韧性最好的天然纤维:原文+全文翻译：日本对孟山都的转基因抗除草剂大豆安全性评估的检查结果报告

Inspection of the Safety Assessment of Genetically Modified, the Roundup Tolerant Soybean:

Monsanto’s Dangerous Logic as seen in the Application Document submitted to Japan.

by: Masaharu Kawata

(Assistant Professor, School of Science, Nagoya University, Japan)

  Mandatory labeling of genetically modified food became effective this April in Japan.  Japanese consumers can then be able to choose non-genetically modified food by the label in any shop and store.  But,there would be consumers who dare not to choose it for their confidence in Health and Welfare Ministry's safety approval or who purchase processed food containing genetically modified ingredient without knowing it.  Safety assessment of genetically modified crop whose unknown risk with artificially modified gene is raised cannot be too cautious.  This is the report of inspection carried out on Safety Assessment Application Documents submitted by JAPAN- MONSANT for their herbicide torelant soybean that was approved as food by the Japanese Ministry of Health and Welfare and as animal feed by the Ministry of Agriculture in 1996.

(1) Information disclosure is but nominal

  Application documents to the Ministry of Health and Welfare is kept by Food Safety Association, one of the extra-governmental organization, and is made available for inspection in their Tokyo and Osaka offices.  However, inspection is only allowed 10am to 12am and 13pm to 16pm 3 days a week, and no copying or photographing is permitted.  Transcription by hand is the only method we could take for accurate scrutiny as was in Edo-Samurai period in old Japanese times.  The application submitted by Monsanto for "Roundup Ready soybean" consists of 10 volumes, which pile up to 1 meter high.  Moreover, the third section on are all in English.  It was impossible for us to transcribe all of them in given time.  We were all told 40 people in 10 days and managed more than 500 pages focusing on important points for best possible effectiveness.  When information disclosure law enacted this April, we must watch closely whether it would make things easier or even harder with excuse of company secrets and keep the decision making of safety assessment to a few bureaucrats and their academics.  Unless safety assessment is made open to public, concerns over genetic engineering will be even worse. 

(2) What is herbicide resistant soybean by Monsanto?

  In growing soybean, well-planned weed and pest control is important to get desired harvest.  Low input cultivation becomes possible if soybean itself has herbicide resistance and dust cropping is done without complexity.  Monsanto has endeavored, in vain, till 1990s to achieve this goal by creating soybean mutant, which is resistant to their best selling organic-phosphate herbicide Roundup in which the glyphosate is its active ingredient.  The resistant strains created had seriously damaged enzymatic activity of EPSPS(5-enol-pyruvylshikimate-3-phosphate synthase : one of the enzymes work to synthesize aromatic amino acid, Tyrosine, Phenylalanine, Tryptphan) and caused growth defect of soybean itself.  The genetic engineering technology was becoming popular then, and was naturally selected tried to introduce gene from different organism to soybean.  Purported herbicide resistant bacterium was found in the sewage water of the glyphosate factory of Monsanto in USA.  This Agrobacterium tumefaciens named as CP4 strain is a kind of soil bacterium, which could synthesize aromatic amino acid in the presence of glyphosate.  The amino acids sequence of the enzyme is largely different from that of any plants, and is called class II EPSPS (refer to as CP4EPSPS hereafter).   

  The bacterial gene generally can not work in plant cell by just inserting it to the genome, because the genetic switch called promoter of prokaryotes and that of eukaryote is different.  Then a powerful promoter from "Cauliflower Mosaic virus" called 35S promoter was connected to the target gene.  Next gene engineering was to connect a small protein called "signal peptide" which carries the CP4EPSPS protein to where the enzyme is supposed to function, in this case chloroplast.  This gene of signal peptide was taken from flower petunia.  A part of plant cancer virus gene called NOS, which make a signal to stop gene read through is also required.  Thus created "Roundup tolerant soybean gene" is a completely artificial gene that never exists in natural kingdom (figure 1) which would be never existed in natural evolution.

  In addition to these modifications of the genetic construct, Monsanto artificially had to change genetic codons for efficient translation of the CP4ESPS gene in soybean plant.   The 239 nucleotides out of total of 1,365 (17.51%) were manually converted to other bases (though mostly in the third letter) in order for the protein synthetic machinery of soybean cell to decipher the bacterial gene across the species barrier.  Thus, the Roundup Tolerant soybean came to possess a gene unlike either the prokaryotic gene or the eukaryotic gene.   It is with reason that gene modified plants are called "the Frankenstein plants" in Europe.  Focal point of safety assessment is whether such soybean with artificially modified gene is the same as the conventional one. 

(3) A mystery of "The samples used for analysis and animal dietary test were cultivated without herbicide application".

  The Roundup Ready soybean marketed is usually applied with the herbicide Roundup.  But surprisingly enough, our inspection revealed that both the gene modified soybean 40-3-2 strain and conventional strain A5403 were NOT sprayed with Roundup herbicide in their cultivation.  Monsanto produced only small amount sample with Roundup on the side to test residual glyphosate in the harvested forage, hay and seed.  All the soybean of a few thousand kilograms used in safety experiments was harvested not sprayed.  The reason is not stated in the documents. 

  The data obtained with such samples may be therefore not valid to guarantee safety of soybean that human and animals take in the real life, not just because of the residue glyphosate is a toxin to kill plants by inhibiting plant enzyme EPSPS.  Effects on other metabolic pathway must be taken into account particularly when such artificial genes are inserted.  For consumers, the test results using different sample than marketed soybean may be meaningless.

(4) Incomplete analysis of introduced protein CP4EPSPS.

  It is expected that CP4EPSPS protein expressed in the bio-engineered soybean have the same amino acid sequence as the soil bacterium from which the gene was originated.  This can only be verified when soybean produced protein is isolated and the amino acid sequence is determined, because exchanging genes between bacteria and higher organism can sometimes result in partial amino acid change and/or post-translational modification after expression.  Before inspection we presumed that amino acid sequence of soybean CP4EPSPS was determined.  However, to my surprise, it was not.

  What Monsanto has determined was only 15 amino acids from N-terminal of the protein, which was expressed in E.coli.  The rest of the sequence was presumed one from the nucleotide sequence of the bacterial DNA.  Then only 3.3% of expected total of 455 amino acids was decided, and the protein is not of soybean!  ELISA test described in the documents is the only method to verify antigenic equivalence of proteins.  But antigenic similarity itself does not prove the amino acid sequences are the same.  The true face of CP4EPSPS protein in the soybean that we are taking is still unknown.

(5) The E.coli expressed protein is used for acute toxicity test too on rat.

 CP4EPSPS protein used for acute toxicity test on rat is also came from that produced by E.coli harboring CP4EPSPS plasmid.  Monsanto excuses in the application document that to obtain large amount of CP4EPSPS protein from soybean is difficult.  This is unacceptable because there is a possibility that the inserted gene work differently in soybean than was in the original bacterium, and therefore the expression product may be different from that of soybean.  Moreover, according to the application document, 0.238mg of CP4ESPS protein is detected in one gram of genetically modified 40-3-2 soybean, which is enough concentration to extract without problem.  This again is the typical "All for the conclusion" approach by Monsanto.  This kind of problem could be resolved if all CP4ESPS amino acid sequence in soybean had been analyzed and confirmed equal as the bacterium.  The experiment looks like conducted on the presumption that the other soybean proteins are the same as the non-GM soybean as long as the CP4EPSPS is not toxic.  If so, this is too easy and one-sided approach.  The core of this problem is whether the soybean gene gets affected from insertion of foreign gene or not.  The series of experiments described is incoherent on the fundamentals. 

(6) Insufficient feeding experiment and intentional neglect of “wrong “ data.

  Animal feeding test is important for safety assessment.  Then Monsanto conducted the experiments for animals as rat, cow, chicken, catfish and quail.  However, the scale of experiment is less than adequate.

For example, in rat experiments, raw and toasted soybean both genetically modified and non-modified were fed to mere 10 rats each group and feeding period is only limited to 28 days.  Toxicity across generation or chronic toxicity may not be detectable by these limited experiment size and duration.

  Under these insufficient experiments however, the data for body and organ weight of lever, kidney and testicles show obvious difference in the male rats between both groups, wild A5403 and bio-engineered 40-3-2 soybean.  

  Raw soybean fed group showed no difference. But toasted soybean 40-3-2 fed male group weighed 6.7% less body weight than A5403 fed group and 13% less than commercial feed mix fed group at the end of test periods, 28 days.  Though this difference is described as statistically significant in the data sheet, the conclusion ignores these results and states that "no statistical significance is observed."  

  The experiments are far from satisfactory in its sample size and the statistic method used. Our group transcribed all raw data and redid statistical analysis using Turkey multiple method. The result again showed the apparent growth obstacle for the body and kidney weight in male rats group fed with toasted 40-3-2 soybean.  I wondered why there is no such difference in female rats group.  The answer to this question seemed to be the amount of the feed intake where male took 25-30g/day, female rats took only 18-20g (approx. 70% of male)/day.  It is highly possible that female rats also showed significant growth difference if experiment is conducted in much larger scale and with longer feeding period.

(7) Misguided interpretation and disregard of data in chemical analysis.

  Chemical analysis of the components from both normal and genetically modified is important to certify so-called substantial equivalence.  

  We found a highly intended misinterpretation ignoring obvious data difference between A5403 and 40-3-2 hybrid in the documents.  Analysis of raw soybeans showed no difference between gene modified 30-4-2 and non-modified A5403 soybean.  Difference is observed in toasted soybeans.  Besides such main components like water content, protein, fat, fiber and ash, the analysis detected trypsin-inhibitor, lectin and urease which are called harmful physiologically active substance as feed.  Urease is used an indicator of protein denaturation by heat treatment.  

  Obvious difference appeared when after toasted with actual feed processing condition (108℃, 30min).  The concentration of total protein and potassium were not changed, but concentration of trypsin-inhibitor, urease, and lectin have significantly higher in the toasted 30-4-2, the glyphosate-tolerant bean compared to that of A5403 normal bean.  These physiologically active substances remained active even after heat treatment in the genetically modified soybean, though those of herbicide sensitive normal bean were easily denatured and inactivated.  The high activity of these elements does not usually satisfy the feed standard.

  Monsanto took this result as "only the modified soybeans are toasted insufficiently in the experiment", and returned and asked re-treatment of the sample to Texas A & M who processed the beans.  Monsanto ordered the condition of re-toast as 220 ℃ for 25min, which is considerably higher than normal processing of 100℃,10 minutes.  However re-toasting further widened the difference in the activity between the two strains.  Another hybrid 61-67-1,which is another genetically modified soybean inserted with bacterial CP4EPSPS, showed highly resistant property to the heat.  

  Scientist would usually conclude in such case that there is substantial difference between the two.  But Monsanto dared to challenge this common sense, and concluded again the second toasting is still not enough.  In the end, they toasted twice further and got the result they wanted, all proteins were denatured and inactivated. With this result, they concluded that genetically modified and non-modified soybeans have equivalent properties.

  No protein can withstand repeated heat treatment and stay active.  This is a common knowledge of protein chemistry.  The argument at normal feed processing condition is required and no more, no less.   Monsanto based their argument on their presumption " they can't be different" and their need "there shouldn't be difference".  Their translation of the experiment is "Safe is the conclusion" attitude and not at all scientific.  The English data volume did not show analysis data of third and fourth heat treatment, but the Japanese summary volume, as if there were data, has a graph showing after loss of activity and described that "data from insufficient heat treatment is not adopted" and "No substantial difference observed."  If you review only Japanese summary volume and not look into English data volume, you would be ushered to the conclusion of "Safe."

  However we could found in the first and the second analyses data of toasted soybean a fact indicative of regular heat treatment. Granulated soybean, when heated, lose weight as water and other volatile components evaporate, and as the result, relative concentration of non-volatile substance such as total protein and ash increases.  The data shows clearly that the gene modified 40-3-6 and 61-67-1 and non-modified A5403 gone through same level of heat treatment.  The decrease of water content also certify this facts.

(8) Residual herbicide in crop increase, then the safety standard should be changed high level, Monsanto’s conclusion.

  By adopting the Roundup tolerant soybean, the herbicide concentration will rise in the soybean plant and seed, because the herbicide is directly sprayed on the plant by postemergence application before harvest.   The Monsanto studied in detail how will be the results by changing factors like spraying times, concentration of the active ingredient glyphosate, duration of harvest after spraying, and cultivation places.  The data show clearly that the concentration of glyphosate and AMPA (a degraded substance of glyphosate) in forage and hay increase greatly by postemergence application of the herbicide compared to that of conventional preemergence application, although the residual concentration in the plant differed from place to place.  The largest value of the combined glyphosate and AMPA was 40.187 ppm in forage which is higher than the US safety standard of 15 ppm in forage and hay in 1994 when FDA and USDA accepted the application documents.  The maximum combined concentration of glyphosate and AMPA in soybean seed was 13.178 ppm, which is less than 20 ppm of the US standard at that time.  The residual concentration increased in accordance of the application increased from twice to three times.           Then cultivating Roundup ready soybean may sometimes violates the US safety standard.  We found a surprising description in the document to dissolve the problem.

  In final conclusion, Monsanto say that “the maximum combined glyphosate and AMPA residue level of approximately 40 ppm in soybean forage resulting from these new uses exeeds the currently established tolerance of 15 ppm.  Therefore, an increase in the combined glyphosate and AMPA tolerance for residues in soybean forage will be requested.”  They know very well that adoption of herbicide tolerance crop needs higher safety standards.  In effect, the US tolerance standard of combined glyphosate and AMPA in soybean forage was changed to 100 ppm after they approved the genetically engineered soybean.

  As to Japanese government, they revised the safety standard of combined glyphosate and AMPA in soybean seed to 20ppm in April 2000, from old standard of 6 ppm according to the request of US government.  Of course Japan become could import soybean from USA without worrying about of violation of the law by this decision.

 Thus, Monsanto, in their rush to verify safety, patch worked the results of experiments and analyses that are full of voids like a puzzle and asserted safe with manipulation of the results.  They requested, if necessary, even the revision of safety standard.  We found the facts showing inadequate and incomplete safety assessment described above in the application document by Monsanto even in our limited work under difficult situation.  The process of genetic recombination and the results of other animal experiment remained not inspected yet. 

 Monsanto informed US soybean importing countries in May 2000 that they found Roundup resistant soybean has two extra fragments of the CP4EPSPS gene in the genome.  They were there since the first FDA approval in 1992, and all the GM soybeans supplied worldwide contain this gene fragments.  Monsanto asserts that these fragmented genes do not create unknown protein since they have any open reading frame or termination signal around them.  But such basic facts comes to light 8 years after the approval is a sure indication of how incomplete the genetic recombination of crop is, and how dangerous safety assessment can be to rely only on company’s information and data.  We doubt it very much if at all government experts in charge at the Japanese Ministry of Health and Welfare for safety assessment had a good sense to have concluded as safe on the bases of such incomplete application.

【译文】

Inspection of the Safety Assessment of Genetically Modified, the Roundup Tolerant Soybean:　　

Monsanto’s Dangerous Logic as seen in the Application Document submitted to 　　Japan　　.　　

--对孟山都的转基因抗除草剂大豆安全性评估的检查结果：在作为提交给日本申请的文件中，孟山都的危险逻辑显而易见　　

by: Masaharu Kawata 　　

(Assistant Professor, 　School　 of 　Science　, 　　Nagoya University　, 　Japan　　)　　

 --作者：Masaharu Kawata　　

（助理教授，理学院，名古屋大学,日本）　　

在日本，今年四月，孟山都对转基因食品标明的做法生效了。日本的消费者可以按标签在任何店铺和商店选择非转基因食品。然而，还是会存在这样的消费者，他们因对厚生省的批准的安全证书的信心方面的原因而依旧不敢选择非转基因食品，或者还存在这样的情况：采购了含转基因配方的加工过的食品却不知道。人工转基因有未知的危险，对种植的转基因作物的安全评估再谨慎也不过分。本文是对有由JAPAN- MONSANT递交的转基因抗除草剂大豆安全性评估的申请文件而得出的检查报告，在1996年，这些转基因耐除草剂大豆被日本厚生省批准作物食品和被日本农业省批准作为动物饲料。　　

（1）信息披露有名无实。　　

递交给厚生省的申请文件存放在食品安全协会，食品安全协会是属于政府部门以外的一个组织。检查活动要到该组织设在东京和大阪的办事处进行。然而，每星期只有三天允许进行检查，上午10点到12点，下午13点到16点，不许拷贝和照相。为了检查精准，我们只能采用手工抄写的方法了，就好像回到了日本江户武士时期的旧时代。这部孟山都为他的耐除草剂大豆而递交的申请由十卷组成，堆起来高一米。而且，第三章以后都是用英语写的。在规定的时间里我们根本不可能抄得全。为了到达最高效率，我们总共40个人10天聚焦重点设法完成了500多页。信息披露法律在今年四月颁布，我们密切注意信息披露是变得容易还是以公司秘密为借口，依旧决定安全评估只面向几个官僚和他们的学者，从而使得信息披露愈加困难。除非安全评估面向公众开放，否则对遗传工程会更加的担心。　　

（2）孟山都的抗除草剂大豆是什么？　　

在大豆生长过程中，妥善安排控制杂草和害虫事务对获得理想的收成是很重要的。要是大豆本身抗除草剂，给作物喷洒除草剂做起来就不复杂，那么低耕作投入就成为了可能。孟山都为此努力着，但一直没有结果，直到20世纪90年代大豆的突变异种的问世，才达到了这个目标。这种大豆的突变异种对孟山都自己卖得最好的有机碳酸盐除草剂有耐药性。这种除草剂的有效成分是草甘膦。这个被创造出来的抗性品系的酶EPSPS的活性受损严重（5-enol-pyruvylshikimate-3-phosphate 合酶：一种合成芳香族氨基酸，酪氨酸，苯基丙氨酸，色氨酸的酶）从而导致大豆本身的生长缺陷。那时很流行的遗传工程就很自然的被用来把其他不同的有机体的基因引入大豆。据说在美国的孟山都草甘膦工厂的废水里发现了耐除草剂的细菌。 这个被命名为CP4品种的根瘤土壤杆菌是一种在有草甘膦的场合能合成芳香族氨基酸的土壤细菌。这种酶的氨基酸排列的序列不同于任何植物，因此被命名为class II EPSPS (也叫做CP4EPSPS 来世)。　　

仅仅把细菌的基因插入到基因组，细菌的基因还不能起作用，因为叫作原核生物的启动子以及真核细胞的启动子的基因开关是不同的。然后，一个来自“花椰菜花叶病毒”的强大的启动子，叫作35S启动子，被连接到了目标基因。接着基因工程就把携带CP4EPSPS 蛋白质的叫作“信号肽”的小蛋白质，连接到被期望能起作用的酶上，本例中是叶绿素。信号肽的基因来自牵牛花。被称为NOS的植物肿瘤病毒基因的一部分，用来产生基因读出时的停止信号的东西也是必须的。因此，被创造出来的“抗除草剂大豆的基因”是完全彻底的人造基因，这种基因在自然界根本不存在（图一），在自然界的进化过程中也不会出现。　　

除变动基因结构外，孟山都还必须人为地改动基因的密码子，为的是让大豆植物中的CP4ESPS基因有效转化。1365种核苷酸基中的239种核苷酸基（占17.51%）被手动地改为其他基（尽管绝大多数在第三个字母上的）为了让大豆细胞的蛋白质的合成机制能够跨过物种障碍解释病毒基因。因此，抗除草剂大豆就有了一种既不像原核细胞的基因也不像真核细胞基因的基因。正由于这样，在欧洲，转基因植物被叫做“弗兰肯斯坦植物”（ 弗兰肯斯坦--毁灭创造者自己之物）。安全评估的焦点集中在人工转基因大豆是不是等同于传统意义上的大豆。  

（3）“用于分析和动物喂养试验的样本在培养时没有用除草剂”的秘密。　　

市场上交易的抗除草剂大豆通常都是应用在有除草剂的状态下的。但是太出人意料了，我们的检查发现，转基因的40-3-2品种大豆和传统的A5403品种大豆在培养时居然都没喷洒过除草剂。孟山都只用除草剂生产一小部分的样品另外作为检测收割的饲料，干草和种子中残留的草甘膦之用。几千公斤的用于安全性实验的收割下来的大豆，全部都没有喷洒过除草剂。原因在文件中没有说明。　　

由这样所谓的样本得到的数据也就不能有效的保证人类和动物在现实的生活中食用这样的大豆会是安全的，这就不仅仅是因为草甘膦的残留毒素通过抑制植物酶EPSPS足以杀死植物的问题了。当这样的人工基因被插入后，对其它代谢途径的影响必须给予特别考虑。对消费者而言，使用与市场交易的大豆不一样的大豆来做样本而得出的结果是毫无意义的。　　

（4）对引入的蛋白CP4EPSPS的分析不完备。　　

在通过生物工程而产出的大豆的CP4EPSPS蛋白预期应该和它的起源土壤细菌有相同的氨基酸序列的表达。而这只有在大豆的蛋白质是分离的并且氨基酸序列是确定的情况下才能被证实，因为细菌和更高等的生物之间的基因交换，有时会导致氨基酸在表达后出现改变和（或）翻译后修饰。在检查前我们做了这样的假设：大豆的CP4EPSPS氨基酸序列是确定的。然后，让我们吃惊的是，事实上并非如此。　　

孟山都能确定的仅仅是15种N-terminal蛋白质中的氨基酸，这些氨基酸被表达在大肠杆菌中。其余的序列据推测来自细菌DNA的核苷酸序列。然而期望的455种氨基酸中仅有3.3%是明确的，而且都不是大豆蛋白！在文件中描述的酶连锁免疫试验是唯一可以证实蛋白抗原等价的方法。但是抗原本身的相似性却证明不了它们的氨基酸序列也是完全一样的。我们食用的大豆中的CP4EPSPS蛋白的真面目依然是个谜。　　

（5）被用来表达蛋白的大肠杆菌也被用于鼠的急性中毒试验。　　

 用于鼠的急性中毒试验的CP4EPSPS蛋白同样来自于带有CP4EPSPS质粒大肠杆菌。孟山都在申请文件中的借口是要从大豆中获得大量的CP4EPSPS蛋白是困难的。这是不可接受的。因为插入到大豆中的基因和在原先杆菌中的基因有可能起不同的作用。此外，按照这部申请文件的说法，既然可以从1克的转基因40-3-2大豆检测出0.238毫克的CP4ESPS蛋白质，那么也就说明浓度足够，提取毫无问题。这是孟山都的又一个“一切为了结论”的典型做法。如果大豆中的所有CP4ESPS氨基酸序列被分析证实是和细菌的等同的，那么这个问题也算解决了。试验看起来是在这样的假设上实施的，假定只要CP4EPSPS是无毒的，那么其余的大豆蛋白和非转基因蛋白是相同的。如果是这样的话，这个方法也就太容易和太片面了。这个问题的核心是大豆基因是否会受到外部插入基因的影响。基本上，整个实验的描述都是语无伦次，不合逻辑的。　　

（6）不充分的喂养实验和对“错误数据”的蓄意疏忽。　　

动物喂养试验对安全性评估很重要。孟山都用鼠，牛，小鸡，鲶鱼和鹌鹑做实验。然而实验规模却比应有的规模要小。　　

例如，在鼠类实验中，每组仅有10只鼠，投喂生的和烤熟的转基因和非转基因大豆，喂养的时间长度也限于28天。在这样受限制的实验规模和这样受限制时间长度里，跨代的毒性和慢性的毒性是不可能检测出来的。　　

然而这不充分实验也显示出两组的雄性鼠，投喂野生A5403大豆的和投喂生物工程的40-3-2大豆的，在身体和肝脏，肾脏和睾丸的器官的重量的数据上有明显的差异。　　

投喂生大豆的组没有任何的差异。但是，在28天的实验的最后一天，投喂烤熟的40-3-2大豆的雄性组比投喂A5403的组体重轻6.7%，比用商用饲料混合喂养的组体重轻13%。尽管这些差异在数据一览表中被描述为是有统计显著性的，但结论却忽视这些数据还声称这些数据“看起来没有统计显著性”。　　

在样本数量和统计方法上这些实验根本不符合要求。我们小组抄录了所有的原始数据，用Turkey multiple 方法重做了统计分析。结果再次显示，投喂烤熟的40-3-2大豆的雄鼠组，在体重和肾脏重量方面有明显的生长障碍。我们搞不明白，为什么在雌性组之间没有这样的差异。问题的答案似乎是投料的数量，雄性的每天投喂25-30克，雌性鼠每天仅投喂18-20克（大约是雄性的70%）。很有可能，要是实验做得规模足够大喂养的时间足够长的话，雌性鼠同样也会出现明显的生长差异的。　　

（7）被误导的翻译和被忽视的化学分析数据。　　

对自然的和转基因的两者成分的化学分析对证实所谓的实质等同是重要的。　　

我们在这个文件中发现存在忽略A5403 和 40-3-2 混合种之间明显的数据差异的有意的误导。对生大豆的分析显示转基因30-4-2和非转基A5403大豆之间没有任何的差异。烤熟的大豆可以发觉差异。除了像水含量，蛋白质，脂肪，纤维和灰分这些主要的成分外，分析还检测到胰岛素抑制剂，外源凝集素和脲酶这些当做食物时被叫作对生理学活动上有害的物质。脲酶用来作为加热处理时蛋白质变性的指示剂。　　

在真实的饲料加工条件下（108℃,30分钟）烤熟后，出现了明显的差异。全蛋白的浓度和钾的浓度没有改变，但是和 A5403 自然的大豆相比，烤熟的抗草甘膦大豆30-4-2，其胰岛素抑制剂，脲酶和外源凝集素的浓度要明显高。这些生理活性物质甚至在加热处理后的转基因大豆里依然保持活性，而那些对除草剂敏感的正常大豆里这些成分却容易变性和失去活性。这些成分的高活性通常不满足饲料的标准。　　

孟山都被这个结果说成是“仅仅是因为这些转基因大豆在实验中烤得不充分”，然后退回并要求处理这些豆子的Texas A & M对样品进行重新处理。孟山都指定的重烤的条件是220℃，25分钟，大大高于正常处理条件的100℃，10分钟。然而，进一步的重烤却拉开了两个品种在活性上的差异。另一个混合种61-67-1，插入细菌CP4EPSPS的另一种转基因大豆，表现出顽强的耐热性。　　

科学家在如此的情况下通常会得出两种之间实质性差异的结论。而孟山都却敢挑战这个常识，再一次得出第二次还是没有烤充分的结论。最后，他们又烤了两次才得到了他们想要的结果，所有的蛋白质都已经变性了也失去了活性了。用这个结果，他们得出了转基因大豆和非转基因大豆具有相同的特性的结论。　　

任何蛋白都抵抗不了反复的加热处理而保持活性。这是蛋白化学过程中的基本知识。论证是需要建立在不偏不倚的正常的饲料加工条件上的。孟山都是根据“它们不可能不同”的他们自己的假设和“不存在差异”的他们自己的需要去进行论证。他们对实验的解释采用的“安全就是结论”的那种态度是根本不科学的。英文数据卷没有显示第三和第四次加热处理的分析数据，但日文的摘要卷中却好像有数据的样子，还配了图显示失去活性之后的图还说“来自不充分的加热处理得数据不会采用”和“没任何可见的实质性差异”。要是只检查日文摘要卷而没有查找英文数据卷的话，你就会被引导而得出“安全”的结论。　　

无论如何我们还是在第一次和第二次的烤大豆的分析数据中发现了常规的加热处理得事实。颗粒状的大豆，加热后，由于水分和其他其他挥发分成分的蒸发，重量会减轻，从而导致如全蛋白这样的非挥发物质的相对浓度的增加。数据清楚的表明转基因的 40-3-6 及 61-67-1和非转基因的A5403做了同样级别的加热试验。水分的减少证实了这样的事实。　　

（8）孟山都公司的结论，因为作物中除草剂残余量升高了，所以安全标准中残留水平应该变高。　　

由于除草剂是在出苗后到收割之前直接喷洒在植物上的，所以采用了抗除草剂的大豆，在大豆 植物和种子中的除草剂浓度将会增加。孟山都通过改变诸如喷洒次数，有效成分草甘膦的浓度，喷洒后收割的持续时间和培植的地点等因数，详细地研究了导致这样结果的原因。数据清楚的显示了尽管不同地点的植物的残留浓度有所不同，苗期使用除草剂和正常的出芽前使用除草剂相比，草料和干草中草甘膦和AMPA（一种退化的草甘膦物质）的浓度大大地增加了。在草料中草甘膦和AMPA结合物的最大值为40.187 ppm，高于1994年美国食品管理局和美国农业部接受申请文件时的在草料和干草中为15 ppm的美国安全标准。大豆种子中最大的草甘膦和AMPA结合物浓度是13.178 ppm，低于当时的20ppm的美国标准。按照这个申请，残留物浓度从2到3倍地变大。那么培养抗除草剂大豆有时就可能违反了美国的安全标准。在文件中我们发现了使人惊讶地使问题淡化的描述。　　

在最后的结论部分，孟山都说“最大的草甘膦和AMPA结合物的残留在大豆草料中的水平大约是40 ppm，这是由于超过当前确立的容许值为15ppm的新的使用而引起的。所以，就需要增大在大豆草料中的草甘膦和AMPA结合物的残留容许值。”他们清楚地知道采用抗除草剂作物需要更高的安残留水平全标准。事实上，在核准转基因大豆后，美国的大豆草料中含草甘膦和AMPA的容许标准已经改为100ppm。　　

至于日本政府，他们按照美国政府的要求，在2000年四月，修改了安全标准，将大豆种子中草甘膦和AMPA的容许值从原来的6ppm改20ppm。当然，根据这个决定，日本就可以从美国进口大豆而就不用担心有违法律的问题了。　　

就这样，孟山都，在匆匆忙忙的安全论证过程中，给像谜一般的漏洞百出的实验和分析的结果打补丁，再用篡改的结果给出安全断言。甚至我们在检查工作受限制的这种困难条件下，我们还是在由孟山都递交的安全性申请文件中发现了上述的明显的安全性评估不充分和不完备的事实。而基因重组的过程和其他的动物实验还没进行检查。　　

在2000年五月孟山都通知进口美国大豆的国家，宣称在抗除草剂大豆基因组中发现了两个额外的CP4EPSPS基因片段。其实，自1992年美国食品及药品管理局首次核准后，这些片段就一直存在，全世界范围内供应的转基因大豆里都有这样的基因片段。孟山都宣称这些基因片段不会产生不可知的蛋白，因为这些基因都有开放可读的架构或者有围绕这些基因的终止信号。但是，这样的基本事实在核准后的8年才大白于天下，这明确无误的表明作物的基因重组是多么的不完备，依赖于公司的信息和数据得出的安全性评估是多么的危险。我们非常怀疑具有良好判断力的负责安全性评估的日本厚生省的专家在如此不完备申请的基础上会居然会得出安全的结论。　　

孟山都的抗除草剂大豆的安全性评估必须重新评价，这是我们的结论。