Example: An Age Distribution Model for the Production of Antibiotics (A Segregated

Model)
Secondary metabolites, including most fungal antibiotics, are either produced late in
the growth phase (the idiophase) or in the stationary phase and are often classified as
non-growth associated products. Antibiotics produced by unicellular microorganisms
usually only appear when the cell ceases dividing. When the organism is filamentous,
secondary metabolites are formed even though the cell dry weight may still be increasing.
This occurs because cells located in filaments may no longer be dividing, have entered the
iodiophase and produce the desired metabolite, while cells at the tips ofthe hyphae continue
to grow and are do not produce product. Examples of this behavior include production of
antibiotics such as rifamycin, the polyene anitbiotics, and chloramphenicol.
One simple approach to modelling the production of such secondary metabolites,
which does not require detailed information of the complex metabolic pathways in volved,
is to incorporate the concept of cell age into the model. We can consider two age groups of
cells: "mature" cells capable of product synthesis, and "immature" cells unable to form
product37• The growth rate of mature cells is described by Monod-type k.inetics (with con-
stants a, p, whereas immature cells age at a first-order rate (a 12) to form mature cells. A
schematic of the growth and maturation process is given in Figure 3.29.

~ lmmature cell

\-·-·~·..
-e
.---p;;~~Ct , ''M ~ ~
Mature cell

lmmature cells
cell
formed doubling

Figure 3.29. The fonnation of mature cells by a first-order aging process. Mature cells
undergo fission to produce immature cel/s.

Denoting the immature cell concentration as X 1, the concentration of mature cells by

X2, and the product by P, the following set of mass balances for batch culture can be written :
dX, 2aSX2
dt =-a12X 1 + s+p (3. 111 )

dX2 aSX2
dt =ª12X1 - S + p (3. 112)

dS aSX2
Tt=-rs+p (3. 113)

dt
=
k dt
dP (.!. )dX
2
(3.1 14)

The rate of growth of the cell population (X 1 + X2) follows a Monod-type dependence on
substrate (S), while the rate of product fonnation is assumed to be proportional to the number
of mature cells. The cell yield coefficient is l/y, and the rate of product fonnation is pro-
portional to the rate of fonnation of mature cells. This model can simulate the production
k.inetics of the cyclic decapeptide gramicidin S, a bacteria! antibiotic produced by Bacillus
brevis . Gramicidin S synthetase activity increases dramatically in the late logarithmic phase
of growth, resulting in gramicidin S production in the late logarithmic and stationary phases
of growth. The k.inetics of growth and product fonnation are illustrated in Figure 3.30.

(37) The model described is that ofBlanch, H-W, and P.L. Rogers, Biotech. Bioeng. 13, 843 (1 97 1).
 1.2

• • • • •
0.5
dry weight
• V>
e¡¡;
<O

~ 0.8 0.4~ ~·
.<:!] ~ g:
;:: substrate
<O :::>
en 0.6
·a; 0.3 ~::::.
3:
~
s~
<I> -
a 0.4 gramicidin S 0.2"§- ::-:-

.. .. ..
:::>
~
0.2 0.1

o o
o 2 4 6 B 10 12 14
Time (hours)

Figure 3.30. The predictions of the age-dependent model for gramicidin S production by
Bacillus brevis. The parameters werefitted to the experimental data. Values ofthe constants
were a 12 = 0.905 hr" 1, a.= 1.190 hr- 1· ~ = 0.75 gil, Y= 0.51gm/gm,k=3.79 gm cellslgm
gramicidin S.

Evaluation of the constants in this model from batch culture data can be simplified by
introducing a "maturation time" (~) . The rate at which cells mature or enter the idiophase
38

depends on the fonnation rate of immature cells ata petiod t.,,, earlier. We can express this
relationship as

(3.115)

The rate of formation of product can now be written in tenns of the cell concentration
dP
dtl,= (l)dX
k drl1 -r., (3 .116)

Equation (3.116) can be integrated, assuming no product is initially present, and the inoculum
contains a cell concentration X0 :

P(t) = ( ~ }x(t-tm)-XJ (3. 11 7)

A graphical tria! and error procedure can be employed to estímate k and tm. By plotting P(t)
against X at times (t - ~) earlier, the value of the maturation time is established when a
linear relationship is found for sorne value of t.,,,. The slope of the linear relationship yields
the constant k. Table 3.15 illustrates the values of maturation times detennined for vatious
antibiotic ferrnentations this way.

(38) O.E. Brown and R.C. Vass. Biotech. Bioeng.15, 321 (1973).
Table 3.15. Cell maturation times and values of( l/k) for severa/ antibiotic fermentations. 39
Antibiotic Pro- Microorganism Maturation time Constant ( 1/k)
duced (hr)

Candidin Strepromyces virido- 14 7.5 (µg polyene/µg DNA)

flavus
Candicidin Streptomyces griseus 17.5 21.3 (µg polyene/µg DNA)
Chloramphenicol Streptomyces vene- 12 0.2 g chloramphenicol/gm
zuela e ce lis
Rifamycin Streptomyces mediter- 55 1 x 105 units rifamycin/gm
anei ce lis
Gramicidin S Baci/lus brevis 2 0.237 mg gramicidin/mg
ce lis
Penicillin P. chrysogenum 40 2.2 x 105 units penicil-
lin/gm cells