74 
to r tlj6 NIH of the delay, in certification. According to Boyer, the insulin team 
“kept oil' asking" if the plashiid had been certified and he told them it had not. 
BdjmrsthtSs that he n^vet encouraged anyone to go ahead prior “tb certification. 
Thus the sbiirce or encouragement for the team to go ahead prior to certification 
remains obscure. ' 
A different account from Martin's is given by William Rutter, a member* of the 
insulin team and chairman of the UCSF Department of Biochemistry and Bio* 
physics, In its memorandum filed with the UCSF biosafety committee, the insulin 
team had said it would use as vectors pCRl and any other vectors that might in 
future be approved by the NIH recombinant DNA committee. When the NIH 
committee approved pBR322 on 15 January. Rutter says— (the committee gave 
tentative approval on 15 January and full approval on 23 June) — he therefore 
assumed that pBR322 was Sanctioned for use, since he was not then aware of the 
NIH distinction between, approval and certification. 
. The UCSF biosafety committee did not learn until May that the pBR322 
experiment had takfen place. Researchers doing recombinant DNA experiments 
are required to file a description of the experiment for committee approval. But, 
as Rutter has said, the memorandum filed by the insulin team did not mention 
pBR322 Specifically. Researchers using the P3 laboratory at UCSF are also 
required to sign a logbook describing their experiment. Yet Martin says that when 
he inspected the logbook at the outset of his investigation, he found “nothing 
recorded." Those in charge of the facility “were not being compulsive enough in 
sfeeing that people were filling in the logbook”^ — a situation which has now been 
corrected* Martin adds. 
Cleaver, Martin’s successor as biosafety committee chairman, told Science that 
two entries from the insulin team are recorded in the logbook between 1 February, 
when the logbook was instituted (the P3 lab officially opened on 9 November 
1976), and the end of April. An entry on 1 February notes in the column headed 
“vector” that pCRl will be Used, and the second entry on 23 April gives pMB9 
as the vector. The pBR322 experiment, according to Rutter, took place after 
1 February. The vector was not mentioned, he says, because the 1 February 
entry referred to the general experiment already described in the memorandum 
filed with the biosafety committee, in which the team had said it would use pCRl 
and other improved vectors, and the logbook has to be signed only for each experi- 
ment, not tor each use of the laboratory. “The signing in of the logbook meant 
the insulin cloning experiments in general. The vectors were not designated spe- 
cifically, and that was just human error,” says Rutter. Rutter’s laboratory r o 
vided the insulin gene for the experiment ; it was members of Goodman’s 
laboratory — Goodman was away until mid-April — who performed the cloning 
experiments. 
How should a local biosafety committee respond to an incident of this sort? 
“I would have expected that the biohazards Committee to have investigated the 
whole thing," says Gartland. Martin did conduct an investigation, and he took 
action both with the NIH and the insulin team. ‘T felt comfortable we had 
resolved the question and eliminated the possibility of it happening again,” he 
say|. But in fact, written documents of the committee record criticism only of the 
The pBR322 experiment raises no question of hazafd but it does raise the 
possibility that the insulin team might have gained an unfair advantage over 
other researchers who had abided by the NIH rules. Another team at Harvard is 
also working on the same problem. Members Of the UCSF team say that they 
gained no Information from the pBR322 experiment which was helpful to the 
later experiment with pMB9. As it happens, the Harvard team was not neck-and- 
neck with UCSF because it has not even now published any results. 
As far as is known the pBR322 experiment is the only occasion on which the 
NIH rules governing recombinant DNA research have been broken. The re- 
searchers say that the breach was the result of innocent error, a statement not 
refuted by the available evidence. The experiment presented no hazard to public 
health nor, in the event, was any unfair advantage gained over competitors. As 
for the UCSF biosafety committee, its response included action to ensure against 
repetition of the incident, although not a full public account. The committee’s 
discussion of the experiment, as reflected in the minutes of its 20 May meeting, is 
confuted to an attempt — unsupported by available evidence — to ascribe the error 
to confusion generated by NIH. But both the experiment and the biosafety 
committee's response to it occurred in circumstances to which researchers Were 
then still adapting, and for which there were few, if any, precedents. 
[Appendix B — 123] 
