APPENDIX D— 11 
and biological barriers to the dissemina- 
tion of the potentially hazardous agents, 
(hi) The level of containment provided 
by these barriers is to match the esti- 
mated potential hazard for each of the 
different classes of recombinants. For 
projects in a given class, this level is to 
be highest at initiation and modified 
subsequently only If there is a substan- 
tiated change In the assessed risk or in 
the applied methodology, (iv) The guide- 
lines will be subjected to periodic review 
(at least annually) and modified to re- 
flect improvements in our knowledge of 
the potential biohazards and of the avail- 
able safeguards. 
In constructing these guidelines it has 
been necessary to define boundary con- 
ditions for the different levels of phj’sical 
and biological containment and for the 
classes of experiments to which they ap- 
ply. We recognize that these definitions 
do not take Into account existing and 
anticipated special procedures and Infor- 
mation that will allow particular experi- 
ments to be carried out under different 
conditions than indicated here without 
sacrifice of safety. Indeed, we urge that 
individual investigators devise simple 
and more effective containment proce- 
dures and that study sections give con- 
sideration to such procedures which may 
allow change in the conta,inment levels 
recommended here. 
It is recommended that all publications 
dealing with recombinant DNA work in- 
clude a description of the physical and 
biological containment procedures prac- 
ticed, to aid and forev/arn others. who 
might consider repeating the work. 
II. CONTAINMENT 
Effective biological safety programs 
have been operative in a variety of labo- 
ratories for many years. Considerable in- 
formation therefore already exists for the 
design of physical containment faxllitles 
and the selection of laboratory proce- 
dures applicable to organisms carrying 
recombinant DNAs (4-17). The existing 
programs rely upon mechanisms that, for 
convenience, can be divided Into two 
categories: (1) a set of standard prac- 
tices that are generally used in micro- 
biological laboratories, and (il) special 
procedures, equipment, and laboratory 
installations that provide physical bar- 
riers which are applied In varying degrees 
according to the estimated biohazard. 
Experiments on recombinant DNAs by 
their very nature lend themselves to a 
third containment mechanism — namely, 
the application of highly specific biologi- 
cal banders. In fact, natural barriers do 
exist which either limit the infectivity of 
a vector or vehicle (plasmid, bacterio- 
phage or vlnis) to specific hosts, or Its 
dissemination and survival In the envi- 
ronment. The vectors that provide the 
meaas for replication of the recombi- 
nant DNAs and/or the host cells in which 
they replicate can be genetically designed 
to decrease by many orders of magni- 
tude the probability of dissemination of 
recombinant DNAs outside the labora- 
tory. 
As these three means of containment 
are complementary, different levels of 
containment appropriate for experiments 
with different recombinants can be es- 
tablished by applying different combina- 
tions of the physical and biological bar- 
riers to a constant use of the standard 
practices. We consider these categories of 
containment separately here in order 
that such combinations can be conveni- 
ently expressed in the guidelines for re- 
search on the different kinds of recom- 
binant DNA (Section III) . 
A. Standard practices and training. 
Tlie first principle of containment is a 
strict adherence to good microbiological 
practices (4-13). Consequently, all per- 
sonnel directly or indirectly involved in 
experiments on recombinant DNAs must 
receive adequate instruction. This should 
Include at least training in aspectic tech- 
niques and instruction in the biology of 
the organisms used in the experiments 
so that the potential biohazards can be 
understood and appreciated. 
Any research group working with 
agents with a known or potential bio- 
hazard should have an emergency plan 
which describes the procedures to be 
followed if an accident contaminates per- 
sonnel or environment. The principal in- 
vestigator must ensure that everyone in 
the laboratory is familiar with both the 
potential hazards of the work and the 
emergency plan. If a research group Is 
working with a known pathogen for 
which an effective vaccine is available, all 
workers should be immunized. Serologi- 
cal monitoring, where appropriate, 
should be provided. 
B. Physical containment levels. A va- 
riety of combinations (levels) of special 
practices, equipment, and laboratory in- 
stallations that provide additional physi- 
cal barriers can be formed. For example, 
31 combinations are listed in “Labora- 
tory Safety at the Center for Disease 
Control” (4) ; four levels are associated 
with the “Classification of Etiologic 
Agents on the Basis of Hazard” (5) , four 
levels were recommended in the “Sum- 
mary Statement of the Asilomar Con- 
ference on Recombinant DNA Molecules” 
(3) ; and the National Cancer Institute 
uses three levels for research on onco- 
genic viruses (6). We emphasize that 
these are an aid to, and not a substitute 
for, good technique. Personnel must be 
competent In the effective use of all 
equipment needed for the required con- 
tainment level as described below. We 
define only four levels of physical con- 
tainment here, both because the accuracy 
with which one can presently assess the 
biohazards that may result from recom- 
binant DNAs does not warrant a more 
detailed classification, and because addi- 
tional flexibility can be obtained by com- 
bination of the physical with the biologi- 
cal barriers. Though different In detail, 
these four levels (PKP2<P3<P4) ap- 
proximate those given for human etio- 
logic agents by the Center for Disease 
Control (i.e., classes 1 through 4; ref. 5), 
In the Asilomar summary statement (I.e., 
minimal, low, moderate, and high; ref. 
3) , and by the National Cancer Institute 
for oncogenic viruses (I.e., low, moderate, 
and high! ref. 6), as Is Indicated by the 
P-number or adjective In the following 
headings. It should be emphasized that 
the descriptions and assignments of 
physical containment detailed below are 
based on existing approaches to contain- 
ment of hazardous organisms. 
We anticipate, and indeed already 
know of, procedures (14) which enhance 
physical containment capability in novel 
ways. For example, miniaturization of 
screening, handling, and analytical pro- 
cedures provides substantial containment 
of a given host-vector system. Thus, such 
procedures should reduce the need for 
the standard types of physical contain- 
ment, and such innovations will be con- 
sidered by the Recombinant DNA Mole- 
cule Program Advisory Committee. 
The special practices, equipment and 
facility installations indicated for each 
level of physical containment are re- 
quired for the safety of laboratory work- 
ers, other persons, and for the protection 
of the environment. Optional items have 
been excluded: only those items deemed 
absolutely necessary for safety are pre- 
sented. Thus, the listed requirements 
present basic safety criteria for each 
level of physical containment. Other 
microbiological practices and laboratory 
techniques which promote safety are to 
be encouraged. Additional information 
giving further guidance on physical con- 
tainment is provided in a supplement to 
the guidelines (Appendix D). 
PI Level (Minimal) . A laboratory 
suitable for experiments involving re- 
combinant DNA molecules requiring 
physical containment at the PI level is 
a laboratory that possesses no special 
engineering design features. It is a labo- 
ratory commonly used for microorga- 
nisms of no or minimal biohazard under 
ordinary conditions of handling. Work in 
this laboratory is generally conducted on 
open bench tops. Special containment 
equipment is neither required nor gen- 
erally available in this laboratory. The 
laboratory is not separated from the gen- 
eral traffic patterns of the building. Pub- 
lic access is permitted. 
The control of biohazards at the PI 
level is provided by standard microbio- 
logical practices of which the following 
are examples: (i) Laboratory doors 
should be kept closed while experiments 
are in progress, (ii) Work surfaces should 
be decontaminated daily and following 
spills of recombinant DNA materials, 
(iii) Liquid wastes containing recom- 
binant DNA materials should be decon- 
taminated before disposal, (iv) Solid 
wastes contaminated with recombinant 
DNA materials should be decontami- 
nated or packaged in a durable leak- 
proof container before removal from the 
laboratory, (v) Although pipetting by 
mouth is permitted. It is preferable that 
mechanical pipetting devices be usecL 
When pipetting by mouth, cotton- 
plugged pipettes shall be employed, 
(vl) Eating, drinking, smoking, and stor- 
age of food in the working area should 
be discouraged, (vil) Facilities to wash 
hands should be available, (vlll) An in- 
sect and rodent control program should 
be provided, (lx) The use of laboratory 
gowns, coats, or uniforms is discretionary 
with the laboratory supervisor. 
