为数众多的“捣蛋”序列
Partha S. Saha, Simon J. Newkirk, Changfeng Yao, Wenfeng An
Department of Pharmaceutical Sciences, South Dakota State University
《新英格兰医学杂志》2017年7月27日发表了一篇关于可移动DNA与人类疾病的综述[1]。近年来，我们对可移动DNA元件在多种人类疾病中的作用有了越来越丰富的认识，因此这篇综述的发表非常及时。
在该综述中，作者Kazazian和Moran涵盖了许多重要的主题，包括（1）人类基因组中可移动DNA的种类；（2）作为诱变剂的逆转座子；（3）逆转座的机制；（4）逆转座介导的基因组重排；（5）体细胞与癌症中的逆转座事件；（6）宿主抵御可移动子的机制。在这篇对该综述的评论中，我们主要讨论了发育期间可移动DNA的活性，特别是这类DNA与人类繁衍之间千丝万缕的联系。
为数众多的“捣蛋”序列
说到可移动DNA，就不得不提LINE-1——一种在人类基因组中极为丰富的重复序列。部分LINE-1是逆转座子，即LINE1的DNA被复制成RNA，随后这些RNA会被LINE1编码的逆转录酶复制回DNA，插入基因组。根据逆转座过程的发生时间和细胞特性，LINE-1插入可被分为体细胞或生殖细胞系事件。
图1. LINE-1逆转录转座的模型[1]
LINE-1的RNA由位于其5'端非编码区的启动子转录。其RNA被导出至胞质，并在胞质中进行翻译。LINE-1编码的蛋白ORF1p和ORF2p通过称为顺势优先的过程结合至LINE-1编码的RNA上，形成胞质复合体。这一复合体的组分（至少ORF2p和LINE-1的RNA）可以进入细胞核，此时，ORF2p内切核酸酶（endonuclease，EN）在共有序列（例如，5'-TTTT/A-3'）处切开染色体DNA双螺旋的一条链，暴露3'端羟基。ORF2p逆转录酶使用此基团复制LINE-1的RNA并将其整合入染色体新位置。现在还不知道插入位点处的第二条DNA链是如何被切割，以及LINE-1 DNA的第二条链如何合成，但ORF2蛋白可能介导了这些过程。非自主RNA，如细胞内（生成已加工假基因的）mRNA、SVA 的RNA及Alu的RNA，其逆转录转座也需要ORF2蛋白。ORF1蛋白可能辅助SVA和Alu的逆转录转座。RNP表示核糖核蛋白颗粒。本示意图改编自Richardson等人的论文。
在癌症活检标本中发现的插入大多发生在体细胞中，而且对癌细胞特异[2]。一项对结肠癌患者进行的近期研究显示，体细胞插入一般起源于供体LINE-1元件。在正常情况下，这些元件被DNA甲基化介导的转录沉默牢牢地限制在原地；但在特殊情况下，它们可以逃脱这种限制而复制到新的插入位点[3]。
与体细胞事件相反，生殖细胞系中的插入可以遗传，并能在世代之间传递，引发各种遗传性疾病的风险[4]。正如这篇综述中讨论的那样，最著名的病例就是A型血友病，它是被确认的第一种由LINE-1插入导致的疾病，突变的基因是凝血因子VIII。
图2. 一名A型血友病男童的LINE-1逆转录转座[1]
图A显示了家系图谱。橙色实心方块表示患病男童，其他符号表示未患病的家族成员。男童的母亲未携带F8（编码凝血因子Ⅷ的基因）第14外显子上的插入。在图B中，分离自男童母亲DNA的LINE-1元件前体有上述插入。这一前体是22号染色体上的全长LINE-1。较短的致病性插入的序列与部分LINE-1前体完全相同。元件前体5'端的棕色箭形表示LINE-1内部启动子的转录起始位点。TSD表示靶点复制，TSD旁边的浅灰色箱形表示基因组DNA。图A中的示意图改编自Kazazian等人的论文，图B中的示意图改编自Dombroski等人的论文。
这还是个问题
科学家们在生殖细胞系LINE-1插入的发生时间问题上仍然争吵不休。理论上，生殖细胞系插入可以发生在生殖细胞发育的任何阶段（图3）。在人和啮齿动物中，生殖细胞系起源于胚胎发育早期的一小群原始生殖细胞（primordial germ cells，PGC）。这些PGC会迁移到生殖嵴，增殖、分化成四倍体精母细胞或卵母细胞，并最终通过减数分裂，成熟为单倍体精子或卵细胞。
对于男性，生殖细胞系LINE-1插入可发生于PGC、精母细胞、精子或之间的任意阶段。然而，生殖细胞系插入也可能发生于PGC特化之前。例如，如果逆转座事件发生在随后分化成PGC的前体细胞中，就可能产生这种情形。在这种情况下，如果某些前体细胞的子细胞形成体细胞系，那么在某些体细胞中也可能存在相同的插入。
图3. LINE-1在生殖细胞系插入的时间。
原始生殖细胞（PGC）创始细胞簇出现于胚胎发育早期，迁移到生殖嵴，增殖、分化并最终成熟为精子细胞。根据当前的研究，图中用闪电符号标注出易发生LINE-1逆转座的时间点。
科学家对某个问题存在争议，往往是因为这个问题很难研究。对于LINE-1插入的争论也不例外。新的LINE-1内源性插入发生的频率相对较低，这就给在已有大量LINE-1存在的基因组背景下检测新的插入带来了一个重大挑战。
为了便于检测插入频率，早期研究利用了经过基因工程改造的LINE-1转基因模型，以帮助研究者定量鼠类发育过程中的逆转座频率[5-7]。令人惊讶的是，在使用天然LINE-1启动子调控LINE-1转基因的模型中，在早期胚胎中观察到了频繁的插入[7]。同时，在这类模型鼠类子代中未发现可遗传的插入[7]。这与使用组成型启动子（constitutive promoter）调控LINE-1转基因的小鼠模型形成了鲜明的对比[5,6]。这些结果表明，可能有一种机制保护原始生殖细胞及其前体不受逆转座的过度影响。尽管其可能性尚未被试验验证，但极有可能的是，此类转基因在这些细胞中被转录沉默了。
有关LINE-1逆转座发生时间的另一种观点来自于Faulkner团队的一项近期研究[8]。通过第二代测序，该团队发现了共计11例新的全长LINE-1插入。他们确认，内源性的逆转座能够发生在胚胎发生早期，雌性胚胎（5/11）中的发生频率似乎比雄性胚胎（1/11）高。6例早期胚胎插入中，有4例在生殖细胞系中也能找到。其他3例插入似乎发生在雄性胚胎的早期PGC中。一例插入被认定发生在亲代中任意一方的晚期生殖细胞系中。总之，Faulkner等人的研究提供了引人注目的证据，证明新的LINE-1插入可以发生在更大的发育时间范围内——从早期胚胎到早期PGC，再到晚期生殖细胞，且雄性和雌性均可发生。然而，受较小的样本数目所限，我们仍不能确定大多数逆转座事件发生在哪个发育时期。
新发现的LINE-1调控因子
为了保护基因组的完整性，生物体进化出了多重防护机制以限制逆转座活动。男性生殖细胞中最重要的两个保护神是DNA甲基化和与PIWI家族蛋白相互作用的RNA（PIWI-interacting RNA，piRNA）通路[9]。一方面，逆转座子的新甲基化需要DNMT3L （DNA methyltransferase 3-like）蛋白。Dnmt3l缺陷小鼠患有雄性特异性减数分裂阻滞和不育[10]。另一方面，piRNA是生殖细胞中大量表达的小型非编码RNA。在胚胎雄性生殖细胞中，两种PIWI蛋白（MILI和MIWI2）与逆转座子来源的piRNA相互作用，致使逆转座子发生转录沉默（通过将DNA甲基化和H3K9me3靶向至逆转座序列）和转录后沉默[9]。在小鼠敲除模型中，piRNA通路的突变体必然导致雄性不育，并伴有逆转座抑制解除和精子发生阻滞。正如综述中所述，我们需要进一步研究解除逆转座表达的抑制是否引起减数分裂灾难（meiotic catastrophe）或导致逆转座子数目的增加。然而，最近发表的两篇论文为这些问题提供了重要的见解[11,12]。
在一项研究中，我们将一种新型单拷贝LINE-1转基因导入了piRNA缺陷的基因背景（如，Mov10l1敲除）中[11]。与以前的转基因模型相比，这一新型LINE-1报告转基因的优点是它受内源性小鼠LINE-1启动子调控。由于Mov10l1缺陷的生殖细胞没有再次甲基化，因此我们证实，piRNA通路调控该启动子。
在这项研究中，我们也开发了一种高灵敏度的微滴式数字PCR（droplet digital PCR），用来对LINE-1转基因插入定量。使用这项方法，我们发现，直到出生后第14天，第一群雄性生殖细胞开始减数分裂成为精母细胞时，在Mov10l1敲除的小鼠睾丸中的逆转座活性才发生了改变。这时，插入数目比对照组野生型小鼠提高了70倍。这种增高的插入水平会保持到成年期。我们分选了成体睾丸中的细胞组分，确定逆转座增加特异性发生于精母细胞的细线期和偶线期（比野生型高144倍）[11]。因此，我们的研究确定，解除逆转座子表达的抑制的确会导致piRNA缺陷的生殖细胞中逆转座的增加。
在Mov10l1敲除小鼠的睾丸中，逆转座的急剧增加正好发生于生殖细胞死亡之前[11]。这一发现是否表明解除逆转座表达抑制导致了piRNA缺陷动物中的减数分裂灾难和雄性不育呢？迄今为止，对于这个问题仍然没有定论。但多方证据表明，减数分裂方面的表型不是插入形成突变的结果。首先，当我们给予小鼠可有效降低LINE-1逆转座的核苷逆转录抑制剂时，表型没有被拯救[11]。其次，我们使用本研究组的转基因模型推测了内源性逆转座的频率。估计结果，内源性插入的频率是每个细胞2.3次。我们的结论是，这一水平的插入形成的突变过低，无法导致精母细胞死亡[11]。第三，Bourc’his及其同事进行的独立研究尝试直接对来自内源性LINE-1的逆转座进行定量[12]。通常，由于基因组拷贝的数量过多，定量新发内源性LINE-1插入的方法灵敏度不足。虽然如此，与我们的结果一致，这些作者没有在Dnmt3l-缺陷的生殖细胞中发现大量逆转座活性[12]。
那么，解除逆转座抑制是如何在减数分裂表型和雄性不育中发挥作用呢？Bourc’his及其同事提供了一些有趣的证据，支持了LINE-1激活可能在染色质水平发挥作用12。与野生型对照相比，Dnmt3l缺陷的生殖细胞的染色质图谱发生了改变：逆转座子序列上的抑制性H3K9me2修饰缺失及激活性H3K4me3修饰增加。这些生殖细胞染色质图谱上的改变与在逆转座子序列中形成减数分裂双链断裂有关。通常，在减数分裂重组位点处没有逆转座子。这种异常分布可能是Dnmt3l缺陷的精母细胞中染色体联合失败及后续减数分裂过程中断的原因[12]。
毫无疑问的是，要进一步定义逆转座子在减数分裂过程中的作用，科学家还需要进行更多研究。例如，LINE-1的ORF2的蛋白具有核酸内切酶活性剂已知的细胞毒性[9]，但我们从未探索过高表达ORF2蛋白如何影响生殖细胞发育。另一方面，我们的近期研究也表明，可移动DNA与男性生育力之间有着互相依存、互惠互利的关系。简言之，逆转座过多可对生殖细胞产生不利影响，导致不育。反之，为了繁殖下一代，逆转座必须是适度的，且宿主要具备生育力。
基于我们的结果，我们预测，许多生殖细胞系插入起源于逆转座子调控水平折中、但仍有生育能力或生育力较低的个体[11]。这些个体可能以有隐睾、少精或原位睾丸癌病史的患者呈现，而这些患者往往与基因变异体或piRNA通路基因的表达降低相关。若这些预测得到证实，那么我们对LINE-1逆转座在发育中的发生时间，以及LINE-1介导的基因变异的来源和影响方面的知识将会显著改进。在雄性生殖细胞发育的晚期阶段，在精母细胞和精细胞中新的插入可能不会对这些生殖细胞本身的状况即时产生影响。然而，若传播到子代，这样的插入是有害的，会造成遗传性疾病。因此，生殖细胞系中的新逆转座事件可能是人类群体中存在的罕见有害基因变异体的重要来源。
A review on “Mobile DNA and Human Disease” was published on the July 27 issue of NEJM [1]. It is a very timely piece given the increased recognition of the contribution of mobile DNA elements in a variety of human diseases.
In the review, Drs. Kazazian and Moran covered many important topics, including (1) classes of mobile DNA elements in the human genome, (2) retrotransposons as mutagens, (3) mechanism of retrotransposition, (4) retrotransposon-mediated genomic rearrangements, (5) retrotransposition events in somatic cells and cancer, and (6) host defense mechanisms against mobile elements. In this commentary, we wish to further highlight the activities of mobile DNA during development, especially its connection to male infertility, by discussing recent findings in this specific area.
LINE-1 insertions can be categorized as either somatic or germline events depending on the timing and cell specificity of the retrotransposition process. Insertions found in cancer biopsies are mostly somatic and specific to the cancer cells [2]. As shown in a recent study of colorectal cancer patients, somatic insertions typically originate from donor LINE-1 elements that have escaped transcriptional silencing by DNA methylation [3]. In contrast to somatic events, germline insertions are heritable and can be passed from generation to generation, causing the risk of various heritable disorders [4]. Most notably, as discussed in the review, hemophilia A was the first disease identified to be caused by a mutagenic LINE-1 insertion in the factor VIII gene.
The timing of germline insertions remains controversial. In theory, germline insertions can arise at any stage during germ cell development (Figure 3). In both humans and rodents, the germ cell lineage is specified as a small cohort of primordial germ cells (PGCs) at an early stage in the developing embryo. These PGCs migrate to the future gonads, proliferate, differentiate and become tetraploid spermatocytes or oocytes, and eventually mature into haploid sperm or eggs through meiosis. Thus, for males, germline insertions can occur in either PGCs, spermatocytes, sperm, or any stages in between. In these cases, the insertions are only present in germ cells. However, germline insertions may also take place prior to PGC specification. This scenario is possible if a retrotransposition event occurs in the progenitor cells that later differentiate into PGCs. In this case, the same insertion may also be present in some somatic cells if some daughter cells of the progenitor cells form somatic lineages.
The timing of germline insertions has been difficult to study. The relatively low frequency of de novo endogenous insertions poses a significant challenge for detection against the genomic background of abundant preexisting LINE-1s. As a result, earlier studies took advantage of genetically engineered LINE-1 transgenes, which facilitate the quantification of retrotransposition frequencies during rodent development [5-7]. Strikingly, in models using LINE-1 transgenes regulated by the native LINE-1 promoter, insertions were frequently detected in early embryos [7]. On the other hand, no heritable insertions were found in the progeny [7], which is in sharp contrast to mouse models using LINE-1 transgenes that are regulated by constitutive promoters [5,6]. These results suggest that PGCs and progenitors of PGCs may be protected from excessive retrotransposition activities. Most likely, the transgenes are transcriptionally silenced in these cells, although this possibility has yet to be experimentally verified.
Additional insights about the timing of LINE-1 retrotransposition came from a recent study by the Faulkner group [8]. A total of 11 de novo full-length LINE-1 insertions were identified by using a NextGen sequencing approach. Endogenous retrotransposition was confirmed to take place in early embryogenesis, likely more frequently in female embryos (5 out of 11 insertions) than in male embryos (1 out of 11). Four of these six early embryonic insertions were also found in the germline. Three other insertions appeared to arise in the early PGCs of male embryos. One insertion was determined to be a late germline event in either parent. The developmental origin of the eleventh insertion was not clear. Taken together, the Faulkner study furnished compelling evidence that new LINE-1 insertions could occur in a broad developmental time range, from early embryos, to early PGCs, to late-stage germ cells, as well as in both males and females [8]. However, limited by the small sample size, no definitive conclusions can be drawn as to which developmental stage(s) supports the most retrotransposition events.
To protect genomic integrity, multiple defense mechanisms have evolved to restrict retrotransposition activities. In male germ cells, the two most important defense mechanisms are DNA methylation and the PIWI-interacting RNA (piRNA) pathway [9]. De novo methylation of retrotransposons requires the DNA methyltransferase 3-like (DNMT3L). Dnmt3l-deficient mice suffer from male-specific meiotic arrest and infertility [10]. piRNAs are small non-coding RNAs abundantly expressed in germ cells. In fetal male germ cells, two PIWI proteins (MILI and MIWI2) interact with retrotransposon-derived piRNAs, acting to transcriptionally (via targeting DNA methylation and H3K9me3 to retrotransposon sequences) and post-transcriptionally silence retrotransposons 9. In mouse knockout models, mutants of the piRNA pathway invariably exhibit male infertility, which is accompanied by derepression of retrotransposons and spermatogenic arrest. As mentioned in the review, whether the derepression of retrotransposon expression causes meiotic catastrophe or leads to increased retrotransposition requires further study. However, two very recent publications provided important insights with regards to these questions [11,12].
In one study, we introduced a novel single-copy LINE-1 transgene into a piRNA-deficient genetic background (e.g., Mov10l1 knockout) [11]. The advantage of this new LINE-1 reporter transgene over previous transgenic models is that the LINE-1 transgene is controlled by an endogenous mouse LINE-1 promoter. We confirmed that this promoter was regulated by the piRNA pathway, as evidenced by the lack of remethylation in the Mov10l1-deficient germ cells. In this work, we also developed a highly sensitive Droplet Digital PCR based approach to quantify insertions by the LINE-1 transgene. Using this approach, we found that retrotransposition did not change in the knockout testes until postnatal day 14 when the first cohort of male germ cells has commenced meiosis and become spermatocytes. At this time point, insertions increased 70-fold relative to the control wild-type mice. The elevated level of insertions persisted into the adult stage. Using sorted cell fractions, we determined that the increase in retrotransposition occurred specifically in leptotene and zygotene spermatocytes from the adult testes (144-fold higher than that of the wild type counterpart) [11]. Therefore, our study established that the derepression of retrotransposon expression does lead to increased retrotransposition in piRNA-deficient germ cells.
The dramatic increase in retrotransposition immediately precedes germ cell death in the Mov10l1 knockout testes [11]. Is this observation an indication that the derepression of retrotransposon expression causes meiotic catastrophe and male infertility in piRNA-deficient animals? Thus far, the answer to this question is not conclusive but several lines of evidence suggest the meiotic phenotype is not the result of insertional mutagenesis. First, there was a lack of phenotypic rescue when mice were treated with a nucleoside reverse transcriptase inhibitor, which was effective in reducing LINE-1 retrotransposition [11]. Second, using our transgenic model, we extrapolated the frequency of endogenous retrotransposition. The insertion frequency was estimated to be 2.3 per cell. We concluded that this level of insertional mutagenesis was too low to be the cause of death of spermatocytes [11]. Third, a separate study by Dr. Bourc’his and colleagues attempted to directly quantify the change in retrotransposition from endogenous LINE-1s [12]. Typically, assays quantifying new endogenous LINE-1 insertions do not have sufficient sensitivity due to the overwhelming number of genomic copies. Nevertheless, in agreement with our results, these authors did not find massive retrotransposition in Dnmt3l-deficient germ cells [12].
Then, how might derepression of retrotransposons contribute to the meiotic phenotype and male infertility? The study from Dr. Bourc’his and colleagues provided some intriguing evidence that supports a potential role of LINE-1 activation at the chromatin level [12]. As compared to the wild-type control, Dnmt3l-deficient germ cells manifest an altered chromatin landscape: the loss of repressive H3K9me2 modification and the gain of activating H3K4me3 modification at retrotransposon sequences. These changes in germ cell chromatin landscape are correlated with the formation of meiotic double-strand breaks in retrotransposons. Normally, retrotransposons are not present in meiotic recombination sites. This abnormal distribution may be responsible for synaptic failure and subsequent interruption of meiotic progression in Dnmt3l-deficient spermatocytes [12].
Clearly, additional studies are required to further define the role of retrotransposons during meiotic progression in piRNA knockouts. One of the unexplored areas is how the overexpression of LINE-1 ORF2 protein, which has an endonuclease activity with known cytotoxicity, affects germ cell development [9]. On the other hand, our recent study also suggests that the relationship between mobile DNA and male fertility is analogous to a two-way street. Simply put, too much retrotransposition can be detrimental to germ cells and leads to infertility. Conversely, in order to propagate to the next generation, retrotransposition must be modest and the host needs to be fertile. Based on our results, we predict that many germline insertions originate from individuals who have partially compromised retrotransposon control but remain fertile or subfertile [11]. These individuals may be presented as those with a history of cryptorchidism, oligozoospermia, or testicular carcinoma in situ, which are frequently associated with genetic variants or reduced expression of piRNA-pathway genes. If these predictions are borne out, our understanding of the developmental timing of LINE-1 retrotransposition as well as the origin and impact of LINE-1 mediated genetic variation will be significantly revised. De novo insertions in spermatocytes or spermatids, the late stages of male germ cell development, may not have an immediate impact on the fitness of the germ cells themselves. However, such insertions can be harmful if propagated to progeny, causing heritable disorders. Therefore, germline de novo retrotransposition events can be a substantial source for rare, harmful genetic variants that are present in the human population.
专家介绍
安文锋，现任美国南达科塔州立大学药学院副教授，兼该学院首席Markl癌症研究学者。长期从事逆转座子分子生物学及遗产学方面的研究。在北京医科大学（现北京大学医学部）获医学学士和流行病学硕士学位，从密西根大学获微生物及免疫学博士学位，曾在约翰霍普金斯大学医学院的Jef Boeke实验室从事博士后研究，研究核心为LINE-1逆转座子遗传学和表观遗传学调控及其在生殖细胞发育和癌变过程中的功能。详细研究方向可查实验室网址：https://www.sdstate.edu/pharmacy-allied-health-professions/laboratory-wenfeng-1
欢迎来稿
如果你
关注科研进展，熟谙临床医学
洞察内涵规律，志在传播交流
我们期待
你将NEJM的医学之美，讲与我听
submission@nejmqianyan.cn
参考文献
[1] Kazazian HH, Jr., Moran JV. Mobile DNA in Health and Disease. N Engl J Med 2017;377:361-70.
[2] Burns KH. Transposable elements in cancer. Nat Rev Cancer 2017;17:415-24.
[3] Scott EC, Gardner EJ, Masood A, Chuang NT, Vertino PM, Devine SE. A hot L1 retrotransposon evades somatic repression and initiates human colorectal cancer. Genome Res 2016;26:745-55.
[4] Hancks DC, Kazazian HH, Jr. Roles for retrotransposon insertions in human disease. Mob DNA 2016;7:9.
[5] Ostertag EM, DeBerardinis RJ, Goodier JL, et al. A mouse model of human L1 retrotransposition. Nat Genet 2002;32:655-60.
[6] An W, Han JS, Wheelan SJ, et al. Active retrotransposition by a synthetic L1 element in mice. Proc Natl Acad Sci U S A 2006;103:18662-7.
[7] KanoH, Godoy I, Courtney C, et al. L1 retrotransposition occurs mainly in embry ogenesis and creates somatic mosaicism. Genes Dev 2009;23:1303-12.
[8] Richardson SR, Gerdes P, Gerhardt DJ, et al. Heritable L1 retrotransposition in the mouse primordial germline and early embryo. Genome Res 2017;27:1395-405.
[9] Newkirk SJ, An W. L1 Regulation in Mouse and Human Germ Cells. In: Cristofari G, ed. Human Retrotransposons in Health and Disease: Springer International Publishing; 2017:29-61.
[10] Bourc'his D, Bestor TH. Meiotic catastrophe and retrotransposon reactivation in male germ cells lacking Dnmt3L. Nature 2004;431:96-9.
[11] Newkirk SJ, Lee S, Grandi FC, et al. Intact piRNA pathway prevents L1 mobilization in male meiosis. Proc Natl Acad Sci U S A 2017;114:E5635-E44.
[12] Zamudio N, Barau J, Teissandier A, et al. DNA methylation restrains transposons from adopting a chromatin signature permissive for meiotic recombination. Genes Dev 2015;29:1256-70.
联系客服
