Issues Resolved 
“Cosmids" as Vectors under EK1 and EK2 Conditions 
The Recombinant DNA Advisory Committee and 
Its Host-Phage Subcommittee recently examined 
the "cosmid" for use as a cloning vector. Briefly, a 
cosmid IS a hybrid vector composed of a plasmid 
plus the sticky ends (cos sites) of lambda; it can be 
packaged in vitro as a phage but grown in £. co//as 
a plasmid. (Examples of this system and additional 
information can be obtained from a publication by 
J. Collins and B. Hohn in the Proc. Nat. Acad. Sci., 
USA, vol. 75 (1978) 4242-4246). 
For use under EK1 conditions, it will be necessary 
to follow procedures that apply to EK1 host-vector 
systems (see NIH Guidelines). Also, the cosmid 
should not contain material which can give rise to 
conjugal plasmids in the cosmid host. 
The Host-Phage Subcommittee has also recom¬ 
mended a set of criteria which will be used for the 
acceptance of a cosmid system for EK2 host- 
vector containment. In applying for approval of an 
EK2 cosmid system the investigators should sub¬ 
mit to ORDA the following information: 
1. Since the cosmid will persist in its host in the 
plasmid state, the system would have to pass 
all tests for a plasmid-host system. Since the 
cosmid requires a lambda sensitive host, x1776 
will not be useable. Therefore, whatever new 
host is selected it will have to meet necessary 
criteria regarding plasmids and plasmid 
transfer. 
2. Another problem which might be anticipated 
with the cosmid system would be the release of 
phage particles carrying the cosmid during the 
initial packaging or during some later manipula¬ 
tions. Therefore, a limit is placed on any such 
production of particles of less than lO^ total 
particles in any given experiment (i.e., in vitro 
packaging). 
3. In addition to transfer of the cloned piece by 
mobilization of the plasmid, the possibility of 
transfer by infection of the cosmid-containing 
strain with lambda and subsequent packaging 
of the cosmid and its cloned piece must be con¬ 
sidered. This might result in infection of wild- 
type lambda-sensitive strains. The probability 
of escape via this route is the product of four 
factors: 
a) Survival of the host strain outsidethe labora¬ 
tory. 
b) The frequency with which the cosmid- 
carrying strain meets lambdoid phages in the 
outside environment. This might occur either by 
infection or Hfr transfer. 
c) The frequency with which an escaping 
lambda particle containing a cosmid plus 
cloned segment meets a second lambda- 
sensitive host. 
d) The probability of packaging the cosmid and 
cloned piece after infection of the cosmid- 
containing host. This last number should be 
calculated from the following experiment: 
A model recombinant carrying a detectable 
marker of a size appropriate for packaging by 
lambda should be constructed and grown in the 
proposed EK2 host. This strain should be 
infected with lambda at an moi < 10, and the 
resulting phage lysate used to infect a lambda 
lysogen. These infected lysogens, after 
appropriate incubation for expression of the 
cloned marker, should be plated to determine 
the number of cells which have received the 
cloned marker. 
The results of this experiment can be 
expressed as: 
# of cloned markers transferred to new cells 
(= output) 
# of cloned markers in original strain (= input) 
An appropriate positive control to demon¬ 
strate the ability to recover the cloned marker 
after infection of the recipient ceils should also 
be performed. 
Data on (a) will be submitted as part of the 
necessary testing of the plasmid-host. An esti¬ 
mate of (c) has been made previously, of 
approximately <10-3. (d) should be experimen¬ 
tally determined for the cosmid proposed as 
EK2 vector, (b) has not been previously esti¬ 
mated, but data on the density of lambdoid 
phages in sewage or other environments could 
be used for such an estimate. The product of 
these four factors should be less than 1 /lO®. 
16 
