节点文献

薜荔之传粉小蜂

The Pollinating Fig Wasps of Ficus Pumila

【作者】 陈艳

【导师】 陈小勇;

【作者基本信息】 华东师范大学 , 生态学, 2009, 博士

【副题名】我国东南沿海薜荔小蜂隐存种及分布格局

【摘要】 共生是最具影响力的种间相互作用之一,对生命各个层次都有深刻的且常常是决定性的影响;共生体系的长期维持引发了一系列的进化问题,也因此成为进化理论的一大挑战。榕和榕小蜂系统是目前所知的最严格的专性共生体系之一,两者异常丰富的物种多样性和宿主生活型为共生机理研究提供了理想的体系。榕-榕小蜂的共生关系古老且多样,早期的形态学研究显示1种榕往往只有1种对应的传粉小蜂,分子系统学研究也证实在较高的分类等级上两者间具有高度的协同枝系发生和协同进化关系,这种协同关系被应用到物种水平,“一对一”原则被认为在该系统中具有普适性。然而,近年来越来越多的案例证实许多榕属物种拥有1种以上的传粉小蜂,也有同一传粉小蜂为不同榕授粉的实例。特别是分子标记的应用,揭示出大量传粉小蜂隐存种,比例甚至达到所涉及榕属物种的60%;加之宿主转移现象和榕天然杂交后代的发现,冲击着榕-榕小蜂在物种水平高度对应的观点,Machado等也因此提出一个更为松散的进化模式,认为榕与榕小蜂的协同进化仅发生在两者群组之间,在榕群组内宿主转移现象可能经常发生。种间专一性屏障的松弛将极大地影响共生双方的动态,哪怕是低频率的例外事件也会导致严重的进化后果。Machado等的进化新模式主要依据巴拿马和墨西哥2个榕组及其对应的传粉小蜂,但有研究证实在某些地区和某些榕属类群中,榕-榕小蜂在物种水平高度对应,协同成种在其进化中占主导地位。因此要正确评价种间专一性在榕-榕小蜂共生体系中的真正地位,不同榕属类群和不同地区非一对一事件的发生频度及产生原因是不可或缺的基本信息。薜荔小蜂为薜荔及其变种爱玉子的专性传粉小蜂,至今形态学研究仍未发现它们在任何地区有差异,但薜荔和薜荔小蜂的分类地位飘浮不定,不同的研究者将它们归入不同的属,且在爱玉子人工栽培中薜荔小蜂能否为之有效传粉也有不同的结果。为探究薜荔传粉小蜂的物种组成、地域分布和系统发生关系,我们应用线粒体COI和Cytb、核核糖体ITS2基因对我国东南沿海岛屿及邻近大陆31个种群,331个个体进行了遗传组分、遗传多样性和种群扩张动态分析,并在GenBank中检索相关序列联合构建系统树,以期揭示薜荔传粉小蜂在该地区的物种组成、形成原因和可能的扩散路线;评价薜荔与其传粉小蜂间的物种专一性;评判岛屿和大陆生境对小蜂种群动态的影响;为印度-澳洲区系榕-榕小蜂专性对应程度提供案例;为协同进化模式研究提供参考。论文主要结论如下:1)薜荔传粉小蜂在我国东南沿海地区的遗传组成并不均一,可以划分为3个独立的枝系。3个枝系无论是线粒体基因还是核基因都超出了种内差异的范畴,达到物种分异水平;3个枝系没有明显的形态差异,为3个隐存种,分别为Sp.A、Sp.B、Sp.C。不同基因和不同方法均将31个种群的331个个体区分为3个枝系,枝系间的遗传分异明确。STUCTRUE软件模拟100万次将个体分配到各枝系的概率都在95%以上;系统树中3个单系的支持率都达99%;在巢式枝系结构中,3个枝系在95%的置信区间内完全分离。不同基因和不同方法区分得到的3个枝系在单倍型和个体水平都完全对应。比较3个枝系COI序列分化显示,平均枝系内K2P距离为0.57%,而枝系间分别达6.63%,10.94%和10.69%,任2个枝系间的序列分异都达到平均枝系内遗传差异的10倍以上。这不仅超过了区分动物物种的COI barcode标准距离域,也高于类似研究中榕小蜂隐存种对间的遗传距离,其线粒体基因达到物种分化水平。3个枝系的ITS2序列显示出致同进化的特征,枝系内部基本共享相同的ITS2单倍型,而不同枝系对应的单倍型间相差7-13个核苷酸。尽管枝系三有2个对应的ITS2单倍型,但它们仅相差1个碱基,且其中1个单倍型仅局限在1个南方种群。ITS2的单倍型格局从核基因水平证实3个枝系的物种性质。3个枝系达到物种分化水平的核基因证据还来源于Cytb的NUMTs序列,在Cytb单倍型和NUMTs的联合系统树中近期的NUMTs序列没有枝系的交叉,证实枝系间已完全隔离。2)薜荔传粉小蜂3个物种互为姊妹种。在薜荔小蜂COI和Cytb单倍型与其他小蜂相应序列联合构建的系统树中,薜荔小蜂序列聚为1个大单系的支持率为99%;2个物种中同时存在Cytb基因的NUMTs序列也说明它们为近缘种。联合Cytb单倍型和NUMTs序列进行系统重建,NUMTs古老单倍型聚在外类群位置,以之为根的系统树进一步证实了3个物种近缘的系统发生关系;而来源于不同物种的NUMTs序列在外类群中聚为单系更进一步确认它们互为姊妹种的性质。3)薜荔传粉小蜂3个物种在我们东南沿海地区有一定的地域分布和物候差异,生境偏好、地理和时间隔离可能是其物种分化的原因。薜荔传粉小蜂Sp.A完全分布于北部种群,Sp.B主要分布于南部种群,而Sp.C主要分布于北部的岛屿种群。地域分布差异反映了3个物种的生境偏好,纬度、地形地貌和岛屿大陆环境等差异导致的不同选择压力,山系和海洋导致的隔离都在薜荔小蜂祖先种群的分化中起到重要作用,小蜂先于薜荔的成种事件最终导致3种传粉小蜂在不同地域的共存。3个物种的物候有一定差异,Sp.C的出飞期较Sp.A晚10天左右,这种差异甚至逆纬度梯度存在,尽管这个差异在目前条件下尚不足以隔离物种,但在成种过程中起到一定作用。4)Sp.A非随机分布,呈现出明显的大陆-岛屿格局,岛屿群来源于大陆群的向外辐射。Sp.A广泛分布于北部的27个种群,STUCTURE和NCPA软件将之区分为大陆和岛屿2个下级组,X~2置换检验表明2组非随机分布,AMOVA分析揭示大陆群与岛屿群间的差异达到显著水平,它们在单倍型组成、遗传多样性和种群动态上都有不同。与大陆群相比较,岛屿群表现出更明显的种群扩张表征,在巢式枝系图中呈现出更典型的星状结构、更多的单倍型末枝。失配分布分析和统计扩张检验都显示岛屿群经历过或正经历着扩张和瓶颈效应,而大陆群的Tajima’s D和Fs统计量都没有显著偏离0。比较遗传多样性可以看到,尽管岛屿群的样本量为大陆群的3倍,但其遗传多样性无论在整体水平还是种群平均水平均低于大陆群,特别是平均序列差异。COI岛屿特有单倍型间最多相差3个点突变,而大陆特有单倍型间可达12个。系统重建也显示大陆单倍型更为古老,它们在有根树中都深埋于内部节点。从COI单倍型格局看,所有岛屿特有单倍型都以1个点突变围绕着2个频率最高的大陆岛屿共享单倍型,且都分布在相应大陆种群的东北方向,这不仅符合小蜂的生活史特性,也与杭州湾风力场特征一致。岛屿生境剧烈波动和小种群效应导致的更频繁的岛屿种群灭绝事件、大陆种群的远距离救援、以及其后的定居、种群扩张和相对独立条件下的进化,构成了Sp.A典型的大陆-岛屿格局。5)宁德及其以南地区为Sp.B和Sp.C的潜在冰期避难所。Sp.B和Sp.C在南方和北方种群都有分布,但南方种群表现出更高的遗传多样性,拥有更多的古老特有单倍型。就物种C而言,宁德具有最多的特有单倍型,3个特有单倍型间相差8-11个点突变;而在其他11个种群中,仅桃花岛有1个特有单倍型,后者的特有单倍型与分布最广泛的共享单倍型间只相差1个点突变。从种群动态看,宁德为长期存在种群,失配分析和中性检验的各统计量都显著拒绝其经历过近期统计扩张的假设。宁德及其以南区域很可能为物种B、C潜在的冰期避难所,该地区的武夷山系为薜荔小蜂冰期种群提供了保护地,冰期后,小蜂沿山系从西南向东北方向扩散,直至薜荔分布区北缘。

【Abstract】 Mutualisms are one of the most important ecological interactions,with strong influences onalmost all levels in biological systems.Their long term persistence raises many challengingevolutionary questions.Figs and their pollinating wasps,with their great species richness anddiverse host life forms,are among the most tightly integrated pollination mutualisms known,andprovide a model system for developing and testing theories of mutualism evolution.The mutualisms of fig and fig wasps are both ancient and diverse.Results frommorphological studies suggested high species specificity between the two interacting partners.Initial molecular phylogenetic studies also suggested co-cladogenesis and coevolution at highertaxonomic levels.These studies led to the adoption of a predominant assumption of one host figtree to one pollinating wasp.However,more and more exceptions have been revealed by recent,especially molecular,studies and breakdowns in the one to one relationship may even representthe majority of relationships.In addition,the discovery of host switches of pollinating wasps andnatural hybrids between fig trees has further undermined the empirical basis of the one-to-one rule.Machado et al.(2005) have proposed a new evolutionary model,saying that the relationshipbetween fig and fig wasps is not strict-sense cospeciation between individual pairs of interactingfigs and wasps,but broad-sense coevolution between related groups,with host-shifting commonwithin groups of figs that are imperfectly defined genetically because of hybridization.However,their new model was based only on results from Ficus sections Pharmacosycea and Americana inthe Americas and may not be applicable elsewhere because some fig groups in other regions havedifferent relationships with their pollinators,with strict host specificity and cospeciation asdominant patterns.Consequently,the frequency in breakdowns of the one to one relationshipwithin different fig groups,in different regions,is essential for the understanding of the relativeimportance of the competing finer-scale cospeciation or broad-sense coevolution models.Wiebesia pumilae (Hill),originally described from Hong Kong,is the specific pollinator ofFicus pumila L.var.pumila and its variant var.awkeotsang,and shows no known morphological variation from place to place.However,the taxonomic status of F.pumila and W.pumilae areunresolved.The host was placed in Synoecia section Rhizocladus or Ficus section Ficus based ondifferent studies,while the pollinator was assigned to the genera Blastophaga or Wibesia bydifferent researchers.In addition studies in Taiwan and Fujian have found that some pollinatorsderived from var.pumila did not enter receptive figs of var.awkeotsang which suggested thatmore than one species of pollinator might be present.We employed three genes,COI (mitochondrial),Cytb (both mitochondrial and its NUMTssequences) and ITS2 (nuclear ribosomal) to investigate the extent of divergence,distributions andphylogenetic relationships within the Wiebesia pollinators of F.pumila in Southeastern China.Wesampled 331 individuals from 31 mainland and island populations.Using the softwareSTRUCTURE,ANeCA,PAUP and ARLEQUIN,we analysed the degree of genetic divergence,genetic diversity and demographic expansion dynamics.Integrating corresponding sequences inthe GenBank data base,we reconstructed phylogenetic trees.We also assessed the speciesspecificity of F.pumila and its pollinators,and estimated the influence of habitat differencesbetween mainland and islands on W.pumilae populations.The main results were as follows:1) The pollinators of F.pumila in Southeastern China showed deep genetic divergence,andcould be divided into three distinct clades.The genetic distances between any two clades exceededthose among conspecific individuals based on both mitochondrial and nuclear genes.Nomorphological differences were evident between the three clades.They should be recognized asthree cryptic species,named as Sp.A,Sp.B and Sp.C.The 331 wasp individuals were consistently placed into the three identical clades,irrespectiveof the genes and methods employed.All three clades were well defined genetically.Theprobabilities of assignment of individuals to each clade,assigned by STRUCTURE with1,000,000 repetitions,were all more than 95%.The three clades could not be linked togetherwithin the 95% statistical parsimony criterion by TCS integrated into ANeCA.The bootstrappercentages of the three monophyletic clades in different phylogenetic trees are all 99%.The divergences of COI between any two clades were more than ten times the averagedifferences within clades.The Kimura-2-parameter differences between clades were 6.63%, 10.94% and 10.69% respectively,while the average of those within each clades was only 0.57%.The distances between any two clades were not only beyond the standard COI barcoding thresholdfor animal species,but also higher than the known maximum found in similar studies on crypticfig wasps.The nuclear genes of W.‘pumilae’also showed deep divergence.Different individuals sharedthe same ITS2 haplotypes within two clades,while the haplotypes in the three different cladesdiffered from each other by 7 to 13 mutations.Although two ITS2 haplotypes were detected in oneof the three clades,they were highly homogenized with just one different nucleotide out of 454bpsequences.Moreover one of these two haplotypes was limited to just one population.Thedistribution of ITS2 haplotypes coincided with the concerted evolution of the ribosome gene,further demonstrating interspecies differences among the three clades.NUMTs of Cytb verified complete isolation among the three clades.The recent NUMTshaplotypes derived from the individuals from different clades clustered with corresponding Cytbhaplotypes only within each clade in the combined phylogenetic trees,rejecting the possibility ofrecent gene flow among clades.2) The three cryptic pollinators of F.pumila are sister species.When clustered with the corresponding sequences of other pollinating fig wasps in GenBank,all the W.pumilae haplotypes were fixed in one monophyly with bootstraps of 99%,based on bothCOI and Cytb.The co-occurrence of NUMTs sequences converted from the same fragment ofmitochondrial genome,Cytb,in two of the three species,also suggested close relationship amongthem.In addition,the NUMTs haplotypes derived from different clades clustered together inoutgroups,illustrating their recent common ancestry.3) The three cryptic species display divergence in distribution and phenology.Habitatpreferences,geographical and perhaps temporal isolation could explain the co-existence of threepollinators in Southeastern China.Wiebesia‘pumilae’Sp.A was widespread in almost all the northern populations,Sp.B wasdistributed mostly in southern populations,while Sp.C was mostly found in the northern islandpopulations.Their distributions suggested different habitat preferences,possibly related tolongitude,landforms,and mainland versus island environments.Habitat specialization and geographic barriers may have promoted the division and isolation of pollinators withoutequivalent host plant speciation.Sp.A and Sp.C displayed small differences in phenology,the former emerging about 10 daysearlier than the latter,independent of longitude.The difference in emergence time could havefavoured divergence among ancient populations of W.‘pumilae’.4) Sp.A is distributed unequally in mainland and island populations,suggesting colonizationfrom the mainland to the islands.Sp.A was found in 27 of the 31 populations.Chi-square tests revealed a nonrandomdistribution of these populations,with two repartitioned groups mainly on the mainland andislands respectively,as defined by STUCTURE and ANeCA.AMOVA analysis revealedsignificant differences between mainland and island groups.Compared with the mainland group,the island populations displayed more obvious signs ofdemographic expansion.The island-specific haplotypes showed a typical star-like network withmuch more tip haplotypes.Recent or undergoing demographic expansion and bottleneck eventsin the island group were verified by both mismatch distribution analysis and neutrality tests,while Tajiam’s D and Fs statistics did not deviate significantly from zero in the mainland group.The island group showed lower genetic diversity than the mainland,despite larger samplesizes of populations and individuals.Both the total and average population genetic diversities inthe islands were lower than in the mainland group,especially the mean number of pairwisedifferences.The maximum difference in mainland-specific COI haplotypes was 12 mutations,while that of the islands was just 3,indicating an ancient existence of mainland haplotypes.Furthermore,all the mainland specific haplotypes were settled in deep inner nodes inphylogenetic trees,further affirming their ancient origin.All island specific COI haplotypes were linked to two shared haplotypes with one mutation.The two shared haplotypes were distributed widely in both mainland and island populations,withthe highest frequencies among all haplotypes.The distribution pattern of these two sharedhaplotypes was in concordance with the characteristics of sea-land breezes in Hangzhou Bay.Many factors may have contributed to the mainland-island pattern of Sp.A,such as morefrequent extinction events on the islands due to greater population fluctuations and smallerpopulation sizes,individual radiations from mainland populations,colonization,expansion and independent evolution in the relatively isolated island habitats.5) There were potential glacial refugia for Sp.B and Sp.C,in Ningde and the southernregions.The southern populations of Sp.B and Sp.C had higher genetic diversity and more ancientspecific haplotypes than elsewhere.Four population-specific Sp.C COI haplotypes were detected,three in Ningde,one in Taohua.The three specific haplotypes in Ningde were distinguished fromeach other by 8 to 11 mutations,while that in Taohua was linked to the most frequently sharedhaplotype by 1 mutation.Mismatch distribution analysis and neutrality tests detected nosignificant recent expansion events in the Ningde population and all statistics showed that it was astable population that had existed for a long period.So Ningde and other southern regions werepotential glacial refugia of W.‘pumilae’.The Wuyi Mountains there could provide protection forthe populations during glacial periods.After the last glacial maximum,wasps may haverecolonized north-eastern regions along the Wuyi Mountains.In conclusion,F.pumila is pollinated by three genetically distinguishable cryptic sisterspecies in Southeastern China.The three species coexist over much of the range of their host plant,with no evidence of hybridization,but vary in their relative abundance within the range of theirhost.Some differences in their biology were detected.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络