New York Medical College 
Department of anatomy 
Basic science Building 
Valhalla. New York 10585 
October 23, 1979 
Dr. Donald S. Fredrickson 
Director 
National Institutes of Health 
Bethesda, MD 20014 
Dear Dr. Fredrickson: 
I would like to point out several aspects of the recently reported Rowe-Martin 
tumorigenicity risk-assessment experiments (Israel, et al, Science 205, 1140 
(1979)) that I believe have not received adequate discussion either at the 
Recombinant DNA Advisory Committee or in the scientific community at large: 
(1) Restriction enzyme-generatea ^agments of the polyoma genome retain the 
tumorigenicity exhibited by polyoma DNA even after incorporation into plasmids 
or lambda particles. Therefore, working with subgenomic fragments does not 
represent am increased level of containment as may have been thought previously. 
Indeed, a recent paper from Rowe’s and Martin's laboratories (Israel, et al, 
PNAS , 76, 3713 (1979)) shows that preparations of polyoma DNA cam actually have 
their tumorigenicity increased to 100% (vs. 19% in controls) by treatment with 
certain restriction enzymes. Therefore, working with subqenomic fragments 
actually represents a new level of hazard. 
(2) Since polyoma DNA amd subgenomic fragments thereof are tumorigenic in the 
absence of residual infectivity, the argument that working with the DNA is no 
more hazardous than working with the whole virus is no longer tenable. Specifically, 
the generation of infective particles and subsequent spread through the tissue of 
an exposed host is not necessary for a tumor to be formed; action on a single cell 
of a host and subsequent replication of that transformed cell is all that is 
necessary. 
(3) A corollary of (2) is that the capsid of an intact virus may actually 
represent a form of containment which is breached in splicing experiments. There 
is good reason to believe that the host range for tumorigenicity may be extended 
by the use of subgenomic fragments, since unlike the capsid, DNA probably does not 
observe cell surface differences. 
(4) The insertion of plasmids or lambda particles containing polyoma DNA 
fragments into bacteria represents a new route into the ecosystem and a new 
reservoir for variation of a material which is manifestly an agent of animal 
carcinogenesis . 
(5) The fact that the splice products are less effective than the whole 
virus in promoting tumors in polyoma's normal host is irrelevant when considered 
in the context of the creation of qualitatively new routes of dissemination of 
this tumorigenic agent. 
[ 359 ] 
