Microbial Maintenance: A Critical Review on Its Quantification
- 2.3k Downloads
Microbial maintenance is an important concept in microbiology. Its quantification, however, is a subject of continuous debate, which seems to be caused by (1) its definition, which includes nongrowth components other than maintenance; (2) the existence of partly overlapping concepts; (3) the evolution of variables as constants; and (4) the neglect of cell death in microbial dynamics. The two historically most important parameters describing maintenance, the specific maintenance rate and the maintenance coefficient, are based on partly different nongrowth components. There is thus no constant relation between these parameters and previous equations on this subject are wrong. In addition, the partial overlap between these parameters does not allow the use of a simple combination of these parameters. This also applies for combinations of a threshold concentration with one of the other estimates of maintenance. Maintenance estimates should ideally explicitly describe each nongrowth component. A conceptual model is introduced that describes their relative importance and reconciles the various concepts and definitions. The sensitivity of maintenance on underlying components was analyzed and indicated that overall maintenance depends nonlinearly on relative death rates, relative growth rates, growth yield, and endogenous metabolism. This quantitative sensitivity analysis explains the felt need to develop growth-dependent adaptations of existing maintenance parameters, and indicates the importance of distinguishing the various nongrowth components. Future experiments should verify the sensitivity of maintenance components under cellular and environmental conditions.
KeywordsMicrobial Biomass Relative Growth Rate Relative Death Rate Biomass Increment Microbial Dynamic
Describing microbial dynamics is of great importance for many different applications. Microbial dynamics determines production levels in food industry, waste removal in reactors, pollution cleanup in soils and sediments, as well as the biogeochemical cycles to major extents. Although some kinetic approaches of microbial dynamics do not explicitly account for changes in microbial biomass, a majority of the kinetic descriptions incorporates some measure of microbial growth. Most microbiologists acknowledge that, apart from microbial growth, some measure of “maintenance” is needed to provide proper descriptions of microbial kinetics. Over the years, maintenance has been quantified in various, partly contradictory, ways. These contradictions are caused by four interrelated issues.
The first issue that lies at the heart of matter is the definition of maintenance. Maintenance has been defined as “the energy consumed for functions other than the production of new cell material” . This definition includes all nongrowth components and does not provide insight in its underlying processes, which complicates its general application. The nongrowth components thus included and determined empirically when measuring maintenance are (1) shifts in metabolic pathways, (2) energy spilling reactions, (3) cell motility, (4) changes in stored polymeric carbon, (5) osmoregulation, (6) extracellular losses of compounds not involved in osmoregulation, (7) proofreading, synthesis and, turnover of macromolecular compounds such as enzymes and RNA, and (8) defense against O2 stress [6, 30, 45, 50]. The opinion on which component is the dominant nongrowth component differs from osmoregulation , turnover of macromolecular compounds [6, 27, 50], to energy spilling reactions .
Second, and even more important, is that not all these nongrowth components are, physiologically speaking, part of maintenance—when maintenance is defined as a basic or endogenous metabolism. Endogenous metabolism was introduced  around the same time as “maintenance,” and the variables were frequently used interchangeably (e.g., ). Physiological maintenance comprises energy costs of osmoregulation, cell motility, defense mechanisms, and proofreading and internal turnover of macromolecular compounds. As indicated previously , physiological maintenance should not include shifts in metabolic pathways, storage of polymers, or extracellular losses. Some of the physiological maintenance components—such as shifts in metabolic pathways—seem to be better in place when corrected for in the cell growth yield, Y G, which was defined as the amount of cell growth per amount of substrate consumed . Physiological maintenance dominates the nongrowth losses of energy under starvation conditions .
Third, even though several incomparable nongrowth components were combined into a single variable, maintenance evolved as a biological kinetic constant through time. Originally, none of the microbial kinetic parameters used up to the present were initiated as constants. The most popular equation used to describe microbial growth, the Monod equation , was introduced as a purely empirical relationship resembling an adsorption isotherm. However, many microbiologists viewed the Monod equation as something that has an inherent meaning. From there, it is a small step to consider the variables involved as constants for the species considered. When measurements of maintenance (e.g., [14, 32]) showed deviations from a constant maintenance, the concept itself was not questioned, but an additional growth-rate-dependent maintenance parameter was introduced . Similar evolution occurred with growth yield Y G and the energy yield (Y ATP ) , which was introduced based on a correlation between biomass production and ATP availability . These were subsequently “proven” to be constants (e.g., ), whereas Stouthamer et al.  concluded that Y G is not a biological constant, based on measured shifts in metabolic pathways.
The fourth issue arises from the fact that cell death is hardly considered in microbial dynamics. Flux measurements do not distinguish cell death from intracellular turnover. Although Herbert  described maintenance as similar to a negative relative growth rate, only some applied models (see, e.g.,  for a review) consider it as a “relative death rate.” Explicit inclusion of death and lysis can have serious implications for the interpretation of maintenance effects .
These four issues have led to several inconsistencies in the quantification of “maintenance” and in equations describing microbial kinetics in general. The aim of this article is to describe the various measures of maintenance, to show the apparent relationship between the different measures and to quantify the dynamics of maintenance when distinguishing the various components included in maintenance to increase the understanding, description, and prediction of microbial dynamics.
Estimators for Maintenance
Specific Maintenance Rate
As shown in Eq. (1), the definition of specific maintenance rate is completely analogous to the definition of μ. Maintenance a is an imaginative decay rate to account for a diversion of substrate flux from growth, but does not necessarily lead to additional losses of substrate. In fact, a includes osmoregulation, extracellular losses and turnover terms, and neglects the occurrence of the other nongrowth components listed in the Introduction. Given its mathematical definition, a also includes relative death rates when a is determined experimentally from biomass and compound balances, e.g., in chemostat experiments. Estimates derived for either a or relative death rates thus overlap, and researchers should avoid applying both simultaneously in a kinetic model (as done in ). It also implies that a includes nongrowth components other than physiological maintenance, whereas some physiological maintenance processes are not incorporated. More complicated models are needed to estimate real physiological maintenance rates.
Relation between Specific Maintenance Rate and Maintenance Coefficient
Equation (9) implies that there is no constant relation between the two parameters describing maintenance (Fig. 2), because Y app depends on μ if maintenance occurs. Without sophisticated experimental tools, it is impossible to tell which parameter is nonconstant, because the two approaches on maintenance address partly the same and partly different nongrowth components and even cell death, whereas shifts in metabolic pathways are neglected by both approaches. For each component, the contribution to overall nongrowth losses may shift both in absolute and in relative terms with growth conditions. Therefore, the two ways of describing “maintenance” should be seen as two independent measures approaching a complex phenomenon from different directions. Combining empirical parameter estimates from both approaches in one model (as done frequently, e.g., ) should thus be done with great reservation and care. In reality, it is highly probable that given the different components combined in “maintenance,” partial losses occur in both biomass and in consumption. This will be extended and quantified below.
Only a few studies tried to separate the maintenance estimators. Servais et al.  quantified the relative release rate of label that had been incorporated into DNA as a relative death rate or specific maintenance rate (sensu ) assuming that DNA is not subject to maintenance (sensu ). Servais et al.  also showed that the temperature dependence of this specific maintenance rate was slightly less than generally found for relative growth rates.
Maintenance as a Minimum Substrate Concentration
Influence of Nongrowth Components on Overall Maintenance Dynamics
A Conceptual Model
Ideally, a maintenance description should distinguish physiological maintenance from other nongrowth components and incorporate the dynamics of each component. Unfortunately, experimental data on the contribution of individual nongrowth components—let alone information on changes in contributions as a function of growth conditions—are scarce. Mason et al.  introduced a maintenance model that was based on death and cell lysis. Lysis was incorporated in decreased Y values. Unfortunately, their four compartments (active cells, nonviable active cells, dead cells, and extracellular products) are hard to distinguish, as indicated by the same authors, and led to numerous unknown kinetic parameters that had to be fitted. Beeftink et al.  also derived kinetic equations by accounting for death and maintenance separately. However, their critical assumptions on constant total maintenance energy and the applied modulation between decay and maintenance-associated catabolism are questionable given the variability in maintenance requirements.
In this section, an analysis is introduced that builds upon existing formulations and that separates the individual nongrowth components to the extent allowed by the limited experimental data. To allow application in kinetic studies, the conceptual model contains a minimum number of parameters that may be estimated from kinetic studies. Moreover, only single species cultures are considered. Microbial physiological processes were simplified and lumped using existing formulations as long as these did not contradict first principles. Present experimental data do not justify further specification. The conceptual model is by no means quantitatively correct, but allows us to analyze the sensitivity of microbial dynamics to various nongrowth components and may be refined when experimental data become available.
Most theoretical approaches assume that Y G is constant for a given substrate S. However, this is not appropriate if shifts in metabolic pathways occur . Varying Y G as a function of growth conditions would thus quantify the last nongrowth component. Unfortunately, there are no experimental data that allow general quantification of this variation in Y G .
Relating the Conceptual Model to Maintenance Parameters
The described conceptual model explicitly incorporates all nongrowth terms listed in the Introduction: d allows for excretion, leakage, and cell death; x i allows for storage and growth-dependent activities; and m p for the various physiological maintenance processes, and the variation in Y G allows for shifts in metabolic pathways. It allows the quantification of nongrowth components as a function of growth conditions and can be related to existing maintenance parameters:
Sensitivity Analysis of the Conceptual Model
The above analysis of “maintenance” estimators shows that there is no simple solution to reconcile the different approaches and terminologies. First, the analysis underlines the importance of distinguishing and explicitly describing and quantifying the various components of maintenance to understand its dynamics and the differences between maintenance estimates. This was hardly done in the past. The conceptual model—based on common sense microbiological assumptions such as inactive vs reactive fractions and Monod kinetics, simplified to apply to quasi steady states only—incorporated all nongrowth components and showed that maintenance is a dynamic process that depends nonlinearly on relative death rates, relative growth rates, physiological maintenance, and growth yield. The first two variables dominated the variability in the overall maintenance. This is not to say that physiological maintenance is unimportant, because part of the sensitivity of m p on overall maintenance goes through μ r , see Eq. (17). The dominance of μ on m tot variability led to an almost linear dependence and was mainly due to the mechanistic distinction between an inactive and an active microbial fraction. Similar linear relationships between μ and maintenance have been incorporated in empirical formulations (e.g., [32, 39]).
This is not meant to indicate that the presented conceptual model is quantitatively correct, but it considers and quantifies all nongrowth components and is, contrary to earlier descriptions on maintenance, internally consistent in the sense that it avoids combinations of contrasting descriptions of maintenance. In its present form, the conceptual model is highly helpful as it provides estimates on the sensitivity of maintenance on the underlying components and shows maintenance dynamics. These characteristics may make the approach attractive as a point of departure to develop and validate mechanistic models on maintenance.
Second, the analysis made clear that the difference between physiological maintenance (or endogenous metabolism sensu ) and other nongrowth conditions is crucial. Explicit consideration of this difference in models would greatly improve our understanding of the phenomenon, whereas neglect of this factor has led to many misunderstandings on the quantification of maintenance and its dynamics as a function of growth conditions.
Third, the analysis pinpointed inconsistencies in current formulations of maintenance, i.e., the specific maintenance rate and the maintenance coefficient. These variables are principally incompatible and there is no constant relation between the two. Schulze and Lipe's  derived relationship between the two, which was applied by others later on, is incorrect. Comprehension and correct use of these differences is crucial to correctly describe maintenance.
Finally, an important illustrated aspect of microbial dynamics is the variability in kinetic macroscopic “constants” with growth conditions as shown for the overall maintenance coefficient. Moreover, the analysis showed how the dynamics of Y G is intermingled with maintenance. This problem is more general than discussed above, because Y G is derived empirically as the maximum yield after correcting for maintenance according to a presupposed formulation on the variation of maintenance with cultivation conditions. By this procedure, Y G becomes by definition a constant, whose value depends on the chosen formulation for maintenance. This variability in kinetic parameters and its separation needs reevaluation.
All results outlined above were obtained from macroscopic microbial features. It should be noted, however, that the conclusions do not change when considering energy balances instead, i.e., when applying Y ATP and m E as yield and maintenance coefficients (e.g., [12, 52]) instead of Y G and m. The equations to calculate m E  are fully analogous to Eq. (8a), with the same conceptual problems and assumptions of constancy. Moreover, similar to Y G , Y ATP is not constant even when considering maintenance .
Implications for Microbial Ecology
Many microbial ecologists have a clear concept on the processes entangled in physiological maintenance or endogenous metabolism, but this is not what is measured empirically. The introduced conceptual model has quantified how physiological maintenance in concert with other nongrowth components determines empirical maintenance estimates. This conceptual model also provides possible explanations for the wide variety of maintenance estimates among microorganisms. The most sensitive variable was the estimate of relative death rates, which was taken independent of growth conditions although that is a major simplification. A decrease in relative death rates with relative growth rate, as found experimentally , leads to even stronger nonlinearities in m tot with growth conditions. Surprisingly, estimates of m tot were relatively insensitive to variation in physiological maintenance, which might imply that this is relatively unimportant in explaining the variety in m tot among microorganisms. The variability in m tot and its nonlinearity needs to be considered by experimental microbial ecologists and should lead to adjustments in the equations used to describe microbial dynamics.
Apart from explanations of measured maintenance dynamics, the conceptual model also shows the need for experimental data on the contribution of the various nongrowth components, a need felt already by Pirt . The presented model may be used to validate—components of—maintenance. Attempts to quantify physiological maintenance costs already exist , and may be integrated with measurements on the dynamics of relative death rates, e.g., through the application of flow cytometry. If, in addition, future experimental studies would make a stringent separation of effects on Y G and physiological effects on maintenance, then the contribution of physiological maintenance costs, growth yield dynamics, and losses of biomass on m tot with changing growth conditions may be quantified. This would validate and improve the presented conceptual model and would strongly improve our knowledge on maintenance and nongrowth components in general.
In conclusion, the general definition of maintenance has led to a partly independent development of several maintenance estimators. Each maintenance parameter approaches the phenomenon from a different perspective and is thus fundamentally different. Previous attempts to relate the parameters are incorrect. The review of studies on maintenance showed that, although they started initially as constants, growth-dependent adaptations of maintenance variables were developed over time. A conceptual model that explicitly described the various nongrowth components also showed a strong dependence of overall maintenance on the relative growth rate. Apart from relative growth rates, overall maintenance depended in a nonlinear way on the combination of relative death rates, physiological maintenance, and growth yield. This analysis emphasizes that the components underlying maintenance should be considered and distinguished explicitly. Future experiments should verify the sensitivities of maintenance on the actual cellular and external environmental conditions.
Nancy de Bakker, Hans Scholten, Reinoud Segers, and three anonymous reviewers provided constructive comments on an earlier version of this manuscript.
- 10.Ghosh, S, Pohland, FG (1974) Kinetics of substrate assimilation and product formation in anaerobic digestion. J Water Pollut Control Fed 46: 748–758Google Scholar
- 15.Herbert, D (1958) Some principles of continuous culture. In: Tunevall, G (Ed.) Recent Progress in Microbiology, VII Intern. Congr. for Microbiology, Stockholm, pp 381–396Google Scholar
- 17.Ivanova, TI, Nesterov, AI (1988) Production of organic exometabolites by diverse cultures of obligate methanotrophs. Microbiology 57: 486–491Google Scholar
- 21.Kugelman, JJ, Chin, KK (1971) Toxicity, synergism, and antagonism in anaerobic waste treatment processes. In: Pohland, FG (Ed.) Anaerobic Biological Treatment Processes. American Chemical Society, Washington DC, pp 55–90Google Scholar
- 22.Lawrence, AW, McCarty, PL (1969) Kinetics of methane formation in anaerobic treatment. J Water Pollut Control Fed 41: R1–R17Google Scholar
- 29.Mason, CA, Bryers, JD, Hamer, G (1986) Activity, death and lysis during microbial growth in a chemostat. Chem Eng Commun 45: 163–176Google Scholar
- 33.Novak, JT, Carlson, DA (1970) The kinetics of anaerobic long chain fatty acid degradation. J Water Pollut Control Fed 42: 1933–1943Google Scholar
- 34.O'Rourke, JT (1968) Kinetics of Anaerobic Treatment at Reduced Temperatures. Stanford University, CaliforniaGoogle Scholar
- 35.Panikov, NS (1995) Microbial Growth Kinetics, 1st edn. Chapman and Hall, LondonGoogle Scholar
- 36.Pavlostathis, SG, Giraldo-Gomez, E (1991) Kinetics of anaerobic treatment. Water Sci Technol 24: 35–59Google Scholar
- 41.Powell, EO (1967) The growth rate of microorganisms as a function of substrate concentration. In: Powell, EO, Evans, CGT, Strange, RE, Tempest, DW (Eds.) Microbial Physiology and Continuous Culture. HMSO, London, pp 34–56Google Scholar
- 42.Richter, O, Diekkrüger, B, Nörtersheuser, P (1996) Environmental Fate Modelling of Pesticides. VCH, WeinheimGoogle Scholar
- 50.Stouthamer, AH, Bulthuis, BA, van Verseveld, HW (1990) Energetics of growth at low growth rates and its relevance for the maintenance concept. In: Poole, RK, Bazin, MJ, Keevil, CW (Eds.) Microbial Growth Dynamics. IRL Press, Oxford, Special Publications Society for General Microbiology, vol. 28, pp 85–102Google Scholar
- 54.Tros, ME, Bosma, TNP, Schraa, G, Zehnder, AJB (1996) Measurement of minimum substrate concentration (S min) in a recycling fermentor and its prediction from the kinetic parameters of Pseudomonas sp. Strain B13 from batch and chemostat cultures. Appl Environ Microbiol 62: 3655–3661PubMedGoogle Scholar
- 55.van Niel, EWJ, Pedro Gomes, TM, Willems, A, Collins, MD, Prins, RA, Gottschal, JC (1996) The role of polyglucose in oxygen-dependent respiration by a new strain of Desulfovibrio salexigens. FEMS Microbiol Ecol 21: 243–253Google Scholar