方舟子的Blog

英国《自然》2016年8月8日
记者：David Cyranoski

（方舟子翻译）

一种据称能替代流行的CRISPR–Cas9系统的基因编辑技术实际上是否有效的争议正在升级。

三个月前，石家庄河北科技大学一个生物学者韩春雨报告说，NgAgo这种酶能够用于编辑哺乳动物的基因。现在，越来越多的科学家在抱怨说他们无法重复韩的结果，虽然有一个告诉《自然》说他能重复出来。

韩说他每天收到几十个骚扰电话和短信，嘲笑他并告诉他他的职业生涯结束了【译按：韩的这个说法无法证实而且极其可疑。因为韩威胁那些反映重复不出其结果的人，他们在网上大多匿名，不太可能去给韩打骚扰电话。】，但是他深信这项技术是可靠的。他告诉《自然》，他已在8月8日应遗传信息存档网站Addgene的要求向其提交了一份详细的实验步骤，并希望这将有助于重复出其工作。【译按：韩提交的这个实验步骤新增加的内容就是他以前已在中文论坛回应过的“要注意预防支原体感染”之类，并无实质性内容。】发表该研究的《自然·生物技术》正在调查此事。

这涉及到重大利益。在过去的几年内，CRISPR–Cas9系统已改变了生物学。但是它也让科学家们渴望有其他方法扩充基因编辑工具：NgAgo是几种已出现的新方法中的一种。“我们很多人真的是在为其欢呼并希望它能行。”哈佛医学院的遗传学家George Church说。

CRISPR–Cas9用一小段遗传序列来引导一种酶在特定的位置切割DNA。受其启发，韩查阅文献寻找其他能被引导的蛋白质“剪刀”，发现一个叫Argonaute的蛋白质家族符合要求。其他人此前已经把这类蛋白质列为有可用于基因编辑的潜力。

在那篇论文中，韩的团队报告说，使用一系列变化多样的遗传序列来引导这些蛋白质中的一种，NgAgo，能编辑人类细胞中的八个不同基因，并把基因插入染色体中特定的位点。

韩说，关键是，NgAgo特异性很强，只切割目标基因，而CRISPR–Cas9有时会编辑错误的基因。而且，CRISPR–Cas9需要在靠近切割位点的地方有一段特定的遗传序列来引发其切割，而NgAgo不需要，这样就扩大了其可能的用途，韩补充说。

在中国，对这项研究的最初反应是一片颂扬，包括中央电视台也去韩的实验室采访。韩说，这让他承受不起。他是一个孤僻的人。他的爱好包括收集茶叶和弹奏古琴。他不喜欢旅行，从没有离开过中国：今年3月去杭州拜访一个合作者是他活到42岁第一次坐飞机。在其论文发表之前，“没有人知道我”，韩说。他在其实验室和附近一家餐馆接受了《自然》的采访。

7月初，对韩春雨研究的怀疑开始出现。当时一个以揭露造假科学家出名的前生物化学学者方是民（方舟子）在他主办的网站新语丝撰文说，他接到了试图重复韩春雨研究的努力都失败的报告，指控说韩的论文是无法被重复的。随后在各个中文网站上有了越来越多的对韩的批评。

7月29日，这一争议走向了国际。当时，澳大利亚国立大学的一名遗传学家Gaetan Burgio在其博客上贴出了他试图重复该实验但是失败的详细情况。他的博文通常只有几十个点击，但是这篇博文的点击涨到了超过5000。

同一天，马德里西班牙生物技术国家中心的遗传学家Lluís Montoliu给国际转基因技术学会的同事发去电子邮件，建议大家“放弃任何涉及使用NgAgo的项目”“以免浪费时间、金钱、动物和人员”。这封邮件被泄漏并张贴在方的网站上。

从那以后，英国爱丁堡MRC再生医学中心的分子生物学家Pooran Dewari发起的网上调查发现，只有9个研究者说NgAgo有效【译按：已知这9人大部分是水军，例如其中有5人其实是同一个人，在十几分钟内连续投票。】，而97人说它无效。

在一个网上聊天群里，有两个研究者一开始报告说NgAgo有效，现在说他们搞错了。新德里CSIR-基因组学和整合生物学研究所分子生物学家Debojyoti Chakraborty说他重复了韩论文中的一部分工作，这部分描述使用NgAgo敲除一个被转入细胞中的编码荧光蛋白的基因。荧光减弱了，所以Chakraborty假定NgAgo让荧光基因失效了。但是在对DNA做了测序之后，他没有发现基因编辑的证据。他现在说，荧光的减弱一定是由别的因素导致的。

海德堡德国癌症研究中心基因组学博士生Jan Winter说他有相似的经历。“我将在接下来的几周再试试这个实验，不过我认为它不会有效。”他说。

韩说他只有在由他实验室培养的细胞中该系统才有效，在他买来的细胞中则无效。他后来发现买来的细胞受一种叫支原体的细菌污染，他说其他人可能有相同的问题。他补充说，有些研究生也许实验做得太快，使用试剂不小心。Winter不同意：“我认为问题不是出在这些重复不出来的科学家们做错了什么。”

在中国，有一个不属于韩研究小组的研究者由于不想纠缠到公共争议中而不愿公开姓名，他告诉《自然》说他在几种细胞中测试了NgAgo，发现它能在想要的位点引入基因突变，而且他用测序证实了这一结果。他补充说，这个过程的效率不如CRISPR–Cas9，需要优化来提高效率。“但是，简单地说，它有效。”他说。【译按：此人貌似就是中科院上海神经所研究员仇子龙，此前一直在网上公开力挺韩春雨，并攻击质疑韩春雨的人，为什么上了《自然》反而要匿名呢？】

还有两个中国科学家也要求匿名，说他们初步的结果显示NgAgo有效，但是还需要用测序来确认。【译按：还没有测序确认，就不能说有效。】

“它也许会有效，”Burgio说，“但是即便如此，它也太难搞了，不值得去搞。它不会超越CRISPR，希望太渺茫了。“

NgAgo的失败“会令人失望，但是我们还些事要做，看看Argonaute系统是不是在某些情况下能有效，”荷兰瓦赫宁恩大学微生物学家John Van der Oost说，他是2014年一篇分析Argonaute蛋白的论文的共同作者，这篇论文为它们在基因编辑中的用途奠定了基础。

本周，《自然·生物技术》给《自然》的新闻团队发来一份声明，说“几个研究者”已联系该刊报告说他们无法重复出韩春雨的结果，“本刊正按照既定的步骤调查这个问题”。该刊一个发言人拒绝评论调查的性质或期限。

据中国国家媒体报道，河北科技大学说它将要求韩在一个月内重复实验并能被第三方确认。

http://www.nature.com/news/replications-ridicule-and-a-recluse-the-controversy-over-ngago-gene-editing-intensifies-1.20387

NATURE | NEWS

Replications, ridicule and a recluse: the controversy over NgAgo gene-editing intensifies

As failures to replicate results using the CRISPR alternative stack up, a quiet scientist stands by his claims.

David Cyranoski

08 August 2016
SHIJIAZHUANG, CHINA

A controversy is escalating over whether a gene-editing technique proposed as an alternative to the popular CRISPR–Cas9 system actually works.

Three months ago, Han Chunyu, a biologist at Hebei University of Science and Technology in Shijiazhuang, reported that the enzyme NgAgo can be used to edit mammalian genes. Now, an increasing number of scientists are complaining that they cannot replicate Han’s results — although one has told Nature that he can.

Han says he receives dozens of harassing calls and texts each day, mocking him and telling him that his career is over — but he is convinced that the technique is sound. He also told Nature that he had submitted a detailed protocol to the online genetic-information repository Addgene on 8 August, at Addgene’s request, and hopes that this will help efforts to reproduce his work. Nature Biotechnology, which published the research, is investigating the matter.

The stakes are high. Over the past few years, the CRISPR–Cas9 system has transformed biology. But it has also made scientists hungry for other methods to expand the gene-editing toolkit: NgAgo is one of several that have emerged. “A lot of us are really cheerleading and hoping that it works,” says geneticist George Church of Harvard Medical School in Boston, Massachusetts.

CRISPR–Cas9 uses small genetic sequences to guide an enzyme to cut DNA in a particular location. Inspired, Han looked through the literature for other guidable protein ‘scissors’, and came across a family of proteins called Argonaute, or Ago, that fitted the bill. Others had flagged the proteins as potential gene editors.

In the paper, Han’s team reports using a wide variety of genetic sequences to guide one of these proteins, NgAgo, to edit eight different genes in human cells and to insert genes at specific points on chromosomes (F. Gao et al. Nature Biotechnol. 34, 768–773; 2016).

Crucially, NgAgo very specifically cut only the target genes, says Han, unlike CRISPR–Cas9, which sometimes edits the wrong genes. And whereas CRISPR–Cas9 requires a certain genetic sequence to be near the cutting site to initiate its activity, NgAgo does not, which could broaden its potential applications, adds Han.

The initial reaction to the work in China was laudatory, including a visit to the lab by China Central Television. It was overwhelming, says Han, who is a reclusive figure. His hobbies include collecting teas and playing an ancient stringed instrument called the guqin. He doesn’t like to travel and has never left China: a trip to visit a collaborator in Hangzhou in March was the first time the 42 year old had boarded a plane. Before his paper came out, “I was completely unknown”, says Han, who spoke to Nature at his laboratory and a nearby restaurant.

Doubts about the research first surfaced at the beginning of July, when Fang Shimin, a former biochemist who has become famous for exposing fraudulent scientists, wrote on his website New Threads (xys.org) that he had heard reports of failed reproduction efforts, and alleged that Han’s paper was irreproducible. Criticism grew on various Chinese sites.

On 29 July, the controversy went inter-national when Gaetan Burgio, a geneticist at the Australian National University in Canberra, posted thorough details of his failed attempts to replicate the experiment on his blog. Normally, his posts get a few dozen hits, but this one spiked to more than 5,000.

On the same day, geneticist Lluís Montoliu, at the Spanish National Centre for Biotechnology in Madrid, e-mailed his colleagues at the International Society for Transgenic Technologies to recommend “abandoning any project involving the use of NgAgo” to “avoid wasting time, money, animals and people”. The e-mail was leaked and posted on Fang’s website.

Since then, an online survey by Pooran Dewari, a molecular biologist at the MRC Centre for Regenerative Medicine in Edinburgh, UK, has found only 9 researchers who say that NgAgo works — and 97 who say that it doesn’t.

Two researchers who initially reported success with NgAgo in an online chat group now say that they were mistaken. Debojyoti Chakraborty, a molecular biologist at the CSIR-Institute of Genomics and Integrative Biology in New Delhi, says that he repeated a specific section of Han’s paper that described using NgAgo to knock out a gene for a fluorescent protein that had been introduced into a cell. The glow was reduced, so Chakraborty assumed that NgAgo had disabled the gene. But after sequencing the DNA, he found no evidence of gene editing. He now says that the reduction in fluorescence must have had some other cause.

Jan Winter, a PhD student in genomics at the German Cancer Research Center in Heidelberg, says that he had a similar experience. “I will retry the experiment in the upcoming weeks, but so far I think it won’t work,” he says.

Han says that he has only got the system to work on cells cultured in his laboratory, and it failed in cells that he purchased. He later found the purchased cells to be contaminated with bacteria called mycoplasma, and says that others might be having the same problem. He adds that some graduate students might be working too fast and not being careful with reagents. Winter disagrees: “I do not think it is a problem of the scientists doing something wrong.”

One researcher in China, who works independently from Han’s research group and who doesn’t want his name to be entangled in the public controversy, told Nature that he had tested NgAgo in a few kinds of cell and found that it was able to induce genetic mutations at the desired sites — a finding that he verified by sequencing. He adds that the process was less efficient than CRISPR–Cas9, and requires tweaking to improve the efficiency. “But, in short, it worked,” he says.

Two more Chinese scientists, who also asked not to be named, say they have initial results showing that NgAgo works but still need to confirm with sequencing.

“It might, might work,” says Burgio, “but if so, it’s so challenging that it’s not worth pursuing. It won’t surpass CRISPR, not by a long shot.”

The failure of NgAgo “would be disappointing, but then there is work for us left to do to see whether other Argonaute systems can get it to work somehow,” says microbiologist John Van der Oost of Wageningen University in the Netherlands, a co-author of the 2014 analysis of Argonaute proteins that laid the groundwork for their use in gene editing (D. C. Swaarts et al. Nature 507, 258–261; 2014).

This week, Nature Biotechnology sent a statement to Nature’s news team, saying that “several researchers” have contacted the journal to report that they cannot reproduce the results, and that “the journal is following established process to investigate the issues”. A spokesperson declined to comment on the nature or duration of the investigation. (Nature Biotechnology is published by Nature’s publisher, Springer Nature; Nature’s news and comment team is editorially independent of the publisher’s research editorial teams.)

Hebei University says that it will ask Han to repeat the experiment so that it can be verified by an independent party within a month, according to Chinese state media.

Nature 536, 136–137 (11 August 2016) doi:10.1038/536136a