Biotechnology Advances 21 (2003) 417430Know-how and know-why in

biochemical engineering

U. von Stockar a,*, S. Valentinotti a, I. Marison a,C. Cannizzaro a, C. Herwig b

aLaboratory of Chemical and Biochemical Engineering, Swiss Federal Institute of Technology (EPFL),

CH-1015 Lausanne, SwitzerlandbPharmaplan Engineering AG, Altkirchstrasse 8, CH-4054 Basel, Switzerland

Abstract

This contribution analyzes the position of biochemical engineering in general and bioprocess

engineering particularly in the force fields between fundamental science and applications, and

between academia and industry. By using culture technology as an example, it can be shown that

bioprocess engineering has moved slowly but steadily from an empirical art concerned with mainly

know-how to a science elucidating the know-why of culture behavior. Highly powerful monitoring

tools enable biochemical engineers to understand and explain quantitatively the activity of cellular

culture on a metabolic basis. Among these monitoring tools are not just semi-online analyses of

culture broth by HPLC, GC and FIA, but, increasingly, also noninvasive methods such as midrange

IR, Raman and capacitance spectroscopy, as well as online calorimetry. The detailed and quantitative

insight into the metabolome and the fluxome that bioprocess engineers are establishing offers an

unprecedented opportunity for building bridges between molecular biology and engineering

biosciences. Thus, one of the major tasks of biochemical engineering sciences is not developing new

know-how for industrial applications, but elucidating the know-why in biochemical engineering by

conducting research on the underlying scientific fundamentals.

D 2003 Elsevier Inc. All rights reserved.

Keywords: On-line monitoring; On-line midrange IR spectroscopy; On-line Raman spectroscopy; Capacitance

spectroscopy; Biochemical engineering sciences; Role of biochemical engineering; Scientific fundamentals of

www.elsevier.com/locate/biotechadvbiochemical engineering

0734-9750/03/$ - see front matter D 2003 Elsevier Inc. All rights reserved.

doi:10.1016/S0734-9750(03)00058-2

* Corresponding author. Tel.: +41-21-693-31-91; fax: +41-21-693-36-80.

E-mail address: urs.vonstockar@epfl.ch (U. von Stockar).

1. Introduction

Biotechnology started out as a craft. For thousands of years, mankind exploited

empirical know-how handed down from generation to generation to produce bread, beer,

wine and the like. Throughout this long period, the evolution of production procedures

was so slow that it could hardly be perceived because new know-how had to be obtained

in a completely empirical way due to the total lack of know-why. This state of affairs only

started to change relatively recently when around 1700, Anton van Leeuwenhoek invented

the first optical microscope. The discovery of new and unsuspected microscopic life forms

ensuing from his invention marked the starting point of biology as a biotechnology-

relevant science. A steady development of biological knowledge and science followed,

pioneered by eminent personalities such as Louis Pasteur, Robert Koch, and Alexander

Fleming among many others.

In the last 30 years, the development of fundamental and molecular biology started to

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430418accelerate in an exponential fashion due to breakthroughs such as genetic engineering,

hybridoma and animal cell culture technology, genomics and proteomics. As a result,

biotechnology today could not be further away from a craft; it has given rise to a high tech

industry par excellence. The biotechnology products sold on the market are often the result

of cutting-edge scientific research.

Due to the explosion-like development of fundamental biological sciences, biochemical

engineering finds itself in a very different situation. While biochemical engineering and

biotechnology designated quite similar fields 30 years ago in that they both meant the

science and art of using living cells and their constituents in order to obtain useful

industrial products or services, the term biotechnology has a much wider meaning today,

comprising all sorts of activities based on biological sciences and using some kind of

technical means, be it industrial development and production, applied or fundamental

research, or medical activities. This raises the intriguing question of whether biochemical

engineering is just the collection of know-how needed for industrial realization of

biotechnology results or whether it is a science in its own right that must be taught in

academia (Fig. 1). The question is especially pertinent for all process-relevant activities of

Fig. 1. Biochemical and chemical engineering in the force field between science and applications.

biochemical engineering such as bioprocess development and design, bioprocess engineer-

ing, equipment design and engineering, and manufacturing.

The question is of utmost importance to academic biochemical engineers because many

university managers and research policymakers appear to be so impressed by the recent

successes of fundamental and molecular sciences that they tend to lose sight of engineer-

ing and biochemical engineering. At the same time, academic biochemical engineering

programs tend to shift their emphasis ever more from the process related to biomolecular

aspects of biochemical engineering (Chaudhuri, 1997). As a result, there is a trend in many

countries to outsource at least the bioprocess-related aspects of biochemical engineering

from academia to polytechnics or industries, a trend that was well characterized by

Wandrey (2001) in an editorial titled appropriately Bio- without Technology?

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430 419Fig. 2. Traditional brewingmodern monitoring. Reprinted with permission from Rose (1981), (Copyright: J.

Brenneis).

The aim of this contribution is to analyze the position of biochemical engineering in

general and bioprocess engineering particularly in the force field between applications

and science by showing that it evolved gradually from a know-how-based type of

approach to a science elucidating more and more know-why.

2. The shift of bioprocess engineering from know-how to know-why

The evolution of biochemical engineering towards science may be illustrated using

culture technology as an example. Whereas the kettles used for beer brewing in Fig. 2 are

still very traditional, the figure also shows the presence of measuring probes and of a

modern control panel. The need to monitor and to control cultures has given rise to a more

widespread use of probes and control schemes in bioreactors as well. By developing and

Semi-online analysis EtOH GC 30 mg/l 6/h

Acetic acid GC 30 mg/l 6/h

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430420Acetaldehyde GC 30 mg/l 6/h

Glucose FIA 10 mg/l 10/h

Sucrose FIA 50 mg/l 10/h

EtOH FIA 50 mg/l 10/h

NH4+ FIA 18 mg/l 10/happlying more and more sophisticated online and at-line monitoring methods, biochemical

engineers obtain an ever more complete picture of why cultures behave as they do, which

enables them to develop new and more efficient know-how.

Table 1 shows a list of online and semi-online analysis methods that are available today

in many academic biochemical engineering laboratories. Amongst them are conventional

probes for pO2 and pH and off-gas analysis. In addition, a large number of metabolites

may nowadays be monitored by semi-online analysis based on repetitive sampling of the

culture at relatively high frequencies and on an analysis of the samples by GC or by FIA

(Herwig et al., 2001). Because taking large numbers of samples is often avoided in the

biological process industries, biochemical engineers have been looking for alternative

noninvasive methods to obtain the same type of information. One suitable method is the

use of online reaction calorimetry (von Stockar and Marison, 1991). By applying this

technique to a 300-l fermentation process, Voisard et al. (2002) have demonstrated very

clearly that calorimetric online measurements are a relatively simple and highly appro-

priate method to monitor industrial cultures online and also to control fed-batch processes.

Another class of online instruments that have recently generated a lot of interest are

spectroscopic measurements. Table 2 shows four types of spectroscopic online analyses

that have been investigated in the authors laboratory. Mid-infrared spectroscopy is able to

Table 1

Current on-line monitoring techniques for cellular cultures

Compound Method Detection limit Frequency

Sensors H+ pH electrode Continuous

O2 pO2 electrode Continuous

Off-gas O2 Paramagnetic 0.01% Continuous

CO2 IR 0.01% Continuous

measure the concentration of a number of different metabolites simultaneously at high

frequency (Doak and Phillips, 1999; Fayolle et al., 2000; Pollard et al., 2001; Sivakesava

et al., 2001). Because the spectra of the metabolites of interest normally overlap, it is

necessary to work with so-called multivariate calibration models (Martens and Naes,

1988). Fig. 3 shows that if this model is correctly constructed, it is possible to follow

online quite a number of different metabolites (Kornmann et al., submitted for publica-

tion). The graph also demonstrates the reliability of the respective signals during spiking of

the culture with important metabolites, such as ethanol or acetic acid. The respective

Table 2

Spectroscopic techniques for on-line monitoring of cellular cultures

Analyte Spectroscopy type Detection limit Frequency (min 1)

Carbohydrates,

amino acids,

organic acids

Mid-IR with ATR 100500 mg/l 0.33

Biomass Dielectric 105 cells/l 2

Carotenoids Raman 1 mg/l 0.1

Vitamins, cofactors Fluorescence 1

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430 421signals reproduced the peaks in a faithful way, but the spiking did not disturb other

unrelated signals, thus showing the correctness of the calibration.

Biomass cannot be seen with infrared spectroscopy. One way to follow cell concen-

tration is to use dielectric or capacitance spectroscopy (Kell et al., 1990; Olsson and

Nielsen, 1997). Cannizzaro et al. (in press a,b) show that this technique is especially

promising if exploited as a real spectroscopy. Systematically scanning over a range of

frequencies while doing the measurements and using multidimensional calibration models,

they were able to follow not only the cell counts, but also the viability and even to get

online information on the size of the cells.Fig. 3. Monitoring four metabolites simultaneously by IR spectroscopy during a culture of Gluconoacetobacter

xylinus. Keys: measured concentrations (g/l) of fructose (E), acetate (D), ethanol (z), and gluconacetan (.).Lines: predictions by IR spectroscopy. At 28 h, a pulse of ethanol was administered that was rapidly metabolized

into acetate, which was consumed in turn. Reprinted with permission from Kornmann et al. (submitted for

publication).

Another interesting form of spectroscopy is Raman spectroscopy, which picks up

conjugated double bonds occurring typically in the membranes. It is thus possible to

measure online the concentration of carotenoids with a high detection sensitivity (Weesie

et al., 1999; Cannizzaro et al., in press a,b).

The online analytical tools just described were applied to a culture undergoing standard

transient experiments (Fig. 4). In standard transient experiments on continuous cultures

growing initially at steady state, the metabolism is challenged by forcing the culture

through a transient either by applying a pulse to the culture by a shift-up or shift-down of

the dilution rate or a substrate feeding rate. Such standard transient experiments are a

highly powerful way to investigate the metabolism of a culture (Duboc et al., 1998;

Herwig et al., 2001). It is possible to obtain fundamental and qualitative insight into

physiology and metabolism, and quantitative data on the kinetics of regulation. It has also

been shown that such experiments permit rapid and effective strain and mutant character-

ization and yield quantitative data for mathematical models, for rapid and rational

bioprocess development, and for control (Duboc et al., 1998). Fig. 4 shows the transients

that occur if a culture of Saccharomyces cerevisiae growing at a subcritical dilution rate is

subjected to a shift-up to a higher dilution rate but which is still subcritical. This means

that both values of the dilution rate will allow continuous cultures of S. cerevisiae to grow

by a completely oxidative metabolism, i.e., by pure respiration. One could thus expect that

the signals simply evolve in an exponential way from a lower level to a higher level

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430422because of the shift-up. However, Fig. 4 clearly shows that a small overshoot of glucose

occurs and that the other concentrations and signals show pronounced excursions with

overshoots. The fact that the carbon evolution rate increases much faster than the oxygen

uptake rate clearly indicates a transient excursion into reductive metabolism for many

Fig. 4. Shift-up in dilution rate from 0.07 h 1 to 0.22 h 1 in a culture of S. cerevisiae. Glucose (z), ammonium(.), biomass (o), specific CO2 evolution (E) and oxygen consumption ( ) rates. Reprinted with permissionfrom Herwig and von Stockar (2003).

hours. Fig. 5 demonstrates that reductive overflow metabolites such as ethanol and acetic

Fig. 5. Shift-up in dilution rate from 0.07 h 1 to 0.22 h 1 in a culture of S. cerevisiae. Ethanol (n), acetic acid(+), respiratory quotient (w). Reprinted with permission from Herwig and von Stockar (2003).

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430 423acid are secreted in the fermentation broth and disappear again after a number of hours.

In order to obtain more insight and to understand why the culture behaves as shown in

Figs. 4 and 5, it is useful to consider the rates at which a given metabolite is consumed or

produced by the cells rather than the concentration profiles. The concentration data may beFig. 6. Bioreactor with system boundary for computing consumption and production rates.

transformed into rates online by performing material balances over the whole fermenter

(Fig. 6) and by correctly accounting for the accumulation effects inside the fermenter

(Herwig et al., 2001). Even more insight into the culture behavior may be obtained by

Fig. 7. Opening the black box: metabolic flux analysis (MFA) allows an estimation of intrinsic metabolic reaction

rates.

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430424interpreting it in terms of intrinsic metabolic rates rather than exterior consumption and

production rates (Fig. 7). This type of analysis permits a partial opening of the black box

of the cell and an online observation of the metabolic activity going on inside. For this

purpose, it is necessary to draw a very simple metabolic map and then to calculate the rates

rj of the metabolic reactions from the exterior conversion rates ri by online metabolic flux

analysis. Fig. 8 shows the simplified metabolic map that was drawn for the shift-up

experiment just discussed. As depicted, the metabolism was represented in an extremely

simple way by one anabolic reaction ( qan), one catabolic reaction representing glycolysis

( qcat), two reactions representing oxidative ( qcatox) and reductive ( qcatred) catabolism ofFig. 8. Simple online metabolic flux model for aerobic culture on glucose for yeast. Reprinted with permission

from Herwig and von Stockar (2003).

pyruvate, oxidative phosphorylation ( qoxphos), and the production of ethanol ( qEtOH) and

of acetic acid ( qNAc). The volumetric rates rj of these reactions may be computed online

(Herwig and von Stockar, 2002a,b), so that a quantitative picture of the metabolic activity

of the culture is obtained in real time.

Fig. 9 shows this picture as a function of time in terms of specific rates qj, which were

obtained off-line using biomass concentration data. Although the glucose uptake rate

increases dramatically fast after shift-up (not shown), it cannot quite cope with the increase

in glucose feeding rate, thus explaining the small glucose-concentration peak in the broth.

Due to the dramatic glucose uptake rate just after the shift-up, catabolism is also seen to

increase enormously during almost an hour ( qcat, Fig. 9a). Oxidative catabolism through

the TCA cycle qcatox, as well as oxidative phosphorylation qoxphos jump to a higher value

immediately after shift-up, but as this value is insufficient to match the dramatic increase

of qcat, a large fraction of pyruvate overflows into reductive metabolism ( qcatred). There

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430 425Fig. 9. Intrinsic metabolic rates as a function of time during transient growth. (a) qcat (bold black line), qcatox(black line), qoxphos (fine black line), qcatred (bold gray line); (b) qcatred (bold black line), qEtOH (black line), qHAc(gray line).

seems to exist another bottleneck for qEtOH, which cannot cope with qcatred, as shown in

Fig. 9b. The difference thus overflows into acetic acid ( qHac).

Over time, qcatox and qoxphos increase exponentially to meet the oxidative demand of

qcat (Fig. 9a). This reduces the overflow reactions such that first qHac and then qEtOHcome to a halt. At that point, which is reached 6 h after the start of the experiment, qcatoxand qoxphos are strong enough to handle the glucose influx oxidatively, and now

correspond to their new steady state values. Nevertheless, they increase further, making

qcatred, qHac and qEtOH negative. The disappearance of these two metabolites from the

culture is thus not only due to washout, but also to actual consumption by the culture. As

soon as the last traces of ethanol have disappeared from the fermentation broth at about

9.5 h, qcatox and qoxphos fall back to their steady-state values. Fig. 9 illustrates that using

modern analytical online tools, biochemical engineers are in the position of getting a fair

amount of insight into the metabolic behavior of their cultures right during the

fermentation process.

3. Know-why in biochemical engineering and molecular biosciences

Fig. 10 puts the observation of the previous section into a wider perspective. It

shows a schematic view of a single living cell with its genome, proteome, metabolome

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430426Fig. 10. Towards an understanding of the cell at the molecular level.and fluxome. Biochemical engineers are most interested of course in the lower part of

the figure. They have to be able to predict which sort of substrate is consumed at

what rate and internally transformed into what interesting product under which

conditions. But by developing tools that enable them to obtain insight into the

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430 427internal metabolic activity and also by fundamental research on metabolomics, or

fluxomics, performed in several laboratories in Europe, biochemical engineers already

developed quite extensive quantitative understanding of the inner workings of the

cellular metabolism.

Biomolecular sciences, on the other hand, work more from the top downward on this

scheme. Starting from molecular biology, they have developed an impressive array of

highly efficient tools for large-scale genome analysis. New tools are being developed in

the present postgenomic era that will allow them to investigate also the proteome.

Functional genomics and proteomics, based on advances in microarray techniques, on

capillary electrophoresis, on 2-D gel electrophoresis, on mass spectrometry and on an

enormous effort in bioinformatics, now permit genome-wide exploration of transcriptional

and expression profiles (Ryu and Nam, 2000; Ye et al., 2001; Shoemaker and Linsley,

2002; Panda et al., in press). There is little doubt that these tools will allow biomolecular

sciences to predict relevant features also of the metabolome and of the fluxome. It is thus

certainly possible that in the future there will be a coalescence of the research efforts by

biochemical engineers, working from the bottom upwards in Fig. 10, and of biomolecular

scientists, working from the top downwards.

Biochemical engineers are already involved in many activities at the frontier between

biomolecular sciences and biochemical engineering, and are undoubtedly playing a key

role in bringing about the coalescence referred to above. Methods such as oligonucleo-

tide-directed mutagenesis, DNA shuffling, combinatorial chemistry and peptidomimetics

are harnessed by biochemical engineers for rapid and efficient discovery of new drugs

(Weng and De Lisi, 2002; Ryu and Nam, 2000). Together with protein engineers, they

also form the basis for developing more efficient biocatalysts for biomolecule produc-

tion processes. Biochemical engineers are, however, also involved in building bridges

from biomolecular sciences all the way to bioprocess engineering. The challenge in

predicting the phenotype and the metabolic capacities and behaviour in a quantitative

way exceeds very clearly the current potential of a purely molecular approach due to

the phenomenal complexity of cells and biology. The fact that this endeavour requires

considering cells and organisms as hugely complex systems (Palsson, 2000) explains

why many biochemical engineers are active in metabolic pathway analysis, metabolic

engineering and systems biology. Their engineering background enables them to tackle

complex systems and they are able to draw from nonmolecular fundamentals of

engineering sciences. Once the frontiers between biomolecular sciences and bioprocess

engineering have been bridged, biochemical engineers will be able to predict the

behavior of the cultures quantitatively under actual application conditions, e.g., in

bioreactors, directly from the molecular knowledge of the genome. It is obvious that

such an understanding of the cell at a molecular level would enable biochemical

engineers to develop their bioprocesses, biosystems, applications and new know-how

much more efficiently and rationally, based not on empirism, but on fundamental know-

why. The opportunity that now exists to build bridges from bioengineering to the

biomolecular sciences makes biochemical engineering a particularly exciting subject

today.

Based on these statements, Fig. 11 positions biochemical engineering in the force

fields between science and application (vertical axis) and between academia and industry

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430428(horizontal axis). In this schematic drawing, biochemical engineering denotes not only

the biomolecular aspects of the field, but explicitly also the nonmolecular aspects

including bioprocess engineering. Industry is clearly interested in real-world applica-

tions. It usually has its own science programs that feed scientific knowledge into the

development of applications. Academia, on the other hand, is mainly concerned with

basic and molecular sciences, but it also continuously produces new ideas, new

molecules and new know-what for applications. These are then transferred into industry

and into the real world through technology transfer. This transfer process is often much

longer and more complicated than what scientists only familiar with basic and molecular

science imagine. That is exactly why engineers are needed in the process. Their typical

systemic and multidisciplinary engineering approach and profound knowledge of what is

needed in the real world and on the market enables engineers to catalyze the transition

of discoveries from basic and molecular science into real-world applications. This can of

course be done in a more or less empirical way, or in a scientific way. An example is

bioprocess development in the biotechnology industry, which is often done under huge

Fig. 11. Biochemical engineering between sciences and applications.time pressures and is constrained by complex regulatory issues. Therefore, there may not

always be the time to optimize bioprocesses in a scientific and rational way. This has led

to the paradox situation that biotechnology offers high-tech products at the cutting edge

of science but tends to produce them using very traditional process technology that has

not changed much in the last 30 years or so. An example is erythropoietin production by

Amgen, which was scaled up simply by multiplying a relatively inefficient laboratory

culture technique, namely roller bottles.

However, if biotechnology as a whole, including engineering activities such as

bioprocess development, is to move away from an art and become a thoroughly scientific

endeavor, then engineering activities must themselves be based on scientific fundamentals

and on solid know-why and not only on empirical know-how.

Although trivial to engineers, many scientists are not aware that the scientific basis

underlying biochemical engineering activities is quite different from molecular and

fundamental sciences. It is much wider, and may or may not, or may not yet, have a

molecular basis. The opportunity existing today to build bridges between bioengineering

biology as a complex system. Equally important is a certain knowledge of industrial

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430 429practices and needs if young biochemical engineers want to fulfill their role of catalyzing

the transition from molecular discoveries to real-world applications. Only the existence of

strong academic curricula in engineering, emphasizing both modern biomolecular and

nonmolecular engineering sciences, will make sure that an adequate number of chemical

and biochemical engineers are trained who will be able to tackle empirical and traditional

know-how in industry from a scientific point of view and who will establish a rational and

scientific basis for industrial procedures, thus making empirical and trial-and-error-based

search for know-how obsolete.

4. Conclusions

The role of biochemical engineering sciences is to provide a sound scientific basis for

the engineering activities, i.e., a sound basis of know-why. The aim of this quest is to

provide tools for rational, rapid and efficient development of processes, products and

services, i.e., for the efficient development of novel know-how.

The mission of chemical and biochemical engineering alike could be formulated as the

advancement of the scientific fundamentals of engineering for a rational, sustainable and

safe approach to the development of products, processes and services in the chemical,

pharmaceutical and life science industries. These scientific fundamentals, underlying both

chemical and biochemical engineering, must become increasingly molecular in nature.

However, in part due to the phenomenal complexity of the biological world and real-world

systems, they are wider than the molecular sciences and must comprise the principles of

nonmolecular engineering sciences. It is precisely one of the fascinating challenges of

todays biochemical engineers to not only elucidate the scientific basis of their engineering

profession, but in doing so also to build the necessary bridges between engineering and

biomolecular sciences.and biomolecular sciences, and thus to construct a molecular basis, is one of the

particularly fascinating challenges that lie ahead.

This is also the reason for the need for strong academic programs in biochemical

engineering sciences. The main task of academic biochemical engineering is not primarily

to develop new know-how for industrial applications, but to elucidate the know-why by

conducting research on the scientific fundamentals underlying biochemical engineering

activities. Although thoroughly scientific, this type of research is not usually conducted by

fundamental scientists for two reasons. First, todays competitive climate in academic

research forces fundamental scientists to pursue topics that are recognized as hot in

fundamental sciences and does little to motivate them to deviate into engineering issues. In

addition, it takes a good understanding of research needs in the real world of industrial

environment to define meaningful research programs. Conducting research into the

fundamentals of engineering science is thus usually performed by researchers with a

strong engineering background.

Similar remarks hold for teaching. Although biomolecular aspects will increase in

importance in engineering education, the nonmolecular engineering sciences and methods

remain crucial for biochemical engineers, who will have to elucidate our comprehension of

References

U. von Stockar et al. / Biotechnology Advances 21 (2003) 417430430Cannizzaro C, Rhiel M, Marison I, von Stockar U. On-line monitoring of Phaffia rhodozyma fed-batch process

with dispersive Raman spectroscopy. Biotechnol Bioeng 2003a [in press].

Cannizzaro C, Gugerli R, Marison IW, von Stockar U. On-line biomass monitoring of CHO perfusion cultures

with scanning dielectric spectroscopy. Biotechnol Bioeng 2003b [in press].

Chaudhuri JB. Biochemical engineeringpast, present and future. TIBTECH 1997;15(October):3834.

Doak DL, Phillips JA. In situ monitoring of an Escherichia coli fermentation using a diamond composition ATR

probe and mid-infrared spectroscopy. Biotechnol Prog 1999;15:52939.

Duboc P, Cascao Pereira LG, von Stockar U. Identification and control of oxidative metabolism in Saccharomyces

cerevisiae during transient growth using calorimetric measurements. Biotechnol Bioeng 1998;57:6109.

Fayolle P, Picque D, Corrieu G. On-line monitoring of fermentation processes by a new remote dispersive middle-

infrared spectrometer. Food Control 2000;11:2916.

Herwig C, von Stockar U. Quantitative analysis of the oxidative metabolism in HXK2 and REG1 deletion

mutants of Saccharomyces cerevisiae. Enzyme Microb Technol 2002a;31:50515.

Herwig C, von Stockar U. A small metabolic flux model to identify transient metabolic regulations in Saccha-

romyces cerevisiae. Bioprocess Biosyst Eng 2002b;24:395403.

Herwig C, von Stockar U. Quantitative comparison of transient growth of S. cerevisiae, S. kluyveri, and K. lactis.

Biotechnol Bioeng 2003;81:83747.

Herwig C, Marison IW, von Stockar U. On-line stoichiometry and identification of metabolic state in dynamic

process conditions. Biotechnol Bioeng 2001;75:34554.

Kell DB, Markx GH, Davey CL, Todd RW. Real-time monitoring of cellular biomassmethods and applications.

TrAC Trends Anal Chem 1990;9:1904.

Kornmann H, Duboc P, Niederberger P, Marison IW, von Stockar U. The influence of residual ethanol on the

growth of Gluconoacetobacter xylinus. Appl Microbiol Biotechnol 2003 [submitted for publication].

Martens H, Naes T. Multivariate calibration. New York: Wiley; 1988.

Olsson L, Nielsen J. On-line and in-situ monitoring of biomass in submerged cultivations. Trends Biotechnol

1997;15:51722.

Palsson BO. The challenges of in sileco biology. Nat Biotechnol 2000;18:114750.

Panda S, Sato TK, Hampton GM, Hogenesch JB. An array of insights: application of DNA chip technology in the

study of cell biology. Trends Cell Biol 2003 [in press].

Pollard D, Buccino R, Connors N, Kirschner T, Olewinski R, Saini K, et al. Real-time analyte monitoring of a

fungal fermentation, at pilot scale, using in situ mid-infrared spectroscopy. Bioprocess Biosyst Eng 2001;

24:1324.

Rose AH. The microbiological production of food and drink. Scientific American; 1981. p. 94104 [September

issue].

Ryu DDY, Nam D-H. Biomolecular engineering, a new frontier in biotechnology. J Mol Catal B Enzym

2000;10:2337.

Shoemaker DD, Linsley PS. Recent developments in DNA microarrays. Curr Opin Microbiol 2002;5:3347.

Sivakesava S, Irudayaraj J, Demirci A. Monitoring a bioprocess for ethanol production using FT-MIR and FT-

Raman spectroscopy. J Ind Microbiol Biotechnol 2001;26:18590.

Voisard D, Pugeaud P, Kumar AR, Jenny K, Jayaraman K, Marison IW, et al. Development of a large-scale

biocalorimeter to monitor and control bioprocesses. Biotechnol Bioeng 2002;80(2):12538.

von Stockar U, Marison IW. Large scale calorimetry and biotechnology. Thermochim Acta 1991;193:21542.

Wandrey C. Bio- ohne Technik? Chem Ing Technol 2001;73(7):787.

Weesie RJ, Merlin JC, Lugtenburg J, Britton G, Jansen FJHM, Cornard JP. Semiempirical and Raman spectro-

scopic studies of carotenoids. Biospectroscopy 1999;5(1):1933.

Weng Z, De Lisi C. Protein therapeutics: promises and challenges for the 21st century. TIBTECH 2002;20(Jan-

uary):2935.

Ye RW, Wang T, Bedzyk L, Croker KM. Applications of DNA microarrays in microbial systems. J Microbiol

Methods 2001;47:25772.

Know-how and know-why in biochemical engineeringIntroductionThe shift of bioprocess engineering from know-how to know-whyKnow-why in biochemical engineering and molecular biosciencesConclusionsReferences