SUMMARY OF THE PROPOSED GUIDELINES FOR RESEARCH.
ON RECOMBINANT DNA MOLECULES

i. INTRODUCTION pp/-3

1. Robert Sinsheimer: In the advent of a new biological
research, the research process is changing from one
of analysis to one of synthesis. The complications are
great. The process is an irreversible one. Vectors will
escape and there will be no control. We will not be able
to stop production as we could with DDT.

2. Daniel Callahan: The underlying question is how to
establish a proper policy bias. There is no moral
obligation to do this research although it certainly is
commendable but there is a moral obligation to do no harm;
and given the uncertainty of the hazards, one should incline
to caution. Does the burden of proof lie with the scientist
to show no harm or with the public to show that there is
harm? The benefit of the doubt should be to the “worriers"
and not the researchers. This should be the policy bias,
and the guidelines should be reviewed accordingly. With
this in mind, the guidelines seem to be reasonable and
prudent in the conduct of this research.

A. Definition of Experiments Included Within the Guidelines

The proposed guidelines concern experiments in which different segments
of DNA chains are joined by biochemical techniques and the resulting
recombined DNA molecules are then inserted into living cells in which
the molecules can be reproduced. The recipient cells are called the
hosts." In general, one segment of the recombined DNA molecule will
’ be derived from an extra-chromosomal genetic element common to the

“host" species. That segment of DNA is referred to as the "vector."
The other segment of the recombined DNA is referred to as "foreign"
DNA, since it is not normally chemically linked to the vector. "Host"
cells will usually be either single-celled microorganisms, such as
bacteria, or single cells of higher organisms (plants and animals)
grown under special laboratory conditions.
In nature, exchange and recombination of genetic information (that is,
DNA segments) a consequence of the sexual mode of reproduction and
generally occurs only between individual organisms of the same species.
The unique feature of recombinant DNA experiments is that they provide
a means for combining genes from diverse species. This unique feature
promises revolutionary potential both for the investigation of basic
biological processes and for approaches to important practical problems
in medicine and agriculture. The same unique feature is the basis for
concern over the potential biological hazards that might result from
such novel genetic.recombinants. ~

B. Principles Upon Which the Guidelines Are Based

The guidelines are consistent with conclusions formulated at the
International Conference on Recombinant DNA Molecules held at
Asilomar Conference Center, Pacific Grove, California, in February
1975.

l. Certain experiments are judged to present such serious potential
hazard that they should not be attempted at this time. a 7

ye

2. Other experiments can be undertaken provided that they,afford _
new kn or benefits not readily obtained by ¢onventional
methodology and provided that appropriate safeguards are incor-
porated into the experimental design. Such safeguards should
protect laboratory workers, the community, and the environment

from possible infection with potentially hazardous agents. The

proposed barriers to dissemination of the agents are twofold:
(1) physical containment of organisms containing recombinant
DNA; and (2) biological containment, that is the use of "hosts"
and "vectors" with limited ability to survive in natural
environments. The two types of barriers are complementary .
and are to be used in combination. coun 1A
3. Recognizing that present relevant knowledge is limited, the
various types of experiments caa-be-ranked with regard to

expected degree of potential hazard. The level of containment
achieved by combined physical and biological barriers should

be chosen to match the estimated potential hazard. Containment
should be high at the start and should be modified subsequently

only if there is a substantiated change in the assessed risks.

4. The guidelines should be reviewed at least annually and
modified to reflect new knowledge.
II. METHODS OF CONTAINMENT

1.

2.

David Baltimore: He asks whether the proposed guidelines
are an appropriate response to the hazard posed by
recombinant DNA molecules and whether the process that

led to the formulation of guidelines was a responsible

one. He notes that recombination of DNA molecules has gone
on for aeons; he urges the focus must be on those DNAs that
are a potential hazard, not focus on the mere joining of
DNA molecules. The hazard is defined as the potentiality
of artificially joined DNA molecules escaping from the
laboratory and disseminating in the general population.

He believes the guidelines are appropriate to meet that
hazard. He urges that the guidelines be promulgated as
quickly as possible. He cites the scientific interest

and the medical justification for these experiments to

go on.

Robert Sinsheimer: Too little attention has been given

to hazards that might result from new combinations of

genes that shall be created in this research. Research
increases the opportunities for potentially dangerous

gene combinations by many, many orders of magnitude that
might occur sporadically in nature. In this regard
biological containment may not be adequate. The key element
of concern is the irreversibility of the process that will
become magnified as more investigators in laboratories
pursue research in this area. Rather than going the global
approach, the guidelines might be better served by approaching

each desired ‘ ; .
possible. objective in as specific and safe a manner as
A. and B. Physical Containment PP ae JQ

1. Peter Barton Hutt: The concept of physical containment
is both too imprecise and too subject to the vagaries
of human fallibilities. Thus physical containment can
only be applied when risk is non-existent. When there
is any potential risk, protection must rest with biological
containment. Consideration of the Pl through P3 levels
should be undertaken. Perhaps Pl and P2 should be combined
into a single level of physical containment. Another
possibility is to upgrade some of the experiments to a
P3 or P4 level that are currently under Pl and P2 levels.
The intended requirements for Pl through P3 levels should
be spelled out. Perhaps Dr. Barkley who raised several
points in his presentation could review this section for
additional comments. Description of acceptable practices
should be as clear as possible and might well include
stated minimum levels of scientific training and experience
necessary to conduct the experiments.

1. Philip Handler: What does physical containment mean?
Are Pl and P2 levels really physical containment? Should
we not consider P3 as the minimum level incorporating
‘Pl and P2? This research must go forward but the pace
must be set by Dr. Fredrickson.

Four levels are defined and designated Pl, P2, P3, and P4 in order of
increasing containment. Each higher level assumes the practices of
the lower ones. The specifications reflect existing technology as
developed for work with known pathogenic organisms. It is anticipated
that technical developments will lead to novel alternate procedures
for achieving physical containment.

© Pl (minimal): Strict adherence to the standard microbiological
practices widely used in -research and clinical laboratories
for work with moderately pathogenic organisms. Appropriate
training of all personnel is required.

@ P2 (low): Access to the laboratory is limited to authorized
and informed personnel when potentially hazardous organisms
are being used and until after appropriate decontamination.
Pipetting by mouth is prohibited, and specific precautions
are required for procedures expected to release aerosols
containing potentially hazardous material.
e P3 (moderate): Laboratories.are separated from areas used
for less hazardous experiments and access is limited to those
who work therein. Laboratories should have air pressure
lower than that of the surrounding areas to limit the flow
of airborne organisms into surrounding areas. Exhaust air
from laboratories should be appropriately discharged or
decontaminated prior to recirculation. Biological safety
cabinets: should be used for all transfer operations and
for all procedures likely to produce aerosols. Gloves are
to be worn and all vacuum lines protected by filters.

Alternate procedures, affording at least equivalent physical
containment, are suggested for situations in which laboratory
air conditions cannot be controlled in the specified manner.

e P4 (high): Special facilities designed to contain highly
infectious and hazardous microorganisms are required. Such
facilities involve isolation by means of airlocks, negative
pressure environments, clothing changes and showers by
personnel, biological safety cabinets, and decontamination
of all air as well as all liquid and solid wastes. Only a
limited number of such facilities exist in the Dnited States.

C. Biological Containment

The nature of these barriers will depend on the particular "host" and
"vector" used in each experiment since the barrier is defined by the
limited ability of the "host" and its resident recombined DNA to survive
in natural environments. In general, the choice of "host" and "vector"
will be determined both by the nature of the experiment and by the
required containment. Therefore, experiments are grouped below
according to possible “host-vector” systems.

D. Publication

The committee strongly recommends that all publications dealing with
recombinant DNA work include a description of the physical and
biological containment procedures used.
III. EXPERIMENTAL GUIDELINES Pp 1 - (2

A.

Experiments That Should Not Be Performed at This Time

i. Any experiments in which a portion of the DNA to be joined is
derived from highly pathogenic organisms (classes 3, 4, and 5 of
the "Classification of Etiologic Agents on the Basis of Hazard,"
Center for Disease Control, USPHS, Atlanta, Georgia).

2. Experiments in which a portion of the DNA to be joined
contains genes for production of highly toxic agents.

3. Experiments in which a portion of the DNA to be joined is
derived from a plant pathogen if the resulting host may acquire
increased virulence or the ability to infect previously
unsusceptible species.

4. Widespread or uncontrolled release into the environment of
any organism containing a recombinant DNA molecule unless a
series of controlled tests leave no reasonable doubt of safety.

1. LeRoy Walters: The guidelines are least clear in defining
prohibition on experiments where antibiotic resistance
may occur, Further clarification is necessary.

5. Transfer of genes conferring drug resistance to micro-

‘organisms not known to acquire such resistance naturally when

such resistance may compromise clinical use of the drug in
medicine or agriculture.

6. Large-scale experiments with recombinant DNAs known to
result in the formation of harmful products. Exceptions

are permissible for specific experiments of direct societal
benefit, provided that they are expressly approved by the NIH
Recombinant DNA Molecule Program Advisory Committee.
B. Guidelines for Permissible Experiments

5.

6.

1.

Wallace Rowe: The risks are real. Tests are necessary

to determine the nature_gf these risks including a
disturbance in the; fegilation of the E. coli, the introduc-
tion of DNA from E.~eeli-into other cells and the dispersion
of- toxic products from the E. coli.

Roy Curtiss: He outlines possible tests for the risks as
the probability of escape, the probability of survival
in the outside environment; and the probability of the
manifestation of adverse consequences. In this regard
we are not only talking about the protection of man but
the protection of the entire biosphere.

Philip Handler: The NIH should undertake to develop a
program of study of the most important hazardous experiments
This should be done at P4 facilities at the NIH or

Fort Detrick. These experiments would provide scientific
evidence to defuse the debate on the potential hazards.
Guidelines can be modified accordingly based on that
evidence.
i ; -
1. Criteria for use of Escherichia coli, strain K-12 as “host

, 3.

yh

MM

Robert Sinsheimer: Experiments should be done with micro-
organisms that have no association with man or other
animals. These organisms should have a very restricted
range of genetic exchange:at a very limited biological
niche of viability. ,

1.

Peter Barton Hutt:

3.

Paul Berg: Because we are-so familiar with the metabolism
and genetic systems of E. coli, it provides the best
opportunity to achieve the benefits in this research.
Further, it's the most likely candidate far providing

the safetest host and vector systems. Developing an
entirely new host-vector system will not answer questions
of risk. The ominous and catastrophic scenarios raised
concerning E. coli could equally be raised with any

known organism. ,

vos. .

Philip Leder: He has constructed an EK2 host-vector
‘system and it has been recently announced in the nucleic
acid recombinant scientific memorana. Testing of this
system is currently underway. He expects to submit it
shortly to the Advisory Committee for certification.

What would be the cost in not using

i foster the
E. coli? Can incentives be created to
development of EK2 and EK3 systems. Should there be a

ai) , 2-year limit on EK1 to for

ce development of EK2 and EK3?

If EKl is to be used, there must be as detailed an
explanation as possible of that system and its use.
ef

3)

The conmon bacterium Fscherichta colt, strain K-12, is the "host"
of choice at the present time.: Well-studied extrachromosomal DNA
molecules known to reside and reproduce in this “host" are avail-
able for use as "vectors." These include plasmids and bacterio-—
phages. Extensive knowledge of Escerichia colt and its plasmids
and bacteriophages currently affords the most fruitful approach
to the design of containable "host-vector" systems.

Three containment classes for E. coli strain K-12 "host-vector"
systems are defined. The classes are named EK (for E. eolt K-12)
1, 2, and 3 in order of decreasing ability to survive in natural
environments (that is, increasing containment).

EK1: This class includes most of the currently known and available
EB. colt K-12 “host-vector" systems. Present knowledge suggests
a low probability for dissemination of plasmid or bacteriophage
"vectors" carrying recombined "foreign" DNA. This conclusion
is based in part on the following information: (1) #. eolt
K-12 is not known to colonize the normal bowel; (2) Certain
plasmids known to be useful as "vectors" have only limited
ability to be transferred to other EF. eolt strains commonly
found in animal (including human) bowels and sewerage;

(3) Bacteriophages which specifically infect FE. colt, and
which lend themselves well to use in these experiments, are
available. Some of these bacteriophages have limited ability
either to escape the “host” cell or to reinfect common (not K-12)
E. coli strains.

Biological containment criteria using E. coli K-12 host-vectors

 

EK? host vectors - These are host-vector systems that can be esti-

mated to already provide a moderate level of containment, and include

most of the presently available systems. The host is always E. coli K-l2,

and the vectors include nonconjugative plasmids [e.g., pSC101, ColEl or

derivatives thereof (17-24)] and variants of bacteriophage A (25-27).

The E. colt K-12 nonconjugative plasmid system ts taken

as an example to illustrate the approximate level of contain-

ment referred to here. The avatlable data from experiments

_ tnvolving the feeding of bacteria to hwnans and calves (28-30)

indicate that E. colt K-12 did not usually colonize the normal

bowel, and exhtbited little, if any, multiplication while
passing through the alimentary tract even after feeding high
. doses (t.G., 10° to 107? bacteria per human or calf). However,
general extrapolation of these results may not be warranted

because the implantation of bacteria into the intestinal tract

 

re "depends on a number of @GrGneLawe , such as the nature of the in-

testinal flora present in a given invidivual and the physiological

 

_state of the inoculion. Moreover, since viable E. coli K-12 can

 

be found in the feces after humans are fed 10° bacteria in

 

broth (28) or 3x 10° bacteria protected by suspension in milk

 

(29) s_ transduc tional: and conjugational transfer of the plasmid

vectors from E. eolt K-12 to resident bacteria in the fecal

 

matter before and after excretion must also be considered.

 

Since most conjugative

 

plasmids in nature are repressed for expression of donor

fertility, the frequency at which nonconjugative plasmids are

mobilized and transferred by this sequence of events in vtvo
ts difficult to estimate. However, tn calves fed on an

anttbtotic-supplemented diet, it has been estimated that

 

such triparental noneconjugative R plasmid transfer occurs

 

at frequencies of no more than 10719 to 10-12 per 24 hours

 

per calf (30). In terms of considering other means for

 

plasmid transmission in nature, it should be noted that

 

transduction does operate in vivo for Staphylococcus

 

 

aureus (32) and probably for E. coli as well. However,

 

no data are avatlable to indicate the frequencies of plasmid

transfer in vivo by either transduction or transformation.
These observations indicate the low probabilities for
posstble dissemination of such plasmid vectors by accidental
tngestton, which would probably involve only a few hundred

or thousand bacteria provided that at least the standard

 

practices (Section II-A above) are followed, particularly { i

 

the avotdance of mouth pipetting. The possibility of colont-

zatton and hence of transfer are increased, however, tf the

 

normal flora in the bowel te disrupted by, for example, anti-

 

biotite therapy (33). For this reason, persons receiving such

 

therapy should not work with DNA recombinants formed with any

E. colt K-12 host-vector system during the therapy pertod and

 

for seven days thereafter; similarly, persons who have achlorhydria

 

or who have had surgical removal of part of the stomach or bowel

should avotd such work, as should those who require large doses

 

of antacids.

The observations on the fate of E. colt K-12 in the
human alimentary tract are also relevant to the containment
of recombinant DNA formed with bacteriophage » vartants.
Bacteriophage can escape from the laboratory etther as

mature infectious phage particles or in bacterial host cells
in which the phage genome is carrted as a plasmid or prophage.
The fate of E. colt K-12 host cells carrying the phage genome

as a plasmid or prophage ts similar to that for plasmid-
containing host cells as discussed above. The survival of
the phage genome when released as infectious particles
depends on their stability in nature, their infeetivity and
on the probability of subsequent encounters with naturally
occurring h-sensitive z. colt strains. Although the pro-

bability of survival of 4 and ite infeetton of resident intestinal

 

EB. colt in animals and humans has not been measured, it ts

 

estimated to be small given the high sensitivity of to the
low pH of the stomach, the tnsuscepttbility to infection of
smooth E. colt cells (the type that normally resides in the
gut),.the tnfrequency of naturally oceurring h-sensttive

E. colt (34) and the failure to detect infective i particles
in human feces after ingestion of up to 1011 ) particles (35).
Moreover, 4 particles are very sensitive to desstcatton. |

Establishment of i as a stable lysogen is a frequent

 

event (109 to 1071) for the attt intt eI* phage so that this

mode of escape would be the preponderant laboratory hazard;

While not exact, the estimates for containment
afforded by using these host-veetors are at least as
accurate as those for physical containment, and are
sufficient to indicate that currently employed plasmid
and A veetor systems provide a moderate level of btio-
logteal containment. Other noneconjugative plasmids
and bactertophages that, in association with E. colt

K-12 can be estimated to provide the same approximate

 

level of moderate containment are included in the

EK1 elass.
ee

EK2: This class consists of combinations of modified £. colt K-12
"hosts" with modified plasmid or bacteriophage "vectors. The
modifications will be achieved by classical genetic manipulation

\S L and must be such that the survival rate of the recombined

“foreign” DNA fragment, in natural environments, is less than
one in 108. Testing procedures for verification of the proper-
ties of EK2 systems are described. (EK2 systems are being

developed and tested and it is anticipated that they will
be available shortly. Responsibility for certification
of putative EK2 systems presently lies with the NIH
Recombinant DNA Molecule Program Advisory Committee.)

EK2 host-vectors - These are host-vector systems that have been
genetically constructed and shown to provide a high level of biological
containment as demonstrated by data from suitable tests performed in

the laboratory. The genetic modifications of the E. coli K-12 host

 

and/or the plasmid or phage vector should not permit survival of the
cloned DNA fragment in other than specially designed and carefully
regulated laboratory environments at a frequency qreater than 10-8.
This absolute measure of biological containment has been selected
because it is a realistic measurable entity. Indeed, by testing

the contributions of preexisting and newly introduced genetic pro-
perties of vectors and hosts, individually or in various combinations,

it should be possible to/substantial indicate/ if not prove, that

the specially designed host-vector system can provide a margin of

 

biological containment in excess of that required.

For EK2 plasmid vectors, no more than one in 108 host cells con-
_taining a chimeric plasmid should be able to perpetuate the cloned DNA
fragment under nonpermissive conditions designed to represent the natural
environment either by survival of the original host or as a consequence
of transmission of the cloned DNA by transformation, transduction or
conjugation to a host with properties common to those in the natural

environment.

 
In the construction of EK2 plasmid-host systems it te
important to use the most stable mutations avatlable, preferabi::
deletions. Obviously, the presence of all mutations contributing
to higher degrees of biological containment must be verified
pertodically by appropriate tests. In testing the level of
biologteal containment afforded by a proposed EK2 plasmid-host
system, tt ts important to design relevant tests to evaluate
the survival of the cloned DNA under conditions that are posstble
tn nature and that are also most advantageous for tts perpetuation.
For example, one might conduct a triparental mating with a primary.
donor possessing a derepressed F-type or I-type conjugative plasmid,
the safer host with AbioH-asd, dapD8, Agal-chl", AthyA, deoC, trp and
hsdS mutations and a plasmtd vector carrying an easily detectable
inserted gene such as for ampicillin resistance or trp’, and a
secondary recipient that is Sut hsdS trp (t.e., permissive for the
recombinant plasmid). Such matings would be conducted in a medium
lacking diaminopimelie acid and thymine and survival of the Ap™
or trp? marker in any of the three strains followed as a funetion
of time. Survival of the cloned marker by transduction could also
be evaluated by introduzing a known generalized transducing phage

into the system. Similar experiments should also be done
using a secondary reciptent that is restrictive for the
recombinant plasmid as well as with primary denors possess-
ing repressed conjugative plasmids with tneompattbi lity

group properties like those commonly found in enteric
microorganisms. Since a common route of escape of plasmid-
host systems in the laboratory might be by accidental
ingestion, it ts suggested that the same types of experi-
ments be conducted in suitable animal-model systems. In

addition to these tests on survival of the cloned DNA, tt

 

would be useful to determine the survival of the host strain
under nongrowth conditions such as in water and as a funetion

of drying time after a culture has been spilled on a lab bench.

For EK2 phage vectors, no more than one in 108 recombinant phage
particles should be able to perpetuate the cloned DNA fragment under
non-permissive conditions designed to represent the natural environ-
ment either (a) as a prophage or plasmid in the laboratory host used
for phage propagation or. (b) by surviving in natural environments
and transferring the cloned DNA to a host (or its resident lambdoid

prophage) with properties common to those in the natural environment.

The phenotypes and genetic stabilittes of the mutations
and chromosome alterations ineluded in these r-host systems
indicate that containment well in excess of the requtred
107? or lower survival frequency for the eloned DNA fragment

should be attained. Obviously the presence of all mutations

 

contributing to this high degree of biologtcal containment
must be vertfied periodically Ly appropriate tests. Laboratory

tests should be performed with the bacterial host to measure

all posstble routes of escape of eLloned DNA such as the

 

_ frequency of lysogen formation, the frequeney of plasmid

 
| formation and the survival of the lysogen or ecarrter bacteriun.
Similarly, the potential for perpetuation of the cloned DNA
fragment carried by infectious phage particles could be tested
by challenging typteal wild-type E. colt strains or a h-sensitive
nonpermissive laboratory K-12 strain, espectally one Zysogente

for a lambdoid phage...

. EK3: These are EK2 systems for which the increased containment
45 ' has. been independently confirmed by appropriate tests in
r animals. EK3 systems are not presently available.
2. Classification of experiments using the E. coli K-12

containment systems

In the following classification of containment criteria for different

kinds of recombinant DNAs, the stated levels of physical and biological

containment are minimums. It is recommended that higher levels of
- biological containment (EK3 > EK2 > EK1) be used if they are available

and are equally appropriate for the purposes of the experiment.

<a> Shotgun Experiments

These experiments involve the production of recombinant DNAs
between the vector and the total DNA or (preferably) any partially

purified fraction thereof from the specified cellular sourte.

2. Donald Brown: The guidelines are too strict and inflexible.
Requiring EK2 containment for plant, vertebrate and viral
_DNAs places.in effect an indefinite moratorium on research.
Basis for this is that there are at present no EK2 vectors
nox. 4 clear mechanism_of how stems will be certifi
facilities. . ye

1. Robert Sinsheimer: Shotgun experiments with eukaryotic
DNA incorporated into E. coli are especially troublesome.
A safer and potentially reversible method is to incorporate
such DNA into animal viruses where there is already in
place a viable defense. One such example that comes to
mind is cowpox. Here vaccination is possible and if the
hazard were perceived, all infected animals and tissue
culture cells could be destroyed.

4. Allen Silverstone: He recommends banning of all shotgun
experiments with mammals and birds. For all of the
eukaryotes P4 containment is recommended. He believes
the possibility of cancer-causing DNA in lower organisms
is a potential hazard which requires the highest physical
containment. In the case of mammals he also recommends
a ban on the use of chemically purified DNA.
 

 

 

 

 

 

Cold-blooded vertebrates — P2 physical containment + an EK2
Vertebrates

 

 

 

 

 

If the plant Carries a known Pathogenic agent or makes a Product known

to be dangerous to any Species, the containment Should be raised to P3
v0 | Ota tn —

physical] containment + an EK2 host-vector.
0 9X] -Y3-(44) prokaryotic DNA recombinants
prokaryotes that exchange genetic information with E. coli -
The level of physical containment is directly determined by the rule
of the most dangerous component (see introduction to Section III).

Thus Pl conditions can be used for DNAs from those bacteria in Class ]

of ref. 5 ("Agents of no or minimal hazard....") which naturally ex-

 

change genes with E. coli; and P2 conditions should be used for such

 

bacteria if they fall in Class 2 of ref. 5 ("Agents of ordinary potential

hazard...."), or are plant pathogens. EK1 host-vectors can be used for

 

* bef ined as observable under optimal laboratory conditions by transformation,

transduction, phage infection and/or conjugation with transfer of phage, plasmid
and/or chromosomal genetic information.
all experiments requiring only Pl physical containment; in fact, ex-
periments in this category can be performed with E. coli K-12 vectors
exhibiting a lesser containment (e.g., conjugative plasmids) than EK]
vectors. Experiments with DNA from species requiring P2 physical con-
tainment which are of low pathogenicity (for example, enteropathogenic

Escherichia coli, Salmonella typhimurium, and Klebsiella pneumoniae)

 

can use EK] host-vectors, but those of moderate pathogenicity (for
example, Salmonella typhi, Shigella dysenteriae type I, and Vibrio

cholerae) should use EK2 host-vectors.°

*the bacteria which constitute Class 2 of ref. 5 ("Agents of ordinary potential
hazard....") represent a broad spectrum of etiologic agents which possess dif-
ferent levels of virulence and degrees of communicability. We think it appro-
priate for our specific purpose to further subdivide the agents of Class 2 into
those which we believe to be of relatively low pathogenicity and those which are
moderately pathogenic. The several specific examples given may suffice to
illustrate the principle. The Committee has asked the Center for Disease Control
to help prepare a complete list of Class 2 agents subdivided into low and moderate
pathogenicity that could be distributed to interested parties.
Prokaryotes that do not exchange genetic information with

E. coli - The minimum containment conditions for this class
“consist of P2 physical containment + an EK1 host-vector, and apply
when the risk that the recombinant DNAs will increase the patho-
genicity or ecological potential of the host is judged to be minimal.
Experiments with DNAs from pathogenic species (Class 2 ref. 5 plus
“plant pathogens) should use P3 + EK2.

Experiments extending the range of resistance to therapeutically

 

useful drugs and disinfectants should use P2 + EK2 containment or higher

 

_ depending on the virulence of the donor.
g 2s , (iii) Characterized clones of DNA recombinants derived from

shotgun experiments

4. The Environmental Defense Fund: They too urge that E. coli
“mot be used. They also are extremely concerned about use
of the terms, "harmful" and "harmless," in describing genes
and their products. They point out that accidents will
occur and would like as broad a definition for “harmful”
as possible to minimize possible accidents. They recommend
that EKi strains never be used except in circumstances
where there is no possible risk to humans.

 

When a cloned DNA recombinant has been rigorously characterized"

and there is sufficient evidence that it is free of harmful genes,"

 

then experiments involving this recombinant DNA can be carried out
under P1 + EK] conditions if the inserted DNA is from a species that

exchanges genes with E. coli, and under P2 + EK] conditions if not.

 

“the terms "characterized" and "free of harmful genes" are unavoidably
vague. But, in this instance, before containment conditions lower than
the ones used to clone the DNA can be adepted, the investigator must obtain
approval from the granting agency. Such approval would be contingent upon
data concerning: (a) the absence of potentially harmful genes (e.g.,
sequences contained in indigenous tumor viruses or which code for toxic
substances), (b) the relation between the recovered and desired segment
(e.g., hybridization and restriction endonuclease fragmentation analysis

where applicable), and (c) maintenance of the biological properties of
the vector.

Urol oheert myers ‘\ &
0%

1.

<b> Purified-cellular DNAs other than plasmids, bacteriophages,

and other viruses

Donald Brown: The guidelines for vertebrate "shotgun"
experiments are too stringent and rigid. They
establish a moratorium on gene isolation from all
vertebrates. He urged the use of hybridization
methods at reduced levels of physical and biological
containment. Further, the requirements for chemically
purified DNA are not achievable. There can be no
guarantees for 99% purity. The Committee did not
consider how purified DNA components differ from

bulk DNA.

 

 

ee

The formation of DNA recombinants from cellular DNAs that have been

enriched? by physical and chemical techniques (i.e., not by cloning) and

which are free of harmful genes can be carried out under lower contain-

ment conditions than used for the corresponding shotgun experiment. In

general, the containment can be decreased one step in physical containment

(P4 + P3 + P2 + P1) while maintaining the biological containment specified

for the shotgun experiment, or one step in biological containment

(EK3 + EK2 > EK1) while maintaining the specified physical contain-

ment--provided that the new condition is not less than that specified

above for characterized clones from shotgun experiments (Section <a>--

iii).

A DNA preparation is defined as enriched if the desired DNA represents at least
99% (w/w) of the total DNA in the preparation. The reason for lowering the con-
tainment level when this degree of enrichment has been obtained is based on the
fact that the total number of clones that must be examined to obtain the desired
clone is markedly reduced. Thus, the probability of cloning a harmful gene
could, for example, be reduced by more that 10°-fold when a nonrepetitive gene
from mammals was being sought. Furthermore, the level of purity specified here
makes it easier to establish that the desired DNA does not contain harmful genes.
<c> The following classifications refer to cases where the
~ "foreign" DNA is itself derived from plasmids, bacterio-

Q-3¥ phages, or viruses.
> (i) "Foreign" DNA from viruses that infect animals.
P4 and EK2, or P3 and EK3 are to be used. P3 and EK2

may be used after purification by cloning and demon-
stration that only harmless viral genes are present.

(ii) "Foreign" DNA from viruses that infect plants.

P3 an P2 and EK2,

  
  
  

 

"Foreign" DNA from purified \(99% pure)/eukaryotic

organelle DNAs.

From primates: P3 and EK1 or P?2 and EK2.
Other: P2 and EK1.

(iv) "Foreign" DNA from prokaryotic plasmids or

bacteriophages. .

If the "foreign" DNA is from a species that exchanges
genetic information with F. colt and is known not to
contain harmful genes, Pl and ERI may be used. Other-
wise, the containment defined in <a> (ii) above is
required.
host-vector" systems other than

BE. coli

932 3. Experiments with prokaryotic "

l. Richard Goldstein: As previously noted he recommends E. coli
be banned as a host~vector within the next two years. He
urges that the development of alternative prokaryote systems
be developed and suggests one possibility is B. subtilis.

Other prokaryotic host-vector systems are at the speculative,
planning, or developmental stage, and consequently do not warrant
detailed treatment here at this time. However, the containment
criteria for different types of DNA recombinants formed with E.
coli K-12 host-vectors can, with the aid of some general principles
given here, serve as a guide for containment conditions with other

host-vectors when appropriate adjustment is made for their different

habitats and characteristics.

In general, the strain of any prokaryotic species used as the
host should conform: to the definition of Class 1 etiologic agents
given in ref. 5 (i.e., “Agents of no or minimal hazard...."), and
the plasmid or phage vector should not make the host more hazardous.
In addition, it is recommended that the newly developed host-vector
Systems offer some distinct advantage over the E. coli K-12 host-
vectors--for instance, thermophilic organisms or other host-vectors
whose major habitats do not include humans and/or economically im-
portant animals and plants. Appendix A gives a detailed discussion

of the B. subtilis system, the most promising alternative to date.
At the initial stage, the host-vector should exhibit at least
a moderate level of biological containment comparable to EK] systems,
and be capable of modification to obtain high levels of containment
comparable to EK2 and EK3. The type of confirmation test(s) required
to move a host-vector from an EK2-type classification to an EK3-type
will clearly depend upon the preponderant habitat of the host-vector.
-For example, if the unmodified host-vector propagates mostly in, on,
or around higher plants, but not appreciably in warm-blooded animals,
modification should be designed to reduce the probability that the
host-vector can escape to and propagate in, on, or around such plants,
or transmit recombinant DNA to other bacterial hosts that are able to
occupy these ecological niches, and ijt is these lower probabilities
which should be confirmed. The following principles should be followed
in using the containment criteria given for experiments with E. coli
K-12 host-vectors as a guide for other prokaryotic systems. Experi-
ments with DNA from prokaryotes (and their plasmids or viruses)

should be classified according to whether the prokaryote in question

exchanges genetic information with the host-vector or not, and the con-
tainment conditions given for these two classes with E. coli K-12
host-vectors applied. Transfer of recombinant DNA to plant pathogens
can be made safer by using nonreverting, doubly auxotrophic, non-
pathogenic variants. Experiments using a plant pathogen that affects
elements of a local flora will require more stringent containment
than if carried out in areas where they are not common.

Experiments with DNAs from eukaryotes (and their plasmids or
viruses) can also follow the criteria for the corresponding experi-

ments with E. coli K-12 vectors if the major habitats of the qiven
 

host-vector overlap those of E. coli. ‘If the host-vector has a major

 

habitat that does not overlap those of E. coli, (e.g., root nodules
in plants), then the containment conditions for some eukaryotic
recombinant DNAs should be increased (for instance, higher plants
and their viruses in the preceding example), while others may be

reduced,
4. Experiments with eukaryotic "host-vector'" systems

1. Maxine Singer: The use of SV40 virus is troublesome.
If it is not to be banned altogether, it should probably
be used only at P4 levels.

2. Peter Barton Hutt: In the case of SV40, if it is to
be used at all, it should be done only at the P4 level.
Because this virus is known to cause cancer in animals,
the burden is on the scientist to show there is no
possibility whatever of harm to man when it is used in
these experiments. If it is to be used, its use should
be fully defended in the preamble in the guidelines.

7y\ <a> Animal host-vector systems - Because host cell lines generally
o have little if any capacity for propagation outside the laboratory, the
( primary focus for containment is the vector, although cells should also be
derived from cultures expected to be of minimal hazard. Given good micro-
biological practices, the most likely mode of escape of recombinant DNAs
from a physically contained @aboratory is carriage by humans; thus vectors
should be chosen that have little or no ability to replicate in
human cells. To be used as a vector in a eukaryotic host, a DNA

molecule should display all of the following properties:

 

(1) It should not consist of the whole genome of

 

any agent that is infectious for humans or that replicates

 

to a significant extent in human cells in tissue culture.
(2) Its functional anatomy should be known--that is,
there should be a clear idea of the location within the
molecule of:
a) the Sites at which DNA synthesis originates
and terminates,
b} the sites that are cleaved by restriction
endonucleases,
c) the template regions for the major gene products.
(3) It Should be well studied genetically. It is desirable
that mutants be available in adequate number and variety, and

that quantitative studies of recombination have been performed.

(4) The recombinant should be defective, that is, its pro-

 

pagation as a virus is dependent upon the presence of a comple-
menting helper genome. This helper should either (a) be integrated
into the genome of a stable line of host cells (a situation that
would effectively limit the growth of the vector to that particular
cell line) or (b) consist of a defective genome or an appropriate
conditional lethal mutant virus (in which case the experiments
would be done under non-permissive conditions), making vector and
helper dependent upon each other for propagation. However, if

none of these is available, the use of a non-defective genome

as helper would be acceptable.

Currently only two viral DNAs can be considered as meeting

these requirements: these are the genomes of polyoma virus and

 

SV40.
Of these, polyoma virus is highly to be preferred. SV40
is known to propagate in human cells, both in vivo and in vitro,
and to infect laboratory personnel, as evidenced by the frequency
of their conversion to producing SV40 antibodies. ‘Also, SV40
and related viruses have been found in association with certain
human neurological and malignant diseases. SV40 shares many
properties, and gives complementation, with the common human
papova viruses. By contrast, there is no evidence that polyoma
infects humans, nor does it replicate to any significant extent
in human cells in vitro. However, this system still needs to
be studied more extensively. Appendix B gives further details
and documentation.
Taking account of all these factors, it is proposed that:
1. Polyoma Virus
a Recombinant DNA molecules consisting of defective
polyoma virus genomes plus DNA sequences of any non-
pathogenic organism, including Class 1 viruses (5), can
be propagated in or used to transform cultured cells
in P3 conditions; appropriate helper virus can be used

if needed. Whenever there is a choice, it is urged

that mouse cells, derived preferably from embryos, be
used as the source of eukaryotic DNA. Polyoma virus

is a mouse Virus and recombinant DNA molecules contain-
ing both viral and cellular sequences are already known
to be present in virus stocks grown at a high multiplic-
ity. Thus, recombinants formed in vitro between polyoma
virus DNA and mouse DNA are presumably not novel from

an evolutionary point of view.
b. Such experiments can be done, under P4 conditions, if the
recombinant DNA contains segments of the genomes of Class
2 animal viruses (5). Once it has been shown by suit-

‘ able biochemical and biological tests that the cloned

recombinant contains only harmless regions of the viral °

 

 

genome (see Section IIIB-2-c-i) and that the host rang Vallee
aT

of the polyoma virus vector has not been altered, ex- wad
periments can be continued under P3 conditions. yar”

2. SV40 Virus
a. Defective SV40 genomes, with appropriate helper, can
, .7 be used in P4 conditions as a vector for recombinant DNA
| molecules containing sequences of any non-pathogenic
organism or Class I virus (5), (i.e., a shotgun type
experiment) ; established lines of cultured cells should

be used.

[om

. Such experiments can be carried out in P3 conditions if

the non-SV40 DNA segment is (a) a purified? segment of

 

prokaryotic DNA lacking toxigenic genes, or (b) a segment

 

“The DNA preparation is defined as purified if the desired DNA represents
at least 99% (w/w) of the total DNA in the preparation, provided that it
was achieved or verified by more than one procedure.
of eukaryotic DNA whose function has been established,
which does not code for a toxic product, and which has
been previously cloned ina prokaryotic host-vector
system.

. A recombinant DNA molecule consisting of defective SV40

DNA lacking substantial segments of the late region,

{o

plus DNA from non-pathogenic organisms or Class I

viruses (5), can be propagated as an autonomous cellular
element in established lines of cells under P3 conditions
provided that there is no exogenous or endogenous helper,

and that it is demonstrated that no infectious virus fo’
particles are being produced. Until this has been
demonstrated, the appropriate containment conditions

specified in 2. a. and 2. b. shall be used.

. Recombinant DNA molecules consisting of defective SV40

{a

DNA and sequences from non-pathogenic prokaryotic or
eukaryotic organisms or Class I viruses (5) can be
used to transform established lines of non-permissive
cells under P3 conditions. it must be demonstrated that
fd infectious virus particles are being produced; rescue
of SV40 from such transformed cells by co-cultivation
or transfection techniques must be carried out in P4
conditions.

3. Efforts should be made to ensure that all cell lines are

free of virus particles and mycoplasma.
Since S¥40 and polyoma are limited in their scope to

act as vectors, chiefly because the amount of foreign DNA
that the normal virions can carry probably cannot exceed
2x 10° daitons, we urge that consideration be given

to the development of systems in which recombinants can

be cloned and propagated purely in the form of DNA, rather
than in the coats of infectious agents. Plasmid forms

of viral genomes or organelle DNA should be explored as

possible cloning vehicles in eukaryotic cells.

p39 lt Plant host-vector systems

1. Milton Zaitlin: In the guidelines plants have only a
peripheral role; however, they could become extraordinarily
important in recombinant DNA studies. Cloning of higher
plant DNA and incorporating this DNA into bacteria is well
covered by the guidelines. However, the guidelines are

_ Silent on the incorporation of foreign DNA into the genome
of higher plants. There are five DNA plant viruses that
could play a significant role here. These viruses are
transmitted by aphids and are a problem with regard to physical
containment. Physical containment in the guidelines is

ay designed to prevent bacteria from getting out of the facility
and to to this, negative air pressures are specified. But
positive air pressures are required to avoid drawing insects
into the facility. Guidelines might be modified to specify

ys positive air pressures under these circumstances. The guide-
Ta . lines also are silent on how to deal with a new plant once
per ws it is produced. Procedures need to be outlined for testing
+0 « new plants for any undesirable characteristics they might
' ( , have. The guidelines, therefore, should include restrictions

on the release of plants from containment facilities until they
can be tested through several generations. Plants should be
. tested to insure they are not toxic to animals which might be
likely to eat them. In plants where the capacity for nitrogen
fixation is introduced, testing must be done to insure that
these plants don't have a competitive advantage over cher -
plants and become pests in themselves,
Cells in tissue cultures, seedlings, or plant parts (e.g., tubers,
stems, fruits, and detached leaves) or whole mature plants of small
species (e.g., Arabidopsis) can be handled under the P1-P4 containment
conditions that we have specified previously. However, work in most
plants poses additional problems. P2 physical containment conditions

-can be provided by: (i) the best insect-proof greenhouses, (ii) appro-
priate disinfection of contaminated plants, pots, soil, and runoff
water, and (iii) adoption of the other standard practices for microbio-
logical work. P3 physical containment can be sufficiently approximated
by confining the operations with whole plants to growth chambers like those
used for work with radioactive tsatopes, provided that (i) “such chambers
2 - are modified to produce s(eantive oeseure environment with the exhaust
air appropriately filtered, and (ii) that other operations with infectious
materials are carried out under the specified P3 conditions. The P2
and P3 conditions specified earlier are therefore extended to include

these cases for work on higher plants.

The host cells for experiments on recombinant DNAs may be cells
in culture, in seedling or plant parts, or in whole plants. Cells in
whole plants that cannot be adequately contained should not be used
as hosts for shotgun experiments at this time, and attempts to infect
whole plants with DNA recombinants cloned elsewhere should not be
initiated until their effects on host cells in culture, seedlings,

or plant parts have been studied.
Organelle or plasmid DNAs or DNAs of viruses of low pathogenicity
to plants may be used as vectors. In general, the same preference
criteria for selecting host-vectors given in the preceding secticn on
animal systems apply to plant systems, where organelle and plasmid DNAs
can be grouped together as offering the potential of highly contained
vectors that should be investigated.

Experiments on recombinant DNAs formed between the initial moderately
contained vectors and DNA from cells of species in which the vector DNA
can replicate, either autonomously or as an integrated segment of the
cell's genome, should use P2 physical containment--provided that the
source of the DNA is itself not pathogenic or known to carry pathogenic
agents, or to produce products dangerous to plants. In the latter cases,
of if the vector is an unmodified virus of low pathogenicity, the ex-
periments should be carried out under P3 conditions.

Experiments on recombinant DNAs formed between the above vectors
and DNAs from other species can also be carried out under P2 if that DNA

has been purified?® and determined not to contain harmful genes. Other-

 

10the DNA preparation is defined as purified if the desired DNA represents
at least 99% (w/w) of the total DNA in the preparation, provided that it
was achieved or verified by more than one procedure.
wise, the experiments should be carried out under P3 conditions if the
source of the inserted DNA is not itself a pathogen, or known to carry
such pathogenic agents, or to produce harmful products--and under P4
conditions if these conditions are not met.

The development and use of host-vector systems that exhibit a
high level of biological containment permit a decrease of one step in

the physical containment specified above (P4 + P3 + P2 + Pl).

<c> Fungal or similar lower eukaryotic host-vector systems

The containment criteria for experiments on recombinant DNAs using
these host-vectors most closely resemble those for prokaryotes, rather
than those for the preceding eukaryotes, in that the host cells usually
exhibit a capacity for dissemination outside the laboratory that is
similar to that for bacteria. We therefore consider that the contain-
ment guidelines given for experiments with E. coli K-12 and other pro-
karyotic host-vectors (Sections IIIB-] and -2, respectively) provide
adequate direction for experiments with these lower eukaryotic host-
vectors. This is particularly true at this time since the development

of these host-vectors is presently in the speculative stage.
IV. IMPLEMENTATION

B cing Revseck