As the likelihood of subpopulation formation is increasing with

elongated generation times [11], continuous processes especially

suffer from subpopulation formation. Diverse microbial continu-

ous processes and their effects on productivity have been summar-

ized in literature already [17, 18]: Productivity over process time

resulted in a bell shaped curve: upon induction of inducible pro-

motors, productivity increased up to a stable level; however, ongo-

ing subpopulation diversification is believed to promote a fast

decrease of recombinant protein formation.

Unlike microbial continuous cultivation, continuous biopro-

cessing has already been implemented for mammalian host cell

lines, and the first products have been commercialized [19–

22]. As mammalian cell lines are known to propagate at slower

growth rates than microbial cells, subpopulation effects might not

be visible during “conventional continuous process times” (i.e.,

3–6 weeks process duration) [10, 11]. As industry accomplished

stable continuous processes with mammalian cell lines, the micro-

bial production sector is aiming at realizing stable continuous

processes as well. Results indicate cascaded continuous cultivation

to outperform (conventional) chemostat cultivation in regard to

long-term stable productivity [23–25]. Chemostat cultivation is

limited in its biomass production, as dilution rate and fed substrate

have to be adapted in order to avoid host cell washout [26, 27]. As

cascaded continuous cultivations uses two sequentially operated

chemostat processes, higher biomass can be achieved compared to

conventional chemostat processes [25].

Cascaded continuous cultivation comprises two sequentially

continuously operated reactors without additional requirements

for cell retention [10, 25]. Thereby a spatial separation of biomass

growth and target protein formation can be achieved. Reactor

one (i.e., stage one) is conventionally used for biomass growth

only. Noninduced cells are transferred to the second stage where

an additional feed is applied for induction. Recombinant product

can be harvested as a bleed stream from the second stage.

For E. coli BL21(DE3), pET plasmids are frequently employed

controlling target gene expression under control of the lac promo-

tor [28, 29]. Induction is thus restricted to either isopropyl-β-D-1-

thiogalactopyranoside (IPTG) or the natural inducer allolactose,

formed by fed lactose [30, 31]. IPTG has shown beneficial results

when used for short induction times (i.e., fed-batch cultivations)

[28, 32]. On the other hand side, IPTG induction was described to

exhibit toxic effects on host cells, especially visible at higher gener-

ation times [33]. Lactose induction facilitated a more stable pro-

ductivity than IPTG induction for both chemostat and cascaded

continuous cultivation. Still, when feeding lactose, carbon catabo-

lite repression (CCR) is a well-known phenomenon occurring in

substrate co-feeding [34, 35]. The glucose-lactose diauxic growth

causes decreased lactose uptake rates when glucose is present in

224

Julian Kopp and Oliver Spadiut