Appl Microbiol Biotechnol (1990) 33:72-75
Applied
Water activity: a fundamental parameter
of aroma production by microorganisms
Patrick Gervais
Laboratoire de Biologie Physico-Chimique, Ecole Normale Sup~rieure de Biologie Appliqu~e fi la Nutrition et fi l'Alimentation,
Campus Universitaire Montmuzard, F-21000 Dijon, France
Received 19 October 1989/Accepted 8 December 1989
Summary. The water activity of the medium, which is
analogous to osmotic pressure in liquid medium, is a
fundamental parameter for the mass transfer of water
and solutes across the cell membrane. The control of
this parameter could be used to modify the metabolic
production or excretion of a microorganism, as demon-
strated in this work for aroma production by a fungus
and a yeast.
Introduction
The control of numerous parameters such as tempera-
ture, pH and oxygen concentration is generally re-
quired for submerged and solid state fermentations, al-
though the basic mechanisms of their biological actions
are not well understood. An increase in temperature
will modify molecular configurations by breaking van
der Waals or hydrogen bounds and so will lead to den-
aturation of native biological molecules such as en-
zymes, proteins, and DNA. This action is well de-
scribed by the Arrhenius equation for enzymes but the
application of this equation to the whole microorgan-
isms is not valid. The heat transfer properties of differ-
ent parts of microorganisms are not well-known, and
these properties vary with temperature. Moreover, there
is an important interaction between temperature and
water activity dependence of microorganisms due to
the calorific property of water.
A kinetic model which relates the rate constant of
death of microbial cells to water activity (aw) and tem-
perature has been proposed (Moser 1988) using the fol-
lowing equation:
EA "aw
k = k = .aw-exp RT
where the constants k~ and EA are calculated from the
experimental values of aw. MacMeckin et al. (1987)
have proposed a kinetic approach to the temperature
Offprint requests to: P. Gervais
.. Microbiology
Biotechnology
© Springer-Verlag 1990
dependence of microorganisms. Such kinetic models
have also been developed for pH and O2 effects on mi-
croorganisms (Andreeva and Biryukov 1973). In spite
of the lack of structured information, all these parame-
ters undoubtedly have an influence on cell metabolism
and can be measured and controlled by using existing
acute sensor such as thermometers, pH meters and oxy-
meters.
The effect of the water activity or water potential of
a medium on microorganisms is quite well known.
There is no active transport of water in the cell and wa-
ter moves as a function of the gradient of water poten-
tial between the intra- and extracellular medium. Water
activity in homogeneous solutions is related to osmotic
pressure in liquid media by a logarithmic equation
(Griffin 1981) and is determined by molecular solute-
water interactions. In heterogeneous media, water ac-
tivity is related to the solute concentration, capillary
forces, and absorption properties of the insoluble solid
substratum.
An increase in the osmotic pressure (or decrease in
the water activity for a heterogeneous medium) of ex-
tracellular medium leads to a two-phase cell response
(Zimmermann 1978; Steudle et al. 1983): (a) a very fast
but passive exit of cell water for a few seconds, with a
corresponding decrease in cell volume for protoplasts
or decrease in tugor pressure for cells with walls; (b) a
second compensation phase during which the cell al-
lows the input of permeable solutes or the biosynthesis
of intracellular metabolites. This accumulation in re-
sponse to a hydric stress has been identified in microor-
ganisms, plant and animal cells (Gilles 1975; Hellebust
1976; Luard 1982). The solutes synthesized, such as
proline may interfere with enzymatic systems that mod-
ify the Na + pump and thus the cell membrane permea-
bility. Other solutes, such as polyols, may not interfere
with the enzymatic systems of the cell, and act only by
increasing the cell osmotic pressure.
Numerous experiments have demonstrated the in-
fluence of water activity on development and metabol-
ism of microorganisms on both liquid and solid media
(Scott 1957; Mossel 1975; Troller 1980). In all cases,
 73

optimum growth and metabolic production occurred at =~

media a~ values o f less than 1. It has been reported that ~
low water activity increases the maintenance metabol- -~ 12-
~:-
ism o f cells (Esener et al. 1981) and, using experimental
data, we have recently developed a structural model " 10-
(Gervais et al. 1988a) that expresses growth rate as a ~
function of water activity o f the medium. This model "~ 8-
can be applied to moulds, yeasts or bacteria cultivated ~
on liquid or solid media. "~ 6-
In spite o f the importance of water activity and due -~
to the lack o f a sensor, no control of water potential or 7 4-
osmotic pressure in order to optimize metabolic pro- o
duction in classical fermentation processes has been c
2-
developed. Only a few applications should be noted, ~
such as for glycerol production by Dunaliella (Avron =
..~

~ ' "~ . . . . . f i ~ ..... ~ 1

and Ben-Amotz 1979), or for glutamic acid production ~. 0.94 0.95 0.96 0.97 0.98 0,99 1.00
by Corynebacterium glutamicum melassecola (Dubreuil ~
a~
1985). Mattiasson and Hahn-H~igerdal (1982) have pro- ~
posed a model for immobilized cells in which de- <
creased water activity results in new metabolic behav- Fig. 1. Influence of water activity (aw) on aroma production ([])
iour initiated by a changed balance between reduced and on aroma release (Q) by Sporidiobolus salmonicolor cultivated
for 20 days on liquid medium: the shaded area corresponds with
nicotinamide adenine dinucleotide and reduced nico- accumulated aroma
tinamide adenine dinUcleotide phosphate.
The influence of aw on aroma production was stud-
tivation, the relative humidity of the circulating air was fixed at
ied on two systems: the.methyl ketone production of a 80%. Thus, the aerial part of the mycelium was subjected to an
filamentous fungus, Trichoderma viride, and the lactone immediate stress.
production of a yeast, Sporidiobolus salmonicolor. Such
a study allows the use o f aw as a controlling factor o f
aroma production. Results

A r o m a production versus aw of the liquid or solid me-

Materials and methods
dia are shown in Figs. 1 and 2. In both cases, produc-
Culture in liquid medium. 7/-Decalactone production by the yeast tion values were dramatically affected by the water ac-
S. salmonicolor F and T was studied in 250-ml flasks and in a 2-1 tivity variation of the media. In the liquid medium, 7/-
fermentor. This lactone has a very low volatility so the main part decalactone production was high for aw = 0.99-0.97 and
of the production stays in the culture medium previously reported markedly decreased above and below this range. In the
(Gervais and Battut 1989). In order to extract F-decalactone, the solid medium, 2-heptanone production increased as the
culture medium was centrifuged for t0 rain at 4000 rpm. Fifty mil-
aw value of the medium decreased. Such results were
lilitres of this medium were then filtered through a separative li-
quid chromatographic non-polar column (Seppak C18, Waters, obtained by analysis o f the extracellular m e d i u m and
Milford, Mass, USA). The column was then eluted with 2 ml hex- resulted from the excretion availability of the cell,
ane before chromatographic analysis. which combines intracellular production and permea-
In order to generate a hydric stress in the liquid medium, a bility properties.
process of successive dilutions was set up in flasks as well as in
the fermentor. A preliminary experiment fixed the dilution time at
I8-
3 h in order to prevent interaction of the growth of the microor-
ganism with the aroma analysis. ~ 16~
Culture on gelose. 2-Heptanone production by the fungus T. viride
TS was studied in Roux flasks on a gelosed medium previously
quoted (Gervais et al. 1988b); continuous aeration (50 ml/h) per-
mitted analysis of the highly-volatile heptanone produced (Tallu ,- I0-
~ ,
1986). In the present study, sequential headspace chromatograph- g ~
ic analysis was used (Gervais et al. 1988b).
Description of the chromatographic analyser, which was the
same for both aromas, and the chromatographic conditions have
o
~. 6 i
been reported previously (Gervais et al. 1988b). The selected wa-
ter activity values in both media were obtained using glycerol as a
depressor. The fixed a~ values were verified at the beginning and
°i
~ 4
~ 2~

end of the culture using a hygrometer (Novasina, Zurich, Switzer- ~ ~ ~

~ ~ ~ ~ ~ i
land) and an osmometer (Roebling, Berlin, FRG). Variations in 0.960 0.965 0.970 0.975 0.9~0 0.985 0.990
water activity during incubation never exceeded 0.01 for all cul-
a~
tures studied, due to metabolism and evaporation.
In the solid substratum the effects of hydric stress were stud- Fig. ~. Influence of a~. on aroma production by Trichoderma viride
ied on a culture of T. viride TS initially at aw = 0.99. After 5 h cul- TS cultivated for 5 days on solid medium: DM: dry matter
74
II lI [I
"~-¢~03 "~'~ "!
~ ~
I -~ -~
-~ ~ This decrease corresponded to enzyme inhibition that
=~ 36 - =
8
~ . ~I"
~O 34
~-
~ I
~::
~- ~:
&
~C
Time (hour)
Fig. 3. Influenceof hydric stress on aroma production by T. viride
TS: ~, with stress; rq, without stress. The s h a d e d area corre-
sponds with accumulated aroma; R.H.: relative humidity
In order to distinguish between these two proper-
ties, the intracellular aroma content was examined in
further experiments. An initial osmotic phase of water
input occurred when cells were placed in hypotonic
media. In order to prevent mechanical damage to the
membrane due to volume expansion, a second phase of
active osmotic molecule output also occurred. The ac-
cumulation intensity was evaluated by the difference
between the quantity of extracellular aroma before and
after the hypotonic stress. This method is not quantita-
tive and gives only a fraction of the intracellular mole-
cule content.
Results presented in Fig. 3 for the solid medium
show that a decrease in water potential caused a drastic
decrease in extracellular heptanone production. The
hydration properties of the medium, especially water
activity, are therefore involved in aroma production by
the fungus. The return to the initial hydration condi-
tions after 10 h aeration at a relative humidity of 80%
corresponded to an increase in aroma production. Ac-
cumulation of the aroma during the stress period in re-
lation to osmoregulation by the fungus is represented
by the shaded area.
For the liquid medium, intracellular accumulation
was studied by diluting the media. Each initial culture
medium (from aw --- 0.94 to aw ~ 1) was diluted with wa-
ter to reach a comparative level of water activity (-~ 1).
In all cases this dilution led to an increase in the release
of ~,-decalactone, especially for the values of 0.97-0.99
as shown in Fig. 1. In this range the same 7/-decalactone
quantity was produced by the yeasts with maximum ac-
cumulation occurring at aw =0.97, which corresponds
to severe osmotic conditions. The release of aroma due
to dilution decreased for high aw values ( 2 1) as well as
for low water activity values (0.94).
The osmotic effect of the accumulation of aroma in-
side the cells has recently been demonstrated by experi-
mental evidence showing the presence of this com-
pound inside the cell. This was done by grinding the
cells with a glass bead cell homogenizer. It may be con-
cluded that most 7/-decalactone released by the cell oc-
curred when the water activity ranged between 0.97 and
0.99 and decreased both above and below this range.
could be induced by too high as well as by too low a
water potential level, as previously reported (Grajek
and Gervais 1987).
Discussion
In both types of media it seems that the extracellular
aroma production rate is influenced by the amount of
hydration of the medium. When the external osmotic
pressure is increased, cells have to equilibrate their in-
ner medium with the outer medium to prevent osmotic
stress due to passive exit of water. Three ways are pos-
sible: facilitating diffusion of small solutes such as
glycerol across the membrane, synthesizing molecules
in the inner medium or modifying the permeability of
the membrane. Heptanone and 7/-decalactone synthesis
can be related to this process of osmoregulation by liv-
ing cells. However, it has been verified that 7/-decalac-
tone accumulation only plays a very minor role in this
osmoregulation. The maximum intracellular concentra-
tion of this compound could be evaluated to 200 ppm
(v/v), which is about 1000 times smaller than the quan-
tity necessary to decrease the intracellular aw from
about 1 to 0.97. So the main part of the osmoregulation
must be assumed to be by the synthesis or transport of
other molecules such as salts, amino acids or polyols.
The observed variation of 7/-decalactone accumula-
tion in the aw value range 0.97-0.99 may only be due to
variation in cell permeability under the influence of aw.
It may also be concluded that maximum cell release of
heptanone and ~,-decalactone occurred when the aw
value of the external medium was low enough to gener-
ate an osmotic gradient with the inner cell medium, but
is not too low to inhibit enzyme kinetics.
Variation in water activity in liquid or solid fermen-
tation media could lead to great modifications in intra-
cellular accumulation and in extracellular excretion of
aromas produced by yeasts or moulds. Enzyme inhibi-
tion and cell membrane permeability variation are in-
volved in such mechanisms. This conclusion may be of
great interest for optimizing metabolite production
such as aromas, antibiotics, and enzymes by the mean
of aw control. A recently-developed sensor that allows
continuous measurement of the osmotic pressure of a
liquid medium could be used with this aim (Gervais
1989).
It could be of interest to maintain a high cell meta-
bolites concentration during a concentrated medium
step of a culture on solid as well as on liquid substrata
in order to release the aroma during a following diluted
step. The use of permeabilising agents might also be en-
visaged.

References
Andreeva LN, Biryukov VV (1973) Analysis of mathematical
models of the effect of pH on fermentation processes and their
use for calculating optimal fermentation conditions. Biotech-
nol Bioeng Symp 4:61-76
Avron M, Ben-Amotz A (1979) Metabolic adaptation of the alga
Dunaliella to low water activity. In: Shilo M (ed) Strategies of
microbial life in extreme environments. Dahlem Konferenzen,
Berlin pp 83-91
Dubreuil P (1985) Cin+tiques et mod~lisation de la fermentation
glutamiqueo Thbse de Docteur-Ing~nieur, Institut National Po-
lytechnique de Lorraine, Nancy
Esener A, Bol G, Kossen N, Roels JA (1981) Effect of water activ-
ity on microbial growth. In: Moo Young M, Robinson CW,
Vezina C (eds) Advances in Biotechnology, vol 1. Pergamon
Press, Toronto, pp 339-344
Gervais P (1989) New sensor allowing continuous water activity
measurements of submerged or solid-substrate fermentations.
Biotechnol Bioeng 33:266-271
Gervais P, Battut G (1989) Influence of water potential on ~'-
decalactone production by the yeast Sporidiobolus salmonico-
lor. Appl Environ Microbiol 55, 11:2939-2943
Gervais P, Grajek W, Bensoussan M, Molin P (1988a) Influence
of the water activity of a solid substrate on the growth rate and
sporogenesis of filamentous fungi. Biotechnol Bioeng 31:457-
463
Gervais P, Belin JM, Grajek W, Sarrette M (1988b) Influence of
water activity on the aroma production by Trichoderma viride
growing on solid substrate. J Ferment Technol 66:403-407
Gilles R (1975) Mechanisms of ion and osmoregulation. In:
Kinne O (ed) Marine ecology, vol 2, part 1. Wiley-Inter-
science, Chichester, pp 259-347
Grajek W, Gervais P (1987) Influence of water activity on the en-
zyme biosynthesis and enzyme activities produced by Trichod-
erma viride TS in solid-state fermentation. Enzyme Microb
75
Technol 9: 658-662
Griffin DM (1981) Water and microbial stress. Adv Microb Ecol
5:91-136
Hellebust JA (1976) Osmoregulation. Ann Rev Plant Physiol
27:485-505
Luard EJ (1982) Growth and accumulation of solutes by Phytoph-
thora cinnamomi and other lower fungi in response to changes
in external osmotic potential. J Gen Microbiol 128:2583-
2590
MacMeckin T, Chandler RE, Doc PE, Garland CD, Olley J, Pu-
tros S, Rotkowsky DA (1987) Model for combined effect of
temperature and salt concentration/water activity on the
growth rate of Staphylococcus xylosus. J Appl Bacteriol
62: 543-550
Mattiasson B, Hahn-Hagerdal B (1982) Microenvironmental ef-
fects on metabolic behaviour of immobilized cells. A hypothe-
sis. Eur J Appl Microbiol Biotechnol 16:52-55
Moser A (1988) Temperature dependence, water activity, and en-
thalpy/entropy compensation. In: Bioprocess technology, ki-
netics and reactors. Springer-Verlag, Berlin, Heidelberg, New
York, pp 198-204
Mossel DAA (1975) Water and microorganisms in foods. In:
Duckworth RB (ed) Water relations of foods. Academic Press,
New York, pp 347-361
Scott WJ (1957) Water relations of food spoilage microorganisms.
Adv Food Res 7:83±127
Steudle E, Tyerman SD, Wendler S (1983) Water relations of
plant cells. In: Marcelle R, Clijsters H, Poucke M van (eds)
Effects of stress on phytosynthesis. Martinus Nijhoff, The Ha-
gue, pp 95-109
Tallu B (1986) Production d'ar6me. DEA Sciences des Aliments,
Universit6 Clermont-Ferrand
Troller JA (1980) Influence of water activity on microorganisms in
foods. Food Technol 34:76-82
Zimmermann U (1978) Physics of turgor and osmoregulation.
Ann Rev Plant Physiol 29:121-148