You are on page 1of 10

Food Control 29 (2013) 290e299

Contents lists available at SciVerse ScienceDirect

Food Control
journal homepage:

Predictive microbiology theory and application: Is it all about rates?

Tom McMeekin*, June Olley, David Ratkowsky, Ross Corkrey, Tom Ross
Food Safety Centre, Tasmanian Institute of Agriculture/School of Agricultural Science, University of Tasmania, Hobart 7001, Tasmania, Australia

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 9 March 2012
Received in revised form
23 May 2012
Accepted 1 June 2012

We review early work on the microbial growth curve and the concept of balanced growth followed by
commentary on the stringent response and persister cells. There is a voluminous literature on the effect
of antibiotics on resistance and persistence and we call for a greater focus in food microbiology on the
effect of biocides in the same context. We also raise potential issues in development of resistance arising
from sourceesink dynamics and from horizontal gene transfer. Redox potential is identied as crucial
in determining microbial survival or death, and the recently postulated role for reactive oxygen species in
signalling also considered.
Traditional predictive microbiology is revisited with emphasis on temperature dependence. We
interpret the temperature vs growth rate curve as comprising 11 regions, some well-recognised but
others leading to new insights into physiological responses. In particular we are intrigued by a major
disruption in the monotonic rate of inactivation at a temperature, slightly below the actual maximum
temperature for growth. This non-intuitive behaviour was earlier reported by other research groups and
here we propose that it results from a rapid metabolic switch from the relaxed growth state to the
stringent survival state.
Finally, we envision the future of predictive microbiology in which models morph from empirical to
mechanistic underpinned by microbial physiology and bioinformatics to grow into Systems Biology.
2012 Elsevier Ltd. All rights reserved.

Predictive microbiology
Systems biology
Biological rates
Stringent response
Persister cells
Growth rate etemperature response
Mechanistic model

1. Introduction
This paper is based on a presentation of the same title given at
7ICPMF in September 2011 (McMeekin, Olley, Ratkowsky, & Ross,
2011). In the presentation we discussed rates and time scales in
microbiology, early studies on microbial growth rate curves and the
concept of balanced growth. This was followed by consideration of
the stringent response and persister cells, topics which are underrepresented in the food microbiology literature. Attention was then
focused on modelling studies with emphasis on temperature
dependence models. Finally, we asked if the quantitative information embedded in predictive models could be effectively integrated
with microbial physiology and omics technologies.
In answer to the question posed in the presentation title, . is it
all about rates? we were condent to respond positively when the
question addressed the use of predictive models to estimate the
safety and shelf-life of foods. However, our response was equivocal
when the question was framed in the context of integrating
predictive models with microbial physiology studies and the data
deluge emanating from omics studies. Here we extend the scope

* Corresponding author.
E-mail address: (T. McMeekin).
0956-7135/$ e see front matter 2012 Elsevier Ltd. All rights reserved.

of the 7ICPMF presentation by incorporating recent literature (much

of it published since September 2011) in an attempt to provide
a denitive response to the objective of successful integration of
traditional and futuristic predictive modelling.
2. Rates: all pervasive and all persuasive?
Time scales in microbiology range from milliseconds for enzyme
catalysed reactions (Stockbridge, Lewis, Yuan, & Wolfenden, 2010)
to doubling times of 7 min for Clostridium perfringens to days,
weeks or months for psychrophiles growing under optimal conditions to more than 3.5 billion years for life to reach the current level
of adaptive evolution. The last time frame has been achieved only as
a result of the sub-second rates of enzyme catalysed reactions. For
example the enzyme catalysed decarboxylation of orotdine-50 phosphate, the nal step in the synthesis of pyrimidines and thus
nucleic acids, has a half-life of 0.017 s. Without the enzyme the halflife of the same reaction at 20  C is 78 million years (Stockbridge
et al., 2010).
In microbiology we consider rates, with time as the universal
denominator to describe the development and decline of microbial
populations. Much of the early work on microbial growth was carried
out by the Paris School (Monod, 1942, 1949) and the Copenhagen
School of Maale, Kjeldgaard, Neidhardt and Schaechter. Schaechter

T. McMeekin et al. / Food Control 29 (2013) 290e299

(2006) provided an interesting retrospective of that groups research

drawing particular attention to the concept of balanced growth,
dened by Campbell (1957) as the condition in which all cell
constituents increase in proportion over the same time interval. Thus,
cell composition varies with growth rate with higher rates of growth
requiring more DNA, RNA and proteins (Kjeldgaard, Maale, &
Schaechter, 1958; Schaechter, Maale, & Kjeldgaard, 1958). An
important conclusion drawn by the authors was that growth rate is
the primary factor determining the physiological state of cells.
Schaechter (2006) also argued strongly for precision to be
maintained in the design and execution of experiments to ensure
balanced growth commenting that: Not infrequently articles
published in the literature include only a casual mention of how the
cultures were indication of indifference is the use of
terms like mid-log phase. This is nearly meaningless. It is the midpoint of what? .. conditions and a growth curve should be published in detail in papers that describe bacterial growth. Prerequisites for specication of the physiological state include exact
denition of media composition, inoculum size and time/temperature combinations for preparation of starter cultures.
To understand more about the origins and signicance of the
SchaechtereMaaleeKjeldgaard experiments and the excitement
generated by the pioneering research of the Copenhagen School,
Cooper (1993) is a must read paper in which the author provided
insights into the experiments, the researchers, and the experimental process. The experimental process also attracted comment
from Neidhardt (1999) who noted that: analysis of problems
relating to the growth rate are as reductionist as any.. Yet, when
growth is the ultimate interest, one cannot delve into single
enzymes, genes or pathways without returning to the whole cell
and [its] coordinated operation. In other words the cell is not
simply a bag of enzymes.
Discussing the concept of solving the cell Cooper (1993)
opined that the Copenhagen school markedly advanced understanding of microbial physiology. Now this pursuit has moved to
the next level with the widespread adoption of molecular techniques in physiological studies as predicted by Schaechter (2006) in
the title of his publication From growth physiology to systems
biology. Balanced growth also provides a physiological interpretation of transition from the lag phase to the exponential growth
phase and then to the stationary phase of growth, respectively as
a nutritional up-shift and a series of nutritional down-shifts. While
the lag and stationary phases are regions of zero net growth rates
for populations of cells, it is important not to lose sight that zero
rates have important physiological and practical implications.
Another member of the Copenhagen School, Neidhardt (1999),
wrote about the obsession with dN/dt (i.e., population growth
rate) but did not subscribe to the view that emphasis on growth had
delayed work on the stationary phase. On the contrary he observed
that understanding growth was a prerequisite for studying the
physiology of the stationary phase noting that the physiology of
non-growth is not the absence of the physiology of growth.
Others argue that the real action in the bacterial population
growth cycle is conned to resolution of the lag phase and the
exponential growth phase where the maximum growth rate is
reached. This may be so for some spoilage and pathogenic bacteria
where the maximum population density (MPD) in a food will be
attained after spoilage has occurred or safety is compromised. But,
with these organisms, as well as with probiotic and protective
cultures, full growth curves also indicate the MPD and provide
information on potential interactions between components of
mixed populations. An example of a non-specic interaction is the
effect described by Jameson (1962) who observed that the rst
group of organisms to reach its MPD would cause competitors to
move from the exponential to the stationary phase i.e. from a rapid


rate to a zero net growth rate. Despite that the lag and stationary
phases of growth are regions of zero or very slow growth they are
crucial to the continuation of a lineage as will be evident when the
physiology of the stringent response and of persister cells are
considered. The physiology of the lag phase identies it as a distinct
growth phase that prepares cells for exponential growth (Rolfe
et al., 2012).
3. The stringent response e paradigm lost in food
Even major changes in the physiology of the cell can occur in
seconds. A good example is the transition from the relaxed
response state to that of the stringent response and the converse
switch which occur in 20e30 s (Cashel, 1975). This is equivalent to
the half lives of guanosine tetraphosphate (ppGpp) and guanosine
pentaphosphate (ppGppp), the alarmones that give an unequivocal
clue to cells to switch from one state to the other (Lund &
Kjeldgaard, 1972). Thus, in a few seconds the purpose of cellular
metabolism is totally reversed from a focus on growth (the relaxed
state) to a focus on survival (the stringent [response] state; SR).
The SR has been shown to play a signicant role in processes as
different as biolm formation, quorum sensing, antibiotic resistance, and virulence regulation (Navarro Llorens, Tormo, &
Martinez-Garcia, 2010). Perhaps, it also provides an explanation
for the Jameson Effect? Here we posit that the component of
a mixed culture to approach MPD rst produces sufcient ppGpp
not only to signal its entry into the SR state but also that of its
competitors. Relief from the SR state can be achieved by inoculation
into a fresh batch culture or prevented by growth in continuous
cultures in which the alarmone is continually diluted and nutrients
are continually added. However, if a continuous culture is set up in
a retentostat (a chemostat with 100% feedback of biomass) very
slow or zero growth rates ensue (Chesbro, Evans, & Eifert, 1979;
Gofn et al., 2010). This phenomenon can again be attributed to the
programmed objective of the stringent response physiological
state: survival at any cost. During transition from the exponential to
the stationary phase the cytoplasmic concentration of the general
stress factor (RpoS) increases in Gram negative cells (Gentry et al.,
2003; Lange & Hengge Aronis, 1991, 1994) and cell density and
quorum sensing compounds have been proposed to have a role in
the mechanism of transition (Hengge-Aronis, 2002). However,
Ihssen and Egli (2004) argued that, as other factors (growth rate,
carbon source availability and metabolite concentration) also
change during the transition, quorum sensing alone is insufcient
to explain the transition. They concluded, on the basis of both batch
and continuous culture experiments, that RpoS expression is not
controlled by quorum sensing, that specic growth rate plays
a prominent role and that ppGpp is a possible intracellular signal
linking RpoS and specic growth rate. From an ecological standpoint this egoistic strategy of self-determination by individual
cells was deemed to be eminently sensible compared with the
more risky option of depending on signals from other cells. The line
of reasoning developed by Ihssen and Egli (2004) was supported by
Potrykus, Murphy, Philippe, and Cashel (2011) who concluded that
ppGpp is the major source of growth rate control in Escherichia coli.
4. Persister cells e stealth bombers in the microbial
survival armoury?
Cells with slow or zero growth rates confer a very distinct
competitive advantage on microbial populations that, as a result of
minimal metabolism, are extremely difcult to inactivate. The term,
bacterial persistence, was introduced by Bigger (1944) who reported the inability of ampicillin to sterilize cultures of


T. McMeekin et al. / Food Control 29 (2013) 290e299

Staphylococcus aureus. Several studies prior to 1944 in which

super-survivors were described were noted by Kolter (1999).
Research on persistence continues apace with organisms including
E. coli (Shah et al., 2006), Pseudomonas aeruginosa (Mulcahy, Burns,
Lory, & Lewis, 2010; Silver, 2011), and Mycobacterium tuberculosis
and Candida albicans (Fauvart, Grootet, & Michiels, 2011).
Balaban, Merrin, Chait, Kowalik, and Leibler (2005) made
a signicant advance using a microuidic device and microscopy
to distinguish between normal, rapidly growing cells and slow
growing persister cells. Persistence was shown to arise from preexisting heterogeneity in bacterial populations with phenotypic
switching occurring between fast growing cells and slow growing
cells. Importantly, this demonstrated that mutation was not the
mechanism of persistence development because cells subcultured from persisters remained sensitive to ampicillin. For
a perspective on non-inherited antibiotic resistance see Levin and
Rozen (2006). A microuidic device also proved useful to track
accelerated antibiotic resistance in connected environments
(Zhang et al., 2011). The device was designed to mimic naturally
occurring bacterial niches in the mammalian body. A spatially
complex microenvironment was suggested to lead to accelerated
evolution as a result of a stress gradient in which a mutant, that
acquires resistance to a local stress, will have increased tness in
the population. The outcome was that four single nucleotide
polymorphisms were established in 10 h from an inoculum containing as few as 100 cells.
Biolms are widely reported in the literature to provide
protection to cells contained in the extracellular matrix making
them more resistant to environmental insults than planktonic cells
(Costerton, Stewart, & Greenberg, 1999). Apart from physical
protection supposedly afforded by the biolm, tolerance to antibiotics is increased as a result of slow growth rates (Gilbert, Collier, &
Brown, 1990). But, this is not a unique feature of biolms and,
contrary to popular opinion, slow growing planktonic cells can also
display enhanced antibiotic tolerance (Spoering & Lewis, 2001).
These workers also showed that, while the biocide peracetic acid
was less effective in killing biolm cells than exponential phase
planktonic cells, stationary phase planktonic cells were more
resistant than biolm cells. Thus, the physiologically slow state and
growth rate of stationary phase cells are key elements in
While the vast majority of persister cell studies have been
concerned with antibiotics, the peracetic acid effects described by
Spoering and Lewis (2001) expose the commonly held misconception that sessile cells are more resistant to sanitisers and
disinfectants used in the food industry without reference to the role
of their physiological state. The discussion in Spoering and Lewis
(2001) is particularly pertinent: Unlike conventional antibiotics
biocides. kill cells directly. Treatment with peracetic acid did not
result in persister cells which can survive antibiotic levels
100e1000 times higher than the minimum inhibitory concentration (MIC). The mechanism of killing by antibiotics is indirect and
involves activation of a programmed cell death response. By
contrast, direct lethal damage caused by biocides resulted in total
kills at only four times the MIC (Lewis, 2001).
A link back to the stringent response is found in Rodionov and
Ishiguro (1998) and references therein, e.g. Sarubbi, Rudd, and
Cashel (1988), in which steady state growth rates were shown to
be affected by ppGpp production. Recent research on the stringent
response and persisters includes Wang and Wood (2011) and
Yamaguchi and Inouye (2011) who demonstrated a role for toxin/
antitoxin systems in biolm and persister cell formation; Kim et al.
(2011) who screened a chemical library to locate a single compound
that caused persisters to revert to antibiotic sensitive cells and
Nguyen et al. (2011) who concluded that starved cells and biolm

cells of P. aeruginosa increase antibiotic resistance by active

responses to starvation rather than the passive effects of growth
arrest. The protective mechanism described was controlled by the
SR and reduced levels of oxidative stress. When the protective
mechanism was inactivated, the sensitivity of biolm cells
increased by several orders of magnitude and the efcacy of antibiotic treatment was markedly increased in experimental infections. The reduction in oxidative stress is consistent with the work
of Belenky and Collins (2011) and Shatalin, Shatalina, Mironov, and
Nudler (2011).
5. Greater focus on biocides in food microbiology?
The number of publications on the role of the SR and persister
cells in surviving antibiotic challenges greatly outnumbers those
concerned with detailed mechanisms of survival of cells when
challenged with biocides. The possible effect of four antimicrobial
substances widely used in the food industry was addressed by EFSA
(2008) which wrote that despite a long history of use, there are
currently no published data to conclude that the application of
chlorine dioxide, acidied sodium chlorite, trisodium phosphate
and peroxy acids to remove microbial contamination from poultry
carcasses will lead to the occurrence of acquired reduced susceptibility to these substances. The same conclusion was reached in
relation to the development of resistance to therapeutic antibiotics
arising from use of the four biocides. However, the panel recognised
that uncertainties exist and recommended that research on other
biocides should continue. Essentially the same conclusions were
reached earlier by an expert panel convened by IFT (Davidson &
Harrison, 2002). A similar opinion was expressed by Meyer
(2006), by Kastbjerg et al. (2009) who measured internal pH
levels in a persistent strain of Listeria monocytogenes and by
Kastbjerg et al. (2010). The response to an acid challenge on a single
cell level was the same whether cells were planktonic or attached.
Thus, persistence of the strain in food processing environments was
deemed unlikely to result from selection of a more acid tolerant
sub-population. Park and Chen (2011) examined inactivation of
STEC strains producing extracellular cellulose by degrading the
cellulose with cellulase and subsequent treatment with acetic and
lactic acids and sanitisers. Cellulase alone had no effect on survival
but strains with reduced cellulose were more susceptible to organic
acids and sanitisers. This nding is consistent with that of Poole
(2002) who reviewed the literature on biocide resistance. In this
the author indicated that impermeability (with target alteration
and efux mechanisms) were the major survival strategies.
5.1. Connected microenvironments and sourceesink dynamics
The fascinating results of Zhang et al. (2011), revealing accelerated evolution in connected microenvironments, requires consideration of similar microcosms in agricultural and food
environments. In the former, examples include: soil pores and
structural features of plant roots, shoots and leaves which are
associated with the vexed question of internalisation of human
pathogens in plant tissues. In food environments, biolms are an
example of connected microenvironments with gradients of pH
and oxygen tension created by microbial metabolism. Similarly,
solid, multi-component foods such as salami also may represent
temporally connected microenvironments as pH, osmotic and
redox conditions change systematically and markedly during
fermentation and maturation.
In addressing the question we should remember that Zhang
et al. (2011) exclusively examined accelerated emergence of
bacterial antibiotic resistance and, as discussed above, properly

T. McMeekin et al. / Food Control 29 (2013) 290e299

applied sanitisers and disinfectants do not appear to lead to

formation of persister cells. Nevertheless, expert panels and individual authors who have written on the potential to induce
increased resistance to commonly used cleaning agents add
a caveat to their general conclusion that this is unlikely. The caveat
normally recommends that further research is required because
current knowledge is decient.
Zhang et al. (2011) discussed the sourceesink model of
evolution that, like the stringent response, does not appear to be
well represented in the food microbiology literature. A sink is
a species habitat in which local mortality exceeds local reproduction. In a sink, a population can only be maintained with continuous migration from the source habitat where reproduction
exceeds mortality (Hermsen & Hwa, 2010). Characteristically, sinks
occur near the limits of a species range. In sourceesink dynamics
(SSD), sources provide poorly adapted immigrants to the sink and
those that can adapt through mutation are able to colonise the sink.
SSD are reported to lead to accelerated bacterial drug resistance,
resistance of fruit ies to insecticides and, relevant to this discussion, the evolution of virulence (Chattopadhyay et al., 2007;
Dennehy, Friedenberg, McBride, Holt, & Turner, 2010; Sokurenko,
Gomulkiewicz, & Dykhuizen, 2006). SSD studies have also
provided advances in understanding evolutionary processes with
several publications on biological tness landscapes revealing
evolutionary pathways and why these pathways were selected
(Lobkovsky, Wolf, & Koonin, 2010; Poelwijk, Kiviet, Weinreich, &
Tans, 2007; Rowe, Wedge, Platt, Kell, & Knowles, 2010).
How are we to interpret sourceesink dynamic situations in
relation to the potential development of biocide resistance? Are
they the result of an adaptive, phenotypic process resulting in
habituated cells which will return to their original state on removal
of the stress? Or, are they the result of mutations which become
xed in a newly competitive population and will be passed on to
progeny of the lineage. Perhaps both strategies are in operation
providing the population with an increasingly diverse array of
options for survival? Laland, Sterelny, Odling-Smee, Hoppitt, and
Uller (2011) offered a philosophical treatment of cause and effect in
biology, observing that proximate (immediate) outcomes such as
physiological responses have traditionally been separated from
ultimate (evolutionary) outcomes (Mayr, 1961).
5.2. Horizontal gene transfer and gene duplication inuence the
rate of prokaryotic evolution
Laland et al. (2011) concluded that it is time for the proximateultimate dichotomy to be replaced in biological sciences by a new
reciprocal concept of causation. They proposed that the microbial survival imperative is serviced by adaptive, phenotypic
responses which are induced rapidly and become redundant
when a crisis passes as retention would place an unnecessary
energy drain on the cell; plus an investment in long term
(evolutionary) responses. Neither of these strategies is static, and
the extent and importance of horizontal gene transfer (HGT) both
in proximate and ultimate outcomes is increasingly recognised
(see Koonin, Makarova, & Aravind, 2001; Smillie et al., 2011 and
references therein). Treangen and Rocha (2011) compared the
roles of HGT and gene duplication showing that the vast majority
of expansions in protein families are due to transfer. Transferred
genes persist longer while duplicated genes are expressed more,
suggesting that transfer and duplication have different (but
complementary) roles in evolutionary adaptation of prokaryotes;
transfer enables new functions and duplication leads to increased
gene dosage.
Smillie et al. (2011) concluded that a global network of gene
exchange among members of the human microbiome was driven


by ecology rather than geography or phylogeny. They reasoned that

genes recently transferred between bacteria occupying welldened niches are likely to reect adaptation to that niche.
Perhaps, this provides additional support for the view of BassBecking (1895e1963) that everything is everywhere but the
environment selects (de Wit & Bouvier, 2006). Other examples
included massive gene exchange between archaeal and bacterial
hyperthermophiles (Aravind, Tatusov, Wolf, Walker, & Koonin,
1998), transfer of carbohydrate-active genes from marine bacteria
to Japanese gut microbiota (Hehemann et al., 2010) and high levels
of HGT by Shewanella baltica in a nutrient rich pelagic environment
leading to its designation by Caro-Quintero et al. (2011) as the
ultimate r-strategist. The potential of HGT to cause rapid transfer
of genetic material was well illustrated by Stecher et al. (2012) who
demonstrated that gut inammation induced by S. Typhimurium
triggered blooms of commensal E. coli. These were infected, with
100% efciency in vivo, by a conjugative colicin plasmid from S.
Typhimurium and out-competed the inhibitory effects of the
commensal microbiota. Conversely, the ecological competition for
colonisation of a niche was won by the indigenous soil microbiota
that prevented establishment of invading E. coli O157:H7 (van Elsas
et al., 2004).
6. Redox potential e an arbiter of growth or death and
associated rates
Above we were concerned with microbial survival and presented several examples of widely contrasting time scales and rates
from rapid physiological responses to eons of evolutionary time.
One of the major events during evolution was the microbial driven
switch from anaerobiosis to the highly oxygenated world that we
currently inhabit. But, while aerobic metabolism was essential for
development of more complex life forms there was also a downside
because of the generation of reactive oxygen species (ROS) that
cause cell senescence and death.
Only recently have we started to realise the enormous
importance of the anaerobic nether regions in soils, sediments
and permafrost where billions of tonnes of carbon are captured in
huge microbial populations. Estimating the extent and rate of
release of carbon dioxide from these carbon sinks is one of the
most active research areas in biology at this time. The applicability of models, originally derived to estimate the microbial
stability and safety of foods, to the often-called big picture
science applications such as climate change, environmental
degradation and bioremediation and biotechnology will be discussed later.
An inevitable consequence of anaerobic conditions is a marked
reduction in energy yield from fermentative metabolic pathways
compared to respiratory metabolism. However, for the microbial
legions stable, anaerobic and cold conditions with minimal maintenance metabolism offer remarkable longevity that some consider
to approach immortality (Price & Sowers, 2004). Hamanaka and
Chandel (2011), Anastasiou et al. (2011) and Guerra et al. (2011)
drew attention to reduction in reactive oxygen species (ROS) and
reducing conditions as factors promoting cell survival. A similar
conclusion was reached earlier by Chesney, Eaton, and Mahoney
(1996) in protecting cells from the effects of chlorine. Intracellular
glutathione protected E. coli against hydrogen peroxide, hypochlorous acid and chloramines and glutathione decient mutants
were approximately twice as sensitive to killing by each of the
While ROS are well known for their role in cell stress, Dickinson
and Chang (2011) suggested that an alternative role as signal
molecules may contribute to physiology and increased tness. The
signalling function is achieved by:


T. McMeekin et al. / Food Control 29 (2013) 290e299

 co-localisation of the ROS sources and targets. Many ROS

species are inherently unstable and reactive, e.g., the cellular
half-life of [OH] is w109 s and for H2O2 is w1 ms;
 modulation of local redox buffering capacity by millimolar
concentrations of glutathione provides substantial redox
buffering for many ROS but reacts too slowly with H2O2. Buffering the latter is achieved by peroxiredoxin proteins that cycle
between reduced dithiol proteins and oxidised disulphide
forms controlled by glutathione and H2O2 respectively.
Reduction of disulphide bonds also features in releasing the
antimicrobial activity of human defensin-1 (hBD1) against the
opportunistic pathogens and gut commensals (Schroeder et al.,
2011) prompting the authors to postulate a general defence
mechanism regulating the colonisation of mucosal membranes and
skin surfaces.
The recent studies referred to above accord with earlier studies
attributing the viable but non-culturable state to ROS (Bloomeld,
Stewart, Dodd, Botth, & Power, 1998; Dodd, Sharman, Bloomeld,
Booth, & Stewart, 1997; Imlay, 1995; Vartoukian, Palmer, & Wade,
7. Quantifying rates in predictive microbiology
To this point the discussion has been concentrated on aspects of
microbial physiology that affect growth and inactivation rates and
rates of recovery from the effects of environmental insults. From
this point we will move to consider predictive models more closely
in sections 8e10. Finally we will attempt to integrate modelling and
ecophysiology through the development of a mechanistic model,
and the enormous opportunities offered by omics technologies,
systems biology and bioinformatics.
The scheme proposed by Whiting and Buchanan (1993) which
classied predictive mathematical models in food microbiology as
primary, secondary or tertiary is logical and has stood the test of
time. Commonly used primary models that describe sigmoidal
bacterial growth curves include the Logistic, Gompertz and Baranyi
models. Details and comparisons of these and other primary
models were presented by McKellar and Lu (2004). Secondary
growth models describe the effect of environmental factors growth
or inactivation rate and, while most attention has been given to
modelling temperature dependence, multi-factorial models have
been developed and the paper by Mejlholm et al. (2010) is in the
must read category as it evaluates the performance of such
models. These include one in which the combined effects of up to
nine environmental factors on the growth of L. monocytogenes were
considered. Ross and Dalgaard (2004) provided a comprehensive
account of secondary models to that time.
Whiting and Buchanan (1993) dened tertiary models as algorithms incorporated into software to integrate the effect of environmental variables on microbial responses and to provide
predictions of the outcomes. The most developed tertiary models
are all now available via the internet and include
 the Pathogen Modeling Program (
 SymPrevius (,
 Seafood Spoilage and Safety Predictor (
 ComBase (
The Microbial Resources Viewer (MRV: http://mrv.nfri.affrc.go.
jp) was developed by the National Food Research Institute, Japan
(Koseki, 2009) to complement ComBase. Both the SSSP and the
MRV provides growth/no growth data sets with growth rate

contour plots. This allows users to identify conditions at the

growth/no growth boundary visually as well as quantitatively and
the user friendly visual feature suggests that Koseki (2009) should
also be placed in the must read category.
8. Modelling the effects of temperature on growth
As noted above, temperature has often been the primary factor
to measure in predictive microbiology because it is the factor likely
to uctuate to the greatest extent as product moves through the
supply chain. The major competitors for interpreting the effects of
temperature on microbial ecology in foods are Arrhenius-type and
lehradk-type models. Whilst the goodness-of-t of these
models have been compared in many studies, differences are often
minimal. The conclusion is that both logarithmic and square root
transformations of data are very effective in homogenising variance. Subsequently, arguments may arise about the relative merits
of a model on the basis of other criteria. For example does the
model have a mechanistic basis? Arrhenius-type models have their
origin in chemical reaction kinetics and this has been sufcient to
endow them with a mechanistic aura despite the fact that they do
not provide an adequate description of biological reaction kinetics.
This view is encapsulated in the title of a paper by Knies and
Kingsolver (2010) Erroneous Arrhenius: modied Arrhenius model
best explains the temperature dependence of ectotherm tness.
Differences of opinion on the goodness-of-t of models also
arise within a particular model type. Huang (2010, 2011) described
a new temperature dependence model for bacterial growth. The
motivation for the study was the discrepancy between Tmin, the
theoretical minimum temperature for growth and the actual
minimum temperature at which growth is observable (MINt). The
author posited that Tmin should reect biological reality (i.e., MINt)
and, in an attempt to make the estimates similar or the same,
changed the exponent of the square-root model from 2.0 to 1.5.
Furthermore, the idea of a theoretical minimum temperature has
an important and universally accepted precedent in science, the
concept of absolute zero, introduced by Lord Kelvin.
To assist in preventing further misinterpretation Ross, Olley,
McMeekin, and Ratkowsky (2011a, 2011b) proposed to retain the
current notation, Tmin, for the theoretical minimum temperature
and introduced the abbreviation MINt for the measurable
minimum temperature. Despite the clear denitions of Tmin and
Tmax by McMeekin, Olley, Ross, and Ratkowsky (1993), uncertainty
about the denition of Tmin as a theoretical temperature was also
evident in the advent of cardinal temperature models. These were
rst proposed by Lobry, Rosso, and Flandrois (1991), Rosso, Lobry,
and Flandrois (1993), and Rosso, Lobry, Bajard, and Flandrois
(1995) in each of which Tmin was dened as the minimum growth
temperature and Tmax as the maximum growth temperature.
Bajard, Rosso, Fardel, and Flandrois (1996) presented the rst report
from the Rosso group in which Tmin and Tmax were correctly dened
as the minimal temperature where growth can theoretically not be
observed and Tmax as the maximal growth temperature above
which growth is theoretically not possible.
Bajard et al. (1996) highlighted the unexpected growth rate
response of L. monocytogenes at suboptimal temperatures. Whilst
E. coli, Clostridium botulinum and P. aeruginosa growth rates were
well described by the square root model at suboptimal temperatures, the response of L. monocytogenes deviated from linearity. A
change in temperature response kinetics was proposed between
10  C and 15  C and square root models were tted to data at
temperatures <10  C and >15  C. The Tmin value using the low
temperature rates was w5  C for both strains, while for the high
temperature range data the estimates were w4  C and 1  C. The
authors emphasised the need to use the low temperature range

T. McMeekin et al. / Food Control 29 (2013) 290e299

estimate of Tmin as the high temperature range estimate would lead

to fail dangerous predictions.
L. monocytogenes was offered and, to the best of our knowledge,
none has been advanced since. Here we propose that the response
is a function of the dual lifestyle of the organism. On one hand it is
an intracellular pathogen and, on the other an environmental
organism with remarkable survival capability. The physiological
basis for this duality may possibly be explained by thermosensitive riboswitches. Cossart (2011) reviewed transcriptional
complexity and RNA regulation in L. monocytogenes noting that
many virulence factors are regulated by PrfA under the control of
a thermosensor. This results in optimal PrfA expression at high
temperatures and translational repression at low temperatures
with maximum expression of virulence genes at 37  C. A thermosensitive link between virulence and nutrient availability in
the transition from saprophyte to parasite was also established by
Toledo-Arana et al. (2009). Conversely, agella activity, an
ecological advantage in saprophytic mode was greater at low
temperatures (Kamp & Higgins, 2009; Mauder, Williams, Fritsch,
Kuhn, & Beier, 2008). The majority of temperature dependence
studies on virulence have been conducted at more than
30  C whereas the change in growth rate noted by Bajard
et al. (1996) occurred at 10  Ce15  C. Data is needed to bridge
the temperature gap as L. monocytogenes moves from an
environmental to a parasitic lifestyle.
8.1. Why select an exponent of 2?
The square root model is a special case of Be
temperature function rst introduced in 1926, in which the expolehradk model became widely
nent is two (Ross, 1993). The Be
used in biology and exponents other than two have been suggested.
Spencer and Baines (1964) used an exponent of one for sh spoilage
but Olley and Ratkowsky (1973) showed that it was accurate only
over a narrow temperature range from 0  C to 6  C. Further work on
deterioration of sh was reported by Ohta and Hirahara (1977) who
found an exponent of two was suitable to describe the temperature
dependence of nucleotide degradation. In turn this led to the
application of square root kinetics to microbial growth rates
(Ratkowsky, Lowry, McMeekin, Stokes, & Chandler, 1983;
Ratkowsky, Olley, McMeekin, & Ball, 1982).
Subsequently, the square root model was extended to include
terms for water activity (McMeekin et al., 1987) and pH (Adams,
Little, & Easter, 1991). Importantly, this did not change selection
of 2 as the appropriate exponent or the estimate of Tmin but MINt
increased with increasing levels of NaCl and decreasing water
From the 1980s the great majority of power-law models have
used an exponent of 2 across several elds of biology. Examples
include Jefferies and Brain (1984), for pollen tube penetration; Li
(1984), for marine phytoplankton photosynthesis (for further
discussion on this see McMeekin et al., 1993, App. 4.5, p. 340); Smith
(1985), for coliform growth rates; Li and Dickie (1987), for marine
phytoplankton growth rate; Blankenship, Craven, Lefer, and
Custer (1988), for C. perfringens growth during cooling and
Grifths and Phillips (1988) for spoilage of milk.
8.2. Why have theoretical Tmin and Tmax values?
So, what are the advantages of the dichotomy of theoretical and
measurable minimum and maximum temperatures? Below we
suggest several, some practical and some that may add to the
development of theory in microbial ecophysiology (McMeekin
et al., 2008).


- The value of Tmin is xed and does not depend on other environmental variables such as water activity or pH. However,
MINt values vary markedly depending on other factors;
- A constant Tmin value provides a convenient way to construct
a relative rate curve, the basis of technologies to predict the
microbiological safety and stability of foods;
- There is a continuum of Tmin values describing the thermal
adaptation of bacteria across all temperature categories, psychrophiles, psychrotropes*, mesophiles and thermophiles;
- A constant Tmin leads to a simple multiplicative model to
describe the combined effect of several hurdles (the Gamma
- The form of the model suggests that the combined effect of
several hurdles is additive and not synergistic when the cells
are growing;
- Synergisitic effects occur at the growth/no growth interface
and beyond when cell physiology switches from growth to
survival mode;
- The gap between Tmin and MINt increases with progressively
harsher conditions due to energy diversion to deal with additional hurdles.
(*n.b., The term psychrotroph has gained almost universal approval
in the food microbiology literature to describe a cold tolerant
microorganism but the etymology of the word is incorrect as
etroph, from which psychrotroph is derived is a sufx referring to
nourishment. The correct sufx to indicate cold tolerant is etrope
or etropic referring to turning, as in heliotrope or geotropic).
9. The elements of a growth rate versus temperature plot for
the full biokinetic temperature range
The response of growth rate to temperature plotted as
the square root of growth rate versus temperature is shown in
Fig. 1. This has the advantage of homogenising the variance and

Fig. 1. Typical plot of the square root of the specic growth rate constant vs.
temperature. 1. Tmin e theoretical minimum temperature for growth; 2. MINt e actual
minimum temperature for growth; 2e3. Region of variability in growth rate estimates
due to the distribution of long response times; 4e5. Region of maximum enzyme
stability surrounding Tmes, the point of maximum enzyme stability (Ratkowsky et al.,
2005); 6. Point at which the rate of increase of the specic growth rate starts to
decline; 7. Point at which the maximum specic growth rate is reached. Growth rate
estimates are variable around this temperature; 8. Start of monotonic decline in
growth rate; 9e10. Region in which monotonic decline is disrupted and inactivation
occurs; 10. MAXt e actual maximum temperature for growth; 11. Tmax e theoretical
maximum temperature permitting growth.


T. McMeekin et al. / Food Control 29 (2013) 290e299

providing a linear response in the sub-optimal region. It also

emphasises the juxtaposition of Tmin and MINt and of Tmax and
9.1. Revelations from a temperature dependence plot
Several features of the growth curve require further comment.
The shaded regions of the curve indicate where estimates of growth
rate are variable, but for different reasons. Between points 2 and 3
variability arises from the distribution of generation times
(described by a gamma function) when the variance is proportional
to the mean response time squared (Ratkowsky, Ross, Macario,
Dommett, & Kamperman, 1996). At point 7 variable estimates are
explained by maximum demand on cell physiology to produce
sufcient proteins and other macromolecules to service the
requirements of new cells. The term optimal temperature may be
considered correct to describe the maximum rate of growth but is
a misnomer for optimal metabolic efciency, which is in the Tmes
region (points 4e5).
The apparently anomalous response at 9e10 where inactivation
occurs at temperatures less than MAXt is counter-intuitive but has
been reported by others (Cornet, Van Derlinden, Cappuyns, & Van
Impe, 2010; Niemela & Oivanen, 1992; Van Derlinden, Lule,
Bernaerts, & Van Impe, 2010). Physiological explanations were
not advanced and here we propose that the perturbation is caused
by rapid induction of the stringent response when the purpose of
cellular metabolism is suddenly and unequivocally reversed from
growth to survival.
Note also the skewed shape of the temperature dependence plot
which indicates that specic growth rates rise gradually to
a maximum and then decline rapidly in the superoptimal region.
The same pattern is also observed when rate is plotted against
temperature (Chen & Shakhnovich, 2010). The rapid decline is due
to increasing irreversible denaturation of proteins as MAXt is
approached and identies the superoptimal temperatures as
a region where large declines in rate are achieved by small shifts in

10. Moving towards a mechanistic model

Ratkowsky, Olley, and Ross (2005) reported on the similarity of
temperature dependence of bacterial growth and the stability of
globular proteins using a kinetic model incorporating thermodynamic terms for protein denaturation. The underlying concept
relied on a term for the positive change in heat capacity during
protein unfolding.
The thermodynamic model was developed using 35 data sets for
psychrophilic, psychrotropic, mesophilic and thermophilic bacteria
resulting in biologically meaningful estimates of the thermodynamic parameters. Subsequently the same 35 data sets were used
by Knies and Kingsolver (2010), Chen and Shakhnovich (2010),
Ghosh and Dill (2010) and Dell, Pawar, and Savage (2011). The
probable mechanism of protein unfolding (denaturation) is
a change in tertiary structure in response to exposure of hydrophobic amino acids to water in the cell.
The thermodynamic approach was further advanced by Corkrey,
Olley, Ratkowsky, McMeekin, and Ross (2012) using a Bayesian
hierarchical modelling approach to analyse 95 data sets (including
those considered earlier) drawn from the Archaea, Bacteria and
Eukarya, i.e., representing the three domains of life on earth. The
new model, as in Ratkowsky et al. (2005), was based on the master
reaction hypothesis and the results implied that temperature
dependent growth rates of all unicellular organisms respond
according to the assumptions of the model.

11. Bioinformatics e the highway to predictive Systems

To date predictive microbiology has been concerned with predicting the rates of biological processes, particularly microbial
growth and inactivation, almost exclusively using empirical
models. More recently we have seen development of mechanistic
models, such as the thermodynamically based temperature
dependence model proposed by Corkrey et al. (2012). We have also
experienced the stampede into omics-based technologies by
which enormous amounts of data are accumulated potentially
allowing detailed descriptions of the inputs and outputs and
control of complex metabolic pathways in cells. Thereby, the era of
systems biology has arrived but its application rests on being able
to deal effectively with the data deluge. Thus, biologists now need
to recruit the quantitative skills of bioinformaticists to provide the
link between data accumulation, organisation and storage, and
interpretation. This is exemplied by a trend in major journals to
spin off separate sections or, in some cases, new journals within
a publishing group to deal with Computational Biology.
Mechanistic models and systems biology sit naturally together
so it is appropriate that they should evolve in parallel. In striving for
a comprehensive description of metabolic co-ordination, how these
are affected by environmental conditions and how they are
controlled and ne-tuned by genetic and epigenetic events in the
cell, it is appropriate that our focus shifts from empirical, predictive
microbiology models to mechanistic, predictive systems biology
models. The latter have the potential to reveal unifying themes in
biological sciences.
The rate of publication of papers linking genomics, proteomics,
metabolomics etc. to microbial ecology and physiology is increasing
markedly and one of particular relevance to the content presented
here is that by Vieira-Silva and Rocha (2010) which dealt with the
implications of growth rate for ecological genomics and metagenomics. The research outcomes were summarised as follows:
Prokaryotes have evolved a set of genomic traits to grow fast,
including biased codon usage and transient or permanent gene
multiplication for dosage effects. .these traits can be used to
predict minimal generation times from the vast majority of
microbes that cannot be cultivated. .. this inference can also be
made with incomplete genomes and thus be applied to metagenomic data to test hypotheses about the biomass productivity of
biotypes and the evolution of microbes in the human gut after
birth. Our results also allow better understanding of evolution
between growth rates and genomic traits and how they can be
manipulated in synthetic biology. Growth rates have been a key
variable in microbial physiology studies in the last century, and we
show how intimately they are linked with genome organisation
and prokaryotic ecology. The authors summary makes a persuasive case that linking ecophysiology with genomics will advance
knowledge rapidly.
Vieira-Silva, Touchon, Abby, and Rocha (2011) also used proteomics to study the effect of growth rate on the evolutionary rates
of essential proteins showing that the relative frequency of highly
expressed proteins in the proteome increased steeply with growth
rate. The positive correlation between expression levels of highly
expressed proteins and growth rates was conrmed in 74 proteobacteria. Analyses of the ribosome and the replisome suggested
that growth-related sequence conservation is associated with
synthesis of highly expressed proteins, consuming a major part of
the cellular energy budget and leading to strong conservation of
sequences coding for highly expressed proteins. For other papers
using proteomics to predict evolutionary trends, ecological distribution and population physiology see Wang, Yafremava, CaetanoAnolles, Mittenthal, and Caetano-Anolles (2007) and Mueller

T. McMeekin et al. / Food Control 29 (2013) 290e299

et al. (2010). The use of proteomics in food safety research was

addressed by DAlessandro and Zolla (2012) and the contribution of
metabolomics and transcriptomics to unravel the stress response of
E. coli was demonstrated by Jozefczuk et al. (2010). Papp, Notebaart,
and Pal (2011) described systems-biology approaches for predicting
genomic evolution judging that the models: Hold the promise of
transforming evolutionary biology into a more predictive discipline with the ability to make specic and reliable predictions on
the outcome of metabolic evolution, both in short term evolution
and on macroevolutionary time scales. The genomic basis of
trophic strategy in marine bacteria was elucidated by Lauro et al.
(2009) and Raes, Letunic, Yamada, Jensen, and Bork (2011)
combined biochemical, geographical and metagenomic data to
develop a molecular trait-based view of ecology.
Finally, considerable concern has been expressed about the
effect of climate change on global patterns of biological diversity
and the search is on for predictors of diversity. Tittensor et al. (2010)
undertook this challenge in the case of marine biodiversity across
taxa. They concluded that across 13 major species groups from
zooplankton to marine mammals that sea surface temperature was
the only environmental predictor highly related to diversity across
all 13 taxa. So, we anticipate that the model of Corkrey et al. (2012),
proposing universal thermodynamic constraints to temperature
limits for microbial, based on protein stability, will nd utility in
assessing the effect of changing temperatures on microbial growth
rates, microbial diversity and species survival in marine, and
perhaps other, environments.
12. Conclusions
In the presentation given at 7ICPMF in September 2011 we were
unable to conclude unequivocally that successful integration of
traditional and futuristic predictive modelling was possible. Now,
after an extensive additional literature search, including many
papers published subsequent to 7ICPMF we are more condent that
this goal is achievable.
The rst reference cited in this paper, Stockbridge et al. (2010),
provided an estimate of time required for evolution of enzymes. So,
having started the paper by considering enzymes, which as
proteins, require to be folded correctly to function we need to
consider what drives the folding process. Is it purely under thermodynamic control or is there a kinetic element involved
(Govindarajan & Goldstein, 1996; Mallam & Jackson, 2012)? The
answer is that, while the rate of protein folding is accelerated by
molecular chaperonins it is essentially a spontaneous process that,
despite being slow proceeds without misfolding (Mallam &
Jackson, 2012). In turn, this conrms the view of the 1972 Nobel
Prize winner Annsen (Annsen, 1973) that the most probable
conformation of a folded protein in any given environment is
determined solely by its amino acid composition. Rothman and
Shekman (2011) highlighted the very signicant advance in
knowledge arising from elucidation of the role of molecular chaperonins in accelerating rates of protein folding noting that this work
was complementary to that of Annsen. Further, it is consistent
with our opening comments on the crucial part played by enzymes
in determining rates of biological evolution.
Now, the loop is nearly closed in that protein folding can be
described by a mechanistic, thermodynamically based, temperature dependence model (Corkrey et al., 2012) which can be applied
equally well to all three domains of unicellular life. At this juncture
we intend to extend our studies to determine if the effect of other
constraints on microbial growth, such a water activity and pH, can
be modelled thermodynamically and to extend temperature
dependence studies to multicellular life forms. Given similar
outcomes it seems reasonable to project widespread applications


for thermodynamically based models not only in microbiology, but

also in protein stability, thermal biology, ecological theory,
biogeochemistry, extremophile biology and climate science. We are
excited by these prospects and pleased that the transition from
empirical to mechanistic predictive models has its genesis, in part,
in food microbiology.
Tom McMeekin thanks the Organising Committee of 7ICPMF for
their kind invitation to attend the conference and their generosity
and hospitality. The authors are indebted to Vasilis Valdramidis for
his sound advice and considerable patience during preparation of
this manuscript. The authors also thank Professor Stanley Brul,
University of Amersterdam, for bringing to our attention the role of
the stringent response to explain phenomenological reports of nonintuitive behaviour in bacterial growth rates. Experimental work on
growth rate decline at super-optimal temperatures was supported
by the Geoffrey Gardiner Dairy Research Foundation, Melbourne,
Victoria, Australia.
Adams, M. R., Little, C. L., & Easter, M. C. (1991). Modeling the effect of pH, acidulant
and temperature on the growth rate of Yersinia enterocolitica. Journal of Applied
Bacteriology, 71(1), 65e71.
Anastasiou, D., Poulogiannis, G., Asara, J. M., Boxer, M. B., Jiang, J.-k., Shen, M., et al.
(2011). Inhibition of pyruvate kinase M2 by reactive oxygen species contributes
to cellular antioxidant responses. Science, 334(6060), 1278e1283.
Annsen, C. B. (1973). Principles that govern folding of protein chains. Science,
181(4096), 223e230.
Aravind, L., Tatusov, R. L., Wolf, Y. I., Walker, D. R., & Koonin, E. V. (1998). Evidence
for massive gene exchange between archaeal and bacterial hyperthermophiles.
Trends in Genetics, 14(11), 442e444.
Bajard, S., Rosso, L., Fardel, G., & Flandrois, J. P. (1996). The particular behaviour of
Listeria monocytogenes under sub-optimal conditions. International Journal of
Food Microbiology, 29(2e3), 201e211.
Balaban, N. Q., Merrin, J., Chait, R., Kowalik, L., & Leibler, S. (2005). Bacterial
persistence as a phenotypic switch. Science, 305(5690), 1622e1625.
Belenky, P., & Collins, J. J. (2011). Antioxidant strategies to tolerate antibiotics.
Science, 334(6058), 915e916.
Bigger, J. W. (1944). Treatment of staphylococcal infections with penicillin. Lancet, ii,
Blankenship, L. C., Craven, S. E., Lefer, R. G., & Custer, C. (1988). Growth of Clostridium perfringens in cooked chili during cooling. Applied and Environmental
Microbiology, 54(5), 1104e1108.
Bloomeld, S. F., Stewart, G. S. A. B., Dodd, C. E. R., Botth, I. R., & Power, E. G. M.
(1998). Comment: the viable but non-culturable phenomenon explained?
Microbiology, 144, 1e2.
Campbell, A. (1957). Synchronization of cell division. Bacteriological Reviews, 21(4),
Caro-Quintero, A., Deng, J., Auchtung, J., Brettar, I., Hoe, M. G., Klappenbach, J., et al.
(2011). Unprecedented levels of horizontal gene transfer among spatially cooccurring Shewanella bacteria from the Baltic Sea. ISME Journal, 5(1), 131e140.
Cashel, M. (1975). Regulation of bacterial ppGpp and pppGpp. Annual Review of
Microbiology, 29, 301e318.
Chattopadhyay, S., Feldgarden, M., Weissman, S. J., Dykhuizen, D. E., van Belle, G., &
Sokurenko, E. V. (2007). Haplotype diversity in source-sink dynamics of
Escherichia coli urovirulence. Journal of Molecular Evolution, 64(2), 204e214.
Chen, P. Q., & Shakhnovich, E. I. (2010). Thermal adaptation of viruses and bacteria.
Biophysical Journal, 98(7), 1109e1118.
Chesbro, W., Evans, T., & Eifert, R. (1979). Very slow growth of Escherichia coli.
Journal of Bacteriology, 139(2), 625e638.
Chesney, J. A., Eaton, J. W., & Mahoney, J. R. (1996). Bacterial glutathione: a sacricial
defense against chlorine compounds. Journal of Bacteriology, 178(7), 2131e2135.
Cooper, S. (1993). The origins and meaning of the Schaechter-Maale-Kjeldgaard
experiments. Journal of General Microbiology, 139, 1117e1124.
Corkrey, R., Olley, J., Ratkowsky, D., McMeekin, T., & Ross, T. (2012). Universality of
thermodynamic constants governing biological growth rates. PloS One, 7(2),
Cornet, I., Van Derlinden, E., Cappuyns, A. M., & Van Impe, J. F. (2010). Heat stress
adaptation of Escherichia coli under dynamic conditions: effect of inoculum
size*. Letters in Applied Microbiology, 51(4), 450e455.
Cossart, P. (2011). Illuminating the landscape of host-pathogen interactions with
the bacterium Listeria monocytogenes. Proceedings of the National Academy of
Sciences of the United States of America, 108(49), 19484e19491.
Costerton, J. W., Stewart, P. S., & Greenberg, E. P. (1999). Bacterial biolms:
a common cause of persistent infections. Science, 284(5418), 1318e1322.


T. McMeekin et al. / Food Control 29 (2013) 290e299

DAlessandro, A., & Zolla, L. (2012). We are what we eat: food safety and proteomics.
Journal of Proteome Research, 11, 26e36.
Davidson, P. M., & Harrison, M. A. (2002). Resistance and adaptation to food antimicrobials, sanitizers, and other process controls. Food Technology, 56(11),
Dell, A. I., Pawar, S., & Savage, V. M. (2011). Systematic variation in the temperature
dependence of physiological and ecological traits. Proceedings of the National
Academy of Sciences, 108(26), 10591e10596.
Dennehy, J. J., Friedenberg, N. A., McBride, R. C., Holt, R. D., & Turner, P. E. (2010).
Experimental evidence that source genetic variation drives pathogen emergence.
Proceedings of the Royal Society B-Biological Sciences, 277(1697), 3113e3121.
Dickinson, B. C., & Chang, C. J. (2011). Chemistry and biology of reactive oxygen
species in signaling or stress responses. Nature Chemical Biology, 7(8), 504e511.
Dodd, C. E. R., Sharman, R., Bloomeld, S. F., Booth, I. R., & Stewart, G. S. A. B. (1997).
Inimical processes: bacterial self-destruction and sub-lethal injury. Trends in
Food Science & Technology, 8(7), 238e241.
EFSA. (European Food Safety Authority). (). Assessment of the possible effect of the
four antimicrobial treatment substances on the emergence of antimicrobial
resistance. The EFSA Journal, 659, 1e26.
van Elsas, J. D., Chiurazza, M., Mallon, C. A., Elhottova, D., Kristufek, V., & Salles, J. K.
(2004). Microbial diversity determines the invasion of soil by a bacterial
pathogen. Proceedings of the National Academy of Sciences of the United States of
America, 101, 7897e7901.
Fauvart, M., De Grootet, V. N., & Michiels, J. (2011). Role of persister cells in chronic
infections: clinical relevance and perspectives on anti-persister therapies.
Journal of Medical Microbiology, 60(6), 699e709.
Gentry, D. R., Ingraham, K. A., Stanhope, M. J., Rittenhouse, S., Jarvest, R. L.,
OHanlon, P. J., et al. (2003). Variable sensitivity to bacterial methionyl-tRNA
synthetase inhibitors reveals subpopulations of Streptococcus pneumoniae
with two distinct methionyl-tRNA synthetase genes. Antimicrobial Agents and
Chemotherapy, 47(6), 1784e1789.
Ghosh, K., & Dill, K. (2010). Cellular proteomes have broad distributions of protein
stability. Biophysical Journal, 99(12), 3996e4002.
Gilbert, P., Collier, P. J., & Brown, M. R. W. (1990). Inuence of growth-rate on
susceptibility to antimicrobial agents e biolms, cell-cycle, dormancy, and
stringent response. Antimicrobial Agents and Chemotherapy, 34(10), 1865e1868.
Gofn, P., van de Bunt, B., Giovane, M., Leveau, J. H. J., Hoppener-Ogawa, S.,
Teusink, B., et al. (2010). Understanding the physiology of Lactobacillus plantarum at zero growth. Molecular Systems Biology, 6.
Govindarajan, D., & Goldstein, R. A. (1996). Why are some protein structures so
common. Proceedings of the National Academy of Sciences, USA, 93, 3341e3345.
Grifths, M. W., & Phillips, J. D. (1988). Prediction of the shelf-life of pasteurized
milk at different storage temperatures. Journal of Applied Bacteriology, 65(4),
Guerra, C., Morris, D., Sipin, A., Kung, S., Franklin, M., Gray, D., et al. (2011). Glutathione and adaptive immune responses against Mycobacterium tuberculosis
infection in healthy and HIV infected individuals. PLoS One, 6(12), e28378.
Hamanaka, R. B., & Chandel, N. S. (2011). Warburg effect and redox balance. Science,
334(6060), 1219e1220.
Hehemann, J. H., Correc, G., Barbeyron, T., Helbert, W., Czjzek, M., & Michel, G.
(2010). Transfer of carbohydrate-active enzymes from marine bacteria to
Japanese gut microbiota. Nature, 464(7290), 908e912.
Hengge-Aronis, R. (2002). Signal transduction and regulatory mechanisms involved
in control of the sigma(S) (RpoS) subunit of RNA polymerase. Microbiology and
Molecular Biology Reviews, 66(3), 373e395.
Hermsen, R., & Hwa, T. (2010). Sources and sinks: a stochastic model of evolution in
heterogeneous environments. Physical Review Letters, 105(24), 248104.
Huang, L. (2010). Growth kinetics of Escherichia coli O157:H7 in mechanicallytenderized beef. International Journal of Food Microbiology, 140(1), 40e48.
Huang, L. (2011). A new mechanistic growth model for simultaneous determination
of lag phase duration and exponential growth rate and a new Be
model for evaluating the effect of temperature on growth rate. Food Microbiology, 28(4), 770e776.
Ihssen, J., & Egli, T. (2004). Specic growth rate and not cell density controls the
general stress response in Escherichia coli. Microbiology-SGM, 150, 1637e1648.
Imlay, J. A. (1995). A metabolic enzyme that rapidly produces superoxide, fumarate
reductase of Escherichia coli. Journal of Biological Chemistry, 270(34),
Jameson, J. F. (1962). A discussion of the dynamics of Salmonella enrichment.
Journal of Hygiene (Cambridge), 60, 193e207.
Jefferies, C. J., & Brain, P. (1984). A mathematical model of pollen-tube penetration
in apple styles. Planta, 160(1), 52e58.
Jozefczuk, S., Klie, S., Catchpole, G., Szymanski, J., Cuadros-Inostroza, A.,
Steinhauser, D., et al. (2010). Metabolomic and transcriptomic stress response of
Escherichia coli. Molecular Systems Biology, 6.
Kamp, H. D., & Higgins, D. E. (2009). Transcriptional and post-transcriptional
regulation of the GmaR antirepressor governs temperature-dependent control
of agellar motility in Listeria monocytogenes. Molecular Microbiology, 74(2),
Kastbjerg, V. G., Larsen, M. H., Gram, L., & Ingmer, H. (2010). Inuence of sublethal
concentrations of common disinfectants on expression of virulence genes in
Listeria monocytogenes. Applied and Environmental Microbiology, 76(1),

Kastbjerg, V. G., Nielsen, D. S., Arneborg, N., & Gram, L. (2009). Response of Listeria
monocytogenes to disinfection stress at the single-cell and population levels as
monitored by intracellular pH measurements and viable-cell counts. Applied
and Environmental Microbiology, 75(13), 4550e4556.
Kim, J.-S., Heo, P., Yang, T.-J., Lee, K.-S., Cho, D.-H., Kim, B. T., et al. (2011). Selective
killing of bacterial persisters by a single chemical compound without affecting
normal antibiotic-sensitive cells. Antimicrobial Agents and Chemotherapy, 55(11),
Kjeldgaard, N. O., Maale, O., & Schaechter, M. (1958). The transition between
different physiological states during balanced growth of Salmonella typhimurium. Journal of General Microbiology, 19(3), 607e616.
Knies, J. L., & Kingsolver, J. G. (2010). Erroneous Arrhenius: modied Arrhenius
model best explains the temperature dependence of ectotherm tness. American Naturalist, 176(2), 227e233.
Kolter, R. (1999). Growth in studying the cessation of growth. Journal of Bacteriology,
181(3), 697e699.
Koonin, E. V., Makarova, K. S., & Aravind, L. (2001). Horizontal gene transfer in
prokaryotes: quantication and classication. Annual Review of Microbiology, 55,
Koseki, S. (2009). Microbial responses viewer (MRV): a new ComBase-derived
database of microbial responses to food environments. International Journal of
Food Microbiology, 134(1e2), 75e82.
Laland, K. N., Sterelny, K., Odling-Smee, J., Hoppitt, W., & Uller, T. (2011). Cause and
effect in biology revisited: is Mayrs proximate-ultimate dichotomy still useful?
Science, 334(6062), 1512e1516.
Lange, R., & Hengge Aronis, R. (1991). Identication of a central regulator of
stationary-phase gene expression in Escherichia coli. Molecular Microbiology,
5(1), 49e59.
Lange, R., & Hengge-Aronis, R. (1994). The cellular concentration of the sigma(s)
subunit of RNA-polymerase in Escherichia coli is controlled at the levels of
transcription translation, and protein stability. Genes & Development, 8(13),
Lauro, F. M., McDougald, D., Thomas, T., Williams, T. J., Egan, S., Rice, S., et al. (2009).
The genomic basis of trophic strategy in marine bacteria. Proceedings of the
National Academy of Sciences of the United States of America, 106(37),
Levin, B. R., & Rozen, D. E. (2006). Opinion e non-inherited antibiotic resistance.
Nature Reviews Microbiology, 4(7), 556e562.
Lewis, K. (2001). Riddle of biolm resistance. Antimicrobial Agents and Chemotherapy, 45(4), 999e1007.
Li, W. K. W. (1984). Microbial uptake of radiolabeled substrates e estimates of
growth-rates from time course measurements. Applied and Environmental
Microbiology, 47(1), 184e192.
Li, W. K. W., & Dickie, P. M. (1987). Temperature characteristics of photosynthetic
and heterotrophic activities e seasonal variations in temperate microbial
plankton. Applied and Environmental Microbiology, 53(10), 2282e2295.
Lobkovsky, A. E., Wolf, Y. I., & Koonin, E. V. (2010). Universal distribution of protein
evolution rates as a consequence of protein folding physics. Proceedings of the
National Academy of Sciences, 107(7), 2983e2988.
Lobry, J. R., Rosso, L., & Flandrois, J.-P. (1991). A FORTRAN subroutine for the
determination of parameters condence in non-linear models. Binary, 4, 86e93.
Lund, E., & Kjeldgaard, N. O. (1972). Protein synthesis and formation of guanosinetetraphosphate. FEBS Letters, 26(1), 306.
McKellar, R. C., & Lu, X. W. (2004). Modeling microbial responses in foods. Boca Raton,
USA: CRC Press.
McMeekin, T., Bowman, J., McQuestin, O., Mellefont, L., Ross, T., & Tamplin, M.
(2008). The future of predictive microbiology: strategic research, innovative
applications and great expectations. International Journal of Food Microbiology,
128(1), 2e9.
McMeekin, T. A., Chandler, R. E., Doe, P. E., Garland, C. D., Olley, J., Putro, S., et al.
(1987). Model for combined effect of temperature and salt concentration/water
activity on the growth rate of Staphylococcus xylosus. Journal of Applied Bacteriology, 62(6), 543e550.
McMeekin, T. A., Olley, J., Ratkowsky, D. A., & Ross, T. (2011). Predictive microbiology
theory and application: is it all about rates? In E. Cummins, J. M. Frias, &
V. P. Valdramidis (Eds.), Seventh International Conference on Predictive Modelling
in Foods e Conference Proceedings (pp. 17e20) Dublin, Ireland: UCD, DIT,
McMeekin, T. A., Olley, J. N., Ross, T., & Ratkowsky, D. A. (1993). Predictive microbiology: Theory and application. Taunton, UK: Research Studies Press Ltd.
Mallam, A. L., & Jackson, S. E. (2012). Knot formation in newly translated proteins is
spontaneous and accelerated by chaperonins. Nature Chemical Biology, 8,
Mauder, N., Williams, T., Fritsch, F., Kuhn, M., & Beier, D. (2008). Response regulator
DegU of Listeria monocytogenes controls temperature-responsive agellar gene
expression in its unphosphorylated state. Journal of Bacteriology, 190(13),
Mayr, E. (1961). Cause and effect in biology. Kinds of causes, predictability, and
teleology are viewed by a practicing biologist. Science, 134(348), 1501.
Mejlholm, O., Gunvig, A., Borggaard, C., Blom-Hanssen, J., Mellefont, L., Ross, T., et al.
(2010). Predicting growth rates and growth boundary of Listeria monocytogenes e
an international validation study with focus on processed and ready-to-eat meat
and seafood. International Journal of Food Microbiology, 141(3), 137e150.
Meyer, B. (2006). Does microbial resistance to biocides create a hazard to food
hygiene? International Journal of Food Microbiology, 112(3), 275e279.

T. McMeekin et al. / Food Control 29 (2013) 290e299

Monod, J. (1942). Recherches sur la croissance des cultures bacterienne. Paris, France:
Herman and Cie.
Monod, J. (1949). The growth of bacterial cultures. Annual Review of Microbiology, 3,
Mueller, R. S., Denef, V. J., Kalnejais, L. H., Suttle, K. B., Thomas, B. C., Wilmes, P., et al.
(2010). Ecological distribution and population physiology dened by proteomics in a natural microbial community. Molecular Systems Biology, 6. http://
Mulcahy, L. R., Burns, J. L., Lory, S., & Lewis, K. (2010). Emergence of Pseudomonas
aeruginosa strains producing high levels of persister cells in patients with cystic
brosis. Journal of Bacteriology, 192(23), 6191e6199.
Navarro Llorens, J. M., Tormo, A., & Martinez-Garcia, E. (2010). Stationary phase in
gram-negative bacteria. FEMS Microbiology Reviews, 34(4), 476e495.
Neidhardt, F. C. (1999). Bacterial growth: constant obsession with dN/dt. Journal of
Bacteriology, 181(24), 7405e7408.
Nguyen, D., Joshi-Datar, A., Lepine, F., Bauerle, E., Olakanmi, O., Beer, K., et al. (2011).
Active starvation responses mediate antibiotic tolerance in biolms and
nutrient-limited bacteria. Science, 334(6058), 982e986.
Niemela, S. I., & Oivanen, P. (1992). Transition from growth to death in populations
of Klebsiella terrigena and other enterobacteriaceae in relation to temperature.
FEMS Microbiology Letters, 91(2), 181e185.
Ohta, F., & Hirahara, T. (1977). Rate of degradation in cool-stored carp muscle.
Memoirs Faculty of Fisheries Kagoshima University, 26, 97e102.
Olley, J., & Ratkowsky, D. A. (1973). Temperature function integration and importance in the storage and distribution of fresh food above the freezing point.
Food Technology Australia, 25, 66e73.
Papp, B., Notebaart, R. A., & Pal, C. (2011). Systems-biology approaches for predicting
genomic evolution. Nature Reviews Genetics, 12(9), 591e602.
Park, Y. J., & Chen, J. (2011). Inactivation of Shiga toxin-producing Escherichia coli
(STEC) and degradation and removal of cellulose from STEC surfaces by using
selected enzymatic and chemical treatments. Applied and Environmental
Microbiology, 77(24), 8532e8537.
Poelwijk, F. J., Kiviet, D. J., Weinreich, D. M., & Tans, S. J. (2007). Empirical tness
landscapes reveal accessible evolutionary paths. Nature, 445(7126), 383e386.
Poole, K. (2002). Mechanisms of bacterial biocide and antibiotic resistance. Journal
of Applied Microbiology, 92, 55Se64S.
Potrykus, K., Murphy, H., Philippe, N., & Cashel, M. (2011). ppGpp is the major source
of growth rate control in E. coli. Environmental Microbiology, 13(3), 563e575.
Price, P. B., & Sowers, T. (2004). Temperature dependence of metabolic rates for
microbial growth, maintenance, and survival. Proceedings of the National
Academy of Sciences of the United States of America, 101(13), 4631e4636.
Raes, J., Letunic, I., Yamada, T., Jensen, L. J., & Bork, P. (2011). Toward molecular traitbased ecology through integration of biogeochemical, geographical and metagenomic data. Molecular Systems Biology, 7, 374.
Ratkowsky, D. A., Lowry, R. K., McMeekin, T. A., Stokes, A. N., & Chandler, R. E. (1983).
Model for bacterial culture growth rate throughout the entire biokinetic
temperature-range. Journal of Bacteriology, 154(3), 1222e1226.
Ratkowsky, D. A., Olley, J., McMeekin, T. A., & Ball, A. (1982). Relationship between
temperature and growth-rate of bacterial cultures. Journal of Bacteriology,
149(1), 1e5.
Ratkowsky, D. A., Olley, J., & Ross, T. (2005). Unifying temperature effects on the
growth rate of bacteria and the stability of globular proteins. Journal of Theoretical Biology, 233(3), 351e362.
Ratkowsky, D. A., Ross, T., Macario, N., Dommett, T. W., & Kamperman, L. (1996).
Choosing probability distributions for modelling generation time variability.
Journal of Applied Bacteriology, 80(2), 131e137.
Rodionov, D. G., & Ishiguro, E. E. (1998). Temperature sensitivity of bacteriolysis
induced by lactam antibiotics in amino acid-deprived Escherichia coli. Journal of
Bacteriology, 180(8), 2224e2227.
Rolfe, M. D., Rice, C. J., Lucchini, S., Pin, C., Thompson, A., Cameron, A. D. S., et al.
(2012). Lag phase is a distinct growth phase that prepares bacteria for exponential growth and involves transient metal accumulation. Journal of Bacteriology, 194(3), 686e701.
lehradk-type models. Journal of Industrial Microbiology, 12(3e5),
Ross, T. (1993). Be
Ross, T., & Dalgaard, P. (2004). Secondary models. In R. C. McKellar, & X. Lu (Eds.),
Modeling microbial responses in foods (pp. 63e150). Boca Raton: CRC Press.
Ross, T., Olley, J., McMeekin, T. A., & Ratkowsky, D. A. (2011a). Some comments on
Huang, L (2010). Growth kinetics of Escherichia coli O157:H7 in mechanicallytenderized beef. International Journal of Food Microbiology, 140: 40e48.
International Journal of Food Microbiology, 147(1), 78e80.
Ross, T., Olley, J., McMeekin, T. A., & Ratkowsky, D. A. (2011b). Reply to: response to
letter to the editor: growth kinetics of Escherichia coli O157:H7 in mechanically
tenderized beef. International Journal of Food Microbiology, 140: 40e48.
International Journal of Food Microbiology, 147(1), 83e84.
Rosso, L., Lobry, J. R., Bajard, S., & Flandrois, J. P. (1995). Convenient model to
describe the combined effects of temperature and pH on microbial growth.
Applied and Environmental Microbiology, 61(2), 610e616.
Rosso, L., Lobry, J. R., & Flandrois, J. P. (1993). An unexpected correlation between
cardinal temperatures of microbial-growth highlighted by a new model. Journal
of Theoretical Biology, 162(4), 447e463.


Rothman, J. E., & Shekman, R. (2011). Molecular mechanism of protein folding in the
cell. Cell, 146(6), 851e854.
Rowe, W., Wedge, D. C., Platt, M., Kell, D. B., & Knowles, J. (2010). Predictive models
for population performance on real biological tness landscapes. Bioinformatics,
26(17), 2145e2152.
Sarubbi, E., Rudd, K. E., & Cashel, M. (1988). Basal ppGpp level adjustment shown by
new spot mutants affect steady-state growth rates and ribosomal RNA ribosomal promoter regulation in Escherichia coli. Molecular & General Genetics,
213(2e3), 214e222.
Schaechter, M. (2006). From growth physiology to systems biology. International
Microbiology, 9(3), 157e161.
Schaechter, M., Maale, O., & Kjeldgaard, N. O. (1958). Dependency on medium and
temperature of cell size and chemical composition during balanced growth of
Salmonella typhimurium. Journal of General Microbiology, 19(3), 592e606.
Schroeder, B. O., Wu, Z., Nuding, S., Groscurth, S., Marcinowski, M., Beisner, J., et al.
(2011). Reduction of disulphide bonds unmasks potent antimicrobial activity of
human beta-defensin 1. Nature, 469(7330), 419e423.
Shah, D., Zhang, Z. G., Khodursky, A., Kaldalu, N., Kurg, K., & Lewis, K. (2006).
Persisters: a distinct physiological state of E. coli. BMC Microbiology, 6, 53. http://
Shatalin, K., Shatalina, E., Mironov, A., & Nudler, E. (2011). H(2)S: a universal defense
against antibiotics in bacteria. Science, 334(6058), 986e990.
Silver, L. L. (2011). A persistent problem. Journal of Medical Microbiology, 60(3),
Smillie, C. S., Smith, M. B., Friedman, J., Cordero, O. X., David, L. A., & Alm, E. J. (2011).
Ecology drives a global network of gene exchange connecting the human
microbiome. Nature, 480(7376), 241e244.
Smith, M. G. (1985). The generation time, lag time, and minimum temperature of
growth of coliform organisms on meat, and the implications for codes of
practice in abattoirs. Journal of Hygiene, 94(3), 289e300.
Sokurenko, E. V., Gomulkiewicz, R., & Dykhuizen, D. E. (2006). Opinion e sourcesink dynamics of virulence evolution. Nature Reviews Microbiology, 4(7),
Spencer, R., & Baines, C. R. (1964). Effect of temperature on spoilage of wet white
sh. I. Storage at constant temperatures between e 1 and 25 C. Food Technology, 18(5), 769e772.
Spoering, A. L., & Lewis, K. (2001). Biolms and planktonic cells of Pseudomonas
aeruginosa have similar resistance to killing by antimicrobials. Journal of
Bacteriology, 183(23), 6746e6751.
Stecher, B., Denzler, R., Maier, L., Bernet, F., Sanders, M. J., Pickard, D. J., et al. (2012).
Gut inammation can boost horizontal gene transfer between pathogenic and
commensal Enterobacteriaceae. Proceedings of the National Academy of Sciences,
109(4), 1269e1274.
Stockbridge, R. B., Lewis, C. A., Jr., Yuan, Y., & Wolfenden, R. (2010). Impact of
temperature on the time required for the establishment of primordial
biochemistry, and for the evolution of enzymes. Proceedings of the National
Academy of Sciences of the United States of America, 107(51), 22102e22105.
Tittensor, D. P., Mora, C., Jetz, W., Lotze, H. K., Ricard, D., Vanden Berghe, E., et al.
(2010). Global patterns and predictors of marine biodiversity across taxa.
Nature, 466(7310), 1098e1107.
Toledo-Arana, A., Dussurget, O., Nikitas, G., Sesto, N., Guet-Revillet, H., Balestrino, D.,
et al. (2009). The Listeria transcriptional landscape from saprophytism to
virulence. Nature, 459(7249), 950e956.
Treangen, T. J., & Rocha, E. P. C. (2011). Horizontal transfer, not duplication, drives
the expansion of protein families in Prokaryotes. PLoS Genetics, 7(1).
Van Derlinden, E., Lule, I., Bernaerts, K., & Van Impe, J. F. (2010). Quantifying the
heterogeneous heat response of Escherichia coli under dynamic temperatures.
Journal of Applied Microbiology, 108(4), 1123e1135.
Vartoukian, S. R., Palmer, R. M., & Wade, W. G. (2010). Strategies for culture of
unculturable bacteria. FEMS Microbiology Letters, 309(1), 1e7.
Vieira-Silva, S., & Rocha, E. P. C. (2010). The systemic imprint of growth and its uses
in ecological (meta) genomics. PLoS Genetics, 6(1), e1000808.
Vieira-Silva, S., Touchon, M., Abby, S. S., & Rocha, E. P. C. (2011). Investment in rapid
growth shapes the evolutionary rates of essential proteins. Proceedings of the
National Academy of Sciences of the United States of America, 108(50),
Wang, X., & Wood, T. K. (2011). Toxin-antitoxin systems inuence biolm and
persister cell formation and the general stress response. Applied and Environmental Microbiology, 77(16), 5577e5583.
Wang, M., Yafremava, L. S., Caetano-Anolles, D., Mittenthal, J. E., & CaetanoAnolles, G. (2007). Reductive evolution of architectural repertoires in proteomes
and the birth of the tripartite world. Genome Research, 17(11), 1572e1585.
Whiting, R. C., & Buchanan, R. L. (1993). A classication of models in predictive
microbiology e reply. Food Microbiology, 10(2), 175e177.
de Wit, R., & Bouvier, T. (2006). Everything is everywhere, but, the environment
selects; what did Baas Becking and Beijerinck really say? Environmental
Microbiology, 8, 755e758.
Yamaguchi, Y., & Inouye, M. (2011). Regulation of growth and death in Escherichia
coli by toxin-antitoxin systems. Nature Reviews Microbiology, 9(11), 779e790.
Zhang, Q. C., Lambert, G., Liao, D., Kim, H., Robin, K., Tung, C. K., et al. (2011).
Acceleration of emergence of bacterial antibiotic resistance in connected
microenvironments. Science, 333(6050), 1764e1767.