BACTERIOLOGICAL REVIEWS, Mar., 1966Copyright © 1966 American Society for Microbiology

Host-Dependent MicrobesJOHN H. HANKS

Johns Hopkins University-Leonard Wood Memorial Leprosy Research Laboratory, Department ofPathobiology, School ofHygiene and Public Health, Baltimore, Maryland

INTRODUCTION ............................................................... 114The Common Feast.......................................................... 115

MICROBE-DEPENDENT MICROBES.................................................. 115Chelate-Dependent Microbes.................................................. 116

Circumvention ofmycobactin requirements ..... ............................... 117Replacement ofchelate requirements......................................... 118

Other Microbial Dependencies................................................. 118

UTILIZATION OF MICROBIAL GROWTH FACTORS DURING INTRACELLULAR GROWTH ...... 118Plants and Animals as Conveyors of Microbial Growth Factors..................... 119

IMPRINTS OF GENETIC CHARACTER AND OF HOST ADAPTATION IN CHELATE-REQUIRINGPATHOGENS............................................................... 119

Major Peculiarities ......................................................... 120

Inhibitions at neutral pH and by complex nutrients............................. 120

Preference for simplified media and low pH................................... 121

Response to peptones...................................................... 121

Elimination oflag......................................................... 122

Discussion................................................................. 123

Adaptation to phagocytic vacuoles ...... ..................................... 123Selective pressures ........................................................ 123Imprints ofhost adaptation ................................................ 123

Evolutionary Implications of Genetic Markers................................... 124

ORIGINS OF HOST-DEPENDENT MICROBES.......................................... 124

Regressions................................................................ 124

Antiquity ofDisease......................................................... 125

Possibilities................................................................ 125

PROBLEMS SHARED BY HOST-DEPENDENT MICROBES................................. 125

Introspections on Circumventing Certain Impediments to In Vitro Study............... 127

Physicochemical environments............................................... 127

Protection oflabile systems ................................................ 128Impounding of metabolites................................................... 128Lend-lease of critical cofactors.............................................. 129

Imponderables............................................................ 129

SUMMARY AND APPRAISAL...................................................... 129

BEDTIME STORY............................................................... 131

LITERATURE CITED............................................................ 132

This diatribe against the injunction of Polonius to Laertes: "Neither a borrower nor a lender be"[Hamlet: Act I, scene iii], is not intended to affront the memory of an inspired author. Shakespeareis exonerated. Possibly he did not consider how little Polonius knew about differences between de-pendencies in men and in microbes.

INTRODUCTION

The origins, character, and possible cultivabil-ity of the host-dependent (i.e., noncultivated)microbes have been challenging topics in theminds of many microbiologists. Although specu-lations regarding origins cannot be substantiatedfrom existing data, they are fascinating. If ques-tions can be raised within an appropriate frame-work, there is always the possibility that an imag-inative experiment can recapitulate one or more

of the evolutionary steps. A second possibility isthat discussions concerning the nature of prob-

lems in host-dependent microbes may suggestprinciples which are useful to their cultivation.My own interest in these apparent impracti-

calities has been enhanced by the fact that thislaboratory adopted the mycobactin-requiringMycobacterium paratuberculosis as a model fromwhich to extract lessons that might be helpfultoward the cultivation of M. leprae. The resultsobtained with this choice of experimental modelhave produced a series of intellectual shocks.These include the findings that this supposedlyfastidious pathogen is but one of a group of free-

114

Vol. 30, No. 1Printed in U.S.A.

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

living microbes which require microbially synthe-sized chelators of heavy metals; that the specificgrowth factor, mycobactin, and other catalysts orsubstrates are used by bacilli growing withinmammalian cells as readily as by those growing ina synthetic medium; that the presumed originalphysiological requirements of this species haveremained so fixed that it has not adapted to majorfeatures of host environment, in spite of pro-longed passage in cows and sheep; that its majorgenetic markers are counterparts of those in well-known physiological groups of free-living mi-crobes; andthat the major imprint of host adapta-tion is the sluggishness in saturating in vitroenvironments with cofactors and metaboliteswhich it synthesizes too slowly from the simplestcompounds.

These observations raise several novel ques-tions. One of these questions is whether factor-requiring (i.e., microbe-dependent) microbes,after adaptation to plants or animals, couldmasquerade as host-dependent, while in facthaving dependencies upon the original donors ofstrictly microbial growth factors and uponphysicochemical features and the more univer-sally used nutriments of the host in which themicrobe has been discovered. It will be seen thatseclusion within hosts or host cells does not iso-late host-dependent microbes from microbiallysynthesized factors. Another question is whetherthe speciation of host-dependent microbes mayhave been highly developed prior to their adapta-tion to plants or animals and whether, given abasis for intelligent search, counterparts of theirprogenitors might be found in nature even today.

Irrespective of theories, the microbiologist whobecomes distrustful of host-oriented conceptswill be struck by four straightforward proposi-tions. The notable specializations which distin-guish microbes from other forms of life concernmembranes which synthesize cell walls. Thesymptoms common to a broad series of host-dependent microbes are delayed or deficient cellwall formation. Plants and animals are not en-gaged in this line of business. Whether for thesupplementation of membrane systems or wall-forming systems, competent microbes, not hosts,constitute the logical source of specialized cofac-tors which may be needed by host-dependentmicrobes, both in vivo and in vitro.

The Common FeastIf one wishes to comprehend the many-faceted

attributes of host-dependent microbes, it is legiti-mate to parasitize other investigators, particularlyauthor-dependent authors who have preparedthoughtful reviews. They have repaid at least a

portion of our universally intellectual debts byciting highlights and providing references to largeareas of pertinent work.

Trager (87) has presented illuminating views ofthe symbiotic, parasitic, and pathogenic depend-encies in an exceptionally broad range of bio-logical forms, including the host-dependent mi-crobes which are indispensable to insects thatdepend upon microbes. For a general survey ofhost dependency, or challenging model systemsand problems, Trager is an invaluable guide.Moulder (60, 61) has prepared two penetratinganalyses of the rickettsiae and the psittacine bac-teria. The qualities in these predators are set forthby considering their energetics, their enzymaticdeficiencies and "leakiness" as related to adapta-tion to animal hosts, and the character of theirgrowth and infectiousness as related to cell wallformation. One of the notable by-products wasthe rescue of these organisms from the humilia-tion of being mistaken for viruses.

Evidence which suggests that host-dependentstates may have existed in many microbes priorto the development of plants and animals will befound in articles cited in Barnett's (3) review ofmycoparasitism. Before venturing into this field,the medical microbiologist must resolve not to bedismayed or befuddled by names and terminol-ogy. Lest the diet be too rich for many, I haveprepared a digest of these matters (see BedtimeStory).

In the present review, several topics will besurveyed to provide backgrounds for eventualarguments, speculations or conclusions. In gen-eral, the procedure will be to develop the rela-tionships between a pathogen and its physiologi-cal counterparts in soil and to determine whysuch a microbe has appeared to be one of themost fastidious of the microorganisms yet culti-vated. Certain of the properties and problems inhost-dependent microbes can then be reviewedwithin new frameworks.

MICROBE-DEPENDENT MICROBESOne step in attaining perspective regarding

host-dependent microbes is to recognize that de-pendent states exist in free-living organisms. Thissection is concerned with classical examples ofdependency upon strictly microbial growth fac-tors which do not have vitaminlike activity forother forms of life. Requirements for L-lysine orwall-forming precursors that are used or suppliedby plant and animal hosts would not qualifyunder this definition. Experimentally inducedmicrobe-dependent states are numerous. Theyinclude the streptomycin dependencies in a seriesof species (see 22), the diaminopimelic acid de-

115VIOL. 30X 1966

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REV.

ficiency in Escherichia coli strain W 173-25 (4),and requirements for tuberculostearic acid intubercle bacilli (49). Curiously enough, the firstdiscovery of a biological growth factor also identi-fied a microbe-dependent microbe (88).

Chelate-Dependent MicrobesIt is interesting that a natural dependency on

microbially synthesized chelators of heavy metalsis as critical to several species of free-living bac-teria and fungi as to a presumed strongly host-adapted mycobacterial pathogen. The implica-tions of this relationship have not beenappreciated for several reasons. First, the growthfactor from competent mycobacteria, discoveredin 1913 by Twort and Ingram (88) and nowknown as mycobactin (28), was long thought tobe significant only for the fastidious and little-studied M. paratuberculosis. Second, the interestin functionally related chelators has centered onstudies on antibiotics and soil microbiology. Theresult has been that all such chelators have beenconsidered "special" rather than "universal"microbial growth factors. The known examples ofchelate-dependent microbes have been discoveredlargely by accident. As will be seen, the deliberateinclusion of microbial chelators during surveys offree-living and host-associated microbes shoulddisclose a variety of additional forms exhibitingthis single type of dependency.

In recent years it has been shown that naturallyoccurring microbial chelators which containthree N-substituted hydroxamic acid groupings(R1-CO-NOH-R2) (52) are essential to the growthof microbes. Evidence of the widespread impor-tance of such chelators came simultaneously fromthree laboratories. Lochhead, Burton, and Thex-ton (57) demonstrated that the iron-chelating"terregens factor" is produced by growth-compe-tent species of Arthrobacter and is required byA. flavescens and A. terregens. Hesseltine et al.(44) showed that coprogen is produced by aspecies of Penicillium and is necessary for thegrowth of the dung-dependent fungus, Piloboluskleinei (45). Neilands (63) isolated ferrichromefrom the smut fungus, Ustilago sphaerogena andfound that it supported the growth of A. ter-regens and P. kleinei (64).New concepts regarding the importance of

microbially synthesized iron chelators and theirrelationships to antibiotics originated from theFederal Institute of Technology in Zurich and theCiba Research Laboratories in Basle. As a resultof these investigations (6, 7), the siderochromeswere divided into two major groups: (i) thesideramines, which function as growth factors,and (ii) the sideromycins, which were considered

to have antibiotic activity because of being antag-onists of the sideramines. A third minor groupwas proposed to include chelators, such as ferri-chrome A, which do not function biologically.Thirteen sideramines have been isolated fromStreptomyces species and from fungi. The nineferrioxamines from Streptomyces species areclosely related. The four from fungi differ andhave been called ferrichrysin, ferricrocin, ferrirho-din, and ferrirubin (50, 51).Although structurally the mycobactin from M.

phlei is a dihydroxamate, it functions as a hexa-dentate chelator because of a phenolic OH groupand the N atom of the oxazoline ring (81). Itserves as a growth factor for all the sideraminerequirers which have been tested. Because of itsphysical and functional equivalence to sider-amines, both Snow (82) and Morrison, Antoine,and Dewbrey (59) have considered it to be asideramine.The heterotrophic character of sideramine re-

quirements in free-living dependent species isdemonstrated by the readiness with which sider-amines, including mycobactin, can be exchangedbetween diverse species (17, 18, 24, 43, 44, 45, 64,70, 98). The one example of apparent specificityis the failure of the mycobacterial pathogens torespond to the water-soluble sideramines.The survival of chelate-requiring microbes in

nature apparently is due to the universal produc-tion of sideramines by growth-competent species,whether they be aerobic or facultative. Hesseltineet al. (45) recovered "coprogen type" stimulatorsfrom 10 of 32 species of yeasts, fungi, strepto-mycetes, and bacteria. Burnham and Neilands(17) demonstrated bound hydroxylamines in thecells of 20 of 25 representative species and stimu-lation of A. flavescens JG-9 by extracts from 37 of50 species. Failures to demonstrate the produc-tion of active chelators by 100% of the growth-competent species probably was due to adequatelevels of iron in many of the media employed forthe cultivation of such a broad spectrum of mi-crobes. Zahner et al. (98) found that limitationof iron accentuates the production of large quanti-ties of desferrisideramines by all the streptomycesand fungi studied and concluded that sideramine-type compounds are produced by all aerobicorganisms. Antoine, Morrison, and Hanks (1)observed that iron limitation was fundamental tothe production of mycobactin by two synthesiz-ing species. Burnham (personal communication)has reached similar conclusions.The role of the sideramines in the transport of

iron and their insertion into respiratory catalystshas been under active study by Burnham (15, 16).The alterations of membrane activities and of

116 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

respiratory pathways during iron restriction andmycobactin production are being investigated byMorrison and associates in this laboratory.

In view of the fact that nearly 40 years wererequired to establish relationships between agrowth factor for a fastidious pathogen and ageneral mode of microbial dependency, it is likelythat yet additional examples of chelate requirersawait discovery.

Within the genus Bacillus an example has re-cently come to light. This involves the response ofB. megaterium to schizokinen, which has beenrecently characterized as a sideramine (66).Schizokinen promotes the growth of the Arthro-bacter JG-9 strain (Lankford, personal communi-cation). Similar effects on A. terregens and A.flavescens have been confirmed by Morrison inthis laboratory.

Circumvention of mycobactin requirements. Inspite of the many billions of sideramine-requiringcells from four genera of microbes which havebeen planted on a variety of media in the absenceof sideramines, there are no reports of the emer-gence of competent mutants.

Insight into the cirvumvention of mycobactinrequirements (58) arose through Morrison's de-sire to examine the differences between strainMj68 of M. paratuberculosis, which strictly re-

0.500

0.400

zCIa

-j4du

0.0

0L300

0.200

0.100

quires mycobactin, and two mycobactin-"inde-pendent" strains (Teps and III-V) which hadbeen grown for 33 years on the Watson-Reid(WR) synthetic medium. The findings regardingthe development and the causes of "independ-ence" were quite unexpected. Although pellicles,i.e., organelles, of Mj68 required 1 ,ug of myco-bactin per ml to make a successful transition fromcomplex media to the WR medium, subsequenttransfers demonstrated that they had become"independent" during the first transfer.

Further investigation of the three strainsshowed that the "independent" growth does notalter the response to submicrogram concentra-tions of mycobactin. Even in low iron mediagrowth is not associated with the production ofdetectable mycobactin. The initiation of growthdepends in part upon low pH and in part uponcomplexes which are formed during the auto-claving of the glucose-containing medium at pH5.5. Autoclaving of key ingredients of the mediumresults in the formation of compound(s) whichshow characteristic absorption maxima at 230 andat 290 mA (Fig. 1).At this point, the evolution of knowledge re-

garding the "hosts" for a fastidious pathogen haddeveloped through three stages, the successiverequirements being: cows or sheep, mycobactin,

200 300 400 S00 600

WAVE-LENGTH (map)FIG. 1. Absorption spectrum produced by 1% glucose in 0.015 M phosphate after autoclaving at pH 5.5 for 15

min at 121 C (Morrison, unpublished data).

117VOL. 30, 1966

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BAc-TERIOL. REV.

and, if pellicles were to be used, the autoclavingof an appropriate medium.

Curiously enough, pellicles grown in the ab-sence of mycobactin have become our standardsources of physiologically active, but mycobactin-requiring, cells. Although the transfer of strainMj68 from complex, mycobactin-containing me-dia to the WR medium may involve the selectionof clones adapted to the new physical environ-ment, it will be shown later that declumped, di-luted suspensions of all three "independent"strains exhibit a strict requirement for mycobac-tin and that the minor adaptations have notselected cells with adequate capacity to saturateenvironments with cofactors and metabolites.The circumventions described illustrate two

basic points regarding the character of M. para-tuberculosis. In the first place, aside from themycobactin requirement, this "fastidious" patho-gen has retained the capacity for a chemosyn-thetic existence. Second, the original genetic de-ficiencies which are related to mycobactin havepersisted in strains Teps and III-V during some33 years of growth in the absence of mycobactin.This stability of requirement is typical of sider-amine-requiring microbes from soil. It seemspointless to argue that intracellular environmentsselected this property while having no similarinfluence on bovine-type tubercle bacilli.

Replacement of chelate requirements. In thecase of the heterotrophic sideramine requirers,hemin has been found to produce growth stimu-lations, provided it is used in narrow ranges ofconcentration. Burton, Sowden, and Lochhead(18) reported the growth of A. terregens in thepresence of heme at 0.1 jig/ml, but not at con-centrations one log higher or lower. Burnhamand Neilands (17) reported a 50% maximal re-sponse by A. flavescens JG-9 to 0.08 jig/ml.Demain and Hendlin (24) reported that Micro-bacterium lacticum responds to hemin below con-centrations of 50 pig/ml.

Subsequently, Morrison conceived that growthstimulation by sideramines might not dependupon special structures but upon the formationof relatively nonpolar metal chelates. Experi-ments with A. terregens and A. flavescens demon-strated that the naturally occurring iron chelatorscan be replaced by synthetic bidentate chelators,such as 8-hydroxyquinoline, salicylaldehyde, andacetylacetone (59). At appropriate concentrationsthe chemical chelators produce typical dose-re-sponse curves. They are toxic in higher concen-trations. With respect to the narrow range ofuseful concentrations they are reminiscent ofhemin. In the choice of biologically active chemi-cal chelators, two properties appear essential: the

compound must chelate ferric ions and the result-ing chelate must be lipophilic.

In respect to dependent microbes in general, itnow is evident that presumed specific physiologi-cal requirements can be met in several ways: bymicrobial growth factors from related or unre-lated species, by empirical discovery of media orconditions which tend to circumvent the require-ment, or by selecting suitable compounds from achemical catalogue.

Other Microbial DependenciesThe frequent, often quantitative, occurrence of

unanalyzed dependencies is illustrated by thefollowing. The catalogue of the American TypeCulture Collection (ATCC) notes that the growthrates of B. subtilis 12695 and E. coli 12696 areincreased 10 times by a growth factor from twostrains of Aspergillus. A personal communicationfrom Mrs. Daley of the ATCC states that the B.subtilis strain was isolated from the intestines of12-day-old chicks which had been fed on mashenriched with the Aspergillus growth factor.Anaerobes seem not to have been tested for the

production or requirement of sideramines. Never-theless, an example of critical requirements forunknown microbial factors in anaerobes is foundin the report of Hardy, Lee, and Nell (42) on theisolation of new types of oral spirochetes. Theseappeared as satellite colonies near contaminantson a plating medium which had been developedas optimal for the known types of spirochetes.The new types have since been propagaged bymeans of unknown factors in culture filtratesfrom a microaerophilic diphtheroid. It will benoted that these oral spirochetes could not haveacquired their apparent host dependencies be-cause of prolonged enclosure within host tissuesor cells. Although they depend upon the humanhost for warmth, moisture, and certain nutrients,they evidently rely upon adjacent competentmicrobes to meet certain additional requirements.

UTILIZATION OF MICROBIAL GROWTH FACrORSDURING INTRACELLULAR GROWTH

Before inquiring whether host-adapted mi-crobes could rely upon plants or animals to con-vey microbial growth factors to intracellularagents of disease, the critical question of the pos-sible impenetrability of host cell membranes mustbe considered.The intermediate position of M. paratuberculo-

sis and the wood pigeon mycobacteria in thespectrum of mycobacterial species and the com-bination of microbe dependency with host adapta-tion enabled Wheeler and Hanks (93) to explore

118 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

two problems. First, it seemed desirable to explainthe fact that the intracellular growth of myco-bacteria had been proportional to their capacitiesfor growth in vitro. The growth-competent patho-gens and tubercle bacilli were known to growreadily in the three major cell types under avariety of conditions. M. paratuberculosis andrelated types had not been studied. M. leprae-murium grew very slowly (19, 69, 90). M. lepraehad not yet been propagated in vitro. Second, inthe case of the noncultivated mycobacteria, thereis the question whether the host cells or extrane-ous compounds are the likely sources of specialfactors which might be useful during intracellulargrowth.The primary issue has been resolved (93).

When securely phagocyted within cells whichhave been washed free from external bacteriaand then surrounded with streptomycin to pre-vent extracellular growth, M. paratuberculosisand the wood pigeon mycobacteria do not de-pend solely upon the metabolites or constituentsof the host cells. External supplies of factorswhich promote the growth of declumped bacilliin synthetic media stimulate the intracellulargrowth of these microbes just as effectively asthough they were in test tubes. The factors stud-ied were mycobactin, iron, C02, and glycerol.Mycobactin could not be shown to modify theappearance or overall metabolism of the hostcells. The dramatic influence of external suppliesof mycobactin on intracellular growth representsa new phenomenon, the intraphagosomal avail-ability of strictly microbial growth factors. El-berg's (27) review contains equally clear evidencethat externally supplied macromolecules, such aslysozyme and antibody globulins, exert theirtypical effects on securely phagocyted brucellae.Finally, Chang (18) and Garbutt (30) have re-ported the slow in vitro growth of M. leprae intissue cells. Both groups emphasize that multi-plication does not depend upon items which arerequired to maintain healthy cells, but upon un-known factors in serum or other components ofthe medium.Taken together, this series of observations

revolutionizes earlier views regarding the role oftissue cells as hosts for intracellular agents.Whether one is interested in nutrition or inhibi-tion, it is evident that the situation for intra-phagosomal microbes is not fundamentally differ-ent from that of the extracellular types, that thedifferent biochemical environments in varioustissues can be important determinants of intra-cellular growth, and that genetically controlledlevels of metabolites or inhibitors in different in-

dividuals can have a marked influence on sus-ceptibility or resistance.

Plants and Animals as Conveyors of MicrobialGrowth Factors

There is no novelty in the fact that the veryexistence of plants and animals depends uponchemical transactions and vitamin syntheses bymicrobes. Since the uptake and distribution ofstrictly microbial growth factors within plantsand animals has not been examined experiment-ally, the evidence in favor of this possibility mustbe circumstantial. The considerations fall intotwo categories: the diversity of compounds nor-mally taken in through rootlets and the gut, andthe question whether compounds of the sameorder of size but unrelated to host economy couldbe excluded.The sap which is piped upward in plants con-

tains minerals, nitrates, vitamins, etc., and willnourish other plants. The sap flowing downwardfails to do so. To be brief, White (94) has con-cluded that the sap which flows upward in plantsis "in most important regards a good soil solu-tion." However, for want of more precise inform-ation on what can be excluded, we shall leaveopen the question whether plants could conveystrictly microbial factors to the noncultivatedmicrobes which they support.

In the case of animals the probabilities are high.To secure an adequate intake of vitamins, pro-teins, lipids, etc., animals are not content withconsuming all manner of natural products. Theyhave wrapped their hollow bodies around a di-gestive tube which houses an astonishing varietyof microbes that engage in chemical transactionsand syntheses. The ruminants have elaborated onthis arrangement by installing way-stations forhighly diversified fermentations. While absorbingthe long list of items essential to host economy,animals are known to take up dyes, drugs, anti-biotics, etc., which are unrelated to host economy.As shown by many studies on atopic sensitization(26, 56), even adults absorb low proportions ofthe antigenic proteins which they ingest.

In view of the total evidence, it seems likelythat animals could provide small but continuoussupplies of metabolites, membrane supplements,or cell wall precursors which are useful solely toapparently host-dependent microbes.

IMPRINTS OF GENETc CHARACTER AND OF HOSTADAPrATION IN CHELATE-REQUIRING

PATHOGENSA. terregens has been used as the "E. colil for

studies on sideramine requirements. Similarly, the

VOL. 30, 1966 119

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BAC=hRIoL. REV.

once noncultivable M. paratuberculosis and thewood pigeon mycobacteria have been employedas models for examining the ways in which hostadaptation may have increased the fastidiousnessof the incompetent mycobacteria. The choice ofthe chelate requirers as models for the host-de-pendent species rests upon the following con-siderations. (i) For 17 years the cultivation of M.paratuberculosis from lesions in the intestinalwalls of cows and sheep had been impossible.There was a lag of 5 years in learning that themycobacteria in wild wood pigeons can be culti-vated by incorporating killed mycobacteria indiagnostic media (see 93). Were it not for theoriginal inspiration of Twort and Ingram (88),both types might remain uncultivated even today.(ii) As in the case of other chelate requirers, themajor deficiency has not been circumvented byrare mutations to competency or by extracts,complexes, and compounds from plants and ani-mals (28, 80, 88). (iii) If deprived of mycobactin,the resultant noncultivable ceLs, at least in somerespects, should be counterparts of those fromnoncultivated species. (iv) They are the mostfastidious mycobacteria which can be studied bylaboratory disciplines such as nutrition, metab-olism, cell culture, etc.

Efforts to evaluate the peculiarities in chelate-requiring mycobacteria thus far have been re-stricted to the physiology of growth. Wheelerand Hanks (93) used suspensions of declumpedcells from three strains of M. paratuberculosis andtwo of wood pigeon bacilli to study parallelismsbetween growth in vitro and within host cells.The sources of these strains, a description of M.paratuberculosis, and the formula of the modifiedWR medium have been given by Morrison (58).The wood pigeon mycobacteria have been char-acterized by Wheeler and Hanks (93). The prepa-ration of bacterial suspensions for inocula,methods for measuring growth, the sources ofmycobactin, and general specifications regardingexperimental conditions have been described(93). The growth requirements defined to dateare those for cell suspensions diluted only fivetimes below the threshold of visible growth, notthose for diagnostic inocula.

Major PeculiaritiesIn general, the pertinent findings in the pres-

ence of mycobactin have been as follows. Growthat neutral pH, whenever obtained, was poor. Atany pH, growth is readily inhibited by complexnutriments from plants or animals. Growth oc-curs most readily in simplified media at pH 5.5 to4.5. The NH4 ion is an optimal source of nitro-gen. Dilute peptones stimulate the onset of

growth without increasing cell yields. The lagprior to the onset of growth has been eliminatedby mycobacterial metabolites, but not by the hostcomponents tested. These six points will be con-sidered in some detail, because they become thebasis for a series of arguments, speculations, andconclusions.

Inhibitions at neutralpH and by complex nutri-ents. After the classical fashion of the medicalmicrobiologist, both Reich and Hanks exploredtheir assumptions that the growth of M. para-tuberculosis Mj68 could be improved by combin-ing mycobactin with complex supplements orhost components at the pH values (6.5 to 7.2)which are conventional for other mycobacteria.The overall results were: more inhibitions thanstimulations, irregular or unsatisfactory growths,and failures to produce cell crops which weresuitable for physiological studies.Hanks found that strain Mj68 differed from

tubercle bacilli in fundamental respects. Even inthe presence of mycobactin at 1 pg/ml, it failedto grow in a synthetic medium containing redblood cells at 2.5% (v/v), a medium known toinitiate growth from minimal numbers of sapro-phytes and the three types of tubercle bacilli (38).It grew slowly and irregularly in the syntheticmedium plus 10% human ascitic fluid fortifiedwith 0.3% purified serum albumin, a supplementknown to be superior to serum for other myco-bacteria. In the absence of the ascitic fluid, poorgrowth occurred in the synthetic base containing0.3% Trypticase (BBL). Supplementation withunknown factors in 0.15% Phytone (BBL) in-creased growth four times. Complete inhibitionoccurred with 1% Trypticase. This inhibition wasnot relieved by the ascitic fluid or by any of sixseparate additives. It was overcome only duringprolonged incubations in the presence of asciticfluid plus 20 ,ug of mycobactin per ml or by de-creasing the oxygen tensions. Irrespective of nu-trients, slow growth and poor yields of cellsindicated that growth was not proportional to thenutrients provided and suggested that the favor-able combinations were merely counteractinginhibitions.

In more detailed studies, Hanks and Wheeleremployed all five strains of chelate requirers inthe WR medium with various supplements andat pH values of 4.5, 5.0, 5.5, 6.5, and 7.5. Nostrain behaved as though adapted to grow in thepresence of the pH values and the complex nutri-ments which occur in the cytoplasmic compart-ment of host cells. For example, diluted inoculaof strain WP9, which grows the most readily ofall strains and is the least subject to inhibition,grew in the WR medium without added myco-

120 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

bactin. Nevertheless, it has a strict requirementfor mycobactin on Loewenstein-Jensen and othercomplex media. Omission of the L-asparagine andthe diethyl glutamic acid from the 8a medium ofReich and Hanks (70) doubled its rates of growth.

Preference for simplified media and low pH.Aside from the requirements for mycobactin andlow pH, the five strains possess synthetic capaci-ties which, though slow, are the equivalent ofthose in tubercle bacilli. By adding fumarate tothe WR medium, Morrison has shown that theyobtain their total nitrogen from the NH4 ion asreadily as from L-asparagine.The apparently fixed requirements for simple

nutrients and for low pH were of special interestto Wheeler and Hanks during their studies onfactors which promote intracellular growth ofthe chelate-requiring mycobacteria (93). In fur-ther efforts to define the role of pH, it was firstconfirmed that autoclaving the WR medium atpH 6.5 and 7.5 was disadvantageous, even whensuch media were adjusted to pH 5.5 before beinginoculated. The experiments which are most in-structive for present purposes were conducted inlots of the WR medium which had been auto-claved at pH 5.5 to standardize their content of"autoclave factors," and then adjusted to providea series ofpH values: 7.5, 6.5, 5.5, and 5.0 or 4.5.

All five strains initiated growth most success-fully in media adjusted to the pH range of 5.5 to4.5. Results with strain MjlII-V are cited as anexample, since this strain had been grown on theWR medium without mycobactin for 30 years.After 10 weeks (mycobactin, 1 ug/ml), cell yields

1TS0-] pH5.5

wTS

. 10- /pH 6.5E @-I 1/ 0

o 5- TSO_pH5/ /

10 /IT_o

TIME IN WEEKS

Fio. 2. Effect of pH and of 0.2% peptone on thegrowth of strain WP9 in the absence of mycobactin(Wheeler, unpublished data). Solid lines: WR mediumat 50% of the original concentration. Three lots wereautoclaved at pH 5.5, then one was adjusted to pH 6.5and one to pH 7.5. Broken lines: as above, but supple-mented with 0.2% Trypticase-Soy peptone (TS).

at pH 6.5 were 12 and 25% of those at pH 5.5and 5.0. At pH 7.5 no growth appeared. StrainMj68 multiplied at pH 4.5 as readily as at 5.5.Its poor responses at pH 6.5 and 7.5 will be il-lustrated later. The most competent strain, WP9,was the only one which grew with or withoutmycobactin and in all segments of such experi-ments within 10 weeks. Data obtained with thisstrain (Fig. 2) probably illustrate a pattern whichwould emerge with larger inocula or with pelliclesof the other four strains. Since strain WP9 was anexception because of its ability to initiate growthat pH 7.5 (even without mycobactin), its distinctpreference forpH 5.5 over 6.5 and 7.5 emphasizesthe importance of low pH for the growth of suchorganisms.

Response to peptones. The initiation of growthand rates of growth of all five strains were stimu-lated by tolerated levels (0.2%) of a peptone(Trypticase-Soy) from combined animal andplant sources. The peptone, however, did notincrease cell crops. Strain WP9, the only strainwhich attained full growth within 10 weeks, hasbeen used to illustrate the acceleration of growthwithout increase in cell crops (see Fig. 2).Although the more fastidious strain Mj68 be-

gan to grow in full-strength WR medium at pH5.5 within 6 weeks (see Fig. 3), this strain ex-hibited osmotic sensitivity when the medium wasused at 50% of the usual concentration. In thediluted medium, a combination of low pH, myco-bactin, and 0.2% peptone was required. Duringthe development of the growths shown in Fig. 3,traces of growth occurred at pH 5.5 and later atpH 6.5 in the absence of peptone. When it wassuspected that these growths might not succeed,microscopic examinations revealed swollen cellsand rods with central or terminal enlargements.These observations suggest that the dilute pep-tones may provide unknown factors that facili-tate cell wall formation.When inhibitions by conventional concentra-

tions of peptones are coupled with osmotic pro-tection by dilute peptones, two types of explana-tion seem to be required. First, since peptonesfacilitate the synthesis of N and deoxyribonucleicacid (DNA) in growth-competent mycobacteria(85), they may contribute to certain biosyntheticpathways without exerting corresponding effectsupon the chelate-dependent systems. Second,dilution of peptones may prevent this state ofimbalance and at the same time provide criticallyuseful precursors for rate-limiting steps. As anexample, Work (96) has demonstrated that thediaminopimelic acid in bacteriological peptonesis derived from microbial contamination. It willbe shown below that only microbial derivatives

121VOL. 30, 1966

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BAC=ERIOL. REV.

have been found capable of eliminating the lagperiods in the chelate-requiring mycobacteria.Having gained experience in certain conditions

which are optimal for the chelate-requiring myco-bacteria, the question of utilization of host com-ponents was resurveyed in collaboration withD. A. Power. The WR medium was employed atoptimal pH (5.0), with asparagine replaced bythe NH4 ion plus malate or fumarate. Inhibitionsagain were demonstrated when the concentrationsof Trypticase or Phytone were increased from 0.2to 0.6% and also when sera which stimulatetubercle bacilli were added at 20% (v/v). Whenorganic acids and Tween 80 were present in theNH4-type WR medium, the addition of L-aspara-gine proved to be inhibitory. It seems apparentthat the stimulation of growth has not been ob-tained by the application of host-oriented con-cepts, and that new approaches are in order.

Elimination of lag. A. terregens, the chelate re-quirer from soil, is regarded as having multi-factorial requirements, since it can achieve practi-cal rates of growth only when supplemented withvitamins or yeast extract in addition to sideram-ines (18). If judged by the criterion of growthwithout excessive lag period, the chelate-requiring

8.01

4.0

2.0

1.0

0.3

V 0.1

I

Media: - WR 100%

___ WR 509%

pHs 0 5.50 6.5A 7.5S

I inoculu sieV visible growth

////Ii

I---/I//I-/I-/II

////

///I/

//

E

E

I

V

2 4 6 8 10 12W I I K S

FIG. 4. Effect of Mycobacterium phlei supernatantfluids and ofgelatin on the lag of strain Mj68 (Powerand Hanks, unpublished data). (A) WR medium (50%)

* in water. (A') WR medium (50%) plus equal volume ofM. phlei supernatant fluid autoclaved at pH 5.5 for 15min at 121 C. (B) WR medium (100%), pH 5.5. (B')

,)\ WR medium plus gelatin (20%).r

0 2 4 6 8 10

FiG. 3. Effects ofpH, diluted medium, mycobactin,and a peptone on the initiation ofgrowth ofstrain Mj68(Wheeler and Hanks, unpublished data). Media: WR(100%) minus mycobactin and WR (50%) plus 3 ,g ofmycobactin per ml; solid symbols: presence of Trypti-case-Soy peptone (0.2%); dottedlines: abnormal growthdue to swollen cells in WR (50%) medium in the absenceof the peptone.

mycobacteria also have multifactorial require-ments. The usual statements regarding slowgrowth are inaccurate. When data are plotted asin Fig. 2 (97), it is evident that essentially one-half of the experimental periods have been due tolag prior to the onset of growth, and that subse-quently growth proceeds at reasonable rates.The long periods of lag indicate a sluggishness

in saturating media and cells with essential co-factors and metabolites. The ease with whichthese impediments can be relieved has been dem-onstrated in two different ways: by "lend-lease"of factors synthesized by more competent myco-bacteria and by "do-it-yourself" through the useof impounded cofactors. As shown in Fig. 4, thelags in onset of growth have been eliminated andrates of growth improved by adding to the WRmedium equal volumes of autoclaved supernatantfluids in which M. phlei had grown and also bymerely adding 20% gelatin to the WR medium.Both methods reduced lag periods by approxi-mately 5 weeks.

a0i.jJ

0a-

S'2a-Wi

I1

122 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

The absence of lag periods within tissue cellsand the stimulating effects of CO2 have beendocumented elsewhere (93). Highly viscous solu-tions of gelatin enclose the bacilli within a multi-tude of artificial acidic vacuoles, perhaps produc-ing situations which are analogous to those in thephagocytic vacuoles of tissue cells. The tests in aviscid menstruum of gelatin were suggested bythe work of Necas (62) and Landman and Halle(54) on the regeneration of complete yeasts andbacteria from protoplasts and L forms. The latterinvestigators suggested, inter alia, that mass con-versions of these labile forms to complete rodsmight be due in part to the retention of wall pre-cursors in close proximity to cell surfaces.

DiscussionBearing in mind that the foregoing information

has been obtained from factor-requiring models,not from host-dependent species, analysis of thefollowing sequence of points may assist in clar-fying prevailing views regarding the origins ofhost-dependent microbes in general. (i) The fac-tor-requiring mycobacteria are adapted to growwithin the phagocytic vacuoles, but not in thecytoplasm of host cells or in body fluids. Thisdisadvantageous property could not have beenselected or imposed by prolonged existence withinhost cells or body fluids. (ii) This and other prop-erties in no way resemble those of tubercle bacilliand brucellae, which may have passed throughanimal hosts for an equivalent period of time.(iii) The interesting peculiarities are simply coun-terparts of those found in well-known groups ofnonpathogenic microbes. (iv) The imprint of hostadaptation is manifest by sluggishness in saturat-ing environments with cofactors.

Adaptation to phagocytic vacuoles. Discussionson intracellular parasitism frequently are mean-ingless because of failures to relate the growthrequirements of infectious agents to the fact thatanimal cells provide two very different physico-chemical environments: that within phagocyticvacuoles and that within cytoplasm.The evidence that the chelate-requiring myco-

bacteria differ decisively from tubercle bacilli andbrucellae and that their genetically determinedqualities should prevent the initiation of growthwithin the cytoplasmic compartment of host cellsseems adequate. The internal milieu of cytoplasmhas an overall pH of 7.2 and is rich in particu-lates, enzymes, proteins, nucleoproteins, and thetrade goods of intermediary metabolism. At cor-responding pH values in vitro, mycobactin is al-ways required. When growth occurs, it is poorand readily inhibited by complex nutrients derivedfrom cells or body fluids.

Phagocytic vacuoles, formed by invaginationsof cell membranes, are compartments providinglipids, organic acids, protein hydrolysates, etc.As judged by color indicators (71) and the pHoptima of the hydrolytic enzymes (46, 94), thepH ranges from 4.5 to 6.0. Within these phago-somal compartments, the pH, the nutritionalbackground, and the impounding of bacterialcofactors should promote the growth of chelate-requiring mycobacteria.

Selective pressures. The significant point re-garding a fixed adaptation to phagocytic vacuolesis that this quality could not have been favored orselected by prolonged existence within animalcells or by intermittent exposures to body fluids.An analysis of the well-known events in thenatural history of such infections will demon-strate that "intracellular" existence could favorthe heterotrophic properties observed in tuberclebacilli and brucellae, but not the properties foundin chelate-requiring mycobacteria.

In the first place, whether because of toxicityor sensitization, the phagosomal membraneswhich enclose infectious agents are susceptible tolysis. The resultant exposures to cell cytoplasmwould be acceptable to tubercle bacilli and bru-cellae, but unfavorable to chelate-requiring myco-bacteria. Second, the intracellular pathogenscannot find a continuous haven within tissuecells. Even the least toxic of the mycobacteria,rat and human leprosy bacilli, damage tissue cells(35). During the interludes before the mycobac-teria can be ingested by a new host cell, exposuresto the unfavorable pH of 7.4 and to the com-plexity and the inhibitions of body fluids are in-evitable. Inability to adapt to these universalfeatures of host environment cannot be attributedto prolonged existence within animal hosts.Under this point, known properties in two

noncultivated mycobacteria may be included.Metabolic and other studies by Hanks and Grayhave shown that serum and body fluids at theirnormal pH cause a rapid disruption of the me-tabolism (33, 39) and of the infectiousness (34) ofM. lepraemurium. Although there are no similardata on M. leprae, it will be seen that its limitedmultiplication in bacteriological media occurs atthe low pH values and under conditions suitablefor the chelate-requiring mycobacteria. Thus, theevidence from both the chelate requirers and thenoncultivated species indicates that their majorpeculiarities have been determined, not by hostcells and body fluids, but by original geneticstraitjackets.

Imprints ofhost adaptation. The foregoing factsand conclusions do not oppose classical evidencethat microbes become modified during their

VOL. 30, 1966 123

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REv.

adaptation to plants and animals. Even with in-ocula which need to multiply only five times toyield visible growth, the mycobacterial modelsare inept in saturating in vitro environments withcofactors. The relief of these impediments solelyby bacterial derivatives suggests that the multi-factorial deficiencies in these types involve sys-tems which are unique to microbes and unrelatedto host economy.

Evolutionary Implications of Genetic MarkersAside from the imprints of host adaptation, the

major peculiarities in chelate-requiring myco-bacteria may best be interpreted as an interestingcombination of properties of free-living and non-pathogenic microbes. The requirement for myco-bactin is an authentic counterpart of the sider-amine requirements in free-living microbes.Requirements for low pH are common in non-pathogenic yeasts and fungi and mandatory inFerrobacillus ferrooxidans (78), where pH valuesas low as 3.0 assist in the uptake of large amountsof iron. Lichstein (55) has suggested that low pHfacilitates the penetration of dicarboxylic acidsinto the propioni bacteria. Organic acids havebeen shown to be of prime importance in stimu-lating the onset and smoothness of growth ofsaprophytic mycobacteria, a tubercle bacillus,and chelate-requiring mycobacteria (67). Otherproperties, such as the requirement for an iron-bearing growth factor (in this instance heme), apreference for the NH4 ion and simplified media,and the stimulation of growth by dilute peptoneswithout increase in cell crop, are classical markersin cultivable ruminocola (13). Furthermore, givena synthetic medium autoclaved at proper pH, andeither mycobactin or a sufficient inoculum, thecomplete capacities for chemosynthesis fromNH4, glucose, glycerol, and organic acids hasbeen retained in these supposedly fastidiouspathogens as faithfully as in the saprophyticmycobacteria.

In brief, the genetic markers of soil and free-living microbes have not been deleted from factor-requiring mycobacteria by a presumed prolongedintracellular existence. Because of original geneticstraitjackets, they have not become "intracellu-lar" parasites at all. Being merely "intraphago-somal" parasites, it seems that the major role ofthe cells in cows and sheep has been in part toreduce natural hazards and in part similar to thatof gelatin, to provide small acidic pockets whichhave impounded mycobacterial cofactors whilepermitting the original rates of synthesis to de-cline.Given this much insight, and having demon-

strated that certain similar properties and prob-

lems occur in the noncultivated species, oneshould be consistent in the pursuit of logic. If onesuspects that evolution has not terminated, heshould speculate that even today the progenitorsof such microbes exist in favorable ecologicalniches and on occasion give rise to the famous,noncultivated mycobacteria in men, mice, andrats; in water buffalo in the Dutch East Indies; inthe green parrots of Mexico, in Pacific salmon,and even in Bolivian bullfrogs (see 36). He alsoshould suggest that those interested in the culti-vation of the host-dependent mycobacteria mayhave been betrayed by having first discoveredsuch types of mycobacteria in animal hosts.

ORIGINS OF HOST-DEPENDENT MICROBES

The lessons learned from competent, requiring,and dependent mycobacteria may not be relevantto all types of host-dependent microbes. It is ofinterest, therefore, to consider whether host-dependent states can be imposed upon growth-competent pathogens by, or within, their hostsand to examine several alternative possibilities.

RegressionsA theory of regression from growth-competent

pathogens to host-dependent microbes, and evento viral particles, fascinated earlier scholars (32,53). Such views encounter insurmountable ob-stacles. A major difficulty is the lack ofinterrelated forms with single, double, or tripledeletions of synthetic capacity. Even the familyof mycobacteria, which is so large and so diversethat new species and types are being discoveredcontinually, fails to provide the required relation-ships. No matter how long bovine-type tuberclebacilli and Johne's bacilli have lived side by sidein the lymph nodes of the cow, the antigenic andphysiological differences are so great that theyindicate separate origins. It is simplest to supposethat evolutionary and speciating factors exertedtheir influence upon the progenitors of these twotypes before they learned to live in the cow.The properties of competent pathogens argue

against the possibility that they could give rise todependent types. The classical pathogens haveretained competency in spite of prolonged andcontinuous exposures to the interiors of plantsand animals. Although these types undergo pro-found metabolic (8) and biochemical differentia-tions (79, 83) in vivo, they defend their capacitiesto regenerate all systems which are useful in vitro.Selection tends to bring the more competent andvirulent clones into predominance. Those withthe greatest ability to synthesize cell walls andcapsules in vivo have superior capabilities toinitiate growth prior to the onset of immune

124 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

response and altered physiology in host tissues.They also are most capable of withstanding thehazards of transmission. It matters not if theyyield less competent clones. These pathogenspresumably have not emerged as separate speciesof infectious agents because of lying in the wake offorms which kill the more susceptible hostsand excite the defenses of the more resistantindividuals. Finally, although immunologicalmechanisms in animal hosts may impair systemsand structures in competent pathogens, suchmechanisms could not explain the occurrence ofhost-dependent microbes in plants.We conclude, therefore, that growth com-

petency in pathogenic microbes establishes adefinite trend against host dependency.

Antiquity ofDiseaseUncritical audiences are delighted to learn that

host-dependent states arise in microbes becauseof prolonged existence in plant and animal hosts.But who will demonstrate that the staphylococci,typhoid bacilli, brucellae, and tubercle bacilliare more recent, whereas rickettsia, psittacinebacteria, and leprosy bacilli are more ancient?As students of infectious disease most of us havefocussed too narrowly on plants and animals. Wehave not perceived that every class of dependencyknown in medical microbiology is exhibited inendless variety among the free-living myco-parasitic fungi (3). These dependencies includethe necessity for proximity in order to acquireunidentified growth factors, intracellular de-pendence upon living cells, insidious syplasticgrowth as wall-deficient forms (transitional Lforms),and even the phage-viral trickof extrudinggenetic units and cytoplasm into new host cells.If it is a question of antiquity, we may assumethat many dependent microbes lived among andupon the earliest forms of life for eons prior tothe emergence of plants and animals.

Possibilitiess Having eliminated the influence of hosts upongrowth-competent pathogens and questioned thetheory of antiquity, we may examine clues ob-tained from experimental studies on microbesthemselves. The mutagenic effects of ultravioletlight cause specific deletions of competency invitro and probably in nature. Certain deletions,e.g., the synthesis of diaminopimelic acid andtuberculostearic acids, create microbe-dependentmicrobes. Whether mutagenic deletions couldinduce requirements for metal chelates and thusexplain a well-known, widespread natural de-ficiency could be examined experimentally. Thequestion is whether microbes thus impaired could

survive in nature by borrowing from adjacentmicrobes and also have the fortune to find andtolerate the havens offered by plants or animals.During the evolutionary process, it matters littlehow many millions of experiments fail. There isalways the possibility that one form of bad luckcan be compensated within an accidentallyfavorable niche.

Given more time and experience, recent tam-perings with evolution by exposing microbes todrugs and antibiotics, both in vitro and in vivo,may yield instructive insights. A succinct sum-mary of the effects of antibiotics on microbialmembranes and of resultant problems in celldivision, wall formation, and cultivability willbe found in Landman's discussion of thepsittacine bacteria (see 31).The likelihood of producing ecologically suc-

cessful alterations in growth-competent patho-gens in vivo seems remote. The complete deletionof competent forms from many individuals leavesthe impaired forms confronted with the problemof emerging in experienced populations while themore readily transmitted forms are immunizingall in whom they gain a foothold.

It is known, however, from the observations ofWittler et al. (95) that antibiotics can inducewall-deficient, tissue cell-dependent states inlabile and potentially pathogenic strains of diph-theroids. A similar range of sub-bacterial formsand the attendant difficulties in the regenerationof complete cells can be induced in vitro (2, 21).To summarize, I would offer the view that

growth-competent microbes are not easily as-sembled and that, having circumvented theendless vicissitudes of creation and of becomingpathogenic, they are not likely to be degradedto consistently host-dependent states. The bestsuppositions appear to be: (i) that innumerabledefective creations failed until competent micro-organisms became established, (ii) that thence-forth they have been sustained by borrowing andby more intimate dependencies, and (iii) thatsome of these labile forms [e.g., those with prob-lems in cell wall formation (see Problems Sharedby Host-Dependent Microbes)] have benefittedfrom the favorable environments later affordedby plants and animals. Since such types shouldkill hosts less readily and alter immune responseless dramatically, there might be left open awider pathway of evolution toward additionaldependencies.

PROBLEMS SHARD BY HOST-DEPENDENT MIcRoBEsThe great variety of problems in energetics,

synthesis, cell wall formation, and "leakiness"in the plasmodia, the rickettsiae, and the

VOL. 30, 1966 125

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REV.

psittacine bacteria have been summarized byMoulder (60). Among the difficulties which haveprevented the study of noncultivated mycobac-teria on a similar basis, the foremost has been amisconception concerning the role of rugged andrelatively impenetrable cell walls and of the wayin which new cells are generated in vivo. Parallel-isms between the growth cycle of the psittacinebacteria, or bovine-type tubercle bacilli, and of M.leprae are instructive.An in vivo dissociation of spheroplast genera-

tion from cell wall formation and the role of wallformation in the pathogenesis of disease areclearly illustrated by Moulder's (61) descriptionof insidious onset of disease during the time whenrapid replications of protoplasts outstrip cell wallformation, and phenomenal increase in toxicityand transmissability as growth slows down andthe replicated units become enclosed withinanalyzable cell walls. Perhaps the term "transi-tional L" forms could be used to interpret therapid replications of genetic units and cytoplasmwithin flexible and slowly dividing membranesand "complete cells" for the "dense bodies"which possess mature cell walls.While studying the infection of rabbits with

bovine-type tubercle bacilli, Brieger, Fell, andSmith (9) were intrigued by the virtual disappear-ance of stainable bacilli and the insidious develop-ment of a fulminating infection before completebacilli could be found again. Though spleenhomogenates were devoid of stainable or cul-tivable mycobacteria, they were highly infectiouswhen passaged to guinea pigs. Electron micro-graphs of sediments obtained by high-speed cen-trifugation of spleen and lung homogenatesrevealed an abundance of 5- to 8-IA spheroplasts(10). Brieger and Glauert (11) stated: "Ultrathinsections of tissues from infected rabbits did notcontain any structures which could be identifieddirectly as bacilli. However, very dense, irregu-larly shaped bodies (0.2-0.8 A) were seen in anumber of cells and were clearly distinguishablefrom the mitochondria and other cell inclusionsof normal cells." Cell cultures of spleen fragmentsexplanted during the "eclipse" contained few,if any, demonstrable acid-fast bacilli during thefirst 6 to 7 days, but then became loaded withmycobacteria at unexplainable rates (9). In1954, I assisted these workers in demonstratingthat during the latent period many cells werealready richly loaded with gram-positive nucleoidbodies, both individually and in short rows.Although M. leprae must be studied as re-

covered from lesions, i.e., without control ofgrowth cycles, its propensity for growth withinweak walls resembles that of the psittacine bac-

teria. Because of the large spherical masses illus-trated by Danielssen and Boeck (23) in 1847 andthe descriptions used by Hansen (41) from 1868 to1873, we must conclude that transitional L formswere the first evidence that microbes cause diseasein humans. The failure to recognize the sig-nificance of L forms in lepromatous leprosy untilrecently is an interesting example of a celebratedinvention causing a total eclipse of learning. It isironic that Neisser (65) in 1879 and throughoutthe remainder of his life took such controversialpride in his staining of "Bacillus leprae" by dyes.(The spheroplasts and L forms are not acid-fast,and their delicate membranes are ruptured duringthe drying of films.) As late as 1910, Unna (89)and others were fascinated by the undiscernedmeaning of the unique "globus" form of growth.However, the staining of rugged rods outlivedthese earlier workers and interesting truths wereneglected by microbiologists until 1963, whenstudies on the cultivation of M. leprae fell intothe hands of two men with sufficient training incytology to use wet mounts routinely.

In early 1963, B. R. Chatterjee in this labora-tory observed that transitional L forms were pro-liferating from single rods placed in a sphero-plasting medium, and he stated that similar formswere abundant in lepromas. He enlarged upon thesignificance of such forms with respect to theinsidious, then often explosive, onset of leproma-tous leprosy and upon possible relationshipsbetween wall-deficient forms and latency of thedisease. In August, C. V. Reich, formerly of thislaboratory, reported from the LWM Laboratoryin Cebu, Philippines, that M. leprae showedapproximately a 10-fold increase in the numbersof globi in vitro. He soon reported that this oc-curred only when the pus from reaction blisterswas placed in cultures of a coccodiphtheroidbacillus "X"; that the membrane-bound formsfrequently resembled a "bag full of marbles";and also that the globi, when ruptured, releasecomplete bacilli and round bodies of about 1 ,uin size.

In retrospect, these two series of observationsclarified the actual implications of Hansen'searliest statements and of isolated observationsby others. Denney's (25) "wet-mount" study in1934 contained descriptions, demonstrations, andmicrographs which prove that flexible membranessurround the L forms of M. Ieprae in vivo. Friere(29) observed the development of transitional Lforms in vitro, but his attention apparently wasdiverted to the study of yeastlike contaminantsin his material. Micrographs shown by W. M.Meyers in 1963 convinced Chatterjee and methat C. K. Becker, a missionary in the depths of

126 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES12

the Congo, also had seen proliferations of transi-tional L forms of M. leprae in organ-type culturesof lepromata before 1960 (5).

Further work by Chatterjee with water-washedrods has demonstrated that spheroplast inducers,such as penicillin, are neither necessary nor de-sirable. Low pH, high C02, and other conditionsuseful to M. paratuberculosis, and tested thus far,are advantageous. Because of the limited materialavailable and the variety of developing forms,the extent of replication of genetic units and cyto-plasm has not been estimated. As measured bythe less successful conversions to rugged rods, thenet increases in bacilli have risen from approxi-mately 10 to 30 to 40 times the inoculated num-bers before growth ceases. This means that opera-tional knowledge has been improved but that thekey factors or conditions required for full andtransplantable growth have not been found.A survey of a spectrum of mycobacteria shows

that the problem of wall formation becomes morecomplicated as one progresses from saprophytesto noncultivated species. During maximal ratesof growth in synthetic media of approximately1 iso-osmol and containing 1 to 3% of Tween 80,the saprophytic M. phlei produces flexible orswollen cells. In the same medium the attenuatedtubercle bacillus RlRv develops an osmoticsensitivity that interrupts growth for some hours(67). In similar circumstance, M. paratuberculosisrequires 3.5 iso-osmols and grows as spheroplasts.In the absence of Tween 80, M. paratuberculosisis handicapped in the WR medium at 1.8 iso-osmoles, but thrives at 3.5 iso-osmoles. In similarmedia the limited multiplication of M. lepraeoccurs in the form of spheres and transitionalL forms.

Although these observations suggest decliningcapacities for wall synthesis, there is also evidencethat the walls constructed by saprophytes andthe tubercle bacilli are less complicated than thoseon the noncultivated species (37). As judged byimpenetrability, the differences are at least oneorder of magnitude. During incubation in physi-ological solutions, the intervals required to stain50% of cells by 0.04 M crystal violet were: forM. phlei, 0.15 hr; for BCG, 0.45 hr; for M.lepraemurium, 4 hr. The periods required forM. leprae from untreated and sulfone-treatedpatients ranged from more than 2 hr to 0.4 hr,respectively. Differences in the periods requiredto stain 100% of rods were much greater, 30 minfor M. phlei and 14 days for M. lepraemurium.For some time, the extreme impenetrability ofthe noncultivated species was regarded as a po-tential impediment to the transport of nutrients.It is exciting to discover that these impenetrabili-

ties are altered during outbursts of new growth.It now appears that supplementation of wallformation has become the more challengingproblem.The slugishness of cell wall formation by cer-

tain host-deependnt microbes will be recognizedas extreme modulations of phenomena whichoccur during differentiations in yeast and fungiand during the rapid growth of microbial cultures.The vulnerability of young as compared withmature cells was described many years ago (74,75). This has since been explained by the fact thatrates of synthesizing genetic units and cytoplasmexceed the rates of cell wall formation, and bythe subsequent gains in wall components whilethe synthesis of internal components is taperingoff (76). In view of the handicaps imposed uponhighly competent microbes by cell wall impair-ments, it may be that one of the major vulnera-bilities in consistently intracellular microbes hasbeen recognized.

Introspections on Circumventing CertainImpediments to In Vitro Study

The problems in host-dependent microbes areso diverse and in some instances so serious thatto discuss their solution would be presumptious.Nevertheless, I would emphasize that, before pro-posing to deal with truly deep-seated problems, afirst consideration is to protect, activate, andsupplement all operable systems. To do so re-quires attention to four principles.

Physicochemical environments. The present in-vestigations illustrate the necessity of relating thedefinable properties of an infectious agent to thetwo very different conditions which prevail in thephagocytic vacuoles and in the cytoplasm of hostcells. Though this compartmentalization is de-stroyed in sensitive or heavily infected cells, it islikely that one compartment or the other willfavor the growth of highly specialized microbesto critical levels. Larger spaces can then besaturated with cofactors that permit more exten-sive growth.With regard to intracellular microbes in gen-

eral, the following considerations are funda-mental. Mammalian cells provide physiologicallybalanced electrolytes, the ratios of major cationsbeing inverse to those in extracellular fluids (77).Internal protein concentrations of 28% increaseosmotic pressures by approximately 1 iso-osmole(77). It will be noted that important osmoticeffects are produced by amphoteric macromole-cules. For 60 years it has been known that sucroseand other nonelectrolytes displace cations fromcell surfaces (47). One or two cations and anexcess of sucrose are merely crutches that prevent

VOL. 30, 1966i 127

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REV.

osmotic rupture and delay cell death. Experimentsby Power and Hanks with M. paratuberculosisin a synthetic medium providing approximately3.5 mammalian iso-osmols demonstrated thatboth spheroplasted forms and whole cells are in-hibited when the osmotic pressures of the WRmedium were further increased by means of NaCl,KC1, NH4Cl, or more concentrated medium.Nevertheless, the increased osmotic pressures dueto adding 20% gelatin permitted prompt growth.In view of the importance of cations for thegeneration of complete cells from spheroplasts(54), and the differing ionic exchange capacitiesof various macromolecules, their influence onionic strengths and ratios cannot be ignored.

Protection of labile systems. Soft agar mushesimpede convection currents, trap CO2, and es-tablish oxygen gradients suitable for aerobic,microaerophilic, or anaerobic growth. Densesolutions of viscid molecules, such as gelatin, aresuperior in several respects: greater osmotic pres-sures, superior impedance of diffusion, and theseries of marked stratifications which developduring incubation in containers that preventevaporation of moisture. This column stabilityundoubtedly establishes gradient concentrationsof oxygen. Such conditions may protect labilemicrobes, supplements, or supplementing sys-tems which do not tolerate adverse or unstableoxidation-reduction equilibriums in vitro.

Impounding ofmetabolites. A hungry boy, whilepassing a banana stand, consoled himself with anancient adage: "The Lord helps those who helpthemselves."Many of the universal shortcomings in host-

adapted microbes might be overcome to a re-markable degree by impounding near cell surfacesthose cofactors and metabolites which are leakedtoo rapidly or synthesized too slowly. Saturationof systems and cells with soluble factors is in-fluenced by inoculum size, by rates of synthesis,by the impounding effects of cell walls or host cellmembranes, etc. Competent microbes synthesizeat high rates and are corsetted within cell walls.Nevertheless, the beginner in microbiology istold that a single microbe can grow on agar, oreven better in poured agar or agar mushes, butthat it may be lost in even a modest sea of thesame composition. Being devoid of cell walls,delicate spirochetes, mycoplasmas, and L formsare propagated in soft agars more readily than inliquids. The cells of higher forms of life are ac-customed to environments which entrap co-factors. Though plant cells have rigid cell walls,they are cultivated as tissue cubes. Mammaliancells are habituated to pockets within a gelatinousmatrix containing hyaluronic acid and mucopoly-

saccharides. In vitro such cells require largeinocula, diffusion barriers, or metabolic assist-ance. In primary explantations, lack of the usualmetabolites and the original gelatinous matrix iscompensated by inocula of 106 or more cellsper milliliter of liquid (73). The cloning of singlecells from fairly competent cell lines dependsupon "feeder layers" of living cells (68) or uponimplantation in microdroplets or capillary tubes(73). The student of apparently obligate intra-cellular microbes will note that what is fair forthe host-cell "goose" might be even more im-portant to the intracellular "gander." The studentof phagocytosis will recognize that properlydevised impounding systems can be the equiva-lent of artificial phagocytic vacuoles. It will beevident that the efficient impounding of singlemicrobial cells substitutes for the inoculation ofa critical mass of cells, yet avoids the rapid de-pletion of factors or nutrients which may bepresent in limited amounts.Agar gels impede diffusion measurably but

inefficiently (54). The earliest discovered demon-stration of markedly superior impounding wasprovided by Necas (62). Employing yeast proto-plasts which could not grow in a liquid medium,but which multiplied as L forms in soft agar(about 1% of cells regenerating walls after 5 daysof colony growth), he found that essentially100% of the protoplasts generated walls during24 hr of incubation in gelatin, provided the con-centrations were in the range of 20 to 40%. Thisresult was not obtained on gelatin surfaces or byadding hydrolysates of 30% gelatin. It dependedupon actually enveloping the protoplasts withingelatin solutions which gelled at the incubationtemperatures. Landman and Halle (54), em-ploying protoplasts from a nonsporulating strainof B. subtilis with strong propensities for re-generating cell walls, showed that 2.5% agar issuperior to 0.9%. However, only 5 to 15% of theprotoplasts planted on 2.5% agar succeeded inregenerating walls, and then only after the "L"forms had multiplied into masses. After demon-strating that gelatin solutions of 15 to 35% per-mitted regeneration of walls of 100% of theprotoplasts, they cited evidence that the im-pedance of diffusion in agar is minimal. From thiswork and a subsequent study by Ryter andLandman (72), it appears that the inefficiencyof protoplast growth as L forms is related to the"ballooning" of protoplasts in the usual osmoticenvironments and to the loss of mesosomal in-vaginations. It was not determined whether rever-sions in gelatin depend primarily upon theimpounding of wall-forming factors or uponosmotic shrinkage and re-establishment of the

128 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

mesosomal crypts that are vital to septum forma-tion and cell division.

In using dense solutions of gelatin to eliminatethe 5-week lag periods in the chelate-requiringmycobacteria, Power and Hanks sought merelyto impound slowly synthesized cofactors. Thebenefits were attributed to the viscosity andosmotic effects of high concentrations of gelatinfor three reasons: (i) the chelate-requiring myco-bacteria are chemosynthetic, (ii) 20% gelatinwas five times more effective than 10% solutions,and (iii) peptone-type fragments of gelatin wouldbe expected to be inhibitory if exceeding 0.2%.As noted earlier, it cannot be assumed that acommercial product such as gelatin is free fromwall-forming supplements.

In many instances synthetic macromoleculesmight provide the necessary critical effects with-out complicating nutritional backgrounds. Thebeneficial effects of methyl cellulose in shake cul-tures of mammalian cells were first attributed tohigh viscosity. However, dilute solutions of lowviscosity grades have since been found to beequally effective (12). Bryant (personal communi-cation) has suggested that adsorption on cellsurfaces may be an additional mode of actionfor such macromolecules.Even though not tested on an adequate basis,

the principles discussed already have been shownto be useful to host-dependent microorganisms.Trager (86) observed that 6% gelatin plus 0.7%bovine serum albumin fraction V improved themaintenance of Plasmodium lophurae. Thepresence of these two proteins enabled themalarial parasites to utilize contributed adenosinetriphosphate (ATP) and pyruvate and to developbeyond the merozoite stage.A practical advantage of dense solutions of

macromolecules over agar is that the physiologicalstate of cells can be studied after the cells havebeen recovered by dilution and washing.

Lend-lease of critical cofactors. Borrowing hasbeen emphasized as a universal expedient amongheterotrophs. The remarkable effects of factorsother than mycobactin on chelate-requiring myco-bacteria have been illustrated in Fig. 4. The rapidformation of satellite colonies of M. paratuber-culosis near unrelated contaminants on diagnosticmedia containing autoclaved cells of competentmycobacteria has been described several times(see 84). Similar effects on hitherto host-dependent microbes are exemplified by Reich'sdiscovery that a coccobacillus assists M. lepraein vitro and by the unknown factors in culturefiltrates which permit the growth of new types oforal spirochetes (42). It might be wise to exploremore widely the principle that competent

microbes may produce metabolites and precursorswhich are uniquely useful to noncultivatedmicrobes.The borrowing of supplements for energy-cap-

turing systems is a fascinating and no longercontroversial topic. The dictum that phos-phorylated intermediates cannot enter wholecells can be demonstrated conveniently, providedone uses competent cells and completes his ob-servations within 30 to 60 min. During the earlyadjudication of such questions it was not knownthat membranes in intracellular microbes are"leaky" or "open." Moulder (60, 61) has citedsome 16 instances of the usefulness of adenosinediphosphate (ADP), ATP, nicotinamide adeninedinucleotide (NAD), acetyl CoA, etc., to rick-ettsiae, the psittacine bacteria, and malarialparasites. Chatterjee and Williams (21), whilestudying the various types of incomplete bacteriawhich arise from the chromatinic bodies of B.megaterium, demonstrated that ATP and yeastextract increase the regeneration of completecells.

Imponderables. Microbes with limitations insystems which normally are used for independentgrowth are likely to be extremely susceptible toinhibitions, unbalanced growth, or the accumu-lation of end products which competent microbesutilize or excrete in harmless forms. The associa-tions between inhibitions and limitations in growthcapacity within the spectrum of mycobacteriahave been reviewed by Hanks and Gray (40) andfurther illustrated in this review. In the case ofthe chelate-requiring mycobacteria, presumptiveevidence of inhibitions by volatile end products ofmetabolism has not been analyzed. The leproma-tous patient has a distinctive odor, not unlikethat ofan incubator full of mycobacterial cultures.This indicates both the production and thedisposal of volatile end products from M. leprae.The various mechanisms by which plant andanimal hosts may oxidize, utilize, excrete, orrespire microbial end products are lacking invitro. Feedback inhibitions may be more in-hibitory to host-dependent types than to compe-tent species.Given conditions or factors which promote

competent performance, it may be hoped thatcertain of these imponderables will become un-important. If not, replacement of one aspect ofhost activity may require considerable ingenuity.

SUMMARY AND APPRAIsALComprehension of the fascinating problems

presented by host-dependent microbes lags farbehind existing information regarding the inter-dependencies between heterotrophic forms of life.

VOL. 30, 1966 129

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REV.

Therefore, it has seemed of interest to test whetherspecialized knowledge of given groups of host-dependent microbes, if examined in terms ofbroader precepts, might afford at least rudi-mentary insight into the origin and character ofhost-dependent states. The results of this ex-amination have led me to challenge the view thathost-dependent microbes have been derived fromgrowth-competent types, and to suggest that theymore likely arose from types which were defectiveprior to their adaptation to plants or animals.One point of departure is to recognize that a

community of symptoms in the intracellular typesof host-dependent microbes is related to mem-branes that yield delayed or defective cell wallformation, and that plant and animal hosts arenot engaged in that line of business. It, therefore,may be useful to look to growth-competentmicrobes for cofactors or supplements whichmay be lacking in, or supplied to, host-dependentmicrobes, both in vivo and in vitro.A further step toward clearer understanding of

the origin and character of host-dependentmicrobes depends upon the fact that loaning andborrowing of strictly microbial growth factors isas natural as the more familiar transactions infactors which are used by all forms of life. Geneti-cally immutable dependencies upon microbialgrowth factors, such as iron chelators, occur in aseries of free-living microorganisms as well asin the most fastidious of the mycobacteria whichcan be maintained on bacteriological media.Similar requirements for unknown, strictly micro-bial, factors occur among soil microbes and indelicate oral spirochetes. Some of the ruminalbacteria are cultivable and others are not. Amongthe fungi that live in soil, or stumps, or any oldplace, many exhibit dependencies upon strictlymicrobial factors or upon living fungi. Givenproper study, the number of noncultivatedmicrobes in nature could well exceed the numberknown to students of infectious disease. In brief,although dependent states should be fostered byprolonged existence within plants or within theparenteral compartments of animals, there islittle reason to assume that they originated bythat mechanism.The deliberate offering of strictly microbial

growth factors and components during surveysof soil, plant rhizospheres, and the miscellaneousnooks and crannies in animals should lead tothe discovery of additional microbe-dependentforms. Considering the potential multitude andthe manifold potentialities of such microbes, itwould be expected that some have learned tolive in intimate association with plants andanimals. As long as the strictly microbial char-

acter of certain deficiencies is not suspected, anysuch microbe can masquerade as host-dependentwhile depending upon microbes for special fac-tors and upon the host for those used moreuniversally.For microorganisms to utilize strictly microbial

growth factors while growing in plants andanimals seems to present no fundamental prob-lem. For example, the rootlets of plants do notexclude other soil extractives while taking upminerals, nitrates, and vitamins. The digestivetract in animals cannot exclude microbial growthfactors while absorbing microbially synthesizedvitamins and antibiotics and even proteins frommilk and egg white. In the case of microbialfactors which are not metabolized rapidly by thehost, the seclusion of a microbe within mam-malian cells presents no impediment to theirutilization. The intraphagosomal utilization ofexternal supplies of iron, glycerol, or mycobactin,a microbial growth factor with a molecular weightof 870, has been clearly demonstrated. Otherstudies with securely phagocyted microbes haveshown that proteins, such as lysozyme andplasma globulins, are not excluded from thiscompartment of mammalian cells. Intraphagoso-mal microbes, therefore, do not depend solelyupon the components or systems of animal cells.Aside from the universal benefits of osmoticprotection and the impounding of cofactors whichare synthesized very slowly, the nutritional situa-tion for intracellular parasites does not differfrom that for the extracellular types as funda-mentally as has been suspected.

Specific clues regarding the general relation-ships outlined above have been obtained bycomparing chelate-requiring microbes from soilwith analogous mycobacterial pathogens, whichwere noncultivable for some years and have sincebeen regarded as highly fastidious, and by analyz-ing the conditions which permit the latter togrow in vitro and within mammalian cells. Re-quirements for pH and simple nutrients demon-strated that these pathogens are adapted to growin phagocytic vacuoles, but not in body fluidsor in the cytoplasmic compartment of tissuecells. This indicates that original genetic straight-jackets, though highly disadvantageous, havenot been modified by prolonged existence inmammalian cells or by repeated exposures tocytoplasm and body fluids. Furthermore, in spiteof selective forces in vivo, the lag periods ingrowth could not be shortened by means of hostcomponents or derivatives from eggs. Thesecaused more inhibitions than stimulations. Mean-while, aside from the chelate requirement, thesepathogens have retained the synthetic capacities

130 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

of saprophytic mycobacteria. In fact, the majormarkers and requirements proved to be thosecommonly found in free-living and nonpathogenicmicrobes.

After attempts to reduce the protracted lagphase in the absence of tissue cells, it was con-cluded that the only imprint of prolonged ex-posure to intracellular habitat was sluggishnessin saturating environments with cofactors andmetabolites. These effects of host adaptation wererelieved completely by two principles which areapplicable to microbes in general:the impoundingof slowly synthesized factors near cell surfaces bymeans of gelatin, or the lending of factors in thesupernatant fluids of growth-competent myco-bacteria.Such observations indicate clearly that "adapta-

tion to an animal host" has played no crucial rolein molding the character of these once apparentlynoncultivable microbes.The question whether growth-competent patho-

gens could give rise to host-dependent typeswithin host environments was raised by thestriking contrasts between the behavior of thechelate-requiring mycobacteria and tuberclebacilli or brucellae, both in vivo and in vitro. Inthe case of growth-competent forms, rapidpassages or prolonged infections promote theselection of increasingly virulent clones. Beingmore readily transmissable, these clones kill offthe susceptible and immunize the more resistantmembers of host populations. Since this processshould block the emergence of less competentsublines, it was concluded that competency, onceacquired, cannot be reversed. An effort was madeto comprehend the possible ways in which funda-mental impairments might be induced in com-petent microbes either before or during theirsojourn in hosts. To include both plant andanimal pathogens, it was necessary to disregardimmunological impairments of cell wall forma-tion in animal hosts. The considerations whichseemed pertinent led to the suggestion that, oncesuccessful forms of life existed, less perfect crea-tions could have been sustained in favorableecological niches and later have found havens forfurther evolution in plants and animals.

Irrespective of theories, universal symptoms independent microbes indicate different types anddegrees of deficiency in membrane physiologyand cell wall formation. During introspectionson the further investigation of dependentmicrobes, primary consideration was given toprotective and supplementing measures. The fourcornerstones contributed by a variety of physio-logical investigations were: (i) knowledge of thecompartments, electrolyte balances, amphoteric

osmotic agents, and components within hostcells; (ii) methods for establishing regionalgradients of oxygen tension and for the protec-tion of labile microbes and supplements; (iii)development of systems for impounding near cellsurfaces the soluble components which sluggishmicrobes may synthesize too slowly; and (iv)lend-lease of supplements and systems fromgrowth-competent microbes. Nearly all theseprinciples can be applied simultaneously. Whenthis has been done as well as possible, the trulydeep-seated problems in synthesis or energetics atlast may emerge to be analyzed.

This discourse represents an effort to openwider the shutters and foggy windows of host-oriented concepts regarding the origins and theneeds of host-dependent microbes. Many of ourprevailing views are inadequate to progressivestudy of the interesting and delicate organismswhich we understand and manage so ineptly. Itis my hope that a mutual re-examination of newperspectives may enable us to see this poorlyilluminated terrain more clearly.

BEDTIME STORYMicrobiologicus var. medicus had been taught

to focus on expensive diseases in values hosts.Although he labored mightily, his tunnel visioncreated the liability of spending a lifetime withoutnoting that the musty odors from the dewygrass, the green-scummed pond, the rusty stumps,and the copse of trees whisper of substancesliberated for the continuance of life and ofhumbleagents which might teach the lessons he strives tolearn. Once upon a time, having struggled longwith the complexities and arguments of his ownworld, he layed aside his telescope to listen to thetall stories of Mycologicus naturalis.

Naturalis, being steeped in mythology, sacri-ficed to medicus first a generous flagon containingan intermediate from the goddess cerevisiae. Hethen considered whether his own experiencescould afford a moral or model to persons as in-nocent as medicus. Finally, he proceeded as fol-lows, quoting always from good authority (3).

Although many writings describe the orderlyprocesses and useful services of the fungi, favoritelegends dwell on departures from the norm. Inancient days, such stories were introduced bystating that fungal communities are happy aboutliving in an industrious manner and not payingtaxes. However, they may be subject to tributeat any time and in such a way that "few groupsare free from parasites, some of which may evenbe their close relatives."

Naturalis hesitated about whether to enlargeupon the growth-competent bandits, lest tales of

VOL. 30, 1966 131

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REV.

stupid cupidity might destroy the fascination hehoped to create in medicus. But, being a realist, hecould not resist calling by name the charactersthat scarcely pause to say "Jack Robinson"before launching their mortiferous antibioticsfrom a great distance, trumpeting against thewalls of Jericho until the tremor of a touch willcrumble them, or sacking the city after directassault. Once inside a residence, they plan tostay only overnight. They keep on their virulentovercoats and boots while burning the furniture.Each knows that by dawn his host will be gonetoward heaven and that he will be going for thedoor. What reckless fellows. They can be keptin custody for years, often becoming less noxious,but always disgracefully greedy to reproduce inglass houses on the slightest pretext and minimalencouragement.

Naturalis took the nodding of medicus tomean that medicus appreciated a retelling of thefolklore which he could understand. He nextthought it timely to quietly offer medicus a mouseof a story designed to keep the mountain of thelatter from falling flat on its prairie.The interesting dependencies are more subtle.

In some instances, Gamma depends upon Beta,who in turn must rely on Alpha. Some sly onesfind it sufficient to live intimately, side by sidewith a provident neighbor. Gonatobotrys, thesimplex, and Calcarisporum, the parasiticum,being perhaps addled by ultraviolet, are knownto be dependent on a growth factor which is madeand also required for the growth of many fungi(92). The elixir to which they are addicted is notknown, but it is true that they must have it, evenif the amount be only one needle amidst a millionhaystalks. [Extracts having "mycotrophein" ac-tivity (92) have since been shown to containbound hydroxylamine and to supply the sider-amine requirement of A. terregens (Morrison,unpublished data).]

Jonesii of the Chaetocladium clan, being devoidof some essential, lays a finger on the more in-dustrious Mucorales. Contrary to some reports,this Good Samaritan does not set jonesii on hisass, but in a shady nook helps him to build a pic-nic table of wonderous design (14).With others it becomes to a question of break-

ing and entering, not from boldness, but fromvital necessity. Among the Chytrids there aremany legends concerning the five Rozella brothers(48). They intrude into a new domicile by drillinga hole through the wall and squeezing throughit with the finesse of a phage, leaving their hatsand overcoats outside. They insinuate themselvesinto the good graces of the host before taking aseat at his table. During reproduction, they are

modestly invisible. The children grow and multi-ply without walling off their genes and cytoplasmin visible integuments. During their youth theyare harmless. As growth slows down they cladthemselves in nylons, hoods, and skates andprepare to make their way into the world. Ifthere is trouble in opening the door, they tearround and round within the hollow shell of hostremnants. If the Rozella brothers and their heirsunderstood righteous indignation, they could suethe psittacine bacteria and M. leprae for infringe-ment of patents.

Medicus, being weary of keeping his mind inconstant focus, could only wonder whether hisfolks had sent him to the wrong kindergarten.He loaded his syringe with mysterious yet power-ful mixtures of drugs and antibiotics. He preparedto contemplate these fanciful tales while com-posing himself for rest.

Sleep well, Microbiologicus medicus, perhapssome later night you can invent a better story.Meanwhile, there is much work for the morrow.

ACKNOWLEDGMENTSI am indebted to B. R. Chatterjee, N. E. Morrison,

D. A. Power, C. V. Reich, and W. C. Wheeler forpermission to cite unpublished data. Special thanksare expressed to N. E. Morrison for assistance withthe section on microbe-dependent microbes, to Mrs.A. M. de Czekala for Fig. 3 and 4, and to both ofthem for many helpful suggestions.

This investigation was supported by Public HealthService grants AI-02998 and AI-0220 from the Na-tional Institites of Allergy and Infectious Diseases and5-F1-GM-14961 to William C. Wheeler from the Di-vision of General Medical Services.

LITERATuRE CITED1. ANToINE, A. D., N. E. MORRISON, AND J. H.

HANKS. 1964. Specificity of improved methodsfor mycobactin bioassay by Arthrobacter terre-gens. J. Bacteriol. 88:1672-1677.

2. BACHMANN, B. J., AND D. M. BONNER. 1959. Pro-toplasts from Neurospora crassa. J. Bacteriol.78:550-556.

3. BARNE-r, H. L. 1963. The nature of mycoparasit-ism by fungi. Ann. Rev. Microbiol. 17:1-14.

4. BAUMAN, N., AND B. D. DAVIS. 1957. Selection ofauxotrophic bacterial mutants through di-aminopimelic acid or thymine deprival. Science126:170.

5 BECKER, C. K., AND E. M. BRIEGER. 1960. Organculture method of producing a leprosy-bacillussuspension (Nimorpel) for use as a lepromin.Intern. J. Leprosy 28:441-443.

6. BICKEL, H. R., E. BOSSHARDT, E. GXUMANN, P.REUSSER, E. VISCHER, W. VOSER, A. WETrSTEIN,AND H. ZXHNER. 1960. Stoffwechselproduktevon Actinomyceten, 26. Mitt. Uber die Iso-lierung und Charakterisierung der Ferrioxamine

132 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

A-F, neuer Wuchsstoffe der Sideramin-Gruppe.Helv. Chim. Acta 43:2118-2128.

7. BICKL, H., E. GXUMANN, W. KELLER-SCHIERLEIN,V. PRELOG, E. VISCHER, A. WErSTEiN, AND H.ZAHNER. 1960. Uber eisenhaltige Wachstums-faktoren, die Sideramine, und ihre Antago-nisten, die eisenhaltigen Antibiotica Sideromy-cine. Experientia 16:129-133.

8. BLOCH, H. 1960. Biochemical properties of viru-lent and avirulent strains of Mycobacteriumtuberculosis. Ann. N.Y. Acad. Sci. 88:1075-1085.

9. BREGER, E. M., H. B. FELL, AND B. R. SwEH.1951. A comparative study of the reaction invivo and in vitro of rabbit tissues to infectionwith bovine tubercle bacilli. II. Observations oncultures of spleen and lymph glands from in-fected rabbits. J. Hyg. 49:189-200.

10. BREGER, E. M., H. A. R. MILES, V. E. CossL.Err,AND R. W. HORNE. 1951. Nature and infectivityof granular forms of the bovine tuberclebacillus. Nature 168:96-100.

11. BREGER, E. M., AND A. M. GLAUERT. 1954. Thedemonstration of tubercle bacilli in ultra-thinsections of infected tissues by electron micros-copy. Tubercle 35:80-85.

12. BRYANT, J. C., V. J. EvANs, E. L. SCHILLING, ANDW. R. EARLE. 1961. Effect of chemically definedmedium NCTC 109 supplemented with metho-cel and of silicone coating the flasks on strain2071 cells in suspension cultures. J. Natl. Can-cer Inst. 26:239-252.

13. BRYANT, M. P., AND I. M. RoBNsoN. 1962. Somenutritional characteristics of predominant cul-turable ruminal bacteria. J. Bacteriol. 84:605-614.

14. BURGEFF, H. 1924. Untersuchungen uber Sexuali-tat und Parasitismus bei Mucorineen I. Botan.Abhandl. 4:1-135.

15. BURNHAM, B. F. 1962. Investigations on themechanism of ferrichrome function. Arch. Bio-chem. Biophys. 97:329-335.

16. BURNHAM, B. F. 1963. Investigation on the actionof the iron-containing growth factors, sider-amines, and iron-containing antibiotics, sidero-mycins. J. Gen. Microbiol. 32:47-121.

17. BURNHAM, B. F., AND J. B. NEiILANDs. 1961. Stud-ies on the metabolic function of the ferrichromecompounds. J. Biol. Chem. 236:554-559.

18. BURTON, M. O., F. J. SOWDEN, AND A. G.LOCHHEAD. 1954. Studies on the isolationand nature of the "terregens factor." Can. J.Biochem. Physiol. 32:400-406.

19. CHANG, Y. T. 1962. Growth of M. lepraemuriumin cultures of mouse macrophages. Intern.Congr. Microbiol. 8th Montreal, p. 122.

20. CHANG, Y. T., AND R. L. NEIKIRK. 1965. Myco-bacterium lepramurium and Mycobacteriumleprae in cultures of mouse peritoneal macro-phages (preliminary results). Intern. J. Leprosy33(Part 2) :586-598.

21. CHATERJEE, B. R., AND R. P. WILLmS. 1963.Growth of cellular forms in cultures of chroma-

tin bodies isolated from Bacillus megaterium. J.Bacteriol. 85:623-627.

22. CoLETSOS, P. J. 1951. De la dependance du Myco-bacterium tuberculose a la streptomycine. Rev.Tuberc. 15:548-555.

23. DANELsEN, D. C., AND C. W. BOECK. 1947. AtlasColori6 de Spedalskhed. Bergen en Norvege.Plate XXIV.

24. DEMAIN, A. L., AND D. HENDLIN. 1959. 'Irontransport' compounds as growth stimulatorsfor Microbacterium sp. J. Gen. Microbiol. 21:72-79.

25. DENNEY, 0. E. 1934. A microscopic study ofMycobacterium leprae. Intern. J. Leprosy 2:275-278.

26. DONNALLY, H. H. 1930. The question of the elimi-nation of foreign protein (egg-white) inwomen's milk. J. Immunol. 19:15-40.

27. ELBERG, S. S. 1960. Cellular immunity. Bacteriol.Rev. 24:67-95.

28. FRANras, J., H. M. MACTURK, J. MADINAvEmA,AND G. A. SNow. 1953. Mycobactin, a growthfactor for Mycobacterium johnei. I. Isolationfrom Mycobacterium phlei. Biochem. J. 55:596-607.

29. FREmE, S. A. 1954. Some new bacteriologicalaspects of leprosy. I. Culture of the etiologicalagent of human leprosy by the glass slide cul-ture method (Mycobacterium shape). ArquivosMineiros Leprol. 14:289-293.

30. GARBUTr, E. W. 1965. Studies of M. lepraemuriumand M. keprae in tissue culture. Intern. J. Lep-rosy 33(Part 2):578-585.

31. GORDON, F. B, AND A. L. QUAN. 1962. Drug sus-ceptibilities of the psittacosis and trachomaagents. Ann. N.Y. Acad. Sci. 98:261-274.

32. GREEN, R. G. 1935. On the nature of filterableviruses. Science 82:443-445.

33. HANKS, J. H. 1954. Influence of physical andchemical factors on the hydrogen transfer ca-pacity of murine leprosy bacilli. Intern. J.Leprosy 22:162-173.

34. HANKS, J. H. 1954. Relationship between themetabolic activity and the infectiousness ofM. lepraemurium; refrigeration studies. Intern.J. Leprosy 22:450-460.

35. HANKS, J. H. 1958. Assay of the fate of mycobac-teria in cell and tissue cultures. Am. Rev. Tu-berc. Pulmonary Diseases 77:789-801.

36. HANKS, J. H. 1960. The biochemical properties ofnoncultivated mycobacteria. Ann. N.Y. Acad.Sci. 88:1087-1092.

37. HANKs, J. H. 1961. Significance of the capsularcomponents on M. leprae and other mycobac-teria. Intern. J. Leprosy 29 :74-83.

38. HANKs, J. H., AND B. EvANS. 1940. The action ofserum, cells and blood on acid-fast bacteria invitro. Am. Rev. Tuberc. Pulmonary Diseases41:620-636.

39. HANKS, J. H., AND C. T. GRAY. 1954. Extracellu-lar inhibitors in leprotic infections and their roleas barriers to experimental transmission. Intern.J. Leprosy 22:461-468.

'v oL. 30, 1966 133

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

BACTERIOL. REV.

40. HANKS, J. H., AND C. T. GRAY. 1956. The meta-bolic properties of mycobacteria and the patho-genesis of mycobacterial disease. Advan. Tu-berc. Res. 7:1-16.

41. HANSEN, G. A. 1874. Undersogelser angaendeSpedalskhedens Arsager. Norsk Mag. Laegevi-denskaben 4:76-79.

42. HARDY, P. H., JR., Y. C. LEE, AND E. E. NELL.1964. Use of a bacterial culture filtrate as an aidto the isolation and growth of anaerobic spiro-chetes. J. Bacteriol. 87:1521-1525.

43. HENDLIN, D., AND A. L. DEMAIN. 1959. An abso-lute requirement for "Iron Transport Factors"by Microbacterium lacticum 8181. Nature 184:1894.

44. HESSELTINE, C. W., C. PIDACKS, A. R. WHITEHILL,N. BOHONOS, B. L. HUTCHINGS, AND J. H.WILLIAMS. 1952. Coprogen, a new growth factorfor coprophylic fungi. J. Am. Chem. Soc.74:1362-1363.

45. HESSELTINE, C. W., A. R. WHITEHILL, C. PIDACKS,M. TEN HAGEN, N. BOHONOS, B. L. HUTCHINGS,AND J. H. WILLIAMS. 1953. Coprogen, a newgrowth factor present in dung, required byPilobolus species. Mycologia 45:7-16.

46. HIRSCH, J. G. 1962. Cine microphotographic ob-servations on granule lysis in polymorphonu-clear leucocytes during phagocytosis. J. Exptl.Med. 116:827-834.

47. H6BER, R. 1946. Influence of some extracellularfactors on cellular activity, p. 307-311. In Physi-cal chemistry of cells and tissues. The BlakistonCo., Philadelphia.

48. KARLING, J. S. 1942. Parasitism among thechytrids. Am. J. Botany 29:24-35.

49. KARLSSON, J. L. 1956. Auxotrophic mutants ofmycobacteria dependent on fatty acids derivedfrom the parent strain. J. Bacteriol. 72:813-815.

50. KELLER-SCHIERLEIN, W. 1963. Stoffwechselpro-dukte von Mikroorganismen, 45. Mitt. Uberdie Konstitution von Ferrirubin, Ferrirhodinund Ferrichrom. Helv. Chim. Acta 46:1920-1929.

51. KELLER-SCHIERLEIN, W., AND A. DEER. 1963.Stoffwechselprodukte von mikroorganismen,44. Mitt. Zur konstitution von ferrichrysin undferricrocin. Helv. Chim. Acta 46:1907-1920.

52. KELLER-SCHIERLEIN, W., V. PERLOG, AND H.ZXHNER. 1964. Siderochrome (NaturlicheEisen (III)-trihydroxamat-Komplexe). Fort-schr. Chem. Org. Naturstoffe 22:279-322.

53. LAIDLAW, P. P. 1938. Virus diseases and viruses.The University Press, Cambridge, England.

54. LANDMAN, 0. E., AND S. HALLE. 1963. Enzy-mically and physically induced inheritancechanges in B. subtilis. J. Mol. Biol. 7:721-738.

55. LICHSTEIN, H. C. 1959. Symposium on initiationof bacterial growth. III. Physiological aspectsof growth initiation. Bacteriol. Rev. 23:261-266.

56. LIPPARD, V. W., 0. M. SCHLOSS, AND P. A.JOHNSON. 1936. Immune reactions induced ininfants by intestinal absorption of incompletely

digested cow's milk protein. Am. J. DiseasesChildren 51:562-574.

57. LOCHHEAD, A. G., M. 0. BURTON, AND R. H.THEXTON. 1952. A bacterial growth factor syn-thesized by a soil bacterium. Nature 172:282-283.

58. MORRISON, N. E. 1965. Circumvention of themycobactin requirement of Mycobacteriumparatuberculosis. J. Bacteriol. 89:762-767.

59. MORRISON, N. E., A. D. ANTOINE, AND E. E.DEWBREY. 1965. Synthetic metal chelatorswhich replace the natural growth-factor re-quirements of Arthrobacter terregens. J. Bac-teriol. 89:1630.

60. MOULDER, J. W. 1962. The biochemistry of intra-cellular parasitism. Univ. Chicago Press, Chi-cago.

61. MOULDER, J. W. 1963. The psittacosis group asbacteria. John Wiley and Sons, Inc., New York.

62. NECAS, 0. 1961. Physical conditions as importantfactors for the regeneration of naked yeastprotoplasts. Nature 192:580-581.

63. NEILANDS, J. B. 1952. A crystalline organo ironcompound from the fungus Ustilago sphaero-gena. J. Am. Chem. Soc. 74:4846-4847.

64. NEILANDS, J. B. 1957. Some aspects of microbialiron metabolism. Bacteriol. Rev. 21:101-111.

65. NEISSER, A. 1879. Zur Aetiologie der Lepra.Breslauer Artzl. Z. 1:200-202, 214-215.

66. POWELL, M. V., B. R. BYERS, AND C. E. LANK-FORD. 1965. Relation of iron and manganeseconcentrations to lag and schizokinen produc-tion by Bacillus megaterium. Bacteriol. Proc.,p. 123-124.

67. POWER, D. A., AND J. H. HANKS. 1965. The effectof organic acids, serum albumin, and wettingagents on lag phase, dispersed growth and pHstabilization in mycobacterial cultures. Am.Rev. Respirat. Diseases 92:83-93.

68. PUCK, T. T., P. I. MARCUS, AND S. J. CIECIURA.1956. Clonal growth of mammalian cells invitro. Growth characteristics of colonies fromsingle Hela cells with and without a "feeder"layer. J. Exptl. Med. 103:273-284.

69. REES, R. J. W., AND E. W. GARBUTr. 1962.Studies on Mycobacteriwn lepraemurium in tis-sue culture. I. Multiplication and growth char-acteristics in cultures of rat fibroblasts. Brit. J.Exptl. Pathol. 43:221-235.

70. REICH, C. V., AND J. H. HANKS. 1964. Use ofArthrobacter terregens for bioassay of myco-bactin. J. Bacteriol. 87:1317-1320.

71. RoUs, P. 1925. The relative reaction within livingmammalian tissues. J. Exptl. Med. 41:379-411.

72. RYTER, A., AND 0. E. LANDMAN. 1963. Electronmicroscope study of the relationship betweenmesosome loss and the stable L state (or proto-plast state) in Bacillus subtilis. J. Bacteriol.88:457-467.

73. SANFORD, K. K., W. R. EARLE, AND G. D.LIKELY. 1948. The growth in vitro of singleisolated tissue cells. J. Natl. Cancer Inst. 9:229-246.

134 HANKS

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

HOST-DEPENDENT MICROBES

74. SHERMAN, J. M., AND W. R. ALBUS. 1923. Physio-logical youth in bacteria. J. Bacteriol. 8:127-139.

75. SHERMAN, J. M., AND W. R. ALBUS. 1924. Thefunction of lag in bacterial cultures. J. Bacteriol.9:303-305.

76. SHOCKMAN, G. D., J. J. KOLB, AND G. TOENNIES.1958. Relation between bacterial cell wall syn-thesis, growth phase, and autolysis. J. Biol.Chem. 230 :961-977.

77. SHOHL, A. T. 1939. Mineral metabolism. ReinholdPublishing Corp., New York.

78. SILVERMAN, M. P., AND D. G. LUNDGREN. 1959.Studies on the chemoautotrophic iron bac-terium Ferrobacillus ferrooxidans. I. An im-proved medium and a harvesting procedure forsecuring high cell yields. J. Bacteriol. 77:642-647.

79. SMITH, H. 1960. Studies on organisms grown invivo to reveal the bases of microbial patho-genicity. Ann. N.Y. Acad. Sci. 88:1213-1226.

80. SMITH, H. W. 1955. Sources of the Phlei growthfactor for Mycobacterium johnei. Nature 175:40-41.

81. SNOW, G. A. 1954. Mycobactin. A growth factorfor Mycobacterium johnei. III. Degradation andtentative structure. J. Chem. Soc., p. 4080-4093.

82. SNOW, G. A. 1965. The structure of mycobactinP, a growth factor for Mycobacterium johnei,and the significance of its iron complex. Bio-chem. J. 94:160-165.

83. STINEBRING, W. R., W. BRAUN, AND A. POMALEs-LEBR6N. 1960. Modified serum resistance ofbacteria following intracellular residence. Ann.N.Y. Acad. Sci. 88:1230-1236.

84. STUART, P. 1954. Satellitism in cultures of Myco-bacterium johnei. Vet. Record 66:305-306.

85. TEPPER, B. S. 1965. Modification of cellular con-stituents during growth of Mycobacteriumphlei. Am. Rev. Respirat. Diseases 92:75-82.

86. TRAGER, W. 1950. Studies on the extracellularcultivation of an intracellular parasite (avianmalaria). I. Development of the organisms inerythrocyte extracts, and the favoring effect ofadenosinetriphosphate. J. Exptl. Med. 92:349-365.

87. TRAGER, W. 1960. Intracellular parasitism andsymbiosis, p. 151-213. In J. Brachet and A. E.Mirsky [ed.], The cell, vol. 4 Academic Press,Inc., New York.

88. TWORT, F. W., AND G. L. Y. INGRAM. 1913. Amonograph on Johne's disease (Enteritis chron-ica pseudotuberculosa bovis). Bailliere, Tindalland Cox, London.

89. UNNA, P. G. 1910. Histotechnik der leprosenHaut. Leopold Voss, Hamburg.

90. WALLACE, J. H., S. D. ELEK, AND J. H. HANKS.1958. Limited multiplication of Mycobacteriumlepraemurium in cell cultures. Proc. Soc. Exptl.Biol. Med. 97:101-104.

91. WEISS, C. 1939. Proteinase and peptidase activityof polymorphonuclear leucocytes, monocytesand epithelioid cells of inflammatory exudates.Am. Rev. Tuberc. Pulmonary Diseases 39:228-231.

92. WHALEY, J. W., AND H. L. BARNETT. 1963.Parasitism and nutrition of Gonatobotrys sim-plex. Mycologia 55:199-210.

93. WHEELER, W. C., AND J. H. HANKS. 1965. Utiliza-tion of external factors by intracellular mi-crobes: Mycobacterium paratuberculosis andwood pigeon mycobacteria. J. Bacteriol. 89:889-896.

94. WHITE, P. R. 1954. The cultivation of animal andplant cells. The Ronald Press Co., New York.

95. WrTTLER, R. G., W. F. MALIZA, P. E. KRAMER, J.D. TuQKETT, H. N. PRICHARD, AND H. J.BAKER. 1960. Isolation of a corynebacteriumand its transitional forms from a case of sub-acute bacterial endocarditis treated with anti-biotics. J. Gen. Microbiol. 23:315-333.

96. WORK, E. 1957. Biochemistry of the bacterial cellwall. Nature 179:841-847.

97. YouMANs, G. P., AND A. S. YouMANs. 1949. Amethod for the determination of the rate ofgrowth of tubercle bacilli by the use of smallinocula. J. Bacteriol. 58:247-255.

98. ZXHNER, H., E. BACHMANN, R. HUTTER, AND J.NUESCH. 1962. Sideramine, eisenhaltige wach-stumsfaktoren aus mikroorganismen. Pathol.Microbiol. 25:708-736.

VOL. 30, 1966 135

on Septem

ber 17, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from