Uniform Nomenclature for Bacterial Plasmids: Proposal170 NOVICK ET AL. was then thought to integrate into theEsche-richiacoli chromosome). However,it sooncame to be used as a generic

BACTERIOLOGICAL REVIEWs, Mar. 1976, p. 168-189Copyright C) 1976 American Society for Microbiology

Vol. 40, No. 1Printed in U.S.A.

Uniform Nomenclature for Bacterial Plasmids: a ProposalRICHARD P. NOVICK,* ROYSTON C. CLOWES, STANLEY N. COHEN, ROY CURTISS III, NAOMI

DATTA, AND STANLEY FALKOWDepartment ofPlasmid Biology, The Public Health Research Institute of the City ofNew York, Inc.,

New York, New York 10016*; Division ofBiology, University ofTexas at Dallas, Richardson, Texas 75080;Department ofMedicine, Stanford University School ofMedicine, Stanford, California 94305; Department of

Microbiology, University ofAlabama in Birmingham, Birmingham, Alabama 35294; Department ofBacteriology, Royal Postgraduate Medical School, Hammersmith Hospital, London, W12 OHS, England;

Department of Microbiology, University of Washington, Seattle, Washington 98195

INTRODUCTION ............ ..................................... 168TERMINOLOGY ................................................. 169PLASMID DESIGNATIONS ............................................... 175STRAIN DESIGNATIONS ................................................ 176PLASMID GENE ABBREVIATIONSPhenotypes ......... ........................................ 177Genotypes ......... ........................................ 179

MOLECULAR REARRANGEMENTS ........................................... 181APPENDICESSuggested Designations for Plasmid-Carried Genes ....... ..................... 182Plasmid Incompatibility Groups ............................................. 185Summary of Recommendations of Demerec et al................................ 185

Summary of Recommendations on Plasmid Nomenclature ..... ................. 186ACKNOWLEDGMENTS........................................................ 186

LITERATURE CITED ............................................... 186

INTRODUCTIONDuring the rapid growth in plasmid research

that has taken place in recent years, a greatdeal of variability has developed in the use ofbasic terminology in publications in this area.In particular, many terms have multiple mean-ings, and there is often a profusion of terms fora particular entity or concept. The resultingconfusion is aggravated by the lack of a system-atic designation and referencing code for plas-mids. In consequence, a given plasmid may bereferred to by several different designations; agiven designation may refer to several plas-mids; and many plasmids are described withoutbeing given any identifying designation at all.A similar state of confusion exists for the desig-nation of plasmid-carried genes.There has been a growing awareness of this

problem for some years, and at a joint Japan-U.S. meeting on bacterial plasmids (Honolulu,November 1972) the problem was briefly dis-cussed, and there was found to be a generalconsensus that it would be worthwhile to at-tempt to develop a uniform system of plasmidnomenclature. The result of this would be, itwas hoped, to promote increased clarity in com-municating scientific results and thus to permitreaders not directly involved in plasmid re-search, as well as those involved, to compre-hend the published material with greater facil-ity.

At the Honolulu meeting, we were named asa working group with the task of preparing aproposal on nomenclature that, after consulta-tion with the group at large, as well as withother plasmid workers who did not happen to beat the meeting, might be published to serve as aset of guidelines in much the same manner aswas done for microbial genetics in general (23).We met in March 1973 and prepared a draft

proposal that was distributed in November 1973to more than 200 scientists, with a request forcomments and suggestions. More than 50 re-sponses were received, most containing helpfulideas, either on the entire proposal or on spe-cific items. Subsequently, the draft proposalwas presented and discussed during the Ameri-can Society for Microbiology Conference on Ex-trachromosomal Elements in Bacteria, held inJanuary 1974. We met after this conference toconsider the written and verbal responses to thedraft proposal and to incorporate these in asecond draft. Two additional meetings of thegroup were held in November 1974 and March1975 to consider and incorporate input fromvarious additional sources during the produc-tion of the present document. The proposal as itappears here has been considerably modifiedfrom the 1973 draft and reflects the broadestconsensus that we have been able to obtain.Our hope is that it is specific without beinginflexible and that it strikes a balance between

168

on Decem

ber 30, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

on D

ecember 30, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

on Decem


mbr.asm

.org/D

ownloaded from

http://mmbr.asm.org/



BACTERIAL PLASMID NOMENCLATURE

the need for full and accurate description on theone hand and brevity for recording purposes onthe other.The major intentions are (i) to clarify obscuri-

ties and resolve ambiguities in the presentusage of terminology in the hope of encourag-ing the adoption of one basic series of clearlydefined terms in place of multiple and oftenconfusing synonyms, and (ii) to recommenduniform systems of designations for plasmidsand for plasmid-carried genes in the hope ofavoiding multiple names for a single geneticelement and vice versa.

It cannot be stated too strongly that the pro-posal is intended as a set of guidelines ratherthan as a set of regulations. Although we be-lieve that a broad consensus on nomenclaturecan ultimately evolve through usage, we appre-ciate the difficulty of obtaining widespreadagreement among scientists working in thiscomplex and rapidly evolving field. It may beuseful to note in this connection that, althoughinitial acceptance ofthe 1966 nomenclature pro-posal by Demerec et al. (23) was less than com-plete, its major conventions have now gainedwide acceptance and have clearly led to clarifi-cation and standardization of the general no-menclature currently used in bacterial ge-netics.This proposal contains two main parts, one

on terminology and one on designations forplasmids and genes. The section on terminol-ogy is restricted to a list of some 30 terms whichwe believed to be particularly problematicaland in serious need of clarification. The sectionon plasmids and gene symbols is considerablymore important in that it attempts to provide asystematic and unambiguous reference systemfor coping with the large number of plasmidsand plasmid-borne hereditary determinants al-ready described in the literature and the stilllarger number that are certain to appear in thefuture.We consider this proposal as basically an am-

plification in one area of the recommendationsby Demerec et al. (23) on bacterial geneticsnomenclature, and we have attempted to buildupon these recommendations whenever possi-ble. Throughout its preparation, therefore, wehave utilized established conventions whereverpossible; we have endeavored to make use ofterms and abbreviations that were already inuse as long as they were clear and appropri-ately descriptive; and when we have found itnecessary to introduce new terms and abbrevia-tions, we have tried to make them as descrip-tive as possible of the phenomenon being con-sidered. Finally, we have attempted to make

our definitions, categories, and concepts broadenough to provide sufficient flexibility to en-compass new discovereries.

TERMINOLOGY(i) Plasmid (synonym: extrachromosomal

genetic element). A plasmid is a replicon thatis stably inherited (i.e., readily maintainedwithout specific selection) in an extrachromo-somal state. Naturally occurring plasmids ofprokaryotes are generally dispensable.

The term "plasmid" was introduced by Leder-berg (48) as a generic term for extrachromo-somal genetic elements, and it is proposed herein that sense. Thus, it replaces "episome" in thegeneric sense (but see below, item ii). Bacterialplasmids appear to constitute a well-differen-tiated taxon. Typical examples are the F plas-mid, colicinogenic plasmids, antibiotic resist-ance plasmids, toxinogenic plasmids, etc. Ineukaryotic systems, the genomes of mitochon-dria and plastids (but not the organelles them-selves) may be considered to be plasmids. Itshould be recognized that extrachromosomalnucleic acid molecules are not necessarily plas-mids; the definition implies genetic homogene-ity, constant monomeric unit size, and the abil-ity to replicate independently of the chromo-some. Thus, the heterogeneous circular deoxy-ribonucleic acid (DNA) molecules of Bacillusmegaterium (9) are not necessarily plasmids,nor are abortive transducing fragments orother abortive exogenotes. Genetic elementsthat are composed of multiples of a monomericplasmid unit (such as certain R [63] and Colplasmids 16]) and others that are composites oftwo or more plasmids (28, 54) are themselvesplasmids so long as they are stable in the extra-chromosomal state.The dividing line between plasmids and

phages (or viruses) is not sharp-for example,the P1 prophage and other similar prophagesare typical plasmids (44) even though the prod-ucts of their vegetative growth are typical bac-teriophages; the replicative forms of coliphagesfl, fd, M13, and their relatives are stably inher-ited in an extrachromosomal state that is char-acterized by the continued production of infec-tive phage particles (41). Nevertheless, typicalplasmids and phages (or viruses) are well differ-entiated; the in-between types could constituteevolutionary intermediates.

(ii) Episome. An episome is a genetic ele-ment that can replicate in either of two alterna-tive states: integrated into or independent of thehost chromosome (45).This definition was introduced to cover the

behavior of F, coliphage X, and ColEl (which

VOL. 40, 1976 169

on Decem


mbr.asm

.org/D

ownloaded from


170 NOVICK ET AL.

was then thought to integrate into the Esche-richia coli chromosome). However, it soon cameto be used as a generic term for all bacterialextrachromosomal elements. This latter usagehas, in the long run, turned out to be confusingor ambiguous for two reasons. First, becausethe episome definition crossed the taxonomicboundary between plasmids and temperatephages, it tended to obscure that boundary andthe very basic differences between these twoclasses of genetic elements. Second, because thedefinition included the requirement that theelement be able to integrate reversibly into thehost chromosome, it was too restrictive andimposed a considerable intellectual burden onthe early plasmid workers who needed a termto describe the elements they were discovering.In consequence, it was sometimes necessary toperform elaborate intellectual contortions to"prove" that a particular extrachromosomalelement was an episome. For example, whenconjugative R plasmids (see items iii and ixbelow) were first identified (2, 60), their simi-larity to F was clear enough; therefore, theywere assumed to be episomes. But they couldnot be shown to integrate. Satisfaction was ob-tained by the finding that they could mobilize(see item xiii) chromosomal genes (71), eventhough this was hardly proof of integration.The situation with the ColI plasmids was simi-lar (12, 61). Other "episomes" could not by anystretch of the imagination be shown to inte-grate -in particular, the evidence upon whichColE1 was originally classed as an episome wasshown to be spurious (13), and the integrationof ColEl has, in fact, never been observed.These various experimental discomforts even-tually resulted in the resurrection of the moregeneral term "plasmid" (13, 53, 55) after 11years of interment.

Since "episome" does not define a taxonomicentity but is instead descriptive of the behaviorof some members of at least two very differenttaxa, we recommend that it not be used as thetaxonomic label for self-replicating extrachro-mosomal elements in bacteria, i.e., as a syn-onym for plasmid (33, 56).

(iii) Conjugative plasmid. A conjugativeplasmid is a plasmid that can bring about thetransfer ofDNA by conjugation (see also itemsxii and xiii). Examples include the Fplasmid ofE. coli; the CoLIb, V, andB plasmids, and manyof the R plasmids of gram-negative bacteria.With the exception ofF (see item vi), naturallyoccurring plasmids identified on the basis oftheir ability to mediate conjugation should bereferred to simply as conjugative plasmidsrather than labeled with some other specificterm.

There are several synonyms for "conjugativeplasmid" currently in use. We suggest thatsome of these be avoided and others deempha-sized because of ambiguity and other reasons.Thus, "transmissible plasmid," which has beendefined as "capable of mediating conjugation,literally means "capable of being transmitted.""Transferon" and "conjugon" are neologisms forwhich there appears to be no special need.Furthermore, they do not readily permit theconstruction of antonyms. "Sex factor" and "fer-tility factor" are ambiguous in that there isoften uncertainty about whether transfer of thefactor itself is inferred or that of other geneticmaterial in the cell, especially chromosomalgenes, or both. These forms also do not readilylend themselves to the construction of anto-nyms. "Factor" is better abandoned when theproperties of the object become better known.The general use of other terms such as "con-jugal fertility factor," "conjugation factor,""autotransfer factor," and "infectious plasmid"is discouraged for one or more of the abovereasons. (However, it is recognized that theneed for variety requires the use of synonyms,so that there are times when some of theseterms may be useful.)

(iv) Nonconjugative plasmid. A nonconjuga-tive plasmid is a plasmid that cannot bringabout the transfer ofDNA by conjugation.The term defective conjugative plasmid has

been used in reference to naturally occurringnonconjugative plasmids that appear to be ge-netically related to conjugative ones (17, 52).This usage is confusing; naturally occurringnonconjugative plasmids should be referred toas defective conjugative plasmids only if theyhave some demonstrable remnant of a conjuga-tive system.

(v) Cryptic plasmid. A cryptic plasmid is aplasmid to which no phenotypic traits havebeen ascribed.

Strictly speaking, the presence of extrachro-mosomal circular DNA is in itself a phenotype.However, this situation is operationally suffi-ciently far removed from the usual sense ofphenotype that it does not flaw the definition.Cryptic plasmids are not to be confused withcryptic prophages; the latter are defective pro-phages that do not express immunity (27).

(vi) F plasmid. F plasmid is the prototype"fertility factor" responsible for conjugation inthe K-12 strains. This term was used by Leder-berg et al. (49), by Cavalli-Sforza et al. (10), andby Hayes (32) in their early studies of bacterialmating. The letter F refers specifically to thisplasmid and thus should not be used to refer toother naturally occurring conjugative plas-mids.

BACTERIOL. REV.

on Decem


mbr.asm

.org/D

ownloaded from



(vii) F' plasmid. An F' plasmid is an F de-rivative incorporating a segment ofthe bacterialchromosome.

This term supercedes "F' factor." An orga-nism in which an F' plasmid has arisen (66) andwhich carries a chromosomal deletion corre-sponding to the segment incorporated by theplasmid is termed a primary F'-containingstrain, and an organism into which an F' plas-mid has been transferred is termed a secondaryF'-containing strain (7). The differentiationinto primary and secondary is intended to em-phasize a situation (the former) in which theplasmid generated may be essential to thegrowth of the particular organism in which itarose.A nomenclature system for F' plasmids is in

existence (50) which suggests that new F' plas-mids be given serial numbers. It is our recom-mendation that F' plasmids be identified by apair of initials as well as by a number (see thesection, Plasmid Designations). The publishedF' numbering system recommends designatingderivatives of a particular F' plasmid by a suf-fix number. We suggest instead that all deriva-tive plasmids be given new plasmid numbersand that suffixes be avoided for reasons givenin the section, Plasmid Designations. See alsothe discussion of this point in Demerec et al.(23).Note that the definition of F' given in De-

merec et al. is inconsistent with the above defi-nition in that it defines F' as a strain carryingan F' plasmid. In practice, this definition hasnot been widely adopted. Inasmuch as the sym-bol F' has generally (and in our view, appropri-ately) been applied to the plasmid rather thanto the strain, it is preferable to use F'+ or F'-containing to refer to the state of harboring anF' plasmid. These usages are consistent withthe distinction between F and F+ originallymade by Demerec et al. (23), in which the for-mer refers to the plasmid and the latter tostrains carrying it.By analogy, the prime symbol appended to

any plasmid designation could be used infor-mally to signify incorporation of a segment ofbacterial chromosome into that plasmid. Thisshould not, however, generate numbering sys-tems for "prime" plasmids separate from thosefor other plasmid derivatives (see section, Plas-mid Designations).

(viii) Hfr. Hfr is the state of harboring aconjugative plasmid that is integrated into thechromosome and consequently is able to pro-mote oriented chromosomal transfer to suitablerecipients.

In general, Hfr strains transfer certain chro-mosomal segments at a greater frequency than

donor strains carrying the same plasmid in theautonomous state. However, it is our view thatthe basic features of the Hfr state are integra-tion and oriented transfer, and so we haveomitted from the definition any requirementinvolving frequency. In addition, we haveframed the definition to include integration andoriented chromosomal transfer involving conju-gative plasmids other than F so as to emphasizethe fundamental similarity of these plasmidsand of the conjugative process they bring about.

(ix) Resistance plasmid (R plasmid). An Rplasmid is a plasmid that carries genetic infor-mation for resistance to antibiotics and/or otherantibacterial drugs. OR plasmid" supercedes 'Rfactor" (see above, item iii) and OR (or r) deter-minant" (in the sense of a nonconjugative plas-mid carrying resistance genes).We have had reservations about formalizing

the elevation to taxonomic rank ofcertain char-acters carried by plasmids merely because theyconstitute salient phenotypes (see section, Plas-mid Designations). However, since two ofthem as such and at least to define them asclearly as possible. It should be noted that plas-mids initially identified as R or Col, etc., arefrequently found later to carry additionallyother types of genes, and so their designationsultimately become quite arbitrary.tionally other types of genes, and so their desig-nations ultimately become quite arbitrary.This being the case, it seems logical to frame

the definition of "resistance plasmid" so that itconforms at least with the general clinical sig-nificance of plasmid-linked resistance. Thus,the letter "R" (or "r") in reference to plasmidsimplies naturally occurring resistance to anti-biotics and other antibacterial drugs ratherthan resistance in general (such as to radiation,bacteriocins, phages, etc.). We recognize thatthe line between antibacterial drugs and otherantibacterial substances may not always besharp; however, we believe that generalizationto all resistance would destroy whatever useful-ness the term may have.

In connection with this definition it is recom-mended that "determinant" (as in "r determi-nant") not be used to refer to an entire plasmidor to a polyfunctional plasmid segment, since adeterminant is a gene or a contiguous set ofgenes specifying a single phenotypic character.

Finally, it is intended that "R plasmid" beused in an informal sense rather than to definea formal taxon. (Other informal classes of plas-mids are outlined in the section, Plasmid Desig-nations.)

(x) Resistance donor (R donor). A resist-ance donor is a (bacerial) strain that is capableof transferring resistance genes, usually by

VOL. 40, 1976 171

on Decem


mbr.asm

.org/D

ownloaded from


172 NOVICK ET AL.

conjugation, to a suitable recipient.Many natural isolates of gram-negative bac-

teria are capable of transferring resistancegenes to other organisms. Some of these carry asingle conjugative plasmid in which the resist-ance genes are incorporated. These have beenreferred to as class I R factors (4) and as plasmidcointegrates (14). Others carry several plas-mids, including one or more conjugative plas-mids and one or more nonconjugative plasmidsthat determine resistance and are mobilized(see item xiii) by the former. These have beenreferred to as class II R factors (4) and as plas-mid aggregates (14). In the absence of informa-tion on the plasmid content of such naturallyoccurring strains, they should be referred tosimply as resistance donors, with the implica-tion that plasmids are involved. Although con-jugation is the usual mode of resistance trans-fer, especially in gram-negative bacteria, spon-taneous transduction appears to be the rule inStaphylococcus aureus (58); a strain capable oftransferring an R plasmid by spontaneoustransduction should also be considered a resist-ance donor.The term "RTF" has been used to refer to the

agent responsible for transfer of resistancegenes by conjugation in strains that we definehere as R donors. For reasons given below (seeitem xv) we suggest that the use of "RTF" beavoided. In cases where a plasmid separatesinto conjugative and nonconjugative compo-nents, these should be referred to simply assuch.

(xi) Col plasmid. A Col plasmid is any plas-mid that carries genetic information for theproduction of a colicin. This term supercedes'Col factor" (see item iii).

(xii) Bacterial conjugation. Bacterial conju-gation is the process of genetic exchange be-tween bacteria, dependent on cellular contact,in which genetic material is transferred fromone organism (the donor) to another (the recip-ient).We suggest that "donor" and "recipient" be

used to specify bacterial mating types and that"male" and "female" generally be avoided be-cause they refer only to the presence or absenceof F; a strain carrying any conjugative plasmidcan express the donor phenotype.

(xiii) Mobilization. Mobilization is the proc-ess by which a conjugative plasmid bringsabout the transfer of DNA to which it is notstably and covalently linked (whether or nottransient linkage is established during suchtransfer is still unclear and in any case irrele-vant to this definition).Examples of mobilization are the transfer by

autonomous F or other conjugative plasmids ofchromosomal segments or of other plasmids

such as ColEl (29) and nonconjugative R plas-mids (3). Where a conjugative and a nonconju-gative plasmid are joined to form a stable coin-tegrate (see item xxv), the transfer of the latterby the former should not be referred to as mobi-lization, nor should complementation betweentra- (see item xv) or other defective conjugativeplasmids. It should be noted that this is apurely operational definition and implies noth-ing about mechanism; there may well be sev-eral mechanisms involved.

It should be noted that a particular plasmidcan legitimately be referred to as non-mobiliza-ble or non-transferable only in reference tothose individual conjugative plasmids withwhich it has been tested.

(xiv) Transconjugant. A transconjugant is abacterial cell that has received genetic materialfrom another bacterium by conjugation. Atransconjugant should be referred to as a re-combinant only if the transferred genetic mate-rial has been inserted into a preexisting repli-con in the recipient. If the transferred materialis perpetuated per se as a plasmid, then the cellis ordinarily referred to as a plasmid transcon-jugant. If such material is not a plasmid anddoes not become incorporated into a residentreplicon, it will fail to replicate and will be lostthrough dilution. In this case, the bacteriumwill ordinarily be referred to as an abortivetransconjugant.We suggest "transconjugant" here because it

conveys the sense of transfer. Other alterna-tives such as "conjugant" (analogous to trans-formant), "conjugatant" (analogous to trans-ductant), and "exconjugant" (currently in use)imply merely that the cell has participated inconjugation, either as donor or recipient.The term transcipient is currently in wide

use, but has at least two different meanings: (i)the consequence of conjugational transfer ofDNA, or (ii) the result of DNA transfer by anymeans. In parallel with transformant andtransductant, the respective consequences oftransformation and transduction, we suggestan equivalent term (i.e., "transconjugant") for arecipient of genetic information via conjuga-tion. Thus, we recommend that "transcipient"be used in the nonspecific sense, i.e., as in iiabove (although we find it difficult to imaginethat this usage will be common).

(xv) Transfer genes. Transfer genes arethose genes carried by a conjugative plasmidthat are responsible for the donor phenotype.The recommended abbreviation for such genesis tra. Other genes commonly carried by conju-gative plasmids but not essential for fertility(e.g., phage inhibition, incompatibility, etc.)should not be referred to as tra genes.The transfer genes of a conjugative R plasmid

BACTERIUL. REV.

on Decem


mbr.asm

.org/D

ownloaded from



have been referred to collectively as a resist-ance-transfer factor, or RTF. A conjugativeplasmid capable of transferring a nonconjuga-tive R plasmid has also been referred to as anRTF. Because of this ambiguity and becausethe designation RTF implies that there is anexclusive class of conjugative plasmids capableof transferring resistance genes, we urge retire-ment of "RTF."

(xvi) Superinfection inhibition. Superinfec-tion inhibition is an operational umbrella termreferring to interference with the entry or es-tablishment of an entering plasmid by a resi-dent plasmid. This term supercedes 'superin-fection immunity" which, by analogy with pro-phage immunity, should be specifically re-stricted to inhibition of replication. Among themechanisms that may be involved in superin-fection inhibition are entry exclusion (see be-low), DNA restriction, incompatibility, and theinhibition of transcription or translation.

(xvii) Entry exclusion. Entry exclusion isinterference by a resident plasmid with theentry ofgenetic material via conjugation (56).This phenomenon has also been referred to as

surface exclusion (1). Our preference for "entryexclusion" is based upon the fact that this iswhat is usually measured experimentally; "sur-face exclusion" is too specific in its implicationof mechanism. Moreover, there is historicalprecedent for the use of "entry exclusion" (56).

(xviii) Segregation. Segregation is the sepa-ration of independent, usually homologous ge-netic elements (e.g., replicas) during cell divi-sion. Failure of normal segregation to takeplace should be referred to as defective segre-gation. Distribution of two or more (usuallynonhomologous) genetic elements to the sameprogeny cell has usually been referred to ascosegregation (literally: "separation together").Since this usage has an uncomfortable aura ofself-contradiction about it, we suggest that it bereplaced by co-distribution.

"Segregation" has been used rather loosely torefer to a variety of phenomena. Some of theseusages are contradictory to others, and somewould better be supplanted by other terms. Wepropose that "segregation" be confined to thetwo usages listed below, and that the usageslisted in (iii) be abandoned: (i) separation ofhomologous genetic elements during meiosis indiploid cells; and (ii) separation of sister chro-matids during mitosis and, by extension, sepa-ration of any two independent genetic elementsduring cell division in prokaryotes. (iii) Withrespect to plasmids, "segregation" has beenused improperly in at least three ways: (a) torefer to separation of a plasmid into two ormore independent replicons. This phenomenonshould be referred to simply as separation,

which is the better term in our view, since itdoes not imply distribution to progeny cells. (b)"Segregation" has been used to refer to loss ofgenetic material where the lost material is notrecoverable as part of a functional genetic unit.This loss is more accurately referred to as "dele-tion" -it is presumably a process analogous tochromosomal deletion and its greater frequencyand greater relative extent in plasmids is pre-sumably a consequence of the fact that most ofthe plasmid genome is nonessential. (c) "Segre-gation" has also been used to refer to loss of anentire plasmid. This is a contradiction of defini-tion (ii) above; to reconcile this contradiction,one ought properly to speak of loss of an entireplasmid as the result of defective segregationrather than as a segregation event. Thus, mu-tants in which an entire plasmid is lost atgreater than normal frequency are termed"Seg- mutants."

(xix) Plasmid incompatibility. Plasmid in-compatibility is the inability of two differentplasmids to coexist stably in the same host cellin the absence of continued selection pressure.One may speak of incompatibility only when

it is certain that entry of the second plasmidhas taken place, and where DNA restriction isnot involved.Groups of plasmids that are mutually incom-

patible with one another have been variouslyreferred to as incompatibility or as compatibil-ity classes or types. Since the members of sucha group are mutually incompatible, it seemsmore logical to refer to it as an incompatibilitygroup, class, or type.

(xx) Conjugative (or donor) pili. Conjuga-tive pili are conjugation-specific hairlike ap-pendages of bacteria (8).

Conjugative pili have also been referred to as"fimbriae" (30). However, the former term ispreferable on phonetic grounds and, moreover,has some historical precedence.

(xxi) Fertility inhibition. Fertility inhibi-tion is inhibition by one plasmid of conjugativepilus synthesis or of conjugation mediated byanother plasmid when both are present in thesame cell (see Appendix 1 for suggested pheno-typic notations).Most plasmids that exhibit fertility inhibi-

tion also repress synthesis of their own conjuga-tive pili.

(xxii) Donor-specific phages. Donor-specificphages are those that infect only strains carry-ing conjugative plasmids; in all cases so farknown, the role of the plasmid involves entry ofthe phage genome (although not necessarily viaa conjugative pilus).These phages have heretofore been referred

to as male-specific or pilus-specific phages, andit is suggested that these terms be abandoned

VOL. 40, 1976 173

on Decem


mbr.asm

.org/D

ownloaded from


174 NOVICK ET AL.

as general references-the former for the rea-son given in item xii, the latter for the reasonthat penetration by means other than attach-ment to conjugative pili is possible among suchphages. There is, of course, no objection to theterm "pilus specific" if attachment of phage topili has been demonstrated.

(xxiii) Homogenic. Two genetic elements arehomogenic with respect to one another if theyare descended from a common ancestor by aknown sequence of steps.

For example, all mutant derivatives of a par-ticular genetic element should be consideredhomogenic to one another; descendants thatcontain gross rearrangements of gene sequenceshould likewise be considered homogenic solong as no new genetic material has been added(see item xxiv). Where a deletion has occurred,the element carrying the deletion is still homo-genic to its parent. An acceptable synonym iscongenic; "isogenic" is in common use in thesense in which "homogenic" is defined above;however, as it literally means "identical," wethink it is sufficiently misleading in that con-text to warrant the recommendation that itshould be used only in reference to direct repli-cas of a given genome.

(xxiv) Heterogenic. Two genetic elements areheterogenic with respect to one another if theyare not known to have common ancestry.The intention here is to provide an opera-

tional rule ofthumb to facilitate the handling ofsimilar genetic elements when ancestry cannotbe documented. Thus, coliphages 4x174 andS13 are heterogenic, as are coliphages ft andM13, f2 and MS2, staphylococcal plasmids pI524and pI258, enteric plasmids R1 and R100, etc.,even though members of these pairs are quitesimilar to one another.A recombinant between two heterogenic ele-

ments is considered homogenic to each of itsparents with respect to that portion of its ge-nome derived therefrom. Thus, the F plasmid ishomogenic to F gal, coliphage X is homogenic toX gal, and F gal and X gal may be homogenic toeach other with respect to the gal region carriedby each but are obviously heterogenic with re-spect to the F and X portions.(xxv) Cointegrate. A cointegrate is a natu-

rally occurring genetic element composed oftwoor more complete replicons in covalent linearcontinuity where the component replicons areknown to be capable of physically independentreplication. (See reference 14.)

This term is intended to imply that the coin-tegrate has an individually separate replicatorregion corresponding to each of its componentreplicons. It thus refers to chromosomes withintegrated prophages or plasmids (including

BACTERIOL. REV.

the chromosomes of Hfr strains) as well as tocertain composite plasmids. It does not refer toplasmids that have simply acquired one ormore new genes (these are heterogenic recom-binant plasmids until proved otherwise), nordoes it apply to plasmids that are made up oftwo or more regions with distinct functions.Two interesting recently described examples

of cointegrates are certain naturally occurringR plasmids that can replicate from at least twoalternative origins and can reversibly separateinto two (or more) individual plasmids (16, 54),and an in vitro-constructed composite of entericplasmids pSC101 and ColEl that uses one or theother of its component replicons according towhether external conditions favor one or theother (73).

Generally, two sets of terms are in use refer-ring to the formation and breakdown of cointe-grates. Integration and excision are used pri-marily to refer to interactions of smaller repli-cons with the chromosome, whereas cointegra-tion and separation would be preferable termsfor the corresponding events involving two ormore replicons of similar sizes.

(xxvi) Plasmid chimera, or chimeric plas-mid. A plasmid chimera is a recombinant plas-mid derived from two parental genetic elementsobtained from organisms that are ordinarilyunable to exchange genetic information. Plas-mids constructed in vitro or in vivo and con-taining DNA from organisms that can ordi-narily exchange genetic information should bereferred to as hybrid plasmids. A general termsuch as composite plasmid may be used to de-fine a stable recombinant plasmid constructedin vitro that contains DNA from organisms thatmay or may not be able to exchange genetic in-formation ordinarily. "Composite plasmid" is ageneral term, and as such refers to elementsthat may or may not be composed oftwo or moreseparate replicons.

(xxvii) Translocation (or transposition) se-quence. A translocating sequence is a well-defined genetic element, usually ofconstant size,that translocates intact from one genetic locus toanother. Insertion sequences are translocationsequences that have been identified on the basisofthe polar mutations that they cause at the siteof insertion.

Historically, the first such elements identi-fied in bacteria were termed "insertion se-quences" (51, 70). These, of which four havebeen described to date (IS1, IS2, IS3, and IS4),are characterized as different on the bases oflength and nucleotide sequence. So far, theyhave not been found to carry identifiable ge-netic markers.

Subsequently, it has been found that certain

on Decem


mbr.asm

.org/D

ownloaded from



plasmid-c-arried resistance genes undergotranslocation (38, 40, 46). Since these are genet-ically marked, it has been possible to observethe translocation event experimentally. Thetranslocation sequences were found to have in-verted repeat nucleotide sequences homologousto the previously described insertion sequences,and so it has been inferred that the latter causemutation by translocating. Hence, we suggestthat the class designation be "translocation se-quences" and that "insertion sequences" be con-sidered a subclass.Translocation sequences have been referred

to as "transposons" (38) and as "translocons"(40). We question the utility of coining a neolo-gism for this type of genetic element.

Translocation sequences have been abbrevi-ated "Tn" and given letters to indicate theirgene locus (e.g., TnA has been used to indicatea translocation sequence carrying ampicillinresistance [40]). Since "A" refers to the ampicil-lin resistance phenotype, it would be preferableto use the accepted phenotypic notation, Ap, sothat the recommended notation would beTnAp. Inasmuch as more than one such ele-ment may exist, it is suggested that uniqueidentifications be appended, referring to thegenome where the translocation sequence wasfirst identified (e.g., RP4TnAp; however, moreinformal notations may be used in context, solong as the reference is clear; see the sections,Plasmid Gene Abbreviations and MolecularRearrangements).

(xxviii) Plasmid copy number. The plasmidcopy number is the number of molecules of aspecific plasmid per genome equivalent or perhost cell (synonym: plasmid multiplicity).Terms such as multicopy plasmid and oligo-

copy plasmid may be useful in reference toplasmids normally present in many or few cop-ies per genome or cell, respectively. Since thereis a continuum of plasmid copy numbers, theseterms are of limited utility and should be usedonly in a general descriptive sense. The terms"stringent" and "relaxed" replication controlshould not be used as synonyms for oligocopyand multicopy plasmids, respectively (see itemxxix). An organism carrying two or more differ-ent plasmids should be referred to as a di-, tri-,tetra-, etc., or multiplasmid organism or strain.

(xxix) Relaxed and stringent control ofplasmid replication. There are at least fourpossible parameters by which plasmid replica-tion control may be characterized: (i) ability orinability to replicate in the absence of chromo-some replication; (ii) plasmid multiplicity andits variability; (iii) continuity or punctuality ofplasmid replication with respect to the cell divi-sion cycle; and (iv) randomness or regularity of

plasmid replication with respect to choice ofmolecule.The analysis ofplasmid replication control is

a rapidly developing area; although no plasmidhas so far been fully characterized with respectto all four of these parameters, the terms "re-laxed" and 'stringent" plasmid replication con-trol have been used variously to refer to one ormore ofthe four (14, 26, 62). Ultimately, it mayturn out that these terms are not very useful;however, we believe that it is necessary at leastto suggest uniform interim definitions so thatwhen they are used they will have a specificmeaning.

Therefore, we suggest that criterion i aboveshould be the one upon which the definitionsare based, because it is relatively easy to applyand would appear to define a dichotomy ratherthan a continuum. Hence:

(xxixa) Relaxed control of plasmid replica-tion. Relaxed control should be used in referenceto plasmid replication that is not obligatorilycoupled to chromosome-replication.

(xxixb) Stringent control of plasmid replica-tion. Stringent control should be used in refer-ence to plasmid replication that is obligatorilycoupled to chromosome replication.Even here, it must be emphasized that the

terms are useful only as informal phenotypicdescriptions of the behavior of a particular plas-mid in a particular host; they should not beapplied without qualification.

PLASMID DESIGNATIONSIn general, published designations of plas-

mids should be retained except where they areduplicated or ambiguous (see below). There-fore, our recommendations are primarily forthe naming of new and derivative plasmids.The basis of these recommendations is the useof uniquely identified numerical series analo-gous to the numerical seriation of strains rec-ommended by Demerec et al. (23) (recommen-dation 9). To distinguish plasmid designationsfrom bacterial strain designations, we suggestthe letter "p" as a prefix, in addition to the twoletters referring to the naming laboratory.Thus, pXY1234 = plasmid no. 1234 from thecollection of the investigator using the identify-ing initials XY.Any genotypic modification of a plasmid

would engender both a new strain number anda new plasmid number; all derivative plasmids(mutants, deletions, recombinants, etc.) gener-ated by a particular laboratory as well as allnaturally occurring ones isolated by that labo-ratory should be assigned to a single series. It isurged that this rule be applied to derivatives ofplasmids whose designations do not presently

VOL. 40, 1976 175

on Decem


mbr.asm

.org/D

ownloaded from


176 NOVICK ET AL.

conform to this recommendation, as well as tothose whose designations do conform.The addition of hyphenated or other suffixes

to plasmid numbers to indicate derivation isundesirable for several reasons: it inevitablygenerates duplications; it endows plasmid num-bers with generic significance; it generatescomplications such as how to designate recom-binants, etc.; and it makes the numbers forderivative plasmids increasingly unwieldy.Changes in plasmid number would not be

made after simple transfer unless there weresome reason to suspect genotypic change ac-companying the transfer, for example, trans-duction by a phage whose genome is smallerthan the plasmid.Some investigators may wish to use letters

other than "p" to signify "plasmid," e.g., "r" for"resistance plasmid." This is acceptable so longas it is clear. We would urge, however, thatduplication of numbers in a given series beassiduously avoided, i.e., a laboratory shouldnot have both pXY1001 and rXY1001. There isno reason, however, why the same initialsshould not be used for the plasmid series andfor the bacterial strain series of the same labo-ratory. In the event that duplications arise inthe assignment of names, numbers, etc., it isrecommended that these be handled by thestandard taxonomic practice, i.e., the first pub-lication gets priority and the subsequent one(s)must be changed so as to eliminate the duplica-tion.

It is extremely important to keep careful his-tories of plasmids. As hidden molecular rear-rangements may take place from time to time,one would like to be able to trace pedigrees inorder to pinpoint, where possible, the occur-rence of such changes. In this connection, itshould be emphasized that it is highly desirablein publications to quote the original referencefor the source of a plasmid or bacterial strainand its derivation, not merely the worker fromwhom the culture was obtained.Plasmid-containing cultures received from

other laboratories should be treated as follows.(i) If the plasmids are prototypes with well-established names or stock numbers, the lattershould be retained and used in publications. (ii)If they have stock numbers that conform to theabove recommendation, these should likewisebe retained. (iii) If they have numbers that donot conform, they should be given conformingnumbers in consultation with the sending labo-ratory and in publications both old and newnumbers should be listed. (iv) Numbers givento derivative plasmids should conform to theseriation of the laboratory that produces the

derivatives. (v) In general, when using thissystem of seriation, it will be useful in textualmaterial to indicate parenthetically the salientfeatures of a plasmid or strain so designated.Thus, "pXY1234 (a 40-megadalton conjugativeR plasmid) was introduced...," etc.One of the problems that stimulated this se-

ries of recommendations was the duplication ofplasmid designations due to the parallel growthof different series bearing the same numbersand the same prefix letter (usually R). It isurged that the laboratories that have initiatedthese series remedy the situation by changingthem, where possible, to conform to these rec-ommendations. To this end, the prefix "p" fol-lowed by identifying initials should be ap-pended to the extant plasmid designation, re-taining the original number: thus, R124 mightbecome pXY124. Of utmost importance in re-cording such changes is a list of synonyms inthe published tabulations of strains.For general understanding and readability in

context, informal plasmid designations havebeen useful, generally referring to some salientfeature of the plasmid, e.g.: antigen plasmid,bacteriocinogenic plasmid, chloramphenicol re-sistance plasmid (not "chloramphenicol plas-mid"), degradative plasmid, enterotoxinogenicplasmid, hemolysin plasmid, mutator plasmid,pigment-inducing plasmid, resistance plasmid,toxinogenic plasmid, tumorigenic plasmid, andvirulence plasmid.Although such informal designations are

helpful and their use should be continued, theyshould not be used as taxonomic classes and,especially, should never be used in place offormal numerical designations. Abbreviationsfor these phenotypes should be kept as simpleas possible -preferably a capital and two lowercase letters, e.g., Deg, Ent, Tox, etc., andshould be treated in context as trivial nameswith clear reference to unique numerical desig-nations.

STRAIN DESIGNATIONSThis recommendation refers to the designa-

tion of plasmid-carrying strains only, and to thetabulation of such strains for publication.Strain numbers for plasmid-carrying strainsshould be formulated as recommended by De-merec et al. (recommendation 9, see Appendix3). Since a strain number constitutes a com-plete genotype, the introduction of a plasmidengenders a new strain number for the result-ing host-plasmid complex as does any genotypicmodification of an existing host-plasmid com-plex. In listings of derivative strains, ancestralstrain numbers may be indicated as part of the

BACTERIOL. REV.

on Decem


mbr.asm

.org/D

ownloaded from



genotype. This practice greatly simplifies com-plex genotypes and serves also to clarify pedi-grees. Plasmids and prophages should each beenclosed in a separate parenthesis and, where astrain has been cured of plasmid or prophage,this information should be included in the geno-type. A series of K-12 derivatives and theircorresponding genotypes might be listed as fol-lows:K-12 E. coli prototrophC600 K-12(F-(-) thr leu tonA lacY thiXY1000 C600(F'155)(pXY1234)XY1001 XY1000(pXY1234-)In the last strain, if the listing were "C600-(F'155)," there would be no indication that thestrain had once carried, and been cured of,pXY1234.Complex strain lists will often be greatly

simplified by listing host strains and plasmidsseparately, with the implicit understandingthat the construction of plasmid-carrying deriv-atives is without difficulty unless so stated.

It should be noted that there is a naturaltendency for designations of particularly impor-tant strains (and also of plasmids) to becomegeneric types. Some obvious examples of thisare the various (now generic) E. coli proto-types-the B, C, W, and K-12 strains and the Fplasmid, and at least one generation of K-12substrains-CR34, C600, W1485, etc. At somestage in this process, such designations, whichwere, after all, originally stock numbers usedto designate specific strains, gradually losetheir specificity and become used generically.This can sometimes give rise to ambiguity; thusthe notation K-12 is often used in place of K-12(F-) or K-12(X) so that one is unsure which ismeant. This is a generally undesirable practicebut one that is often difficult to inhibit, espe-cially when the formally complete designationsare unwieldy. If strains are listed as in theabove example, always with ready reference tocomplete genotypes, such ambiguities are auto-matically eliminated.

PLASMID GENE ABBREVIATIONSPhenotypes

The plasmid itself, strictly speaking, has nophenotype; rather, the plasmid genes are phe-notypically expressed by the strain carrying theplasmid. Thus, the word "phenotype" should betaken to mean the phenotype of the bacterialcell carrying the plasmid.Plasmid genes should be designated pheno-

typically according to convenience and in amanner designed to avoid ambiguity. There-

fore, in general, we recommend a capital plus alower case letter (or two where necessary). Alist of recommended phenotypic abbreviationsfor the commoner plasmid-carried genes is ap-pended (Appendix 1). This type of notationshould serve, in some cases, to distinguish plas-mid-carried genes from chromosomal loci giv-ing the same phenotype, since the latter mostusually have three-letter abbreviations (e.g.,Smr versus Strr for plasmid and chromosomalstreptomycin resistance phenotypes), and itwill also improve the clarity of listed plasmidphenotypes, since a plasmid carrying resistanceto ampicillin, kanamycin, streptomycin, andsulfonamide will be listed as "Ap Km Sm Su"instead of "AKSSu," as has usually been done.Although it seems useful to distinguish pheno-typically plasmid-linked from chromosomal re-sistance genes because both are common andbecause they usually have different modes ofaction, it is not our intention to set up a rigiddistinction between chromosomal and plasmidgenes in general by assigning a different type ofphenotypic abbreviation to each. Such a dis-tinction would appear to serve no useful gen-eral purpose and, in fact, a glance at Appendix1 will reveal many three-letter plasmid pheno-types.

In view of recent evidence suggesting that allor part of the biosynthetic pathways for certainantibiotics may be plasmid-linked, the questionarises of how to distinguish phenotypic nota-tions for synthesis of a particular antibioticfrom those for resistance to the same com-pound. Although we do not suggest any suchspecific notations, we suggest that synthesis bedenoted by an "S" prefixed to the usual pheno-typic notation for resistance, e.g., Skm woulddenote kanamycin synthesis.There is one conceptual difference between

our recommendation for phenotypic notations-and that of Demerec et al. (23). This differencestems from the fact that a variable and oftenrather large portion of the plasmid genome isnonessential for the existence of the plasmid asan autonomous replicon (the larger the plas-mid, the larger the nonessential regions). Con-sequently, there is great variability amongclosely related plasmids in the carriage of thevarious nonessential genes that determine thephenotype of the plasmid-carrying organisms.Therefore, we suggest that listing of plasmidgenes by phenotype should be taken to signifypositive expression of the phenotype involved.This is in contrast to the usual convention (23),which is to list only those alleles that differ bymutation from the wild type. In the plasmidnotation, however, the absence of a phenotypic

177VOL. 40, 1976

on Decem


mbr.asm

.org/D

ownloaded from


TABLE 1. Example ofplasmid recombinant notationPlasmid Phenotype Genotype Derivation

R100 Tc Cm Sm Su Tra tet+cat+aadA+sul+tra+ Naturally occurringR6 Tc Cm Sm Su Km-Nm Tra tet'cat'aadA+sulaphC'tra' Naturally occurringpXY550 Sms Tra- aadA2 traE8 Sms, Tra- mutant of R100pXY560 Rep(Ts) Tra- repB7 traB4 Rep(Ts), Tra- mutant of R6pXY570 Smr Rep+ Tra+ 550[traB+-tral+repB ]:560[traJ+- Recombinant, pXY550 x pXY560

traE+aadA +aphC+traK tet+cat'sul+

abbreviation implies absence of the expressedphenotype regardless of mechanism. An excep-tion to this would be in a listing of derivativesof a given plasmid for which the entire pheno-type (and/or genotype) need be given only forthe first listing of the plasmid; in listing deriva-tives, then, one need only identify those loci atwhich there are differences from the prototype(see Table 1). It is recommended that this proce-dure be followed as a rule; simply referring toan earlier publication for a complete plasmidphenotype (or genotype) is discouraged. In cer-tain cases it may be desirable to include theabbreviation for an absent gene for clarity oremphasis. In such cases a (-) superscript isnecessary. A (-) superscript may also be usedto indicate a sensitive or negative derivative ofa plasmid originally carrying the gene in itspositively expressed state. The use of subscriptsshould be avoided, since it creates difficulties intypesetting and so adds to journal publicationcosts.In some cases it may be useful to employ

other phenotypic modifiers. This should be doneaccording to the recommendations of Demerecet al. (23), as shown in the following example(note that whereas single-character modifierssuch as r, s, +, -, etc., appear as super-scripts, multiple-character modifiers are bestgiven in parentheses, as shown): Km(Ts) ApTc- Tra(Sus) would be the phenotype of an Rplasmid that is thermosensitive for kanamycinresistance, expresses ampicillin resistance, issuppressor sensitive for conjugational transfer(Sus means responsive to the action of a trans-lational suppressor), and is a tetracycline-sen-sitive derivative of a plasmid originally carry-ing tetracycline resistance. Note that the modi-fiers are not separated from the phenotypesthey modify, but the notations are separated byspaces.New plasmid characteristics should be ab-

breviated briefly and descriptively; e.g., Fbmight be used for a fibrinolysin marker. Newphenotypic abbreviations should be chosencarefully so as to avoid confusion with geneticor biochemical notations already in use. Forexample, Na would be better than Nad for nali-

dixic acid resistance because of the biochemicalmeaning of the latter.Analogue series. The existence of series of

similar but nonidentical genetic traits poses aspecial problem for plasmid nomenclature be-cause of the large numbers of genes involved,because of the especial importance of someof these genes in plasmid biology, and becauseplasmids carrying analogous genes may readilycoexist in the same cytoplasm, giving rise torecombinants, translocations, etc. A formalsolution for this problem in nomenclature isrecommended in the section, Genotypes, below.For phenotypes, a purely operational approachwould appear to suffice. The phenotypic abbre-viation for a gene should reflect the propertyby which the gene was identified. Thus, if a,8-lactamase gene was identified by ampicillinresistance, its phenotype is Ap; if it was identi-fied by penicillin G resistance, its phenotypeis Pc. In cases where it is necessary to differ-entiate, for example, two or more similarresistance genes specifying different cross-resistance patterns, compound phenotypicabbreviations may be useful. For example,Sm would refer to streptomycin resistance,Sm-Sp to streptomycin-spectinomycin resist-ance. In the latter case, the implication wouldbe that one knows that both resistances arespecified by a single gene.Phenotypic abbreviations usually refer to a

single gene or trait; however, any phenotypeas given need not be complete (e.g., additionalcross-resistances, known or unknown, mayexist).Among the analogue series, certain groups

of plasmid traits present special problems andneed to be dealt with individually. These areseries of traits where specificity is an essentialaspect of the phenotype. Three such series arethe determinants of bacteriocinogeny, incom-patibility specificity, and DNA restriction. Wemake specific recommendations for these three,the essence of which is the inclusion in thephenotypic notation of an indication of speci-ficity. It is hoped that this principle will alsobe used for other analogue series where speci-ficity is an essential aspect of the phenotype.

178 NOVICK ET AL. BACTERIOL. REV.

on Decem


mbr.asm

.org/D

ownloaded from



Bacteriocinogeny. The established nomen-clature for colicins (and other bacteriocins) isproblematic from the standpoint of this pro-

posal for two reasons. First, the usual abbre-viation, e.g., ColB, which was originally in-tended to refer to the entire plasmid thatencodes colicin B production, is also used torefer to the phenotype associated with colicinB production. Second, the phenotypic notation,ColB, may cause confusion if translated di-rectly to a genotypic notation (see below),inasmuch as the genotypic notation colB wouldconventionally signify "the B cistron of thecol locus" and not "the production of colicinB."To make the best of a difficult situation, we

suggest the following: (i) that the establishednomenclature for the bacteriocins themselvesbe retained (we offer no suggestion for theeventuality, for colicins, that all 26 lettersget used up); and (ii) that the phenotypes andgenotypes for bacteriocinogeny be abbreviatedin conformity with established practice (see Ap-pendix 1). It is our recommendation, therefore,that the established abbreviations (ColB, etc.)be retained as designations for the entire plas-mid and that the phenotypic and genotypic ab-breviations be different so as to eliminate theexisting ambiguity.

Incompatibility specificity. The phenotypicnotation for incompatibility should consist of"Inc" plus an indicator of specificity. "Inc" isfavored over "Com" since members of the samegroup are mutually incompatible. A list ofknown incompatibility types is given in Appen-dix 2. The existing incompatibility type desig-nations as listed in Appendix 2 are well estab-lished and should be retained. New classesshould be designated by "Inc" plus an arabicnumeral. For the enteric plasmids, this nu-

merical series should start with Inc26 sincethere are 25 established classes at the timeof this writing.Assignment of incompatibility classes should

be made in recognition of the fact that allplasmids that can be maintained in a givenhost species are capable of interacting withone another in vivo. For example, incompati-bility classes should be assigned to nonconjuga-tive plasmids from enteric species with refer-ence to those assigned to conjugative plasmidsfrom these organisms. In general, the assign-ment of a new incompatibility type should notbe made without a test against a completeset of the existing appropriate prototypes.

Restriction-modification. Genetic notationsfor restriction-modification systems have beencomplicated, confusing, and nonuniform. We

should like to propose for phenotypic notationsa modification of those commonly in use (5)(Table 2).

Antibiotic-inactivating enzymes. An ex-tremely important group of plasmid-linkedtraits consists of the antibiotic-inactivatingenzymes. These, in general, occur in a numberof similar but nonidentical forms. On one hand,it is rarely possible to say that two differentplasmids encode the very same enzyme, and,on the other hand, it is rarely possible to sub-divide analogous enzymes having the samemode of action into unambiguous groups (seebelow). Therefore, we recommend that no indi-cator of specificity be appended to the pheno-typic notation for (enzymatic) antibiotic resist-ance (see below for a recommendation ongenotypic notation for these traits).

GenotypesIn general, we suggest that recommenda-

tions 1, 2, and 3 of Demerec et al. (23; seeAppendix 3) be followed for the designation ofloci, genes, and alleles. We have implicitlyadopted recommendations 4 (which has to dowith the designation of plasmids as a whole)and 5 (which suggests that genotypes of plas-mid markers be assigned according to the samerules as those of chromosomal markers).In specifying plasmid genotypes in general,

the absence of any notation for a gene shouldbe taken to indicate the wild-type allele. Inthis, we follow the standard practice for chro-mosome genotype notation. In some cases, itmay be desired to include a notation empha-sizing that a gene is present. Here, a super-script should be used: (+) for the wild type,

MOLECULAR REARRANGEMENTSPlasmid recombinants. Recombinants be-

tween homogenic plasmids should be con-sidered as simple derivatives and designated

TABLE 2. Genetic notations for restriction-modification systems

Recognition sys- Notation in com- New recommendationtem mon use

Phage P1 rp,- Res-(Pl)mpl- Modi(Pi)

Plasmid RIP rR11, Res-(RII)rRII-mMI Res- Mod-(RII)

a An indicator of specificity should appear as aphenotype modifier, but only when necessary forclarity.

b Note that "RII" does not refer to a particularplasmid but rather to a restriction-modification sys-tem that recognizes and acts upon a particular nu-cleotide sequence-specific site.

179VOL. 40, 1976

on Decem


mbr.asm

.org/D

ownloaded from


180 NOVICK ET AL.

and (-) for a mutation where an allele numberhas not been assigned. As recommended byDemerec et al. (23), the designation of an allelesays nothing about the phenotype correspond-ing to that allele. Thus, modifiers such as ts(for thermosensitive), etc., should not be in-cluded in formal genotype specifications butmay be included informally in context fordescriptive purposes.The assignment of genotypes to members of

analogue series of plasmid genes involves aspecial problem that is not generally encoun-tered in dealing with other genomes and thatrequires a specific solution if the literature onplasmids is to be decipherable. This problemhas been stated clearly by Demerec et al. (23)who, however, despaired of its solution. Wequote here their passage on hybrid strains:

The system described above for desig-nating mutant loci and mutation sitespresents no problems as long as all strainsare derived from a single wild type. Asdiscussed under Recommendation 1, alocus is considered mutant if it differs fromthe corresponding locus in the arbitrarilychosen wild-type strain.Thus, a series of mutant loci have been

designated within strains derived from E.coli K-12, another series within strainsderived from E. coli B, still another withinstrains derived from Salmonella typhi-murium, and so on. But what is the geno-type of a hybrid strain, arising from a crossbetween wild-type E. coli K-12 and wild-type E. coli B? Some of its loci will bederived from one wild type, and some fromthe other. If K-12 were considered as thereference strain, the loci inherited from Bwould be mutant, and vice versa. Further-more, the genotype of the hybrid could notbe written until it was known from whichparent each locus was derived.Should it be possible to determine from

which parent a particular wild-type locuswas derived, a symbol could be devised toconvey this information. Most loci, how-ever, are likely to remain unidentified. Insome situations, e.g. when many newstrains are to be derived from a particularhybrid, it will be best to designate thehybrid itself as a new prototype straincomparable to a wild type.Thus, the problem of distinguishing among

analogous genes arises only where differ-ent genomes may interact directly with oneanother-which is very much the case here,

inasmuch as different plasmids, related or un-related, may occupy the same host cell at thesame time and complement one another orrecombine to produce hybrids. To deal withthis situation, a definitive means of differ-entiating among analogous genes carried bydifferent plasmids is required.

Ideally, one would like to give a single geno-typic designation to all analogous loci that arealleles and different ones to all that are not.However, it seems to us useless on theoreticalgrounds to attempt to make such distinctions,for, among naturally occurring analogues,there is really no dividing line between allelicgenes and similar but nonallelic ones. Never-theless, in conducting genetic analyses, onemust have a way of distinguishing, for ex-ample, the traA cistron of F from that ofpXY1234.Our recommendation, then, is that, where

possible, analogous genes should be given thesame genotypic symbol, but in genetic analy-ses, the specific, formal genotypic notation ofrecord should include an identification of theplasmid of origin (e.g., FtraA, pXY1234traA).Because of the unwieldiness of these notations,they should be used in context only wherenecessary to make the distinctions.With respect to antibiotic-inactivating and

other plasmid-coded enzymes, we recommendthat the genotypic notation be solely a reflec-tion of the mechanism of action of the enzyme(see Appendix 1). Where there is a series ofsimilar but nonidentical enzymes having thesame mode of action, such as the 3-lactamases,their genetic determinants should all carry thesame genotype, and individuals should beidentified by reference to their plasmid oforigin, as recommended above. Where a classi-fication of such enzymes has been developed(for example, on the basis of substrate speci-ficity [61a, 65a]), this should not be reflectedin the genotypes (or, for that matter, in thephenotypes), but should be simply indicated incontext as required, e.g., ". . . the RP4bla gene,encoding a type II 34-lactamase . . ." etc. Wherea plasmid specifies several enzymes constitut-ing all or part of a biochemical pathway, thephenotypic and genotypic notation will gen-erally reflect the function of the overall path-way, for example, tol and raf for toluene degra-dation and raffinose fermentation, respectively.As a rule of thumb, unless one is dealing

with mutations and direct plasmid interactionsrequiring the use of genotypes, it will be muchsimpler to refer to plasmid genes by theirphenotypes.

BACTERIOL. REV.

on Decem


mbr.asm

.org/D

ownloaded from



accordingly (see section, Plasmid Designa-tions). Composite or recombinant plasmidscontaining covalently joined segments fromtwo or more heterogenic genetic elements,however, require more elaborate designationsas follows. First, they should always be givennew sequential numbers as noted above (seesection, Terminology, item xxiv). Second, inspecifying the genotypes of such hybrids, it isnecessary to identify the origin, insofar as it isknown, of every gene. It is suggested that thisbe done with square brackets, as shown inTable 1. The hyphens between traB+ and traI+and between traJ+ and traE+ signify the inclu-sion in each case of a block of genes betweenthe two limiting loci according to the knownmap order of the tra operon. The notationsoutside the brackets need merely be sufficientto identify the plasmids in context. Genes ofuncertain parentage in such hybrids shouldbe listed outside the last bracket.

Kilobase (kb) coordinates are physical mapdistances in thousands of nucleotide pairs froman arbitrarily determined point of origin. Forcircular genomes, kilobase distances increasein the clockwise direction, based on an arbi-trary orientation of the map. Kilobase coordi-nates are determined from contour lengthmeasurements in conjunction with the identifi-cation of fixed topographical features of thegenome map. Where kilobase coordinates (68,69) are available from molecular mappingdata, these can also be included, e.g., F[20-67kb traA6]:R100[67/60'-19'/20 kb tetA2] wouldindicate an F-R100 recombinant in which theregion from 20 to 67 kb is from F, and theregion from 67 through 0 to 20 kb has beenreplaced by a segment from R100 with coordi-nates of 60' through 0' to 19'. The junctionpoints are at 67/60' and 19'/20 (the primednumbers refer to locations on the original R100genome and the unprimed numbers refer tolocations on the original F). The plasmidcarries the mutations traA6 in the F regionand tetA2 in the R100 region.

Deletions, insertions, transpositions, trans-locations, and inversions. These molecular-rearrangements constitute structural changesin genotype. Lacking evidence to the contrary,it is safest to assume that such events areunique and therefore to assign to each occur-rence a unique genotypic notation. Of especialimportance in plasmid genetics are two basicclasses of such rearrangements: insertions(including transpositions and translocations)and deletions. In recommending specific nota-

tions for these, we expand the established con-vention of using the Greek delta to designatedeletions by suggesting the Greek omega as adesignation for insertions.Complete genotypic notations for deletions

aid insertions would include numerical seria-tion and reference to the parent plasmid asfollows.

Deletions are conventionally identified by aGreek delta. Where only one gene is known tobe affected, an allele number is given, e.g.,AblaZ7 specifies that allele no. 7 in the blaZgene is a deletion. Where more than one geneis affected, the deletion is specified by a deltaand a number followed by a list of deletedgenes separated by hyphens and enclosed insquare brackets. Alternatively, this may bedone by specifying the first and last genes of adeleted region; however, the circularity ofplasmids may give rise to ambiguity here, andthis possibility should be borne in mind. Ifthe map is known, an en bloc deletion may bespecified by its first and last genes in clock-wise order or from left to right, separated by ahyphen. The delta indicates a single en blocdeletion. If the plasmid has suffered more thanone deletion, each should be indicated sepa-rately. If available, the kilobase coordinates ofdeletions and other salient features of theheteroduplex map may be specified as part ofthe genotype (68, 69). Three examples of plas-mids with molecular rearrangements and theircorresponding genotypes are:

pXY9109 FAtraA3 [63.1-64.0 kb]pXY1112 pI258A7[cad-asa]pXY2101 pSClOlQf4[0 kb:K-12hisA 1.5 kb]

The first is a Tra- F plasmid carrying the trA3allele, which is a deletion encompassing theregion from 63.1 to 64.0 kbs. The second is aderivative ofS. aureus plasmid pI258 with dele-tion no. 7 spanning the region from cad to asa.The third is a pSC101 derivative with an inser-tion at kb 0 of a 1.5-kb segment carrying the E.coli K-12 hisA locus. We have not included anyrecommendation for translocation sequencesand related genetic elements since the study ofthese is so new that such a recommendationwould be premature.

It is hoped that these guidelines for desig-nating molecular rearrangements will be use-ful for other naturally occurring rearrange-ments as well as for artificially constructedrecombinant DNA molecules.

VOL. 40, 1976 181

on Decem


mbr.asm

.org/D

ownloaded from


APPENDIX 1

Suggested Designations for Plasmid-Carried Genesa

Phenotypeb Genotypeb Effect

AcAkcAkAkAp [A, Amp]Asa [AsO42-]Asi [As032-]BiBtBtBtCamCbCdCeIcxCaa [ColA]Cba [ColB]Cda [ColD]Cel [ColEl]Ce2 [ColE2]Ce3 [ColE3]Cia (ColIa)Cib (ColIb)Cka (ColK)Cva (ColV)Cxa (ColX)Cm (C)

CoCpCrCxDm

Dm

Dm

Dm

EbEex [Sex, Sfx]EmEntExtFaFi+(X) [fi+)Fi-(X) [fi-)

Gm [G, Gk]GmGmGm[G, Gk]Hg [Hg2+]HlyHysInc [Com, Mc, Mcr]Km [K]Km [K]

acr

aacAc l

aacCaadCbla [pen, amp]asa

asibisaphA'aacAaacBcam

blaecadblaicx [clx, imm]caa

cbacdacea

cebcec

cia

cibckacva

cxa

cat [cam, cml]

cobblablablaaacA

aacB

aacC

aadBg

Ebreex [traS, sex]erm [ero, ery]entextfusfin

aacA [gat]aacBaacCaadBmer

aphA [kan]aacA [kat, kan]

Acridine resistanceAmikacin resistance (aminoglycoside 6'-N-acetyltransferase)Amikacin resistance (aminoglycoside 3-N-acetyltransferase)Amikacin resistance (aminoglycoside 4"-adenylyltransferase)Ampicillin resistance (,3-lactamase)Arsenate resistanceArsenite resistanceBismuth ion resistanceButirosin resistance (aminoglycoside 3'-phosphotransferase)Butirosin resistance (aminoglycoside 6'-N-acetyltransferase)Butirosin resistance (aminoglycoside 2'-N-acetyltransferase)Biodegradation of camphorCarbenicillin resistance (,8-lactamase)Cadmium ion resistanceCephalexin resistance (,1-lactamase)Immunity to colicin "X"Production of colicin AProduction of colicin BProduction of colicin DProduction of colicin ElProduction of colicin E2Production of colicin E3Production of colicin IaProduction of colicin IbProduction of colicin KProduction of colicin VProduction of colicin XChloramphenicol resistance (chloramphenicol acetyltransfer-

ase)Cobaltous ion resistanceCephalosporin resistance (,8-lactamase)Cephaloridine resistance (,/-lactamase)Cloxacillin resistance (,8-lactamase)3',4'-Dideoxykanamycin B resistance (aminoglycoside 6'-N'-acetyltransferase)

3',4'-Dideoxykanamycin B resistance (aminoglycoside 2'-N-ace-tyltransferase)

3',4'-Dideoxykanamycin B resistance (aminoglycoside 3-N-ace-tyltransferase)

3',4'-Dideoxykanamycin B resistance (aminoglycoside 2"-ad-enylyltransferase)

Ethidium bromide resistanceEntry exclusionErythromycin resistance (ribosomal ribonucleic acid methylase)Enterotoxin productionExfoliative toxin productionFusidic acid resistanceInhibition of conjugational transfer mediated by plasmid "X"Lack of inhibition of conjugational transfer mediated by plasmidGXd

Gentamicin resistance (aminoglycoside 6'-N-acetyltransferase)Gentamicin resistance (aminoglycoside 2'-N-acetyltransferase)Gentamicin resistance (aminoglycoside 3-N-acetyltransferase)Gentamicin resistance (aminoglycoside 2"-adenylyltransferase)Mercuric ion resistance (mercuric reductase)Hemolysin productionHydrogen sulfide productionDetermination of incompatibility specificityKanamycin resistance (aminoglycoside 3'-phosphotransferase)Kanamycin resistance (aminoglycoside 6'-N-acetyltransferase)


on Decem


mbr.asm

.org/D

ownloaded from



APPENDIX 1 (continued)

Phenotypeb Genotype° Effect

aacBaacCaadB

aadCaphA [Ipt]aacBaacCblahsmhss

nahnic

aphA [npt, neo]aacA [neo]aacB [neo]aacC [neoloctblabla [pen, amp]

pif

traaphA

aacB

aacC

rep[seg][tsr]hsrhss

aphA

aphB

aacB

aacCsal

antsegaphC [spt, str]

aadA [sas, str]

aadA

sultetaacA

Kanamycin resistance (aminoglycoside 3-N-acetyltransferase)Kanamycin resistance (aminoglycoside 3'-N-acetyltransferase)Kanamycin resistance (aminoglycoside 2"-N-adenylyltransfer-

ase)Kanamycin resistance (aminoglycoside 4"-adenylyltransferase)Lividomycin resistance (aminoglycoside 3'-phosphotransferase)Lividomycin resistance (aminoglycoside 2'-N-acetyltransferase)Lividomycin resistance (aminoglycoside 3-N-acetyltransferase)Methicillin resistance ((3-lactamase)Modification of DNA by methylase activityModification of DNA requiring site specificity protein for

expression of methylase activityMutator activityBiodegradation of naphthaleneNickelous ion resistanceNitrogen fixationNeomycin resistance (aminoglycoside 3'-phosphotransferase)Neomycin resistance (aminoglycoside 6'-N-acetyltransferase)Neomycin resistance (aminoglycoside 2'-N-acetyltransferase)Neomycin resistance (aminoglycoside 3-N-acetyltransferase)Biodegradation of octanolOxacillin resistance (P-lactamase)Penicillin resistance ((3-lactamase)Interference with phage productionInterference by F with T3 and T7 propagationInterference by R plasmids with coliphage X and coliphage T1

propagationInterference by ColI with T5 propagationPilus synthesis or structureParamomycin resistance (aminoglycoside 3'-phosphotransfer-

ase)Paramomycin resistance (aminoglycoside 2'-N-acetyltransfer-

ase)Paramomycin resistance (aminoglycoside 3-N-acetyltransfer-

ase)Raffinose fermentationReplicationThermosensitiveRestriction of DNA by endonuclease activityRestriction of DNA requiring site specificity protein for expres-

sion of endonuclease activityRibostamycin resistance (aminoglycoside 3'-phosphotransfer-

ase)Ribostamycin resistance (aminoglycoside 5"-phosphotransfer-

ase)Ribostamycin resistance (aminoglycoside 2'-N-acetyltransfer-

ase)Ribostamycin resistance (aminoglycoside 3-N-acetyltransferase)Biodegradation of salicylatesSurface antigen production (e.g., K88)Antimony ion resistancePlasmid segregationStreptomycin resistance (aminoglycoside-3"-phosphotransfer-

ase)Streptomycin resistance (aminoglycoside 3"-adenylyltransfer-

ase)Spectinomycin resistance (aminoglycoside 3"-adenylyltransfer-

ase)Sulfonamide resistanceTetracycline resistanceTobramycin resistance (aminoglycoside 6'-N-acetyltransferase)

Km [K]Km [K]Km [K]

Km [K]LvLvLvMcModMod

MutNahNicNifNm [N]Nm [N]Nm [N]Nm [N]OctOxPc [P]PhiPhiPhi

PhiPilPm

Pm

Pm

RafRepRep [ts]ResRes

Rm

Rm

Rm

RmSalSanSbSegSm [S]

Sm [S]

Sp

Su [Sa]Tc [T]Tm

183VOL. 40, 1976

on Decem


mbr.asm

.org/D

ownloaded from


APPENDIX 1 (continued)

Phenotypeb genotype' Effect

Tm aacB Tobramycin resistance (aminoglycoside 2'-N-acetyltransferase)Tm aacC Tobramycin resistance (aminoglycoside 3-N-acetyltransferase)Tm aadB Tobramycin resistance (aminoglycoside 2"-adenylyltransferase)Tm aadC Tobramycin resistance (aminoglycoside 4"-adenylyltransferase)Tol tol Biodegradation of tolueneTp dfr Trimethoprim resistance (dihydrofolate reductase)Tra tra [fer] Mediating conjugationUv [uvr] Resistance to ultraviolet

a Abbreviations for chromosomal genes often found on plasmids are not given. The reader is thereforereferred to the articles by Taylor and Trotter (72), Sanderson (65), and Hopwood (42).

b Entries in square brackets refer to published synonyms which, it is hoped, will be superceded by themain entries. Where the gene(s) responsible for a given phenotype has not been defined, no genotype symbolis given.

' It is suggested that resistance phenotypes refer to the substance(s) to which the marker confersresistance and that, where appropriate, genotypes constitute an abbreviation of the inactivating enzyme. Itshould be noted that this convention will, in a general way, distinguish typical plasmid resistance genesfrom the usual types of chromosomal resistance markers. It should be reiterated that, in the absence ofinformation on the nature of the plasmid-specified, drug-inactivating enzyme, only the phenotype designa-tion should be used.

"' There have been described seven seemingly different aminoglycoside acetyltransferases, each of whichinactivates a unique spectrum of aminoglycoside antibiotics by acetylation at the 3, 2', or 6' position on theantibiotic molecule. Although the genes coding for each of these could be given a specific cistron designa-tions from A to G, we believe this might be premature since some of the enzymes acetylating the sameposition on the antibiotic molecule might be related to each other by simple mutational alteration of acommon ancestor gene. We therefore suggest that genes specifying 6'-N-acetyltransferases be designatedaacA, genes specifying 2'-N-acetyltransferases be designated aacB, and genes specifying 3-N-acetyltransfer-ases be designated aacC.

e In keeping with recommendations in the section, Plasmid Gene Abbreviations, we have given all /3-lactamases the same genotypic symbol, since they all have the same mechanism of action.

' There have been described three different aminoglycoside phosphotransferases, each of which inacti-vates a unique spectrum of aminoglycoside antibiotics by phosphorylation at the 3', 5", or 3" position, andwe recommend that the genes specifying these enzymes be designated aphA, aphB, and aphC, respectively.

9 There have been described three different aminoglycoside adenylyltransferases, each of which inacti-vates a unique spectrum of aminoglycoside antibiotics by adenylylation at the 3", 2', or 4" position, and werecommend that the genes specifying these enzymes be designated aadA, aadB, and aadC, respectively.


on Decem


mbr.asm

.org/D

ownloaded from



APPENDIX 2Plasmid Incompatibility Groupsa

Plasmids of enteric bacteria

Publishedgroup des- Suggestedignations phenotypic Plasmid prototypes Reference(including notationssynonyms)

FI IncFI F, R386(HH)b 24FII IncFII R100, R1 35Fill IncFIll ColB-K98 35FIV IncFIV R124(HH) 35FV IncFV Folac 18Com5 Inc5 R27(IP)b Y. Chabbert,

personalcommuni-cation

M, Com7c IncM, Inc7c R446b(H1) 11, 37R69(IP)

Com9 Inc9 R71(IP) 11ComlO InclO R72(IP) 67Comli Incl R147(IP) 67A IncA RA1(HH) 34B IncB TP113, TP125 31C, Com6e IncC, Inc6 R4Oa(IP), R55(IP) 20, 67H IncH R27(HH) (=TP117) 22, 31Ia, Coml, IncIa, IncI, ColIb-P9, A, 11, 31, 36

Il" InclIc R144(HH)I2 IncI2 TP114 31ly IncIy R621a(HH) ColIb- 36, 43

IM1420J IncJ R391(HH) 15L IncL R471a(HH) 39N, Com2r IncN, Inc2e N3, R15 190 IncO R'6, R724(HH) 22, 25P, Com4e IncP, Inc4c RP4 21S IncS R478(HH) 39T IncT Rts1 15W IncW S-a, R7K 34

" Although this list is based on all of the data availableat the- time of this writing, it must be regarded as provi-sional because work on the interrelationships among thelisted plasmids has not been complted.

' (HH) and (IP) indicate that the plasmids so designatedbelong to the Hammersmitli Hospital and Institut Pasteurseries, respectively. These series have overlapping numbersthat have not yet been reconciled (see Section, PlasmidDesignations, above).

c It has been necessary to list synonyms in some cases,because these exist in the literature, and general agree-ment on a single designation has not yet been reached. Insome cases, the indication of synonyms is provisional, sincethe prototype plasmids have not all been fully tested forcompatibility with all of the other prototypes listed.

Plasmids of Staphylococcus aureusPlasmid proto-

Group Phenotype types Reference

1 Incl pI524, pI258 592 Inc2 pII147 593 Inc3 pT127 644 Inc4 pS177 645 Inc5 pC221 646 Inc6 pK545 647 Inc7 pUB101a 47; R, Novick, un-

published data

epUB101 is a plasmid carrying genes for penicillinaseand fusidic acid resistance present in strain FAR4 (47).

APPENDIX 3

Summary of Recommendations ofDemerec et al.

1. Each locus of a given wild-type strain isdesignated by a three-letter, lower-case itali-cized symbol.

2. Different loci, any one of which maymutate to produce the same gross phenotypicchange, are distinguished from each other byadding an italicized capital letter immediatelyfollowing the three-letter lower-case symbol.

3. A mutation site should be designated byplacing a serial isolation number after thelocus symbol. If it is not known in which ofseveral loci governing related functions themutation has occurred, the capital letter isreplaced by a hyphen.

4. Plasmids and episomes should be desig-nated by symbols which are clearly distinguish-able from symbols used for genetic loci.

5. Mutant loci and mutational sites on plas-mids and episomes should be designated bysymbols of the same kind as those used forloci and sites on the chromosomes.

6. The description of a strain carrying anepisome should include a statement concerningthe state and/or location of the episome. Thesymbols F-, F+, F' and Hfr should be usedonly to designate the sex factor states as out-lined above, and not to convey informationconcerning the phenotypic properties of matingactivity.

7. Genotype symbols which have alreadybeen published and which conform to the sys-tem recommended above should not be changed.Genotype symbols which do not conform to theabove system should be changed accordingly,and the change should be noted when the newsymbol is first published.

8. Phenotypic traits should be described inwords, or by the use of abbreviations which aredefined the first time they appear in a givenpaper. The abbreviations should be clearly dis-tinguishable from genotype symbols.

9. Strains should be designated by simpleserial numbers. To avoid duplications, differentlaboratories should use different letter prefixes.Strain designations should not be italicized.

10. Strain designations which have alreadybeen published and which conform to Recom-mendation 9 should not be changed. Straindesignations which do not conform to Recom-mendation 9 should be changed accordingly,and the change should be noted when the newdesignation is first published.

11. When a strain is first mentioned inpublication its genotype should be described,and relevant phenotypic information should be

185VOL. 40, 1976

on Decem


mbr.asm

.org/D

ownloaded from


186 NOVICK ET AL.

given. The genotype includes a list of all mu-tant loci and/or mutation sites, a list of epi-somes and/or plasmids, and information con-cerning the state and location of any episome.

(Appendix 3 of the proposal on nomenclaturefor bacterial genetics by Demerec et al. [23].This summary is quoted here for reference; itshould be noted that in the present proposalthe term "episome" [recommendations 4, 5, and6] is updated [see text], and the abbreviationF' [recommendation 6] is redefined [see text].)

APPENDIX 4Summary of Recommendations on

Plasmid Nomenclature1. Each newly described plasmid and each

newly isolated genotypic modification of aknown plasmid is given a unique numericaldesignation of the form pXY1234. Publishedplasmid designations may be retained even ifthey do not conform to this recommendationunless they are ambiguous or have been dupli-cated, in which case they should be changedand the change noted in publications.

2. Bacterial strain numbers refer to strainsas they exist, including plasmid content.

3. Any change in the plasmid content of abacterial strain, including genotypic altera-tions of preexisting plasmids, gives rise to anew strain number.

4. The phenotypic notation for a plasmidgene should consist of a capital and a lower-case letter (or two if necessary) and shouldreflect the phenotypic trait for which the geneis responsible.

5. In the initial listing of a plasmid pheno-type or genotype, all genes known to be presentshould be specified, even those that are presentin their wild-type configuration. Phenotypicdesignations may require an indicator of speci-ficity.

6. Genotypic notations for plasmid genesshould be in the form recommended by De-merec et al. (23) for bacterial genes (recom-mendations 1, 2, and 3). The complete geno-typic identification of a plasmid gene, however,includes as a prefix the name of the plasmidon which the gene was found.

7. Plasmid recombinants should be givennew plasmid numbers. Where the plasmidsinvolved are heterogenic, the complete geno-type of the recombinant should specify thecontributions of each parent by enclosing theappropriate symbols within square bracketsand should indicate also those genes whoseparentage is uncertain by listing the appro-priate symbols outside of the brackets.

8. Plasmid deletions should be indicated bya Greek delta and identified by unique serialnumeration and a list of the deleted genes.

9. Plasmid insertions, transpositions, andtranslocations should be indicated by a Greekomega and identified by unique serial numera-tion and a list of the translocated genes. Thephysical location of the inserted materialshould be specified where known, as shouldits origin.

ACKNOWLEDGMENTSThe preparation of this proposal would not have

been possible without the support and encourage-ment of our colleagues from many countries. Weshould like to thank especially those, too numerousto mention individually, who responded in writingto our request for suggestions.We should like particularly to acknowledge

helpful discussions with Masanosuke Yoshikawaand Toshio Arai who were present at one of ourmeetings (New Orleans, January 1974). We shouldlike also to acknowledge the cooperation and sup-port of Susumu Mitsuhashi and the other Japaneseworkers who held meetings on nomenclature andcontributed a written summary of their sugges-tions, which was most useful in the preparation ofthis document. We should like also to thank JulianDavies, Yves Chabbert, and Susumu Mitsuhashifor their help in formulating the genotypes andphenotypes associated with antibiotic-inactivatingenzymes. Suggestions by Alan Campbell, Alvin J.Clark, Gilbert Howe, and Neil Willetts were mosthelpful in the clarification of a number of termsand the rationalization of some of the proposeddesignations for plasmids and plasmid-carriedgenes.We should like to thank the following pharma-

ceutical companies for their very generous financialsupport: Abbott Laboratories, Bristol Laboratories,Eli Lilly & Co., Merck & Co., Inc., Schering Labora-tories, E. R. Squibb & Sons, Inc., and The UpjohnCo.

Finally, we should like to thank Annabel Howardfor her expert secretarial help.

LITERATURE CITED1. Achtman, M., N. Willetts, and A. J. Clark.

1971. Beginning a genetic analysis of con-jugational transfer determined by the Ffactor in Escherichia coli by isolation andcharacterization of transfer-deficient mu-tants. J. Bacteriol. 106:529-538.

2. Akiba, T., K. Koyama, Y. Ishiki, S. Kimura,and T. Fukushima. 1960. On the mechanismof the development of multiple-drug resist-ant clones of Shigella. Jpn. Med. J. 1866:46-50.

3. Anderson, E. S., and M. J. Lewis. 1965.Characterization of a transfer factor asso-ciated with drug resistance in Salmonellatyphimurium. Nature (London) 208:843-849.

4. Anderson, E. S., and E. J. Threlfall. 1970.

BACTERIOL. REV.

on Decem


mbr.asm

.org/D

ownloaded from



Change of host range in a resistance factor.Genet. Res. 16:207-214.

5. Bannister, D., and S. W. Glover. 1970. Theisolation and properties of non-restrictingmutants of two different host specificitiesassociated with drug resistance factors. J.Gen. Microbiol. 61:63-71.

6. Bazaral, M., and D. R. Helinski. 1968. Char-acterization of multiple circular DNA formsof colicinogenic factor El from Proteusmirabilis. Biochemistry 7:3513-3519.

7. Berg, C. M., and R. Curtiss III. 1967. Trans-position derivatives ofan Hfr strain ofEsche-richia coli K-12. Genetics 56:503-525.

8. Brinton, C. C., P. Gemski, and J. Carnahan.1964. A new type of bacterial pilus geneti-cally controlled by the fertility factor ofE. coli K12 and its role in chromosometransfer. Proc. Natl. Acad. Sci. U.S.A.52:776-783.

9. Carlton, B. C., and D. R. Helinski. 1969.Heterogeneous circular DNA elements invegetative cultures of Bacillus megaterium.Proc. Natl. Acad. Sci. U.S.A. 64:592-599.

10. Cavalli, L. L., J. Lederberg, and E. M. Leder-berg. 1953. An infective factor controllingsex compatibility in Bacterium coli. J. Gen.Microbiol. 8:89-103.

11. Chabbert, Y. A., M. R. Scavizzi, J. L. Witchitz,G. R. Gerbaud, and D. H. Bouanchaud.1972. Incompatibility groups and the classi-fication offi- resistance factors. J. Bacteriol.112:666-675.

12. Clowes, R. C. 1961. Colicin factors as fertilityfactors in bacteria: Escherichia coli. Nature(London) 190:988-989.

13. Clowes, R. C. 1963. Colicin factors and epi-somes. Genet. Res. 4:162-165.

14. Clowes, R. C. 1972. Molecular structure ofbacterial plasmids. Bacteriol. Rev. 36:361-405.

15. Coetzee, J. N., N. Datta, and R. W. Hedges.1972. R factors from Proteus rettgeri. J.Gen. Microbiol. 72:543-552.

16. Cohen, S. N., and C. A. Miller. 1969. Multiplemolecular species of circular R-factor DNAisolated from Escherichia coli. Nature (Lon-don) 224:1273-1277.

17. Cooper, P. 1971. Interaction of a colicinogenicfactor with a resistance factor and with afertility factor F in E. coli K12. Genet. Res.17:151-159.

18. Datta, N. 1975. Epidemiology and classifica-tion of plasmids, p. 9-15. In D. Schlessinger(ed.), Microbiology -1974. American Societyfor Microbiology, Washington, D.C.

19. Datta, N., and R. W. Hedges. 1971. Com-patibility groups among fi- R factors. Na-ture (London) 234:222-223.

20. Datta, N., and R. W. Hedges. 1972. R factorsidentified in Paris, some conferring genta-micin resistance, constitute a new compati-bility group. Ann. Inst. Pasteur 123:879-883.

21. Datta, N., R. W. Hedges, E. J. Shaw, R. B.Sykes, and M. H. Richmond. 1971. Prop-

erties of an R factor from Pseudomonasaeruginosa. J. Bacteriol. 108:1244-1249.

22. Datta, N., and J. Olarte. 1974. R factors instrains of Salmonella typhi and Shigelladysenteriae 1 isolated during epidemics inMexico: classification by compatibility. Anti-microb. Agents Chemother. 5:310-317.

23. Demerec, M., E. A. Adelberg, A. J. Clark,and P. E. Hartman. 1966. A proposal for auniform nomenclature in bacterial genetics.Genetics 54:61-76.

24. Dennison, S. 1972. Naturally occurring Rfactor, derepressed for pilus synthesis, be-longing to the same compatibility group asthe sex factor F of Escherichia coli K-12.J. Bacteriol. 109:416-422.

25. Evans, J., E. Galindo, J. Olarte, and S.Falkow. 1968. /3-Lactamase of R factors. J.Bacteriol. 96:1441-1442.

26. Falkow, S., J. A. Wohlhieter, R. V. Citarella,and L. S. Baron. 1964. Transfer of episomicelements to Proteus. II. Nature of lac+Proteus strains isolated from clinical speci-mens. J. Bacteriol. 88:1598-1601.

27. Fischer-Fantuzzi, L., and E. Calef. 1964. Atype of X prophage unable to confer im-munity. Virology 23:209-216.

28. Fredericq, P. 1969. The recombination of coli-cinogenic factors with other episomes andplasmids, p. 163-174. In G. E. W. Wolsten-holme and M. O'Connor (ed.), CIBA sym-posium. Bacterial episomes and plasmids.J. & A. Churchill Ltd., London.

29. Fredericq, P., and M. Betz-Bareau. 1953.Transfert genetique de la propriete colicino-gene en rapport avec la polarite F desparents. C. R. Seances Soc. Biol. Paris147:2043-2045.

30. Grindley, J. N., and E. S. Anderson. 1971.I-like resistance factors with the fi+ char-acter. Genet. Res. 17:267-271.

31. Grindley, N. D. F., J. N. Grindley, and E. S.Anderson. 1972. R factor compatibilitygroups. Mol. Gen. Genet. 119:287-297.

32. Hayes, W. 1953. Observations on a transmissi-ble agent determining sexual differentiationin Bacterium coli. J. Gen. Microbiol. 8:72-88.

33. Hayes, W. 1964. The genetics of bacteria andtheir viruses. John Wiley & Sons, NewYork.

34. Hedges, R. W., and N. Datta. 1971. fi- Rfactors giving chloramphenicol resistance.Nature (London) 234:220-221.

35. Hedges, R. W., and N. Datta. 1972. R124 anfi+ R factor of a new compatibility class.J. Gen. Microbiol. 71:403-405.

36. Hedges, R. W., and N. Datta. 1973. Plasmidsdetermining I pili constitute a compatibilitycomplex. J. Gen. Microbiol. 77:19-25.

37. Hedges, R. W., N. Datta, J. N. Coetzee, andS. Dennison. 1973. R factors from Proteusmorganii. J. Gen. Microbiol. 77:249-259.

38. Hedges, R. W., and A. Jacob. 1974. Trans-position of ampicillin resistance from RP4

187VOL. 40, 1976

on Decem


mbr.asm

.org/D

ownloaded from


188 NOVICK ET AL.

to other replicons. Mol. Gen. Genet. 132:31-40.

39. Hedges, R. W., V. Rodriguez-Lemoine, andN. Datta. 1975. R factors from Serratiamarcescens. J. Gen. Microbiol. 86:88-92.

40. Heffroa, F., R. Sublett, R. W. Hedges, A.Jacob, and S. Falkow. 1975. Origin of theTEM beta-lactamase gene found on plas-mids. J. Bacteriol. 122:250-256.

41. Hofuan-Berling, H., and R. Maze. 1964.Release of male specific bacteriophages fromsurviving host bacteria. Virology 22:305-313.

42. Hopwood, D. A. 1967. Genetic analysis andgenome structure in Streptomyces coelicolor.Bacteriol. Rev. 31:373-403.

43. Howarth-Thompson, S., H. M. Heffernan, andM. D. Buchan. 1974. Two compatibilitygroups among ColI plasmids. J. Gen. Micro-biol. 81:279-282.

44. Ikeda, H., and J. Tomizawa. 1968. ProphageP1, an extrachromosomal replication unit.Cold Spring Harbor Symp. Quant. Biol.33:791-798.

45. Jacob, F., and E. L. Wollman. 1958. Les6pisomes, elements g~netiques ajoutes. C. R.Acad. Sci. Paris 247:154-156.

46. Kopecko, D. J., and S. N. Cohen. 1975. Site-specific recA-independent recombination be-tween bacterial plasmids: involvement ofpalindromes at the recombinational loci.Proc. Natl. Acad. Sci. U.S.A. 72:1373-1377.

47. Lacey, R. W., E. Lewis, and V. T. Rosdahl.1974. Evolution of plasmids in vivo in astrain of Staphylococcus aureus. J. Med.Microbiol. 7:117-125.

48. Lederberg, J. 1952. Cell genetics and heredi-tary symbiosis. Physiol. Rev. 32:403-430.

49. Lederberg, J., L. L. Cavalli, and E. M. Leder-berg. 1952. Sex compatibility in E. coli.Genetics 37:720-730.

50. Low, K. B. 1972. Escherichia coli K-12 F-prime factors, old and new. Bacteriol. Rev.36:587-607.

51. Malamy, M. H. 1970. Some properties ofinsertion mutations in the lac operon, p.359-373. In J. R. Beckwith and D. Zipser(ed.), The lactose operon. Cold Spring Har-bor Laboratory, Cold Spring Harbor, NewYork.

52. Meynell, E., G. G. Meynell, and N. Datta.1968. Phylogenetic relationships of drug-resistance factors and other transmissiblebacterial plasmids. Bacteriol. Rev. 32:55-83.

53. Nagel de Zweig, R., D. N. Anton, and J. Puig.1962. The generic control of colicinogenicfactors E2, I and V. J. Gen. Microbiol.29:473-484.

54. Nisioka, T., M. Mitani, and R. C. Clowes.1969. Composite circular forms of R-factordeoxyribonucleic acid molecules. J. Bac-teriol. 97:376-385.

55. Novick, R. P. 1963. Analysis by transductionof mutations affecting penicillinase forma-

tion in Staphylococcus aureus. J. Gen.Microbiol. 33:121-136.

56. Novick, R. P. 1969. Extrachromosomal in-heritance in bacteria. Bacteriol. Rev. 33:210-263.

57. Novick, R. P., and R. Brodsky. 1972. Studieson plasmid replication. I. Plasmid incom-patibility and establishment in Staphylo-coccus ajreus. J. Mol. Biol. 68:285-302.

58. Novick, R. P., ;nd S. I. Morse. 1967. In vivotransmission of drug resistance factors be-tween strains of Staphylococcus oureus. J.Exp. Med. 125:45-59.

59. Novick, R. P., and M. H. Richmond. 1965.Nature and interactions of the geneticelements governing penicillinase synthesisin S. aureus. J. Bacteriol. 90:467-480.

60. Ochiai, K., T. Yamanaka, K. Kimura, and0. Sawada. 1969. Inheritance of drug resist-ance (and its transfer) between Shigellastrains and between Shigella and E. colistrains. Jpn. Med. J. 1861:34-46.

61. Ozeki, H., and S. Howarth. 1961. Colicinefactors as fertility factors in bacteria: Sal-monella typhimurium, strain LT2. Nature(London) 190:9B6-988.

61a. Richmond, M. H., G. W. Jack, end R. B.Sykes. 1971. The f-lactamases of Gram-negative bacteria including Pseudomonads.Ann. N.Y. Acid. Sci. 182:243-257.

62. Rownd, R. 1968. The replication of resistancetransfer factors. In S. Falkow (ed.), Interna-tional symposium on infectious multipledrug resistance. U.S. Govt. Printing Office,Washington, D.C.

63. Rownd, R., H. Kasamatsu, and S. Mickel.1971. The molecular nature and replicationof drug resistance factors of the entero-bacteriaceae. Ann. N.Y. Acad. Sci. 182:188-206.

64. Ruby, C., and R. P. Novick. 1975. Plasmidinteractions in Staphylococcus aureus: non-additivity of compatible plasmid DNA pools.Proc. Natl. Acad. Sci. U.S.A. 72:5031-5036.

65. Sanderson, K. E. 1972. Linkage map ofSalnso-nella typhimurium, edition IV. Bacteriol.Rev. 36:558-586.

65a. Sawai, T., K. Takahashi, S. Yamagishi, andS. Mitsuhashi. 1970. Variant of penicillinasemediated by an R factor in Escherichia coli.J. Bacteriol. 104:620-629.

66. Scaife, J., and A. P. Pekhov. 1964. Deletionof chromosomal markers in association withF-prime factor formation in Escherichia coliK-12. Genet. Res. 5:495-498.

67. Scavizzi, M. R. 1973. Nouveaux groupes d'in-compatibilit6 des plasmides. Intkr~t dans lesepidemies de creche A Escherichia coliO111:B4. Ann. Microbiol. (Inst. Pasteur)124B:153-167.

68. Sharp, P. A., S. N. Cohen, and N. Davidsqn.1973. Electron microscope heteroduplexstudies of sequence relations among plas-mids of Escherichia coli. II. Structure of

BACTERIOL. REV.

on Decem


mbr.asm

.org/D

ownloaded from


BACTERIAL PLASMID NOMENCLATURE 189

drug resistance (R) factors and F factors. J.Mol. Biol. 75:235-255.

69. Sharp, P. A., M.-T. Hsu, E. Ohtsubo, andN. Davidson. 1972. Electron microscopeheteroduplex studies of sequence relationsamong plasmids ofEscherichia coli. I. Struc-ture of F-prime factors. J. Mol. Biol. 71:471-497.

70. Starlinger, P., and H. Saedler. 1972. Insertionmutations in microorganisms. Biochimie 54:177-185.

71. Sugino, Y., and Y. Hirota. 1962. Conjugalfertility associated with resistance factor Rin Escherichia coli. J. Bacteriol. 84:902-910.

72. Taylor, A. L., and C. D. Trotter. 1972. Linkagemap of Escherichia coli strain K-12. Bac-teriol. Rev. 36:504-524.

73. Timmis, K., F. Cabello, and S. N. Cohen.1974. Utilization of two distinct modes ofreplication by a hybrid plasmid constructedin vitro from separate replicons. Proc. Natl.Acad. Sci. U.S.A. 71:4556-4560.

VOL. 40, 1976

on Decem


mbr.asm

.org/D

ownloaded from


ERRATAThe Hemolysins of Staphylococcus aureus

GORDON M. WISEMANDepartment ofMedical Microbiology, University ofManitoba, Winnipeg, R3E OW3 Canada

Volum3 39, no. 4, p. 321, column 2, line 3: "Bernheimer et al. (112) studied" should read "Klaineret al. (112) studied."Page 324, Table 5, column heading. "mg ofpacked cells/ml" should read "mg/ml ofpacked cells."Page 327, Table 7, opposite the reference to Gow and Robinson's paper, under column heading

"Sp act (HU) of protein or nitrogen/mg": "5 x 105/mg protein" should read "5 x 108/mg protein."

Recalibrated Linkage Map of Escherichia coli K-12BARBARA J. BACHMANN,* K. BROOKS LOW, AiD AUSTIN L. TAYLOR

Yale University School ofMedicine, New Haven, Connecticut 06510,* and University of ColoradoMedical Center, Denver, Colorado 80220

Volume 40, no. 1, p. 141. The genetic symbol for the loci specifying the subunits ofRNA polym-erase was incorrectly given as pro. The relevant text should have read: "The determinants forRNA polymerase subunits will be called rpo (J. G. Scaife and R. S. Hayward, personal communi-cation). Thus some ofthe most familiar old drug resistance markers now appear on the map undernew symbols: rif is now rpoB," . . .Page 166, the following reference should be inserted: "732. Wright, M. 1971. Mutants ofEsche-

richia coli lacking endonuclease I, ribonuclease I, or ribonuclease II. J. Bacteriol. 107:87-94."

Uniform Nomenclature for Bacterial Plasmids: a ProposalRICHARD P. NOVICK,* ROYSTON C. CLOWES, STANLEY N. COHEN, ROY CURTISS III,

NAOMI DATfA, AND STANLEY FALKOWDepartment ofPlasmid Biology, The Public Health Research Institute of the City ofNew York, Inc.,

New York, New York i0016*; Division ofBiology, University of Texas at Dallas, Richardson,Texas 75080; Department of Medicine, Stanford University School of Medicine, Stanford,California 94305; Department of Microbiology, University ofAlabama in Birmingham,

Birmingham, Alabama 35294; Department ofBacteriology, Royal PostgraduateMedical School, Hammersmith Hospital, London, W12 OHS, England; Department

of Microbiology, University of Washington, Seattle, Washington 98195

Volume 40, no. 1, p. 171, column 2, following line 12 of section (ix), insert: them, R and Col,are already in near universal use, it appears unavoidable to acknowledge..At the end of same paragraph, delete the partial sentence beginning with "-tionally" and

ending with "arbitrary."Page 179, column 2, main heading "MOLECULAR REARRANGEMENTS" and 3 lines

following: move to p. 181, top of column 1.The publisher regrets these errors, which were introduced at the final correction stage.

525

Documents

Uniform Nomenclature for Bacterial Plasmids: Proposal170 NOVICK ET AL. was then thought to integrate into theEsche-richiacoli chromosome). However,it sooncame to be used as a generic