Bacteria as Multicellular Organisms They differentiate into various cell types and form highly regular colonies that appear to be gUided by sophisticated temporal and spatial control systems

by James A. Shapiro

ithout bacteria, life on earth where they form organized multicellu­ it is a Simplification that is likely to could not exist in its present lar structures that function as facto­ prove invalid in many cases. How ex­ Wform. Bacteria are key play­ ries for nitrogen production. At about ceptional-or how common-are mul­ ers in many geochemical processes, the same time Sergei Winograd sky, ticellular features in bacteria? In in­ including the fundamental nitrogen, working in Paris, elucidated the role vestigating this question I have con­ carbon and sulfur cycles, which are bacteria play as decomposers of cellu­ cluded that most-perhaps virtually critical to the circulation of life's basic lose in the global carbon cycle. Wino­ all-bacteria lead multicellular lives. elements. If these processes were to gradsky was also one of the first mi­ Examples of multicellularity among grind to a halt, the planet's soils, wa­ crobiologists to observe bacteria di­ the bacteria abound; indeed, some of ters and atmosphere would become rectly in the soil, where he found that the complex biochemical processes inhospitable for life. Yet in spite of few exist as isolated cells; most live in performed by bacteria could not be such global importance, the notion groups adhered to soil particles. Simi­ carried out as effectively without or­ has persisted that bacteria are simple lar group behavior was well known in ganized groups. Photosynthesis is a unicellular microbes. the laboratory, where bacteria formed process that illustrates this point in That view is now being challenged. distinctive colonies on petri dishes or several ways. Photosynthetic bacteria, Investigators are finding that in many adhered as organized populations to like green plants, rely on solar energy ways an individualbacterium is more the walls of culture flasks. to convert carbon dioxide into organic analogous to a component cell of a In spite of these early observations, chemicals. One group of photosyn­ · multicellular organism than it is to the image of bacteria as unicellular thetic bacteria, known as the cya­ a free-living, autonomous organism. organisms has persisted over the nobacteria, often grow as connected Bacteria form complex communities, years. In large part this can be at­ chains of cells or as intertwined mats; hunt prey in groups and secrete chem­ tributed to medical bacteriology. Dis­ they contain a form of chlorophyll and ical trails for the directed movement ease-causing organisms are commonly in many ways resemble multicellular of thousands of individuals. identified by isolating a single cell algae. For many years, in fact, the cya­ Already at the beginning of this cen­ of the suspected pathogen, growing nobacteria were thought to be mem­ tury investigators had evidence that a culture from that cell and showing bers of the plant kingdom. The multi­ bacteria live communally in the soil. that the resulting pure culture gives cellular organization aids in light har­ Martinus Beijerinck of the Netherlands rise to the disease in question. The vesting but yields other benefits too. discovered that Rhizobium bacteria in­ possibility that infections of the hu­ Anabaena, an inhabitant of freshwa­ fect the roots of leguminous plants, man body involve multicellular aggre­ ter ponds, is one of the best-known of gations of bacteria is normally not the photosynthetic bacteria. Anabae­ even considered. na is capable of both photosynthesis JAMES A SHAPIRO is professor of mi­ Indeed, many existing theories of and nitrogen fixation. These two bio­ crobiology at the University of Chica­ bacterial growth, physiology and ge­ chemical processes are incompatible go. He earned his bachelor's degree in netics are formulated exclusively in within a single cell because oxygen, 1964 English at Harvard College in and terms of the isolated bacterium. From produced during photosynthesis, in­ was awarded a Marshall Scholarship for an epistemological standpoint this activates the nitrogenase required for postgraduate study at the University of emphasis on the single cell is curious. nitrogen fixation. When nitrogen com­ Cambridge, where he received his Ph.D. in genetics. He spent a year as a post­ In practice most research is carried pounds are abundant, Anabaena is An doctoral fellow at the Pasteur Institute out on cell populations. enzyme strictly photosynthetic and its cells in Paris, almost twoyears as visiting measurement, for example, may be are all alike. When nitrogen levels are professor at the University of Havana based on an extract from 100 million low, however, specialized cells called and several months at Tel-Aviv Univer­ cells, but conclusions based on the heterocysts are produced. The hetero­ sity; he was involved in the U.S.-U.S.S.R. results are often made under the as­ cysts lack chlorophyll but synthesize microbiology exchange program from sumption that every bacterium in a nitrogenase, an enzyme with which 1975 to 1978. In 1973 he joined the population is more or less the same. they are able to convert nitrogen gas faculty at Chicago. This is Shapiro's sec­ ond article for SCIENTIFIC AMERICAN. Such a premise may simplify the inter­ into a usable form. pretation of experimental results, but James W. Golden and Robert Hasel-

82 SCIENTIFIC AMERICANJune 1988

© 1988 SCIENTIFIC AMERICAN, INC korn of the University of Chicago have some species, to form elaborate fruit­ they align themselves side by side and shown that differentiation in Anabae­ ing bodies. Movement is highly coordi­ either move off in the same direction na involves a form of controlled ge­ nated: when the population migrates or rub alongside each other before netic engineering. In the course of het­ over agar, it displaces itself as an separating. Daughter cells can be ob­ erocyst differentiation a specific DNA intact unit. When an individual cell served taking part in a similar routine rearrangement occurs that results in moves a few microns beyond the edge, following cell division. These bacteria the creation of a complete coding se­ it quickly pops back into place as literally keep in touch with each other! quence for one of the subunits of though drawn by an elastic thread. Predator-prey relationships exist in nitrogenase. Such a rearrangement oc­ Even those species of Myxobacteria the microbial world just as they do in curs only in cells that are differentiat­ that do enter dormancy as single-cell the world of larger organisms. Sever­ ing to form heterocysts. Comparable spores are social throughout much of al predatory species of Myxobacteria DNA rearrangements are involved in their life cycle. When two cells of Myx.­ feed by secreting enzymes that dis­ the formation of specialized immune­ ococcus virescens meet, for example, solve the outer cell layer of other mi­ system cells in vertebrates. they go through a characteristic ritual: croorganisms; when the cells burst, In addition there are submicroscop­ ic channels within each Anabaena fila­ ment that connect the two kinds of cells. The transport of cellular prod­ ucts (fixed nitrogen to the photosyn­ thetic cells and photosynthetic prod­ ucts to the heterocysts) takes place by way of these channels. In overall char­ acter, then, it can be said that Anabae­ na functions more like a multicellular organism than a unicellular one: it relies on division of labor among its cells to carry out specialized and in­ compatible chemical processes.

ore spectacular examples of multicellular behavior can be Mfound among the Myxobacte­ ria, the most morphologically complex of all bacteria. Their elaborate fruiting bodies rival those of fungi and slime molds and have long been an object of scientific curiosity. Myxobacteria are social creatures par excellence and their intriguing, almost psychedelic patterns of aggregation and move­ ment have been recorded in a fascinat­ ing series of time-lapse motion pic­ tures produced by Hans Reichenbach of the Society for Biotechnological Re­ search in Braunschweig and his collab­ orators at the Institute for Scientific Film (LW.F.) in Gbttingen. Unlike many bacteria, which periodi­ cally enter a dormant stage as individ­ ual spores, many Myxobacteria never exist as single cells. Instead they enter dormancy in the form of a multicel­ lular cyst that eventually germinates and spawns a ready-made popula­ tion of thou:sands of individuals. Each cyst founds a new population; as the bacteria become more numerous and dense, a number of sophisticated events speCific to multicellularity take place. Trails of extracellular slime are secreted and serve as highways for the directed movement of thousands of cells, rhythmic waves pulse through the entire population, streams of bac­ MULTICELLULAR FRUITING BODY of Chondromyces crocatus, a species in the group teria move to and from the center and Myxobacteria, is magnified268 diameters. The structure consists of a central stalk edges of a spreading colony, and bac­ that branches to form specialized clusters of single-cell spores. When the clus­ teria aggregate at speCific places with­ ters burst, the spores disperse to form new colonies. The micrograph was made by in the colony to construct cysts or, in Hans Reichenbach of the Society for Biotechnological Research in Braunschweig.

SCIENTIFIC AMERICANJune 1988 83

© 1988 SCIENTIFIC AMERICAN, INC ied the genetic basis of communica­ enzyme under study and the recombi­ tion and movement in M xanthus. By nant DNA is inserted into Pseudomo­ searching for mutants that have lost nas, then when the bacteria are grown the ability to spread or to form fruit­ on agar containing those chemicals, ing bodies, they have identifiedspecif­ the amount and distribution of col­ ic regions of the organism's DNA that or is a direct measure of the gene's control for aggregation, motility and expression. differentiation. They found that motil­ When I plated my recombinant ity in M xanthus is under the control strains of P. putida on such indicator of two different systems. The A (for agar, I was astonished to find that adventurous) system enables individ­ every colony exhibited a characteristic ual cells to move across the substrate; flowerlike pattern of staining. I repeat­ the S (for social) system controls the ed the experiment with engineered E. movement of groups of cells. If either coli and with strains of naturally pig­ system is defective, the cells are still mented bacteria, such as Pseudomo­ able to spread, but they do so abnor­ nas cepacia, Serratia marcescens and mally. If both systems are defective, Chromobacterium violaceum. Each the cells cannot spread at all. The two produced its own unique flowerlike systems are surprisingly complex and pattern. The fact that different strains a great deal of speCific genetic infor­ and species produced distinctive colo­ mation is required for each to be oper­ nies (including species that were natu­ ative. A mutation affecting any one of rally colored and not subjected to ge­ 23 different genetic loci will eliminate netic manipulation) gave me reason to the A motility system; at least 10 dif­ believe colony growth in bacteria is a ferent loci control the S system. highly regulated process and is under Kaiser and his colleagues have also some form of temporal control. Subse­ begun to unravel some of the ways M quent studies confirmedthat hypoth­ xanthus cells communicate with one esis. It is now clear to me that colony another. They accomplished this by organization follows certain general combining different mutant cells that rules, which help to explain the exis­ were defective for the same trait but tence of general patterns. in which the defect resulted from mu­ tations at different loci. They found he colonies tend to assume a that when two motility-defective mu­ circular configuration, growing ANABAENA, a photosynthetic cyano­ tants are combined in the same petri outward by adding cells to the bacterium, forms filaments of cells in T dish, they regain motility as long as perimeter. As a colony spreads across freshwater ponds_ Most of the cells are both mutants stay together. Similarly, the agar, it is apparent that the pattern photosynthetic, but when nitrogen lev­ if two sporulation-defective mutants of growth consists of both concentric els are low, heterocysts develop_ The are mixed in culture, they willform and radial elements. The concentric heterocysts, which are capable of ni­ normal fruiting bodies and sporulate. trogen fixation but not photosynthesis, elements are rings that encircle the are larger than the photosynthetic cells In some instances such mixed-cell colony; the radial elements, or sectors, and have special storage granules for complementation is now known to be look like slices of pie. Each sector nitrogen-rich compounds_ The Nomar­ mediated by the production of extra­ consists of outwardly growing proge­ ski micrograph was made by the author; cellular substances; in other cases it ny descended from a common ances­ the cells are enlarged 1,625 diameters_ may result either from direct cell-to­ tor. Some grow better than others and cell contact or from the physical pres­ expand, whereas other sectors merely ence of two complementary cell types. hold their own or even disappear as the myxobacteria absorb their con­ the colony gets bigger. tents. One species, Myxococcus xan­ y own studies of multicellular thus, has evolved a specialized meth­ behavior in bacteria began, as od of prey capture in response to its Mis often the case, as the result aquatic environment. In water it can­ of a chance observation. A little over not release its digestive enzymes be­ five years ago I was experimenting cause they would immediately be di­ with a genetic-engineering tool, de­ luted, as would the prey's nutrients. signed by Malcolm ]. Casdaban of Jeffrey c. Burnham and his colleagues the University of Chicago and his stu­ at the Medical College of Ohio have dents, in order to study enzyme ex­ shown that instead M xanthus con­ pression in Pseudomonas. The tech­ structs spherical colonies containing nique enabled me to join the genes for millions of bacteria. The colonies sur­ certain Pseudomonas putida enzymes round suitable prey organisms, trap­ with the DNA sequence from Esche­ ping them in pockets on the surface of richia coli that encodes the enzyme the sphere, where both the digestive beta-galactosidase. The advantage of enzymes and the prey contents can be beta-galactosidase as a genetic-engi­ effectively corralled. neering tool is that it causes certain FORMATION OF FRUITING BODY in the Over the past decade A Dale Kaiser chemicals to turn color when they are Myxobacteria follows a characteristic se­ of the Stanford University Medical exposed to it. If the beta-galactosidase quence of steps, as is shown from the School and his colleagues have stud- sequence is linked to the gene for the far left in these scanning electron micro-

84 SCIENTIFIC AMERICANJune 1988

© 1988 SCIENTIFIC AMERICAN, INC PREDATORY SPHERESare formed by millions of individual eventually digested within pockets on the sphere's surface. cells of Myxococcus xanthus as a means of capturing prey in an The spheres are enlarged 440 diameters in this scanning aquatic environment. Microscopic prey, such as the cyanobac· electron micrograph made by Jeffrey C. Burnham, Susan A terium Phormidium luridum (inset), stick to the colony and are Collart and Barbara W. Highison of the Medical College of Ohio.

In many cases it is possible to select N. Patrick Higgins of the University of The concentric elements in a colony individual cells from different sectors, Alabama Medical School and I found are less familiar than the sectors and grow them in culture and show that that differences in DNA between dis­ hence more puzzling. The cells in a the cells in one sectorial culture have tinct sectors and concentric zones can concentric zone or ring share a com­ heritable properties differing from sometimes be visualized directly by mon property (such as the level of those of the cells in another sectorial picking colonies up on filter paper, expression of beta-galactosidase ac­ culture; often cells from different sec­ extracting their DNA in situ and then tivity) but are not related by common tors can be distinguished on the basis applying radioactive probes to detect ancestry; they are directly related to of differences in their DNA Recently specific sequences. bacteria in the preceding and succeed-

graphs tracing the morphogenesis of Stigmatella aurantiaca. from the Indiana University campus, the stalk is elaborately At first cells form aggregation centers. As the centers accu­ branched (far right). The S. aurantiaca fruiting body is en­ mulate bacteria they bubble upward until the vertical stalk larged 450 diameters. Micrographs were made by Gabriela M. is complete. In some species, such as this unidentified one Vasquez, Frank Qualls and David White of Indiana University.

SCIENTIFIC AMERICANJune 1988 85

© 1988 SCIENTIFIC AMERICAN, INC ing zones, not to one another. If the organization of the colony into dis­ tinct concentric zones cannot be ex­ plained by heredity, how might it be explained? Some system must exist that bestows common properties on bacteria within a ring and distinguish­ es them from bacteria in other rings. One set of clues to the origins of concentric patterns lies in the differ­ ent ways the sectorial and concentric elements interact with one another. Photographs of colonies often show that concentric rings persist through sectors that grow faster than the rest of the colony. The resulting pat­ tern contains rings that are stretched outward. The stretching shows that the rings are formed not at specif­ ic positions on the agar (that is, at particular distances from the center) but at a specifictime in the course FLOWERUKE COLONIES can be produced by streaking bacterial cultures over an agar plate. Each colony (when it is not crowded) assumes a pattern characteristic of its of colony development. This sug­ strain. Chromobacterium violaceum (top left)naturally produces the pigment viola· gests bacteria have biological clocks cein and forms purple colonies. Serratia marcescens (top right) synthesizes prodigio· that enable them in some way to pro­ sin; it forms bright red colonies once thought to be drops of blood. Pseudomonas gram cellular differentiation at spe­ cepacia (bottom left)is yellow and has a unique surface texture, the result of cell ag­ cific times during development. The gregation at the surface. AnEscherichia coli colony (bottom right) carries genetically rhythmic pulsations of spreading nyx­ engineered DNA sequences encoding the enzyme beta·galactosidase; where the en­ obacterial colonies also reflect the zyme is expressed the colony turns blue. The colonies are at various magnifications. operation of a clocklike mechanism. Both biological clocks and the tempor­ al control of development, previously unknown in bacteria, are important features of higher organisms. Examination of the surface textures of a colony indicates that cellular dif· ferentiation also takes place at the level of cellular aggregation. When light is reflected from a colony, vari­ ous surface textures that are as organ­ ized as the pigmentation patterns and also show radial and concentric ele­ ments become visible. In many cases the two patterns coincide: a sector defined by color may also display a novel surface structure, and a ring may stand out on the basis of both its distinctive color and its topography.

learly colony organization in­ volves more than the distribu­ Ction of cells with simple bio­ chemical differences. In order to study structural patterns more preCisely, I turned to the scanning electron micro­ scope, which visualizes surfaces at high magnification with great depth of field. The scanning electron micro­ graphs revealed that colonies of P. putida and E coli are made up of E.COU colony was grown from a single drop of culture that contained thousands of highly differentiated cells that often cells. The highly regular, intricate pattern of pigmented rings is characteristic of form distinctive multicellular arrays certain genetically engineered E.coli colonies stained for beta·galactosidase activity. Pie·shaped sectors, within which the control of enzyme synthesis has changed, are coincident with the macroscopic or­ also apparent. The sectors at five o'clock and ten o'clock have curved edges, indio ganization of the colony. They also re­ cating that the bacteria within these regions spread faster than the rest of the col· vealed that each colony secretes extra· ony. The concentric rings that run through the ten o'clock sector are displaced out· cellular materials, some of which form ward, suggesting that changes in enzyme activity occurred at similar times both in· a skin or framework over its surface. side and outside the sector. This colony was approximately a centimeter in diameter. It was clear from these studies that

86 SCIENTIFIC AMERICANJune 1988

© 1988 SCIENTIFIC AMERICAN, INC biochemical activity within a bacterial colony is highly organized and spatial­ ly restricted: cells in different regions of the colony have different shapes and biochemical properties. In order to identify the unknown factors that control multicellular growth in bacte­ ria, I began to study swarm colonies. As the name implies, these are col­ onies that grow quickly and cover a large surface area, two characteristics that make them ideal for laboratory experimentation. Swarm behavior can be observed in many taxonomically distant species of bacteria. I have focused on one spe­ cies: Proteus mirabilis. Like the Greek god Proteus, this bacterium assumes different forms, producing striking configurations in a petri dish. Over the years two key features of Proteus colo­ nies have been noted. One is that there are at least two very different types of cells in a colony: long swarmer cells covered with hundreds of flagellae and short nonswarmer cells with few flagellae. The second feature is that colony development occurs as a tight­ ly programmed rhythmic process. Swarm colonies develop from an ini­ tial population of short nonswarmer cells. As the short cells divide, long SCANNING ELECTRON MICROGRAPH of a Pseudomonas putida colony reveals that swarmer cells begin to appear. They extracellular material covers its surface like a skin. This superstructure may in some migrate to the periphery of the colony, way facilitate communication among cells of the colony. One long, curved cell can where they assemble in groups and be seen bridging a crack in the skin; such cells may also be involved in intercel­ ' then pioneer the expansion of the col­ lular communication within the colony. The colony is enlarged 2,300 diameters. ony by moving out in a series of swirls. The flagellae that cover the surface of swarmer cells rotate and in so doing not spread indefinitely; they stop af­ somehow propel the cells. (It is easy to ter moving a set distance and start understand how the spinning flagellae spreading again only after a delay, propel a cell through liquid; how such which may be as long as a few hours, delicate structures are able to propel depending on conditions such as tem­ bacteria over the highly viscous agar perature and the composition of the surface, however, remains a total mys­ agar. Swarming is strictly a multicellu­ tery.) Observing swarm colonies under lar activity; an individual cell that gets a microscope, one sees thousands of separated from the rest is unable to flagellae on dozens of swarmer cells advance over the agar unless it is en­ moving in synchrony, creating oscil­ gulfed by another swarmer group, at lating waves as the cells spread out­ which time it starts to move again. ward from the periphery. Such exqUi­ Activity inside the edge has its own site coordination prompted Alexander periodicities and rhythms but is con­ Fleming, the discoverer of penicillin, nected to the expansion of the colony to wonder in print if Proteus could as a whole. When the swarmer cells possibly have a nervous system! finish one phase of spreading, for ex­ By recording Proteus morphogene­ ample, a series of visible waves com­ sis with a time-lapse video camera I posed of more densely aggregated have been able to identify distinct pe­ cells moves from inside the swarm riodicities in colony growth, confirm­ zone toward the perimeter. Then there ing the findings of earlier workers. I is a thickening of the cell mass in the discovered that intense activity within recently colonized zone and a bub­ LONG SWARMER CELLS can be seen at a swarm colony can be seen both at bling of the surface inside the perime­ the edge of a Proteus mirabilis colony. the advancing edge, where swarmer ter. These and other exquisitely cho­ The cells are preparing to move across cells are rapidly moving outward, and reographed postmigration processes the agar as a group; they do so by ro­ inside the edge, where cell division produce elaborate textures in the tating their flagellae in synchrony. The and streaming activities continue­ form of terraces on the surface of a cells are enlarged 600 diameters in this even when the swarmer cells have fully developed swarm colony. micrograph made by S. A. Sturdza of stopped advancing. Swarmer cells do Watching the swarming process, I the Cantacuzino Institute in Bucharest.

SCIENTIFIC AMERICANJune 1988 87

© 1988 SCIENTIFIC AMERICAN, INC wondered whether it would be possi­ ble to learn about the systems that control this intricate and regular be­ havior. In biological research impor­ tant clues can come from looking at the response of organisms to unusual circumstances. If conditions in the pe­ tri dish changed, would a swarm colo­ ny alter its behavior? Two kinds of evidence suggest that swarming behavior is indeed regulat­ ed. One is that when swarming is inter­ rupted by chemical or physical obsta­ cles, or by interference with the grow­ ing cell mass, specific morphological responses take place. Observation of PROTEUS MIRABillSmutants show geometries that provide clues to the way colony the geometries of the swarm colonies spreading is controlled. The mutant (left)forms regular repeating terraces in a after they have been artificially manip­ nine-centimeter petri dish. If a trench is cut in the agar, however, spreading stops ulated (or following spontaneous acci­ after a few cycles. The fact that swarming is blocked in the shadowed zone indicates dents that deform the regular outlines that a chemical signal must emanate from the center. Morphogenesis appears to be under whole-colony control; it is not regulated solely by the migrating edge. of the colonies) makes it clear that The mutant (right) has lost its circular symmetry and has a markedly different pat­ both temporal and directional con­ tern. The bacteria began by forming thick columns along stress lines in the agar; trols influence colony growth. In par­ after these columns grew to a certain size they ramified into smaller perpendic­ ticular, chemicals diffusingthrough ular processes that in turn formed smaller branches. This pattern suggests that al­ the agar appear to play important though the mutant lacked the ability to produce circular colonies, it retained some roles in guiding colony spread. kind of directional control. The length of this colony was about four centimeters. The second line of evidence is that morphogenesis is under hereditary control. For one thing, each naturally isolated strain of P.mirabilis has its own characteristic mode of swarm­ colony development. For another, one can obtain from these natural isolates various mutants that can still spread but that have geometries markedly different from those of their progen­ itors. Some mutants form periodic swarm terraces that are more closely spaced than those of their parents, whereas others have no terraces at all. One particularly striking mutant has lost its circular symmetry altogether and grows in a branching pattern in­ fluenced by stress lines in the agar. Clearly the genes that regulate mor­ phogenesis have been altered in these mutants, whereas the ability to swarm has not been eliminated. A detailed biochemical explanation for how mor­ phogenetic control systems might op­ erate in Proteus (or in Myx.ococcus or E.coli, for that matter) has not yet

DISTINCTIVE MORPHOLOGY of cells in different zones is demonstrated in a se­ ries of scanning electron micrographs of an E.coli colony. The 68-hour colony (top), formed from an initial population of 100,000 cells, was five millimeters in diameter. When the leading edge is en­ larged 100 diameters (bottom left), a dis­ tinct boundary is visible between two groups of differentiated cells. At great­ er magnification (750 diameters) the out­ ermost zone (bottom right) is seen to be made up of large cells arranged irregu­ larly, whereas the inner zone has smaller cells grouped in roughly parallel arrays.

88 SCIENTIFIC AMERICANJune 1988

© 1988 SCIENTIFIC AMERICAN, INC PURPOSEFUL MOVEMENT of M.xanthus cells is shown in a the flare turns toward the bead at 18 minutes (second frame); series of frames on a video monitor. At the bottom of the at 33 minutes (fourth frame) the flare reaches it. After sensing screen is a latex bead, fivemicrometers in diameter, toward that the bead is inedible the flare will move on. These images which the cells, collectively called a flare, will orient. The tip of were made by Martin Dworkin of the University of Minnesota.

been found. Nevertheless, the hered­ may even be possible to introduce much more is known about the intri­ itary specificity of the developmen­ multicellularity characteristics that op­ cacy and complexity of this group tal phenomena suggest that studying timize productivity or improve the ca­ of organisms. If, as I have proposed these systems will yield important les­ pacity of specific strains to degrade here, bacteria possess elaborate devel­ sons about coordinating the behaviors synthetic compounds. Understanding opmental and behavioral capabilities of large numbers of bacteria in a spa­ the behavior of bacteria may also typical of higher organisms, then it is tially and temporally defined way. make it easier to monitor the release likely that detailed explanations of The view that bacteria are sentient of genetically engineered organisms. how these small cells communicate creatures, able to receive, process and In medicine greater understanding will influence views of information respond meaningfully to external sig­ of bacterial behavior may lead to in­ processing in all organisms. nals, has been gaining ground over creased efficacy of drug treatments. Although bacteria are tiny, they the past two decades as investigators j. William F. Costerton and his col­ display biochemical, structural and spend more time exploring the mys­ leagues at the University of Calgary in behavioral complexities that outstrip teries of bacterial behavior. Alberta recently described a patient scientific description. In keeping with Martin Dworkin of the University of who suffered from recurrent blood­ the current microelectronics revolu­ Minnesota has recently provided a stream infections because a colony of tion, it may make more sense to graphic demonstration of multicel­ Staphylococcus aureus had formed on equate their small size with sophis­ lular responsiveness in the microbi­ the lead of his cardiac pacemaker. In­ tication rather than with simplicity. al predator Myxococcus xanthus. He dividual cells would break away from There is little reason to doubt that found that roaming flares, or groups, the parent colony periodically and in­ insights gained from studying the in­ of M.xanthus cells perceive clumps of fect his bloodstream. Although the in­ teractions of billions of bacteria, living prey bacteria (or even glass or plastic dividual cells were sensitive to penicil­ together in a volume of less than a few beads) on an agar surface, make sharp lin, the colony itself was drug-resis­ cubic millimeters, will enhance under­ turns toward the objects and then tant like many bacterial colonies, it standing of all forms of life. move directly to them. Once there, was protected by a coating of extracel­ the Myxococcus flares are able to tell lular slime. In order to end the chronic FURTHER READING whether the objects are edible or not. infection the only solution was to re­ NATURE Of THE SWARMING PHENOMENON If the objects are edible, the flares stay move the pacemaker (and its colony). IN PROTEUS. Fred D. Williams and Rob­ to feed; if they are not, the flares turn In other cases of bacterial pathogen­ ert H. Schwartzhoff in Annual Review of 32, 101-122; away and continue their searching be­ esis there is a clear correlation be­ Microbiology, Vol. pages 1978. havior. Such purposeful behavior has tween the tendency of cells to aggre­ MYXOBACTERIA: DEVELOPMENT AND CELL traditionally been thought to operate gate and their ability to establish an INTERACTIONS. Edited by Eugene Rosen­ 25 only in larger organisms. infection. It has been known for berg. Springer-Verlag, New York, 1984. years that the organism that caus­ ORGANIZATION OF DEVELOPING ESCHE­ hat practical value, if any, do es gonorrhea, Neisseria gonorrhoeae, RICHIA COLI COLONIES VIEWED BY SCAN­ these findings have? The bio­ forms several types of colonies on NING ELECTRON MICROSCOPY. James A technology industry, eager to laboratory media. Cells from one type Shapiro in Journal of Bacteriology, Vol. W 169, No. 1, pages 142- 156; January, turn to genetically engineered bacteria are virulent, whereas those from an­ 1987. as factories for the production of other are not. Since a successful infec­ The following films are available in the complex biochemicals, will undoubt­ tion requires that the disease organ­ U.S. from Audio-Visual Services, Penn­ edly benefit from the knowledge that ism colonize its host, I suspect path­ sylvania State University: bacterial cells specialize and control ogenesis is directly related to this BACILLUS CIRCULANS-AUFBAU UND VER­ protein synthesis with the aid of inter­ organism's ability to form multicellu­ HALTEN (B. CIRCULANS-GROWTHAND BEHAVIOR). El83. cellular communication signals. In the lar organizations (as reflected in the Silent film, Institut field of biodegradation the application kinds of colonies it builds). fUr den Wissenschaftlichen Film. PROTEUS-BEWEGUNGSVERHALTEN (PRO­ of bacteria to remove toxic chemicals Bacteriology's early pioneers, such TEUS SWARMING BEHAVIOR). Silent film, from polluted soils and water sources as Louis Pasteur, recognized that there E27 1. Institut ftirden Wissenschaftli­ may be enhanced by greater under­ are many lessons to be learned from chen Film. standing of multicellular processes. It these smallest of living cells. Today

SCIENTIFIC AMERICANJune 1988 89

© 1988 SCIENTIFIC AMERICAN, INC