by Stephen T. Abedon

(for a more easily printed version, click here)

[Note: Do not reference this manuscript. Instead, please see the Phage Ecology chapter by Stephen T. Abedon as it is to appear in The Bacteriophages, Edition 2, Oxford University Press. The manuscript submission deadline for that publication is February 1, 2002. This manuscript, as presented, is not in final submission form. Any comments and criticisms on this manuscript sent prior to the submission deadline would be greatly appreciated by the author. Please send by e-mail to abedon.1@osu.edu. Thanks.]

Contents

1. Introduction 2. Phage Organismal Ecology a. The basic phage life cycle b. Phage adsorption c. Infection (the latent period) d. Phage-progeny release e. Phage decay 3. Phage Population Ecology a. Phage latent-period evolution b. Contribution of early adsorbers c. Phage plaque growth 4. Phage Community Ecology a. Community stability b. Refuges c. Slowed adsorption d. Reduced phage productivity e. Synthesis 5. Phage Ecosystem Ecology 6. References

Introduction

Phage ecology is the study of the real-time interactions between phages and environments. These interactions are ecologically important, particularly to the extent that they affect bacteria populations. Here - keeping phages in the fore rather than bacteria or ecosystem functioning - I consider phage organismal, population, community, and ecosystem ecology (Table 1). For complementary approaches to the review of phage ecology see (6,7,11,18,23,39,47,51), plus various recent reviews of aquatic and ecosystem phage ecology (38,53,57,63,64,69,73).

Table 1: Defining Phage Ecology

Ecology	A bacterium is...	Considerations	Experiments
Organismal	...a target, or an entity that impacts on the phage phenotype	Phage anatomy, physiology, and behavior characterized from Darwinian perspective; virion stability, survival, and adsorption; eclipse period, latent period, and burst size; adaptations overcoming barriers to transmission	Single-step growth; adsorption curves; kinetics of phage decay
Population	...an environmental resource	Phage population growth and density; liquid versus spatially structured environments (broth growth versus plaque growth); low versus high phage multiplicity	Single-step growth; adsorption curves; kinetics of phage decay
Community	...a partner in coevolution	Phage-host coevolution; inverse relationship between phage and uninfected-host density; community stability; host resistance; phage host-range breadth and variation; transduction and phage (lysogenic) conversion; competition among different phage species	Phage-host continuous-culture or serial-transfer experiments; in situ observation and experiment
Ecosystem	...a lower trophic level	Phage impact on ecosystem nutrient cycling and energy flow; short circuiting of microbial loop	In situ observation and experiment

Phage Organismal Ecology

The basic phage life cycle. In the hierarchy of ecological disciplines, phage organismal ecology is most closely allied with the molecular (plus physiological and genetic) characterization of phages. While underlain by copious variety and details, the general phage life cycle (Figure 1) basically involves adsorption, infection, and release, plus considerations of phage decay. The study of these processes - especially from the perspective of in situ costs, benefits, expression, and per-infection productivity - is the province of the phage organismal ecologist. More broadly, one can view virus organismal ecology as the study of the adaptations viruses employ to overcome physical, chemical, or biological barriers to their transmission between hosts (44).

Figure 1: General Phage Life Cycle (below).

Phage adsorption. Phage adsorption begins after phage release from infected cells and ends with the uptake of phage genomes into the cytoplasm of adsorption-sensitive hosts. The more rapidly a phage adsorbs a permissive host cell, the greater its likelihood that it will avoid decay (e.g., 19,36,56) and the shorter its overall life cycle (5). Nevertheless, phage mutants displaying faster adsorption than wild type have been isolated from laboratory cultures (30). In addition, by requiring specific adsorption cofactors, some phages, such as phage T4, may be adsorption competent within environments in which healthy hosts are likely (e.g., colons) but adsorption incompetent (or less competent) in environments where healthy hosts are less likely (e.g., sewage) (28,47).

Infection (the latent period). The phage latent period begins with the eclipse, a period during which the artificial lysis of an infected host will not release infective phage particles. Post eclipse of a typical infection I refer to as a period of phage-progeny maturation. During this latter period either infected bacteria release mature phages without lysis (46) or artificial lysis results in the release of phage progeny (35). Maturation, for highly virulent phages, mostly occurs at a constant, linear rate rather than exponentially because the rate of synthesis of virion components is limited by some aspect of the host's anatomy or physiology (40,68). Things are complicated, however, if cells are able to continue to grow and divide during a phage infection since cell growth can increase in number whatever cell components are limiting. At the same time, the synthesis of phage components can have negative effects on host division. These negative effects range from a slowing of host population growth as seen with filamentous phages (46) to a complete cessation of host division as seen with highly virulent phages such as phage T4 (e.g., 2).

A further complication is the length of the period of progeny maturation. With lytic phages the timing of host lysis controls the length of this period, with the timing of lysis, in turn, under the control of phage genes and proteins (e.g., holins; 67). Host nutrition status (40,64,68), temperature (32), and physiological state vis-à-vis the standard bacterial growth curve or chemostat doubling times (7,55,61) can also impact on this timing as can phage-controlled processes such as lysis inhibition and lysis from without (1,2,3,4). In addition, among even synchronously infected cells, it has long been known that the overall duration of the phage latent period can vary within a single culture (13). Consequently, most phages during single-step growth display a non-instantaneous rise, which is the time over which a population of synchronously infected hosts display phage-induced lyses (32,61). Progeny maturation and rise periods associated with chronic (a.k.a., continuous or persistent) infections can be particularly long (46).

Phage-progeny release. So long as a virus particle remains inside an infected bacterium, then it is not free to acquire a new host. An infected host may display significant productivity in terms of the intracellular maturation of progeny virions, but such productivity can pale in comparison with the growth rates that phage populations may achieve via the exponential growth that phage-progeny release makes possible. For most phages the release of progeny phages coincides with the destruction of the parental infected cell (lysis; 67). For filamentous phages, that extrude their phage progeny across the host cell envelope, release does not necessarily result in host-cell death (46).

The phage progeny released from an individual bacterium are collectively referred to as a burst. For phage populations one typically determines a parameter know as burst size that is equal to the total number of phages produced by a single round of phage infection of host cells, divided by the total number of cells that had been infected prior to phage-progeny release. Measured burst size can vary considerably between individual infected cells (33) and even over the course of released-phage storage (20). Burst-size determination is complicated if hosts aggregate or fail to fully separate in culture (12) or if phage release occurs via a mechanism other than host lysis (7).

Phage decay. If one is willing to accept that phages are alive (e.g., 24), then phage decay is equivalent to virion death (a.k.a., "inactivation" or "loss of titer"; 36) or to a lack of infected-cell productivity without reduction to lysogeny. Phage decay (recently reviewed by 64,73) likely limits the impact of phages on bacteria (65) plus imposes important constraints on the evolution of non-temperate phages since it implies that virion populations cannot survive indefinitely in the absence of sufficient densities of susceptible bacteria (e.g., 27,71). Similarly, the evolution of lysogeny must be dependent at least in part on the relative importance of virion decay versus phage and prophage replication rates as, for example, Stewart and Levin (62) suggest with their "hard times" hypothesis.

Phage Population Ecology

Phage population growth. While phage organismal ecology emphasizes per-infection-productivity and phage community ecology has a host-population-dynamics emphasis, the emphasis of phage population ecology (Table 1) is on phage population growth either within bacteria cultures (5) or within individual infected bacteria (25). Like any organism living within a suitable environment possessing sufficient resources, a phage population will increase in number exponentially over time. Phage exponential growth is especially tractable during phage growth within liquid culture (2,21,32). Phage populations that increase in size most quickly should acquire host cells most rapidly. The acquisition (exploitation) of one bacterium by one phage means that one-less unit of bacteria resource is available for exploitation by a second phage. Over the short term, in relatively simple environments, selection within phage populations therefore should be for both more-rapid population growth and more-rapid host-cell acquisition.

Phage latent-period evolution. Certain phage characteristics should contribute to faster phage population growth (e.g., 31). For instance, we should expect evolution to favor decay-resistant virions, rapid adsorption (though, during plaque growth, not necessarily; see below), short eclipse periods (except given selection for pseudolysogeny or true lysogeny), high rates of progeny maturation (balanced, in some cases, e.g., with filamentous phages, against damage to the host resource), and, once initiated, rapid progeny release (ditto). With or without caveats, conspicuously absent from this list is the duration of phage latent periods and periods of progeny maturation, with the length of both a function of lysis timing. Here I consider forces that impact the evolutionary optimization of the duration of phage latent periods.

From an ecological perspective we can distinguish the members of populations into three groups: prereproductive, reproduction, and postreproductive. Postreproductive phages, variously defined, are irrelevant to this discussion. Prereproductive phages are those engaged in either adsorption (including the extracellular search for susceptible bacteria) or the eclipse, since during these periods the phage is not generating mature phage progeny. Reproductive phages are those infecting bacteria during the phage period of progeny maturation. For phages that must lyse their host bacteria to disseminate phage progeny, we may describe a period of progeny maturation as optimal in duration should the latent period giving rise to it result in maximized phage-population growth rates. Too-short latent periods result in insufficient burst sizes to sustain maximal phage population growth while too-long latent periods slow phage population growth by delaying phage-progeny acquisition of new host bacteria.

When prereproductive periods are short, this means that free phages can rapidly find uninfected cells and then rapidly gear up for intracellular progeny maturation. Such conditions should select for rapid infection turnover (short latent periods) such that phage progeny acquire uninfected hosts before those cells are obtained by competing phages. In general then, high host densities and short phage eclipse periods should select for shorter phage latent periods (5). When prereproductive periods are long, by contrast, the reproductive period, once begun, is more valuable thereby resulting in selection for increased per-infection productivity. Thus, low host densities or long phage eclipse periods should select for larger phage burst sizes even at the expense of longer phage latent periods (5).

Contribution of early adsorbers. The impact of changes in host density on phage population growth and latent-period evolution are not as straightforward as one might expect given that a phage cohort's mean time until bacteria adsorption varies directly with host density (Figure 2B). The reason for this complication is a consequence of phage adsorption occurring essentially as an exponential decay in free-phage density (Figure 2A). For any phage cohort released at a given moment into a population of hosts, phage adsorption occurs such that some constant fraction of remaining free phages will adsorb over any given interval. As a consequence, more phages from a given cohort will adsorb during a sooner interval compared with some later interval.

Figure 2A: Exponential phage adsorption and phage population growth (below). Free-phage adsorption (e.g., 5) with log(N = per ml host density) indicated for different curves and k (the phage adsorption constant) = 2.5 x 10^-9 ml/min. Adsorption curves cross the horizontal line at the average phage adsorption time (mean free time) = 1/kN. [Goto 2A, 2B, 2C, 2D, or all four]

Figure 2B: Exponential phage adsorption and phage population growth (below). Mean free time graphed as a function of bacteria density. [Goto 2A, 2B, 2C, 2D, or all four]

If by chance a phage adsorbs to a host earlier rather than later, then the duration that this phage is prereproductive will be shorter and therefore the total duration of that phage's life cycle will also be shorter. The rate of phage population growth is a function of the duration of the phage life cycle, as well as the per-host burst size. Furthermore, earlier-adsorbing phages are potentially greater in number due to the exponential kinetics of phage adsorption plus spend less time susceptible to non-adsorption-related virion decay (above). Consequently, it stands to reason that earlier-adsorbing members of phage-adsorption cohorts will contribute more to the exponential growth of a phage population than later-adsorbing members.

At greater host densities all phages adsorb relatively rapidly such that the variance in phage pre-reproduction duration is not large. However, at lower host densities the timing of the adsorption of the majority of a phage cohort is spread over much longer intervals (Figure 2A), and the contribution of those phages that by chance adsorb hosts earlier becomes increasingly large and important to overall phage population growth. Thus, the average timing of phage adsorption (the phage mean free time; see 5) may very well decline as a direct function of host density (Figure 2B), but phage population growth as a function of host density does not decline as quickly (Figure 2C). This means that while evolution ought to favor phages with longer periods of progeny maturation as host densities decline, the phage latent period that is optimal for phage-population exponential growth should not increase as fast as host densities decline (5; Figure 2D).

Figure 2C: Exponential phage adsorption and phage population growth (below). Log phage density following 1000-min of phage growth as simulated or calculated at different host densities (log-scale and log-transformation are both intentional as presented). A latent period of 25 min, burst size of 75 phages/cell, and adsorption constant as above were used. Simulations assumed exponential phage adsorption (circles) which is equivalent to the adsorption curves in panel A. For calculations it was assumed that individual free phages adsorb after an extracellular search of 1/kN min (squares) as calculated as a function of host density in Figure 2B. [Goto 2A, 2B, 2C, 2D, or all four]

Figure 2D: Exponential phage adsorption and phage population growth (below). Phage latent period that gives rise to maximal phage population growth determined using simulations (adsorption via exponential free-phage decline; circles) and calculations (adsorption for all of free-phage cohort is mean free time in duration; squares) as described for panel C (graph used with permission from ASM). See (5) for discussion of methods. [Goto 2A, 2B, 2C, 2D, or all four]

We would expect similar compromises to hold for phages that release their progeny via extrusion. For such phages, however, the important balance should be between (i) the kinetics of phage maturation and release, (ii) the impact of greater rates of phage release on infected-host replication, and (iii) the overall latent-period duration. For experiments addressing these issues for filamentous phages see (46) and for lytic phages see Abedon (in preparation).

Phage plaque growth. Phage growth may be observed within a simple, spatially structured environment as plaques punctuating an otherwise opaque bacteria lawn embedded within a soft-agar overlay. Phage growth in plaques may be considered to occur in four stages (45): (i) initial adsorption of seeded phages, (ii) initial round of infection, (iii) an "enlargement phase" which involves multiple rounds of adsorption, infection, and release, and (iv) the end of the enlargement phase which typically is associated with physiological changes in the bacteria lawn. Differences between phage growth in plaques versus broth occur throughout the enlargement phase during which the physical structure of solid media (i) slows both phage and host diffusion, (ii) prevents gross environmental mixing, and (iii) probably gives rise to local phage multiplicities that are much higher than one observes over the majority of phage growth in broth. Phage growth within plaques additionally introduces plaque size as a means by which issues of phage fitness may be addressed (e.g., 49,50).

We can imagine at least five selective pressures that act on phages during plaque growth: (i) At the periphery of plaques there should be selection for more-rapid exponential growth, e.g., short phage latent periods when host densities are high (above); (ii) regardless of location within a plaque, during the plaque enlargement phase there should be selection for fast diffusion away from the plaque center such that uninfected hosts surrounding the plaque may be obtained and exploited (essentially the same argument as many suggested explained the classic observation that smaller phages should make larger plaques than larger phages; e.g., 72); (iii) towards the center of plaques - where there is a low prevalence of uninfected hosts - there should also exist a countering selection for greater burst sizes even at the expense of longer latent periods; (iv) throughout the plaque there should be selection exerted by the tendency of phages to decay (48) including by processes of adsorption to cell debris or adsorption to infected cells (the latter due to superinfection exclusion; 3); and (v) there can be selection for maintenance of phage growth despite the physiological aging of the bacterial lawn (e.g., phage T7; 50). Given this myriad complexity, how, where, and when one determines phage fitness during plaque growth is extremely important since different plaque regions may be under different selective pressures that can vary over the course of plaque development.

As a further complication, plaque size does not necessarily correlate with per-infection productivity. It has been hypothesized, for instance, that phages displaying shorter latent periods, even given smaller burst sizes, could display larger plaques (45,74). Longer latent periods resulting in smaller plaque sizes are most commonly (and classically) observed among T-even phages where lysis-inhibition defective (r) mutants display larger plaques and conditionally shorter latent periods than lysis-inhibition competent wild-type phages (34,43). I have also observed larger plaques with phage RB69 (also T-even-like; 8) that appear to be a consequence of reductions in phage latent periods (and burst size) rather than due to changes in phage adsorption rates or other increases in per-infection productivity (Abedon, in preparation). It has additionally been hypothesized (45,74) that reducing host-attachment efficiency given phage-host collision can increase rates of plaque enlargement since with slower adsorption phages might spend less time infecting cells and more time diffusing towards the periphery of plaques. Indeed, one explanation for why phage l lost its tail fiber upon domestication (42) is that reduced adsorption efficiency resulted in the formation of inescapably selectable larger plaques. Sarma and Kuar (60) observed perhaps similar results with cyanophage N-1.

Phage community ecology

Community stability. Phage community ecology (Table 1) emphasizes the bacteria host, e.g., the impact of phages on bacteria densities and the evolution of phage resistance (10,16). Phage community ecology also considers phage-host coevolution, such as the propensity for phages to evolve strategies that counter mechanisms of host resistance. Bacteria evolution of phage resistance can contribute to the stability of phage-containing communities by impeding bacteria extinction. Stability additionally refers to the range in densities of host and phage populations as they oscillate over time, with greater oscillation amplitude (density variance) corresponding to lower community stability.

Phage community stability in the laboratory typically is studied within continuous phage-bacteria cocultures that are commonly, though when phages are present not necessarily correctly (58) referred to as chemostats. A chemostat possesses a reservoir containing sterile media connected to a well-mixed growth vessel containing microorganisms. Flow from reservoir to growth vessel may be controlled via the use of a peristaltic pump, with outflow from the growth vessel occurring at the same rate as inflow. Phage-host communities within chemostats often are more stable than may be accounted for by phage community ecology theory (16,61). In Figure 3A I present a simulation of a relatively unstable chemostat. Note that phages have driven phage-sensitive bacteria to extinction (<10^-2/ml) after about 110 hours of chemostat progression and that, due to outflow from the chemostat growth chamber, phages then decline to extinction about 100 hours later.

Figure 3A: Computer-simulated chemostats (below). Chemostats were simulated employing the method and parameter values of Bohannan and Lenski (14). Time steps here are 1 min rather than 3 min; the initial host and phage densities are 10⁴/ml and 10⁵/ml, respectively; and unless otherwise noted (in subsequent simulations) the limiting nutrient is glucose which is found in the chemostat reservoir at a density of 0.5 mg/liter. Bacteria are presented as solid lines and phages as dotted lines. Phage and bacteria densities during simulations were sampled for inclusion in graphs once every 30 min. These simulated chemostats contain no phage-resistant bacteria or other bacteria refuges from phage attack. Extinction is assumed to occur at or below densities of 10^-2/ml. Bacteria are presented as solid lines and phages as dotted lines. [Goto 3A, 3B, 3C, 3D, or all four]

Refuges. Levin et al. (52) speculated that refuges for sensitive bacteria away from phage attack could increase the stability of phage-host communities, as subsequent experiments have corroborated (61). In such a scheme the extinction of sensitive bacteria is prevented by their hiding, for example, within chemostat wall populations. Following phage-induced lysis of host populations, presumably only those sensitive hosts survive that remain in hiding. Through cell division, these hosts can supply sensitive hosts to the liquid (unrefuged) phase of the chemostat. Once phage populations have declined, due to their outflow from the chemostat, the liquid-phase host populations can grow back to higher population densities.

Slowed adsorption. Bohannan and Lenski (15) describe bacteria that have entered a "genetic" refuge (phage-resistant mutants) as "invulnerable prey". However, since unless a phage's collision with a bacterium results in some degree of phage-host attachment, then a resistant bacterium is not potential prey but instead some relatively inert component of the environment off of which phages "bounce." Wilkinson (70), on the other hand, has suggested a model in which completely resistant bacteria really are invulnerable prey. Here the assumption is that the "predator" species (in this case a Bdellovibrio) may reversible interact with non-prey bacteria by pausing following collision. This delay in detachment extends the bdellovibrio's extracellular search. From the perspective of susceptible bacteria, this delay is equivalent to a reduction in the effective predator density. Wilkinson's conclusion upon modeling such a system is that the presence of non-prey bacteria, even in the absence of metabolic competition with prey bacteria, will result in a stabilization of sensitive-bacteria population densities.

Reductions in phage adsorption rates could similarly result in increased community stability. A partial reduction in host reception to phage adsorption (e.g., T2-partially resistant bacteria, a.k.a.,"less vulnerable" bacteria), for instance, should contribute to an increase in community stability by delaying overall of phage-population attachment to sensitive bacteria (17). Consistently, Figure 3B presents a simulated chemostat for which the phage rate of adsorption has been reduced by one half, and bacteria and phage extinctions are thereby avoided. Again with T2 phages, there apparently is a tendency for these phages to be temporarily adsorption-inhibited (up to weeks at room temperature) following release from host cells (59). This phenomenon could also serve to increase community stability by delaying phage adsorption. Indeed, any host refuge from phage adsorption should reduce phage population productivity by reducing the effective host density (a "numerical" host refuge; 26,61), even if direct physical interactions between refuged (but otherwise phage-sensitive) hosts and free phages are nonexistant. By extension, mechanisms of phage decay, including outflow from chemostat growth chambers, should have the effect of reducing phage number, thereby increasing community stability. Furthermore, phage evolution that counters mechanisms that interfere with phage adsorption or decay should result in a decrease in community stability

Figure 3B: Computer-simulated chemostats (below). Figure shows the same chemostat as Figure 3A though with the phage adsorption constant reduced by one half. [Goto 3A, 3B, 3C, 3D, or all four]

Reduced phage productivity. Host density impacts on community stability by affecting the peak phage densities that follow community-wide host lysis. With more phages than hosts within a batch-culture system, eventually all sensitive bacteria may become adsorbed and lysed (11,26). However, in continuous culture there will be decay in free-phage densities due to outflow from the growth vessel. Consequently, since the rate that host cells are found by free phages is a function of free-phage density (1), there is a race between phage sensitive bacteria survival and free-phage outflow. The lower the peak phage density, the less the bacteria population will be reduced in size due to phage adsorption, and the greater the likelihood that phage adsorption will not reduce the bacteria population to the point of extinction (compare, for example, the peak phage densities in Figure 3A versus Figures 3C and 3D). The smaller the bacteria population available for infection within a chemostat, in turn, the lower the peak phage density (compare Figures 3A and 3C). Bohannan and Lenski (14) demonstrate this point by reducing the bacteria growth potential, through restrictions in the density of a limiting nutrient (glucose) within the nutrient reservoir, and then observing an increase in phage-bacteria community stability. See Figure 3C for a simulated chemostat in which the nutrient density in the chemostat nutrient reservoir has been reduced by one half and note again that the extinction of bacteria and phages is avoided.

Figure 3C: Computer-simulated chemostats (below). Figure shows the same chemostat as Figure 3A but with one-half as much limiting glucose. [Goto 3A, 3B, 3C, 3D, or all four]

The productivity of phage infections, along with their density, together determines peak phage densities. It is well known that phage growth parameters, such as burst size, can vary as a function of host physiology (discussed above). If the stability of chemostats is an inverse function of peak phage density (i.e., more phages = less stability), then reduced infection productivity given reduced nutrient availability should contribute to an increase in community stability (as I will demonstrate elsewhere with chemostat simulations; Abedon, in preparation). Similarly, we might expect that the T-even-phage lysis-inhibition phenotype (1,3) would be destabilizing since it contributes, particularly at higher host densities, to a larger phage burst size. In Figure 3D the impact on community stability of reducing the phage burst size by one-half is explored, with bacteria and phage extinction yet again avoided.

Figure 3D: Computer-simulated chemostats (below). Figure shows the same chemostat as Figure 3A except with the phage burst size reduced by one half. [Goto 3A, 3B, 3C, 3D, or all four]

Synthesis. It is highly likely that phage-host community stability arises from two relatively simple forces: (i) If sensitive hosts cannot be driven to extinction by even excess phage densities, e.g., as is at least approximated with host refuges from phage attack, then sensitive hosts simply will not be driven to extinction by phages. (ii) If sensitive hosts can be driven to extinction given sufficient phage densities, then hosts will be driven to extinction only if sufficient phage densities are present within an ecosystem. There are two corollaries to the second point: (a) At peak phage density, the fewer phages found within an ecosystem, the smaller the negative impact those phages will have on phage-susceptible bacteria populations and the more stable the system (compare Figure 3A with Figures 3C and 3D). (b) Mechanisms that interfere with a phage's attainment of higher peak densities (or with phage impact on individual bacteria) - e.g., more phage decay, more-rapid outflow, partial inhibition of phage adsorption, reduced phage burst sizes - may lead to an increase in the stability of a phage-bacteria community (ditto, plus Figure 3B). Indeed, as noted by E.S. Anderson in 1957 (p. 205) (11), "It is evident that suboptimal conditions for growth of the host cells may restrict phage multiplication in any environment, even when contacts between the virus and its host occur... Anything which restricts the phage titre limits the selective action of phage."

Phage ecosystem ecology

Phage ecosystem ecology (Table 1) encompasses the biotic as well as the abiotic world, in particular the biogeochemical cycling of nutrients and the flow of energy between and through ecosystems, usually with an aquatic emphasis (for recent reviews see 38,53,57,63,64,69,73). Bacteria consume, produce, and store nutrients and energy plus contribute to the decomposition of other organisms. Phage infections contribute to a solubilization of bacteria cells, whether following host-cell lysis or via the conversion of host components into virion particles. Solubilized bacteria, in addition to no longer functioning as consumers, producers, or decomposers, are also less available as food to bacteria grazers (protists or animals) that obtain their nutrients and energy through the ingestion or engulfment of intact bacteria. Since bacteria are the chief consumers of the water-solubilized components of especially aquatic ecosystems, a major consequence of the phage-induced lysis of bacteria is not just a reduction in the productivity of bacteria populations but also a delay in the movement up food chains of bacteria-contained nutrients and energy. Suttle (64) additionally argues that aspects of virus-induced cell lysis in aquatic environments can have significant positive and negative impacts on the abundance of various greenhouse gasses found within Earth's atmosphere.

In their requirement for intact bacteria, phages in a sense are competitors of the bacteria grazers. Due to the host-range constraints observed among all parasites, individual phages also tend to be more specialized than most grazers in terms of what bacteria within a community they may affect (e.g., 37,66). In addition, phages can make direct, positive contributions to the fitness of bacteria hosts through phage conversion or via the transduction of genes from other bacteria. Phage DNA and protein coats, following abortive infection, could even serve as a bacteria nutrient (38).

Phage ecosystem ecology also represents an elaboration on the various issues of organismal, population, and community ecology already discussed. It follows, therefore, that many or all of the complications, caveats, and considerations discussed throughout this review also affect our understanding of the phage impact on ecosystem nutrient cycling and energy flow. In addition, much of the impact of phages on ecosystems has been discerned from the study of aquatic phage biology. However, aquatic systems - since for the most part they are liquid rather than solid, can be moderately well mixed, and also can be quite large - are among the very simplest phage-containing ecosystems. We might expect that other ecosystems, for example soils (22) or biofilms (41), would display greater complexity in terms of the phage impact on nutrient cycling and energy flow. Thus, both literally and figuratively, our understanding of the impact of phages on real ecosystems has barely scratched the surface of phage ecosystem ecology's ultimate goal: Quantifying the impact of phages on nutrient cycling and energy flow throughout the biosphere.

References

1. Abedon, S.T. 1990. Selection for lysis inhibition in bacteriophage. Journal of Theoretical Biology 146:501-511.
2. Abedon, S.T. 1992. Lysis of lysis inhibited bacteriophage T4 infected cells. Journal of Bacteriology 174:8073-8080.
3. Abedon, S.T. 1994. Lysis and the interaction between free phages and infected cells, p. 397-405. In J.D. Karam (ed.), The Molecular Biology of Bacteriophage T4. ASM Press, Washington, DC.
4. Abedon, S.T. 1999. Bacteriophage T4 resistance to lysis-inhibition collapse. Genetical Research 74:1-11.
5. Abedon, S.T., T.D. Herschler, and D. Stopar. 2001. Bacteriophage latent-period evolution as a response to resource availability. Applied and Environmental Microbiology 67:4233-4241.
6. Ackermann, H.-W. 1997. Bacteriophage ecology, p. 335-339. In M.T. Martins, M.I.Z. Sato, J.M. Tiedje, L.C.N. Hagler, J. D”bereiner, and P.S. Sanchez (eds.), Progress in Microbial Ecology (Proceedings of Seventh International Symposium on Microbial Ecology). Brazilian Society for Microbiology,
7. Ackermann, H.-W. and M.S. DuBow . 1987. Viruses of Prokaryotes. CRC Press, Boca Raton, Florida.
8. Ackermann, H.-W. and H.M. Krisch. 1997. A catalogue of T4-type bacteriophages. Archives of Virology 142:2329-2345.
9. Adams, M.H. and F. Wassermann. 1956. Frequency distribution of phage release in the one-step growth experiment. Virology 2:96-108.
10. Allison, G.E. and T.R. Klaenhammer. 1998. Phage resistance mechanisms in lactic acid bacteria. International Dairy Journal 8:207-226.
11. Anderson, E.S. 1957. The relations of bacteriophages to bacterial ecology, p. 189-217. In R.E.O. Williams and C.C. Spicer (eds.), Microbial Ecology. Cambridge University Press, London.
12. Barron, B.A., V.A. Fischetti, and J.B. Zabriskie. 1970. Studies of the bacteriophage kinetics of multicellular systems: a statistical model for the estimation of burst size per cell in streptococci. Journal of Applied Bacteriology 33:436-442.
13. Bentzon, M.W., O. Maaloe, and G. Rasch. 1952. An analysis of the mode of increase in number of intracellular phage particles at different temperatures. Acta Pathologica et Microbiologica Scandinavica - Section B, Microbiology and Immunology 30:243-270.
14. Bohannan, B.J.M. and R.E. Lenski. 1997. Effect of resource enrichment on a chemostat community of bacteria and bacteriophage. Ecology 78:2303-2315.
15. Bohannan, B.J.M. and R.E. Lenski. 1999. Effect of prey heterogeneity on the response of a food chain to resource enrichment. American Naturalist 153:73-82. 16
16. Bohannan, B.J.M. and R.E. Lenski. 2000a. Linking genetic change to community evolution: insights from studies of bacteria and bacteriophage. Ecological Letters 3:362-377.
17. Bohannan, B.J.M. and R.E. Lenski. 2000b. The relative importance of competition and predation varies with productivity in a model community. American Naturalist 156:329-340.
18. Boucher, I. and S. Moineau. 2001. Phages of Lactococcus lactis: an ecological and economical equilibrium. Recent Research Developments in Virology 3:243-256.
19. Bremermann, H.-J. 1983. Parasites at the origin of life. Journal of Mathematical Biology 16:165-180.
20. Bullas, L.R., A.J. Zuccarelli, and R.L. Nutter. 1967. Aging effect on phage particles leading to an increase in burst size. Nature 216:1308
21. Burnet, F.M. 1934. The bacteriophages. Biological Reviews of the Cambridge Philosophical Society 9:332-350.
22. Burroughs, N.J., P. Marsh, and E.M.H. Wellington. 2000. Mathematical analysis of growth and interaction dynamics of streptomycetes and a bacteriophage in soil. Applied and Environmental Microbiology 66:3868-3877.
23. Campbell, A. 1961. Conditions for the existence of bacteriophages. Evolution 15:153-165.
24. Chao, L. 2000. The meaning of life. BioScience 50:245-250.
25. Chao, L., K.A. Hanley, C.L. Burch, C. Dahlberg, and P.E. Turner. 2000. Evolution of virulence in parasites: making hard and soft choices. Quarterly Review of Biology 75:261-275.
26. Chao, L., B.R. Levin, and F.M. Stewart. 1977. A complex community in a simple habitat: An experimental study with bacteria and phage. Ecology 58:369-378.
27. Chopin, A., A. Bolotin, A. Sorokin, S.D. Ehrlich, and M.C. Chopin. 2001. Analysis of six prophages in Lactococcus lactis IL1403: Different genetic structure of temperate and virulent phage populations. Nucleic Acids Research 29:644-651.
28. Conley, M.P. and W.B. Wood. 1975. Bacteriophage T4 whiskers: A rudimentary environment-sensing device. Proceedings of the National Academy of Sciences, U.S.A. 72:3701-3705.
29. Corbin, B.D., G.M. Aron, and R.J.C. McLeon. 2001. Bacteriophage T4 multiplication in an Escherichia coli biofilm. Canadian Journal of Microbiology 47:680-684.
30. Daniels, L.L. and A.C. Wais. 1998. Virulence of phage populations infecting Halobacterium cutirubrum. FEMS Microbiol Ecology. 25:129-134.
31. DeFilippis, V.R. and L.P. Villarreal. 2000. An introduction to the evolutionary ecology of viruses, p. 125-208. In C.J. Hurst (ed.), Viral Ecology. Academic Press, San Diego.
32. Delbrück, M. 1942. Bacterial viruses (bacteriophages). Advances in Enzymology 2:1-32.
33. Delbrück, M. 1945. The burst size distribution in the growth of bacterial viruses (bacteriophages). Journal of Bacteriology 50:131-135.
34. Doermann, A.H. 1948. Lysis and lysis inhibition with Escherichia coli bacteriophage. Journal of Bacteriology 55:257-275.
35. Doermann, A.H. 1952. The intracellular growth of bacteriophages. I. Liberation of intracellular bacteriophage T4 by premature lysis with another phage or with cyanide. Journal of General Physiology 35:645-656.
36. Douglas, J. 1975. Bacteriophages. p.77-133. Chapman and Hall, London.
37. Frank, S.A. 1994. Polymorphism of bacterial restriction-modification systems: The advantage of diversity. Evolution 48:1470-1477.
38. Fuhrman, J.A. 1999. Marine viruses and their biogeochemical and ecological effects. Nature 399:541-548.
39. Goyal, S.M., C.P. Gerba, and G. Bitton. 1987. Phage Ecology. CRC Press, Boca Raton, Florida.
40. Hadas, H., M. Einav, I. Fishov, and A. Zaritsky. 1997. Bacteriophage T4 development depends on the physiology of its host Escherichia coli. Microbiology 143:179-185.
41. Hanlon, G.W., S.P. Denyer, C.J. Olliff, and L.J. Ibrahim. 2001. Reduction in exopolysaccharide viscosity as an aid to bacteriophage penetration through Pseudomonas aeruginosa biofilms. Applied and Environmental Microbiology 67:2746-2753.
42. Hendrix, R.W. and R.L. Duda. 1992. Bacteriophage lamda PaPa: not the mother of all lambda phages. Science 258:1145-1148.
43. Hershey, A.D. 1946. Mutation of bacteriophage with respect to type of plaque. Genetics 31:620-640.
44. Hurst, C.J. and H.D.A. Lindquist. 2000. Defining the ecology of viruses, p. 3-40. In C.J. Hurst (ed.), Viral Ecology. Academic Press, San Diego.
45. Koch, A.L. 1964. The growth of viral plaques during the enlargement phase. Journal of Theoretical Biology 6:413-431.
46. Kuo, M.Y., M.K. Yang, W.P. Chen, and T.T. Kuo. 2000. High-frequency interconversion of turbid and clear plaque strains of bacteriophage f1 and associated host cell death. Canadian Journal of Microbiology 46:841-847.
47. Kutter, E., E. Kellenberger, K. Carlson, S. Eddy, J. Neitzel, L. Messinger, J. North, and B. Guttman. 1994. Effects of bacterial growth conditions and physiology on T4 infection, p. 406-418. In J.D. Karam (ed.), The Molecular Biology of Bacteriophage T4. ASM Press, Washington, DC.
48. Lee, Y., S.D. Eisner, and J. Yin. 1997. Antiserum inhibition of propagating viruses. Biotechnology and Bioengineering 55:542-546.
49. Lee, Y. and J. Yin. 1996b. Detection of evolving viruses. Nature Biotechnology 14:491-493.
50. Lee, Y. and J. Yin. 1996a. Imaging the propagation of viruses. Biotechnology and Bioengineering 52:438-442.
51. Lenski, R.E. 1988. Dynamics of interactions between bacteria and virulent bacteriophage. Advances in Microbial Ecology 10:1-44.
52. Levin, B.R., F.M. Stewart, and L. Chao. 1977. Resource limited growth, competition, and predation: A model and experimental studies with bacteria and bacteriophage. American Naturalist 111:3-24.
53. Martin, E.L. and T.A. Kokjohn. 1999. Cyanophages, p. 324-332. In A. Granoff and R.G. Webster (eds.), Encyclopedia of Virology second edition. Academic Press, San Diego.
54. McLean, R.J., B.D. Corbin, G.J. Balzer, and G.M. Aron. 2001. Phenotype characterization of genetically defined microorganisms and growth of bacteriophage in biofilms. Methods in Enzymology 336:163-174.
55. Middelboe, M. 2000. Bacterial growth rate and marine virus-host dynamics. Microbial Ecology 40:114-124.
56. Murray, A.G. and G.A. Jackson. 1992. Viral dynamics: A model of the effects of size, shape, motion, and abundance of single-celled planktonic organisms and other particles. Marine Ecology Progress Series 89:103-116.
57. Paul, J.H. and C.A. Kellogg. 2000. Ecology of bacteriophages in nature, p. 211-246. In C.J. Hurst (ed.), Viral Ecology. Academic Press, San Diego.
58. Paynter, M.J.B. and H.R. Bungay III. 1971. Characterization of virulent bacteriophage infections of Escherichia coli in continuous culture. Science 172:405-405.
59. Sagik, B.P. 1954. A specific reversible inhibition of bacteriophage T2. Journal of Bacteriology 68:430-436.
60. Sarma, T.A. and B. Kaur. 1997. Characterization of host-range mutants of cyanophage N-1. Acta Virology 41:245-250.
61. Schrag, S. and J.E. Mittler. 1996. Host-parasite persistence: the role of spatial refuges in stabilizing bacteria-phage interactions. American Naturalist 148:348-347.
62. Stewart, F.M. and B.R. Levin. 1984. The population biology of bacterial viruses: Why be temperate. Theoretical Population Biology 26:93-117.
63. Suttle, C.A. 2000a. Cyanophages and their role in the ecology of cyanobacteria, p. 563-589. In B.A. Whitton and M. Potts (eds.), The Ecology of Cyanobacteria: Their Diversity in Time and Space. Kluwer Academic Publishers, Boston.
64. Suttle, C.A. 2000b. The ecology, evolutionary and geochemical consequences of viral infection of cyanobacteria and eukaryotic algae, p. 248-286. In C.J. Hurst (ed.), Viral Ecology. Academic Press, New York.
65. Suttle, C.A. and F. Chen. 1992. Mechanisms and rates of decay of marine viruses in seawater. Applied and Environmental Microbiology 58:3721-3729.
66. Thingstad, T.F. 2000. Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnology and Oceanography 45:1320-1328.
67. Wang, I.-N., D.L. Smith, and R. Young. 2000. Holins: The protein clocks of bacteriophage infections. Annual Review of Microbiology 54:799-825.
68. Webb, V., E. Leduc, and G.B. Spiegelman. 1982. Burst size of bacteriophage SP82 as a function of growth rate of its host Bacillus subtilis. Canadian Journal of Microbiology 28:1277-1280.
69. Wilhelm, S.W. and C.A. Suttle. 2000. Viruses as regulators of nutrient cycles in the sea, p. 551-556. In Proceedings of the 8th International Symposium of Microbial Ecology.
70. Wilkinson, M.H.F. 2001. Predation in the presence of decoys: An inhibitory factor on pathogen control by bacteriophages or bdellovibrios in dense and diverse ecosystems. Journal of Theoretical Biology 208:27-36.
71. Williams, S.T., A.M. Mortimer, and L. Manchester. 1987. Ecology of soil bacteriophages, p. 157-179. In S.M. Goyal, C.P. Gerba, and G. Bitton (eds.), Phage Ecology. John Wiley & Sons, New York.
72. Wilson, G.S. and A.A. Miles. 1946. The bacteriophage, p. 325-350. In Topley and Wilson's Principles of Bacteriology and Immunity. Williams and Wilkins, Baltimore.
73. Wommack, K.E. and R.R. Colwell . 2000. Virioplankton: viruses in aquatic ecosystems. Microbiology and Molecular Biology Reviews 64:69-114.
74. Yin, J. and J.S. McCaskill. 1992. Replication of viruses in a growing plaque: A reaction-diffusion model. Biophysical Journal 61:1540-1549.

Figure 2: Exponential phage adsorption and phage population growth (below).

Figure 3: Computer-simulated chemostats (below).