Figure 1. Screening for microorganisms from a soil sample that can degrade pyrene. The cells are grown isolated on the Petri dish with pyrene as the only C source. The clones that can degrade pyrene and divide are directly observed (Kästner, 1994, personal communication).

Figure 2. Screening for E. coli cells that produce penicillin amidase. A filter paper with a substrate selectively hydrolyzed by this enzyme that yields a yellow product is placed on the Petri dish with isolated cell clones. The clones producing penicillin amidase color the filter paper yellow (here seen as dark spots) above the clone. The cells that produce more enzyme develop the yellow color more rapidly than those producing less of the enzyme.

Figure 3. A fluorescence-activated cell sorter (FACS) for high-throughput screening (HTS) of cells with respect to different properties such as size, DNA content, or content of an intracellular or periplasmic enzyme, E. In the latter case, single cells are transported through a capillary in a solution with a fluorogenic substrate that can only be converted to a fluorescent product by the enzyme E. The cells with the enzyme (marked black) become fluorescent and are detected by the measuring device (detector and computer); this gives a signal to a device for electrostatic selection of these cells, and they are collected into a separate vessel.

Figure 4. Transcriptional, translational, and posttranslational processes that influence the yield of intracellular, extracellular, and periplasmic enzymes. The latter two are synthesized as pre-pro-enzymes. A signal peptide (pre-) is required for the transport through the cell membrane, and is cleaved off by signal peptidases. The biologically active enzyme is written in boldface letters.

Figure 5. The proteolytic processing of penicillin amidase from E. coli (PA _EC, EC 3.5.1.11). The pre-pro-enzyme (ppPA) is one polypeptide chain that from the N-terminal consists of the signal peptide (SP) that directs the periplasmic enzyme to the membrane translocation system, the A-chain, the linker or pro-peptide that must be removed to obtain an active enzyme, and the B-chain. The number of amino acids in these peptides is given in the second line from above. The dominating active PA form in samples of technical enzyme (PA _7.0) from wild-type cells consists of the A- and the B-chain and has an isoelectric point (pI) of 7.0 (Kasche et al., 1999, 2005).

Figure 6. The influence of temperature and medium on the biosynthesis of PA _EC in the wild-type parental strain E. coli ATCC 11105 studied with SDS gel electrophoresis (12% gel) and immunoblotting using a monoclonal antibody against an epitope on the B-chain. In this denaturing electrophoresis, unfolded polypeptides are separated with respect to molecular weight. The immunoblotting visualizes ppPA _EC and polypeptides derived from it with the epitope to which the antibody is bound. The temperature dependence was studied with LB medium; the medium influence (LB and a synthetic M9 medium) with cells grown at 28 °C. Cells were grown for 6 h after the induction with 6 mM phenylacetic acid, homogenized (temperature) or separated into cytoplasmic (C) and periplasmic (P) fractions. In each lane, the same number of cells was analyzed. Pure PA _7.0 and a slowly processing pro-PA (pPA) were used as standards. The intensity of the same polypeptide band increases with its content in the sample. The intensity of the B-chain (Figure 5) with MW 62 kDa is a measure for the synthesis of biologically active PA _EC, and shows that it has a temperature optimum at 26–28 °C and depends on the medium composition. The presence of the pPA and ppPA bands gives no information whether these represent correctly folded structures that can be processed to biologically active PA _EC that is localized in the periplasm. The other immunostained bands represent B-chain fragments with the epitope produced by intracellular proteolysis. The synthesis of the pre-pro-enzyme is hardly influenced by temperature in the range studied. With increasing temperature, the amount lost by misfolding, inclusion body formation, and intracellular processing approaches 100%.

Figure 7. The influence of the Ca ²⁺ content in the medium on the yield of biologically active homogeneously expressed PA _EC (E. coli penicillin amidase) and heterogeneously expressed PA _AF (A. faecalis penicillin amidase) in E coli (BL21(DE3)) cells exponentially grown at 28 °C. Aliquots were removed 9 h (PA _EC) and 12 h (PA _AF) after induction with ITPG, fractionated into cytoplasmic (C) and periplasmic (P) fraction, and separated as native folded proteins with respect to isoelectric point (pI = pH of the focused band) by nondenaturing isoelectric focusing or molecular weight by denaturing SDS gel electrophoresis (12.5% gel). The same standards for PA _EC and pPA _EC were used as in Figure 6. After IEF, a chromogenic substrate (NIPAB) was layered on the gel. The colored bands indicate the existence of more than one active enzyme forms, as is expected from the processing scheme in Figure 5. In this case, the intensity of the band increases with the amount of the enzyme form, as long as not all substrate near the band has been consumed. The dominating form of PA _EC produced in these cells has a pI of 7.3 and a higher specific activity than the dominating form produced by wild-type cells in shake flasks that has a pI of 7.0. After IEF and gel electrophoresis, the gels were immunostained with a monoclonal antibody against an epitope on the B-chain of PA _EC or a polyclonal antibody against PA _AF. This visualizes correctly folded enzymes in IEF, and in SDS gel electrophoresis unfolded polypeptides with epitopes recognized by the antibodies. Without Ca ²⁺, no correctly folded pre-pro- or pro-enzyme is found and the yield of active enzyme was very low (Table 3). The cytoplasmic band with MW » 92 kDa was identified as pPA _EC by N-terminal sequencing. At optimal Ca ²⁺ content, much less intracellular proteolysis was observed in the cells producing PA _AF. Unlike the pre-pro-PA _EC, pre-pro-PA _AF is transported in an unfolded conformation through the Sec translocation system. Its content per cell is higher than that for the Tat translocation system, through which folded ppPA _EF is transported. The residence time in the cytoplasm for ppPA _EC is probably longer than that for ppPA _AF. Thus, the yield of PA _EC is lower than that for PA _AF, due to loss of precursor by intracellular proteolysis.

Figure 8. Production of penicillin amidase from E. coli in recombinant E. coli BL21(DE) cells grown to high cell density (CDW, g l ^–1). The cells were fed exponentially with glucose and different concentrations of Ca ²⁺ at 10 h before induction with ITPG. The cells were induced to produce the pre-pro-enzyme at time 0 (Kasche et al., 2005).

Figure 9. Mechanism for the posttranslational processing of extracellular lipases (L, EC 3.1.1.3) from different microorganisms. SP = signal peptide. The schemes apply for (a) eukaryotic lipases (such as Rh. miehei, Rh. oryzae) and (b) prokaryotic lipases (such as Pseudomonas sp., Burkholderia sp., C. viscosum) (Jaeger, Dijkstra, and Reetz, 1999; Ueda et al., 2002).

Figure 10. Downstream processing steps for the recovery, concentration, purification, formulation, and quality control of soluble biologically active intracellular, extracellular, and periplasmic enzymes. The number of processing steps before the concentration step and the formulation and quality control is given without and (with) removal of nucleic acids.

Figure 11. Mixed-mode adsorbents use different interactions for the adsorption and desorption. Enzymes are bound by hydrophobic interactions to an uncharged adsorbent at a pH > pI. Then the pH, ionic strength I, and the electric field strength e are constant in the electric double layer on the adsorbent surface. When the bulk pH is lowered, the enzyme and the surface charge on the adsorbent become positive, and the physical properties depend on the distance from the surface. With increasing charge on the enzyme and the surface, the repulsion force becomes larger than the hydrophobic interaction, and the enzyme is desorbed. The desorption pH depends on the pI of the enzyme.

Figure 12. The Langmuir adsorption isotherm for the binding of an enzyme to an adsorption site on the surface of a porous particle or a porous membrane. The static capacity n and the apparent dissociation constant K are determined as shown.

Figure 13. The time dependence of the adsorption of fluorescein-labeled α-chymotrypsin to soybean trypsin inhibitor covalently immobilized in porous Sepharose particles (n = 700 µM; K = 12 µM) as a function of the enzyme concentration outside the particle. The concentration profiles within the particle were determined at different times increasing from the bottom at times in seconds given to the right, with CLSM (confocal laser scanning microscopy) (Kasche, Galunsky, and Ignatova, 1999; Kasche et al., 2003): (a) outside fluorescein-labeled α-chymotrypsin concentration 22 µM > K; (b) outside fluorescein-labeled α-chymotrypsin concentration 2.2 µM < K. As expected from Eqs. ( 1) and ( 2), the adsorption is much faster, and the adsorbed protein amount per unit adsorbent volume is higher at the higher concentration, as follows from the line connecting the profiles after 650 s.

Figure 14. Continuous chromatography with a simulated moving bed for the nonisocratic separation of a two-component protein system (blue = A, the target protein that is adsorbed in the adsorbent particles; yellow = B, other proteins that are not adsorbed in the adsorbent). The SMB consists of eight identical columns filled with adsorbent particles whose inlets can be connected to outlets of adjacent columns in the zone or the feed and adsorbent and desorbent buffers. The outlets can be connected to inlets of adjacent columns in the zone or the containers for pure A, pure B, and the waste. The figure shows the distribution of A and B in the zones 1–4 between two consecutive switches of the last column in each zone one step downward. The time between the switches is constant and gives Q, the mobility of the solid phase that moves opposite to the liquid mobile phase. The feed to the first column in zone 2 continues until it is filled with adsorbed A. As the mobility of A in this section must be <Q, the adsorption isotherms (Eqs. ( 1) and ( 3)) of A (and B) must be known to design the process. The column with adsorbed A then moves downward one step and is connected to desorbent 1, and replaces the column from zone 3 without B that is moved to the desorption zone and connected to desorbent 2. At the same time, the regenerated column from zone 1 is connected to the feed, and the first column in the desorption zone, from which A has been desorbed, is moved to zone 1 and connected to the adsorbent buffer. The inlet and outlet ports of the columns that are moved to the next zone must be changed as shown in the figure. This is achieved using multiport valves that are controlled by a computer (Gottschlich and Kasche, 1997).

Figure 15. Chromatographic purification of wild-type A. faecalis penicillin amidase (PA) produced in recombinant E. coli at optimal Ca ²⁺ content as shown in Figure 7, with (b) and without (a) hydrolysis of the nucleic acids in the concentrated clarified homogenate. An anion-exchange Mono Q 10/10 column (GE Healthcare) was used. After equilibration with 30 mM Tris–HCl, pH 7.5 at 2 ml min ^–1, and application of the sample with »3 mg active enzyme in the same buffer, the column was eluted as follows: 0–70 min with a linear salt gradient 0–20% 1 M NaCl in 30 mM Tris–HCl; 60–90 min with 30 mM Tris–HCl containing 1 M NaCl. The inset in (b) shows a Coomassie-stained SDS electrophoretic gel (lane 1: the homogenate; lane 2: the enzyme peak shown by the arrow; lane 3: molecular weight markers). The last peak in (a) contains mainly nucleic acids (Yuryev, 2010, and unpublished results).