
immediately applied to the stigma. The amount of pollen in each cross was controlled
by completely covering the stigma with pollen. Twenty self and twenty outcross
pollinations were made on each plant, which included five plants from each of the
three sites for each species (15 plants per species). The outcross pollinations on each
plant were made using pollen from the other four plants collected from the same site;
five pollinations by each pollen parent were made.

Fruit was harvested when ripe (80-100% blue in color), between 4 and 12
weeks post-pollination. A few fruit that were not ripe at 12 weeks post-pollination
were harvested while still green. Fruit set per plant was calculated as the number of
fruit harvested divided by 20 pollinations x 100. Mature seeds and aborted seeds were
counted from each fruit as in Krebs and Hancock (1991). The proportion of fertilized
ovules that developed into mature seed (% mature seed) in each fruit was calculated as
the number of mature seed / total fertilized ovules (mature + aborted) x 100. For each
plant, seed set was calculated two ways: 1) as the sum of the % mature seed in all
fruit divided by the number of fruit harvested (% mature seed per fruit), and 2) as the
sum of the percent mature seed in all fruit divided by 20 pollinations (% mature seed
per pollination). For % mature seed per fruit, plants with no fruit set were not
included in the analysis. For % mature seed per pollination, pollinations which
resulted in no fruit were considered as fruit with zero percent mature seed; therefore,

plants with no fruit set had zero percent mature seed per pollination. Arcsine

22

transformed fruit set and seed set were compared by t-test between species (mean
comparisons) and between self and outcross pollinations within species (mean
comparisons of per fruit data; paired comparisons of per pollination and fruit set data).
The t statistic and degrees of freedom were calculated for unequal variances where
appropriate (Ott, 1988). Linear correlations were tested between % mature seed per
fruit and fruit set (including self fruit when set) and between self and outcross %
mature seed per pollination.
Ilgflsnghass

Pollen chase experiments are often used to demonstrate that self pollen, when
applied before outcross pollen, fertilizes the ovules and prevents fertilization of the
ovules by the outcross pollen. To test for this effect in these Vaccinium species,
pollinations were made in the Spring of 1995 in the greenhouse on two plants each of
V corymbosum, V. angustifolium, and V. myrtilloides. Two days before the initial
pollinations were made, all open flowers were removed and the plants were covered
with gauze to discourage any insect pollination. We used open flowers or flowers in
which the stigma had elongated but the corolla had not yet opened. Because some of
the plants had limited numbers of flowers, it was necessary to use flowers on day 0
that ranged in age from zero to two days old. By day 3, flowers were three to five
days old. The period of pollen receptivity is reported to last at least four days in V.
corymbosum and four to eight days in the lowbush blueberry (Hall et al., 1979; Eck,
1986)

On day 0, 50 flowers were emasculated on each plant. Several different

control and treatment pollinations were made using 10 flowers for each: 1) self pollen

23

was applied on day 0 (Control 1); 2) outcross pollen was applied on day 0 (Control 2);
3) outcross pollen was applied on day 3 (Control 3); 4) self pollen was applied on
day 0, followed ("chased") by outcross pollen on day 1 (T reatrnent l); 5) self pollen
was applied on day 0, followed by outcross pollen on day 3 (Treatment 2). A single
pollen donor was used for all outcross pollinations on a single mother plant. Pollen
donors were selected randomly with regard to compatibility with the mother plants, but
were limited to plants from the same collection site as the mother, except for that used
to pollinate MBR3.

Fruit was harvested when ripe. Mature seeds and aborted seeds were counted
and the proportion of fertilized ovules that developed into mature seed in each fruit
was calculated as described above. For each plant, the percent mature seed per
pollination for each control and treatment was calculated as the sum of the percent
mature seed in all fruit divided by ten pollinations. Pollinations that did not result in
fnrit set were considered as fruit with zero percent mature seed. The data were arcsine
transformed for analysis. A t-test was used to make mean comparisons between all
treatments and controls within each plant. Again, the t statistic and degrees of
freedom were calculated for unequal variances where appropriate.

Embryonic 3mm

To estimate embryonic genetic load, relative survivorship in the plants of each
species which set self fruit was calculated as the percent mature seed per self
pollination relative to outcross pollination. The number of lethal equivalents per zygote
(2B) was calculated from the relative survivorship (RS) of those plants based on this

equation modified from Morton, Crow, and Muller (1956): ZB = 2(-1/F)ln(RS). The

24

number of lethal equivalents could not be calculated for the plants that did not set self
fruit because relative survivorship in those plants was zero. Because the inbreeding
coefficient, F, is theoretically equal to 1/2 when a non-inbred, diploid plant is selfed,
we used the equation 2B=-4ln(RS) to calculate lethal equivalents in the diploid V.
myrtilloides (Sorensen, 1969; Seavey and Carter, 1994). For tetraploid V. corymbosum
and V. angustifolium, we used the equation ZB=-121n(RS) because F is theoretically
equal to 1/6 when a non-inbred autotetraploid is selfed (Krebs and Hancock, 1991).
The mean percent mature seed per self and outcross pollination, relative survivorship,
and lethal equivalents of the plants included in this analysis were compared between

the species by t-test.

Results

Self mgutgrgss Pollinations

 

Mean self and outcross fruit set and seed set of the fifteen plants for each
species are shown in Table 2. Almost all of the plants (14/15 in each species)
produced outcross fruit, but less than half of the plants produced self fruit. Seven of
the 15 V. corymbosum plants produced self fruit, and only four of the 15 plants
produced self fruit in both V. angustifolium and V. myrtilloides. Fruit set and seed
set were significantly reduced (p<.05) by self pollination in every species, except the
reduction in self percent mature seed per fruit compared to outcross in V. corymbosum
was not significant (p>.10).

There were no significant differences between the species in outcross fruit set

or seed set. However, there were some significant differences between the species

25

Table 2. The average fruit and seed set after self (S) and outcross (OC)
pollinations on fifteen plants from each of three Vaccim'um species, and the results
of pairwise t-test comparisons between the species.

 

% Mature Seed

 

 

 

 

Fruit Set Per Fruit Per Pollination
S OC S OC S OC
V. corymbosum (c) 14.7 39.0 20.2 31.5 2.3 14.0
V. angustifolium (a) 2.7 34.7 13.0 36.9 0.4 13.8
V. myrtilloides (m) 3.3 50.0 5.6 33.0 0.2 17.5
Comparisons: c‘a * ns ns ns * ns
c‘m * ns ** ns "”" ns
a‘m ns ns ns ns ns ns

 

ns not significant, " p < .10, ‘* p < .05

26

following selfing. Self fruit and seed set were lower in one of the tetraploid species,
V. angustifolium, than in the other tetraploid species, V. corymbosum, although the %
mature seed per fruit was not significantly lower. Self fruit set and seed set were
significantly lower in diploid V. myrtilloides than in tetraploid V. corymbosum, but
there was no significant difference in self fruit set or seed set between the diploid V.
myrtilloides and the tetraploid V. angustifolium.

Some important correlations among fruit and seed set parameters were found.
There was a decrease in the % mature seed in the fruit corresponding with a decrease
in fruit set in V. angustifolium and in V. myrtilloides, but not in V. corymbosum
(Figure 2). Also, the % mature seed per self pollination was generally higher in plants
with higher % mature seed per outcross pollination in all three species, although this
correlation was not significant in V. angustifolium (Figure 3).
Palmgiass

The proportion of fertilized ovules that developed into mature seed (% mature
seed) per pollination resulting from the control and treatment pollinations in the pollen
chase experiments are shown in Table 3. There was no significant difference when
outcross pollen was applied on day 0 (Control 2) or on day 3 (Control 3) in any of the
plants, demonstrating that flower age did not have an effect on the % mature seed.
The number of days between the application of self pollen and the application of
outcross pollen did, however, appear to have an effect on seed set. A one day interval
resulted in seed sets comparable to the outcross controls, while a three day interval

resulted in much lower seed sets.

27

Figure 2. Correlation between fruit set and the % mature seed per fruit in the three

Vaccinr'um species: A. V. corymbosum; B. V. angustifolium; C. V. myrtilloides.

% Fruit Set

% Fruit Set

% Fruit Set

Figure 2.

100 -

 

28

 

 

 

 

 

A. V. corymbosum .
so - o ' r=0.221
p=1.000
60 - o o o
O.
40 - o
. O O Q
20 _. .. . e . o
O.
0 I I I I I I I
0 10 20 30 4O 50 60 7O
80 I B. V. angustifolium
C 0
6° “ ' . r=0.533
0 =0.023
40 -:
C
O O
O
20 - o o
O O
O
O O O
0 I I I I I I I I
100 O 10 20 3O 40 5O 60 7O 80
C. V. myrtilloides .
80 - o ..
O
60 - . r=0.638
p=0.004
4O 4 O .
Q .0 O
20 1 0
O O.
o l l I l I I
O 10 20 3O 40 50 60
%MatureSeed/Fruit

29

Figure 3. Correlation between self and outcross % mature seed per pollination in the
three Vaccim'um species: A. V. corymbosum; B. V. angustifolium; C. V.

m yrtilloides.

Self

Self

Figure 3.

30

 

 

A. V. corymbosum
8 T o
C
6 _.
4 _. ° . r=0.611
p=0.015
2 ._ C
o
0 m I *I I I I I

010 20 30 40 50 60 70

6 ._
B. V. angustifolium
O
4 _.
r=0.413
=0.126
2 a
O O

 

 

Orb-«re—r—ar—Or'fisl

 

 

120 51015 20 25 30 35
' C. V. myrtilloides
1.0 4 o o
0.8 ~
0.6 - r=0.565
=0.028
0.4 - . p
0.2 —
O
0-0 W P I I i I
0 10 20 30 4O 50
Outcross

% Mature Seed per Pollination

31

.8. v a s: .8. vs. ..a .2. v m. assume—.9... .2. a.

 

 

 

 

 

a m: m: a m: m: NHLU Q.
as... a... as... as... a... a... NHLH G
3.... ma .1... an. an m: NHaMU
3.8 a... .3... as... m: m: NHaNU Am
3 m: m: m: m: a: Cinnv
.. m: a: a: m: m: 2.30 3
a: a... as... 9...... as... 3.... FHLU AM
m: m: m: m: a: ma muamv 3
2.2.. a: 3.2.. a... a... m: MOLD
a... a... ..33 .3... a a... NULU . C “mcaflanaoo
S c o as E on at man—.8225 5323
2:. of 3.. a: mom 2: 2b .aoaeaao aaafiom
SN 2 ..a ..2 «.2 ..m :8 35 323.5
3. X 2: ..mm m: on 38 39:83.5
o o o c c o :8 can :3
am: SEE mh<< o—Amx‘. aqmo mqmu
wage—zap. .> E:=o%fi=w§ .> gmonEboo .>

 

.6505 8:. on. .53 :03 8.. 388 8258: can .958 05 .«o 285.923 .8... 8353 .N .5838...
u NH 3 «5:53;. n E. mm .9550 u no “N 3280 H No x 3980 "a .86on :80 me $53 025 E «5832.
23 .823 some .8 aoaecéoa ..oa woo... 8:38 Beacon mike"? 3:08.888 omega 5:2. mo £98m .m 633.

32

In every plant, the % mature seed after a 3 day delay of outcross ”chaser"
pollen (Treatrnent 2) was significantly less than the % mature seed after a 1 day delay
of "chaser" pollen (Treatment 1). The % mature seed after Treatment 1 was actually
higher than the % mature seed after outcross pollen alone (Controls 2 and 3), although
it was only significantly higher (.05<p<.10) than Control 2 in MBR3. Treatment 2
was significantly less than Control 2 in the V. angustifolium and V. myrtilloides plants,
although not in the V. corymbosum plants, and it was significantly less than Control 3
in AAT3 and MBR3, although not in the other plants.

The % mature seed was zero, because there was no fruit set, following the
application of self pollen alone (Control 1) in every plant. Seed set following self
pollination (Control 1) was significantly less than outcross pollination on day 0
(Control 2) in all plants, and was significantly less than outcross pollination on day 3
(Control 3) in all plants except CEL3 and MPR3. Control 1 was significantly lower
than Treatment 1 in every plant, but Treatment 2 was not significantly different than
Control 1, except in ABLlO and MBR3 (.05<p<.10).

Emu—9.0m man

The percent mature seed per self and outcross pollination, the relative
survivorship, and the number of lethal equivalents in the self-fertile plants of each
species are shown in Table 4. Plants that did not produce self fruit were omitted from
the analyses. As with the means of all fifteen plants of a species, outcross % mature
seed was not significantly different between any of the species, and self % mature
seed was lower in V. angustifolium and V. myrtilloides than in V. corymbosum.

Relative survivorship was lower in tetraploid V. angustifolium than in tetraploid V.

33

.88on 2.: 8.. 338 2a 3353an 5 $3832..
3.5.5338 9526b 5»... n ma ”KN ”33303.55 25205 EN? H mm ”Xv.
S. vs. 1.... .3. v m 9.. .2. v m .. sewage? .8 8

 

a.
mu

*§

ma

§§§
a.

m:
an
mG

ma

**
*

are ”33.580

 

0.: 22. .3 to $32

3.. 3o... N... 0.. 352

mom 8o... 3. ..o 952

Ca: 0.: 88.8 «.8... and EN 8.8 0.. Sm:
as “8.23%. .>

.3 m. ..o 3. go 2.5.

a: $3 3: e... Ema.

98 «8... SN 3 039‘

95 Nam 88.8 m8... :15 on 3.: no Sm...
A3 ESL—cmamswna .>

we SN... no. 2 5.8

.2 £3 MEN an 25.0

3.. 83 3. a... 3:8

52 2 ..o n8 3 Emu

«.2 53 w: m: :8

a: So... 3. no Emu

3.»: a... 82.8 at... 8.5 3: ..33 w... 2 do
A3 Samoan—boo .>

38030 hum

.2:2e>3~om 3564

ASP—3.5539
mEfloPSEm 03.20%

 

chem com voom 9:534 .x.

 

$53 3383.3”... N was Eamokzgwau K .Eamooiboo x oibmfiom 5 3:28:93 350— .«o
con—8:: 05 38 3.7.8323 2530.. 05 £022.52— 88830 was .28 con 33. 23.2: «.823 SF .v 033.

34

corymbosum, and the number of lethal equivalents was higher in V. angustifolium.
Relative survivorship was also higher in V. corymbosum than in the diploid V.
myrtilloides, but the number of lethal equivalents was the same in these two species.
Relative survivorship was not significantly different in tetraploid V. angustifolium and
diploid V. myrtilloides, and the number of lethal equivalents was significantly higher

in this tetraploid than in the diploid.

Discussion
Winbreeding dggression

There was a marked effect of self pollination on fertility in these three
Vaccinium species. Fewer self pollinations set fruit than outcross pollinations and a
lower proportion of fertilized ovules developed into mature seed in the selfed fruit than
the outcrossed fruit. Because there was a reduction in both fruit set and seed set per
fruit, the effect of self pollination on fertility in these species is represented most
inclusively as seed set per pollination.

The range in self fertility among the V. corymbosum genotypes supports the
earlier conclusion by Krebs and Hancock (1991) that early-acting inbreeding
depression is responsible for the reduction in self fertility in this species. Variable
expression across individuals is a ”hallmark" of inbreeding depression due to
mutational genetic load (Seavey and Carter, 1994). However, the proportion of self-
fertile individuals (7 of 15) and the average °/o mature seed per self and outcross
pollination (Table 2) among the fifteen individuals from three different populations in

this study is considerably lower than that of the twenty-eight individuals from one

35
population in their study (23 of 28, 19.3 % and 46% respectively). In the other two

species, only 4 of the 15 individuals produced self fruit and the % mature seed per self
pollination among those individuals is low (Table 4).

Other studies have suggested that a level of self fertility this low in a species is
due to variable expression of pre-zygotic self incompatibility (Byers, 1995) or is
indicative of late-acting self incompatibility (Waser and Price, 1991). We believe that
there is evidence for early-acting inbreeding depression in these Vaccinium species.

To begin with, some plants of all three species did produce self fruit. Furthermore,
aborted seed was clearly present in addition to mature seed in the self fruit that was
harvested, indicating that the ovules had been fertilized by the self pollen in those
fruit. If there were a completely functional pre-zygotic or late-acting self
incompatibility system, there would be no self seed set. Also, the aborted seed in both
self and outcross fruit ranged in size and shape (round, flattened, or shriveled) within a
fruit and between plants suggesting that seeds were aborted at different stages of
development. One of the main tenets of late-acting self incompatibility is that
embryos are aborted at a uniform stage early in development (Seavey and Bawa,
1986). If seed abortion is due to inbreeding depression, the seed phenotypes should
vary depending on the deleterious genes being expressed in the progeny (W iens et al.,
1987; Krebs and Hancock, 1990 and 1991; Seavey and Carter, 1994).

Incomplete expression of a self-incompatibility (SI) locus in the self-fertile
individuals and complete expression in the other plants might be an explanation for the
complete lack of self fmit set by most of the plants and low fruit set in the others.

According to Byers (1995), incomplete expression of SI accounts for "weak” self

36

compatibility in Eupatorr'um perfoliatum. If there were ”weak" self compatibility,
fewer self pollen tubes would be expected to reach the ovules. In the present study,
the average total number of fertilized ovules was lower in self fruit than in outcross
fruit in V. myrtilloides and V. corymbosum, but the difference was not significant (data
not included). The total number of fertilized ovules was the same in self and outcross
fruit in V. angustifolium. The slight decrease in the number of fertilized ovules in self
fruit could indeed be due to fewer of the self pollen tubes reaching the ovules. The
results of our pollen chase experiment (discussed below) did indicate that self pollen is
less vigorous than outcross pollen. However, the total number of fertilized ovules
could also be lower in self fruit because more ovules were aborted very early in
development and could not be seen to be scored (Krebs and Hancock, 1988).
Regardless, a significant number of ovules still were fertilized by self pollen and the
proportion of those self-fertilized ovules that developed into mature seed was lower
than outcross fertilized ovules.

Although this does not exclude the possibility that SI was being "completely”
expressed in the plants without self fruit, an alternative explanation is that the fruit did
not develop due to the failure of all of the fertilized ovules in a fruit to develop into
mature seed. Wiens et al. (1987) suggested that the number and vigor of maturing
ovules determines fruit survivorship, and developing fruit which contain more maturing
ovules will out-compete fruit on the same plant with fewer maturing ovules. In
blueberries, fruit generally does not set unless at least one fertilized ovule develops
into a mature seed, and a decrease in fruit set is typically associated with a decrease in

swd set (Hall et al., 1979; Vander Kloet and Hall, 1981; Eck, 1986). In this study,

37

all ripe fruit contained at least one mature seed, although seven V. corymbosum fruit
which had not ripened by 12 weeks post-pollination contained only aborted seeds. We
did find a significant positive correlation between fruit set and the proportion of
fertilized ovules that developed into mature seed in the fruit in V. angustrfoh'um and in
V. myrtilloides (Figure 2), indicating that fruit set is reduced when fewer seeds mature.
This correlation is not significant in V. corymbosum. The lack of a correlation
may be because the stage of development when embryo abortion occurs is not uniform
from plant to plant, depending on which deleterious recessives are present. As we
have already mentioned, this may have made it difficult to score the percent mature
seed per fruit in some plants. In fact, one V. corymbosum plant had low self fruit set
(15%), but one fruit contained only one mature seed and no detectable aborted seeds.
In this plant, embryo abortion may have occurred so early that aborted seeds were too
small to detect, and % mature seed per fruit (50.7%) was overestimated. This was
also reflected in low outcross fruit set (30%) and high seed set (59%) in this plant.
In another plant, self percent seed set was very low (6.1%), but self fruit set was very
high (80%) because it included five of those fruit that were harvested at 12 weeks
post-pollination which contained only aborted seeds. The embryos in those fruit
apparently were far enough along in development to cause fruit swelling, but were
aborted before the fruit could ripen. Outcross percent seed set in that plant was 19%
and fruit set was 85%. Variation in the stage of embryo abortion like this may explain
why fruit set and the % mature seed per fruit did not appear to be correlated.
However, the number of mature seed per fruit, not taking into account aborted seed,

was significantly correlated to fruit set (r=.632, p=0.002).

38

The correlation between self and outcross fertility in these three species
(Figure 3) is further evidence of early-acting inbreeding depression. In general,
individuals with high levels of self fertility had high outcross fertility, individuals with
low levels of self fertility had intermediate outcross fertility, and individuals which
were self-infertile had lowest outcross fertility. Other studies in these species have
also indicated a correlation between self and outcross or open-pollinated fertility
(Wood, 1968; Vander Kloet and Hall, 1981; Krebs and Hancock, 1991). This
correlation suggests that maternal embryonic genetic load is being expressed in both
selfed and outcrossed progeny. If there were a self-incompatibility system, either pre-
zygotic or late-acting, we would expect outcross fertility to be unrelated to self
fertility, even if there were variable expression of self-incompatibility.

The results of the pollen chase experiment also provide evidence that ovules
are fertilized by self pollen in all three species. Because there is no difference in the
% mature seed per pollination between Treatment 1 and outcross pollen alone
(Controls 2 and 3), it appears that outcross pollen fertilized the ovules when it
followed one day after self pollen. However, the significant reduction in the %
mature seed set when outcross pollen wasn't applied until three days after self pollen
indicates that self pollen fertilized the ovules in Treatment 2. Unfortunately, we were
limited by the number of flowers available for pollination, and the small number of
pollinations may explain why some of the differences are unexpectedly significant or
not significant. However, these results are very similar to those reported by Krebs and
Hancock (1991) from a pollen chase experiment in one V. corymbosum clone.

Although self pollen appears to be less vigorous than outcross pollen, it is capable of

39

fertilizing the ovules over time. This sort of "cryptic" self-fertility has been reported
in Amsincla'a grandiflora (Weller and Ornduff, 1989). This species also is subject to
early-acting inbreeding depression, and Weller and Ornduff (1991) suggest that
"cryptic" self-fertility may play a role in maintaining higher levels of genetic load in a
population by promoting outcrossing. It should also be noted that the two V.
corymbosum plants and one of the V. myrtilloides plants (MPR3) in the pollen chase
study were plants that did not produce self fruit previously, which suggests that self
infertility in those plants is due to self seed abortion and not ”complete expression" of
self incompatibility.

It is interesting that, while not significantly so (except in MBRB), the
proportion of fertilized ovules that develop into mature seed in Treatment 1 is higher
than the outcross controls. This increase in seed set may be an artifact due to the
small number of pollinations, although it would be coincidentally so in every plant.
Other explanations should be considered. It may be that there is a "mentoring” effect
of the outcross pollen on the self pollen, such that self pollen tubes grow much faster
in the presence of outcross pollen than they do alone. However, our self and outcross
pollinations demonstrate that fertilization by self pollen results in a lower % mature
seed. If self pollen and outcross pollen both fertilized ovules in Treatment 1, we
would have expected a decrease in the % mature seed rather than an increase. Vander
Kloet and Lyrene (1985) did report a decrease in seed set when self and outcross
pollen were applied at the same time to flowers of V. corymbosum. The other
possibility is that the self pollen is the ”mentor" such that the initial stimulus of

pollination by self pollen enhances conditions for fertilization of ovules by the outcross

40

pollen when it is applied one day later. Vander Kloet (1991) also observed in V.
corymbosum that the application of self pollen 10 hours before outcross pollen
promotes seed set by the outcross pollen, but the effect was diminished when outcross
pollen was not applied until 30 hours after self pollen.

It does appear in these Vaccim'um species that self pollen is ”less compatible"
than outcross pollen, but self pollen is not "incompatible”. Incomplete expression of
SI in the self-fertile individuals does not explain our results, including the reduction in
the proportion of self-fertilized ovules that develop into mature seed, the correlation
between self and outcross fertility, or the reduction in the % mature seed per
pollination when self pollen was applied three days before outcross pollen. Also,
although complete expression of 81 could explain no self fruit set, abortion of all
fertilized ovules is also an explanation. We suggest that a combination of both less
vigorous self pollen tube growth, which may reduce the number of fertilized ovules,
and subsequent abortion of the ovules that are fertilized by self pollen accounts for the
absence of fruit set in some individuals. Interestingly, not only the abortion of
fertilized ovules, but the reduced vigor of self pollen may be explained by a very high
level of genetic load in those individuals. Deleterious recessives carried by the parent
may be expressed in the formation of pollen, and the expression of deleterious
recessives in the haploid gametophyte may also be a possibility.

Estimates piembgonic mm

We calculated the number of lethal equivalents per zygote in the plants of each

species that produced self fruit. Lethal equivalents could not be calculated in plants

that did not set self fruit because the relative survivorship in those plants was zero.

41

For all fertilized ovules to be aborted, the number of lethal equivalents in those plants
may be higher than the highest number we were able to calculate. However, some of
those plants may have had the same low levels of self-fertility as the least self-fertile
plants in this study, but we did not detect it with only 20 pollinations. We assume
that the plants which did not produce fruit carry as many or more lethal equivalents as
the highest numbers we calculated.

The theoretical relationship between the number of lethal equivalents and the
relative survivorship in a diploid and an autotetraploid, based on the equations we
used, is illustrated in Figure 4. Although 20 lethal equivalents is enough to leave a
diploid essentially self-infertile, an autotetraploid may maintain up to 60 lethal
equivalents and still have a low level of self-fertility. In fact, relative survivorship will
always be higher in the tetraploid than the diploid, unless the number of lethal
equivalents in the diploid is both below 20 and is one-third or less than the number of
lethal equivalents in the tetraploid. Generally, we expect inbreeding depression to be
lower, even though genetic load is higher, in the tetraploid than in the diploid.

In the present study, there was less inbreeding depression (relative survivorship
is higher) in the tetraploid V. corymbosum than in the diploid V. myrtilloides. This is
expected in theory if the levels of genetic load are the same in a tetraploid and a
diploid. Lower inbreeding depression was also reported in the tetraploid form of
Epilobium angustifolium than in the diploid form of that same species (Parker,
Nakamura, and Schemske, 1995; Husband and Schemske, 1995). However, inbreeding
depression is the same in the tetraploid V. angustifolium and the diploid V.

myrtilloides, suggesting that genetic load is higher in the tetraploid than the diploid.

42

Figure 4. The predicted relationship of diploid and autotetraploid relative survivorship

to the number of lethal equivalents.

43

1.0 a —— Autotetraploid
———- Diploid

\

l

l

l
l
l

.0
a)
1

0.4 ~

Relative Survivorship

0.2 -4

 

 

0.0 -*

 

Lethal Equivalents

Figure 4.

44

VanderKloet (1991) also reported equal effects of inbreeding in diploid and tetraploid
forms of V. corymbosum. Parker, Nakamura, and Schemske (1995) found more
inbreeding depression (lower relative survivorship) in the tetraploid Epilobium
angustrfolium than in the closely related diploid Epilobium ciliatum. Clearly ploidy
level is not the only factor effecting inbreeding depression in these species.

Ecological and life history differences may account for the levels of genetic
load maintained by these species. If the life histories are similar between the two
ploidy levels, as in the different cytotypes of the same species in Epilobium
angustifolium (Parker, Nakamura, and Schemske, 1995), the effect of ploidy level on
inbreeding depression is more likely to fit expectations. However, there are obvious
differences in life histories between the two species Epilobium angustifolium and
Epilobium ciliatum, reported by Parker, Nakamura, and Schemske (1995), which
would accormt for a considerably higher level of genetic load in the tetraploid than in
the diploid. The diploid Epilobr’um ciliatum is an herbaceous annual with high levels
of autogamy and the tetraploid Epilobium angustifolium is a rhizomatous clonal
perennial with low levels of autogamy. Inbreeding depression is typically higher in
perennials than in annuals and in outcrossers than in selfers. A self-fertile annual
eliminates more deleterious recessives through purging than does an outcrossing
perennial (Wiens, 1984; Wiens et al., 1987; Husband and Schemske, 1995). Also,
the accumulation of spontaneous mutations may be higher in clonal species than
nonclonal species. Clonally reproducing species have no means to purge genetic load,
which requires sexual reproduction for the expression of deleterious recessives

(Klekowski, Mohr, and Kazarinova-Fukshansky, 1986; Klekowski, 1988).

45

Vaccinium angustifolium and V. myrtilloides are both rhizomatous, clonal
species like Epilobium angustifolium. This may explain in part thehigher level of
genetic load in V. angustifolium than the other tetraploid, V. corymbosum, and a level
of genetic load in the diploid V. myrtilloides as high as the tetraploid V. corymbosum.
Also, the lower level of genetic load in V. corymbosum compared to V. angustifolium
may be accounted for by a higher level of self fertilization in V. corymbosum.
Although Vaccim‘um species reportedly should not self spontaneously due to floral
structure and protandry, at least one species (V. uliginosum) has been found to be
largely self-pollinated, and the pollen-to-ovule ratio of many of the species suggest
moderate autogamy (V ander Kloet, 1988). In addition, cultivated V. corymbosum in
monoculture produces fruit with varying levels of seed set depending on the cultivar
planted (Krebs and Hancock, 1988; Hancock et al., 1989). The frequency of self
pollination in natural populations would be influenced by many factors, including
population size and density, size and dispersal of clonal propagules, and pollinator
activity and availability. No work to directly assess the rate of natural selfing in the
Vaccinium species has been done. However, only a small amount of selfing may be
necessary for deleterious recessives to be purged from the populations (Lande and
Schemske, 1985; Charlesworth and Charlesworth, 1987).

Although life history traits like these could account for different levels of
genetic load in these species, another possibility to consider is that the number of
lethal equivalents calculated based on relative survivorship does not reflect only
mutational genetic load. Relative survivorship, expressed as self fertility relative to

outcross fertility, assumes that outcross fertility represents maximum fertility and that

46

the reduction in fertility is due only to mutational load. Burbidge and James (1991)
make the point that if deleterious mutations are being expressed in the homozygous
condition following outcrossing as well as selfing, then outcross fertility is less than
maximum; relative survivorship is higher than it would be and the number of lethal
equivalents calculated is an underestimate. In the same manner, if other factors are
affecting self-fertility in addition to mutational load, the effect of mutational load is
overestimated; relative survivorship is lower than it would be and the number of lethal
equivalents calculated is an overestimate.

Burbidge and James (1991) suggest an alternative calculation of the number of
lethal equivalents (LE) that does not rely on relative survivorship: S=(x)LB; LE=log
S/log (x), where S is the proportion of ovules that develop into mature seed following
selfing and x is the proportion of offspring that are not homozygous recessive
following selfing at a single locus. By definition, x is equal to 3/4 or 0.75 in a diploid
and to 5/6 or 0.83 in an autotetraploid. Burbidge and James (1991) found that
estimates of LB using either calculation were generally similar in the Stylidium species
they were studying. We calculated LE in this alternative way for our self fruitful
plants and found that the average number of lethal equivalents in V. corymbosum and
V. myrtilloides (17.5 and 18.8, respectively) were close to the average calculated using
relative survivorship (18.6 and 15.7 in Table 4). However, in V. angustifolium the
number of lethal equivalents calculated by the alternative method was not higher, but
considerably lower (26.1) than when calculated using relative survivorship (37.7).

Possibly, other factors in addition to mutational load are effecting self fertility

in V. angustifolium. It may be that a higher segregational load is carried by this

47

species than the others. In diploids, segregational load is expected to contribute
equally to self and outcross offspring, and thus would not effect the estimate of
mutational load (Morton, Crow, and Muller, 1956). However, in the autotetraploid
Medicago sativa (alfalfa), inbreeding depression in controlled crosses was more severe
than expected based only on the approach to homozygosity, and the loss of inter-allelic
interactions has traditionally been the explanation (Busbice and Wilsie, 1966; Bennett,
1976). More favorable combinations of alleles may have accumulated in V.
angustrfoh‘um than in V. corymbosum. Also, Bingham et al. (1994) recently proposed
that there is a greater potential in autotetraploids than in diploids for complementary
gene action between dominant alleles at heterozygous loci, which complement each
other by masking deleterious recessives that affect the same trait. Perhaps fewer
complementary dominant alleles in V. angustifolium is the reason inbreeding
depression is more severe in this tetraploid.

Charlesworth and Charlesworth (1987) caution against extrapolating the number
of lethal equivalents to the actual number of mutations. They generally rely on
relative survivorship as an indication of genetic load. However, taking into account
the genetic differences between diploids and autotetraploids, including the approach to
homozygosity, the mutational and segregational load carrying capacity, and the
potential for complementary gene action, relative survivorship clearly does not
sufficiently reflect the level of genetic load. Charlesworth and Charlesworth (1987)
suggest that the number of lethal equivalents is a useful measure of the "magnitude” of
inbreeding depression. We suggest that the number of lethal equivalents is a measure

of the ”magnitude” of genetic load, although not necessarily the number of mutations.

48

Conclusion

Inbreeding depression is a factor contributing to the low levels of self-fertility
in these Vaccim'um species. Regardless of estimation technique, genetic load in all
three species in this study is relatively high. In the appendix of their paper, Waser and
Price (1991) argue strongly that self-sterility in Ipomopsis aggregate can not be due to
early-acting inbreeding depression although both reviewers of their paper had
suggested it. In that species, 1 percent of self-fertilized ovules develop into mature
seed. They argue that the number of lethal equivalents that an individual must carry
in order for only 1 % of selfed offspring to survive is impossibly high. The number
they calculate is S=(.75)“3; LE=16. However, most studies which convincingly
demonstrate early-acting inbreeding depression in a species, particularly long-lived
woody perennials, report individuals of that species which do carry 16 lethal
equivalents or more. Two of the individuals of Epilobium obcordatum reported by
Seavey and Carter (1994) had more than 16 lethal equivalents, as did five of the
Psuedotsuga menziesir‘ individuals reported by Sorensen (1969). Franklin (1972) found
six Pinus taeda trees with greater than 16 lethal equivalents. Two of the V.
corymbosum clones reported by Krebs and Hancock (1991) had more than 16 lethal
equivalents. Burbidge and James (1991) suggest that an average plant among the
perennial species of Stylidium may be heterozygous for up to 20 lethal equivalents and
populations of some species of Sty/(diam may carry lethals at 100 different loci.

We discuss or mention many factors which could influence the level of genetic
load: ploidy level in particular, but also population size, population density, pollinator

activity, reproductive system, and mating system. Generally, the same traits which

49

are thought to influence the level of heterozygosity and genetic diversity in a
population (Loveless and Hamrick, 1984; Hamrick and Godt, 1989) can influence
genetic load. It should not be surprising that, like genetic variation, the levels of
genetic load vary among populations and among species and can be very high in

some. As more examples like ours of early-acting inbreeding depression and more
inclusive studies like those by Wiens (1984), Wiens et al. (1987), and Burbidge and
James (1991) appear in the literature, it becomes increasingly apparent that genetic
load is higher and contributing more to self infertility in plants than has generally been

recognized.

50

Literature Cited
Aalders, LE. and Hall, IV. 1961. Pollen incompatibility and fruit set in lowbush

blueberries. Canadian Journal of Genetics and Cytology 3(3):300-307.

Bennett, J.H. 1976. Expectations for inbreeding depression on self-fertilization of

tetraploids. Biometrics 32:449-452.

Bingham, E.T., RW. Groose, DR Woodfield, and K.K. Kidwell. 1994.
Complementary gene interactions in alfalfa are greater in autotetraploids than in

diploids. Crop Science 34(4):823-829.

Broyles, SB. and R. Wyatt. 1993. The consequences of self-pollination in Asclepias

exaltata, a self-incompatible milkweed. American Journal of Botany 80:41-44.

Burbidge, AH. and SH. James. 1991. Post-zygotic seed abortion in the genetic system

of Stylidium (Angiosperrnae:Stylidiaceae). Journal of Heredity 82:319-328.

Busbice, TH. and GP. Wilsie. 1966. Inbreeding depression and heterosis in

autotetraploids with application to Medicago sativa L. Euphytica 15:52-67.

Byers, D.L. 1995. Pollen quantity and quality as explanations for low seed set in small
populations exemplified by Eupaton‘um (Asteraceae). American Journal of

Botany 82(8):]000-1006.

51

Charlesworth, D. 1989. Why do plants produce so many more ovules than seeds?

Nature 338:21-22.

Charlesworth, D. and B. Charlesworth. 1987. Inbreeding depression and its
evolutionary consequences. Annual Review of Ecology and Sytematics 18:237-

268.

Eck, P. 1986. Blueberry. In: QRQ Iiandbookof Fruit Set andDeyelopment. Shaul

P.Monse1ise, Ed. CRC Press, Inc, Boca Raton, Florida p75—84.

Franklin, EC. 1972. Genetic load in loblolly pines. American Naturalist 106:262-265.

Haldane, J.B.S. 1930. Theoretical genetics of autopolyploids. Journal of Genetics.

22:359-372.

Hall, I.V., LE. Aalders, N.L. Nickerson, and SP. VanderKloet. 1979. The biological
flora of Canada. 1. Vaccinium angustrfolium Ait., sweet lowbush blueberry.

Canadian Field Naturalist 93(4):415-430.

Hamrick, 1L. and M.J.W. Godt. 1989. Allozyme diversity in plant species. In: Brown,
A.H.D., M.T. Clegg, A.L. Kahler, and BS. Weir, Eds. Plant Population
Genetics, Breeding, and Genetic Resources. Sinauer Associates, Inc.,

Sunderland, MA.

52
Hancock, J.F., S.L. Krebs, M. Sakin, and TR Holtsford. 1989. Increasing blueberry

yields through mixed variety plantings. Michigan State Horticultural Society

Annual Report 119:130-133.

Hokanson, KB. and J.F. Hancock. 1993. The common lowbush blueberry, Vaccinium
angustifolium Aiton may be an autOpolyploid. Canadian Journal of Plant

Science 73:889-891.

Husband, BC. and D.W. Schemske. 1995. Magnitude and timing of inbreeding
depression in a diploid population of Epilobium angustifolium (Onagraceae).

Heredity 75:206-21 5 .

Kenrick, J.V. V. Kaul, and E.G. Williams. 1986. Self-incompatibility in Acacia

retinodes: site of pollen tube arrest is the nucellus. Planta 169:245-250.

Klekowski, E.J. 1988. Mutation, Developmental Selection, and Plant Evolution.

Columbia University Press, NY.

Klekowski, E.J., 1H. Mohr, and N. Kazarinova-Fultshansky. 1986. Mutation, apical
meristems, and developmental selection in plants. In: Gustafson, J.P., G.L.
Stebbins, and E.J. Ayala, Eds. Genetics, Development, and Evolution. Plenum

Press, NY.

53

Krebs, S.L. and J.F. Hancock. 1988. The consequences of inbreeding on fertility in
Vaccinium corymbosum L. Journal of the American Society for Horticultural

Science. 113(6):914-918.

Krebs, S.L. and J.F. Hancock. 1989 Tetrasomic inheritance of isoenzyme markers in

the highbush blueberry, Vaccim‘um corymbosum. Heredity 63:11-18.

Krebs, S.L. and J.F. Hancock. 1990. Early-acting inbreeding depression and
reproductive success in the hi ghbush blueberry, Vaccinr‘um corymbosum L.

Theoretical and Applied Genetics 79:825-832.

Krebs, S.L. and J.F. Hancock. 1991. Embryonic genetic load in the highbush
blueberry, Vaccinium corymbosum (Ericaceae). American Journal of Botany

78(10): 1427-143 7.

Lande, R. and D.W. Schemske. 1985. The evolution of self-fertilization and inbreeding

depression in plants. I. Genetic Models. Evolution 39(1):24-40.

Lewis, D. 1979. Sexual incompatibility in plants. Studies in biology No. 10. Edward

Arnold, London, pp. 29-31.

Loveless, MD. and J.L. Hamrick. 1984. Ecological determinants of genetic structure

in plant populations. Annual Review of Ecology and Systematics 15 :65-95.

54

Morton, N.E., J.F. Crow, and H.J. Muller. 1956. An estimate of mutational damage in
man from consanguineous marriages. Proceedings of the National Academy of

Sciences USA 42:855-863.

Ott, L. 1988. An Introduction to Statistical Methods and Data Analysis, 3rd ed. PWS-

Kent Publishing Co. Boston, MA pp.l70-l79.

Parker, I.M., RP. Nakamura, and D.W. Schemske. 1995. Reproductive allocation and
the fitness consequences of selfing in two sympatric species of Epilobium
(Onagraceae) with contrasting mating systems. American Journal of Botany

82(8): 1007-1016.

Seavey, SR and KS. Bawa. 1986. Late-acting self incompatibility in angiosperms.

Botanical Review 52:195-219.

Seavey, SR and SK. Carter. 1994. Self-sterility in Epilobium obcondatwn

(Onagraceae). American Journal of Botany 81(3):331-338.

Sorensen, F. 1969. Embryonic genetic load in coastal douglas fir, Pseudostuga

menziesii var. menziesii. American Naturalist 103:389-398.

Vander Kloet, SP. 1988. The genus Vaccim'um in North America. Research Branch,

Agriculture Canada, Publication No. 1828.

55

Vander Kloet, SP. 1991. The consequences of mixed pollination on seed set in

Vaccim‘um corymbosum. Canadian Journal of Botany 69:2448-2454.

Vander Kloet, SP. and IV. Hall. 1981. The biological flora of Canada. 2. V.
myrtilloides Michx, the velvet leaf blueberry. Canadian Field-Naturalist 95:329-

345.

Vander Kloet, SP. and PM. Lyrene. 1985. Self-incompatibility in diploid, tetraploid,
and hexaploid Vaccim'um corymbosum. Canadian Journal of Botany 65:660-

665.

Waser, NM. and M.V. Price. 1991. Reproductive costs of self-pollination in Ipomopsis
aggregata (Polemoniaceae): are ovules usurped? American Journal of Botany

78:1036-1043.

Weller, 8G. and Omduff, R. 1989. Incompatibility in Amsinckia grandiflora
(Boraginaceae): distribution of callose plugs and pollen tubes following inter-

and intramorph crosses. American Journal of Botany 76:277-282.

Weller, 8G. and Ornduff, R. 1991. Pollen tube growth and inbreeding depression in
Amsinckr‘a grandiflora (Boraginaceae). American Journal of Botany. 78(6):801-

804.

56

Wiens, D. 1984. Ovule survivorship, brood size, life history, breeding systems, and

reproductive success in plants. Oecologia 64:47-53.

Wiens, D., C.L. Calvin, C.A. Wilson, C.I. Davern, D. Frank, and SR. Seavey. 1987.
Reproductive success, spontaneous embryo abortion, and genetic load in

flowering plants. Oecologia 71:501-509.

Wiens, D., D.L. Nickrent, C.I. Davem, C.L. Calvin, and NJ. Vivrette. 1989.
Developmental failure and loss of reproductive capacity in the rare

paleoendemic shrub Dedeckera eurekensr's. Nature 338:65-67.

Wood, G.W. 1968. Self-fertility in the lowbush blueberry. Canadian Journal of Plant

Science 48:431-433.

CHAPTER 2

GENETIC VARIATION AND THE EXPRESSION OF GENETIC LOAD

IN NATURAL POPULATIONS OF TWO TETRAPLOID

AND ONE DIPLOID VACCINIUIM SPECIES.

57

58

Abstract

To investigate the reasons for different levels of inbreeding depression in three
Vaccinium species, isozyme variation and open-pollinated seed set in the species were
compared. The number of alleles and the level of heterozygosity, averaged over seven
polymorphic isozyme loci, were lower in the diploid Vaccim'um myrtilloides (2.9;
21.7%) than in two closely related tetraploid species, V. angustifolium (3.4; 57.1%)
and V. corymbosum (3.6; 75.6%). The percentage of polyploid individuals with three
alleles at a single locus (tri-allelic) was 3.0% in V. angustifolium and 7.4% in V.
corymbosum. The proportion of fertilized ovules that developed into mature seed
(percent mature seed) in open-pollinated fruit was highest in V. corymbosum (44.6%),
intermediate in V. mym'llor'des (38.3%), and lowest in V. angustifolium (27.7%).
Based on the levels of inbreeding depression previously observed in these species, the
lower percent mature seed in V. angustifolium was unexpected. Geitonogamous self
pollination was considered as an explanation for the lower percent mature seed set in
V. angustifolium, and for the lower level of heterozygosity in V. angustifolium than V.
corymbosum. The effect of tetrasomic inheritance and the level of heterozygosity on

the expression of genetic load are discussed.

59

Introduction

It is generally expected that tetraploid species will maintain higher levels of
genetic variation than diploid species and, consequently, inbreeding depression should
be less severe in the tetraploid (Haldane, 1930; Bennett, 1976; Thompson and
Lumaret,1992; Bever and Felber, 1992). A number of electrophoretic studies have
compared genetic variation in tetraploids to diploids and most have demonstrated an
increase in the number of alleles per locus, polymorphic loci, and the level of
heterozygosity in tetraploids (Roose and Gottlieb, 1976; Ness, Soltis, and Soltis, 1989;
Soltis and Soltis, 1989; Bayer, 1989; Lumaret and Barrientos, 1990). Exceptions to
this trend have also been reported, but mainly in allopolyploids (Cai, Macdonald, and
Chinnappa, 1990; Watson, Elisens and Estes, 1991; Purdy and Bayer, 1995). While
levels of genetic variation have often been compared between ploidy levels, relatively
few studies have compared inbreeding depression in natural populations of tetraploids
and related diploids (Hokanson and Hancock, in prep; Parker, Nakamura, and
Schemske, 1995).

In a recent greenhouse study of self and outcross seed set on plants from
natural populations, Hokanson and Hancock (in prep.) did compare inbreeding
depression in related tetraploid and diploid Vaccim'um species. In that study, it was
shown that there was less inbreeding depression in one tetraploid, V. corymbosum,
than in the diploid, V. myrtilloides, although the level of genetic load was nearly the
same in these two species. In another tetraploid, V. angustifolium, inbreeding
depression was the same as the diploid and the level of genetic load was considerably

higher. This was contrary to the expectation that inbreeding depression would be

60

more severe in the diploid than in the tetraploids despite a higher level of genetic load
in the tetraploids. That expectation is based on the theoretically slower approach to
homozygosity following selfing in an autotetraploid compared to a diploid.

Two possible explanations for the different amounts of inbreeding depression
were proposed by Hokanson and Hancock (in prep). One possibility is that potential
differences in life histories between these Vaccinium species, particularly reproductive
strategies, may maintain different levels of genetic load by regulating the amount of
load that is purged. For example, species with clonal reproduction presumably have a
reduced capacity to purge genetic load, because the homozygous expression of
deleterious alleles requires sexual reproduction (Klekowski, 1988). Alternatively, self
fertilization acts to eliminate more deleterious recessives in a population through
purging than does outcrossing (Lande and Schemske, 1985; Charlesworth and
Charlesworth, 1987). Presently, little is known about the extent of clonal reproduction
or the rate of self fertilization in natural populations of these three species.

The second possibility is that the complex genetics of the autotetraploid
Species may account for greater inbreeding depression in V. angustifolium than
expected. In autotetraploid Medicago sativa (alfalfa), inbreeding depression also
exceeded expectations based on the approach to homozygosity. The traditional
explanation has been that the potential loss, due to inbreeding, of favorable
interactions among alleles at a single locus in autotetraploids contributes significantly
to inbreeding depression (Busbice and Wilsie, 1966; Dunbier and Bingham, 1975;
Groose et al., 1989). An alternative explanation for the inbreeding depression in M.

sativa, however, was recently proposed by Bingham et al. (1994). Their theory

61

suggests that complimentary gene action (additive and epistatic) among linked loci can
explain the progressive heterosis and the unexpected severity of inbreeding depression
in autotetraploids.

In this study, the reason for the different levels of inbreeding depression
between the species are investigated. Isozyme variation and open-pollinated seed set
are compared between the species, to provide information about levels of
heterozygosity and levels of inbreeding. Comparisons are made between the diploid

and the tetraploid species, and between the two tetraploid species.

Materials and Methods

Isogme Analysis

Isozyme analysis was used to assess levels of heterozygosity and genetic
variation in the three Vaccinium species, V. myrtilloides, V. angustifolium, and V.
corymbosum. Three populations of each species were included in the analysis, and 30
plants were sampled from each population. Descriptions of the species and the
population locations are listed in Hokanson and Hancock (in prep.) Dormant flower
buds were collected from each population of V. myrtilloides and V. angustifolium in
September of 1992, and expanding buds were collected from each population of V.
corymbosum in March of 1995. Buds were transported in plastic bags on ice to the
laboratory. The dormant buds were stored at 4°C for approximately two weeks until
grinding. The expanding buds were ground the day after collection. Buds were also
sampled from up to ten plants from each pepulation that had previously been

transplanted into a greenhouse at Michigan State University (Hokanson and Hancock,

62

in prep). Therefore, a total of 30-40 plants from each population, or 90-120 plants
from each species, were sampled.

Six enzyme systems were used in this study: malate dehydrogenase (MDH), 6-
phosphogluconate dehydrogenase (6PGDH), shikimate dehydrogenase (SKDH),
phosphoglucoisomerase (PGI), phosphoglucomutase (PGM), and alcohol
dehydrogenase (ADH). The extraction and staining protocols (except for ADH), as
well as the gel and electrode buffers are the same as those reported in Krebs and
Hancock (1989). ADH was resolved on the lithium borate, tris-citrate system, pH 8.3
of Scandalios (1969) and stained with the following solution: 100 m1 .1 M Tris-HCl
buffer pH 8.0, 3 ml 95% EtOH, 50 mg NAD, 20 mg MTI‘, 5 mg PMS. Seven
polymorphic loci were scored for the number of alleles per locus, the number of
heterozygous individuals (including di-allelic and tri-allelic heterozygotes), and the
number of tri-allelic individuals. Loci and alleles within loci were designated with
numbers starting with number one as the locus or allele migrating farthest from the
origin. Mdh-l, Mdh-Z, 6pgdh-2, Pgi-Z, Pgm-2, and Adh-2 were previously shown to
segregate as Mendelian loci (Krebs and Hancock, 1989; Breuderle and Vorsa, 1994).
SKDH, a putative Mendelian locus, was scored for what appeared to be two allelic
bands, each with a second ”shadow" band (Figures 1 and 2).

At the monomeric enzyme loci (Pgm-2 and SKDH), heterozygous diploid
individuals (ab) had two bands which stained with equal intensity (no dosage), and di-
allelic heterozygous tetraploid individuals (abbb, aabb, abbb) typically had two bands
which stained with varying intensity depending on the number (or dosage) of each

allele (Figures 1-3). Tri-allelic heterozygous tetraploid individuals (i.e.,aabc) had a

63

Figures 1-6. Five enzyme loci in diploid and tetraploid Vaccim'um, illustrating the
banding patterns of different genotypes. Loci are designated by the numbers on
the left, with superscript letters indicating alleles. Figure l. Monomeric SKDH
in a diploid. Lane A is homozygous for allele a (aa), with a ”shadow" band
below it; Lane B is a heterozygote with the genotype ab, and with "shadow"
bands below both allelic bands. Figure 2. Monomeric SKDH in a tetraploid.
Lane A is homozygous for allele b (bbbb), with a "shadow" band below it;
Lanes B and C are di-allelic heterozygotes with the genotypes aabb and aaab,
respectively. ”Shadow" bands are below both allelic bands. Figure 3.
Monomeric Pgm-2 in a tetraploid. Lane A is homozygous for allele b (bbbb);
Lanes B and C are di-allelic heterozygotes with the genotypes bbcc and abbb;
Lane D is a tri-allelic heterozygote with the genotype aabc. Figure 4.
Dimeric 6pgdh-2 in a tetraploid. Lane A is homozygous for allele b (bbbb);
Lanes B, C, and D are di-allelic heterozygotes with the genotypes aabb, aaab,
and abbb, respectively. Heterodimeric bands are in between the allelic bands.
Figure 5. Dimeric Mdh-l and Mdh-2 in a diploid. Mdh-l: Lanes A and B
are homozygous for allele a (aaaa) and b (bbbb), respectively; Lane C is a
heterozygote with the genotype ab, and with a heterodimeric band in between
the two allelic bands. Mdh-2: All lanes are homozygous for allele c (cc).
Figure 6. Dimeric Mdh-l and Mdh-2 in a tetraploid. Mdh-l: Lane B is
homozygous for allele a (aaaa); Lanes A,C,D, and E are di-allelic
heterozygotes with the genotype aaab, with heterodimeric bands in between the
allelic bands. Mdh-2: Lane A is homozygous for allele c (cccc); Lanes B,C,
and D are di-allelic heterozygotes with the genotypes aacc, accc, and bccc,
respectively, with heterodimeric bands in between the two allelic bands; Lane
D is a tri-allelic heterozygote with the genotype abcc, with a heterodimeric
band in between each pair of allelic bands.

 

1. SKDH-2x 2. SKDH-4x
B A B A C
3. PGM-4x 4. 6PGDH-4x
2a- ~00 .-
b- , , - a . a-
i - - ’ ‘8 8 t 2
2b-
A D C B
5. MDH-2x

 

 

12““

zeii 21:2:

ABC DACEB

 

 

Figures 1-6.

65

band representing each of the three alleles, with varying intensity depending on dosage
(Figure 3).

At the dimeric loci (6pgdh-2, Mdh-l, Mdh-Z, Pgi-Z, Adh-Z), heterozygous
diploid individuals had three bands: one representing each allele (homodimers) which
stained with equal intensity and one heterodimer in between these which stained
asdark or darker than the homodimers (Figure 5). Di-allelic heterozygous tetraploid
individuals also had two homodimers and one heterodimer in between them, but with
various staining intensity depending on dosage (Figure 4 and 6). Tri-allelic
heterozygous tetraploid individuals had three homodimeric bands with heterodimeric
bands between each pair of allelic bands, and varying dosage (Figure 6).

The number of alleles per locus and the number of different alleles in an
individual at a locus were easily scored. However, the number of copies of each allele
(dosage) based on staining intensity was not always clear. Allele frequencies were
only determined at the Mdh-2 locus and the 6pgdh-2 locus because these were the loci
where dosage could be reliably scored in both of the tetraploids (Figures 4 and 6).
These allele frequencies (Table 1) were used to calculate the Hardy-Weinberg expected
heterozygosity (Hap) assuming random mating among individuals within populations

(Table 2). The observed heterozygosity (H ) for those two loci was calculated in

obs
each population as the number of heterozygous individuals divided by the total number
sampled. Differences between observed and expected heterozygosity were tested by

chi-square analysis.

66

._ 332.5 5 ~ 032. com dorm—soon coo :233235

 

 

 

 

ovoo ”who 35o mmoo memo ooo.o ooog ooo._ ooog m

ooo.o mmoo «moo Nvoo «moo mwoo ooo.o ooo.o ooo.o N

ooo.o mw_.o emao mmoo momo ooo.o ooo.o ooo.o ooo.o _ «-392

ooo.o ooo.o ooo.o mmoo ooo.o ooo.o ooo.o ooo.o ooo.o o

ooo.o :oo ooo.o omoo ooo.o ooo.o ooo.o ooo.o ooo.o m

ego mwmo Ebo ooo.o wood Nooo mwoo homo ooo._ v

ooo.o ooo.o ooo.o ooo.o ooo.o ooo.o Zoo ooo.o ooo.o m

vamo momo tho mmmo momo oomo ooo.o Soo ooo.o N

ooo.o ooo.o ooo.o ooo.o ooo.o ooo.o ooo.o ooo.o ooo.o _ min—0mm
222 ..3qu

A00 EU .50 .53. .52 qm< ME: mm: "Am:

Eamonfiboo .> E::o~5:m§ .> mega—«Eb: .>

 

.86on 55535. 025 we mecca—anon fimEomE E 62 Balance 95 3 mflocosoot 22?. .~ 053.

67

Table 2. Equations used to calculate expected heterozygosity at diploid and
tetraploid loci in populations with two alleles (p,q), three alleles (p,q,r), and four
alleles (p,q,r,s).

 

Diploid
Two alleles: 2 pq

Three alleles: 2 pq + 2 qr+ 2 qs

Tetraploid
Two alleles: 4p3q + 4pq3 + bpzq2

Three alleles: 4p3q + 4pq3 + 4p3r + 4pr3 + 4q3r + 4qr3 + 6p2q2 + 6p2r2 + 6q213 + 12p2qr +
lqu’r + 12pqr2

Four alleles: 4p3q + 4pq3 + 4p3r + 4pr’ + 41338 + 4ps’ + 4q3r + 4qr’ + 4q’s + 4qs3 + 41.33 +
4rs3 + 6p2q2 + 6p2r1 + 6pzs2 + 6qu’ + 6qu2 + 6rzs2 + 12p2qr + lqu’r +
1213912 + 12p2qs + 12pqzs + 12pqs2 + 12p2rs + 12przs + 12prs2 + qu’rs +
qur‘s + 12qr32 + 24pqrs

A

68
Seed §e_t

 

To determine the % mature seed in open-pollinated fruit in each species, fruit
were collected from each of the three populations of V. myrtilloides and V.
angustifolium in the summer of 1993, and of V. corymbosum in the summer of 1994.
Ten ripe fruit were collected from each of 20 plants in each population, except the V.
angustifolium Adam's Trail population and the V. myrtilloides Poe Reef population,
where fruit were collected from only ten plants. Fruit were transported in plastic bags
on ice to the laboratory, where they were stored at 4°C. Seeds were removed from
fruit, scored as mature or aborted, and counted (Hokanson and Hancock, in prep). The
total number of fertilized ovules was calculated as the sum of mature and aborted
seeds, and the percent mature seed per fruit was calculated as the number of mature
seed divided by the total number of fertilized ovules x 100. A one-way ANOVA was
used to test for significant differences in the total number of fertilized ovules, the
number of mature seed, and the percent mature seed between populations within

species and between species. Population mean comparisons were tested by LSDWS).

Results
The number of individuals scored, the number of alleles per locus, the percent
heterozygous individuals per locus, and the percent tri-allelic individuals per locus at
each of the seven loci are shown for each species in Table 3. The number of
individuals scored per population varies depending on the resolution of the isozymes
on the gels. In every species, the highest number of alleles was found at the Pgi-2

locus. In V. myrtilloides and V. corymbosum, the highest level of heterozygosity was

69

 

 

E 92. an 93 3 Zn 3 ma - SN mN saw so:
3 Na N g 3 on. N a: - Ne N a :8;
N. QNN s a 3 :N N 8 - a... N N2 3:?
a: NNN N E a: 2a m NN - com N 8 N32
2 on n N” f 4.8 a N: - as e 3 32
o.» New N am 2 3a m a - E. z 8 Non:
MN “.3 a. 3 S 3; m :w - a? N 8 ES:
3. o: a. S i. a: a a . 3 m a» Nicene

£235 230m E22 .05 £235 ossm 332 as 0:215 each £22. .5 253
.x. as a a a as a a a .x. a a

 

amenaboo .>

5232,35 .>

3205:»:— .>

 

360% :03 mo mecca—soon 085

Echo min—.365 $88 Bmfiog 8n moBm> 360% 85585. 085 E moo. Seance :38 «a 5:33» .m 633.

70

also at the Pgi-Z locus, although the percent heterozygous individuals in V.
angustifolium was higher at 6pgdh-2 and Mdh-l than at Pgi-Z. The Mdh-Z locus was
monomorphic in all the diploid individuals sampled, but was polymorphic in the
tetraploids. The number of alleles per locus was not noticeably lower in the diploid
than in the tetraploids at any of the other loci. However, at every locus,
heterozygosity was higher in the tetraploids than in the diploid. Heterozygosity was
also higher at every locus in the tetraploid, V. corymbosum, than in the other
tetraploid, V. angustifolium, except at the Mdh-l locus. No tri-allelic individuals
were observed at the SKDH locus in either of the tetraploid species, but the percent
individuals that were tri-allelic at the other loci was also higher in V. corymbosum than
in V. angustifolium, except at the Mdh-l locus.

The data on seed set within species are shown in Table 4. The total number of
fertilized ovules was significantly higher in the diploid, V. myrtilloides, than in either
of the tetraploids. Seed set, expressed as either the number of mature seed per fruit or
the % mature seed per fruit, was significantly different between all of the species. It
was lowest in the tetraploid V. angustifolium, highest in the tetraploid V. corymbosum,
and intermediate in the diploid V. myrtilloides. The number of alleles and the percent
heterozygous individuals were higher in the tetraploids than in the diploid. Also, the
percent heterozygous individuals and percent tri-allelic individuals were higher in V.
corymbosum than in V. angustifolium.

Genetic variation and seed set data are summarized for each of the three
populations of each species in Table 5. Among the V. myrtilloides populations, the

total number of fertilized ovules was significantly lower in the Poe Reef population

71

Table 4. Mean seed set in three Vaccim'um species. Values are averaged
across individuals from three populations of each species. Lower case letters
represent mean separations between species (LSDOS).

 

 

Total
Fertilized Ovules # Mature Seed % Mature Seed
V. myrtilloides 67.1 a 25.4 a 38.3 b
V. angustifolium 37.5 b 10.3 c 27.7 e

V. corymbosum 39.4 b 17.8 b 44.6 a

 

72

.an 28: e5 330% e ”was;
.62 0go£ :05» :0 $060 Bwfigf

 

 

N... 5. 3 «3 N a New a 3N a 0:. 23 e5
SN. 5. 3: SN 2 a 3 a n: a Na. seam topaz
S. NE. a.“ :0 N a EN 0 0.: a 8% can 5
£53558 .\_
NE. a. 3 30 on a EN a mm a can 3... is?
NNe. £0 no 9% am a «R a 2: a 3m 003— Banana
:a. 33. «N 0% am a SM n 2: a 3; one seem
Eaaobnmamnu x
to. :o. - EN N a at. a SN a a: .80. 8m
as. 2:. - n: ..N a QNM 3 SN a one 82¢ €885
o o - anN «N a N3 a NaN a BR 23 3&8
6.038355 .x.
"to me: .aefiep has: .83.? Bom saw 835 Sofieaoa
s .x. a 85.5 Ban: 83.an ea 88%

a .x. a 38.

 

Anon—mic 860% 55;» £830.38 5030p 3% 6m $000 05
com 083838 38080.. $032 008 633 dog—38 600 8a 850% 03 m0£mommnoc203 “006090 98 32030
360% 5550.03» 085 me mecca—zoom £55 mag—036$ 8060 @0363 comet? 0868‘ Ba .8 30m .m 033.

73

than the other two populations. The number of mature seeds was lowest in the
Blueberry Ridge population, although it was only significantly lower than the Beaver
Lake population. However, the percent mature seed is significantly lower in the
Blueberry Ridge population than in either of the other two populations. The number
of alleles per locus was not different between the populations, but the percent
heterozygous individuals was also lower in the Blueberry Ridge population than in the
other two populations. There was no significant difference in observed and expected
heterozygosity in the three populations.

There were no significant differences between the three V. angustifolium
populations in any of the seed set parameters. Genetic variation was generally similar
in the three populations, although the % tri-allelics was lower in the Blueberry Ridge
population. Observed heterozygosity was lower than expected in all three of the V.
angustifolium populations, although the difference was not significant.

In V. corymbosum, there was no difference between the populations in the total
number of fertilized ovules, but the number of mature seeds was significantly different
in all three populations. It was lowest in Ely Lake and highest in Otis Lake. The
percent mature seed was also lowest in Ely Lake and highest in Otis Lake, although
Ely Lake was not significantly lower than Herbert Hoover. The number of alleles was
generally the same in the three populations, but the percent heterozygous individuals
was lower in the Ely Lake population than in the other two populations. The percent
tri-allelic individuals was considerably higher in the Herbert Hoover population than in
the other two populations. Observed heterozygosity was higher than expected in all of

the V. corymbosum populations, although not significantly.

74

Discussion

The number of alleles per locus and the level of heterozygosity was higher in
both of the tetraploid Vaccinium species than in the diploid. Similar results have been
reported in a number of other autotetraploid-diploid comparisons (Bayer, 1989; Ness,
Soltis, and Soltis, 1989; Soltis and Soltis, 1989; Lumaret and Barrientos, 1990).
There were 2.9 alleles and 21.7% heterozygous individuals per locus in the diploid V.
myrtilloides, compared to 3.4 and 3.6 alleles and 57.1% and 75.6% heterozygous
individuals per locus in the tetraploids V. angustifolium and V. corymbosum,
respectively. The level of genetic variation we found in V. myrtilloides is similar to
that reported in V. myrtilloides by Bruederle, Vorsa, and Ballington (1994) and is
similar to the average of most woody, long-lived, predominantly outcrossing diploid
species (2.1 alleles per locus and 17.0% heterozygosity, Hamrick and Godt, 1989).
Few studies of genetic variation in autotetraploids report the level of heterozygosity
across only polymorphic loci, but the average level of heterozygosity across six
polymorphic loci reported in the autotetraploid Tolmeia menziesz‘i was 53.5% (Soltis
and Riesberg, 1986), which is similar to the level of heterozygosity in V. angustifolium
(57.1 %).

It is interesting that the total number of fertilized ovules per fruit is higher in
the diploid than in the tetraploids (Table 4). Assuming that the total number of
fertilized ovules reflects the total number of ovules available for fertilization, the
higher number of ovules in the diploid may be an evolutionary strategy to ensure a
high level of reproductive output, because more deleterious recessives will potentially

be expressed in the diploid than in the tetraploid even in outcross progeny. In fact,

75

although the percent of fertilized ovules that survive to produce mature seed in the
diploid is intermediate to the two tetraploids, the actual number of mature seed
produced by the diploid is higher (Table 4).

Because diploids have a lower level of heterozygosity than polyploids, and a
corresponding increased potential for the expression of deleterious recessive alleles,
inbreeding depression is expected to be more severe in a diploid. Previous
comparisons of self and outcross pollinations demonstrated that inbreeding depression
was more severe in the diploid V. myrtilloides than in the autotetraploid V.
corymbosum (Hokanson and Hancock, in prep). This difference in inbreeding
depression may also be reflected in the present study by the lower percent mature seed
set found in open-pollinated fruit from V. myrtilloides than from V. corymbosum.
However, while the lower percent seed set in V. myrtilloides open-pollinated fruit
compared to V. corymbosum was consistent with previously measured levels of
inbreeding depression in these species, the lowest percent seed set found in open-
pollinated fruit from V. angustifolium was not. The levels of inbreeding depression
observed previously were the same in V. angustifolium and in V. myrtilloides
(Hokanson and Hancock, in prep), but open-pollinated seed set in this study was
significantly lower in V. angustifolium than in V. myrtilloides.

The percent mature seed set in open-pollinated fruit from V. angustifolium may
be lower than expected because there is a higher level of self pollination and
subsequent abortion of self-fertilized ovules due to inbreeding depression in this
species than in the others. Higher levels of self fertilization might also explain the

lower level of heterozygosity in V. angustifolium compared to V. corymbosum (Table

76

3). In fact, observed heterozygosity was less than expected in every population of V.
angustifolium at both of the loci for which expected heterozygosity could be calculated
(Table 5).

Although self-pollination is generally discouraged in the Vaccim‘um species by
their pendant shaped flowers and protandry, other factors may affect the selfing rate in
individual species. In fact, significant levels of self fertilization have been inferred or
reported in open-pollinated fruit of some Vaccinium species (V anderKloet, 1988). The
densely rhizomatous growth habit in V. angustifolium may be a factor promoting
geitonogamous selfing (Pritts and Hancock, 1984). Although V. myrtilloides is also
rhizomatous, the ramets are much more sparsely distributed than in V. angustifolium,
and flower density is much lower on the large, crown-forming V. corymbosum plants
(Pritts and Hancock, 1985; personal observation). Some insect pollinators will work a
number of flowers in a small space before moving to another patch of flowers (Levin,
1979). The growth habit of V. angustifolium is well suited to this type of pollinator
activity, encouraging the deposition of self pollen.

If there is indeed a higher level of self pollination occurring in V.
angustifolium, it is surprising that the level of inbreeding depression in this species is
so severe. Recent evidence and theory, at the diploid level, suggests that significant
inbreeding depression can be found in populations where there is low to moderate
levels of self fertilization (Lande, Schemske, and Shultz, 1994; Husband and
Schemske, 1995; Kohn and Biardi, 1995). However, this should only be true if very
few of the selfed progeny survive to reproduce. If the reduction in seed set in open-

pollinated fruit is due to self fertilization in V. angustifolium, then many of the selfed

77

progeny in this species do not survive. Yet, enough selfed progeny are surviving to
decrease the level of heterozygosity. Why, then, is inbreeding depression more severe
in V. angustifolium than in V. corymbosum?

The process of purging genetic load must be much more complex in an
autotetraploid than in a diploid, because of tetrasomic inheritance. In diploid
populations with very high inbreeding depression, some selfed progeny will
theoretically survive to purge genetic load if there is sufficient variation in fitness
among them (Lande, Schemske, and Shultz ,1994; Husband and Schemske, 1995).
There should be more variation in the fitness of selfed autotetraploid progeny than
selfed diploid progeny. However, surviving selfed progeny of an autotetraploid will
still maintain a higher level of genetic load than surviving selfed progeny of a diploid.
The frequency of a recessive allele (b) may increase (aabb to abbb) without becoming
homozygous.

The effect of deleterious recessives on a trait will also depend on the number
of loci involved, and the epistatic interactions among these loci. Selection experiments
in ”two allele” autotetraploid populations, which have been manipulated through
crosses to have a maximum of two alleles at any locus, indicate that an increased
frequency of favorable additive alleles is responsible for improved performance of
individuals (Pfeiffer and Bingham, 1983; Bingham et al., 1994). Recent research with
molecular (RFLP) markers in tetraploid potatoes (Bonierbale, Plaisted, and Tanksley,
1993) and in tetraploid alfalfa (Kidwell et al., 1994; Kidwell et al., 1995) also suggest
that heterosis is additive. The recent theory by Bingham et al. (1994) illustrates that,

because of their higher level of heterozygosity, there is a greater potential in

78

autotetraploids than in diploids for favorable dominant alleles with additive effects
among linked genes and for epistatic effects among the dominant alleles at those loci
that mask the expression of mutational genetic load. A logical extension is that
autotetraploids with more heterozygosity will have dominant alleles at more loci to
mask deleterious recessives at other loci, and therefore less severe inbreeding
depression, than autotetraploids with less heterozygosity. Bennett (1976) also
discussed the effect of interactions between genes in autotetraploids and suggested that
this may explain why newly synthesized autotetraploids, which should have two alleles
in equal frequency at most loci, show much less inbreeding depression than older
natural autotetraploids.

A lower level of heterozygosity is also associated with the low percent mature
seed in the Ely Lake population among the V. corymbosum populations, and in the
Blueberry Ridge population among the V. myrtilloides populations. The reduced open-
pollinated percent mature seed in these populations may be due to the expression of
deleterious recessives at loci that are not masked by dominant alleles at any other loci
because of the lower level of heterozygosity, even when the seed set is predominantly
outcrossed. Because the accumulation of favorable alleles and gene interactions have
also been demonstrated in diploids, primarily in maize (Bingham, 1994), there is
reason to expect the same association of lower heterozygosity and increased inbreeding
depression in a diploid.

Previously, the favored explanation for increased heterosis and severe
inbreeding depression in autotetraploid alfalfa has been the potential for higher order

allelic interactions in the tetraploid, i.e., more than two alleles at a locus (aabc or

79
abcd). (Busbice and Wilsie, 1966; Dunbier and Bingham, 1975). The severe

inbreeding depression observed in alfalfa was similar to the theoretical rate at which
first order interactions were lost from tri- and tetra-allelic loci. The existence of tri-
and tetra-allelic individuals at enzyme loci in alfalfa supported this theory
(Quiros,1982). Although tri-allelic individuals were found at every locus except
SKDH in V. angustifolium and V. corymbosum, the frequency averaged across loci was
generally very low, ranging from 0.9 to 10.6 percent among autotetraploid populations,
and no tetra-allelic individuals were identified at any of the loci. If multiple alleles
played a significant role in inbreeding depression in these tetraploids, we would have
expected a higher frequency of multiple-allelic individuals. Furthermore, because there
were more tri-allelics in V. corymbosum than V. angustifolium, we would expect
inbreeding depression to be more severe in V. corymbosum, but it is not.

There is no reason, however, to expect the association of heterozygosity and
genetic load to be consistent between species, or even between populations within a
Species. For example, Krebs and Hancock ( 1991) did find a significant correlation
between the number of alleles per locus per individual and the level of genetic load
among thirty individuals from one of the V. corymbosum populations (Otis Lake).
Essentially, inbreeding depression was more severe in individuals with more alleles per
locus. Krebs and Hancock (1991) had proposed that the number of alleles per
individual is a reflection of the number of deleterious mutations carried by that
individual. Because the correlation is between the number of alleles per locus per
individual and not the number of heterozygous loci per individual, it may also be a

reflection of segregational load, not mutational load, among individuals in this

80

population (Krebs and Hancock, 1990). Regardless of the explanation, the correlation
may be unique to the Otis Lake population. Certainly, the level of genetic load and its
expression will vary depending on the mutation rate and the number of loci effecting a
trait, as well as the epistatic interactions between those loci. The results of the present
study would predict more severe inbreeding depression in tetraploid individuals with
lower heterozygosity, but we did not have data on genetic variation and genetic load in
the same individuals, so we could not test for this correlation in other populations of

V. corymbosum, in Vangusnfolium, or in the diploid V. myrtilloides.

Conclusion

As expected, the level of heterozygosity was considerably higher in both of the
tetraploids, V. angustifolium and V. corymbosum, than in the diploid, V. myrtilloides.
There was also a lower level of heterozygosity in V. angustifolium than in V.
corymbosum. The lower level of heterozygosity in V. angustifolium may be due to a
higher level of naturally occurring self fertilizations, and tetrasomic inheritance may
allow partial self fertilization to reduce the level of heterozygosity without effectively
purging the genetic load. Furthermore, inbreeding depression may actually be more
severe in a population with a lower level of heterozygosity, particularly if there are
epistatic interactions among loci where deleterious recessives are being masked. If
this is true, then inbreeding depression may not be more severe in V. angustifolium
because it has more deleterious recessive alleles, per se, but because of the lower level
of heterozygosity. This, rather than a significantly higher level of genetic load in V.

angustifolium than the diploid V. myrtilloides, may also explain why inbreeding

81

depression in the previous study was the same in these two species, and not less
severe in the tetraploid as predicted theoretically.

Clearly, it would be of interest to determine the actual rates of self fertilization
in the different populations of these three Vaccim'um species. A comparison of self,
outcross, and open-pollinated seed set on individual plants in the field should be a fair
estimate of the selfing rates of those individuals (Charlesworth, 1988). It also would
be interesting to determine what proportion of surviving open-pollinated progeny are
the result of self fertilization. A detailed isozyme analysis of maternal plants and their
progeny, of the sort described by Clegg (1980), would be required to obtain this
information. Finally, a comparison of the level of heterozygosity to seed set in
individual plants, as Krebs and Hancock (1991) did, from different populations of
these three Vaccinium species should provide more insight into the relationship of

heterozygosity and genetic load in diploids and in autotetraploids.

82

Literature Cited
Bayer, RI. 1989. Patterns of isozyme variation in western North American Antennan‘a
(Asteraceaezlnuleae). II. Diploid and polyploid species of section Alpinae.

American Journal of Botany 76:679-691.

Bennett, J.H. 1976. Expectations for inbreeding depression on self-fertilization of

tetraploids. Biometrics 32:449-452.

Bever, JD. and F. Felber. 1992. The theoretical population genetics of autopolyploidy.

Oxford Surveys in Evolutionary Biology 82185-217.

Bingham, E.T., RW. Groose, D.R. Woodfield, and K.K. Kidwell. 1994.
Complementary gene interactions in alfalfa are greater in autotetraploids than

diploids. Crop Science 34(4):823-829.

Bonierbale, M.W., R.L. Plaisted, and SD. Tanksley. 1993. A test of the maximum
heterozygosity hypothesis using molecular markers in tetraploid potatoes.

Theoretical and Applied Genetics 86:481—491.

Breuderle, LP. and N. Vorsa. 1994. Genetic differentiation of diploid blueberry,

Vaccinium sect. Cyanococcus (Ericaceae). Systematic Botany 19(3):337-349.

83

Breuderle, L.P., N. Vorsa, and IR. Ballington. 1991. Population genetic structure in
diploid blueberry Vaccim'um section Cyanococcus (Ericaceae). American

Journal of Botany 78(2):230-237.

Busbice, TH. and CR Wilsie. 1966. Inbreeding depression and heterosis in

autotetraploids with application to Medicago sativa L. Euphytica 15:52-67.

Cai, Q., S.E. Macdonald, and CC. Chinnappa. 1990. Studies on the Stellan'a Iongipes
complex (Caryophyllaceae): isozyme variability and the relationship between
Stellaria Iongipes and S. Iongrfolia. Plant Systematics and Evolution 173:129-

14].

Charlesworth, D. 1988. A method for estimating outcrossing rates in natural

populations of plants. Heredity 61:469-471.

Charlesworth, D. and B. Charlesworth. 1987. Inbreeding depression and its
evolutionary consequences. Annual Review of Ecology and Systematics
18:237-268.

Clegg, MT. 1980. Measuring plant mating systems. Bioscience 30(12)8l4-818.

Dunbier, M.W. and ET. Bingham. 1975. Maximum heterozygosity in alfalfa: results

using haploid-derived autotetraploids. Crop Science 15:527-531.

84
Groose, R.W., L.E. Talbert, W.P. Kojis, and ET. Bingham. 1989. Progressive

heterosis in autotetraploid alfalfa: studies using two types of inbreds. Crop

Science 29:1173-1177.

Haldane, 1.8.8. 1930. Theoretical genetics of autopolyploids. Journal of Genetics

22:359-372.

Hamrick, J.L. and M.J.W. Godt. 1989. Allozyme diversity in plant species. In: Brown,
A.H.D., M.T. Clegg, A.L. Kahler, and BS. Weir, Eds. Plant Population
Genetics, Breeding, and Genetic Resources. Chapter 3, pp. 43-63. Sinauer

Associates Inc, Sunderland, MA.

Husband, BC and D.W. Schemske. 1995. Magnitude and timing of inbreeding
depression in a diploid population of Epilobium angustifolium (Onagraceae).

Heredity 75:206-215.

Kidwell, K.K., ET. Bingham, DR Woodfield, and TC. Osborne. 1994. Relationships
among genetic distance, forage yield, and heterozygosity in isogenic diploid
and tetraploid alfalfa populations. Theoretical and Applied Genetics 89:323-

328.

85
Kidwell, K.K., D.R. Woodfield, E.T. Bingham, and TC. Osborne. 1995. Molecular

marker diversity and yield of isogenic diploid and tetraploid single-crosses of

alfalfa. Crop Science 34:784-788.

Klekowski, 1988. Mutation, Developmental Selection, and Plant Evolution. Columbia

University Press, New York.

Kohn, JR and J. E. Biardi. 1995. Outcrossing rates and inferred levels of inbreeding

depression in gynodioecious Cucurbita foetidissima (Cucurbitaceae). Heredity

75:77-83.

Krebs, S.L. and J.F. Hancock. 1989. Tetrasomic inheritance of isoenzyme markers in

the highbush blueberry, Vaccim'um corymbosum L. Heredity 63:11-18.

Krebs, S.L. and J.F. Hancock. 1990. Early-acting inbreeding depression and
reproductive success in the highbush blueberry, Vaccinium corymbosum L.

Theoretical and Applied Genetics 79:825-832.

Krebs, S.L. and J.F. Hancock. 1991. Embryonic genetic load in the highbush
blueberry, Vaccinium corymbosum (Ericaceae). American Journal of Botany

78(10):l427-1437.

86

Lande, R. and D.W. Schemske. 1985. The evolution of self-fertilization and

inbreeding depression in plants. I. Genetic models. Evolution 39(1):24-40.

Lande, R, D.W. Schemske, and ST. Shultz. 1994. High inbreeding depression,
selective interference among loci, and the threshhold selfing rate for purging

recessive lethal mutations. Evolution 48:965-978.

Levin, DA. 1979. Pollinator foraging behavior: genetic implications for plants. In:
Topics in Plant Population Biology. O.T. Solbrig, S. Jain, G.B. Johnson, PH.

Raven, Eds. Columbia University Press. pp131-153.

Lumaret, R. and E. Barrientos. 1990. Phylogenetic relationships and gene flow
between sympatric diploid and tetraploid plants of Dactylis glomerata

(Gramineae). Plant Systematics and Evolution 169:81-96.

Ness, B.D., D.E. Soltis, and RS. Soltis. 1989. Autopolyploidy in Heuchera micrantha

(Saxifragaceae). American Journal of Botany. 76:614-626.

Parker, I.M., RP. Nakamura, and D.W. Schemske. 1995. Reproductive allocation and
the fitness consequences of selfing in two sympatric species of Epilobium

(Onagraceae) with contrasting mating systems. American Journal of Botany

82(8):]007-1016.

87

Pfeiffer, T.W. and ET. Bingham. 1983. Improvement of fertility and herbage yield by
selection within two-allele populations of tetraploid alfalfa Crop Science

23:633-636.

Pritts, MP. and J.F. Hancock. 1984. Independence of life history parameters in
populations of Vaccim’um angustifolium (Ericaceae). Bulletin of the Torrey

Botanical Club 11 l(4):451-461.

Pritts, MP. and J.F. Hancock. 1985. Lifetime biomass partitioning and yield
component relationships in the highbush blueberry, Vaccinium corymbosum L.

(Ericaceae). American Journal of Botany 72(3):446—452.

Purdy, B.G. and RI. Bayer. 1995. Genetic diversity in the tetraploid sand dune
endemic Deschampsia mackenzieana and its widespread diploid progenitor D.

cespr'tosa (Poaceae). American Journal of Botany 82(1):]21-130.

Quiros, C. 1982. Tetrasomic segregation for multiple alleles in alfalfa. Genetics

101:117-127.

Roose, ML. and LB. Gottlieb. 1976. Genetic and biochemical consequences of

polyploidy in Tragapogon. Evolution 30:818-830.

88

Scandalios, JG. 1969. Genetic control of multiple molecular forms of enzymes in

plants: a review. Biochemical Genetics 3:37-79.

Sotis, DE. and L.H. Riesberg. 1986. Autopolyploidy in T olmr'ea menziesir’
(Saxifragaceae): genetic insights from enzyme electrophoresis. American

Journal of Botany 73:310-318.

Soltis, DE. and PS. Soltis. 1989. Genetic consequences of autOpolyploidy in T olmiea

(Saxifragaceae). Evolution 43:5 86-594.

Thompson, JD. and R. Lumaret. 1992. The evolutionary dynamics of polyploid plants:
Origins, establishment, and persistence. Trends in Research in Ecology and

Evolution 7(9):302-307.

VanderKloet, SP. 1988. The genus Vaccim’um in North America. Research Branch,

Agriculture Canada, Publication No. 1828.

Watson, L.E., W.J. Elisen, and JR Estes. 1991. Electrophoretic and cytogenetic e
evidence for allopolyploid origin of Marshallr’a mohrir’ (Asteraceae). American

Journal of Botany 78:408-416.

"untiltilt