73 
Martin, until recently the chairman of the UCSF biosafety committee. "What 
we do is to try to raise the consciousness of the individuals involved to make them 
respect the guidelines.” Martin spent 2 days investigating the episode and con- 
cluded that a breakdown in communications was responsible. Precise details re- 
main obscure because the minutes of the biosafety committee record only Mar- 
tin's conclusion, and a lengthier account prepared for the minutes by the present 
chairman, James Cleaver, is far from complete. 
The episode revolves about the use of the vectors — virus-like entities known as 
plasmids — which are used to carry genes of experimental interest into bacterial 
hosts. Two particularly useful plasmid vectors, known as pBR322 and pMB9, 
had been developed by Boyer and others. But before either could be used, it had 
to go through a two-stage administrative process, the first of being approved 
by the NIH recombinant DNA committee and the second of being certified by the 
NIH director. Vectors, in other words, are off limits to researchers until certified. 
Plasmid pBR322 was tentatively approved by the NIH committee on 15 Jan- 
uary, finally approved on 23 June, and certified for use on 7 July. Plasmid pMB9 
was certified on 18 April. 
According to both Martin and two members of the insulin team, the episode 
of the earlier experiment was as follows. Early this year, sometime in February, 
an attempt was made to produce copies of the rat insulin gene with pBR322 as 
the vector. The gene was linked to pBR322 and the plasmids inserted into NIH- 
certified bacteria designated EK2 hosts. Some of the bacteria were successfully 
colonized by the plasmids. To verify that clones of colonized bacteria contained 
the rat insulin gene it would have been necessary to carry the experiment to com- 
pletion by extracting and analyzing the DNA sequence of the genes. This step 
was not performed because at that moment, on or around 1 March, the team say 
they learned that the plasmid had not been certified for use. They decided to de- 
stroy all their clones. Further attempts were made to clone the gene with an 
already certified vector, known as pCRl, but without success. 
When pMB9 was certified, team members say, they had all their materials 
ready to go, and repeated the experiment from scratch in the new plasmid. Martin 
says that he inspected the team’s records and has “complete confidence” in their 
statement that the entire experiment was done after the approval of pMB9. 
“When the manuscript appeared so soon, people said ‘How in hell can you do 
it?’ But you can do it — the experiment was very straightforward,” Martin said 
from Great Milton near Oxford, England, where he is at present on sabbatical. 
“I think they realized they would be questioned and that if there were any 
shenanigans they would be really jeopardizing a hell of a lot, not only their own 
careers but the whole advance of science through recombinant DNA technology.” 
In the categories of the NIH rules, the experiment with pBR322 was assigned to 
the highest available containment level short of going off campus to a specially 
secure laboratory. But it is clear that the experiment posed no issue of public 
health since it was performed in the required type of laboratory — a “P3” facil- 
ity — and with a vector which has now been certified as safe. 
The experiment also took place at a time when the NIH rules were still new 
and local procedures for implementing them were still in a formative stage. “This 
was one of the inevitable things that happened as we tried to evolve a new 
system,” comments biosafety committee chairman Cleaver. 
Not wholly plain from the Martin-Cleaver accounts of the episode is how the 
insulin team came to believe it was all right to go ahead with pBR322 prior to 
certification. Howard Goodman, chief of the laboratory which did the cloning and 
sequencing part of the experiment, is out of the country, as he has been for most of 
the past year. One of his postdoctoral colleagues says there was great confusion 
at tjhe time about the status of the plasmid but the reason for the confusion is 
“sort of cloudy now.” 
According to Martin, it was clear that everyone knew that pBR322 had been 
approved but not certified, but the NIH, Martin says, was advising researchers 
that certification was imminent and that they should go ahead. According to the 
minutes of the 20 May meeting of the UCSF biosafety committee, Martin reported 
that the-researchers“had been verbally informed that the certification of an 
approved EK2 vector was imminent and to proceed with its use.” 
This version is strenuously denied by William Gartland, director of the NIH 
Office of Recombinant DNA Activities, and by his only assistant, Daphne Kamely. 
Both say that they would never have advised use of any vector before certification. 
Gartland notes that the team “must have got the vectors from Boyer, who 
certainly knew they were not certified” because Boyer had complained repeatedly 
[Appendix B — 122] 
