Gene Technology and Gene Ecology of InfectiousDiseasesMae-Wan Ho1, Terje Traavik2, Orjan Olsvik2, Beatrix Tappeser3, C. Vyvyan Howard4,Christine von Weizsacker5 and George C. McGavin6

From the 1Biology Dept., Open University, Walton Hall, Milton Keynes, MK7 6AA, UK, 2Departments ofVirology and Microbial Genetics, Institute of Medical Biology, University of Tromsø, Norway, 3Institute forApplied Ecology, Postfach 6226, DE-79038 Freiburg, Germany, 4Foetal and Infant Toxipathology, Universityof Liverpool, Liverpool L69 3BX, UK, 5Postfach 130165, 53061 Bonn, Germany, 6Assistant Curator ofEntomology, Museum of Natural History, Oxford University, Parks Road, Oxford OX1 3PW, UK.

Correspondence to: Mae-Wan Ho, Biology Dept., Open University, Walton Hall, Milton Keynes, MK7 6AA, UK.

Microbial Ecology in Health and Disease 1998; 10: 33–59

According to the 1996 WHO Report, the world is heading for a major crisis in public health as outbreaks of new and re-emerginginfectious diseases are striking at increasing frequencies within the past 10 to 15 years. The current strains of pathogens are moreover,resistant to known treatments; some strains being resistant to all or nearly all drugs and antibiotics. Horizontal gene transfer is nowgenerally recognized to be responsible for the evolution of virulence and the spread of drug and antibiotic resistances. Many pathogenshave crossed species barriers, having acquired genes from phylogenetically distant species that are involved in their ability to causediseases. Recent findings document the extremely wide scope of horizontal gene transfer and the extensive recombination between geneticmaterial from unrelated species that have contributed to the emergence of virulence and antibiotic resistances. The past 15 years coincidewith the development of genetic engineering biotechnology on a commercial scale. Genetic engineering depends on designing vectors forcloning and transferring genes and involves artificially recombining and manipulating genes from unrelated species and their viralpathogens, thereby enhancing the probability for horizontal gene transfer and recombination. The urgent question which needs to beaddressed is the extent to which genetic engineering biotechnology, by facilitating horizontal gene transfer and recombination, iscontributing to the resurgence of infectious, drug-resistant diseases, and will continue to do so if allowed to proceed unchecked. Anenquiry into the possible contribution of genetic engineering biotechnology to the etiology of infectious diseases is all the more pressingin the light of other relevant recent findings indicating that microorganisms genetically engineered for ‘contained use’ may not beeffectively contained. Thus, biologically ‘crippled’ strains of bacteria can survive in the environment to exchange genes with other species;DNA released from cells is not readily broken down in the environment, thereby retaining the ability to transform organisms; some viralDNA can be more infectious than the virus itself; and routine chemical treatments for inactivating pathogenic microorganisms andviruses, before they are discharged into the environment, may be ineffective, leaving a substantial percentage of pathogens in an activeinfectious state. The combination of the different kinds of evidence is sufficiently compelling, especially in view of the precautionaryprinciple, to warrant, at the very least, an independent public enquiry into genetic engineering biotechnology and the etiology of infectiousdiseases. Key words : horizontal gene transfer, virulence, antibiotic resistance, naked DNA.

ORIGINAL ARTICLE

1. THE WORLD HEALTH CRISIS

The world is heading for a public health crisis. At least 30new diseases such as AIDS, ebola, several kinds of hepati-tis and other deadly viruses have emerged over the past 20years (see Table I) [compiled from WHO Report, 1996 (1),and other sources (2)], while old infectious diseases such astuberculosis, cholera, malaria and diphtheria are comingback worldwide. Practically all the pathogens are resistantto drug treatment, many to multiple antibiotics. One thirdof the 52 million deaths from all causes in the world in1995 were due to infectious diseases; over half of these inyoung children. The top killers were tuberculosis (3.1million, mostly adults), malaria (2.1 million, including 1million children), hepatitis B (1.1 million) and AIDS (morethan 1 million). ‘The optimism of a relatively few years

ago that many of these diseases could easily be broughtunder control has led to a fatal complacency among theinternational community,’ says the Director-General ofWHO, Dr. Hiroshi Nakajima in a press release. Emergingdiseases have become a major public health concern duringthe 1990s.

1.1 Ha6e infectious diseases and drug resistance gone uprecently?

Precise epidemiological data are not yet available, but thesigns are that both the incidence and severity of outbreaksof multi-drug resistant pathogens may have experienced asharp upturn within the past 10 to 15 years. For example,Salmonella infections have gone up 20-fold in some coun-tries in Europe since 1980 (3). Similar increases are re-ported for haemorhagic E. coli 0157 food poisoning:

© Scandinavian University Press 1998. ISSN 0891-060X Microbial Ecology in Health and Disease

M.-W. Ho et al.34

between 1986 and 1996, the frequency of infection in-creased 10-fold in England and Wales, and 100-fold inScotland (4).

The first widely used anti-malarial drug, chloroquine,came into use during world war II, and resistance did notappear until the early 1960s. By contrast, the new drugmefloquine, released in 1985, became useless in 60% ofmalaria cases within 5 years (5). Comparable accelerationsin the development of antibiotic resistances have takenplace in the same period. Antibiotics were introduced inthe early 1940s to treat infectious diseases, and resistancedid not appear until the early 1950s (6). Resistance topenicillin, ampicillin and antipseudomonas penicillins inStaphyloccocus aureus went from almost 0% in 1952 tomore than 95% in 1992 (7). By the 1980s, S. aureus hasalso developed high levels of resistance to the syntheticpenicillin, methicillin and all other b-lactams. The newfluoroquinolone antimicrobial, ciprofloxacin, was intro-duced in the mid-1980s, but resistance to it had reachedmore than 80% by 1992. A study carried out by theCenters for Disease Control showed that ciprofloxacinresistance in S. aureus went from less than 5% to greaterthan 80% within one year (7).

By 1990, nearly every common pathogenic bacterialspecies had developed varying degrees of antibiotic resis-tance, often, multiple resistances. These include, besidesStaphylococcus aureus (Toxic Shock Syndrome, postopera-tive infections), Streptococcus aureus (Toxic Shock-likeSyndrome) S. pneumoniae (pneumonia) S. pyogenes(rheumatic fever) Haemophilus influenzae (mennigitis) My-cobacterium leprae (leprosy), Neisseria gonorrhoeae (gonor-rhea), Shigella dysenteriae (dysentry) and several otherspecies of microbes that infect the human gut: E. coli,Klebsiella, Proteus, Salmonella, Serrratia marcescens, Pseu-domonas, Enterococcus faecium, Enterobacteriaceae andVibrio cholerae (cholera) (6). A strain of Staphylococcusisolated in Australia was resistant to 31 different drugsincluding cadmium, penicillin, kanamycin, neomycin,streptomycin, tetracycline and trimethyloprim (8). The var-ious resistance capabilities were due to genes carried ondifferent plasmids (independently replicating units of ge-netic material) that could be separately passed on fromone bacterium to another. Two strains of E. coli isolated ina transplant ward outside Cambridge were resistant to 21out of 22 common antibiotics: imipenen, as well as cefo-taxime, ceftazdime, ciprofloxacin, gentamicin, ampicillin,azlocillin, coamoxiclav, timentin, cephalexin cefuroxime,cefotaxime, cefamandole, streptomycin, neomycin,kanamycin, tobramycin, trimethoprim, sulfamethoxazole,chloramphenicol and nitrofurantoin (9). These multi-resis-tant strains of bacteria are rapidly becoming totally invul-nerable to treatment. Scientists in Japan have alreadyisolated a strain of Staphylococcus aureus that is resistanteven to the last resort antibiotic, vancomycin (10). Recentdata showed that vancomycin resistance in Enterococci

grew from 3% in 1993 to 95% in 1997 in hospitals in SanFrancisco (11) In Italy, erythromycin resistance in Strepto-coccus increased 20-fold just between 1993 and 1995 (12).

1.2 Factors contributing to the recent resurgence ofinfectious diseases

The precise reasons for the resurgence of infectious dis-eases since the 1980s are not known. Many contributingfactors have been suggested (1), among them,

– population growth combined with rapid urbanization,– wars, civil turmoil and natural disasters,– rapid increases in international air travel and the grow-

ing traffic in trade,– expanding areas of human habitation and the conse-

quent environmental destruction,

Table I

New diseases within the past 20 years (incomplete list)

1977: Legionella pneumophila, potentially fatal Legionnaires’disease

1977: Ebola virus, haemorrhagic fever—fatal in up to 80%of cases

1977: Hantaan virus, potentially fatal haemorrhagic feverwith renal syndrome

1977: Campylobacter jejuni, diarrhoeaHuman T-lymphotropic virus I (HTLV-1), T-cell1980:lymphoma-leukemia

1981: Staphylococcus aureus, bacteria, toxic shock syndrome1982: Escherichia coli O157:H7 bacteria, bloody diarrhoea1982: HTLV-2 virus, hairy cell leukaemia

Helicobacter pylori, bacteria, peptic ulcer disease and1983:stomach cancer

1983: Human immunodeficiency virus (HIV), AIDSEnterocytozoon bieneusi, parasite, persistent diarrhea1985:Cyclospora cayetanensis, parasite, persistent diarrhea1986:Hepatitis E virus, epidemics of jaundice in hot cli-1988:mates

1988: Human herpesvirus 6, fever and rashSalmonella typimurium, bacteria, diarrhea1988:Hepatitis C virus, which causes liver cancer as well as1989:liver disease

1989: Ehrlichia chafeensis, bacteria, infectionGuanarito virus, Venezuelan haemorrhagic fever1991:

1991: Encephalitozoon hellem, parasite, conjunctivities1991: New species of Babesia, parasite, infection1992: New strain Vibro cholerae O139, epidemic cholera1992: Bartonella henselae, bacteria cat-scratch disease, bacil-

lus angiomatosisSin Nombre virus, adult respiratory distress syndrome1993:

1993: Encephalitozoon cuniculi, parasite, disseminated disease1994: Sabia virus, which causes Brazilian haemorrhagic

fever1994: Equine morbilivirus, respiratory illness1995: Human herpesvirus 8, Kaposi’s sarcoma in AIDS pa-

tients1995: New monkey pox virus, human to human transmis-

sion, potentially fatal1996: New rabies in Australia1996: New hantavirus, human to human transmission

Gene technology and gene ecology of infectious diseases 35

– social changes including the clustering of young chil-dren in day-care centers and growing numbers of theelderly in nursing homes,

– complacency towards infectious disease in the publichealth sector and collapse of public health systems dueto economic or social crises,

– the overuse and abuse of antibiotics in intensive farmingand medicine.

One factor which has not yet been considered is geneticengineering biotechnology.

1.3 Gene technology and gene ecology

Gene ecology is an emerging discipline. It is prompted byrecent findings that the genetic material—DNA or in someinstances, RNA—can transfer horizontally from one or-ganism to another through the external environment, in-stead of 6ertically by reproduction. Gene ecology is thetotality of how genes function, mutate, move, transfer andrecombine subject to feedback regulation from the inter-connected levels of the genome, the physiology of theorganism and the external ecological conditions.

Gene technology, or genetic engineering, can profoundlydisturb the ecology of genes. The past 15–20 years wit-nessed the rise of genetic engineering biotechnology on acommercial scale. It is a technology for manipulating andtransferring genes horizontally between species that do notnormally interbreed. It is designed to break down speciesbarriers and, increasingly, to overcome the species’ defencemechanisms which degrade or inactivate foreign genes (13,14).

For the purpose of manipulating, replicating and trans-ferring genes, a range of artificial 6ectors have been con-structed. Natural vectors for horizontal gene transfercomprise replicating, often mobile units of genetic mate-rial: viruses, plasmids and transposons. However, otherpieces of DNA can also be taken up by cells and hence actas agent for horizontal gene transfer. While viruses consistof genetic material wrapped in a protein coat, plasmidsand transposons are, to varying extent, naked DNA. Plas-mids generally replicate in the cytoplasm, while transpo-sons are integrated into the chromosome, but can movefrom one site to another in the same or different chromo-some, or from chromosome to plasmid and 6ice 6ersa.Plamids and transposons typically carry virulence genesand genes for antibiotic resistance.

Artificial vectors are made by joining together parts ofthe most infectious natural vectors in order to enhancehorizontal gene transfer. The artificial vectors are ‘crip-pled’, which means that most, if not all of the genescausing disease and infectivity are removed. But it does notmean those functions cannot be supplied by, or regainedfrom, other viruses and parasitic genetic agents that arealways present in the environment and in the cells of allorganisms. The gene to be transferred is as a rule inte-

grated into the genetic material of the vector; viruses,however, can also transfer genes that are not integrated,but merely packaged within the protein coat.

To make a transgenic organism, the vector carrying theforeign gene to be transferred— transgene—is allowed toinfect the cells of the organism. Once inside the cell, thevector can either multiply many copies of itself or becomeintegrated into the genome of the cell. Most artificialvectors possess one or more antibiotic resistance markergenes, so transformed cells can be selected with the appro-priate antibiotic(s).

There are three main routes for horizontal gene transfer:infection with viruses (transduction), through pieces ofgenetic material taken up into cell from the environment(transformation); or by unusual mating taking place be-tween unrelated species (conjugation). As the entire orien-tation of genetic engineering is to facilitate horizontal genetransfer, it is to be expected that antibiotic resistance genesas well as virulence genes will inadvertently spread andrecombine to generate new, drug and antibiotic resistantpathogens.

Indeed, the evolution of virulence and the spread ofdrug and antibiotic resistances are now linked to theextensive horizontal gene transfer and recombinationevents among bacteria and viruses, many of which mayhave occurred in recent years. For example, horizontalgene transfer and subsequent genetic recombination havegiven rise to the new bacterial strains responsible for thecholera outbreak in India in 1992 (15–17), and for theStreptococcus epidemic in Tayside in 1993 (18–21). The E.coli 0157:07 strain involved in the recent outbreaks inScotland is believed to have originated from horizontalgene transfer from the pathogen, Shigella (22). Manyunrelated bacterial pathogens, causing diseases frombubonic plague to tree blight, are now found to share anentire set of genes for invading cells, which have spreadwidely by horizontal gene transfer (23). Similarly, genes forantibiotic resistances have been transferred horizontallyand have recombined with one another to generate multi-ple antibiotic resistances throughout the bacterial popula-tions (24). Antibiotic resistance genes spread readilybetween human beings, and from bacteria inhabiting thegut of farm animals to those in human beings (25). Antibi-otic resistant strains of pathogens have been endemic inmany hospitals for years. In the USA, up to 60% ofhospital-acquired infections are antibiotic resistant (1). Aspecial form of multiple which confer resistance to a widerange of chemically unrelated drugs made its appearanceamong pathogens in the early 1990s (see Section 3.2),although it has been produced in E. coli in the laboratorysince the 1980s (26).

Even more disturbing is the finding that antibiotics canincrease the frequency of horizontal gene transfer 10–10000-fold (27–30). Thus, antibiotics create the very con-ditions facilitating the spread of antibiotic resistance. The

M.-W. Ho et al.36

evolution and spread of antibiotic resistance predate ge-netic engineering, and are largely due to the profligate useof antibiotics in intensive farming, and in medicine (31,32). However, the abuse of antibiotics per se does notaccount for the emergence of new viruses and new virulentstrains of bacteria. Fifty new viruses have been identifiedjust between 1988 and 1996 (33). The urgent question whichneeds to be addressed is the extent to which genetic engi-neering biotechnology, by facilitating horizontal gene trans-fer and recombination, is contributing to the resurgence ofinfectious, drug-resistant diseases, and will continue to do soif allowed to proceed unchecked.

This question is all the more important in the light ofyet other recent findings indicating that microorganismsgenetically engineered for ‘contained use’ may not beeffectively contained, while our regulators are contemplat-ing further relaxations of the guidelines on contained use1.Thus, biologically ‘crippled’ strains of bacteria have actu-ally been found to survive in the environment to exchangegenes with other species (reviewed in (34, 35)). Naked viralDNA may be more infectious than the virus itself (36).DNA released from cells is not readily broken down in theenvironment, thereby retaining the ability to transformorganisms (reviewed in (34, 37, 38)); and routine chemicaltreatments for inactivating pathogenic microorganisms andviruses, before they are discharged into the environment,may be ineffective, leaving a large percentage of pathogensin an active, infectious state (39). The combination ofevidence is sufficiently compelling, especially in view of theprecautionary principle, to warrant, at the very least, anindependent public enquiry into the possible contributionsof the industry to the etiology of infectious diseases.

This paper will review the following topics:

(a) The origins and nature of antibiotic resistance and itsspread by horizontal gene transfer and recombination.

(b) The mechanisms and scope of horizontal genetransfer.

(c) The evolution of virulence and its spread by horizontalgene transfer.

(d) Genetic engineering and horizontal gene transfer.(e) Other recent findings relevant to the possible contribu-

tion of genetic engineering biotechnology to the etiol-ogy of infectious diseases.

On account of the enormity of the subject areas, it isimpossible for us to be exhaustive in our review. Our goalis to supply enough information for readers to judge forthemselves whether the strength of the evidence togetherwith the precautionary principle will justify the publicenquiry proposed.

2. ANTIBIOTICS AND EMERGING PATHOGENS

When antibiotics were first introduced, microbial geneti-cists did not believe antibiotic resistance would pose any

problems, based on the very low spontaneous mutationrates observed (24). Mutation—random changes in thegenetic material—and selection—survival of the fittest—were thought, by most biologists, to sufficiently explain allcharacteristics of organisms. If the mutation rates are verylow, then ‘selection pressure’ would have to be very strongfor the character to spread.

Soon after antibiotics were used in medicine, it wasdiscovered that livestock lived longer and grew faster whenthey were fed antibiotics to overcome infections. This ledto ‘prophylactic’ treatments. In fact, antibiotics were rou-tinely fed to promote growth, to chickens, cattle, pigs anddairy cows, which also extended the shelf-life of meat,poultry, eggs and dairy product. In the 1970s. it wasdiscovered that chickens given high doses of antibioticsresulted in resistant Salmonella strains in both the meatand the eggs. By 1983, mutant strains of antibiotic resis-tant Salmonella that attacked human beings had emerged(6). The emergence of new pathogens is strongly associatedwith mutant strains that cross species barriers.

One of the clearest and most disturbing examples ofmutant bacteria crossing from domestic animals to humanbeings, and of the transformation of a previously benignstrain into a pathogen, is Escherichia coli, a commonbacterium inhabiting the intestine of all human beings andmany other mammalian species. Most of the time, E. coliis harmless, which is why it has been the major tool forgenetic engineers, who have used it and its plasmids rou-tinely to clone genes since the 1970s (40). E. coli is themost intensively manipulated organism, and genes fromspecies in every Kingdom have been transferred to andcloned in it. Perhaps it is not surprising that E. coli hasemerged as a major pathogen.

The most prominent among pathogenic strains, E. coli0157: H7, appeared in 1982, which, far from being benign,caused dangerous haemorrhages of the colon, bowel andkidneys of human beings; and it broke out suddenly andsimultaneously in several states in the USA (41). Like mostE. coli strains in the 1980s, it was moderately resistant tothe antibiotics ampicillin and tetracycline. Since then,many outbreaks have occurred all over the world, and inincreasing frequencies. A mass outbreak in Japan in 1996affected 10000, with 12 deaths in children. A series ofoutbreaks in Scotland in 1997 claimed 20 lives and madehundreds ill (42, 43).

Genetic fingerprinting—a technique that fragments theDNA using specific enzymes and reading the fragments—backs up epidemiological evidence that E. coli 0157:H7arose recently; and up to 1993, was endemic only to cattlein USA, Canada and Great Britain, and absent in cattle inthe Far East (Thailand) (44). E. coli mutates fairly rapidly,hence the low degree of genetic variation observed amongthe 0157:H7 isolates before 1993 suggests they originatedrecently from a single clone. Most cases came from con-taminated meat. However, outbreaks have also been asso-

Gene technology and gene ecology of infectious diseases 37

ciated with cheese, salami, raw vegetables, unpasteurizedapple juice and water (45). The E. coli 0157, like otherwater-borne pathogens such as Cryptosporidium, Giardiaand Shigella, had developed resistance to chlorine ((6),p.430). The mass outbreaks in Japan in 1996 were thoughtto be linked to white radishes (43), which may have comefrom contaminated soil. The specific toxicity of the strainis attributed to the Shiga-like toxin genes acquired byhorizontal gene transfer from Shigella (22, 46). One ofthese genes, VT1, is almost completely identical in basesequence to that of Shigella, indicating again, that the genetransfer was a very recent event.

Studies of dozens of emergent species of pathogensshowed that genes for antibiotic resistance and virulenceoften reside in the same regions of the bacterial DNA, inplasmids or transposons, which may be passed horizon-tally from one species to another, picking up extra genes,recombining and generating novel combinations in theprocess (6, 24).

2.1 A profusion of mechanisms and genes for antibioticresistance

A profusion of biochemical mechanisms and genes areinvolved in antibiotic resistances, fully matching, if notsurpassing the variety of targets at which antibiotics areaimed (7, 24, 46–48). Most antibiotics have a single pri-mary target, usually a single step in the synthesis ofmacromolecules essential to the bacterial cell, and prefer-ably do not affect their eukaryotic host cells. Thus severalclasses of antibiotics are aimed at different steps in synthe-sizing the bacterial cell wall. These include the penicillinsor b-lactams, vancomycin and ristocetin, D-cycloserine,fosfomycin and bacitracin. Other classes of antibiotics areaimed at bacterial enzymes involved in the synthesis ofDNA, RNA and proteins. These enzymes differ from theireukaryotic counterparts, although the effects of the antibi-otics may not always be so selective, and are accompaniedby varying degrees of toxicity for the eukaryotic host.Metronidazole, quinolones, nalidixic acid and novobiocintarget DNA and DNA synthesis, while actinomycin andrifamycin inhibit transcription (RNA synthesis). Transla-tion (protein synthesis) is targetted by a wide variety ofantibiotics: streptomycin and other aminoglycosides suchas kanamycin, neomycin, gentamicin, tobramycin, aikacinand netilmicin; tetracyclines, puromycin, chloramphenicol,erythromycin, lincomycin and clindamycin, and fusidicacid. In addition, antimetabolites such as sulfonamidesand trimethoprim inhibit the synthesis of essentialmetabolites.

Resistances to every class of antibiotics have been iden-tified (see Table II) [modified from (24)], the genes respon-sible may be found on the chromosome or in plasmids.

One of the commonest mechanism of resistance is thedegradation or modification of the antibiotic by specificenzymes. Thus, the b-lactams are degraded by lactamases

Table II

A summary of mechanisms for antibiotic resistance

AntibioticMechanism

b-lactams1. Modification of target toreduce or eliminatebinding of antibiotic

ErythromycinLincomycinStreptomycin

2. Inactivation of antibioticby enzyme

b-lactamsa. HydrolysisErythromycin

b. Derivatization AminoglycosidesChloramphenicalFosfomycinLincomycin

b-lactams3. Sequestration of antibioticby protein binding

Fusidic acid

4. Metabolic bypass SulfonamidesTrimethoprim

5. Binding of antibiotic by Bleomycinspecific immunity protein

6. Overproduction of Sulfonamidesantibiotic target

Trimethoprim

7. Reduced uptake Chloramphenicol

8. Active efflux from cell TetracyclineChloramphenicolMultiple resistance to broadranges of antibiotics

all of which have arisen from one ancestral gene. On theother hand, aminoglycosides are inactivated by specificmodifying enzymes encoded by 30 or more different genes,while chloramphenicol is inactivated by many acetyltrans-ferases encoded by at least a dozen genes. Mutation ofspecific antibiotic target is also frequently involved inresistances. For example, streptomycin is overcome inMycobacterium by mutations in the ribosomal RNA(rRNA) molecules required for protein synthesis, so thatthey have reduced affinity for streptomycin. Antibioticsmay be inactivated simply by sequestration, or bindingpermanently to specific proteins. A metabolic bypass mayappear to a reaction blocked by the antibiotic, or theinhibited protein may simply be overproduced by thebacteria in order to overcome the inhibition. Finally,resistance may be conferred by reduced uptake of theantibiotics or by increased rates of pumping them out ofthe cell-active efflux mechanisms.

Active efflux mechanisms are especially important forthe multiple resistance to broad ranges of structurallyunrelated antibiotics and drugs that have appeared in

M.-W. Ho et al.38

pathogenic bacteria within the past 10 years. These will bedealt with in Section 3.2.

Quite often, several biochemical mechanisms are in-volved in giving protection against a single antibiotic,resulting in a very high level of resistance. Cross resis-tances are also common within a class of antibiotics suchas the aminoglycosides, so at least one kanamycin resis-tance gene used as a genetic marker has been found toconfer cross-resistance to two new generation aminogly-cosides, amikacin and tobramycin (49). A recently docu-mented case of antibiotic cross resistance is even moreserious. Enterococcus faecium inhabiting the gut of turkeyson farms using avoparcin as growth promoter are found tohave developed cross-resistances to vancomycin and tei-coplanin (50), the last defence against multidrug resistantStaphlococcus aureus.

Where do the genes conferring antibiotic resistancescome from? And how do they manage to spread soquickly? Let us first address the question of the origin ofthe genes.

3. ORIGINS OF ANTIBIOTIC RESISTANCE GENES3.1 Biochemically specific mechanisms

There are several sources of antibiotic resistance genes[reviewed by Davies (24)]. The first is from organismsproducing the natural antibiotic, such as penicillin, whichalso produce the inactivating enzyme. As new b-lactamsare introduced, the genes encoding the b-lactamases un-dergo mutational changes, often involving only one basesubstitution and a single amino-acid difference in theenzyme. That would be sufficient to enable the enzyme torecognize the new drug and to break it down. In this way,the enzyme extends its spectrum of resistances to newdrugs. Numerous mutational variants of b-lactamase havealready been identified, showing the ease with which ap-propriate mutational changes can arise in this gene. Thevariants derive from different genes which are themselvestraced to a single ancestral gene. The versatility of theb-lactamase gene is such that when a series of inhibitorswere made against the enzyme and used in combinationwith the b-lactam antibiotic—a kind of belt and bracesapproach—mutant enzymes arose, which not only breakdown the new antibiotic, but are, at the same time insensi-tive to the inhibitor.

More often, however, it is not at all clear where theantibiotic resistance genes come from. New genes seem toappear from nowhere. For example, the 30 or more genesinvolved in resistances to different aminoglycosides are notrelated to one another. Some are thought to have arisenfrom mutations of pre-existing genes encoding enzymescatalyzing normal metabolic reactions in the cell. Othersmay involve activation of previously ‘cryptic’ or hiddengenes in the bacterial genome (see Section 3.2 below). Orthey may have been acquired by gene transfer from an-other bacterium.

3.2 Acti6e efflux mechanisms

Active efflux mechanisms are due to 20 or more genes in atleast 4 gene families (see Table III) [modified from (47)],members of which can be found on plasmids or on thechromosomes. The genes code for transporter proteinssituated across the cell membrane, that pump substancesout of the cell, using ATP or protons as the source ofenergy. Recently, it has been shown that the genes forseveral efflux pumps are modulated in response to environ-mental conditions, and that many of the systems are alsoinducible by the substances to which they give resistance.

The chromosomally encoded efflux genes in E. colibelong to one or another of four inducible operons,marRAB, soxRS, emrAB and acrAB (reviewed in (26)).There is a great deal of homology among the genescomprising different operons. For example, the proteinsencoded by the regulatory genes, marR and emrR are twoof the 15 proteins belonging to the MarR family, membersof which are found, not only in Gram negative bacteriainhabiting the gut, but also in Gram positive bacteria, andthe mycobacteria, including plant and animal pathogens,as well as free-living species.

One member of the MarR family, HpcR, found insoil-dwelling E. coli C strains, is a repressor of an operonencoding the enzymes that break down homoprotocate-chuate (Hpc), a phenolic compound derived from plants.The repressor protein, HpcR. binds to the operator andrepresses the synthesis of enzymes that breakdown Hpcwhen Hpc is absent. But when Hpc is present, the repres-sor protein binds to it instead, freeing the operator, so thattranscription and synthesis of the enzymes can takes place.These observations suggest that naturally occurring nox-ious substances in the environment are the original induc-ers of the multiple antibiotic resistance operons.

Further observations in support of the idea are thatother naturally occurring compounds, such as salicylicacid, dinitrophenol and naphtho-quionones, stimulate theexpression of both soxS and marRAB. The natural naph-thoquinolones, plumagin and menadine, are effective in-ducers at relatively low concentrations, and mutantslacking both the marRAB and soxRS operons are notablyhypersensitive to those compounds, while sensitivity tosalicylic acid is only marginally affected. The marRABoperon expression is enhanced at 30 °C compared with37 °C, as consistent with its origin in free-living bacteria(26). The multiple antibiotic resistance operons have,therefore, evolved from free-living bacteria. The entericbacteria have simply reactivated or mutated the genes inthe operons to allow them to cope with drugs and antibi-otics. This class of mechanisms is reminiscent of the multi-drug resistance which all mammalian and plant cells areable to develop in the presence of sublethal doses ofnoxious substances (see next Section).

Gene technology and gene ecology of infectious diseases 39

Table III

Examples of acti6e efflux systems in bacteria

CCCP, carbonyl cyanide m-chlorophenylhydrazone, SDS, sodium dodecylsulphate

OrganismTransporter Antibiotics

1a. Major facilitator in Gram-positive bacteria*Streptomyces rimosusOtrB OxytetracyclineBacillus subtilis TetracyclineTet(L)various cocciStreptomyces coelicolorMmr MethylenomycinS. coelicolorActII Actinorhodin

TetracenomycinStaphylococcus glaucescensTcmANorA Staphylococcus aureus Fluroquinolones, basic dyes, puromycin chloramphenicol, tetraphenylphosphonium

Quaternary ammonium compounds, basic dyesS. aureusQacAB. subtilisBmr Basic dyes, chloramphenicol, puromycin, fluoroquinolones

1b. Major facilitator in Gram-negative bacteria*TetracyclineE. coliTetAChloramphenicolCmlA Pseudomonus caeruginosaBicyclomycinE. coliBcr

E. coliEmrB CCCP, nalidixic acid, tetrachlorosalicylanilide phenylmercury acetateE. coli CCCP, phenylmercury acetateEmrD

2. RND familyAcreE E. coli Basic dyes, SDS, erythromycin, novobiocin, fusidic acid, tetracycline, mitomycin C

As above, plus othersE. coliEn6DP. aeruginosaMexB Tetracycline, chloramphicol, fluroquinolones, b-lactams, pyoverdine

3. Smr familyQuarternary ammonium compounds, basic dyesS. aureusSmrAs aboveQacE Klebsiella aerogenesMethyltriphenylphosphonium, basic dyes.E. coliM6rC

4. ABC transportersStaphylococcus epidermidisMsrA 14- and 15-membered macrolides

Daunorubicin, doxoribucinStreptomyces peucetiusDrrA, BTylosinTlrC, B Streptomyces fradiae

* Gram positive and negative bacteria are the traditional major divisions of bacteria. The Gram positive bacteria are surrounded by athick peptidoglycan cell wall, which, however is permeable to small molecules; Gram negative bacteria, such as E. coli, are surroundedby a second membrane, which functions as an effective barrier to small molecules.

4. ‘RANDOM’ VERSUS ‘DIRECTED’ MUTATIONS

To measure the rate of ‘spontaneous’ mutations to antibi-otic resistance in a bacterial strain, a small amount of adilute suspension of bacteria is spread on a nutrient agar-plate impregnated with a concentration of the antibiotichigh enough to kill nonresistant bacteria, so that onlythose with a pre-existing mutation which happens to con-fer resistance will survive and form colonies. The sponta-neous mutation rate for antibiotic resistance estimated inthis way is typically 10−9 or lower, and is about the samefor all other genes in the genome. This is the rate ofrandom mutations, in other words, mutations that have nodirect correlation to the environment, or the ‘selective’regime. Based on these estimates, microbiologists did notbelieve that resistance to antibiotics would pose a problemwhen antibiotics first came into use (24). So why didantibiotic resistance evolve and spread so rapidly?

The usual explanation for antibiotic resistance, accord-ing to neo-Darwinian theory, is that it is due to the intense‘selection pressures’ exerted by the antibiotics, so that evenextremely low rates of spontaneous random mutations thathappen to confer resistance to the antibiotic will be rapidlyselected for, while those which do not happen to have themutation will die out (see for example (32)). Antibioticresistance has been assumed to be a classic case of thenatural selection of random mutations. What actuallyhappens is something else.

First of all, the mutations are not random in the sensethat they pre-exist in the population, before exposure toantibiotics. All bacteria, with the exception of those thatlive in hot springs, are killed by boiling or cooking, whichis therefore the surest way to prevent infection. They canalso be killed by high enough concentrations of antisepticsor antibiotics to which they are susceptible. However, at

M.-W. Ho et al.40

lower concentrations of the noxious agent, the bacteriummay fail to grow at first, but after some time, develop therequired resistance to the drug which enables it to grow,multiply and form colonies. This ability to develop drugresistance is common to cells of all species—bacteria,fungi, plant and animal without exception (51–54). It is aphysiological response shared by all the cells or organismsin the population, and is not due to selection of pre-exist-ing random mutations. It is similar to the ‘directed muta-tions’ or ‘adaptive mutations’ (see Section 5 below),whereby starving, non-growing bacteria eventually acquiremutations that enable them to feed on new substrates(55–57). Secondly, competition is irrelevant among bacte-ria in the evolution of antibiotic resistance: they share theirmost valuable assets for survival—genes encoding resis-tance mechanisms against antibiotics—by passing them onhorizontally between species.

In fact, the evolution of antibiotic resistance turns out tobe a paradigm case of the ‘fluid genome’ in the newgenetics, in which genes and genomes can mutate andundergo other changes when the ecological conditions callfor them, and genes can even travel through the externalenvironment from one organism to another (35, 58). Weshall first consider ‘adaptive mutations’ in the broadercontext of the lessons which have been learned frommolecular genetics within the past two decades.

5. ‘NATURAL’ GENETIC ENGINEERING AND‘ADAPTIVE MUTATIONS’

In a recent review, Shapiro (59) offers an appropriateperspective of bacterial evolution that sums up the recentdiscoveries of molecular genetics on how genomes areorganized, on the biochemical mechanisms of geneticchange and the cellular control networks regulating geneticchange.

Bacterial genomes are organized in modular fashion atsuccessive levels of a genetic hierarchy. At the lowest levelof the hierarchy are genes composed of different domains(or subunits) that are reshuffled and recombined to givedifferent genes. Different combinations of similar genesgive rise to multigene units or operons, and differentcombinations of operons make up larger units or ‘regu-lons’ for complex functions such as chemotaxis, symbiosisand protein secretion. Regulons may coincide with inde-pendently replicating ‘replicons’, such as plasmids andviruses, each with its own signal for replication (origin ofreplication). Thus, mosaicism, or cut-and-splice geneticrecombination is involved in every level of genome organi-zation, and plays a key role in generating genetic diversityrapidly. But these mechanisms are by no means random.

Bacteria may have ‘little tolerance for purely randomvariability’ (59). Bacterial cells possess a wide range ofrepair and proof-reading functions to remove accidentalchange to DNA sequences and correct errors resulting

from physiological insults. But the same cells also possessnumerous biochemical mechanisms for changing and reor-ganizing DNA, suggesting that such changes are the resultof the ‘natural genetic engineering’ ability of bacteria torespond to environmental challenges.

The possibility of ‘natural’ genetic engineering mecha-nisms responding to the environment gained considerablesupport in the discovery of ‘adaptive mutation’, in whichstarving bacteria acquire the appropriate genetic mutationenabling them to metabolize a new substrate, when thesubstrate is present in the ‘selective’ medium. A widevariety of mechanisms are involved in generating adaptivemutations, including point mutations (single basechanges), deletions or insertions of nucleotides, conver-sions of gene sequences from one to another, site-specificintegration of genetic material, DNA rearrangement andtransposition (gene jumping) [reviewed in (59)].

There is much current debate as to whether those muta-tions are specifically ‘directed’ by the environmental chal-lenge, or simply the consequence of a hyper-mutable stateof stressed cells that generates many mutations indiscrimi-nately at a high rate, increasing the chances of a lucky hit.In either case, however, it means that genes and genomeschange as the result of built-in mechanisms that are subjectto metabolic and environmental feedback regulation (57).As Shapiro points out, the ability to activate those mecha-nisms under stress can significantly accelerate evolutionarychange in crisis without threatening genetic stability underordinary circumstances. He surmises that further studies ofmobile genetic elements and DNA rearrangements willultimately uncover truly directed mutations. Just as signal-transduction systems of the cell can direct the transcrip-tional apparatus to specific locations in the genome toactivate those genes, so it is possible that the mutationalapparatus(es) can be similarly directed to mutate specificgenes.

Shapiro regards the evolution of antibiotic resistanceand virulence as clear examples of ‘natural’ genetic engi-neering mechanisms at work, and goes no further thanthat. This can easily lead human genetic engineers tojustify what they do on grounds that bacteria themselveshave been at it, probably for hundreds of millions of years.That kind of reasoning is fallacious and dangerous. Asrecent studies indicate, ‘normal circumstances’ for the vastmajority of microbes in the environment as well as in ourbodies may be a non-proliferating but metabolically active,non-virulent state (60, 61). This state is associated withlayers of extracellular matrices or ‘biofilms’ in which typi-cally mixed communities of microbes are embedded, wherethey enjoy a cooperative, multicellular existence rather likethe cells and tissues in our body [see Ho (35), Chapter 14].

Furthermore, there is nothing ‘natural’ about the ‘man-made’ environmental conditions that have provoked mi-crobes to engineer their genes and genomes. The dischargeof toxic wastes, the destruction of the environment, and

Gene technology and gene ecology of infectious diseases 41

the overuse and abuse of antibiotics, have already thrownthe ‘natural’ ecology of microbes seriously out of balance.And over the past 20 years, the extensive manipulation ofgenes and genomes by human genetic engineers is verylikely to have exacerbated the situation. The ecology ofmicrobes consists, not only of the usually understoodinterrelationships with other organisms, but also of geneswithin the genome, of the genome within the physiologicalstate of the cell and the organism, which is in constantintercommunication with the external environment. Dis-ease, we are told, cannot be understood apart from theecological context (62). More importantly, it must beunderstood in terms of the nested levels in the ecology ofgenes (36) and microbes, which is inextricably entangledwith the ecology of human beings.

6. HORIZONTAL GENE TRANSFER

The major role of horizontal gene transfer in spreadingantibiotic resistance has only been recognized since the late1980s, when a Conference on Antibiotic Resistance wasco-sponsored by the Environment Protection Agency andthe National Science Foundation of the USA to look intothe mechanisms and factors affecting such gene transfer(63). The first definitive evidence for horizontal gene trans-fer came from DNA sequence analysis of the genes forneomycin-kanamyin resistance from Staphylococcus au-reus, Streptococci and Campylobacter sp. (64) which werefound to be essentially identical. Identity of sequence notonly gives evidence for horizontal gene transfer, but alsoindicates that the transfer is a very recent event. Furtherevidence of horizontal gene transfer was subsequently ob-tained for a wide range of antibiotic resistances, by DNAsequencing or other methods of characterizing DNA (65–70).

6.1 The scope of horizontal gene transfer

The scope of horizontal gene transfer is essentially theentire biosphere (13, 14, 36, 38, 71–77). Genes are foundto have been transferred, not only among microorganismsand viruses, but among eukaryote species and across tradi-tional Kingdoms of organisms. Among eukaryotes, themain transfers have involved transposable elements (36,77) such as the P element among Drosophilids, and trans-posable introns among lower eukaryotes and bacteria (36).Some especially promiscuous elements such as marinerhave been found in arthropods and vertebrates, includingprimates and humans, where it is associated with a neuro-logical wasting disease (78). Gene transfer mediated bytransposable elements do not require DNA base sequencehomologies.

Transfers have occurred from bacteria to higher plantsand 6ice 6ersa. The best known example is the directdemonstration of transfer between the soil bacterium,Agrobacterium and plants. In a process bearing a strong

resemblance to conjugation between bacteria (see Section7.3), the tumour (T) segment of the tumour-inducing (Ti )bacterial plasmid is transferred and incorporated into theplant genome (79, 80). However, it must be noted thatnearly all cases of directly demonstrated horizontal genetransfer, especially those involving phylogenetically distantspecies, made use of already modified hybrid shuttle vec-tors that can transfer between species and replicate inboth. These shuttle vectors possess signals for replication(origins of replication) in more than one species as well asthe signal for DNA transfer (origin of transfer), and arehence much more likely to be successful in horizontaltransfer than unmodified plasmids found naturally. The Tiplasmid is, indeed, the basis of a gene transfer vectorsystem widely used for genetically engineering crop-plants(see Table IV).

Cross-Kingdom horizontal gene transfer by conjugationhas also been demonstrated between bacteria and yeastusing shuttle vectors derived from broad host-rangepromiscuous plasmids (see Section 7.3) which are alreadytransferable between many bacterial species (71, 74, 81).Such vectors can even substitute for the Ti plasmid intransferring genes from Agrobacterium to plants (82). Re-cent evidence documented the direct transfer of transgenesand marker genes from transgenic plants to soil fungi (83)and soil bacteria (84), indicating that secondary, unin-tended gene transfers can occur from genetically engi-neered crop-plants which are now released commerciallyinto the environment. Despite the title of their publication,Schlutter et al. (84) actually observed a high ‘optimal’ genetransfer frequency of 6.2×10−2 in the laboratory, fromwhich they ‘calculated’ a frequency of 2.0×10−17 underextrapolated ‘natural conditions’. The natural conditions,are of course, largely unknown (36). In a recent Workshopat the German Environmental Protection Agency, tworesearch groups reported secondary transfer of antibioticresistance marker genes from transgenic plants to soilmicroorganisms (85, 86). Debris and exudates from trans-genic plants will, therefore, be expected to transform bac-teria and other microorganisms in the soil. E. coli cellshave been successfully transformed by T-DNA, simplyusing total DNA from a transgenic plant; and the T-DNAvector was recovered as a circular plasmid in the trans-formed cells (79). In addition, other vectors such asviruses, insects and nematodes can also mediate unin-tended, secondary horizontal gene transfer (36).

Mammalian cells are known to take up foreign genes bya number of mechanisms including anastomosis, endosym-biosis, phagocytosis (87, 88) and also endocytosis, a processdependent on specific cell-membrane receptors binding toDNA or to macromolecules (most frequently proteins)complexed to DNA, so that the DNA can be internalized(36). It has even been suspected that gene transfer frombacteria to animal cells can occur by conjugation, a matingprocess previously thought to be confined to bacterial cells

M.-W. Ho et al.42

(see Section 11.3). The ability of certain viruses (retro-viruses) to capture and transfer genes among mammaliancells is well-known and is associated with many animalcancers (see entries in (89)). Numerous illegitimate recom-binations requiring no sequence homologies are due toretroviral insertion into the genome transferring geneshorizontally between individuals of the same or differentspecies (36). In short, a given gene from any organism hasthe potential to spread to all other organisms in thebiosphere (13, 14).

However, DNA uptake by itself is not sufficient forsuccessful gene transfer. Important barriers operate afterthe uptake of foreign DNA which break down, or other-wise block the integration, use and maintenance of theforeign genes (40, 77). Plasmid DNA requires an origin ofreplication that is recognized and used by the host cellbefore it can be replicated and maintained. Similarly,promoter sequences which can be recognized by the hostare necessary for expressing the viral genes that make thevirus, as well as expressing any other foreign genes carriedby the virus. (Though, of course, origin of replication andpromoter sequences can be acquired by recombination.)However, degraded DNA coding for parts or domains ofproteins, or short regulatory motifs that bind transcriptionfactors or otherwise enhance or repress gene expressionmay also be integrated, with significant genetic effectsincluding cancer (36). Recombination of domains is al-ready known to be a major factor in generating newproteins (59).

While the host cell DNA has its own origins of replica-tion, methylation patterns (chemical markers) and pro-moter sequences which are specific to the species, artificialgene transfer vectors and other manipulated DNA haverecombined origins of replication and promoter sequencesfrom different species which can be recognized by differenthost species. In addition, they are designed to overcomethe restriction systems which break down and ‘silence’foreign DNA. Thus, genetic engineering may ha6e creatednew a6enues and possibilities for increasing the scope ofhorizontal gene transfer. We shall consider this in moredetail in Section 10.

The scope of horizontal gene transfer among bacterialspecies has to be appreciated in the light of the revisedphylogeny of the living world based on comparing riboso-mal RNA (rRNA) sequences, a procedure first proposedby Carl Woese (90). Ribosomal RNAs are the most con-servative molecules in evolution, and have not been subjectto horizontal transfer between species. For these reasons,phylogenies based on rRNA sequences are generally con-sidered to be the most reliable.

Comparisons of rRNA and other conserved sequencesgive a universal phylogenetic tree (Fig. 1) (91), revealingthe true phylogenetic status of bacterial species. The tree istripartite, with Bacteria (Eubacteria), Archaea (Archebac-teria) and Eucarya (Eukaryotes) forming three more or

less equal Domains. The traditional Kingdoms—Plants,Animals and Fungi (Coprinus)—are but twigs on a smallperipheral branch of the Eucarya Domain, each twig hav-ing the same phylogenetic status as traditional Orders ofbacteria. Within the Bacteria Domain, E. coli (a Gram-negative bacterium) is as distant from the Gram-positivebacteria, Bacillus, or Heliobacterium, as Homo sapiens isfrom the single-celled ciliate, Paramecium.

But even the revised phylogeny of the living worldgreatly underestimates the genetic diversity of Bacteria.The known bacterial species are less than 1% of all speciesthat exist, as more than 99% cannot be cultivated in thelaboratory by routine techniques (91). This highlights thedangers of releasing genetically engineered microorganismsor DNA into the environment.

All known pathogens belong to the Bacteria and someto the Eucarya Domain. None belongs to Archaea despitethe fact that some archaean species share our environmentand are part of our normal bowel flora. Strauss andFalkow (92) suggest there may be more fundamental dif-ferences between organisms that can become pathogensand those that do not. Another reason, however, may bethat, until fairly recently, archaean organisms have notbeen genetically manipulated to the same extent as havebacterial species. Will the current trend towards increasingmanipulation of archaean species also turn previouslybenign organisms into pathogens? Judging from the recenthistory of genetic engineering, it may only take a few yearsof intensive genetic manipulation of new (archaean) speciesfor new pathogens to emerge. There are essentially nobarriers left to horizontal gene transfer and recombination,particularly as the result of genetic engineering.

7. MECHANISMS OF HORIZONTAL GENETRANSFER

There are three main mechanisms for gene transfer inbacteria, 1) by direct uptake of DNA (transformation) 2)through genes being carried by viruses that infect bacterialcells (transduction) and 3) through a mating process re-quiring cell to cell contact (conjugation). These processeshave been studied most extensively in E. coli, although it isbecoming clear that what is true for E. coli does notnecessarily apply to other species. The scope of our presentignorance is enormous, but what is known already gives usserious cause for concern.

7.1 Transformation frequencies are high

Transformation among microbes in the environment hasbeen extensively reviewed recently (36–38), showing that itis extremely widespread. The process involves DNA beingreleased into the environment, its binding to the surface ofbacteria, entry into the cell and recombination with thebacterial chromosome, or in the case of plasmids, reconsti-tuting the plasmid. All these stages are facilitated by

Gene technology and gene ecology of infectious diseases 43

Fig. 1. The universalphylogenetic tree of the livingworld.

bacterial proteins. Both chromosomal and plasmid DNAare able to transform bacteria, the frequency of transfor-mation varying with the physiological state of the cell,diffusible factors excreted from the bacteria and the pres-ence of salts in the environment. In addition, the nature ofthe DNA, its source, size, and state may also be important.Generally, transformation is most frequent if the DNA isfrom the same species, as its successful integration into therecipient genome depends on the degree of homology, i.e.the similarity in base sequence between the transformingDNA and the host DNA. It also depends on the DNAbeing resistant to the restriction enzymes in the host cellwhich break down foreign DNA. In one experiment, briefheat treatment (9 min at 48.5 °C) dramatically increasedthe competence of cells for receiving foreign DNA by 4orders of magnitude (73). This was attributed to the heatinactivation of restriction enzymes, thereby increasing the

success of gene transfer. Cross species, cross-genera andeven cross-order transfers have been observed with chro-mosomal DNA. Plasmid DNA, in particular, has effectedcross-Kingdom transformations among Eubacteria, Pro-teobacteria and Cyanobacteria (37).

DNA is not only released into the environment whenthe cells die, but is actively excreted by living cells duringgrowth. Some species export DNA wrapped in membrane-bound vesicles. The DNA in a culture slime can be morethan 40% of the dry weight. Thus, the environment isextremely rich in DNA. Marine water contains between0.2 to 44 mg per liter, and fresh water contains between 0.5to 7.8 mg per liter; while freshwater sediment has an upperconcentration of 1 mg per gram. Although enzymes break-ing down DNA (deoxyribonucleases, DNases) are found inthe environment, DNA is protected from degradation byadsorbing to detritis, humic acid, and in particular, clay

M.-W. Ho et al.44

and sand particles. Adsorbed DNA is equally efficient intransforming cells. Thus, the half-lives of DNA in soil is9.1 h for loamy sand soil, 15.1 h for silty clay soil and 28.2h for clay soil. While half-lives (time for half of the DNAto be broken down) in waste water are typically fractionsof an hour, those in freshwater and marine water are 3 to5 h, with high values of 45 to 83 h on the ocean surface,and extremely high values of 140 and 235 h for the marinesediment (37). Adsorption of DNA to solid particles is avery rapid process, which means that DNA released intothe environment can survive indefinitely and maintain itspotential to transform other organisms. Transformationmay be a major route of horizontal gene transfer: frequen-cies obtained under different environmental conditions,using artificial 6ectors and markers are found to be gener-ally quite high, many ranging between 10−2 to 10−5

transformant per viable cell. In fact, transformation fre-quencies are often higher under natural conditions than inthe laboratory. In a recent study in soil microcosms,activity due to earthworms was found to significantlyenhance gene transfer between spatially separated bacterialspecies innoculated into the soil (93). A special form ofbidirectional transformation by cell contact and fusion isnow known to be widespread (38).

The release, or escape of naked recombinant DNA intothe environment is a growing ecological problem, not onlyin the form of transgenic organisms, but also in vaccines,gene therapy vectors, and antisense oligonucleotide—shortsequences of RNA complementary to the base sequence inthe transcript of the genes—and ribozymes—RNA thatact as enzymes to splice out intron sequences from tran-scribed genes—used in various forms of gene therapy (36).As mentioned earlier, even short sequences can have sig-nificant genetic effects when integrated into the genome oforganisms for which they are not intended. The hazard ofnaked DNA is highlighted by the discovery that some viralDNA may infect non-permissive hosts for which the virusitself is noninfective (36). Intravenous injection of naked,circular polyoma viral DNA into rabbits and mice gave apowerful genetic expression with the development of anti-bodies and the production of new virus particles (94),whereas the native virus had no effect. Similarly, intra-venous injections of recombinant plasmids in liposomesgave gene expression in several organs including theovaries, indicating the possibility of an inherited geneticchange (36). These findings are the basis for developingmany recombinant DNA vaccines (95), for which ecologi-cal risk assessment has hardly even been considered.

DNA vaccines are inherently dangerous due to theirability to infect non-permissive hosts, thus crossing widespecies barriers, with many opportunities for recombina-tion to generate new viruses. Recombinant DNA vaccinesare worse, as even small modifications of a virus can resultin significant, unpredictable changes in its host range andvirulence (36). In addition, they are more prone to recom-

bine with other viruses to generate new viruses. Similarhazards are associated with the gene therapy vectors whichare currently being developed (see also Section 11.3).

7.2 Transduction is substantial in aquatic en6ironments

Transduction takes place by means of bacteriophageswhich are usually host specific, and are hence not expectedto transfer genes between unrelated species. However,promiscuous phages with broad host ranges are capable oftransducing genes between unrelated species (38). A bacte-riophage consists of viral DNA wrapped in a protein coat,which binds to specific receptor sites on the bacteria. Oncebound to the receptor sites, the bacteriophage injects itsDNA into the bacterium. The internalized phage DNAeither directs the synthesis of many bacteriophages whicheventually break open (lyse) the cell, or the phage DNAmay insert itself into the bacterial chromosome (as pro6irusor prophage) to become replicated with the chromosome.The bacteriophage can transfer genes by mispackaging apiece of bacterial DNA in the viral protein coat, or bypackaging a viral DNA that has a piece of bacterial DNAincorporated into it. When such a virus infects a secondcell, the bacterial DNA is transferred and may then be-come integrated into the new bacterial chromosome. Thetransducing phage may serve as a vehicle for a transposonthat can then establish itself in an unrelated host, eventhough the phage DNA cannot do so, and hence con-tribute to cross-species, cross-genera gene transfer. Also,bacteriophages themselves evolve by horizontal gene trans-fer and recombination, leading to broadening of their hostranges (96). One study (97) found as many as 108 bacterio-phages per milliliter in aquatic environments, so that one-third of the total bacterial populations is subjected to aphage attack every 24 h, each capable of transducingforeign genes into the cell. Recently, Miller (98) reports1011 virus particles per milliliter in fresh and marinewaters.

Phages are now known to be intimately involved in thehorizontal transfer of entire blocks of virulence genes intothe bacterial chromosome, turning a previously benignstrain into a pathogen in a single step. Co-integrates ofphage and plasmids are also involved in the trafficking ofblocks of virulence genes (see Section 9.1).

7.3 Conjugation is essentially promiscuous

Conjugation was previously thought to be a species-spe-cific process requiring complex, complementary interac-tions between the genetic ‘donor’ and the genetic‘recipient’. Conjugation depends on certain conjugativeplasmids, extrachromosomal pieces of DNA that are repli-cated independently of the chromosome. They carry genesrequired for the process of conjugation. These plasmidsfall into different incompatibility groups. Ippen-Ihler andSkurray (99) listed seven or more incompatibility groupsof over a hundred known plasmids, most of them carrying

Gene technology and gene ecology of infectious diseases 45

one or more antibiotic resistance genes. Plasmids in thesame incompatibility group cannot be maintained in thesame bacterial cell at the same time. However, there arecommon mechanisms for gene transfer that are inter-changeable. One of these codes for the pilus, a tube-likestructure connecting the donor with recipient cells in con-jugation, through which plasmid DNA passes from thedonor to the recipient. In other cases, such as the genesinvolved in the transfer of the plasmid, the gene productsare specific to the plasmid and its host, and not inter-changeable. Several of the incompatibility groups comprise‘promiscuous plasmids’ with broad host ranges.

Conjugative plasmids can integrate themselves into thebacterial chromosome. In such cases, when conjugationtakes place, the integrated plasmid will drag the bacterialchromosome with it across the conjugation tube fromdonor to recipient, leading to transfer of bacterial DNA aswell as plasmid DNA. Donor DNA will then recombinewith recipient DNA to generate new genetic recombina-tions. The DNA is thought to be passed in a singlestranded form, which is resistant to breakdown by en-zymes attacking double-stranded DNA (nucleases and re-striction enzymes), thus ensuring a high success rate ofgene transfer. The single-stranded DNA then directs thesynthesis of the complementary strand to restore double-stranded DNA in both donor and recipient.

Recent studies indicate that conjugation has an extraor-dinarily wide host range, involving diverse, complex mech-anisms that are not yet well understood (38, 100).‘Promiscuous’ plasmids overcome species barriers, trans-ferring genes promiscuously between phylogenetically dis-tant species. At least several of the promiscuousconjugative plasmids have been exploited in constructingartificial shuttle vectors between E. coli and other distantspecies. One promiscuous conjugative plasmid is routinelyused in Agrobacterium-mediated gene transfer to maketransgenic plants (see Table IV). Shuttle vectors are essen-tially unstoppable, as they typically carry the origins ofreplication for both species as well as the origin of con-jugative transfer. The strength of conjugative transfer is inthe single-stranded DNA transferred directly, which es-capes destruction by restriction enzymes. While the shuttlevectors are constructed to require ‘helper’ transfer func-tions provided by other plasmids, those plasmids can bepresent in the donor or recipient, or in a third speciesaltogether. These ‘tripartite’ matings can easily occur in allenvironments, which are full of known and especially,unknown bacterial species.

In addition to conjugative plasmids, conjugative trans-posons mediate their own transfer to recipient cells duringconjugation and insert into the recipient chromosome (101,102). They can also jump from one site to another on thesame chromosome, or from chromosome into plasmidsand 6ice 6ersa. These have also been exploited in construct-ing gene transfer vectors (see Table IV). A special class of

such conjugative transposons, ‘integrons’ (103–105) arethought to be especially involved in the evolution of themultiresistance R plasmids within the past five decades(24). An integron carries its own gene encoding an enzyme,integrase, that catalyzes the site-specific integration of an-tibiotic resistant gene ‘cassettes’ within the integron, suchthat each inserted ‘cassette’ is provided with a ready-madepromoter for expression. An integron can carry severalcassettes, each encoding a different antibiotic resistance.The integron can facilitate recombination between cas-settes to form exotic gene-fusions that code for multifunc-tional proteins. Integrons can also jump from the bacterialchromosome into plasmids and become transferred toanother bacterium during conjugation.

During conjugation, genetic material not only goes fromdonor to recipient, but also in reverse, from recipient todonor, in a recently discovered process, retrotransfer (38).

7.4 Effect of en6ironmental pollutants on horizontal genetransfer

Horizontal gene transfer is sensitive to many physical andchemical factors. In routine laboratory procedures forgenetic transformation, heavy metals, singly or in combi-nation, are used to greatly increase the competence of cellsfor transformation, as are enzymes such as muraminidasesor proteases. Freezing and thawing, treatment with EDTAor other chelating agents, exposing cells to electric fields,or to DNA wrapped in liposomes (vesicles made frommembrane lipids), are all effective in increasing the fre-quency of transformation. Yet, very few experimental in-vestigations have been carried out into the effects ofenvironmental pollutants on horizontal gene transfer. Thisis a glaring omission in view of the hundreds, if notthousands, of xenobiotics that currently circulate in ourenvironment. And to which genetic engineering biotech-nology is releasing, or discharging large volumes ofagents—transgenic plants and animals, microorganisms,vectors, vaccines, and a range of naked DNA and oligonu-cleotides—all capable of taking part in horizontal genetransfer (36). We shall define xenobiotics as compoundsreleased by human beings into the environment in concen-trations that create harmful effects.

Xenobiotics can affect horizontal gene transfer in twoways (36). Some xenobiotics mutate DNA, thus influenc-ing its uptake into cells. Kipling and Kearsey (106) havereported examples of minor changes in a DNA sequencewhich alter the host range of a transferable genetic ele-ment. Other xenobiotics can affect the cell membraneand/or intracellular functions thereby influencing the abil-ity of cells to take up DNA, or to engage in conjugationand other forms of gene exchange.

As mentioned in the Introduction, antibiotics can in-crease horizontal gene transfer 10 to 10000-fold. Antibi-otics are heavily used in medicine and intensiveagriculture, especially in fish farming, and constitute an

M.-W. Ho et al.46

important class of xenobiotics that already affect horizon-tal gene transfer significantly. It is known that ancestralantibiotic resistance genes have originated from antibiotic-producing microorganisms themselves. The possibility hasbeen suggested (24), therefore, that what we call antibioticsmay actually be sex hormones for the bacteria, enhancingconjugation between cells. The process is naturally regu-lated by the product of the ‘antibiotic resistance’ gene,which shuts off the signal. Thus, the use of antibiotics notonly leads to directed mutations of antibiotic resistancegenes (see Section 5), it promotes the spread of those verygenes among pathogens. The case for phasing out the useof antibiotics in intensive agriculture and fish farming isoverwhelming. Antibiotics are also routinely used in ge-netic engineering biotechnology to select for transformedcells and recombinants.

The importance of synergistic interactions betweenxenobiotics is increasingly recognized. Recent research hasshown that chemicals that are apparently harmless whentested singly, can have enhanced effects in mixtures at verylow concentrations (107). As xenobiotics never occursingly in the environment, it is necessary to test drugs andchemicals in combination for realistic risk assessment. Butthis is a logistically and financially impossible task. To testjust 1000 of the commonest toxic chemicals in combina-tions of three (at a standardized dosage) would require atleast 166 million different experiments. An increasing num-ber of scientists are calling for a drastic reduction in thenumber of chemicals to which we are exposed. They wishto see the introduction of reverse onus, in which a chemi-cal must be regarded as potentially harmful until provenotherwise. Xenobiotics are not made by living organisms,possibly for the good reason that they are incompatiblewith healthy cellular activities. The synergistic effects ofxenobiotics on horizontal gene transfer are, for the mostpart, unknown and unpredictable. The precautionary prin-ciple demands that we should strive to limit the release/es-cape of genetically engineered organisms, naked DNA andRNA sequences into the environment as much as possible.

8. HORIZONTAL TRANSFER OF ANTIBIOTICRESISTANCE GENES

There could be very few barriers remaining to horizontalgene transfer. Antibiotic resistance genes, especially thosecarried on R (Resistance) plasmids and transposons, can inprinciple, cross species as well as genera and Kingdomsand even Domains. Evidence for transfer is obtainedwherever and whenever it has been looked for, either bysequence analyses or by direct experimental investigations,as mentioned in Sections 1 and 6. We cite further examplesbelow.

Tetracycline resistant genes are now found to be sharedbetween many genera of bacteria, one of them, the TetLgene, is found in Actinomyces, Clostridium, Enterococcus,

Listeria, Peptostreptococcus, Streptococcus, Staphylococcusand Bacillus. Another gene, the TetK, is found in Bacillus,Clostridium, Enterocococcus, Eubacterium, Listeria, Pep-tostreptococcus, Staphylococcus and Streptococcus. (108–110). More recently, both TetL and TetK have been foundfor the first time in soil bacterial species belonging to thegenera Mycobacterium and Streptomyces, which can causesoft-tissue and skin infections (111). Antibiotic producingStreptomyces are believed to be the ancestral source ofmany aminoglycoside antibiotic resistance genes, but it isnow gaining more of these genes back from the otherspecies. These findings are of particular significance, asthey ‘suggest the potential for the spread of an antibioticresistance gene into all environmental mycobacteria, in-cluding Mycobacterium leprae.’((111), p. 1411).

Even chromosomally encoded penicillin-resistance genes(penicillin-binding proteins, pbps, required for the uptakeof penicillin) are readily transferred. Transfers of pbp geneswere found between Streptococcus pneumoniae and Neisse-ria gonorrhea which subsequently recombined to generatenew hybrid genes (112).

9. HORIZONTAL GENE TRANSFER AND THEEMERGENCE OF NEW AND OLD VIRULENTSTRAINS

The very same genetic mechanisms for horizontal genetransfer have been involved in the emergence of virulenceamong old and new pathogens since the mid 1980s. Forexample, there has been an increase since then, in sporadiccases of very severe invasive Streptococcus pyogenes infec-tions in Europe, North America and elsewhere. In ananalysis of 108 isolates from patients in the USA withStreptococal toxic-shock syndrome (113). A toxin fre-quently found is encoded by a gene, SpeA, belonging to abacteriophage, which has become inserted into the bacte-rial genome. This gene has spread horizontally amongdifferent strains of S. pyogenes. In addition, genes encod-ing membrane surface proteins which determine bindingproperties and virulence of the bacteria, have also under-gone numerous recombination events subsequent to hori-zontal gene transfer, giving rise to mosaic structures of thegenes. This is superimposed on the accumulation of pointmutations acquired as the result of interactions of thepathogens with the immunological system of the humanhost, which make antibodies against the membrane surfaceproteins of the bacteria. A similar picture emerged onanalysis of group A streptococci isolated from a cluster ofcases of serious infections over a 3-month period inTayside, Scotland in 1993 (21). DNA sequence analysis ofvirulence factors suggests that they may have been recentlyacquired by horizontal gene transfer.

In the past, only Vibrio cholerae strains of the type O1are known to have caused epidemcs, while non-O1 strainsare associated with sporadic cases. However, the recent

Gene technology and gene ecology of infectious diseases 47

epidemic in Asia was caused by a non-O1 strain,Vibrio cholerae O139. It turns out that the new strainis identical to an earlier pandemic strain, O1 EL Tor,except for proteins involved in the capsule and the O-antigen synthesis (which is used in typing the strains).This was due to the acquisition of DNA inserted into,and replacing part of the O antigen gene cluster. Thissuggests that O139 arose by horizontal gene transferfrom a non-O1 strain into a type O1 strain (15–17).

Pathogenic mycoplasmas possess proteins called ad-hesins and related accessory proteins which are re-quired for adhering to cells and subsequent diseasedevelopment. Mycoplasma genitalium, implicated in ure-thritis, pneumonia, arthritis and AIDS progression, wasfound to encode one adhesin and 2 adhesin-relatedproteins that shared substantial sequence similarities, aswell as genome organization, with those of anotherspecies, M. penumoniae (114). These were attributed tohorizontal gene transfer events between the two spe-cies.

Most disturbing of all is the recent discovery that atleast 10 unrelated bacterial pathogens, causing diseasesfrom bubonic plague to tree blight, share an entire setof genes for invading host cells, which have spread byhorizontal gene transfer, as they reside as a block, of-ten in the bacterial chromosome, forming ‘pathogenic-ity islands’ (see Section 9.1) (23). More than 20 genesare involved in a system that secretes damagingproteins directly into host cells. They were first discov-ered in Yersinia, a genus of several species that causehuman disease including bubonic plague and intestinalinfections. At least some of the proteins are typical ofeukaryotic cells, and are targeted at eukaryotic bio-chemistry (115), leading to the speculation that thegenes were originally acquired from eukaryotic cells.We shall look at some of these systems in more detaillater.

Like antibiotic resistance genes, many virulence genesreside on transposons or plasmids. For example, theST-enterotoxin genes of E. coli are part of a transpo-son; while important virulence factors of Gram-nega-tive pathogens (Shigella flexneri, Salmonella spp.,Yersinia spp.) as well as Gram-positive pathogens(Clostridium tetani) are encoded in plasmids. Other fac-tors, including the Shiga-like toxins of E. coli, thecholera toxin of Vibrio cholerae, diphtheria toxin ofCorynebacterium diphtheriae, neurotoxins of Clostridiumbolulinum and cytotoxin of Pseudomonas aeruginosa, areencoded by bacteriophage (116, 117). All these toxinsand virulence factors obviously have the capacity tospread horizontally and to create new pathogens by re-combination.

Virulence factors are often also located on the chro-mosome, where they form large blocks, which came tobe called pathogenicity islands.

9.1 Pathogenicity islands

The term ‘pathogenicity island’ (Pai ) refers to large chro-mosomal regions in pathogenic bacteria that encode viru-lence genes acquired by bacteria from unrelated organisms.Hacker et al. (118) reviewed the subject recently. Theydefine a Pai according to the following criteria:

1. Carriage of (often many) virulence genes.2. Presence in pathogenic strains, and absence or sporadic

distribution in less-pathogenic strains of one species ora related species.

3. Different G+C content in comparison to DNA ofhost bacteria, indicative of a foreign source.

4. Occupation of large chromosomal regions (often morethan 30 kb).

5. Represents compact, distinct genetic units, oftenflanked by direct repeats (like many viruses andtransposons).

6. Association with tRNA genes and/or insertion se-quence (IS), a special class of mobile genetic elements,at their boundaries.

7. Presence of (often cryptic) mobility genes (IS elements,integrases, transposases, origins of plasmid replication).

8. Instability.

A bacterial pathogen may have one or more pathogenic-ity islands. For example, uropathogenic (causing infectionof the urinary tract) E. coli strain 536 has two Pais : Pai Iencodes only a-haemolysin, responsible for haemolysis,while Pai II carries the a-haemolysin gene, (hly), as well asthe P-related fimbriae gene (prf) involved in attachment ofthe bacterium to the host cell. Pai I is flanked by the selCgene encoding the selenocysteine-specific tRNA and twodirect 16 bp repeats, while Pai II is associated with leuX,which encodes a leucyl tRNA, and two direct 18 bprepeats. The direct repeats in both cases are attachmentsites for bacteriophages (fR73 in Pai I, and P4 in Pai II).Cryptic integrase genes of these phages are located nearthe respective tRNA genes. Thus, the Pais probably origi-nate from integrated prophages which have picked up arange of virulence genes to turn them into pathogenicityislands. In addition, sequences homologous to the originsof replication of plasmids, parts of IS elements, and ‘re-combination hot spots’ (sites which are most frequentlyinvolved in recombination) are also present. All theseindicate that Pais are highly mosaic, and many recombina-tion events have contributed to their evolution. Have someof these recombinations been inadvertently created bygenetic engineering?

Phages, plasmids and pathogenicity islands are involvedin rapid evolutionary changes of bacteria by ‘quantumleaps’. Pais often exhibit features of both plasmids andphages, reflecting the formation of co-integrates of bothtypes of extra-chromosomal elements. The acquisition ofnew genes following horizontal gene transfer results in the

M.-W. Ho et al.48

generation of new variants of pathogens. Drug-resistantstrains of Mycobacterium tuberculosis and Streptococcuspenumoniae, the new V. cholerae serotype 0139, E. coli0157, Haemophilus influenzae biotype aegypticus, are ex-amples of such new types of pathogens generated byhorizontal gene transfer.

Enteropathogenic bacterial strains (infecting the gas-trointestinal tract), including E. coli 0157:H7, have theability to remodel the actin cytoskeleton of infected cells,producing so-called ‘attaching and effacing’ (AE) lesions inthe host cell. A 35.4 kb pathogenicity island in enteropathicE. coli, called the locus of enterocyte effacement, (LEE), hasbeen found to contain all the known genes necessary for AElesions (119). Several features of the LEE suggest that it hasbeen acquired recently by the enteropathogenic E. coli andother AE bacteria. These include, (i) a G+C content of38.4%, compared with the 51% of the bacterial chromosomeas a whole. (ii) the strict conservation of LEE base se-quences within AE pathogens of 3 genera of bacteria,including the distantly related Hafnia al6ei, and (iii) inser-tion within selC tRNA gene, the same site of insertion forpaiI in uropathogenic E. coli and the bacteriophage fR73.The LEE is one among a whole class of genetic systems(type III systems) that are extremely sophisticated in ex-ploiting host cell functions (see below).

9.2 Exploitation of host cell functions by bacterialpathogens

Recent studies reveal the common strategies used by awide range of pathogenic bacteria as well as the uniquetactics adopted by individual species in establishing infec-tion. Many host cell functions are exploited by the patho-gens, including signal transduction pathways andrearrangements of the cytoskeleton (115).

Adhesion to host cells is the first stage in disease devel-opment. Bacterial pathogens produce a range of proteins,adhesins, which mediate adhesion to host cells. Theseadhesins interact with surface membrane components ofmammalian cell, many of which serve as receptors for thebacterial adhesins. The host cell is often an active partici-pant in adhesion by synthesizing the receptors for adhe-sion. For example, enteropathogenic E. coli adheres tointestinal epithelial cell surfaces by destroying host mi-crovilli and rearranging the cytoskeleton to form apedestal on the host cell surface on which the bacteriumresides (the AE lesions). To achieve this, the bacteriumsecretes at least two proteins, EspA and EspB, whichactivate host cells by inducing a series of events involved insignal transduction: calcium flux, inositol phosphate pro-duction, phosphorylation of the amino-acid tyrosine in amembrane protein, and finally, rearrangement of the cy-toskeleton. The secretion of the proteins that activateepithelial cells is mediated by a specialized secretion systemcalled the type III secretion system, version of which arebeing identified in an ever-increasing number of human,

animal and plant pathogens. They allow the bacteria toattach to cells, to enter the cells, or to prevent themselvesfrom being swallowed by phagocytic cells.

Type III secretion systems comprise at least 20 geneproducts which are essential for the virulence of thesepathogens; the genes encode both secreted effector proteinsand the machinery necessary for secretion and transloca-tion into target cells. Enhanced secretion of the virulencefactors often occurs after contact with host cell surfaces.Although the secreted virulence factors differ amongpathogens which cause different diseases, the secretionmachinery is often interchangeable. The same type IIIsystem is involved in invading cells by activating hostsignalling pathways, leading to appropriate changes in thecytoskeleton which internalize the pathogens. One type ofinternalization mechanism is shared by Yersinia and Liste-ria, while another is shared by Salmonella and Shigella. InSalmonella, the type III secretion system responsible forinvading host cells forms a pathogenicity island on thechromosome, while in Shigella, the system is found in alarge virulence plasmid.

In addition to invasion systems for entering epithelialcells, Yersinia species have evolved mechanisms to a6oiduptake by phagocytic cells. The anti-phagocytosis strategyrelies on the expression of Yop proteins, which also in-volves type III secretion systems. One protein, Yop Eparalyzes the cellular cytoskeleton, while another, Yop His a broad-spectrum tyrosine kinase that interrupts phos-photyrosine signaling associated with phagocytosis. Yetanother protein, Yop O, has homology to mammalianenzymes, serine and threonine kinases involved in signaltransduction, and like Yop H, is targetted to the innersurface of the host-cell plasma membrane.

Among Gram-positive bacteria, Listeria monocytogenesenters mammalian intestinal cells by means of a surfaceprotein, internalin. Internalin contains multiple copies of a22-amino acid leucine-rich tandem repeat (LRR), a featureof several eukaryotic proteins that are generally involvedin protein–protein interactions. Another surface protein,InlB, belonging to the internalin multigene family anddisplaying similar LRRs, mediates entry into cultured hep-atocytes, Hela cells and Chinese Hamster Ovarian cells.

Type III secretion systems have also been identified inthe plant pathogens Ralstonia, Pseudomonas, Xanthomonasand Erwinia, and are involved in their ability to causediseases (120).

10. HAS HORIZONTAL GENE TRANSFERINCREASED RECENTLY?

The evidence is now overwhelming that horizontal genetransfer has been responsible for both the rapid spread ofantibiotic resistances and for the emergence of virulentstrains of pathogens in recent years. Has horizontal genetransfer increased in scope and frequency since geneticengineering biotechnology began?

Gene technology and gene ecology of infectious diseases 49

In the Introduction, we have alluded to the apparentlyaccelerated rates in the evolution of drug and antibioticresistance and of new pathogens since the 1980s. Further-more, sequence identity or near-identity in both virulenceand antibiotic genes from unrelated species indicate that ahigh proportion of the horizontal gene transfers may bevery recent.

A study carried out in 1983 on strains of bacteriaisolated between 1917 to 1954, at least ten years into theantibiotics era, but well before genetic engineering gotunderway, showed that none of the strains carried anyantibiotic resistance (121). By contrast, it takes one to twoyears in the post-genetic engineering era for resistance todevelop and spread, rendering new antibiotics useless in 4years (11, 25). However, 24% of the strains from thepre-genetic engineering era encode genetic information forthe transfer of DNA from one bacterium to another; andfrom at least 19% of the strains, conjugative plasmidscarrying no antibiotic resistance were transferred to thelaboratory strain E. coli K12 (121). Thus, while conjuga-tive plasmids and other mechanisms for horizontal genetransfer may predate both the antibiotics and geneticengineering eras, their prevalence and the range of drugsto which they confer resistance have greatly increased,especially since genetic engineering got underway.

It would also be of interest to screen these strains fromthe pre-genetic engineering era for pathogenicity islands,which will contribute to our understanding of how patho-genicity islands have evolved since.

10.1 Barriers to horizontal gene transfer

The frequency of horizontal gene transfer may therefore bedetermined, not so much by the actual mechanisms thatexist for mediating transfer as on the existence of barriersthat act against the process (77, 122, 123). In bacteria, amajor mechanism for limiting horizontal gene transfer isrestriction—the destruction, or inactivation of foreignDNA. There are 11 restriction/modification systems inenteric bacteria which fall into three families, hsdR, hsdMand hsdS. The alleles (different variants) of the hsdRfamily are responsible for destroying foreign DNA, whilethose of hsdM and hsdS families are responsible for methy-lating, thereby protecting DNA molecules that escaperestriction and facilitating their recombination into thegenome. A second barrier is the mismatch DNA repairsystem encoded by the mutL, mutS and mutH genes of E.coli and S. typhimurium, which severely inhibits recombi-nation of nonhomologous DNA. However, another ge-netic system, the error-prone SOS system, activated understress, stimulates such recombination. In addition, as men-tioned in Sections 6.1 and 7.3, there are origins of replica-tion and origins of transfer specific to DNA from differentspecies, so that even if DNA is transferred, they cannot bereplicated and maintained, so that the chances of integra-tion into the chromosome would also have been much

reduced, unless, of course, the DNA has been manipulatedto combine origins of replication from different species.

Another