Summary Statement of the Asilomar Conference on Recombinant DNA Molecules

Paul Berg, David Baltimore, Sydney Brenner, Richard O. Roblin, and Maxine F. Singer

doi:10.1073/pnas.72.6.1981 1975;72;1981-1984 PNAS

This information is current as of December 2006.

www.pnas.org#otherarticlesThis article has been cited by other articles:

E-mail Alerts. click heretop right corner of the article or

Receive free email alerts when new articles cite this article - sign up in the box at the

Rights & Permissions www.pnas.org/misc/rightperm.shtml

To reproduce this article in part (figures, tables) or in entirety, see:

Reprints www.pnas.org/misc/reprints.shtml

To order reprints, see:

Notes:

Proc. Nat. Acad. Sci. USAVol. 72, No. 6, pp. 1981-1984, June 1975

Summary Statement of the Asilomar Conference on Recombinant DNAMolecules*

PAUL BERGt, DAVID BALTIMORE$, SYDNEY BRENNER§, RICHARD 0. ROBLIN III¶, ANDMAXINE F. SINGER||

Organizing Committee for the International Conference on Recombinant DNA Molecules, Assembly of Life Sciences, National ResearchCouncil, National Academy of Sciences, Washington, D.C. 20418. t Chairman of the committee and Professor of Biochemistry,Department of Biochemistry, Stanford University Medical Center, Stanford, California; * American Cancer Society Professor of Micro-biology, Center for Cancer Research, Massachusetts Institute of Technology, Cambridge, Mass.; § Member, Scientific Staff of the MedicalResearch Council of the United Kingdom, Cambridge, England; ¶ Professor of Microbiology and Molecular Genetics, Harvard MedicalSchool, and Assistant Bacteriologist, Infectious Disease Unit, Massachusetts General Hospital, Boston, Mass.; and II Head, Nucleic AcidEnzymology Section, Laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Bethesda, Maryland

1. INTRODUCTION AND GENERAL CONCLUSIONS

This meeting was organized to review scientific progress inresearch on recombinant DNA molecules and to discussappropriate ways to deal with the potential biohazards of thiswork. Impressive scientific achievements have already beenmade in this field and these techniques have a remarkablepotential for furthering our understanding of fundamentalbiochemical processes in pro- and eukaryotic cells. The use ofrecombinant DNA methodology promises to revolutionize thepractice of molecular biology. Although there has as yet beenno practical application of the new techniques, there is everyreason to believe that they will have significant practicalutility in the future.Of particular concern to the participants at the meeting

was the issue of whether the pause in certain aspects ofresearch in this area, called for by the Committee on Re-combinant DNA Molecules of the National Academy ofSciences, U.S.A. in the letter published in July, 1974**should end; and, if so, how the scientific work could be under-taken with minimal risks to workers in laboratories, to thepublic at large, and to the animal and plant species sharingour ecosystems.The new techniques, which permit combination of genetic

information from very different organisms, place us in anarea of biology with many unknowns. Even in the present,more limited conduct of research in this field, the evaluationof potential biohazards has proved to be extremely difficult.It is this ignorance that has compelled us to conclude that itwould be wise to exercise considerable caution in performingthis research. Nevertheless, the participants at the Conferenceagreed that most of the work on construction of recombinantDNA molecules should proceed provided that appropriatesafeguards, principally biological and physical barriers ade-

* Summary statement of the report submitted to the Assembly ofLife Sciences of the National Academy of Sciences and approvedby its Executive Committee on 20 May 1975.Requests for reprints should be addressed to: Division of MedicalSciences, Assembly of Life Sciences, National Academy ofSciences, 2101 Constitution Avenue, N.W., Washington, D.C.20418.** Report of Committee on Recombinant DNA Molecules:"Potential Biohazards of Recombinant DNA Molecules," Proc.Nat. Acad. Sci. USA 71, 2593-2594, 1974.

1981

quate to contain the newly created organisms, are employed.Moreover, the standards of protection should be greater atthe beginning and modified as improvements in the method-ology occur and assessments of the risks change. Furthermore,it was agreed that there are certain experiments in which thepotential risks are of such a serious nature that they oughtnot to be done with presently available containment facilities.In the longer term, serious problems may arise in the largescale application of this methodology in industry, medicine,and agriculture. But it was also recognized that future re-search and experience may show that many of the potentialbiohazards are less serious and/or less probable than we nowsuspect.

IX. PRINCIPLES GUIDING THE RECOMMENDATIONSAND CONCLUSIONS

Although our assessments of the risks involved with each ofthe various lines of research on recombinant DNA moleculesmay differ, few, if any, believe that this methodology is freefrom any risk. Reasonable principles for dealing with thesepotential risks are: (i) that containment be made an essentialconsideration in the experimental design and, (ii) that theeffectiveness of the containment should match, as closely aspossible, the estimated risk. Consequently, whatever scale ofrisks is agreed upon, there should be a commensurate scale ofcontainment. Estimating the risks will be difficult and in-tuitive at first but this will improve as we acquire additionalknowledge; at each stage we shall have to match the potentialrisk with an appropriate level of containment. Experimentsrequiring large scale operations would seem to be riskier thanequivalent experiments done on a small scale and, therefore,require more stringent containment procedures. The use ofcloning vehicles or vectors (plasmids, phages) and bacterialhosts with a restricted capacity to multiply outside of thelaboratory would reduce the potential biohazard of a par-ticular experiment. Thus, the ways in which potential bio-hazards and different levels of containment are matched mayvary from time to time, particularly as the containmenttechnology is improved. The means for assessing and balanc-ing risks with appropriate levels of containment will need tobe reexamined from time to time. Hopefully, through bothformal and informal channels of information within ind be-tween the nations of the world, the way in which potentialbiohazards and levels of containment are matched would beconsistent.

1982 Summary Statement: Berg et al.

Containment of potentially biohazardous agents can beachieved in several ways. The most significant contribution tolimiting the spread of the recombinant DNAs is the use ofbiological barriers. These barriers are of two types: (i) fas-tidious bacterial hosts unable to survive in natural environ-ments, and (ii) nontransmissible and equally fastidious vec-tors (plasmids, bacteriophages, or other viruses) able togrow only in specified hosts. Physical containment, ex-emplified by the use of suitable hoods, or where applicable,limited access or negative pressure laboratories, provides anadditional factor of safety. Particularly important is strictadherence to good microbiological practices which, to a largemeasure can limit the escape of organisms from the experi-mental situation, and thereby increase the safety of theoperation. Consequently, education and training of all per-sonnel involved in the experiments is essential to the effec-tiveness of all containment measures. In practice, thesedifferent means of containment will complement one anotherand documented substantial improvements in the ability torestrict the growth of bacterial hosts and vectors could permitmodifications of the complementary physical containmentrequirements.

Stringent physical containment and rigorous laboratoryprocedures can reduce but not eliminate the possibility ofspreading potentially hazardous agents. Therefore, investi-gators relying upon "disarmed" hosts and vectors for addi-tional safety must rigorously test the effectiveness of theseagents before accepting their validity as biological barriers.

1II. RECOMMENDATIONS FOR MATCHING TYPES OFCONTAINMENT WITH TYPES OF EXPERIMENTS

No classification of experiments as to risk and no set of con-tainment procedures can anticipate all situations. Given ourpresent uncertainties about the hazards, the parametersproposed here are broadly conceived and meant to provideprovisional guidelines for investigators and agencies con-cerned with research on recombinant DNAs. However, eachinvestigator bears a responsibility for determining whether,in his particular case, special circumstances warrant a higherlevel of containment than is suggested here.

A. Types of containment

1. Minimal Risk. This type of containment is intended forexperiments in which the biohazards may be accuratelyassessed and are expected to be minimal. Such containmentcan be achieved by following the operating procedures recom-mended for clinical microbiological laboratories. Essentialfeatures of such facilities are no drinking, eating, or smokingin the laboratory, wearing laboratory coats in the work area,the use of cotton-plugged pipettes or preferably mechanicalpipetting devices, and prompt disinfection of contaminatedmaterials.

2. Low Risk. This level of containment is appropriate forexperiments which generate novel biotypes but where theavailable information indicates that the recombinant DNAcannot alter appreciably the ecological behavior of the re-cipient species, increase significantly its pathogenicity, orprevent effective treatment of any resulting infections. Thekey features of this containment (in addition to the minimalprocedures mentioned above) are a prohibition on mouth

use of biological safety cabinets for procedures likely toproduce aerosols (e.g., blending and sonication). Thoughexisting vectors may be used in conjunction with low riskprocedures, safer vectors and hosts should be adopted as theybecome available.

3. Moderate Risk. Such containment facilities are intendedfor experiments in which there is a probability of generatingan agent with a significant potential for pathogenicity orecological disruption. The principle features of this level ofcontainment, in addition to those of the two preceding classes,are that transfer operations should be carried out in biologicalsafety cabinets (e.g., laminar flow hoods), gloves should beworn during the handling of infectious materials, vacuumlines must be protected by filters, and negative pressureshould be maintained in the limited access laboratories.Moreover, experiments posing a moderate risk must be doneonly with vectors and hosts that have an appreciably impairedcapacity to multiply outside of the laboratory.

4. High Risk. This level of containment is intended forexperiments in which the potential for ecological disruption orpathogenicity of the modified organism could be severe andthereby pose a serious biohazard to laboratory personnel orthe public. The main features of this type of facility, whichwas designed to contain highly infectious microbiologicalagents, are its isolation from other areas by air locks, anegative pressure environment, a requirement for clothingchanges and showers for entering personnel, and laboratoriesfitted with treatment systems to inactivate or remove bio-logical agents that may be contaminants in exhaust air andliquid and solid wastes. All persons occupying these areasshould wear protective laboratory clothing and shower ateach exit from the containment facility. The handling ofagents should be confined to biological safety cabinets inwhich the exhaust air is incinerated or passed through Hepafilters. High risk containment includes, in addition to thephysical and procedural features described above, the use ofrigorously tested vectors and hosts whose growth can beconfined to the laboratory.

B. Types of experiments

Accurate estimates of the risks associated with differenttypes of experiments are difficult to obtain because of ourignorance of the probability that the anticipated dangerswill manifest themselves. Nevertheless, experiments involvingthe construction and propagation of recombinant DNAmolecules using DNAs from (i) prokaryotes, bacteriophages,and other plasmids, (ii) animal viruses, and (iii) eukaryoteshave been characterized as minimal, low, moderate, and highrisks to guide investigators in their choice of the appropriatecontainment. These designations should be viewed as interimassignments which will need to be revised upward or down-ward in the light of future experience.The recombinant DNA molecules themselves, as distinct

from cells carrying them, may be infectious to bacteria orhigher organisms. DNA preparations from these experiments,particularly in large quantities, should be chemically in-activated before disposal.

1. Prokaryotes, Bacteriophages, and Bacterial Plasmids.Where the construction of recombinant DNA molecules andtheir propagation involves prokaryotic agents that are knownto exchange genetic information naturally, the experiments

Proc. Nat. Acad. Sci. USA 72 (1975)

pipetting, access limited to laboratory personnel, and the

Conference on Recombinant DNA Molecules 1983

can be performed in minimal risk containment facilities.Where such experiments pose a potential hazard, morestringent containment may be warranted.Experiments involving the creation and propagation of

recombinant DNA molecules from DNAs of species thatordinarily do not exchange genetic information, generatenovel biotypes. Because such experiments may pose bio-hazards greater than those associated with the original organ-isms, they should be performed, at least, in low risk contain-ment facilities. If the experiments involve either pathogenicorganisms or genetic determinants that may increase thepathogenicity of the recipient species, or if the transferredDNA can confer upon the recipient organisms new metabolicactivities not native to these species and thereby modify itsrelationship with the environment, then moderate or highrisk containment should be used.Experiments extending the range of resistance of established

human pathogens to therapeutically useful antibiotics or dis-infectants should be undertaken only under moderate or highrisk containment, depending upon the virulence of theorganism involved.

2. Animal Viruses. Experiments involving linkage of viralgenomes or genome segments to prokaryotic vectors and theirpropagation in prokaryotic cells should be performed onlywith vector-host systems having demonstrably restrictedgrowth capabilities outside the laboratory and with moderaterisk containment facilities. Rigorously purified and character-ized segments of non-oncogenic viral genomes or of the de-monstrably non-transforming regions of oncogenic viral DNAscan be attached to presently existing vectors and propagatedin moderate risk containment facilities; as safer vector-hostsystems become available such experiments may be performedin low risk facilities.Experiments designed to introduce or propagate DNA from

non-viral or other low risk agents in animal cells should useonly low risk animal DNAs as vectors (e.g., viral, mitochon-drial) and manipulations should be confined to moderate riskcontainment facilities.

3. Eukaryotes. The risks associated with joining randomfragments of eukaryote DNA to prokaryotic DNA vectorsand the propagation of these recombinant DNAs in pro-karyotic hosts are the most difficult to assess.A priori, the DNA from warm-blooded vertebrates is more

likely to contain cryptic viral genomes potentially pathogenicfor man than is the DNA from other eukaryotes. Conse-quently, attempts to clone segments of DNA from suchanimal and particularly primate genomes should be performedonly with vector-host systems having demonstrably re-stricted growth capabilities outside the laboratory and in amoderate risk containment facility. Until cloned segments ofwarm-blooded vertebrate DNA are completely characterized,they should continue to be maintained in the most restrictedvector-host system in moderate risk containment laboratories;when such cloned segments are characterized, they may bepropagated as suggested above for purified segments ofvirus genomes.

Unless the organism makes a product known to be danger-ous (e.g., toxin, virus), recombinant DNAs from cold-bloodedvertebrates and all other lower eukaryotes can be constructedand propagated with the safest vector-host system availablein low risk containment facilities.

Purified DNA from any source that performs known func-tions and can be judged to be non-toxic, may be cloned withcurrently available vectors in low risk containment facilities.(Toxic here includes potentially oncogenic products or sub-stances that might perturb normal metabolism if producedin an animal or plant by a resident microorganism.)

4. Experiments to be Deferred. There are feasible experi-ments which present such serious dangers that their perform-ance should not be undertaken at this time with the currentlyavailable vector-host systems and the presently availablecontainment capability. These include the cloning of re-combinant DNAs derived from highly pathogenic organisms(i.e., Class III, IV, and V etiologic agents as classified by theUnited States Department of Health, Education and Welfare),DNA containing toxin genes, and large scale experiments(more than 10 liters of culture) using recombinant DNAs thatare able to make products potentially harmful to man,animals, or plants.

IV. IMPLEMENTATION

In many countries steps are already being taken by nationalbodies to formulate codes of practice for the conduct of ex-periments with known or potential biohazard.tt,T$ Untilthese are established, we urge individual scientists to use theproposals in this document as a guide. In addition, there aresome recommendations which could be immediately anddirectly implemented by the scientific community.

A. Development of safer vectors and hosts

An important and encouraging accomplishment of the meetingwas the realization that special bacteria and vectors whichhave a restricted capacity to multiply outside the laboratorycan be constructed genetically, and that the use of theseorganisms could enhance the safety of recombinant DNAexperiments by many orders of magnitude. Experimentsalong these lines are presently in progress and in the nearfuture, variants of X bacteriophage, non-transmissible plas-mids, and special strains of Escherichia coli wvill becomeavailable. All of these vectors could reduce the potential bio-hazards by very large factors and improve the methodology aswell. Other vector-host systems, particularly modified strainsof Bacillus subtilis and their relevant bacteriophages andplasmids, may also be useful for particular purposes. Quitepossibly safe and suitable vectors may be found for eukaryotichosts such as yeast and readily cultured plant and animalcells. There is likely to be a continuous development in thisarea and the participants at the meeting agreed that improvedvector-host systems which reduce the biohazards of recom-binant DNA research will be made freely available to allinterested investigators.

B. Laboratory procedures

It is the clear responsibility of the principal investigator toinform the staff of the laboratory of the potential hazards of

tt Advisory Board for the Research Councils, "Report of theWorking Party on the Experimental TManipulation of the GeneticComposition of 'Micro-Organisms. Presented to Parliament bythe Secretary of State for Education and Science by Commandof Her Majesty, January 1975." London: Her Majesty's Sta-tionery Office, 1975, 23pp.At National Institutes of Health Recombinant DNA MoleculeProgram Advisory Committee.

Proc. Nat. Acad. Sci. USA 72 (1975)

1984 Summary Statement: Berg et al.

such experiments before they are initiated. Free and opendiscussion is necessary so that each individual participating inthe experiment fully understands the nature of the experimentand any risk that might be involved. All workers must beproperly trained in the containment procedures that are de-signed to control the hazard, including emergency actions inthe event of a hazard. It is also recommended that appropriatehealth surveillance of all personnel, including serologicalmonitoring, be conducted periodically.

C. Education and reassessment

Research in this area will develop very quickly and themethods will be applied to many different biological problems.At any given time it is impossible to foresee the entire rangeof all potential experiments and make judgments on them.Therefore, it is essential to undertake a continuing reassess-ment of the problems in the light of new scientific knowledge.This could be achieved by a series of annual workshops andmeetings, some of which should be at the international level.There should also be courses to train individuals in therelevant methods since it is likely that the work will be takenup by laboratories which may not have had extensive ex-perience in this area. High priority should also be given toresearch that could improve and evaluate the containmenteffectiveness of new and existing vector-host systems.

V. NEW KNOWLEDGE

This document represents our first assessment of the potentialbiohazards in the light of current knowledge. However, littleis known about the survival of laboratory strains of bacteriaand bacteriophages in different ecological niches in the outside

world. Even less is known about whether recombinant DNAmolecules will enhance or depress the survival of their vectorsand hosts in nature. These questions are fundamental to thetesting of any new organism that may be constructed. Re-search in this area needs to be undertaken and should be givenhigh priority. In general, however, molecular biologists whomay construct DNA recombinant molecules do not undertakethese experiments and it will be necessary to facilitate col-laborative research between them and groups skilled in thestudy of bacterial infection or ecological microbiology. Workshould also be undertaken which would enable us to monitorthe escape or dissemination of cloning vehicles and their hosts.Nothing is known about the potential infectivity in higher

organisms of phages or bacteria containing segments ofeukaryotic DNA and very little about the infectivity of theDNA molecules themselves. Genetic transformation of bac-teria does occur in animals, suggesting that recombinantDNA molecules can retain their biological potency in thisenvironment. There are many questions in this area, theanswers to which are essential for our assessment of the bio-*hazards of experiments with recombinant DNA molecules.It will be necessary to ensure that this work will be plannedand carried out; and it will be particularly important to havethis information before large scale applications of the use ofrecombinant DNA molecules is attempted.

The work of the committee was assisted by the NationalAcademy of Sciences-National Research Council Staff: ArtemisP. Simopoulos (Executive Secretary) and Elena 0. Nightingale(Resident Fellow), Division of Medical Sciences, Assembly ofLife Sciences, and supported by the National Institutes ofHealth (Contract NO1-OD-5-2103) and the National ScienceFoundation (Grant GBMS75-05293).

Proc. Nat. Acad. Sci. USA 72 (1975)