ECOLOGICAL GENETICS OF THE MOSS

PHYSCOAHTRELLA PATENS

Erica Jane HooPer

Submitted in accordance with the requirements for the degree of Doctor of

Philosophy

The University of Leeds

Faculty of Biological Sciences

September2008

The candidateconfirms that the work submittedis his/her own and that appropriate credit hasbeen given where referencehas been made to the work of others.

This copy hasbeen suppliedon the understandingthat it is copyright material and that no quotation from the thesismay be publishedwithout proper acknowledgement. Acknowledgements

Firstly I would like to thankmy supervisorsAndy Cumingand Bill Kunin, for all their support,guidance and encouragement. This hasbeen immensely valuable to mewhilst conductingand completing this project,and without which I wouldhave beenunable to finish.I amalso extremely grateful to YasukoKamisugi, for all her helpand advice with laboratorywork, particularly the AFLP analysis.Thanks also go to ClaireSmith, for herhelp with planttissue culture.

I am very grateful to all the bryologists who have assistedin finding bryophyte populations,without which I would have beencompletely unable to undertakethis research.Particular thanks go to SamBosanquet, David Chamberlain,Gorden

Rothero,John Blackburn, HansBlom and RachelO'Hara for accompanyingme on my hunt for populations.I am also very grateftil to Chris Preston,David Long,

David Holyoak, Mark Pool, Paul King, Tom Blockeel, Ron Porley, Angela Newton and Mark Lawley, for supplying details of populations.Thanks go to all the landownersand wardenswho gavetheir permissionfor plant collectionsto take place.

Many thanksgo to Jeff Duckett at QueenMary, University of London, for his help with identifying Physcomitrella readeri, without whose expertise the moss would remain a mystery. Thanks also go to both Jeff Duckett and Sylvia Pressel, for supplying beautiftil SEM imagesof the P. readeri plants.

I would like to thank my colleaguesat the University of Leeds,for providing both their friendship and support,along with a stimulating environmentin which to work.

Lastly I will thank my family for their enthusiasmand endlessencouragement, and for accompanyingme on cold, wet, field excursions.I particularly thank my partner for all his crucial supportand understanding,and for spotting the odd sporophyte when I was aboutto give up. iii

Abstract

Molecular geneticstudies using the model bryophytePhyscomitrella patens have advancedthe body of knowledgesurrounding plant functional geneticsand molecularbiology, yet very little is known aboutthe ecology and population geneticsof this species.Although the bryophytesare the secondlargest group of plants, there is little information regardingthe population geneticsof bryophytesin general.To addressthese issues I have conductedthe first study into the population geneticsof P. patens, whereplants from eight populationsin Britain havebeen collected. Samplingwithin thesepopulations was conductedaccording to a hierarchicalscale, so as to assessnot only the level of geneticvariation within populations,but how this is structuredspatially. Analysis of the plants collectedwas conductedusing amplified fragmentlength polymorphism (AFLP) analysis.No spatial geneticstructure was found within populationsof P. patens, and it is hypothesisedthat the natureof the ephemeralaquatic habitats that P. patens occupiesmay accountfor this finding. In this thesisa novel method for studying the mating systemsoperating within bryophytepopulations has been proposed, which exploits the dominanthaploid stageof the bryophyte life cycle. This methodology hasbeen applied to natural populationsof P. patens, and evidenceof mixed mating has beenobserved. Bryophytes are often overlookedor under-recordedin their natural environment,and distribution datawithin Great Britain is likely to be inaccuratefor a large numberof species.This issueis highlighted in this thesis,as a bryophyte speciesnew to Europehas been discovered. iv

Table of Contents

Acknowledgments i ...... Abstract iii ...... List Tables of ...... vii List Figures of ...... vii List Appendices of ...... x List Abbreviations Used of ...... xi Chapter I GeneralIntroduction I ...... 1.1 Introduction 1 to project ......

1.2 Genetic analysis of population structure and evolutionary processes 3

1.2.1 Fine-scale 3 spatial structure ...... 1.2.2 Mating 5 systems within populations ...... 1.2.2.1 The AFLPs in 6 use of population genetic analysis ...... 1.3 Bryophyte biology life history 8 and ...... 1.3.1 Geneticdiversity in bryophyte 10 populations...... 1.3.2Bryophyte 11 mating systems...... 1.4 Population 12 geneticsof model species...... 1.4.1 P. 14 patens -a useftil ecological model ...... 1.4.1.1 Biology P. 15 of patens ...... 1.4.1.2 Phylogenetics 16 and taxonomy...... 1.4.1.3 Historical 19 perspective...... 1.4.1.4 Genomics 20 ...... 1.5 Aims the 22 of project ...... Chapter 2 General Materials Methods 24 and ...... 2.1 Study 24 sites ...... V

2.2 Sample DNA 30 culture and extraction ...... 2.3 AFLP 30 analysis......

Chapter3 Geneticstructure within and betweenpopulations of

P. 32 patens ...... 3.1 Introduction 32 ...... 3.2 Materials 33 and methods...... 3.2.1 AFLP 34 profile scoring and error-ratetesting ...... 3.2.2 Data 35 analysis...... 3.2.2.1 Geneticdiversity linkage disequilibrium 35 and ...... 3.2.2.2 Genetic between 36 structurewithin and populations...... 3.3 Results 38 ...... 3.3.1 Geneticdiversity linkage disequilibrium 39 and ...... 3.3.2 Genetic between 42 structurewithin and populations...... 3.4 Discussion 45 ...... Chapter4 P. in Europe 51 readeri ...... 4.1 Introduction 51 ...... 4.2 Materials 53 and methods...... 4.2.1 Populationdetails 53 and samplecollection ...... 4.2.2 Sample 53 culture ...... 4.2.3 Nuclear ITS 57 region analysis ...... 4.2.3.1 Amplification 57 and sequencing ...... 4.2.3.2 Sequence 58 alignment and analysis ...... 4.2.4 Morphologcial 59 analysis ...... 4.3 Results ...... 59 4.3.1 ITS 59 sequenceanalysis ...... 4.3.2 Morphological 63 analysis...... vi

4.4 Discussion 65 ......

Chapter 5 A method for estimation of the mating systems operating

P. 68 within populations of patens ...... 5.1 Introduction 68 ...... 5.2 Materials 71 and methods ...... 5.2.1 Trial known 71 using progeny ...... 5.2.1.1 DNA AFLP 72 extraction and analysis of progeny ...... 5.2.2 AFLP 73 profile scoring ...... 5.3 Analysis 73 of natural populations ...... 5.4 Results 74 ...... 5.5 Trial known 74 using progeny ...... 5.6 Analysis 76 of natural populations ...... 5.7 Discussion 79 ...... Chapter 6 General Discussion 82 ...... 6.1 Genetic diversity 83 and spatial genetic structure ......

6.2 Mating in P. 84 systems patens ...... 6.3 P. in Europe 86 readeri ...... 6.4 Combining 87 molecular genetics and ecology ...... Literature 90 cited ...... vii

List of Tables

Table 2.1 Site details P. 26 of patens populations......

Table 3.1 Populationssampled, including the number of sporophytescollected at

34 eachpopulation ......

Table 3.2 Proportion of polymorphic loci, Nei's gene diversity (Hj) and the

index S A 41 standardised. of association( A)within patens populations ......

Table 3.3 Analysis (AMOVA) 43 of molecularvariance ......

Table 4.1 Provenancedata for in 54 samplesused phylogeny reconstruction......

Table 5.1 Table detailing the polymorphic loci observedwithin a sporophyte producedfrom a crossbetween the Villersexel and Gransdenstrains of P. patens.

Polymorphic loci AFLP fragment 76 are namedusing the size of the ......

Table 5.2 The number of sporophytescollected in eachof the threepopulations that contained segregating loci, and the density of P. patens plants surrounding the

78 sporophytes sampled......

ist o Figures

Figure 1.1 Overview of AFLP protocol. a) genomic DNA is digested with restriction enzymesEcoRI and Mel; b) adapterligation; c) PCR pre-amplification with one selectivenucleotide; d) PCR selectiveamplification with two selectivenucleotides. 7

Figure 1.2 Diagram depicting the life cycle of the mossP. patens Taken from

Schaefer& Zryd (2001) 9 ...... viii

Figure 1.3 Phylogenyof the diplolepideous-alternatemosses, based on 18SrDNA, trnL-F, rhcL, and rps4 sequences,taken from Cox et al., (2000). The position of the

Funariales, by F. hygrometrica,is indicated 18 as represented ......

Figure 2.1 Location of populationsof P. patens (e) sampled.The distribution of

Cambridgeshireis in detail in b) 25 populationswithin shown greater ......

Figure 2.2 Photographs of example P. patens populations. B= Llanfihangel Gobion,

D= PapercourtMarshes, E= OuseWashes, F= Wicken Fen 27 ......

Figure 2.3 Positionsof P. patens sporophytesthat were collectedfrom Pant-y-llyn in

Cannarthenshire.Each labelled point indicatesclusters of 1-3 sporophytesseparated by distances less 50cm 29 of than ......

Figure 3.1 between The relationship pairwise F, t between populations sampled and geographicdistance between populations, calculated using a subsetof samplesin a preliminary analysis.Fg betweenplants from Lindley Wood Reservoirand all other populations= A, and calculationsbetween all other populations=0...... 39

Figure 3.2 Correlationbetween Nei's genediversity within populationsof P. patens

(excluding H) P=0.088 40 population and population area.r= -0.714, ......

Figure 3.3 The relationshipbetween pairwise F,t betweenpopulations sampled and distance between P=0.081 43 geographic populations. r=0.396, ......

Figure 3.4 Kinship coefficient autocorrelationat a) over all sites,and b) to h) within populationsA to G. Mean kinship coefficient of eachdistance class is plotted againstmaximum classdistance. Significant coefficients are indicatedas P<0.05 =

*, P<0.0 1=** and P<0.00 1=***, basedon the null hypothesisof kinship ix coefficient = 0. The upper and lower 95% confidenceintervals are shown by dashed lines 44 ......

Figure 4.1 Location P. in West Yorkshire, UK 55 of readeri population ......

Figure 4.2 Panoramicphotograph of Lindley Woods Reservoir,where P. readeri

found in November2006 56 was ......

Figure 4.3 ITS I maximum likelihood phylogram. Likelihood =- 1132.Bootstrap values from the maximum likelihood analysisare shown abovebranches, with bootstrapvalues from the parsimonyanalysis shown below (I tree, 159 steps,C. I.

0.86 R.I. 0.80) 61 and = ......

Figure 4.4 ITS2 maximum likelihood phylogram.Likelihood = -1225. Bootstrap values from the maximum likelihood analysisare shown abovebranches, with bootstrapvalues from the parsimonyanalysis shown below. (I tree, 167 steps,C. I. =

0.89 R.I. 0.84) 62 and = ......

Figure 4.5 Sectionof the BLAST sequencealignment between P. readeri from

Lindley Wood (top strand)and P. readeri from Australia (bottom strand).The

in ITS I is highlighted 63 single-basemismatch the region ......

Figure 4.6 a-f SEM images of P. readeri from Lindley Wood Reservoir. a-c) sporophyte,d) leaf, e) stoma f) stomataat baseof seta.Scale bars: a, c= 400ýtm,b, d= 100pm, lOpm, f= 30ptm 64 e= ......

Figure 4.7 a-b. Light micrographshowing entire leavesfrom a) P. patens and b) P.

from Lindley Wood Reservoir 64 readeri ...... x

Figure 5.1 Screenshotdepicting an exampleof polymorphismsbetween the pooled and referencegametophytes produced from the GransdenX Villersexel cross.Three

fragments 254,264 265 bp 75 polymorphic loci correspondto AFLP of size and ......

List of Appendices

Appendix I Table detailing potential populations of P. patens that were visited... 106

Appendix 11Description P. 112 of patens ......

Appendix III Description P. 113 of readeri ......

Appendix IV Screenshots depicting polymorphisms between the pooled and

reference gametophytes produced from the Gransden X Villersexel cross.

Polymorphic fragment indicated 114 sizes are ......

Appendix V Table detaling eachsporophyte analysed in Chapter5, including the

numberof polymorphic loci and the density of P. patens plants aroundeach 115 sporophyte......

Appendix VI Screenshotsdepicting polymorphismsbetween the pooled and

referencegametophytes from sporophytescollected from natural populationsof P. patens, as detailedin Chapater5. SporophyteI. D's and polymorphic fragmentsizes

118 are given......

Appendix VII Screenshotdepicting locus EAGMCT229 119 polymorphic ...... xi

List of Abbreviations Used

AFLP Amplified fragmentlength polymorphism

AMOVA Analysis of molecular variance

Cl Confidenceinterval

C.I. Consistencyindex

bp Basepairs

cDNA ComplementaryDNA

EST Expressedsequence tag

FAM Blue fluorescentdye (Applied Biosystems)

GPS Geographicalpositioning system

HEX Greenfluorescent dye (Applied Biosystems)

IBD Isolation by distance

ITS Internal transcribedspacer

NED Yellow fluorescentdye (Applied Biosystems)

PCR Polymerasechain reaction

QTL Quantitativetrait loci

RAPD Randomamplified polymorphic DNA

R.I. Retentionindex

SCALP Sequencecharacterised amplified length polymorphism

SEM Scanningelectron microscope Chapter 1 General Introduction

1.1 Introduction to project

The bryophytesare a diverseand ecologically interestinggroup of plants,

comprising25,000 speciesoccupying a wide variety of habitats(During & van

Tooren, 1987).They are the secondlargest group of land plants in terms of species

numberand in many ecosystemsthey are the main primary producers(During & van

Tooren, 1987;Mishler, 1988).Despite the diversity and ecologicalimportance of

this group they are often overlookedin population studies,and little information

regardingthe evolutionary geneticsof bryophyteshas been published (Innes, 1990).

Bryophytesare often missedor incorrectly recordedin their natural environment,so

for many groupsof mosseslittle information is availableon their distribution,

ecology and population biology. During the courseof this project, a moss species new to Europehas been found. This specieshas been verified as Physcomitrella

readeri (C. Milell. ) Stone& Scott, a speciespreviously only recordedin the U. S.A,

Japan,Australia and New Zealand.

Knowledge surroundingthe molecular geneticsof bryophyteshas increased dramatically in recentyears, with the developmentof the mossPhyscomitrella patens (Hedw.) Bruch & Schimp.as a model organism.The Physcomilrella Genome

Programmehas generatedover 300,000expressed sequence tags (ESTs) and has recently publishedthe first draft genomesequence assembly (Rensing et aL, 2008)

Other model organismsand their relatives,such asArabidopsis thaliana, Drosophild melanogasterand Escherichia coli, have increasinglybeen utilised in ecologicaland evolutionary studiesto investigateaspects such as adaptation,genetic diversity and 2 reproductivestrategies in natural populations(as reviewedby Mitchell-Olds 2001;

Shimizu, 2002; Travis, 2006).

One area of current focus in plant population ecology is that of the spatial genetic structure within plant species, and how this relates to features of the habitat they occupy. This field has received much attention in vascular plants throughout the last

20 years, with many studies focussing explicitly on how genetic variation is partitioned spatially within a species. There has been an increasing trend towards using mapped individuals as the operational taxonomic unit in population genetic studies, rather than randomly sampling within populations, which are often arbitrarily and inaccurately defined (see Manel et aL, 2003 for a review). However, this is an approach that has only very recently been applied to bryophyte populations

(e.g. Snqll et aL, 2004), and little work into population structure in general has been undertaken outside commonly studied moss genera such as Sphagnum (e.g.

Gunnarsson et aL, 2005; Stenoien & SAstad, 1999; Thingsgaard, 2001).

Additionally, little has been done to clarify reproductive processesin bryophyte populations, yet this aspect of organismal life history has one of the most significant impacts on levels of diversity, genetic structure and population behaviour (Shaw,

2000).

Model organisms, such as P. patens are a useful starting point when addressing questions surrounding the genetic processes in evolutionary ecology as a wealth of information and tools is available for these organisms. This investigation aims to be the first study into the population genetics of P. patens, thus allowing the detailed molecular information available for this species to be related to its ecology. 1.2 Genetic analysis of population structure and evolutionary

processes

1.2.1 Fine-scale spatial structure

Advances in the accuracy, resolution and scale of the molecular tools used for

population genetic studies have allowed utilisation of molecular markers in fine-

scale population studies (Manel et aL, 2003). When using these tools to investigate

the structure within populations at differing spatial scales, the distance over which

genetic variation within populations is distributed and the levels of gene flow within

a population can both be determined.Plant populationsare often highly spatially

structured,which can be due to a number of reasons(reviewed in Ennos,2001).

Plantsare generally static, and geneticdispersal is by meansof pollen and seeds,or

in the caseof bryophytesby spores,sperm and vegetativefragments. Thus gene

flow can be limited and geneticisolation by distance(1131)) occurs within populations(Vekernans & Hardy, 2004; Wright, 1943).Many plants also exhibit

vegetativedispersal, which will lead to a patchy distribution of genotypes.In

addition, spatial structurewithin populationscould be observeddue to sampling

eventsupon colonisationof new areasor regenerationof populations(Ennos, 2001).

How gene flow operates within species greatly affects not only the spatial structure of populations, but also the ability of a population to adapt to local conditions, with a subsequent possibility of independent evolution of those populations resulting in speciation (Slatkin, 1985). When spatially dependent processes, such as 1131),have been identified in plant populations, they can be correlated with landscape and population features, such as population size or fragmentation (Manel el aL, 2003).

This information can be of use in conservation biology to predict the effect that 4 changesin habitat could have on population geneticprocesses (Escudero et aL,

2003; Storfer el aL, 2007).

There are a numberof different methodsfor assessingthe spatial geneticstructure within plant populations(see reviews in Escuderoet aL, 2003; Manel et aL, 2003; and Storfer et aL, 2007). Commonly usedpopulation analyses,such as Analysis of

Molecular Variance,or AMOVA (Excoffier et aL, 1992)and estimatesof genetic differentiation and distancebetween populations, have long beenused to infer the hierarchicalstructure of geneticdiversity within and betweenpopulations (Escudero et aL, 2003). The discretemapping of sampledindividuals and subsequentgenetic profiling, allows far more sophisticatedstatistical analysis to be conducted.Spatial autocorrelation,that is the associationof one variable at one location with the same variable at anotherlocation (Schabenberger& Gotway, 2005), can thereforebe examined.A widely usedmethod of investigatingspatial autocorrelationis a Mantel test (Mantel, 1967).Here spatial clusteringof variablesis examinedby the constructionof two distancematrices, one with geneticdistance between individuals and one with geographicdistance between individuals. The correlationbetween thesetwo matricesis examined,and a randomisationprocedure is usedto determine if clusteringbased on geneticsimilarity can be explainedby spatial position

(Escuderoet aL, 2003; Schabenberger& Gotway, 2005). The disadvantageof this approachis that spatial geneticautocorrelation does not necessarilyfollow a linear pattern (Escuderoet aL, 2003). An enhancementof a Mantel test is to partition measurementsof geographicdistance into distanceclasses. Spatial autocorrelation

analysiscan then be conductedbased on theseclasses rather than linear distance,

with dataoften presentedas a correlogram(Escudero et aL, 2003). 5

1.2.2 Mating systems within populations

The mating patternsin plant species,and specifically the mating systemsoperating within populations,are a significant featureof a species'ecology. The mating systemof any speciesor population is not static,but can vary in responseto changingenvironments and evolutionary forces.Changes in the mating systemcan have important consequencesfor the geneticstructure and behaviourof that population (Barrett & Eckert, 1990).Although somespecies, such as dioecious specieswith separatemale and female plants, are obligate out-crossers,many plant speciesexhibit levels of inbreeding.Classic population genetictheory dictates, particularly in animal populations,that out-crossingis evolutionarily favourableto inbreeding.This is becauseinbreeding may result in inbreedingdepression, which is measuredby the fitness of selfed offspring comparedto thosethat are the product of out-crossing(Ritland, 1990).Inbred individuals are thought to be reducedin fitness comparedto out-crossedindividuals due to two reasons;that heterozygosityis superiorto homozygosityat any one locus, and that inbreedingmay result in an accumulationof deleteriousalleles in a population (Charlesworth& Charlesworth,

1987).However, due to the propensityof plants for self-fertilization, predominant inbreedinghas been found to be a stablestate in many plant populations(Lande &

Schemske,1985). Studies into the mating systemsoperating within plant populationshave found that mixed mating, or the presenceof both inbreedingand out-crossingin a population, also frequently occurs (Barrett & Eckert, 1990;

Goodwillie et al., 2005; Vogler & Kalisz, 2001; Walter, 1986).Despite the

importanceof mating systemsto the ecology and evolution of a species,knowledge

is limited with regardsto within-population variation of mating systemsin plants

(Cruzan, 1998). 6

1.2.2.1 The use of AFLPs in population genetic analysis

The use of neutral molecular markers in plant diversity studies has greatly enhanced the range of genetic studies that can be conducted (Cruzan, 1998), and amplified fragment length polymorphism (AFLP) analysis is increasingly being favoured as the marker of choice. The technique involves restriction digestion of DNA followed by selective amplification of the resulting fragments (see Figure 1.1), which can then be viewed on a gel or automated sequencer (Blears et al., 1998; Ritland & Ritland,

2000; Vos et aL, 1995). The main advantage of using AFLPs in population studies is that the numberof loci scoredand polymorphismsidentified in eachreaction is high

(Blears,et aL, 1998).This is important when studying diversity within populationsas the geneticvariation betweenindividuals may be low. AFLP is also favouredas it is a robust and reliable geneticmarker, which is highly reproducibledue to the specificity of restriction enzymes.it is more reliable than similar techniquessuch as

RAPD, which is more sensitiveto the reactionconditions (Blearset al., 1998).

AFLPs provide dominantmarkers, and in diploid tissueheterozygotes cannot be distinguished.Ideally in population geneticstudies co-dominant markers such as microsatelliteswould be usedto study geneticstructure and the inheritanceof alleles in most species(Ennos, 2001). However, when using microsatellitesfar fewer loci can be scrutinised,and in addition they are expensiveand time consumingto develop.In specieswhere the dominant stageof the life cycle is haploid, suchas in mosses,the disadvantagesof using a dominantmarker are negated,and AFLPs provide an ideal marker system. 7

a) ______

b)

EcoRl adapter Msel adapter

C) +1 ý Msel primer EcoRl primer

+2 Msel primer EcoRl primer 7

Figure 1.1 Overview of AFLP protocol. a) genomic DNA is digested with restriction enzymes EcoRI and Msel; b) adapter ligation; c) PCR pre-amplification with one selective nucleotide; d) PCR selective amplification with two selective nucleotides. 8

1.3 Bryophyte biology and life history

The bryophytesare one of the oldest groupsof land plants, with the fossil record placing their divergencefrom the vascularplant lineageat about 450 million years ago, althoughmolecular evidence suggests that this divergencemay have occurred as early as approximately700 million yearsago (Heckmanet aL, 2001). The defining featureof this divergenceis that the mosses,along with other plant groups suchas algaeand livcrworts, have a contrastinglife cycle comparedto vascular plants. Bryophytesspend the majority of their life cycle as haploids in the gametophyticstage, with the sporophyticstage being short lived and dependenton the gametophyte.As the life history of a specieshas been shown to have a profound impact on the level and distribution of geneticvariation within natural populations

(Hamrick & Godt, 1990),populations of bryophytesmay be expectedto exhibit quite different characteristicsto vascularplants. A depiction of this life cycle in the mossP. patens, is shown in Figure 1.2.

All land plants exhibit an alternationof generationsbetween the sporophyticstage and the sexualgametophytic stage, however in vascularplants the gametophytic stagehas been greatly reduced.The fossil record showsthat in early land plants the gametophyticand sporophyticstages were similar in terms of complexity, with the reduction of the gametophytein vascularplants developingafter the transition to land (Kcnrick & Crane, 1997;Taylor et aL, 2005). The bryophytesrepresent the earliestgroup of plants to diverge in the land plant phylogeny, so they therefore provide an opportunity to study the evolutionaryprocesses associated with the conquestof land (Rensinget aL, 2008; Renzagliaet aL, 2007). 9

nd iN

'64 «a :3 -0

pc T) C: 2 CV) CYJ 0 C) .Ccl 0 "0 6 -6 ý4 LLJ E ca -E CA r = a- _0 0 E () 0 LLI q

0 CL-j - 0 o ( a) 2ý clý CU a)

rp I &EE L vl-- () C,j I A

C), C) co 'C 0 0 (1)

(1) CD F-w21/ >- 16 E mrI. E >1 N>1 (t IIS? Co 0- cl) (1) "'o CL Ca cr 001, E 0 E 'Fa 0 cc CL Cl E 01 M Cc (nCD T 90E ýccyý a) 7a V) 0 - 2 -j -. (D c -J 10

1.3.1 Genetic diversity in bryophyte populations

The consequencesof a dominanthaploid life stagein bryophyteshas long been thought to have significant implications for levels of diversity in mosses,and until recently there was a strongly held view that bryophyteslacked the "evolutionary potential" (Wyatt et aL, 1989b)of vascularplants. There was generalacceptance that diploid organismspossess greater genetic variation, and thus a greater evolutionarypotential than haploid organismsdue to heterozygosityin diploids

(Ennos, 1990).However, a number of reviewershave taken issuewith this assumption(e. g. Ennos, 1990;Wyatt et aL, 1989b),and early studiesof the population geneticsof bryophytesusing isozymevariation have shown haploid organismsto be no lessgenetically variable at the population level than vascular plants (Cummins& Wyatt, 1981;Innes, 1990;Wyatt et aL, 1989a;Yamazaki,

1981). However, it hasbeen suggestedthat there are problemsin interpreting isozymedata, and that they do not necessarilyaccurately detect genetic variation

(Hassel& Gunnarsson,2003; Stenoien& SAstad,1999). With the developmentof other molecularmarkers such as microsatellitesand RAPDs, that record genetic differencesdirectly at the DNA level, high levels of geneticdiversity within bryophytepopulations have againbeen found (e.g. Skotnicki et aL, 1999;Stenoien

& Flatberg,2000; Wilson & Provan,2003). It has beenargued that as most molecularmarkers are selectivelyneutral, the levels of geneticvariability found do not reflect the evolutionaryrate of bryophytepopulations (Stenoien & Flatberg,

2000; Stenoien& SAstad,1999; Yamazaki, 1981).Perhaps more convincingly, where diversity levels havebeen studiedusing quantitative characters,the levels of diversity shown by molecular data are confirmed (Shaw et aL, 1987). 11

1.3.2 Bryophyte mating systems

In bryophytesthere is a wide rangeof mating systemsoperating within species, owing to the unique life history of theseplants (Shaw,2000). Bryophyte mating systemsfall into two generalcategories; dioicous specieswith separatemale and female gametophytes,and monoicousspecies with both sexeson one gametophyte.

Monoicousbryophytes can also be divided into a numberof different forms; with archegoniaand antheridiaeither in the sameinflorescence (synoecious), the same florescencebut separatedby bracts (paroecious),or on the samegametophyte but different stems(autoecious) (Shaw, 2000). The proximity of antheridiaand archegoniaon the samegametophyte will thereforehave consequencesfor the likelihood of the occurrenceof inbreedingin a species.It is likely that all monoicous bryophyteshave at leastan element,if not to a significant degree,of inbreeding within populations,and no bryophyte specieshave beenfound to be self- incompatible(Shaw, 2000).

Bryophytescan not only undergo"true" self-fertilisation Le. the mating of gametes producedfrom the sameindividual (intragametophyticselfing), but also intergametophyticselfing; the mating of gametesproduced from haploid individuals producedfrom the samediploid parent(Klekowski, 1972).Intergametophytic selfing is equivalentto self-fertilisation in seedplants, which resultsin a reduction in heterozygosityin the diploid stageof 50% in one generation,and a reduction of 90% in 10 generations(Hedrick, 1987a;Hedrick, 1987b;Shaw, 2000). In contrastto this, intragametophyticselfing results in a completeloss of heterozygosityin one generation(Hedrick, 1987a;Hedrick, 1987b;Shaw, 2000). It is intuitive thereforeto concludethat monoicousbryophytes should be lessgenetically diversethan dioicous species(Wyatt et aL, 1989b).This hasnot been found to be the casehowever when 12

comparisonshave been conducted (Cronberg et aL, 2006). It is thought that this is

becausesexual reproduction is rare in dioicous bryophytes(Cronberg, 2002), with

male and female gametophytesoften spatially separated(Longton, 1976).This is

supportedby evidencethat sporophytesare frequently producedin monoicous

species,but rarely in dioicous species(Longton & Schuster,1983).

Along with the effect of diversity levels within bryophyte populations, the mating

systems in operation will also have an effect on gene flow within a population and thus the scale at which spatial structure is observed. In monoicous species, where

inbreeding is frequent, small scale spatial structuring would be expected (Ennos,

2001). If true inbreeding does not occur, then crosses with siblings are likely as gamete dispersal has been shown to be restricted to distances of tens of centimetres in bryophyte populations (Anderson & Lemmon, 1974; Longton, 1976). However, as monoicous bryophytes produce sporophytes more profusely than dioicous

species, gene flow mediated by spores, which could travel tens of metres or ftirther, may be more common.

1.4 Population genetics of model species

Populationand ecologicalgenetic studies tend to focus on ecologically interesting organismsor organismsthat are potentially useful, such as crop wild relatives.The focusesof evolutionary studiesare, however,beginning to change.As well as identifying and analysingthe evolutionaryprocesses that occur within natural populations,researchers also aspireto identify the genesthat affect these evolutionaryprocesses (for a review seeFeder & Mitchell-Olds, 2003). This marriageof ecology and moleculargenetics has beentermed "evolutionary and ecological functional genomics" (Feder& Mitchell-Olds, 2003). To conduct 13 molecular studiesrequires a considerablefinancial investment,with well equipped laboratoriesand skilled researchersrequired. Therefore to investigatethe functional genomicsof the wide numberof organismsthat havebeen the focus of prior ecological studieswould be untenable,especially since many such specieshave been poorly characterisedin terms of their molecularbiology. By contrast,to focus ecologicaland evolutionary studieson appropriatemodel organisms,which have beenthe targetsof moleculargenetic studies,would havemany benefits,Gene function in model organismshas classicallybeen studied by creatingmutants in a limited numberof laboratory"wild-type" strains,and identifying genesthrough traditional "forward" and "reverse" genetics.The numberof mutantsthat can be createdis limited by the phenotypeof the wild type strains,so in order to apply such an analytical approachit is necessaryto obtain a wide variety of accessionsfrom various habitatsand locationswhich would enablethe identification of ecologically relevantphenotypic traits (Alonso-Blanco& Koornneef,2000).

The ecologicalgenetics of a number of model speciesand their closerelatives have been studied,however the diversity of model organismsavailable for ecological studiesis low, particularly amongplants, in which the majority of studieshave investigatedthe population geneticsof A. thaliana (Jorgensen& Mauricio, 2004;

Miyashita et aL, 1999;Sharbel et aL, 2000) and its close relatives.The genetic variation found in plants sampledfrom natural populationshas beenused to investigateevolutionary interestinggenes (Johanson et aL, 2000), and quantitative trait loci (QTL) mapshave beenproduced (see Alonso-Blanco & Koornneef,2000;

Tonsor et aL, 2005 for reviews). Work on animal species,particularly D. melanogaster,has been far more prolific, with a long historical body of work to supportthis (for a review seeMousette & Derome,2004), and organismsthat are 14

easily manipulatedin the laboratory,such as Saccharomycescerevisiae also have the potential to be useful evolutionary models (Landry et aL, 2006).

If investigations into the molecular ecology of model species are to be extrapolated generally, a wide diversity of species need to be studied (Travis, 2006). In plants, genetic studies have mainly been restricted to A. thaliana or crop plants, which provide limited coverage of the diversity of plant life. The model moss, P. patens, therefore provides an alternative valuable model in which to combine ecological and evolutionary studieswith the many moleculartools available for this species.P. patens is a useful comparativemodel to A. thaliana and other flowering plants, as it offers the potential to explore characteristicsfundamental to the evolution of land plants (Nishiyamaet aL, 2003; Waters,2003; Zimmer et aL, 2007).

1.4.1 P. patens -a useful ecological model

P. patens has grown in popularity as a model speciesdue to the experimentally amenablecharacteristics of its life cycle, biology and genetics.Haploid protonernal and gametophytictissue is easily cultured under laboratory conditions,and the dominanceof this haploid stageallows for easyidentification of recessivemutants

(Cove et aL, 1997).Gametophytic tissue can be regeneratedboth from sporesand fragmentsof protonernaltissue, and many plants can be cultured with relative speed and little spacerequirements, particularly if cultured on agar.The wealth of knowledgeon the culture of P. patens in the laboratory,and the ability to easily manipulatethe resultantgametophytes, should help make the developmentof laboratorybased ecological studiesmore straightforward.

The life cycle of the most commonly usedstrain, named"Gransden" after its site of collection near GransdenWood in Cambridgeshire,takes around three to four 15

months to complete(Cove, 2005), and yearsof culture in the laboratoryhave

resultedin a loss of fertility in this strain. However, antheridiaand archegoniacan

be producedin the laboratory,and the Gransdenstrain is still able to undergoboth

self- and cross-fertilisation.During the courseof this work, wild accessionshave

been shown to be far more fertile than Gransden,with the production of gametangia

occurring severalweeks earlier than in the Gransdenstrain.

P. patens is also an excellent model for "reverse genetic" studies as targeted

transgeneinsertion by homologousrecombination occurs with high frequency,in

contrastto most other plant species(Schaefer & Zryd, 1997).Therefore, interesting

genes can be more easily manipulated for functional investigation than in any other

plant species.

1.4.1.1 Biology of A patens

A patens inhabits ephemeral aquatic habitats; namely ponds, lakes, ditches and

streamsin the northernhemisphere. Populations are found in Europe,from north of

the Mediterraneanto Scandinavia,Siberia, and North America (Hill et aL, 1994;

Smith, 2004; Tan, 1979).For a detaileddescription of P. patens seeAppendix 11,

which is taken from Smith (2004). Plantsdevelop in early summer,but their

appearanceis dependenton the populationsbeing submergedin winter, and on

water levels subsequentlyfalling in summer(Cove, 2005). Sporophytesare

producedfrom late summerto early winter. P. patens is classifiedas occasionalin

Smith (2004), but it is likely to be a relatively under-recordedmoss (S. Bosanquet, pers. comm.). Further information regardingthe habitatsthat this speciesoccupies is

requiredin order to utilise fully its potential as a model species,however its 16 distribution in Europe,Japan and the USA facilitates study by global research groups.

P. patens is monoicous,with antheridiaand archegoniaon the samegametophore, where self-fertilisation can occur (Cove, 1983).The production of sporophytesin laboratory strainsof P. patens has beenshown to be dependenton a day length of about eight hours (Hohe el aL, 2002) and a temperature range of 15-19'C (Engel,

1968).In theseconditions each sporophyte produces between three and four thousandspores (Engel, 1968).No study has investigatedthe biology of P. patens in natural populationsor in an enviromnentother than the laboratory,so very little is known about its ecology.

1.4.1.2 Phylogenetics and taxonomy

P. patens is a member of the Funariaceae, which are in the order Funariales in the

Bryopsida.This includesthe mosseswith a diplolepideous-alternateperistorne, which describestwo alternateconcentric rings of toothlike structuresat the capsule opening (Cox et aL, 2000). A phylogeny of bryophytes, indicating the position of the Funariales, is shown in Figure 1.3. No phylogeny of the Funariales or

Funariaceae has been produced. The Funariaceae comprise six genera;

Physcomitrella Bruch & Schimp., Aphanorrhegma Sull., Entosthodon Schwager.,

Funaria Hedw., Physcomitrium (Brid. ) Brid. and Pyramiduld Brid. Physcomilrella comprisesonly two species,P. patens and P. readeri. Togetherwith P. patens, the

Funariaceaeincludes a secondwell studiedspecies, Funaria hygrometrica.

The Funariaceaehave posedproblems for taxonomistsdue to a large variability in sporophytemorphology, which is unusualin mosses.The family is characterisedby gametophyticcharacters, including thin-walled obovateleaves with inflated terminal 17 cells. Sharedsporophytic characters include stomataenclosed by a single guard cell

(Fife, 1985).Members of this family also readily hybridise, with P. patens forming hybrids in natural populationswith Aphanorrhegmaand Physcomitrium,which has led to suggestionsthat theseare the samegenus (Bryan, 1957).There is still controversyover the naming of P. patens, and it is classifiedin The Moss Flora of

Britain and Ireland (Smith, 2004) asAphanorrhegma patens (Hedw.) Bruch &

Schimp.This is after the re-interpretationof the Funariaceaeby Lindberg (1864), which includedPhyscomitrella within Aphanorrhegma(in Fife, 1985).

Physcomitrella,as first describedby Bruch, Schimperand Gfimbel (1836-1855;in

Fife, 1985),remains as a separategenus in many other regional floras, including the

Flora of the USA (Goffinet, 2007), and important floras from Japan(Ochi, 1968) and New Zealand(Scott & Stone, 1976; Stone& Scott, 1973).This nomenclature hasbeen retained for the purposesof this study as it is most commonly recognised by the scientific community as Physcomitrella. 18

Diphyscium foliostin 100 Therotia lWolia Diphysciales Funana hygrometrica Fjunariales 24 71mmia ausMaca Timmisles loo rimmia sibitica 12 7 Biyobrittonia longipes Encalyptales 89 Encalypta rhaplocatpa Drummonda obtusifolia 'q 10" Scouteria aquatica 0 61 92 4 Wardia hygromotdca -liaploltpidene" Totfula 81 rurahs PWomitrium gardneri 3, Zygodon irdermedius L3 Ulota Orthotrichales Orthotrichum lyaflil 671 im- 9 E Splachnum arripullaceum EI 19 Tayloria fingulata Splachnsles 100 93 100 1 1 Leptobqm pyriforme Bryacese iz Meesia uliginosa Nicesiacege 92 Amblyodon dealbatus 9 Philonotis 98 Bartramia Bartramisceike d99 22 Rhacocarpuspurpurascens lledwigialts Hedwigia cillata Pohlia cruda Mielichhofada elongata tE 100 Mierichhofedabryoides Z 100'Oot E Leucolepis nwnzt*esi*i Mnium 97F-439 Plagiomnium Leptostomtin macrocarpan Leptostomataccae 1 ssi 33 Rhodobtytm giganteum E Bryum 1 too 5 9 Mielichhofeiia macrocarpa 30 99 Anomobryumlutaceum Aulacomnium tLrgidum Aulacomniaecte Pyrrhobryum vallis-gratiae Rhizogoniteene 12 r- Racopitum Racopilaccae Hypnodendron llypnodendraceae 99= 100 Bescherellia brevifolla Cý,rtopodaccae OrthodontiL#n fineare Hookoria r73 124- Hypopterygitm tama fisci Fontinalis axioyrotica Leucodontales ý i- Plagiothecium Lndulatum 1 Bfachytheclum 74 Hypnoles _ Hypnum lindbergii Rhytidiadelphussquanrsus

Figure 1.3 Phylogenyof the diplolepideous-alternatemosses, based on 18SrDNA, trnL-F, rbcL, and rps4 sequences,taken from Cox et al., (2000). The position of the

Funariales,as representedby F. hygrometrica,is indicated. 19

1.4.1.3 Historical perspective

Mosseshave long beenthe focus of geneticand developmentalstudies, with mosses utilised as important model systemsas early as the 1940's (seeCove et aL, 1997for a review). The first significant genetic studieswith P. patens were conductedby

Engel (1968), who producedmutants of P. patens using chemical and X-ray mutagenesis.Further significant studiesdid not comeuntil the 1970's,when P. patens was adoptedas a model by David Cove at the University of Leeds,and

Wolfgang Abel at the University of Hamburg (Schaefer& Zryd, 2001). Important developmentsduring this period included the regenerationof plants from protoplasts and the resultantability to conductcomplementation studies using protoplast fusion

(Grimsley et aL, 1977a;Grimsley et aL, 1977b).Shortly after this, self-sterilebut cross-fertilemutants were developed,which allowed test crossesto be conducted

(Courtice et aL, 1978).Further studiesof P. patens focussedon its developmental biology, with a numberof developmentalmutants identified (seeCove & Knight,

1993 for a review).

P. patens becamethe focus of moleculargenetic experimentation in the late 1990s, with the discoverythat when transformationwas attemptedvia polyethyleneglycol transformationof protoplasts,DNA containing homology with endogenous sequencesintegrated preferentially at the cognateloci. This "gene targeting" occurredby homologousrecombination, a featureabsent in other model plants. This allows far greaterversatility in P. patens than other plant species,as specific genes can be targetedfor very precisemutagenesis, revealing insights into gene function

(seeReski, 1999;and Schaefer& Zr?d, 2001 for reviews). 20

1.4.1.4 Genomics

P. patens has been found to possess27 chromosomes in the wild type strain (Reski et aL, 1994),and following the assemblyof a first-draft completegenome sequence, its genomehas been estimatedto contain over 35,000 genemodels (Rensinget aL,

2008). An important resourcethat could be utilised in evolutionary studiesof bryophytes,is the P. patens expressedsequence tag (EST) database,which hasbeen usedto annotatethe first draft of the genomesequence assembly (Rensing et aL,

2008). ESTsprovide an insight into possibly evolutionarily significant genes,as they provide information about the genesexpressed in particular tissues,at particular times and under particular conditions (Bouck & Vision, 2007). Along with information about genefunction, ESTs can also be usedto producemolecular markersthat could be utilised in population geneticsstudies, and a number of microsatellitemarkers have recently beendeveloped using ESTs from P. patens

(von Stackelberget aL, 2006).

The function of a number of interestinggenes has beendiscovered in P. patens, including genesthat could havebeen important in the transitionsto life on land. P. patens is tolerant to a numberof environmentalstresses, and the P. patens EST databasecontains sequences that are homologousto stress-relatedgenes in other species, including those that are up-regulated in response to dehydration (Frank et aL, 2005). P. patens is a dehydrationtolerant species,and plants have beenshown to have the ability to regenerateafter losing 92% of their fresh weight after dehydration treatments(Frank et aL, 2005). Conversely,desiccation tolerant speciesare able to regenerateafter losing over 99% of their fresh weight, and they possessspecific mechanismsthat reducethe effect of desiccationand repair the cellular damage causedby it (Oldenhof et aL, 2006; Oliver et aL, 2000). To fully explore the 21 mechanismsthat allowed plants to coloniseland, desiccationtolerant speciesshould be studied,and bryophytessuch as Tortula ruralis are being developedas models for this purpose(Oliver et aL, 2000; Oliver et aL, 2005).

Although the ftinctions of somegenes have been inferred through interrogationof the genomesequence and EST database,the function of the majority of genesin P. patens remainsunknown (Rensinget aL, 2008). A significant step on the road to relating the genomesequence to geneftinction hasbeen the production of a high density linkage map (Kamisugi et aL, 2008). This linkage map was producedusing both AFLP and microsatellitemarkers, and the AFLP markersused have been anchoredto the genomesequence by convertingthe AFLP loci into sequence characterisedamplified length polymorphisms(SCALPs). The developmentof these

AFLP markersprovides a potentially significant resourcefor ecological and evolutionary genetics.Any AFLP loci that are found to be of interestthrough geneticassociation with evolutionarily important traits could potentially be usedto identify sequencedgenes subsequently found to be presentas markersin the linkage map. 22

1.5 Aims of the project

The model mossA patens is one of the most well resourcedmodel plant speciesand is certainly the most well-characterisedbryophyte, in terms of its molecularbiology.

Nevertheless,there is virtually nothing known of the ecology and population geneticsof this plant. To extendthe variety of studiesthat can be conductedand increasethe utility of the molecular knowledgethat hasbeen gleanedfrom this model, more information about the diversity of this speciesin its natural environmentmust be obtained.The laboratory strainsof P. patens that are currently studiedinclude a number of P. patens accessions,termed "Physcotypes". They compriseisolates from widely different parts of the world, but only a few sporophyteshave beencollected from eachlocality. Although probably unintentional,these Physcotypes are likely to have beenselected based on their phenotypesor suitability for laboratory culture (Alonso-Blanco& Koornneef,2000).

Consequently,a larger collection of plants representativeof the geneticvariation in different populationsis required,if the molecular tools available for geneticanalysis of Physcomitrellaare to be effectively mobilised in the study of its population genetics.This requirementhas underpinned the investigationdescribed in this thesis: the first study of the geneticdiversity of P. patens in a variety of natural populations,using multiple samplesderived from a numberof sites.

P. patens is an interestingmodel in which to answerevolutionary questions.Being a haploid, inbreeding,and ephemeralbryophyte, it has a life history in contrastto most well studiedplant species.In the chaptersahead I will examinethe level of geneticvariation within natural populationsof P. patens and how this diversity is partitioned spatially within and betweenspecies. I will also addresshow the mating systems operating within populations of bryophytes can be investigated, by trialling 23 a novel method of analysis, and applying this to natural populations of P. patens. I have also identified a population of P. readeri in West Yorkshire, which is a bryophyte new to Europe. By conducting this study I hope to characterise the population genetic processesoccurring within natural populations of P. patens, so that this may ftu-ther the understanding of the ecology and evolution of ephemeral plants and bryophytes in general, about which little is currently known. 24

Chapter 2 General Materials and Methods

2.1 Study sites

BetweenAugust 2005 and November2006, maturesporophytes of P. patens were collectedand population datawere recordedfrom nine populationsin Great Britain.

Populationsare generallydefined by the limits of the pool, lake, ditch or other such habitat in the vicinity that they occupy, so populationsare thereforediscrete but hugely variable in size. Populationswere chosenaccording to a hierarchicalspatial scaleprotocol, with regionsand populationsseparated by distancesof 5-20km, 50-

1OOkm and I OOkm-400km.Population details are given in Table 2.1with their distribution in the UK shown in Figure 2.1. Photographsshowing examplesof the habitatsthat P. patens occupiesare given in Figure 2.2.

When identifying possibletarget populations,a total of 42 locationswere examined, basedon previous recordsof P. patens at theselocations. Of theseonly II locations were found to contain P. patens plants. The details of theselocations and the historical recordsof P. patens are given in Appendix 1.The successof discoveryof populationswas generally dependanton the ageof the record, and often plants were not found in locationswhere the recordswere over 10 yearsold. However,

PapercourtMarshes in Surreywas found to contain a large population of P. patens, yet the record was datedfrom 1973.In many populationsthat had only recently been recorded,evidence of P. patens was still not found. In certain cases,such as at

BoshertonLakes in Pembrokeshireand severallocations in Cambridgeshire,exact

GPS grid referenceswere provided by bryologists who had found P. patens only a few yearspreviously. In theseinstances it is likely that the water levels in the lakes, 25 reservoirs and strearns in which the populations inhabit were too high, and that the

POPLIlationswere submerged. Other locations were found to contain established plant communities, with little evidence ofaquatic areas, indicating that the area that the population had previously occupied had since become too dry fior P. palens. It can be concluded that P. ptilens is very much dependent on a regular cycle of inundation and exposure for the persistence ol'populations.

Figure 2.1 Location ot'popLilations of P. palctis (e) sampled. The distribution of' populations within Cambridgeshirc is shown in greater dctail in b). 26

m CN rn rn cý rq en en 00 00 rn CD clý le "0 ýo 10 110 Cl% 00 00 Glý le ON 0 00 kf) 00 IC 110 00 ItT C) ý rn \M r- clý -4 CD rn -ZJ Q rn le rn 00 .-r. ýIM vi vi -4 Mt "0 CD Ilt rn rn r- "0 --4 ýo rn Ilt CD 3 rn rq rA * Z CD CD v w cn Ln Ln )--3 w4 (A

"C Ici

0 %Z. i -2 ig ig "0 .!:0 r. li -2 0 "C cl. = (k) ci 0

ce ,

Z,. zi "Ci "0 " Ci 10 0 . 1-, r GM 1.1 0 r

0

ri "Zi

0 0 0 Z: LZ 4 u Z) Ln 9.) -'-, -0 rA = jm. - Ln %Z.,Q CU 0 -bd0 CD. ýg:i u Czu . r. 00 ce M M cu> > *-ý 04 12.4s-2 P.4 (D L gL4 CD cn b P. "0

i- 27

LL.

cz

I

'V

LL 4 M

. "a : czý t fa 1-1

F

. 1.

" -"

$1'

«

LLJ IQ ý1- 28

Sporophyteswere collectedaccording to a hierarchicalspatial scaleprotocol within populations,with the sporophytescollected that were separatedby distancesin the region of 10cm- 50cm, I m-5m and 20m - 50m within eachsite. An exampleof the sampling structureis shown in Figure 2.3. Within-site distanceswere measured betweenconsecutive samples and bearingswere taken betweensamples, so the points where sampleswere collected from could be convertedto x and y coordinates,to producea plan of eachsite and enabledistances between each pair of samplesto be calculated.Mature, undehiscedspore capsules were collectedand storedin 1.5ml microcentrifugetubes. Population area was recordedby walking aroundthe perimeterof the population with a geographicalpositioning systemand recordingthe perimeterposition at regular intervals.The areawithin this perimeter was then calculatedusing ArcGIS 9.1 (ESRI, Redlands,California, USA).

The density of P. patens around each individual collected was also measured, using quadratsof I Ox10 cm and IxIm. The quadratswere madeup of a7x7 grid comprising49 cross-overpoints, and the occurrenceof a P. patens plant beneath eachpoint was recorded,with the samplecollected as the centrepoint. 29

30

25

20

8 15

10

05 10 15 20 25 30 Distance(m)

Figure 2.3 Positions of P. patens sporophytes that were collected from Pant-y-Ilyn in

Carmarthenshire.Each labelled point indicatesclusters of 1-3 sporophytesseparated by distancesof lessthan 50cm. 30

2.2 Sample culture and DNA extraction

Intact sporecapsules were rinsed in sterile distilled water, and then transferredto domesticbleach (ca. 5% free C12)for five minutesin a sterile 1.5ml microcentrifage tube. Capsuleswere rinsed twice in sterile distilled water, transferredto a clean tube and crushedin Iml sterile distilled water, to releasespores. Spores were germinated on BCD medium (Knight et aL, 2002) containingNI-14 (5mM ammoniumtartrate) and I OmM CaC12under constantillumination at 25'C. Following germination, individual plants were transferredto BCD medium containingNI-14 and ImM CaC12. and grown under constantillumination at 25'C. For DNA extraction,individual plants were grown on medium overlaid with cellophanefor isolation of DNA as describedby Knight et aL (2002).

2.3 AFLP analysis

AFLP analysis(Vos et al., 1995)of eachplant was conductedas describedby

Myburg and Remington(2000). Approximately 100-200ngDNA was digestedwith

EcoRl and Msel and ligated with the adaptersequences described by Myburg &

Remington(2000).

AFLP pre-amplification reactions were conducted using EcoRl and MseI primers with a single selective nucleotide (EcoRl: GACTGCGTACCAATTCN; Msel:

GATGAGTCCTGAGTAAN) and selective amplification reactions were conducted with either one or all of the following primer combinations (see Chapters 3 and 5 for specific details): EcoRI+ACIMsel+CG, EcoRI+AGIMsel+CT and

EcoRI+CGIMsel+TA. The primer combinationsselected were chosenas they producedeasily scorabletraces with polymorphic loci, basedon a trial of AFLP 31 primer combinationsby Kamisugi et al. (2008) for production of the P. patens genetic linkage map. TheseEcoRI primers were fluorescentlylabelled with FAM,

HEX and NED respectively,so that the length of fragmentscould be scored automaticallyusing a capillary sequencer.PCR reactionswere conductedas follows accordingto a modified method of Beffes (2001). EachPCR reaction (I 5gl ) containedPCR buffer, 0.2mM dNTP mix, 18ngEcoRI primer, 90ng MseI primer,

0.15ttl Taq polymerase,and 1.5ttl pre-amplificationproduct (diluted 1:50). Cycle parameterswere 12 cycles of denaturationat 94T for 10 seconds,annealing at 65'C for 30 secondsand extensionat 720Cfor 60 seconds- 0.7'C/cycle, followed by 24 cycles of denaturationat 940Cfor 10 seconds,annealing at 560Cfor 30 secondsand extensionat 72'C for 60 seconds+I second/cycle,with a final extensionat 72'C for

2 minutes.Amplified AFLP fragmentswere then electrophoreticallyresolved using an ABI 3130 GeneticAnalyser (Applied Biosystems,Foster City, USA). 32

Chapter 3 Genetic structure within and between populations of

patens

3.1 Introduction

P. patens provides an interesting model in which to study evolutionary questions as it is an ephemeral,inbreeding species, which has a numberof implications for the population geneticsof this organism.P. patens residesin aquatichabitats such as reservoirsand ponds (Smith, 2004), with recedingwater levels exposingpopulations

(Cove, 2005). Thus P. patens populationscould be expectedto exhibit featuressuch as low levels of geneticdiversity and high fecundity, as observedin ephemeral flowering plant species(Barrett & Husband,1997).

An important aspectof the population geneticsof plant speciesis the spatial structureof plant geneticvariation within and betweenpopulations, yet this hasbeen little studiedin bryophytepopulations. Where this hasbeen investigated it hasbeen studied in dioicous (Cronberg,2002; Cronberget aL, 1997;Gunnarsson et aL, 2005) or epiphytic bryophytes(Snall et aL, 2004), which would be expectedto exhibit very different population biology to a monoicousephemeral such as P. patens.

The spatial geneticstructure of individual populationscan be studiedin plant populationsusing spatial autocorrelationanalysis whereby a measureof genetic similarity betweenpairs of individuals, suchas the kinship coefficient, is correlated with the position of individuals in a mappedpopulation (Heywood, 1991;Vekemans

& Hardy, 2004). There are a numberof different methodsfor estimatingthe genetic relationshipbetween a pair of individuals when using spatial autocorrelation 33 analysis,such as thoseby Ritland (1996), Queller and Goodnight (1989), Li et al.

(1993), Hardy and Vekernans(1999), Lynch and Ritland (1999) and Wang (2002).

However, Loiselle et aL (1995) has developedan estimatorbased on that of Hardy and Vekemans(1999), which is itself basedon Moran's I statistic,that performs well in highly selfing species(Vekemans & Hardy, 2004), and it has beenused successfully in spatial autocorrelation analysis of a population of an epyphytic bryophyte by Snall (2004).

In this study I investigatedthe level of geneticdiversity within populationsof P. patens, along with the spatial geneticstructure both within and betweenpopulations.

I have usedamplified fragmentlength polymorphism (AFLP) analysis,as it is a robust and reliable geneticmarker which is highly reproducible,with many loci amplified in individual polymerasechain reactions(Blears et al., 1998;Ritland &

Ritland, 2000). Maximising the numberof loci to be sampledis important when studying diversity within POPulationsas the geneticvariation betweenindividuals may be low. I also investigatedthe spatialpartitioning of geneticdiversity between populationsby structuringthe distancebetween populations at differing spatial scalesas well as by samplingwithin populationsat multiple spatial scales.

3.2 Materials and methods

For a descriptionof the populationsof P. patens sampled,see Chapter 2, Table 2.1.

For details of the samplingprocedure, sample culture, DNA extractionand AFLP analysis,see the relevant sectionsin Chapter2. The numberof sporophytessampled from eachpopulation studiedin this chapteris given in Table 3.1. 34

Table 3.1 Populationssampled, including the numberof sporophytescollected at eachpopulation.

Population Population/ No. of ID Location sporophytes

A Pant-y-Ilyn 19

B Llanflhangel 34 Gobion C Stokesay 20

D Papercourt 32 Marshes E Ouse Washes 27

F Wicken Fen 24

G Swavesey 15

H Mepal 3

I Lindley Wood 37 Reservoir

3.2.1 AFLP profile scoring and error-rate testing

AFLP fragments to be scored were initially filtered using Genemapperg Software v3.7 (Applied Biosystems).Only unambiguousfragments that were between70 and

500 bp in length were included in the analysis.The fragment fluorescenceintensity, or peak height data that was producedby Genemappergwas then convertedinto a genotypetable, where AFLP loci were scoredas either present(1) or absent(0) in eachindividual, using the free r-script basedsoftware, AFLPScore version 1.3

(Whitlock et aL, 2008). AFLPScorenorinalises peak height databetween 35 individuals, and scoresloci as either presentor absentby determiningthe optimum scoring conditionsbased on error-ratetesting. The error rate testing procedure determinesthe loci to be scoredbased on the averagepeak height at eachlocus, or the locus-selectionthreshold, and the relative height at which fragmentsare scored as presentor absent,or the genotype-callingthreshold. The locus-selectionthreshold and the genotype-callingthreshold were chosenbased on the % mismatcherror rate, and the number of loci retained.A binary matrix of AFLP profiles was then producedby AFLPScorebased on thesecriteria.

To estimatethe rate of error within the AFLP procedure,each step of the process was conductedon two replicatedsamples from 24 individuals. Each of these24 sampleswere sub-dividedinto two separatetissue samples, with separateDNA extractions,AFLP reactionsand profile scoringwas then conductedindependently on eachset of 24 samples.

3.2.2 Data analysis

3.2.2.1 Genetic diversity and linkage disequilibrium

Within-population diversity estimatesbased on allele frequencieswere calculated using AFLP-SURV version 1.0 (Vekemanset al., 2002), basedon methodsused by

Lynch and Milligan (1994). The proportion of polymorphic loci at the 5% level, and

Nei's genediversity index (Nei, 1973)were calculatedfor eachpopulation. The correlationbetween the Nei's genediversity index within eachpopulation and population areawas testedby calculating the Spearmanrank correlationcoefficient using XLSTAT (Addinsoft, 2008). The presenceof linkage disequilibrium, or the non-independenceof loci, was testedfor using the programmeLIAN (Haubold &

Hudson,2000). LIAN testsfor statistically independentassortment of allelesby first 36 computing the numberof loci at which eachpair of samplesdiffers. LIAN then calculatesthe varianceof the mismatchvalues (VD), and comparesthis to the

linkage (V,,). The hypothesisHo VD Ve, varianceexpected under equilibrium null : = was testedby a Monte Carlo simulation test using 10,000re-samplings of the data

(seeHaubold & Hudson,2000). A measureof linkage, the standardisedindex of association(I-' A ), was also calculatedin LIAN as:

(VD ,A-s=1 1-1 V,

Where I is the numberof loci analysed.

3.2.2.2 Genetic structure within and between populations

To analysegenetic differentiation within and amongpopulations, Wright's Fst

(Wright, 1978)was calculatedand an Analysis of Molecular Variance (AMOVA,

Excoffier et al., 1992)was performedin ARLEQUIN (Schneideret al., 2000). The significanceof variancecomponents and Fst was testedusing a non-parametric permutationmethod, as describedin Excoffier et al. (1992), using 2000 permutationsof the data. PairwiseFst valuesbetween populations were calculated using the free statistical software package SPAGeDi (Hardy & Vekemans, 2002).

The relationshipbetween pairwise Fst and geographicaldistance between populationswas testedby conductinga Mantel test using XLSTAT.

The kinship coefficient betweenpairs of individuals within and amongpopulations was calculatedaccording to Loiselle et al. (1995) using SPAGeDi (Hardy &

Vekemans,2002). Kinship coefficients,or coancestrycoefficients, are basedon the 37 probability of identity of two homologousalleles, pi andpj, sampledrandomly within two mappedindividuals, i andj (Hardy & Vekemans,2002), or as defined in

Sn5ll et aL(2004) for haploid bryophyte individuals as;

j5)(pi- j5) n+

where the first term is the expectedvalue of py, T is the averagefrequency of the allele in the sampleand k =n(n - 1)/2 is the total numberof possiblepairwise connectionsbetween n samples.The secondterm adjustsfor bias attributableto finite size, and results in py having an expectedvalue of zero for a population in

Hardy-Weinberg equilibrium. The fine-scalegenetic structure amongst all individuals and within populationswas estimatedusing spatial autocorrelation analysis(see Vekemans & Hardy, 2004), which was also calculatedin SPAGeDi.

All individuals within the datasetwere first pooled and the kinship coefficients of all pairs of individuals were binned into physical mappeddistance classes of

10m, 10-50m,50-500m, 500m -I 00krn and 100-400km.SPAGeDi calculatedthe meankinship coefficient for eachdistance class, and the significanceof this kinship coefficient (significantly different from 0) was testedby performing 2000 permutationsof the data.The observedmean kinship coefficient was then compared to the 95% confidenceintervals (CIs) producedfrom the permutedkinship coefficients.This procedureis analogousto a Mantel test (Hardy & Vekemans,

2002). This spatial autocorrelationanalysis was then repeatedwithin populations, excluding population H from Mepal in Cambridgeshire,as only three sporophytes were collected from this population. Distanceclasses of <1Orn, 10-50m, 50-1 00m,

100-200mand 200-500m,were usedin theseanalyses, however not every distance

LEMSONAIFR. qITV 38 classwas usedin all analyses,as the power of this procedureis dependenton the number of pairs within eachdistance class. Only distanceclasses with at least30 pairs in eachwere usedfor eachwithin-population analysis.

3.3 Results

After initial analysisof a subsetof sporophytescollected from all populations studied,one population, at Lindley Wood Reservoirin West Yorkshire, was found to have very different AFLP profiles to thoseobtained from samplesfrom all other populations.A preliminary analysisof pairwise Fstbetween populations (calculated using SPAGeDi), and geographicdistance (Figure 3.1), showedthat the inclusion of plants from the population at Lindley Wood distorts the data. It has sincebeen confin-nedthat the samplesfrom Lindley Wood were not infact P. patens, but P. readeri, and further details of this are discussedin Chapter4. 39

1

A AA AAAA 0.8

0.6

25 L 0.4

0.2

0

-n i 0 50 100 150 200 250 300 350 Distance (km)

Figure 3.1 The relationshipbetween pairwise Fstbetween populations sampled and geographicdistance between populations, calculated using a subsetof samplesin a preliminary analysis.Fst betweenplants from Lindley Wood Reservoirand all other populations= A, and calculationsbetween all other populations= m.

3.3.1 Genetic diversity and linkage disequilibrium

From a total of 233 loci that were scoredusing the threeprimer combinations,152 of thesewere found to be polymorphic (65.5%). The AFLP analysiserror rate, averagedover all loci, was 3.00%, which is within reasonablebounds for AFLP analysisin plant populations(Bonin et aL, 2004). The proportion of polymorphic loci, expressedas a percentage,in eachpopulation is shown in Table 3.2 along with estimatesof Nei's genediversity. Theseestimates were plotted againstpopulation area(Figure 3.2). There is a negativecorrelation betweenpopulation areaand the 40 genediversity index within populations(correlation coefficient, r, = -0.716), however this correlationis not significant (P = 0.088). There is no significant difference betweenthe genediversity indices in populationsB to F, however,the level of diversity within population A from Pant-y-Ilyn in Carmarthenshire,was higher than that found in other populations.The genediversity index within this population was shown to be 0.11875,compared with an averagegene diversity index of 0.04536for all other populations.

Significant linkage disequilibrium was found in the sevenpopulations analysed

(P<0.001in all cases),and the standardisedindex of associationwas similar in all, ranging from 0.0 122in OuseWashes to 0.0540in Pant-y-Ilyn (Table 3.2).

0.14

0.12

0.1

LI) 0.08

0.06 Y) (L) z 0.04

0.02

0iIIIIi 0 1000 2000 3000 4000 5000 6000 7000 Populabon area (M2)

Figure 3.2 Correlation betweenNei's genediversity within populationsof P. patens (excluding population H) and populationarea. r= -0.714, P=0.088. 41

0 Ci.

C> C) 't %ýo eq en 00 't r- en 00 C14 tn

cu en Nt ý10 r- t- N W) C) 00 0 W) C) 110 t- 110 r- r- C) ýT C) C) C) C) C) C) "0 C) C) C) C) C) r. C; 6 (6 6 C; ci 6 -, C> C14 00 00 "Ci Iw W') en en -. 4 Q) r- en kf) It 10 I'll ON W) V5 00 c) p Itý,p It"o ItW) eqkil Itr C> C> 0 o 0 0 Cý) C> Ici c; c; c; c; 0 -0 r. 0

en (0 r- en (2ý 00 0 CN 6 cli cli it C14 r-

0 01.

Z (4-4 0

0 0 Z ON Itt C) C14 r- Itt tn A cq m C-4

0 0 0 . .2 IZ ,a 0 W

r. ýj cl M .0 cl 0 w 03 w U Q

cl 0 0 C3 0

0 P64 42

3.3.2 Genetic structure within and between populations

The resultsof the AMOVA analysis,which was usedto test for genetic

differentiation betweenpopulations of P. patens,are shown in Table 3.3. These

resultsindicate that most of the variation was partitioned within populations,at

93.18%,compared to only 6.82% amongstpopulations. A significant fixation index

(Fst) of 0.062 indicatesthat there is only moderategenetic differentiation between populations(Wright, 1978).

The relationshipbetween pairwise Fst and geographicaldistance between

populations is shown in Figure 3.3. Analysis of this relationship using a Mantel test

indicatedthat there was no significant relationshipbetween pairwise Fst and

geographical distance (r = 0.396, P=0.081).

Graphical representationsof the kinship coefficient spatial autocorrelationanalysis

are shown in Figure 3.4. When all individuals from all populationswere pooled,

spatial structurewithin and betweenpopulations was detected.For distanceclasses

of 1-10m, 10-50m,and 50-500m,significant positive kinship coefficientswere

observed.A significant negativekinship coefficient was also observedin the

distanceclass 100-400km.For within population analysisof kinship coefficients alone, only population B, at Llanfihangel Gobion in Monmouthshire,contained a distanceclass, at 10-50m,with a kinship coefficient significantly different to zero. 43

Table 3.3 Analysis of molecularvariance (AMOVA)

Sourceof d.f SumOf Variance % P variation squares Among 6 101.292 0.45122 6.82 <0.001 Populations Within 161 992.952 6.16740 93.18 Populations Total 167 1094.244 6.61862

0.16 "U 0.14 U UU 0.12

U 0.1

0.08 LLU 0.06 "U 0.04 "U " "U U"" 0.02 'I U

U 0 II

-002 0 50 100 150 200 250 300 350 Distance (km)

Figure 3.3 The relationshipbetween pairwise Fst betweenpopulations sampled and geographicdistance between populations. r=0.396, P=0.08 1. 44

a) b) 0.25 0.25

c 0.15 0.15

005 0 Oý05 ------05 .E -0 -0.05 c ýz -015 ------0,15 -0.25 -0 25 ---- 0001 0.1 10 10 100 1000 Distance (m) C) 0.25 0.25 C

0.15 0,15

0 0.05 005 ------&05 -0,05 -- - c ------ýý 15 -0 -015

-0.25 -0.25 10 100 1000 10 100 1000 e) 0.25 025 D E 0.15 0.15

0.05 005 ------0 ------1ý------u - 77 E -005 ------0 -0.05 c -0.15 -0.15

-0.25 -0.25 10 100 1000 10 100 1000 h) 0.25 0.25 F G 0.15 0.15

=: 0,05 ------OD5

-&05 ------0ý05

-0,15 -0.15

-025 -0,25 10 100 1000 10 100 1000 Distance (m) Distance (m)

Figure 3.4 Kinship coefficient autocorrelation at a) over all sites, and b) to 11)within populations A to G. Mean kinship coefficient of each distance class is plotted against maximurn class distance. Significant coefficients are indicated as P<0.05 =

*, P<0.01 = ** and P<0.001 = ***, based Oll the IILIII hypothesis ol'kinship coefficient = 0. The Lipper and lower 95",0 confidence intervals are shown by dashed

I IIIes. 45

3.4 Discussion

The relatively low levels of geneticvariation observedin most of the P. patens populationsare consistentwith the levels of variation observedin other ephemeral plant species(reviewed in Barrett & Shore,1989). Although low levels of genetic diversity are often associatedwith small population size, this doesnot appearto be the casein populationsof P. patens. Barrett and Husband(1997) found that population size had no significant effect on diversity when using allozyme loci to measureheterozygosity in populationsof Eichhorniapaniculata, an aquaticplant occupying ephemeralhabitats. However, the authorsdid find a significant positive correlationbetween the harmonic meanpopulation size over a three year period and heterozygositywithin populations.As populationsstudied were not static over this time but variable in size, any measureof heterozygositytaken in a single year would havebeen affected by the size of the populationsin previousyears, particularly if the population was much smaller.The authorsof this study also found a significant positive correlationbetween population size and diversity in morphological characters.This may be due to the fact that allozymes,like other molecularmarkers such as AFLPs, are selectivelyneutral, and analysingmorphological characters under strong selectionpressure may yield different results.

The relatively higher level of geneticvariation observedwithin population A in

Pant-y-Ilyn, was, however,unexpected, as this was also the smallestpopulation studied.This could reflect the stability of this population, in terms of population age, along with the frequencyand extent of flooding in this habitat, when comparedwith the other populations.An alternativeexplanation, as discussedabove, is that the current size of populationsof P. patens may not be representativeof population 46

sizesin previousyears. Studies that assessthe impact of population age on levels of

diversity in bryophytepopulations are rare due to the complexitiesin accurately

determiningthe time of population establishment.However this hasbeen

investigatedby Cronberg(2002), who assessedlevels of diversity in populationsof

Hylocomiumsplendens in an archipelagowith islandsof known age. In this study

geneticdiversity was found to increasewith increasingage, which could be

explainedby the constantrecruitment of new allelesby populations.

Although no data are availableto corroboratethese theories, it is likely that where an ephemeralpopulation is relatively stable,and in the caseof this P. patens population, where flooding in the habitat is infrequent,that a higher level of genetic

variability may be accumulatedor maintained.To datethere havebeen very few

studiesinto the dynamicsof geneticstructure within ephemeralplant populations.A

number of studieshave investigatedthe level of geneticdiversity within speciesthat

occupy different successionalstages (reviewed in Barrett & Shore, 1989),and

significantly lower level of within population diversity havebeen found in species

that inhabit early successionalstages, compared to thoseoccupying mature habitats.

However, as thesestudies generally involve different speciesat different

successionalstages, it is difficult to extrapolatethese results.

When examining the spatial structureof the pooled populations,a significant kinship

coefficient was only observedwithin distanceclasses of lessthan 500m, or within

populations,with no spatial structurebetween populations. This indicatesa strong

restriction to geneflow within populations.A similar pattern in terms of the

distancesat which isolation by distanceoccurs in bryophyte populationswas also

found by Snqll (2004), with a significantly positive kinship coefficient found at 47 distancesbetween samples of up to 300m. However, in analysesof individual populations,the only population to exhibit a significant kinship coefficient within a distanceclass was population B at Llanfihangel Gobion. This suggeststhat populations are not genetically structured within populations, or more accurately, within the pools they occupy.The significant kinship coefficient in population B from 10-50mfurther corroboratesthis, as this population was actually divided into two subpopulations,separated by dry land that is not flooded, and by distances greaterthan 50m. A significant kinship coefficient at distancesup to 50m but not in classesof greaterdistances, suggests that thesesub-populations are genetically distinct.

In pooled population spatial autocorrelationanalysis, a negativekinship coefficient was found in distancesbetween individuals of greaterthan I OOkm,as well as within population B at Llanfihangel Gobion, at distancesof 100-200m.A negativekinship coefficient, which suggeststhat the kinship coefficient is lower than that expected betweenany pair of individuals picked at random from a sample,seems unlikely, and this finding may be an artefactof the methodology.However, there may be strong isolation by distance,particularly at large distanceclasses relative to the analysedsample, which could result in negativekinship coefficients relative to the whole samplefrom all populations.A negativekinship coefficient at distanceclasses of over 300m was observedby Snall et al. (2004), in an epiphytic bryophyte population, as well as by Bonnin et al. (2001) in Medicagopopulations.

Linkage disequilibrium was found in all populationsof P. patens analysed,which is as expectedin a selfing species.The level of linkage disequilibrium found was either within or slightly higher than that found in other monoicousbryophyte species 48

(Gunnarssonet aL, 2005; Snall et aL, 2004), howeverit was generally less than that of dioicous bryophytes(Grundmann et aL, 2007; Grundmannet aL, 2008). As the increasedlevel of linkage disequilibriurn found in dioicous bryophytescan generally be explainedby the clonal natureof thesespecies, the comparativelylower linkage disequilibrium, along with the lack of spatial structurefound within populations, suggeststhat asexualdispersal is not commonin P. patens.

The level of geneticdifferentiation betweenpopulations of P. patens, where most geneticvariation is partitioned within, ratherthan betweenpopulations is consistent with levels observedin other bryophyte species(Gunnarsson et aL, 2005; Zartman et al., 2006). However, the lack of spatial structurefound within populationsis more unusual.As P. patens is known to readily self fertilise in the laboratory,and as gene flow through gametedispersal is thought to be severelyrestricted in bryophytesin general(Shaw, 2000) it may be expectedthat natural populationsmay be highly spatially structuredwith patchy genotypicdistribution (Bonnin et aL, 2001).

However, few studieshave investigatedspatial autocorrelationanalysis in either selfing speciesor bryophyte populations.Most analysesof fine-scalespatial structurewithin plant populationshave focussedon largely out-crossingspecies, occupying stablehabitats with genedispersal by seed.As P. patens occupiesvery different habitatsand exhibits a very different life history to thesewell studied species,the trend observedin populationsof P. patens could be expectedto be very different. As the ephemeralhabitats that this speciesoccupies can be extremely variable in both size and existencefrom one year to the next, it could be expected that taxa with such ecology would show little within-population spatial structure.

There havebeen no other investigationsinto the fine-scalespatial structurewithin ephemeralplant populations,but Bonnin et aL (2001) has usedkinship 49 autocorrelationanalysis within populationsof an inbreedingMedicago species.

Here, a high level of spatial structurewas only found in the sub-populationthat was subjectedto the leastdisturbance, and no spatialgenetic structure was found in a sub-populationthat was ploughedregularly. If the populationsof P. patens within the pondsand lakesthat they occupy are periodically submergedthroughout the fruiting season,this action could act as disturbancein a similar way. Although gametedispersal may be restrictedin mosses,spores have the potential to be carried in air or water over many metresor kilometres.Thus sporedispersal could account for much of the lack of geneticstructure within populations,as alleles could be dispersedalmost at random after sufficient rainfall.

This study is the first investigationinto the population geneticsof P. patens, and one of few into the population of ephemeralbryophytes. Populations of P. patens have beenshown to exhibit geneticstructure that is symptomaticof its ephemeralnature, which may provide further avenuesfor investigation.AFLPs were used successfully in this study to investigateboth the level of geneticdiversity and the structureof this diversity within populationsof P. patens. There is a degreeof error within this analysisthat is not so presentin molecularmarkers such as simple sequencerepeats and isozymes,that may causea low level of spatial structuringwithin populationsto be missed.However, this is negatedby the efficiency of AFLPs given that they provide a large numberof loci for utilisation. Also the fact that spatial genetic structurewas found betweenpools a few hundredmetres apart, but not within the pools, further confirms the role of pool mixing in breakingup spatial genetic structurewithin populationsof P. patens when they occupy continuoushabitats. The lack of spatial structurefound within populationsof P. patens is consistentwith that 50 found in other speciesof bryophytesanalysed by using AFLP analysis(Sndll et aL,

2004), which suggeststhat the results found in this study are robust. 51

Chapter 4 P. readeri in Europe

4.1 Introduction

The mossgenus Physcomitrella Bruch & Schimp.currently consistsof one species recordedin Europe,P. patens (Hedw.) Bruch & Schimp.P. patens has a geographicaldistribution within the northernhemisphere, extending from the U. S.A to Europe(excluding the Mediterranean),Scandinavia and Siberia (Smith, 2004;

Tan, 1979).The only other acceptedspecies in the genus,P. readeri (C. Well. )

Stone& Scott, hasup until now beenrecorded only from California, U. S.A,

Australia, New Zealandand Japan(Dixon, 1926;Goffinet, 2007; Ochi, 1968; Scott

& Stone, 1976).The taxonomic statusof P. readeri is contentious;it hasbeen classified as a subspeciesof P. patens. Tan (1979) arguedfor the subdivision of P. patens into four subspecies,based mostly on geographicaldistribution. Thesewere; subsp.patens, comprisingmost plants from the Northern hemisphere,subsp. readeri from Australia, subsp.californica from the U. S.A and Japanand subsp.magdalenae from Rwanda.P. readeri is now acceptedas a distinct speciescomprising the following synonyms;Ephemerella readeri C. Milell., Physcomitridiumreaderi (C.

Well. ). Roth, P. austro-patensBroth. ex Roth, P. patens ssp.californica (Crum &

And.) B. C. Tan and P. californica Crum & And. (Goffinet, 2007; Stone& Scott,

1973).P. readeri is mainly distinguishedfrom P. patens by a shorterleaf midrib, or costa,and argumentsfor its designationas a separatespecies are partly basedon the geographicaldistribution of plants with this feature(Goffinet, 2007).

Where natural populationsof P. readeri havebeen found, they occupy the inundation zonesof reservoirsand lakes,on bare, exposedmud after water levels 52 have fallen, and they havebeen found with P. patens in the U. S.A (Goffinet, 2007).

During collection of plants from a populationof P. patens in West Yorkshire in

November2006,1 collectedplants with sporophytesthat appearedto differ from typical P. patens. Further analysishas shownthe plants sampledto not be P. patens, but P. readeri, thereforerecording the first population of this speciesin Europe.The areavisited, at Lindley Wood reservoir in the WashburnValley, was initially visited as P. patens had beenrecorded at the reservoirin 2001 (Hodgetts,N., pers. comm.).

However,with further analysisof the plants, differencesbetween those collected from Lindley Wood reservoir and other samplesof P. patens in the collection were discovered.

AFLP profile analysis(see Chapter 3) was conductedon plants cultured from spores from the collected sporophytes.Kinship coefficientswere calculatedbetween all plants analysed,including a minimum of 13 plants from eachof the 7 populationsof

P. patens in the UK and 14 plants cultured from the sporophytcscollected at Lindley

Wood reservoir.The AFLP profiles from plants collected in this population were found to differ greatly from the profiles of plants from the other 7 populationsof P. patens sampledin the UK, as well as accessionsfrom Germany,France, Switzerland and the Ukraine.

Upon culturing the Lindley Wood samplesfor production of gametophytictissue and sporophytes,further morphologicaldifference betweenthese plants and P. patens becameevident. To clarify the identity of the plants from Lindley Wood reservoir,analysis of the internal transcribedsequence (ITS) region of ribosomal

DNA was conductedon an individual plant cultured from sporophytescollected at

Lindley Wood Reservoir,along with confirmed P. patens plants and 4 other species of bryophyte within the Funariaceae.This chapterdetails the findings of this 53 analysis,which along with morphologicalobservations, confirm the discoveryof a speciesof bryophyte new to Europe.

4.2 Materials and methods

4.2.1 Population details and sample collection

Sporophytesof P. readeri were collected from a population at Lindley Wood reservoir in West Yorkshire, UK in November2006 (OS grid referenceSE210498).

For the samplingprocedure followed, seeChapter 2. Figure 4.1 showsthe location of the population in the UK, and a photographdepicting the site is given in Figure

4.2.

4.2.2 Sample culture

For details of sporeculture from intact sporophytes,see Chapter 2. When mature plants had beenobtained, individual P. readeri plants were cultured on BCD medium (Knight et aL, 2002) containing I mM CaC12,and grown at 150Cfor 7 hours light over a 24 hour period. Reducingthe temperatureand day length in the growth chamberis routinely usedto induce sporophyteproduction in P. patens (Cove,

2005), and this was trialled in P. readeri, as sporophytesgreatly aid morphological analysis.Individual P. patens plants were also cultured as above,to enablea morphologicalcomparison with the plants collected from Lindley Wood reservoir.

For ITS sequenceanalysis, the following plants were cultured for DNA extraction, as detailedin Chapter2: P. readeri from Australia and Japan,P. patens from the UK and France,Physcomitrium pyriforme (Hedw.) Hampeand Entosthodonobtusus

(Hedw.) Lindb. Provenancedata for thesesamples can be found in Table 4.1. Two plants collected from the Villersexel population in Francewere included in this 54 analysis,as the two lab strainspropagated from theseplants have been shown to be quite different. One strain, termed"K4" hasbeen shown to be very similar to

Gransden,whilst the secondstrain, "KY is the most genetically divergent strain of the P. patens physcotypesavailable (seeKamisugi et aL, 2008).

Table 4.1 Provenancedata for samplesused in phylogeny reconstruction

Species Population/ County/Region Country Location P. readeri Lindley Wood West Yorkshire England Reservoir

P. readeri Melton Reservoir Victoria Australia

P. readeri Masu pond Okayama-shi Japan

P. patens Gransden Cambridgeshire England

P. patens Villersexel Franche-Comtd France

P. pyriforme Carbondale Illinois U. S.A

E. obtusus St Agnes Comwall England 55

-f

Figure 4.1 Location of P. i-eadei-i population in West Yorkshire, UK. 56

-0 E

1.0

"a 0

-0

:3 57

4.2.3 Nuclear ITS region analysis

4.2.3.1 Amplification and sequencing

DNA was extracted from plant tissue as described in Chapter 2. The ITS I and ITS2 regions of the nuclear genome were first amplified from each sample using PCR as follows. The primer sequencesused to amplify the entire ITS region were;

TCGTAACAAGGTTTCCGTAGGTG at the 3' end of the 18S rDNA gene and

ATTTTCAAGCTGGGCTCTTTCC at the 5' end of the 26S rDNA gene.

Each PCR reaction(50pl) containedPCR buffer, 0.2mM dNTP mix, 0.3ýM eachof forward and reverseprimer, I gl Taq polymerase,and 2gl pre-amplification product

(maximum concentration20ng/ttl). The PCRswere then carried out with the following conditions over 35 cycles; pre-melt for 2 minutesat 94*C, denaturationat

94'C for 20 seconds,annealing at 58'C for 40 secondsand extensionat 72'C for 60 seconds,with a final extensionat 720Cfor 10 minutes.PCR productswere purified using a QlAquick PCR purification kit (QIAGEN Ltd, Crawley, West SussexU. K).

The purified productswere than cycle-sequencedby the IntegratedGenomic and

GeneExpression Analysis Facility at the University of Leeds,with a separate reaction for eachof the amplification primers. Productsof the cycle-sequencing were run on an ABI 3130 GeneticAnalyser (Applied Biosystems,Foster City,

U. S.A).

The sequenceobtained from the Villersexel "KY strain was found to be a mixed trace containing more than one sequence.The PCR product producedfrom this strain was then cloned into the EcoRV site of pBluescript by Dr Y. Kamisugi, and two separatesequences were provided for this strain. 58

4.23.2 Sequencealignment and analysis

The sequenceswere viewed and assembledusing the ABI trace viewer in BioEdit, version 7.0.9.0 (copyrightO 1999-2005Tom Hall), with the ITS I and ITS2 regions treatedseparately. Successfully amplified regionswere first automatically aligned in a matrix using ClustalW version 2 (Ramu et aL, 2003). The resulting alignments were then manually modified in the SequenceAlignment Editor in BioEdit

(copyrightO 1997-2007Tom Hall) The ITS region sequencefrom F. hygrometrica

Hedw. was obtainedfrom GenBank(Benson et aL, 2006) and included in the alignment matrix.

After sequencealignment, most parsimoniousphylogenetic trees were produced using the phylogeneticanalysis package TNT (Goloboff et aL, 2008), which has been made freely available by the Willi Hennig Society. This was done by performing a branch-and-boundsearch using the implicit numerationoption in TNT, which guaranteesto find all trees.To test supportfor eachclade, bootstrap analysis was performed in TNT, using 1000 replicates.

Maximum likelihood analysiswas also conducted,using the free softwarepackage

TREEFINDER (Jobb,2008). Optimal substitutionmodels for eachITS region were automaticallyselected using the model proposerdialog in TREEFINDER. Of the 32 modelsproposed, an HKY+G model of evolution (Hasegawaet aL, 1985)was selectedfor eachITS region, which was then usedto reconstructphylogenies in

TREEFINDER. Bootstrapanalysis was performedusing 1000replicates. 59

4.2.4 Alorphologcial analysis

Plantscultured from the sporophytescollected at Lindley Wood reservoir were comparedwith the P. patens plants currently in culture in the laboratory. Scanning

Electron Micrographs(SEM) were taken of the Lindley Wood plants using methodologydescribed in Presselet aL, (2007) by J. Duckett. Leaf specimenswere photographedusing a digital cameraunder an Olympus SZX 12 microscope.

Comparisonswere madebetween the two specieswith regardsto sporophyteand leaf morphology.

4.3 Results

4.3.1 ITS sequence analysis

Figures4.3 and 4.4 show the maximum likelihood phylogramsproduced using the

ITS I and ITS2 sequencesfrom the 10 samplesanalysed. The most parsimonious treesare not shown as the topologiesobtained were identical using both methods.

Both treesplace the samplecollected from Lindley Wood reservoir with A readeri from Japanand Australia in separateclades to the P. patens plants analysedfrom the

UK and France.Figure 4.5 also depictsthe relevant sectionof sequencealignment betweenthe known P. readeri samplefrom Australia, and the samplefrom Lindley

Wood, producedusing BLASTN (Altschul et al., 1990).This figure showsthat within thesetwo regionsthere is a single-basemismatch between these samples, in

ITS 1. This result is in contrastwith the percentageidentity betweenthe Lindley

Wood sampleand the P. Patenssample from the UK, where the percentageidentity is only 86% (averagedover all charactersin BLAST alignment).The ITS I and ITS2 region sequencesfrom the Gransdenand Villersexel K4 strainsof P. patens were 60 identical, indicating that there is little variation within species.However, the K3 strain differed, with a percentageidentity betweenthe K4 strain and the two sequencesderived from the K3 strainsof 99%. This strain has beenshown to be the most geneticallydistinct when comparedto all other samplesof P. patens in laboratoryculture, and has thus beenused to producethe P. patens genetic linkage map (Kamisugi et aL, 2008). Theseresults corroboratethe identification of the

Lindley Wood plants as P. readeri, as opposedto P. patens. 61

P. readeri (Australia)

P. readen (Lindley Wood) 88 98

58 P. readeri (Japan)

P. pyriforme

E. obtusus

P. patens (Villersexel "K4") 100

P. patens (Villersexel "KY-l)

99 P. patens (Gransden) 99

P. patens (Villersexel "K3"-2)

F. hygrometrica 1

H gure 4.3 ITS I maximum likelihood phylogram. Likelihood =- 1132. Bootstrap values from the maximum likelihood analysis are shown above branches, with bootstrap values from tile parsimony analysis shown below (I trce, 159 steps, C. 1. -

0.86 and R. 1. ý 0.80). 62

P. patens(Gransden)

(Villersexel X4")

(Villersexel"KY-l)

(Villersexel "KY-2)

P. readen (Australia)

P. readed (Lindley Wood)

P. readeri(Japan)

obtusus

F, hygrometrica 1

Figure 4.4 ITS2 maximum likelihood phylogram. Likelihood = -1225. Bootstrap valUeSfrom the maximuni likelihood analysis are shown above branches, with bootstrap values from the parsimony analysis shown below. (I tree, 167 steps, C. I. -

0.89 and R. I. = 0.84) 63

Score = 852 bits (443), Expect = 0.0 Identities = 555/556 (99ý), Caps = 0/556 (0 ) ýýrand=Plus/Plus

ý"Lery 1 CACACACAAAAACTGCAAACAACCCCCTTTTTGAGAACTAI'ATCTTATTI'G'I'A('AAAA(: ý.'

-t 1 CACACACAAAAACTGCAAACAACCCCCTTTTTGAGAATTATATCTTATTTCI'ACAAAACC

Figure 4.5 Section of the BLAST sequence alignment between P. wadet-i from

Lindley Wood (top strand) and P. reatlei-i from Australia (bottorn strand). The single-base mismatch in the ITS I region is highlighted.

4.3.2 Morphological analysis

Figure 4.6 depicts the SEM photographs of the Lindley Wood plants. The sporophyte of P. readeri has a seta that IS IIILICIIlonger than those ol'P. patens.

Figure 4.7 sho\,N, s a light micrograph of a leaffrom both a Lindley Wood plant and

P. pcitens. The P. pitens leaf exhibits a vvell defined costa, extending to the apex, whereas the Lindley Wood plants have a 111LIchshorter costa, extending only 2/3 the length of tile leaf. This short costa is a diagnostic feature of P. readeri (Goll-met,

2007). For a detailed description of P. ivadel-i see Appendix III, which is adapted from Scott & Stone ( 1976) and Smith (2004). 64

4

Figure 4.6 a-f. SEM images of P. readeri from Lindley Wood Reservoir. a-c) sporophyte, d) leaf, e) stoma f) stomata at base of seta. Scale bars: a, c= 400pm, b, d=I OOpm,c=I Opm, f= 30pm.

ab

______

25 ý21

Figure 4.7 a-b. Light micrograph showing entire leaves from a) P. patens and b) P. readet-i from Lindley Wood Reservoir. 66

ephemeralecology, where gametophyticand subsequentlysporophytic material appearsin autumnfollowing the dropping of water levels in aquatichabitats. This winter inundationhas been found to be essentialfor the re-appearanceof gametophytesin speciesof bryophytesand pteridophytesthat occupy the same habitats,(Duckett & Duckett, 1980;Hill et aL, 1994;Holyoak & Bryan, 2005), and it is likely that this is also the casefor P. readeri. Many of the bryophytesthat occupy reservoir inundationzones also havea very restricteddistribution in the UK,

for exampleMicromitrium tenerum,Physcomitrium eurystomum, Physcomitrium sphaericumand speciesof Ephemerumhave been recorded from only a few locations (Hill el aL, 1994; Holyoak & Bryan, 2005). It is therefore likely, as with the fore-mentionedtaxa, that P. readeri is native to the UK, but under-recorded.As

P. readeri occupieshabitats that are for large periodsof time submergedand in- accessible,it is likely that populationsof this specieshave beenexcluded from speciessurveys. Whilst looking for populationsof P. patens in autumn2006, many habitatsthat may also havebeen suitable for P. readeri were visited, yet no other populationsof this specieswere discovered.It is thereforelikely that P. readeri is rare in Britain.

Although the population of P. readeri at Lindley Wood reservoir was the only population found during the year, it was a particularly large population, and it occupiedthe entire extent of baremud in the reservoir,which coveredan areaof

65,581 m2. Further discoveriesof P. readeri may thereforebe expectedwith greater exploration of reservoirsand lakesduring the summermonths. It is likely to be

found in other reservoirsnearby in the WashburnValley, suchas Swinsty and

FewstonReservoirs, but water levels in thesereservoirs were very high in autumn

2006, which may accountfor the absenceof P. readeri during this period. Like other 67 rare bryophytesoccupying similar habitats,P. readeri is especiallyvulnerable, particularly if water managementstrategies keep water levels high, which would result in the disappearanceof this species. 68

Chapter 5A method for estimation of the mating systems

operating within populations of A patens

5.1 Introduction

As statedin previouschapters, there havebeen relatively few studiesinto the

molecular geneticsof bryophyte populations,and amongthese the investigation of

bryophyte mating systemshas beenparticularly neglected.However, the mating

systemof any species,especially that operatingwithin natural populations,has

important implications for the evolutionary ecology of thesepopulations (Barrett &

Eckert, 1990;Seymour et aL, 2005).

In diploid species,Wright's fixation index (Fis) hascommonly beenused to estimate

the levels of selfing within a population. This approachhas also beenused to

estimatethe out-crossingrate exhibited in the diploid stagein bryophytes(Eppley et

aL, 2007). Eppley (2007) estimatedthe rate of selfing in a number of populationsof

both dioicous and monoicousmoss species, by calculating inbreedingcoefficients

using allozyme markers.In this study, the authorsnot only characterised"selfing" as

the mating of gametesproduced from the sameindividual (intragametophytic

selfmg), but also as the mating of gametesproduced from haploid progeny from the

samediploid parent(intergametophytic selfing). The authorsfound that selfing rates

were far higher in monoicousspecies than in dioicous species,but that there was

evidenceof mixed mating systemsin somepopulations of dioicous species,with

greaterthan expectedlevels of intergametophyticselfing. Eppley et aL (2007)

concludedthat mixed mating occursin populationsof both monoicousand dioicous bryophytes,suggesting that further studiesinto the mating systemsof bryophytesare 69 necessary,as the level of inbreedingand out-crossingin different speciesis largely unknown. However,the methodsused in this study are not as applicableto bryophytesas to seedplants, particularly where geneticanalysis is attemptedas diploid sporophytetissue is not readily cultured.

Innes (1990) studiedthe level of genotypicdiversity within populationsof the dioicous mossPolytrichumjuniperinum, and confirmed that sexualreproduction was present.The authoralso utilised both the haploid gametophyteand the diploid sporophyteto further study the mating systemswithin thesepopulations. In this investigationthe sporophytictissue from plants sampledin the populationsstudied were genotyped,using isozymes,along with the attachedmaternal gametophytic tissueand the male gametophytictissue from within the populations.As the genotype of the maternal parent is easily identified as it is attached to the sporophyte, the genotype of the male parent was inferred from the residual loci.

Using this methodology Innes (1990) was able to determine that female gametophytes were usually fertilised by male gametophytes that were close by, and at most within a few metres of the female plant. Roads & Longton (2003) also compared the genotypes of maternal gametophytes to their attached sporophytes from populations of the monoicous mosses Phascum cuspidatum and Pottia truncala. No variation was found between sporophytes and maternal gametophytes, suggesting a high degree of inbreeding in these populations.

The predominanceof the haploid stagein bryophytescan assistin identifying the breedingsystems found in bryophytepopulations, as haploid siblings can easily be isolatedfrom the diploid sporophyte.Differing degreesof inbreedingwithin a population will result in differing numbersof sporophytesthat contain either homozygousor heterozygoushaploid siblings, with a true inbreedingevent 70 producing homozygoussporophytes and geneticallyidentical gametophytes(Shaw,

2000). This approachhas been utilised to studythe mating systemswithin natural populationsof fungi (e.g. Ennos& Swales,1987; Free et aL, 1996;Marra et aL,

2004; Milgroorn et aL, 1993;Milgroom et aL, 1992)and lichens (e.g. Murtagh et aL,

2000; Seymouret aL, 2005).

The purposeof this study was to investigatea methodfor estimatingthe mating systemsthat are presentin natural populationsof P. patens. As P. patens is a monoicousbryophyte (Reski, 1998),with both archegoniaand antheridiaon the samegametophyte, there is the possibility that selfing can occur, and P. patens has beenshown to producehomozygous offspring due to selfing (Cove, 1983).This was done by first analysingthe genotypesof the gametophytesreleased from a sporophytethat was the result of a known crossin the laboratory.The same methodologywas then applied to natural populationsof P. patens,to determinethe level of out-crossingor inbreedingwithin thesepopulations.

As gametedispersal has beenfound to be locally restrictedwithin most bryophyte populations,with estimatessuggesting that where cross-fertilisationdoes occur that it is betweenplants separatedby distancesof on averageno more than 10cm.

(Longton, 1976),it can be assumedthat plant density would have a pronounced affect on the probability of crossfertilization. Studiesinvestigating the effect of density on out-crossingrates in angiospermshave found that increaseddensity both increased(Wolff el aL, 1988)as well as decreased(Ellstrand el aL, 1978)out- crossingrate. The effect of densitywas thereforealso investigatedin this study. 71

5.2 Materials and methods

5.2.1 Trial using known progeny

To determineif the methodologyto be usedis effective at identifying progeny that are the result of an out-crossingevent, the methodwas trialled using the progeny of a known cross.The gametophytesanalysed in this trial were the progeny of a cross betweentwo laboratorystrains of P. patens: the standardlaboratory strain,

"Gransden"and a genotypically distinct strain "Villersexel KY. Theseprogeny havebeen used to producea geneticlinkage map of P. patens (seeKamisugi et aL,

2008 for details of the crossingprocedure), and they will from henceforthbe referredto as the mappingpopulation. A sampleof 22 gametophytesfrom this population was randomly selected,and the following protocol was used.Individual plants were cultured from laboratory stocksas describedin Chapter2.

The methodologyto be trialled involved pooling progeny to be genotyped,and identifying the presenceof heterozygousindividuals within that pool by comparing the pooled AFLP profile to that of a referenceindividual from the samesPorophyte.

Using this methodology,any additional loci not found in the referenceindividual immediately identifies that sporophyteas the result of a cross,or a "discernible" mating, as describedby Shawet aL, (1981). Where profiles of the pooled sample and referencesample are found to be identical, this is termeda non-discernible mating, as the sporophytecould be the product of either a crosswith a genetically identical individual, or a self fertilisation event.

A similar methodologyhas beentrialled by Marra et aL, (2004), where the authors were investigatingthe mating systemswithin populationsof the chestnutblight fungus, Cr)phonectriaparasitica. In this study the ascosporesfrom a tree canker 72 were pooled, with the DNA fingerprint of the pooled ascosporescompared to the maternalfingerprint, to determineif the progenysegregate at the DNA fingerprint loci. The authorsfound this methodto be successful,and there is thereforepotential for this methodologyto be applied to bryophytesystems, with a great reduction in the experimentaltime and expensethat is involved with genotypingprogeny individually.

5.2.1.1 DNA extraction and AFLP analysis of progeny

The 22 plants selectedwere divided into two replicategroups of II individuals, with

10 plants pooled and one referenceplant in eachgroup. 10 plants were estimatedto be a sufficient numberof individuals to analysein order to identify segregating alleles, as sporophytesthat are the result of an out-crossingevent will comprise progeny with a genotyperatio of 1:1. Thereforethe probability of identifying any one segregatingallele from 10 individuals will be 1-(0.5)10= 0.999. DNA was extractedfrom the pooled plants and the referenceindividual as describedin Chapter

2.

AFLP analysiswas conductedon eachof the pooled samplesand referencesamples, with eachgroup analysedin a separateAFLP reaction.AFLP analysiswas

conductedas describedin Chapter2, using the EcoRI+ACIMseI+CGFAM labelled primer combination.This primer combinationwas chosenas it displayedthe lowest

error rate of the three combinations,at 1.92%,as calculatedin Chapter3. AFLP

analysiswas also conductedon the parentalstrains, Gransden and Villersexel, so

that the sourceof any segregatingalleles could be identified. 73

5.2.2 AFLP profile scoring

AFLP profiles from the pooled and referencesamples were visualisedusing

Genemapper(2)Software v3.7 (Applied Biosystems).The AFLP profiles of each

pooled samplewere comparedto the referencesamples. Polymorphic loci were

scoredas thosepresent in the pooled samplebut absentfrom the reference,

indicating a segregatingallele within the sporophyte.Only unambiguousfragments

that were between70 and 500 bp in length were included in the analysis.

5.3 Analysis of natural populations

To estimatethe mating systemspresent in naturalpopulations of P. patens, the methodabove was applied to 56 sporophytesfrom threepopulations. The populationswere selectedas they encompassedcontrasting areas of both high and

low densityP. patens plants. The populationschosen were; Llanfihangel Gobion

(PopulationB), PapercourtMarshes (Population D) and OuseWashes (Population

E), as detailedin Chapter2. In Llanfihangel Gobion and PapercourtMarshes 10 sporophyteswere selectedrandomly from within both the high and low density areas.From OuseWashes only 8 sporophyteswere available from eacharea, so all were used.

The AFLP genotypeanalysis was conductedas above,with 10 gametophytesper sporophytecultured and pooled, and a single referencegametophyte cultured. DNA extractionand AFLP analysiswas conductedas above. 74

5.4 Results

5.5 Trial using known progeny

Loci with segregatingalleles were identified in both replicatesof pooled gametophytesfrom the mappingpopulation. From a total of 148 scorableloci, 6 polymorphic loci were identified when the pooled samplewas comparedto the referencesamples, and both replicatesof pooledgametophytes exhibited the same polyinorphisms.A screenshot depicting an exampleof thesepolymorphic loci is given in Figure 5.1, with screenshots depicting all polymorphic loci given in

Appendix IV. The presenceof the alleles in the AFLP profile of either Gransdenor

Villersexel, i. e. the sourceof thesesegregating alleles, is detailedin Table 5.1. 75

ri

C) N C,) (4

C,)

C)

-j

C,) C)

0 C) C

C)

C

C

C) C)

C)) cd) C C) C) krý C) C)) C)

> 76

Table 5.1 Table detailing the polymorphic loci observedwithin a sporophyte producedfrom a crossbetween the Villersexel and Gransdenstrains of P. patens.

Polymorphic loci are namedusing the size of the AFLP fragment.

Polyrnorphic loci Present in Villersexel Present in Gransden (length in bp) 225 Yes No

254 Yes No

264 Yes No

265 Yes No

322 No Yes

339 No Yes

5.6 Analysis of natural populations

Of the 56 sporophytesanalysed, unambiguously polymorphic loci were identified in

3 sporophytes,with 5 polymorphic loci identified in total. The samplederived from one population, Ouse Washes, did not contain a sporophyte with polyinorphic loci.

The numberof sporophytesfound to contain polymorphic loci in eachpopulation, along with the averagedensity of P. patens surroundingeach sporophyte, is shown in Table 5.2. The details of eachsporophyte analysed, including the number of polymorphic loci and the densityof P. patens plants aroundeach sporophyte selectedin this analysis,are detailedin Appendix V. Screenshotsshowing each polymorphic locus are given in Appendix IV. As the numberof sporophytes 77 containingpolymorphic loci was so low, any effect of density could not be measured. 78

C. -a t= to0 0QC., E E! -= C) W) --4 Mt C14 ef) W) 0Uý 140 kn ir- r- r- W') en C-4 C; c; C; C;

10 9) JE -0 = 10r_ C) r. to 0 2 (. ) ri '= El 00r4 %1,) 10t- ý:cnr r-t- tn > en V-4 -4

10 ý ýcE =0 0 z

rA 0. ;5 0

0 t. 0 0 0 0 0 Q.. ýa E C:) o C-4

Iti ý ei R- 7 fi Q W . 0

=$ ; il_ 0 0 0 0 E 00 cl rAwl u ý: -Z 0

C) 2 . 0 -Ig 0. o to g Cd 0 0 J.- wW t= 0 t) cf) 0 V %a..Eb - Q (A r. 0 ,a J (A "m '. " ", cd ýE0 0 ýa 0.C, 9L4 V'J' R2 in. -u u -ý4u JD :E C,:ý o 0 ý4 -i cn 79

5.7 Discussion

The resultsof the trial using a known crossbetween two strainsof P. patens indicate that this methodologycan be usedto identify sporophytesthat are the product of an out-crossingevent. The methodologyhas also been found to be repeatable,as identical polymorphismswere found betweenboth the replicatesanalysed and the referencesamples. When natural populationswere analysed,a low level of variation within the sporophytessampled was observed,with only three sporophytesfound to contain polymorphic loci. However, this doesindicate that out-crossingdoes occur in natural populationsof P. patens, and this suggeststhat mixed-matingsystems operatewithin them.

The samplecollected from the OuseWashes population containedsporophytes with no polymorphic loci. A low level of sporophytescontaining polymorphic loci would be expectedin populationsof P. patens, as self-fertilisation is known to occur with high frequencyunder laboratoryconditions. In selfing bryophytes,not only can true inbreedingoccur, but a crossbetween siblings from the samesporophyte

(intergametophyticself ing) would likely be frequent.This has the potential to greatly reduceheterozygosity within a sporophyte,even if no true inbreedingoccurs

(Hedrick, 1987a;Hedrick, 1987b).

A homozygous,inbreeding population, such as is probablein OuseWashes, maybe stablein bryophyte speciessuch as P. patens.Taylor et aL (2007) studiedthe effects of inbreedingon fitriess using the dioicous mossCeratodonpurpureus and the monoicousmoss F. hygrometrica.By conductingexperimental crosses of thesetwo species,the authorsfound that inbreedingdid have an effect on the fitness of C purpureus but not F. hygrometrica.In somespecies, inbreeding may evenbe 80 advantageous,by purging the populationof deleteriousrecessive alleles (Lande &

Schcmske,1985). F. hygrometrica,like P. patens,is a weedy bryophyte in the

Funariaceae,which colonisesdisturbed environments (Taylor et al., 2007). Due to the natureof the habitatsthat both thesespecies occupy, self ing may be favoured so as to allow re-colonisation.Intragametophytic selfing would then purge the population of deleteriousalleles in one generation(Hedrick, 1987a).Ephemeral populationswould be expectedto experiencegenetic bottlenecks, which would also reducedeleterious alleles in the population (Taylor et aL, 2007; Kirkpatrick & Jame

2000).

Although no relationshipbetween density and the mating systemin populationsof

P. patens could be ascertainedin this study, future studiesinto the breedingsystems of bryophytesand plants in generalshould aim to investigatethe impact of both environmentaland geneticinfluences. Lande & Schemske(1985), characterised plants speciesas either predominantlyselfing or predominantlyout-crossing.

However,plant speciesshow greatvariability in out-crossingrates both within and betweenpopulations (reviewed in Walter, 1986),so speciescannot easily be classed as either selfing or out-crossing.Many different factors,including genetic factors suchas flower size and colour, and ecologicaland environmentalfactors such as density, mode of pollination and altitude, havebeen shown to have a significant impact on the mating systemswithin populations(reviewed in Barrett & Eckert,

1990).Therefore care should be takenover extrapolatingfindings from single population studies.

Wolff el aL (1988), found that the timing of analysishad a significant impact on the out-crossingrates observed in a population of Plantago coronopus.The authors sampledover a period of 6 weeks,and they found a significant decreasein out- 81 crossingrates in the plants sampledlater in the study,mainly due to a decreasein flowering, which is analogousto a decreasein density.Therefore in bryophytes, particular note shouldbe taken of the growth stageof the plants in the population, and in particular the presenceof matureantheridia and archegonia.In practice this could be very time consuming,particularly when taking into accountthe number of populationsand repeatedsampling that would be requiredto investigatethe effect of this on out-crossingrates. In a study suchas this, a more readily recordablevariable maybethe numberof maturesporophytes in the focal area.Studies investigating how ecologicaland environmentalfactors impact on mating systemsin natural populationshave the potential to be extremelycomplex (Barrett & Eckert, 1990).

Therefore,P. patens, as a well establishedexperimental system, has the potential to assistwith advancementin this areathrough laboratorybased studies and controlled field experiments.

The low geneticdiversity levels within populationsof P. patens may poseproblems for fin-therstudies into the mating systemsof this species,as the number of polymorphic loci that can be observedis low. As AFLPs provide a large number of loci for scrutiny, thesemarkers are most appropriatefor suchan investigation.The methodologyused in this study could be most applicablein an initial screento identify populationsof interest,as pooling a number of gametophyticprogeny from a sporophytebefore conductingAFLP analysisis much quicker and cheaperthan genotypingthe progeny individually. As the number of populationswith either no within-sporophytePOlymorphic loci or low levels is likely to be high, this method could be usedto identify only the populationswhere mixed mating can be observed, which maybeof evolutionary interest.Full genotypingcould then be conductedon individual progeny from sporophytesfrom thesepopulations. 82

Chapter 6 General Discussion

Model speciesprovide ideal systemsin which to explore evolutionary questions,as there are often numerousmolecular resources available for the study of such species, along with establishedlaboratories and culturetechniques. General evolutionary questionscan thereforebe answeredwith greaterefficiency in model speciesthan in other organisms.It is howeverimportant for a wide variety of model speciesto be available for such studies,with the diversity of natural systemsfully represented,in order to extrapolateany findings (Travis, 2006). The startingpoint in any evolutionary geneticstudy using any systemmust be to investigatethe level of geneticvariation within that system,along with an estimationof how this is distributed.The level of diversity found within populationsof a speciesreflects its past ability to respondto changingenvironments and evolutionarypressures, and likewise it will also determinethe ability of that speciesto respondto thesechanges in the future. An estimationof how geneflow operateswithin natural populations, and thereforehow this diversity is structuredspatially, will determinethe way in which populationswill respond(Jorgensen & Mauricio, 2004). Therefore,if aspects of a species'biology that are of evolutionary interestare to be studied,the genetic processesthat affect theseaspects must be determined. 83

6.1 Genetic diversity and spatial genetic structure

I have investigatedthe levels of diversity and spatialgenetic structurein natural populationsof P. patens in the UK. There wasno correlationbetween the level of diversity in the populationand size of the areathe population occupies,which contradictspopulation genetic theory. However,these theories have largely been developedfor organismswith very different life historiesto P. patens. As P. patens occupiesephemeral habitats, where the size of the population areais likely to alter from one year to the next, the size of the populationsat the time they were sampled may not be representative of historical populations sizes. Therefore, an extension of this study could be to measure both population area and genetic diversity levels within these populations in ftiture years, as levels of diversity have been shown to correlate with the harmonic mean of population size (Barrett & Husband, 1997).

No spatial geneticstructure was found within populationsof P. patens, whereas theory suggeststhat populationsof bryophytessuch as P. patens should exhibit a degreeof spatial structuredue to an expecteddegree of inbreedingand restricted gametedispersal in bryophytes(Anderson & Lemmon, 1974; Longton, 1976).The lack of spatial structurefound could be due to limitations of the AFLP technique, where the error rate observedcould restrict the resolution of the genetic differentiation that this marker can detect.However, a similar lack of spatial structurehas been found in other bryophyte species(e. g. Snall et al., 2004) and this finding could be relatedto the ecology of P. patens. If geneflow in populationsof

P. patens is primarily mediatedby spores,then periodically rising water levels in the habitatsthat P. patens occupiescould dispersespores over a great distance,leading to a lack of spatial structure.This may also accountfor the finding that genetic diversity is partitioned within ratherthan betweenpopulations, as if P. patens 84 populationsare connectedto a degreeby waterways,gene flow betweenpopulations could be common.As very little is known aboutthe dispersalof sporesin bryophytes,or the regenerationof bryophytepopulations from sporesversus gametophytictissue, then this aspectof the ecologyof P. patens could provide interestingavenues for research(Innes, 1990;Thingsgaard, 200 1).

One limitation of this study was the samplesizes available for within-population spatialautocorrelation analysis. When all the samplesanalysed were pooled, this provided a large data seton which to conductspatial autocorrelation analysis.

However,to derive conclusionsabout the populationstructure in individual populationsis lessrobust than that with the pooledsample, given the samplesizes used.After conductingthis study I would ideally havere-sampled from a single population in 2007, to enablea within-population study with a larger samplesize.

However, extremelyhigh rainfall in summer2007 madethis impossibleas water levels in the populationsdid not recede.

6.2 Mating systems in A patens

The mating systemsoperating within threepopulations of P. patens havebeen investigated,and a novel procedurefor estimationof this proposed.This procedure was found to be proficient in identifying progenythat are the result of an out- crossingevent, when testedusing progeny from a known cross.Haploid species suchas bryophytesprovide convenientmodels in which to investigateplant mating systems,as the haploid progeny from diploid sporophytescan be screenedfor segregatingalleles, thus identifying sporophytesthat are the result of an out-crossing event.Although the progeny from only threeP. patens sporophyteswere found to contain segregatingalleles, this indicatesthat there is at leasta degreeof out- 85 crossingin populationsof P. patens. It is likely that more than three sporophytes analysedin this chapterwere the result of out-crossingevents, as only those crosses betweengenetically different individuals as determinedby the AFLP loci can be detected.

As little is known of the mating systemsin bryophytes,P. patens can be usedas a model to investigatethe levels of inbreedingand out-crossingin bryophytesgiven differing environmentalconditions or populationcharacteristics, as the findings of this study suggestthat mixed-matingsystems operate in populationsof P. patens.

The proceduretrialled provides a relatively inexpensiveand quick method of

conductinginitial analysisand identifying populationsof interest,before proceeding

with more in depth analysisinvolving screeningindividual progeny.As P. patens

can easily be cultured in the laboratory,artificial populationsof P. patens could be

set up and mating systeminvestigations conducted. For example,gamete dispersal

distancecould be estimatedby settingup populationsof P. patens with plants of

known genotypeat known spatial positions,and then screeningthe progeny of any

sporophytesproduced for segregatingalleles at target loci. 86

63 A readed in Europe

Possiblythe most unexpectedfinding in this projectwas the discovery of a moss new to the bryophyte flora of Europe,with P. readeri found in a reservoir in West

Yorkshire. The subsequentmolecular analysis of P. readeri and P. patens indicates that thesespecies should possibly be placedin separategenera, but further analysis using a representativeset of speciesfrom the Funariaceaeis requiredto confirm this.

The finding of P. readeri in West Yorkshire highlights how little is known about the distribution of many bryophyteswithin the UK. Speciesthat occupy habitatssuch as

P. readeri and P. patensare particularly vulnerableto being missedin surveysand thus under-recorded.Whilst looking for populationsof P. patens in the UK, I have found that populationscannot always be found in the samelocality that they have been found previously, primarily due to the water levels in the habitatswhere they are found. It is likely that this is partly the reasonwhy little is known of the ecology and evolution of the plantsthat occupysuch habitats, and why many are classedas rare. Thus P. patens offers a potential model for investigationsinto the ecology and evolution of other ephemeralbryophytes. 87

6.4 Combining molecular geneticsand ecology

This study hasbeen the first investigationinto the population geneticsof P. patens.

AFLP analysiswas usedsuccessfully, with allele frequenciesidentified for a number of polymorphic loci. The AFLP primer combinationsthat were used in this study were also usedto constructthe genetic linkagemap of P. patens (Kamisugi et aL,

2008). Further work could involve scrutiny of thoseloci that were found to be polymorphic in natural populations,which also correspondto loci with identical characteristics(i. e. primer combinationused and fragmentlength) that are polymorphic loci in the mappingpopulation. The polymorphic AFLP loci that were usedto producethe linkage map correspondto DNA fragmentsof known length, but not known DNA sequence.However, the DNA sequenceof eachlocus usedhas beenpredicted by analysisof the sequencedP. patens genomeby Kamisugi et aL(2008) using the SCALPHuntersoftware. Therefore, the DNA sequence correspondingto the AFLP loci usedin this study can also be predictedwith possibleinference of genefunction basedon sequencehomology to known genes.

An exampleof suchan exerciseis describedbelow.

For all polymorphic loci that were characterisedin Chapter3 of this project, candidateswere searchedfor in the P. patens linkage map. From thoseloci identified as having identical characteristicsto a locus on the linkage map, the most

frequentlypolymorphic locus, with an allele frequencyamong the pooled samplesof

0.33, was locus EAGMCT229 (EcoRI+AG, Msel+CT, 229 basepairs in length). An elcctropherogramdepicting this polymorphic locus is given in Appendix VII.

SCALPHunteridentified this locus as a candidatesequence on sequenceScaffold 38

in linkage group I of the P. patens linkage map. Using this predictedsequence, a

BLAST searchof the P. patens genomewas conducted.From this searcha predicted 88 genemodel was identified using the Physcomitrellagenome browser

(http://genomejgi-psf. org/Phypal_I/Phypal_l. home.html). Inspectionof this gene model suggestedthat the computationallydefined model was incorrect, in that the existing EST clonesshowed additional exons.Sequence alignment of the predicted genemodel and the possiblehomologous ESTs was conducted,with a new gene model predictedfrom cDNAs in the P. patens EST database.This genemodel was found to contain a putative protein coding sequence.This polypeptideis a member of a small (2 genes)multigene family in P. patens.This family of proteins is highly conservedin evolution, and encodesa "lipid raft" protein involved in membrane structureand modification.

To identify genesthat maybeof evolutionary significancein P. patens, further samplingof a wide variety of populationsfrom differing environmentsand habitat types is required.If this is achieved,possible ftiture work that combineswhat is known of the molecular geneticsof P. patens with its ecology could be possible.If a large numberof plants are sampledwithin many populations,analyses with molecularmarkers such as AFLPs can be usedto more accuratelyidentify interestingloci (Luikart et al., 2003). With a large number of both samplesand loci,

"outlier loci" can be identified. Outlier loci are loci that exhibit behaviourthat deviatesfrom behaviourpredicted in simulations(Luikart et aL, 2003). For example, theseoutlier loci could be identified basedon simulationsof expectedFst values

(e.g. in Wilding et aL, 2001). In an inbreedingspecies such as P. patens, where the level of diversity, and thus the numberof polymorphic loci, within a population is low, a greatmany loci would be requiredin order to identify outlier loci. Such a study would best be conductedon a populationpreviously identified as containing a high level of geneticvariation. Basedon current knowledgeof diversity within 89 populationsof P. patens,the Villersexel populationin Francemay be a suitable candidatefor sucha study.At presentonly two accessionshave been collected from this location, howeverthese accessions have been shown to be highly divergent basedon AFLP and ITS sequenceanalysis (see Chapter 4 and Kamisugi et aL,

2008). Further analysisinto the geneticsof the population at Villersexel may thereforeprovide a diversity of evolutionarily interestinggenes. As two ITS sequencetraces were found in the K3 strain of Villersexel, suggestinghybridisation hastaken place,this could potentially provide a meansof studying hybridisation and introgressionin P. patenspopulations.

P. patens hasthe potential to be utilised as a model for both molecular geneticand ecologicaland evolutionary studies,as is the casewith other model organismssuch asA. thaliana and D. melanogaster.The study of natural variation within P. patens combinedwith moleculargenetic investigations may provide an insight into the early evolution of land plants. However,P. patens is not only useful as a model bryophyte,as consideringthe ephemeralhabitats that it occupies,it also representsa group of ecologically interestingplants. Little is known of the population geneticsof both ephemeralplants and bryophytes,yet only by studying the full breadthof plant diversity can we fully describethe evolutionaryprocesses that occur within natural populations. 90

Literature cited

Addinsoft (2008) XLSTAT Pro version 2008.7.03available at

http://www. xlstat.com.

Alonso-Blanco C, Koomneef M (2000) Naturally occuring variation in Arabidopsis:

An underexploitedresource for plant genetics.Trends in Plant Science5,22-

29.

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local

alignmentsearch tool. Journal ofMolecular Biology 215,403-4 10.

AndersonLE, Lemmon BE (1974) Geneflow distancesin the moss Weissia

controversaHedw. Journal of the Hattori Botanical Laboratory 38,67-90.

Barrett SCH, Eckert CG (1990) Variation and evolution of mating systemsin seed

plants. In: Biological Approaches and Evolutionary Trends in Plants (ed.

Kawano S), pp. 229-254.Academic Press,London.

Barrett SCH, HusbandBC (1997) Ecology and geneticsof ephemeralplant

populations:Eichhorniapaniculata (Pontederiaceac)in northeastBrazil.

Joumal ofHeredity 88,277-284.

Barrett SCH, ShoreJS (1989) Isozymevariation in colonizing plants. In: Isozymesin

plant biology (eds.SoItIs DE, SoItIs PS),pp. 106-126.Dloscorldes Press,

Portland,Oregon.

BensonDA, Karsch-Mizrachi1, Lipman DJ, Ostell J, Wbeeler DL (2006) GenBank.

Nucleic Acids Research3 4, DI 6-D20.

BerresME (2001) Generalfluorescent AFLP (fAFLP) protocols.University of

Wisconsin Departmentof Zoology, Madison, Wisconsin.Available at

http://ravel. zoology. wisc. edu/sgaap/AFLP-html/fAFLP_protocols. htm.

Accessed02/03/2006. 91

BlearsMJ, De GrandisSA, Lee H, TrevorsJT (1998)Amplified fragment length

polymorphism(AFLP): A review of the procedureand its applications.

Journal ofIndustrial Microbiology & Biotechnology21,99-114.

Bonin A, Bellemain E, EidesenPB, et aL (2004)How to track and assess

genotypingerrors in population geneticsstudies. Molecular Ecology 13,

3261-3273.

Bonnin 1, Ronfort J, Wozniak F, Olivieri 1 (2001) Spatialeffects and rare

outcrossingevents in Medicago,truncatula (Fabaceae).Molecular Ecology

10,1371-1383.

Bouck A, Vision T (2007) The molecular ecologist's guide to expressed sequence

tags.Molecular Ecology 16,907-924.

Bruch P, Schimper WP, GUmbel T (183 6-185 5) Bryologia Europaea seu genera

Aluscorum Europaeorum monographice illustrata E. Scweizerbart, Stuttgart.

Bryan VS (1957) Cytotaxonornicstudies in the Epherneraceaeand Funariaceae.The

Bryologist 60,103-126.

CharlesworthD, CharlesworthB (1987) Inbreedingdepression and its evolutionary

consequences.Annual ReviewofEcology and Systematics18,237-268.

Courtice GRM, Ashton NW, Cove DJ (1978) Evidencefor the restrictedpassage of

metabolitesinto the sporophyteof the mossPhyscomitrella patens (Hedw.)

Br. Eur. Joumal 10,191-198. . ofBryology

Cove D (1983) Geneticsof bryophyta.In: New Manual ofBryology (ed. Schuster

RM), pp. 222-231. Hattori BotanicalLaboratory, Nichinan.

Cove D (2005) The mossPhyscomitrella patens. Annual Reviewof Genetics39,

339-358.

Cove DJ, Knight CD (1993) The mossPhyscomilrella patens, a model systemwith

potential for the study of plant reproduction.The Plant Cell 5,1483-1488. 92

Cove DJ, Knight CD, LamparterT (1997) Mossesas model systems.Trends in

Plant Science2,99-105.

Cox CJ, Goffinet B, Newton AE, ShawAJ, HeddersonTAJ (2000) Phylogenetic

relationshipsamong the diplolepideous-alternatemosses (Bryidae) inferred

from nuclearand chloroplast DNA sequences.The Bryologist 103,224-241.

Cronberg N (2002) Colonization dynamics of the clonal moss Hylocomium

splendenson islandsin a Baltic land uplift area:Reproduction, genet

distribution and geneticvariation. Journal ofEcology 90,925 -935.

CronbergN, Molau U, SonessonM (1997) Geneticvariation in the clonal bryophyte

Hylocomium splendens at hierarchical geographical scales in Scandinavia.

Heredity 78,293-30 1.

CronbergN, RydgrenK, Okland RH (2006) Clonal structureand genet-levelsex

ratios suggest different roles of vegetative and sexual reproduction in the

clonal mossHylocomium splendens. Ecography 29,95-103.

CruzanMB (1998) Geneticmarkers in plant evolutionary ecology.Ecology 79,400-

412.

Cummins H, Wyatt R (1981) Geneticvariability in natural populationsof the moss

Atrichum angustatum.The Bryologist 84,30-38.

Dixon HN (1926) Funariaceae.In: Studiesin the bryology qfNew Zealand,with

special referenceto the herbarium ofRobert Brown. New ZealandInstitute.

Bulletin No.3, Part IV, pp. 191-192.New ZealandInstitute, Wellington.

Duckett JG, Duckett AR (1980) Reproductivebiology and population dynamicsof

wild gametophytesof Equisetum.Botanical Journal of the Linnean Society

80,140.

During HJ, van Tooren BF (1987) Recentdevelopments in bryophyte population

ecology. Trendsin Ecology and Evolution 2,89-93. 93

Ellstrand NC, Torres AM, Levin DA (1978) Density and the rate of apparent

outcrossingin Helianthus annuus(Asteraceae). Systematic Botany 3,403-

407.

Engel PP (1968) The induction of biochemicaland morphologicalmutants in the

mossPhyscomitrella patens. AmericanJournal ofBotany 55,438-446.

Ennos RA (1990) Population genetics of bryophytes. Trends in Ecology & Evolution

5,38-39.

EnnosRA (2001) Inferencesabout spatialprocesses in plant populationsfrom the

analysisof molecularmarkers. In: Integrating Ecology and Evolution in a

Spatial Context (eds. Silvcrtown J, Antonovics J), pp. 45-71. BIackwcII

Science,Oxford.

Ennos RA, Swales KW (1987) Estimation of the mating system in a fungal pathogen

Crumenulopsis sororia (Karst. ) Groves using isozyme markers. Heredity 59,

423430.

Eppley SM, Taylor PJ,Jesson LK (2007) Self-fertilization in mosses:A comparison

of heterozygotedeficiency betweenspecies with combinedversus separate

sexes.Heredity 98,3 8-44.

Escudero,A, Iriondo JM, Torres ME (2003) Spatial analysisof geneticdiversity as a

tool for plant conservation.Biological Conservation113,3 51-365.

Excoffier L, SmousePE, Quattro JM (1992) Analysis of molecularvariance inferred

from metric distancesamong DNA haplotypes:Application to human

mitochondrial DNA restriction data. Genetics131,479-491.

FederME, Mitchell-Olds T (2003) Evolutionary and ecological functional

genomics.Nature ReviewsGenetics 4,649-655.

Fife AJ (1985) A genericrevision of the Funariaceae(Bryophyta: Musci) Part I.

Journal ofthe Hattori Botanical Laboratory 58,149-196. 94

Frank W, RatnadewiD, Reski R (2005) Physcomitrellapatens is highly tolerant

againstdrought, salt and osmotic stress.Planta 220,384-394.

Free SJ, Holtz BA, Michailides TJ (1996) Mating behaviourin field populationsof

Moniliniafructicola. Mycologia 88,208-211.

Goffinet B (2007) Physcomitrella.In: Flora offorth America (ed. Commitee

FoNAE), pp. 194-196.Oxford University Press,New York and Oxford.

Goloboff PA, Farris JS, Nixon KC (2008) TNT, a free program for phylogenetic

analysis.Cladistics 24,1-13.

Goodwillie C, Kalisz S, Eckert CG (2005) The evolutionary enigmaof mixed

mating systems in plants: Occurrence, theoretical explanations and empirical

evidence.Annual ReviewofEcology, Evolution and Systematics36,47-79.

Grimsley NH, Ashton NW, Cove DJ (1977a) Complementation analysis of

auxotrophic mutants of the moss, Physcomitrella patens, using protoplast

fusion. Molecular and General Genetics155,103-107.

Grimsley NH, Ashton NW, Cove DJ (I 977b) The production of somatichybrids by

protoplastftision in the moss,Physcomitrella patens. Molecular and General

Genetics154,97-100.

GrundmannM, Ansell SW, RussellSJ, Koch MA, Vogel JC (2007) Genetic

structureof the widespreadand commonMediterranean bryophyte

Pleurochaete squarrosa (Brid. ) Lindb. (Pottiaceae) - evidence from nuclear

and plastidic DNA sequencevariation and allozymes.Molecular Ecology 16,

709-722.

GrundmannM, Ansell SW, RussellSJ, Koch MA, Vogel JC (2008) Hotspotsof

diversity in a clonal world - the Mediterraneanmoss Pleurochaete squarrosa in Central Europe.Molecular Ecology 17,825-838. 95

GunnarssonU, HasselK, S6dcrstr6mL (2005) Geneticstructure of the endangered

peatmoss Sphagnum angermanicum in Sweden:A result of historic or

contempomry proccsses? The Bryologist 108,194-203.

Hamrick X, Godt MJW (1990) Allozyme diversity in plant species.In: Plant

population genetics,breeding andgenetic resources(eds. Brown AHD,

Clegg MT, Kahler AL, Weir BS), pp. 43-63. SinauerAssociates, Sunderland,

M. A.

Hardy OJ, VekernansX (1999) Isolation by distancein a continuouspopulation:

reconciliationbetween spatial autocorrelation.and population genetics

models. Heredity 83,145-154.

Hardy OJ, VekemansX (2002) SPAGeDi: a versatilecomputer program to analyse

spatial geneticstructure at the individual or population levels. Molecular

Ecology Notes 2,618-620.

HasegawaM, Kishino H, Yano K (1985) Dating of the human-apesplitting by a

molccular clock of mitochondrial DNA. Journal ofMolecular Evolution 22,

160-174.

HasselK, GunnarssonU (2003) The use of inter simple sequencerepeats (ISSR) in

bryophytepopulation studies.Lindbergia 28,152-157.

Haubold B, HudsonRR (2000) MAN 3.0: detectinglinkage disequilibrium in

multilocus data.Bioinformatics 16,847-848.

HeckmanDS, GeiserDM, Eidell BP, et aL (2001) Molecular evidencefor the early

colonization of land by fungi and plants.Science 293,1129-1133.

Hedrick PW (1987a)Genetic load and the mating systemin homosporousferns.

Evolution 41,1282-1289.

Hcdrick PW (1987b) Populationgcnetics of intragametophyticselfing. Evolution 41,

137-144. 96

Heywood JS (1991) Spatialanalysis of geneticvariation in plant populations.

Annual ReviewofEcology andSystematics22,335-355.

Hill MO, PrcstonCD, Smith ME (1994) Atlas oftheBryophytes ofBritain and

Ireland Harley Books, Colchester.

Hohe A, RensingSA, Mildner M, Lang D, ReskiR (2002) Day length and

temperaturestrongly influence sexualreproduction and expressionof a novel

MADS-box gene in the moss Physcomitrella patens. Plant Biology 4,595-

602.

Holyoak DT, Bryan V (2005) Ephemerumhibernicum sp. nov. (Bryopsida:

Ephemeraceae)from Ireland. Joumal ofBryology 27,89-95.

Innes DJ (1990) Nficrogeographicgenetic variation in the haploid and diploid stages

of the mossPolyfrichumjuniperinum Hedw. Heredity 64,331-340.

Jobb G (2008) TREEFINDER version of March 2008. Munich, Germany.

Distributed by the author at www. treefinder.de.

JohansonU, West J, Lister C, et aL (2000) Molecular analysisof FRIGIDA, a major

determinantof naturalvariation in Arabidopsis flowering time. Science290,

344-347.

JorgensenS, h1auricioR (2004) Neutral geneticvariation amongwild North

American Arabidopsis is populationsof the weedy plant thaliana not . geographicallystructured. Molecular Ecology 13,3403-3413.

Kamisugi Y, von StackelbergM, Lang D, et aL (2008) A sequence-anchoredgenetic

linkage map for the mossPhyscomitrella patens. ThePlant Journal 56,85 5-

866.

Kcnrick P, CranePR (1997) The origin and early evolution of plants on land. Nature

389,33-39. 97

Kirkpatrick M, JameP (2000) The effects of a bottleneckon inbreedingdepression

and the geneticload. TheAmerican Naturalist 155,154-167.

Klckowski EJ (1972) Geneticalfeatures of fernsas contrastedto seedplants. Annals

ofthe Missouri Botanical Garden59,13 8-151.

Knight CD, Cove DJ, CurningAC, Quatrano,RS (2002) Moss genetechnology. In:

Molecular Plant Biology (eds. Gilmartin PM, Bowler C), pp. 285-301.

Oxford University Press,Oxford.

Lande F, SchemskeDW (1985) The evolution of self-fertilization and inbreeding

depressionin plants 1.Genetic models. Evolution 39,2440.

Landry CR, Townsend JP, Hartl DL, Cavalieri D (2006) Ecological and

evolutionary genomics of Saccharomyces cerevisiae. Molecular Ecology 15,

575-591.

Li CC, Weeks DE, Chakaravarti A (1993) Similarity of DNA finger-prints due to

chanceand relatedness.Human Heredity 43,45-52. 6fvers. Lindberg SO (1864) UppstAllningaf familjen Funariaceae. AfK. Vet.-Akad. Foerh. 21,5 89-600.

Loiselle BA, Sork VL, NasonJ, GrahamC (1995) Spatial geneticstructure of a

tropical understoryshnib, Psychotria officinalis (Rubiaceae).American

Journal ofBotany 82,1420-1425.

Longton RE (1976) Reproductivcbiology and cvolutionary potcntial in bryophytes.

Journal ofthe Hattori Botanical Laboratory 41,205-223.

Longton RE, SchusterRM (1983) Reproductivebiology. In: New Manual of

Bqology (cd. SchusterRM), pp. 386-462.Hattori Botanical Laboratory,

Nichinan. 98

Luikart G, EnglandPR, Tallmon D, JordanS, TaberletP (2003) The power and

promiseof populationgenomics: From genotypingto genometyping. Nature

ReviewsGenetics 4,981-994.

Lynch M, Milligan BG (1994) Analysis of populationgenetic structurewith RAPID

markers. Molecular Ecology 3,91-99.

Lynch NI, Ritland K (1999) Estimation of pairwiserelatedness with molecular

markers.Genetics 152,1753-1766.

Manel S, SchwartzMK, Luikart G, TaberletP (2003) Landscapegenetics:

Combining landscapeecology and populationgenetics. Trends in Ecology &

Evolution 18,189-197.

Mantel N (1967) The detectionof diseaseclustering and a generalizedregression

approach. Cancer Research 27,209-220.

Marra RE, Cortesi P, Bissegger M, Milgroorn MG (2004) Mixed mating in natural

populationsof the chestnutblight fungus,Cryphonectriaparasitica.

Heredity 93,189-195.

Milgroom MG, Lipari SE, EnnosRA, Liu Y-C (1993) Estimationof the outcrossing

rate in the chestnutblight ftingus, Cryphonectriaparasitica. Heredity 70,

385-392.

Milgroom MG, Lipari SE, Powell WA (1992) DNA fingerprinting and analysisof

population structure in the chestnut blight fungus, Cryphonectriaparasifica.

Genetics131,297-306.

M ishler D (1988) Reproductiveecology of bryophytes.In: Plant Reproductive

Ecology: Patterns and Strategies(eds. Lovett Doust J, Lovett Doust Q, pp.

285-306.Oxford University Press,Oxford.

Mitchcll-Olds T (2001) Arabidopsis thaliana and its wild relatives:A model system

for ecology and evolution. Trendsin Ecology and Evolution 16,693-700. 99

Miyashita NT, KawabeA, Innan H (1999) DNA variation in the wild plant

Ar,abidopsis thaliana revealedby amplified fragmentlength polymorphism

Genetics 152,1723-173 1.

MousetteS, DcromeN (2004) Molecular polymorphismin Drosophila

melanogasterand D. simulans: What havewe learnedfrom recent studies.

Genetica120,79-86.

Murtagh GJ, Dyer PS, CrittendenPD (2000) Sex and the single lichen. Nature 404,

564.

Myburg AA, RcmingtonDL (2000) Protocol for High-ThroughputAFLP analysis

using LI-COR IR2 Automated Sequencers.Forestry and Agricultural

Biotechnology Institute (FABI), University of Pretoria. Available at

http: //www. up.ac. za/academic/fabi/eucgenomics/euc-mapping/AFLP-Protoc

ol. pdf. Accessed 02/03/2006.

Nei M (1973) Analysis of genediversity in subdividedpopulations. Proceedings of

the National AcademyofSciences 70,3321-3323.

NishiyamaT, Fujita T, Shin T. et aL (2003) Comparativegenomics of

PhYS comilrella patens gametophytictranscriptome and Arabidopsis thaliana:

Implication for land plant evolution. Proceedingsof the National Academyof

Sciences 100,8007-8012.

Ochi H (1968) A revision of the family Funariaceae(Musci) in Japan.Japanese

Journal ofBotany 20,1-34.

Oldcnhof H, NVolkcrsNVF, Bowman JL, Tablin F, Crowe JH (2006) Freezingand

desiccationtolcmnce in the mossPhyscomitrella patens: An in situ Fourier

transform infrared spectroscopicstudy. Biochimica et BiophysicaAcla 1760,

1226-1234. 100

Oliver MJ, Velten J, Wood AJ (2000) Bryophytesas experimentalmodels for the

study of cnvirorunentalstress tolerance: Tortula ruralis and desiccation-

tolerancein mosses.Plant Ecology 151,73-84.

Oliver NIJ, Velten J, Mishler BD (2005) Desiccationtolerance in bryophytes:A

reflection of the primitive strategyfor plant survival in dehydratinghabitats?

Integrative and Comparative Biologv 45,788-799.

PresselS, MatchamHW, Duckett JG (2007) Studiesof protonemalmorphogenesis

in mosses.XI. Bryum and allied genera:a plethoraof propagules.Journal of Biyology29,241-258.

QuclIcr DC, GoodnightKF (1989) Estimatingrelatedness using geneticmarkers. Evolution 43,258-275.

Ramu C, Hideaki S, TadashiY, et aL (2003) Multiple sequencealignment with the

Clustal series of programs. Nucleic Acids Research 31,3497-3500.

RensingSA, Lang D, Zimmer AD, et aL (2008) The Physcomitrellagenome reveals

evolutionary insights into the conquestof land by plants.Science 319,64-49.

RenzagliaKS, SchuetteS, Duff RJ, et aL (2007) Bryophytephylogeny: Advancing

the molecularand morphological frontiers. TheBryologist 110,179-213.

Reski R (1998) Development,genetics and molecularbiology of mosses.Botanica Acta 111,1-15.

Reski R (1999) Molecular geneticsof Physcomitrella.Planta 208,301-309.

Reski R, FaustM, Wang X-H (1994) Genomeanalysis of the mossPhyscomitrella

patens (Hedw.) B.S. G. Molecular and General Genetics244,352-359.

Ritland C, Ritland K (2000) DNA-fragment markersin plants. In: Molecular

Methodsin EcoloU. (ed. Baker AJ), pp. 208-234.Blackwell ScienceLtd.,

Oxford. 101

Ritland K (1990) Inferencesabout inbreedingdepression based on changesof the

inbrcedingcooefficicnt. Evolution 44,1230-1241.

Ritland K (1996) Estimatorsfor pairwise relatednessand individual inbreeding

coefficients. GeneticalResearch 67,175-185.

RoadsE, Longton RE (2003) Reproductivebiology andpopulation studiesin two

annualshuttle mosses.Journal of the Hattori Botanical Laboratory 93,305-

318.

Schabenberger0, Gotway CA (2005) Statistical MethodsfOrSpatial Data Analysis

Chapman & Hall, Boca Raton.

SchaeferDG, Zryd J-P (1997) Eff icient genetargeting in the mossPhyscomilrella

patens. ThePlant Journal 11,1195-1206.

Schaefer DG, Zryd J-P (200 1) The moss Physcomitrella patens, now and then. Plant

Physiology 127,1430-1438.

SchneiderS, RoessliD, Excoffier L (2000) Arlequin: A softwarefor population

geneticsdata analysis.Ver 2.000. Geneticsand Biometry Lab, Dept. of

Anthropology, University of Geneva.

Scott GAM, StoneIG (1976) PhyscomitrellaB. S.G. In: The mossesofsouthern

Australia, pp. 260-262.Academic Press,London, New York, SanFrancsico.

SeymourFA, CrittendenPD, DickensonMJ, et aL (2005) Breedingsystems in the

lichm-forming fungal genusCladonia. Fungal Geneticsand Biology 42,

554-563.

SharbelTF, Haubold B, Mitchell-Olds T (2000) Geneticisolation by distancein

Arabidopsis thaliana: Biogeographyand postglacialcolonisation of europe.

Molecular Ecology 9,2109-2118. 102

ShawAJ (2000) Populationecology, populationgenetics, and microevolution. In:

BryophyteBiology (eds. ShawAJ, Goff met B), pp. 369-402. Cambridge

University Press,Cambridge.

Shaw DV, Kah1crAL, Allard RW (1981) A multilocus estimatorof mating

parametersin plant populations.Proceedings of the National Academyof

SciencesýUSA. 78,1298-1302.

ShawJ, Antonovics J, AndersonLE (1987) Inter- and intraspecificvariation of

mossesin toleranceto copperand and zinc. Evolution 41,1312-1325.

Shimizu KK (2002) Ecology meetsmolecular genetics in Arabidopsis.Population

ecology44,221-233.

Skotnicki hIL, Selkirk PM, Ninham JA (1999) RAPD analysisof geneticvariation

and dispersal of the moss Bryum argenteum in Ross Island and Victoria

Land, Antarctica. Polar Biology 21,417-422.

Slatkin hi (1985) Geneflow in natural populations.Annual ReviewofEcology and

Systematics16,393430.

Smith ME (2004) Themossflora ofBritain and Ireland, secondcdn. Cambridge

University Press,Cambridge.

Sn5ll T, FogelqvistJ, Ribeiro PJJ,Lascoux M (2004) Spatialgenetic structure in

two congenericepiphytes with different dispersalstrategies analysed by

threedifferent methods.Molecular Ecology 13,2109-2119.

StenoienHK, FlatbergKI (2000) Geneticvariability in the rare norwegianpeat moss

Sphagnumfroendelagicum. The Biyologist 103,974-801.

StenoicnHK, SAstadSM (1999) Geneticstructure in threehaploid peat mosses

(Sphagnum).Heredity 82,391-400.

StoneIG, Scott GAM (1973) Name changesin Australian mosses.Joumal of

Bi)-oloo, 7,603-605. 103

Storfer A, Murphy NIA, EvansJS, el aL (2007) Puttingthe 'landscape'in landscape

genctics.Heredity 98,12 8-142.

Tan BC (1979) A new classificationfor the genusPhyscomitrella B. S. G. Journal of

the Hattori Botanical Laboratory 46,327-336.

Taylor PJ, Eppley SINI,Jesson LK (2007) Sporophyticinbreeding depression in

mossesoccurs in a specieswith separatesexes but not in a specieswith

combinedsexes. American Journal ofBotany 94,1853-1859.

Taylor TN, Kcrp H, HassH (2005) Life history biology of early land plants:

decipheringthe gametophytephase. Proceedings of the National Academyof

Sciences 102,5892-5897.

ThingsgaardK (2001) Populationstructure and geneticdiversity of the

amphiatlantic haploid peatmoss Sphagnum affine (Sphagnopsida). Heredity

87,485496.

Tonsor SJ,Alonso-Blanco, C, KoornneefM (2005) Genefunction beyondthe single

trait: natural variation, geneeffects, and evolutionaryecology in Arabldopsis

thaliana. Plant, Cell and Environment 28,2-20.

Travis J (2006) Is it what we know or who we know? Choice of organismand

robustnessof inferencein ecology and evolutionarybiology. TheAmerican

Naturalist 167,303-314.

VckcmansX, BeauwcnsT, LemaireM, Roldan-Ruiz1 (2002) Data from amplified

fragmentlength polymorphism (AFLP) markersshow indication of size

homoplasyand of a relationshipbetween degree of homoplasyand fragment

sizc. Molecular Ecology 11,139-151.

VckcmansX, Hardy OJ (2004) Ncw insights from finc-scalespatial gcnetic

structurcanalysis in plant populations.Molecular Ecology 13,921-935. 104

Vogler DNV,Kalisz S (2001) Sex amongthe flowers: The distribution of plant

mating systems.Evolution 55,202-204. von Stackclbcrghl, RcnsingSA, Reski R (2006)Identification of genic moss SSR

markcrsand a comparativeanalysis of twenty-four algal and plant gene

indicesreveal spccies-specificrather than group-specificcharacteristics of

microsatellites. BAfC Plant Biology 6,9.

Vos P, HogersR, B Icckcr M. et aL (1995) AFLP: A new techniquefor DNA

fingerprinting.Nucleic Acids Research23,4407-4414.

Walter DNI (1986) Is theredisruptive selectionfor sclf-fertilization? TheAmerican

Alaturalist 128,421426.

Wang J (2002) An estimatorfor pairwise relatednessusing molecularmarkers.

Genetics 160,1203-1215.

Waters ER (2003) h1olecularadaptation and the origin of land plants. Molecular

Phylogenelicsand Evolution 29,456463.

Whitlock R, HippersonH, Mannarelli M, Butlin RK, Burke T (2008) An objective,

rapid and reproduciblemethod for sýoringAFLP peak-heightdata that

minimizes genotypingerror. Molecular Ecology Resources,In press.

Wilding CS, Butlin RK, GrahameJ (2001) Differential geneexchange between

parapatricmorphs of Littorina saxatilis detectedusing AFLP markers.

Joumal ofEnvironmental Biology 14,611-619.

Wilson PJ, ProvanJ (2003) Effect of habitat fragmentationon levels and patternsof

geneticdiversity in natural populationsof the peat mossPolytrichum

commune.Proceedings of the Royal Societyof London B. 270,881-886.

Wolff K, Friso B, DammcJMNI (1988) Outcrossingrates and male sterility in

natural populationsof Plantago coronopus.Theoretical andApplied

Genetics76,190-196. 105

Wright S (1943) Isolation by distance.Genetics, 114-13 8.

Wright S (1978) Evolution and the geneticsofpopulations. Vol. 4. Variability

%ithin and amongnatural populations.University of Chicago Press,Chicago.

NVyattR, Odrzykoski U, StoncbumcrA (1989a)High levels of geneticvariability in

the haploid mossPlagiomnium ciliare. Evolution 43,1085-1096.

Wyatt R, Stoncbumer A, Ordrzykoski U (1989b) Bryophyte isozymes: Systematic

and evolutionaryimplications. In: Isozymesin Plant Biology (eds. Soltis DE,

Soltis PS), pp. 221-239.Chapman & Hall, London.

YarnazakiT (1981) Genic variabilities in naturalpopulation of haploid plant,

Conoccphalum conicum. 1. The amount of heterozygosity. Japanese Journal

of Genetics56,373-383.

Zartman CE, McDaniel SF, Shaw AJ (2006) Experimental habitat fragmentation

increaseslinkage discquilibrium but doesnot affect geneticdiversity or

population structurein the Amazonianliverwort Radulaflaccida. Molecular

EcoloDý15,2305-2315.

Zimmer A, Lang D, Richardt S, el aL (2007) Dating the early evolution of plants:

Detectionand molecularclock analysesof orthologs.Molecular Genetics

and Genomics278,393-402. 106

E3

0 0 Z Z Z

2 CD C> CD CD "0 "0 .,7 C> CD CD CD CD CD Z! z2 CD CD C>a; M:alý2 f4 CD CD CD

r- u vi ci 00 rn ß_ CD %0 le kn r- C %0 %0 vl (D C> 00 00 00 tu "0 U2

rn rn 120 C> r-

r_ 0% m r- C) 0 0 r- . le 9 9 IZ)

rs .2 dm

-1 "0 "0 2-

2-

iz c 190 12 F2 *Z rý 0 0 0

«li 107-

8

0%.

ýo

0 rA 0 0 0 u 0 Z r. Z >ý Z

%0 1.0 140 "0 C> C) c> CD CD CD CD CD C> CD CD CD CD Cl! rq V2 rA Z" rq -- I clý all all 0, % CD alý (Z C> CD C> Vlb 00 00 CD CD CD CD (D CD CD

rn cý C>

roi 0 r4 m vlä r- all rn r- (D - - 00 alý clý vi %0 \o CD C> cý 00 rn rn Ilt v»i r- CD r- r- le 't r- vi vi V'b le rn cn cn Gn Ln Ln (A Ln Ln

%0 le 140 00 C) CD 00 00 00 o\ 0 CD

ON

Z 9 M vi r- CD rq

fi 0 C) ei

E E E E 0 0 cn0 vý0 uý0 vý0 > > > P-9

rA

0 0 -äg ei it 0 .9 C) 9) "Ci 0 fi 9.1 108

%0 %0 %0 V% %0 %0 %0 CD C> c: h (D (Z C> (D CD C> C: ) CD CD CD CD CD Z, Z! !ri 1-ý! rq -- P CD 00 CD CD C) r*A Vlb (> 00 00 CN clý

a vi 00 00 alý \o - clý et (D (> r- 00 ýo 00 00 NO lit 10 "0 rq r4 0 m ýo m r- m11 vi %0 vl 0 3 3 u2r_ Z 0 F-b»i E- E- vi 0vý cn (n

r-i vläCD "0 CD (Z V Z! 00 CD C> Irl J. e r4 izz 9 52 Z, rq --"0 C) c= CD

tA u cn Gn

92. iz ci 0 10

1.' -Z 01 0 rA 1.1-i ý5 »E! to ci

r_ cý. Ln > 0 1ýi Wi 1 109

cn fi 0 0 0 0 Z > Z Z Z Z Z

*4o %D %0 1.0 C> C> C> CD CD C> CD C) V CD

f 4 C> C2(D cliC> C).- C> C> C> all -4

C> G Z C> CN r- C> %D %0 ;,vi 4rb V) (, A "0 0 V) t- KZ) le d' CD rn rn "0 CD rn C) CY v

CD (Z) C> C) 00 r- CD C> CY\ Z 0 (> C> (D 00 2ý r (: r4 (14 . "0 rn

ci u c) 0 0 9) ta ta to to to :p -. i -1 10 Ici "0 0 iz 0 E . 0 cu

0

0 (A

ri (13 -c zi 1.4 rýfi 4--

E f! 2 92 r. r3 :e .4

t41 i- , fi ei > 10 cu lu ce Z 4 110

-13 CIO

ci E

V3

0 0 0 z z z

ýo %0 C) C) CD

00 00

vý r- (> r- 30 CD (: V, rq %0 8 m Do 0% 0% (> r- (D "0 vi 0 ;Z (: le 00 00 Zoo CD rn er (: 2 Cl% C> w Z) Gn u2-b cn v2 cn u2 v2 Gn

(> (D CD C> Ch C> Q C> (: C> c2 Z! r_ 0% (4Z! r4 -- CD -- _X

JM : c02

rA %A b. 30 0m 0Z) 9.) C)

x 9x %A CD. rA

lua %. ce x c E -1 0 190 Z j ri r_ G im > E2 (A III

0 0 0 0 Z Z Z Z

%0 %0 %D 1.0 C) (D (D C:) C> CD C> CD i2 r4 ic2

C> 00

vi ir- r, i ch ui r r_ 2 Ln (nzi cn rn

f4

Z Z .Xc Vlb A- cz

401 Je

: r3

Itj

0 e3 iýZ Q

12

Z; JD 0 112

Appendix 11Description of P. patens

(rakcn from Smith, 2004)

Ph)-sconzitreitapatens(Hedw. ) Bruch & Schimp.

Ephemeralgreen patches or scatteredplants, to 2.5mm high. Leaveserect or erect- patentwhen moist, ovatc to lanceolatc,shortly acuminate;margins plane, bluntly dcntatefrom about the middle; costaending in or below apex;cells narrowly rcctangularbelow, shorterabove, 15-20pmwide in mid-leaf, 1-2 marginal rows narrowcr. Sctaevery short, a 100 prn long-,capsules immersed, cleistocarpous, globose,with short blunt beak;spores 26-32ILm. By streams,ditches, ponds, lakes, reservoirs,bottoms of dried-out ponds,wet tracks and woodlandrides. Lowland.

Occasionalin England,rare in Wales,very rare in Scotland,extending north to

Dunbartonand Kint)rc, rare in Ireland. EurosiberianTemperate. Europe, except the

Mediterraneanregion, north to southernScandinavia, Yenesei, N. America. 113

Appendix III Descriptionof P. readeri

(Adapted from Scott & Stone, 1976,and Smith, 2004)

Physcomitrellareaderi (C. Milell. ) Stone& Scott

Ephemeralgreen plants, 0.5 - 4mm high. Leaveserect when moist, oblanceolate, spathulatcor obovatewith a short apiculus,margins plane and entire near the base, serrateabove; costa ending far short of the apex;cells large and thin-walled, rectangularor rhomboidalto six-sided,20-30pm wide and mostly 2-3 XI but sometimesmuch larger in the lower part of the leaf and adjoining the costa.Setae very short, usually half the length of the cleistocarpouscapsule; capsules very variable in size, globular or oblong, apiculate,dark brown to black at maturity.

Stomataat baseof the capsulenot immersed,no line of dehiscenceon the capsule.

Sporesdark rusty brown, 3045pm, spinulosewith slightly hookedspines.

Paroccious,anthcridia in clustersof 4-5 in the leaf axils just below the archegoniaat the apex.

Growing in wide patcheson mud or silt in recently flooded areas.UK population in

West Yorkshire, also Australia, New Zealand,U. S.A. and Japan. 114

7;

!Z

Q. m 04Ln C.. N 115

Appendix V Table dctaling eachsporophyte analysed in Chapter5, including the numberof polymorphic loci and the density of P. patensplants around each sporophytc.

Population/Density Sporophyt Polymorphic Density Density designation e I. D. loci (length in (fraction fill) (fraction fill) bp) 1XIM IOXIOCM Papercourt 24.4.2a 0.653 0.408 marshes/High' 24.4.3a 0.714 0.286 24.4.4a 0.735 0.265 24.5.2a 0.531 0.449 24.5.4b 0.653 0.306 24.6.3a 0.367 0.776 24.7.1a 0.388 0.286 24.8.4a 0.571 0.102 24.10.2a, 0.469 0.163 24.10.3b 0.571 0.245 Papercourt 24.9.2a. 0.163 0.265 marshes/Lo%v2 24.1.2b 0.143 0.204 24.1.4a, 0.184 0.163 24.11.2b 0.163 0.163 24.12.1a 0.122 0.102 24.12.3a 0.143 0.122 24.2.1a 0.122 0.245 24.2.5a 322 0.184 0.082 24.3.1a 0.184 0.082 24.3.2a 0.143 0.082 116

Ouse 25.1.3a 0.204 0.143 washes/High' 25.12.3a 0.245 0.245 25.2.1b 0.163 0.531 25.2.2a 0.184 0.612 25.2.3b 0.143 0.449 25-3.1a 0.122 0.245 25.5.1b 0.163 0.510 25.6.1b 0.184 0.306 25.6.2a 0.184 0.327 Ousewashes/Low2 25.8.1a 0.020 0.102 25.9.1a 0.020 0.286 25.11.2 0.041 0.122 25.11.3a 0.041 0.122 25.12.1b 0.041 0.286 25.13.2b 0.020 0.102 25.14.2a 0.041 0.204 25.14.3a 0.041 0.163 25.7.1a 0.041 0.163 Llanfihangel 26.3.2a 0.204 0.306 Gobion/ High' 26.4.1a 0.204 0.204 26.4.2a 0.245 0.367 26.4.3a 0.408 0.265 26.5.1a 0.102 0.327 26.6.3 0.122 0.143 26.8.1a 0.122 0.143 26.8.3a 0.122 0.245 26.10.1b 0.122 0.367 26.13.1a 152,260,279 0.122 0.367 117

Llanfdlangel 26.12.1a 0.041 0.163 GobionILo,. v2 26.12.2b 0.041 0.122 26.12.3b 0.061 0.082 26.13.2a 0.082 0.184 26.14.1 0.061 0.204 26.14.2a 0.041 0.122 26.14.3a 0.061 0.286 26.15.1c 0.040 0.122 26.15.3b 199 0.020 0.204 26.9.3b 0.041 0.061 I Sporophytes within a high density area 2 Sporophytcs within a low density area 118

I, -0 t;jý_- E c2. pN CA ;, CD cu

C'4

C.,

CO Cd C%l Ln zL, c2-. I

I-

k

I.

I,

LM

!

r- .Z (a fill.]

fl

C4 119

Appendix VII Screenshot depicting polymorphic locus EAGMCT229.