Biot?h. Adv. Vol. 9, pp. 217-240, 1991 0734- 9750/91 $0.00 + .50 Printed in Great Britain. All Rights Reserved. ? 1991 Pergamon Press pic
BACTERIAL CELL DISRUPTION: A KEY UNIT
OPERATION IN THE RECOVERY OF
INTRACELLULAR PRODUCTS SUSAN T. L. HARRISON
Department of Chemical Engineering, University of Cambridge, Pembroke Street, Cambridge, CB2 .3RA, U.K. Present Address: Department of Chemical Engineering, University of Cape Town, Private Bag, Rondebosch 7700, South Africa
The need for microbial cell disruption has hindered the large scale production of commercial biotechnological products of intracellular derivation. The intracellular nature of many recombinant products and the potential use of the bacterial storage product, PHB as a commodity thermoplastic have
renewed interest in the improvement of this unit operation. This paper provides a review of processes of a mechanical, physical, chemical or biological nature used for cell disruption on both the laboratory and large scale. Applicability of the techniques to large scale operation is discussed. Modification of existing processes is suggested for the reduction of energy requirements and improved process economics. The requirements for the liberation of granular intracellular products such as inclusion bodies and virus-like yeast particles are distinguished from those for the liberation of soluble products,
mainly proteinaceous in nature.The integrated nature of the process with both upstream and downstream processes is addressed. Finally, the recent approach of selective liberation of soluble products of interest is reviewed.
cell disruption, intracellular product release, cell permeabilisation, high pressure homogenisation, high
speed bead mills, chemical lysis, enzymic lysis, differential product release 217
S. T. L. HARRISON 218 INTRODUCTION Increased interest is being shown in the efficient and cost-effective release of intracellular bacterial products from their host micro-organisms. This is due to the predominantly intracellular location of products of recombinant DNA technology, the overproduction of proteins as inclusion bodies and the potential use of the intracellular storage product poly-13-hydroxybutyrate (PHB) as a biodegradable
thermoplastic. The nature of the cell disruption process may influence the extent of product recovery, the ease of the subsequent purification steps, the nature of the suspensions to be processed and the
form and quality of the final product. These combined recovery processes are of prime importance in process economics. It is estimated that in the production of recombinant proteins, 45% of the equipment costs are associated with product recovery while only 14% can be attributed to the fermentation process (27). The ratio of variable costs of recovery to those of fermentation vary from I to 3 for enzyme and antibiotic recovery up to 10 for the recovery of intracellular recombinant insulin (19, 27).
Laboratory scale disruption processes have been well established for many years. However, processes
available for the large scale disruption of bacteria remain few. In addition, application of such processes to the production of commodity products is frequently cost-prohibitive. In this review, a brief description of the bacterial cell envelope is given to identify key components in its disruption. A
classification of cell disruption methods employed on both the laboratory and large scale follows in which assessment of the applicability of the processes to both large scale operation and specific product
recovery is given. The integrated nature of the process with operations both upstream and downstream
is considered. This initiates discussion of the recent developments in which the preferential release of the desired product is sought in place of general disruption and in which chemical or enzymic treatments are coupled to mechanical cell disruption to enhance its efficiency.
THE BACTERIAL CELL ENVELOPE The cell envelope of micro-organisms is a semi-rigid structure providing sufficient intrinsic strength to protect the cell from osmotic lysis. Additionally, it forms a biologically active boundary between the micro-organism and its external environment. In a typical Gram-negative bacterium, such as E.coli,
this consists of an elastic semi-permeable cytoplasmic membrane (innermost), an inter-membrane periplasmic space, a thin rigid wall layer comprised of peptidoglycan, and a lipid-protein outer membrane bilayer. Gram-positive bacteria such as Bacillus lack the outer membrane component, but in
tum possess a more dominant peptidoglycan structure. These envelope structures are readily distinguished by their staining characteristics. While the composition of these envelopes have been well
reviewed (24, 30, 35, 63), a brief summary is useful to identify the nature of the structures to be disrupted.
The cytoplasmic membrane provides the major interactive barrier between the internal cell environment
and the bulk media. Typically, the bilayer structure is approximately 4 nm wide (59) and is comprised predominantly of phospholipid and protein. This biophysico-chemical system actively maintains
BACTERIAL CELL DISRUPTION 219 concentration gradients, houses transport systems and is involved in ATP generation. It does not provide any significant structural strength and is readily disrupted by osmotic shock in the absence of the structural component layers.
The rigid peptidoglycan layer forms the basic structural framework of the cell envelope, providing its mechanical strength. The basic peptidoglycan structure, similar in all bacteria, is comprised of linear polysaccharide chains of alternating N-acetyl-0-glucosamine (NAG) and N-acetyl-muramic acid (NAM) residues joined by -(1-4) glycosidic bonds. The chains are cross-linked by a tetrapeptide of
the basic structure L-alanyl-D-glutamyl-L-R3-D-alanine attached to the C3 lactic acid side chain of the
NAM residue. The L-R3 amino acid residue may be one of several di-amino acids, frequently
diarninopirnelic acid. The peptide branches of the parallel chains are further cross-linked. The resulting
rigid grid structure acts as a single macromolecular network to provide the shape, tensile strength and osmotically protective nature of the cell envelope. Its strength is governed by the frequency of peptide
chains and their crosslinking (14, 24, 30, 35). The peptidoglycan layer in Gram-negative bacteria is 1.5 to 2.0 nm in thickness and typically accounts for 10 to 20% of the cell envelope in terms of dry mass. It is spatially distinct from the cytoplasmic membrane and forms the major resistance to cell breakage. The cell envelope of Gram-positive bacteria is composed of 50-80% peptidoglycan, associated with teichoic acids, presenting a greater structural resistance to breakage.
The outer membrane specific to Gram-negative bacteria is comprised of protein, lipopolysaccharide and
phospholipid. It separates the peptidoglycan layer from the bulk medium environment, thereby preventing their interaction. It is readily distinguished from the cytoplasmic membrane by the
anisotropic nature of its bilayer configuration and its narrow spectrum of protein molecules. Divalent cations play an essential role in its stabilisation (24, 35, 63).
In conclusion, three layers of the cell envelope require consideration in the rupture of Gram-negative bacterial cells. The outer membrane protects the inner layers from direct chemical attack. The peptidoglycan layer provides the mechanical strength of the cell. The cytoplasmic membrane may be considered the biochemical boundary of the cell and the major player in permeability.In Gram-positive
cells, the peptidoglycan layer is not protected from the external environment by an outer membrane, but
provides greater structural strength.
While this review is chiefly confined to the disruption of bacteria, it is useful to consider the gross distinctions between their cell envelopes and those of other micro-organisms of importance in biotechnology, in particular the yeasts. The basic structural components of the yeast cell wall are glucans, mannans and proteins which form a crosslinked polysaccharide-protein structure, the molecular details of which remain under debate (24). The overall structure is somewhat thicker than that of bacteria, typically 70 nm for Saccharomyces cerevisiae. As with bacteria, the role of the
membrane structure is predominantly biological. It is thus seen that while similarities may be drawn
between the mechanical disruption of these crosslinked polysaccharide-protein macromolecules on the
disruption of bacteria and yeasts, the applicability of chemical and biological disruption studies are confined to a particular wall structure.
S. T. L. HARRISON 220
CLASSIFICATION OF BACTERIAL DISRUPTION PROCESSES Broadly speaking, processes that have been applied to bacterial cell disruption can be classified as mechanical, physical, chemical or biological. The classification of techniques reponed in the literature
is summarised in Figure 1.
Mechanical methods Non-mechanical methods
I I II physical chemical biological as solid in suspension
II I -lysozyme-heat pressure mechanical -alkali
agitation -freezing -acid -Cytophaga I HPH -osmotic -detergents lysing enzymes
colloid mill shock -solvents -mutanolysin
impingement -dessication -EDTA -autolysis
-gas de- -antibiotics -phage
compression -chaotropic -wall inhibitors
grind pressure -ultrasonic agents
cavitationI I -bead mill Hughes press -hydrodynamic
-mortar+ X-press cavitation
Characterisation of Cell Djsruptjon Techniques.Fj ure 1
CELL DISRUPTION BY MECHANICAL MEANS Processes involving either solid shear (eg. bead milling, extrusion offrozen cells) or liquid shear (eg.
high pressure homogenisation) have been most frequently used in cell disruption. Some have found
application on a large scale. In general, the equipment used for large scale cell disruption has been
modified from that used for panicle size reduction or the formation of emulsions in other industries.
Common disadvantages of mechanical disruption include high capital investment and energy costs.
BACTERIAL CELL DISRUPTION 221 protein inactivation by shear-associated denaturation at air-liquid interfaces (8) and excessive heating due to energy dissipation.
Djspersjon and Colloid Mills Dispersion and colloid mills are used for the disruption of lightly bonded clusters and agglomerates as well as for the formation of emulsions. They operate on the principle of high speed fluid shear and cause very little grinding of individual particles. In general, the resultant particle size is 1 J.l.m (74). As
there is a high dissipation of energy, considerable heating occurs. Colloid mills may have cone or disc rotors that are smooth or grooved and rotate at some 3600 rpm. The rotor is separated from the fixed stator by a minimum of 25 11m. The disruption of microbial cells has been demonstrated. High
throughput rates (10m3/h) and continuous operation are possible, but complications of heating and
rotor-stator wear have been reponed (77).
Hi2h Pressure Homo2enisation High pressure homogenisation (HPH), employing a valve and impactor arrangement, is one of the most widely known methods for large scale cell disruption. A positive displacement piston pump is
used to draw the cell suspension through a check valve into the pump cylinder. On return of the piston, the suspension is forced through the adjustable annular gap of a discharge valve and impinges on an
impact ring. The discharge pressure, regulated by a spring-loaded valve rod, controls the position of the valve in relation to the valve seat.
The use of the HPH for microbial cell disruption has been mainly reponed for yeasts (23, 28, 43).
From the study of Saccharomyces cerevisiae using a pressure range of 10-54 MPa, a biomass
concentration of 84-210 kg dry massjm3 and a temperature range of 5-4QOC (43), the following were
i. Protein release (a measure of cell disruption) is temperature dependent.
ii. The process is independent of biomass concentration in the range 84 to 170 kg dry massJm3.
A dependence may be found at higher concentrations.
iii. In general loss of enzyme activity is not found on homogenisation (28).
iv. The rate of release of specific enzymes with respect to the extent of total protein release is dependent on the intracellular location of these enzymes with periplasmic enzymes most readily released (28). v. Protein release, R, is first order with respect to the number of passes, N, through the homogeniser:
_dR_k'R(1) dN- ?
vi. The dependence of protein release on operating pressure, P, can be expressed as a function of the pressure raised to an exponent.
S. T. L. HARRISON 222 A combination of (v) and (vi) results in the correlation (43):
(2) 1n Rm = kNPa( ) Rm- R
where R is the protein released (kg protein/kg biomass), Rm is the maximum protein available for
release, k is the rate constant and a is the pressure exponent. The applicability of these findings to the disruption of Gram-negative bacteria have been confmned in a study employing Alcaligenes eutrophus
in which cell concentrations of 95-260 kg dry mass/m3 were used (38). The dependence on pressure is
reduced with increasing operating pressure in excess of 50 MPa (23, 48). As disruption at these high pressures is largely independent of biomass concentration, the cell concentration should be kept as high as possible to achieve a high efficiency.
Table 1. The Effect of Micro-oreanism on Pressure required for Disruption by Hieh
Operating pressures quoted are those required to achieve 50% cell disruption on a single
pass using the Stansted cell disrupter (47).
Organism Type of Organism Shape Size Pressure applied
E.co/i Gram -ve bacterium X 0.5 15 2-4 rod
B.subtilis Gram +ve bacterium rod 1.5-3 X 0.8 24
L.casei Gram +ve bacterium rod 4 X 0.4-0.7 31
Gram +ve bacterium ovoid 1.02 diam. Streptococcus 150 cocci faecalis
S.cerevisiae yeast oval 7-12x5-8 150
fungi filament. 68 Aspergillus fumigaris
Fusarium sp. fungi filament 68
Chorel/a algae 48
Values of the constants, pressure exponent a and rate constant k of Equation 2 reported in the literature suggest a dependence of the level of cell disruption achieved on the nature of the organism and its culture conditions. Keleman and Sharpe (47) have used a homogeniser fitted with a ball and seat valve (Stansted cell disrupter) to determine the relative forces required to disrupt a range of microbial cells.
Their results are tabulated in Table 1. All micro-organisms investigated showed a sigmoidal profile
BACTERIAL CELL DISRUPTION 223
with respect to operating pressure with little disruption being observed below a threshold pressure. However, distinctly different pressures were required to disrupt different micro-organisms. Gram? negative bacteria are easier to disrupt than Gram-positive bacteria and filamentous fungi, which in turn
are easier to disrupt than unicellular yeasts. Ease of disruption appeared to be related to the cell wall
composition, size and shape and growth phase. The increased ease of disruption of the Gram-negative bacterium Alcaligenes eutrophus with increased growth rate has been clearly shown (38). In addition,
culture conditions such as media composition also influence the efficiency of breakage (33).
Ultramicroscopic studies of the disruption of Gram-negative bacteria by HPH (40) have indicated two
stages in the disruption process. In the initial stage, cell rupture is observed as discrete fractures in the electron dense peptidoglycan structure. The second phase involves the disintegration of the cell structure and the total liberation of its intracellular contents.
Conditions of high velocity shear, sudden decompression with resultant cavitational stresses, turbulence and impingement are expected to exist within the process. Rapid pressure release (11, 12, 22, 29), impingement on a stationary plate (23) and hydrodynamic cavitation (36) are apparent as the
significant causative mechanisms of cell rupture. By modification of the impact ring in the high pressure homogeniser, Keshavarz Moore et al (49) have illustrated a decrease in cell disruption with
increasing distance before impact. Indeed, only 20% of the disruption obtained after a single pass under normal conditions is achieved in the absence of the impact ring. This suggests that in the conventional valve and impactor design, cell rupture may be chiefly attributed to impingement on a stationary plate.
Commercial homogenisers may be operated at a maximum pressure of 55 MPa with a capacity of up to
53 m3Jh. For operation at higher pressures (100 MPa), the capacity is severely reduced (0.11 m3/h); hence multiple passes are generally required on a large scale. Multiple passes are accompanied by the micronisation of cell debris which may hinder further downstream processing.
Impjn ement Jets In fluid energy or jet mills, there is a high energy release and a high order of turbulence which causes particles to grind upon themselves and hence rupture (74). These jets may either impinge on a solid surface or, in a counter-current design, impinge on each other. Engler and Robinson (23) report efficient disruption of Candida utilis by the impingement of a high velocity jet of suspended cells
against a stationary surface and show it to be a first order process with respect to number of passes. Indeed, it is postulated to provide the primary mechanism of cell rupture in the high pressure homogeniser (49). Recently the effective use of counter current impingement jets in the disruption of both yeast and bacteria has been favourably compared to HPH (50). Factors influencing the process include nozzle geometry, inter-nozzle distance, jet velocity, number of passes and morphology of the micro-organism. A 60% disruption of wild type E.coli and a 95% disruption of a recombinant E.coli
are reported on a single pass.
224 5. T. L. HARRISON Hieh Sneed Bead Mills
The bead mills, originally developed for the pigment and dye-stuff industry, provide grinding and dispersion by inter-particle collision and solid shear. The bead mill consists of either a vertical or a horizontal grinding chamber containing rotating discs or impellers mounted on a motor driven shaft. These accelerate the glass or plastic beads to supply a grinding action. The grinding chamber has a sieve plate or similar device to separate and retain the beads. An efficient cooling system, usually in the
form of a cooling jacket (and possibly a cooled impeller shaft), is required to dissipate the heat generated. Horizontal units are preferred for cell disruption as the grinding action in the vertical units may be reduced by the fluidising effect of the upward fluid flow on the beads (15, 48).
Studies of cell rupture using high speed bead mills have been carried out on yeasts (17, 41, 56, 58,
69), and both Gram-positive and Gram-negative bacteria (41, 69). This complex process is influenced
by a wide range of parameters relating to the number and energy of impacts taking place, the energy transfer to the grinding elements, liquid shear, hydrodynamics and mixing. These parameters include bead diameter, density and loading, cell concentration in feed, flow rate of feed, agitator speed and configuration, geometry of the grinding chamber and temperature. In addition, the residence time, nature and state of the cell envelope, and size of different organisms will affect the process.
In general, smaller bead sizes are more effective. Typically for yeast cells, beads of 0.2-2.8 mm diameter are used with the range 0.25-0.5 mm being preferred. By using larger beads, enzymes located
in the periplasmic space can be preferentially released whereas smaller beads are required for the release
of cytoplasmic enzymes (69). Disintegration of bacteria in a bead mill is hampered by their size. Gram?
negative bacteria such as E.coli and A.eutrophus occupy approximately 1% of the volume of a yeast
cell. Reduction of bead size is required for efficient disruption. This is limited by the tendency of small
beads to float. The outcome is the necessity for repeated passes or increased residence time (Table 2)
and a decrease in processing rate by as much as a factor of 10 (69).
Cell breakage shows first order kinetics in which the rate of protein release is directly proportional to the amount of unreleased protein. In a batch reactor, this is represented by:
(3)m(R:R) = kt
where R is the protein released, Rm is the maximum protein available for release, t is the time of
treatment and k is a first order rate constant (17, 56). For continuous operation, this has been
expressed by viewing a mill as a series of CSTRs as seen in Equation 4 (56):
+ (f k ?d)1 (4)(R:R) = j
225BACfERIAL CELL DISRUPTION
where 't is mean residence time and j is the number of CSTRs in series. If the mill operates as a plug flow reactor as illustrated by Mao and Moo-Young (58), factor f is unity. Should back mixing occur, f
is fractional. The rate constant is dependent on agitator speed, flow rate of suspension, yeast concentration, bead size, bead volume and temperature. This subject has been extensively reviewed (15, 24, 51, 69).
The advantage of high speed bead mills in the recovery of a granular product is the thorough disintegration of the cells ensuring that the granules are released from the peptidoglycan structure. The
drawbacks for bacterial rupture include reduced suitability for small microbial cells and the micronisation of cell debris, thereby complicating further separations.
Table 2. Effect of Type and Size of Mjcro-oreanjsm on Cell Djsruptjon jn a Hi :h
Speed Bead Mm (65).
Micro-organism Residence Time for Complete Disruption
Aerobacter aerogenes 96
Bacterium cyclo-oxydans 35
Escherichia coli 85
Penicillium chryogenum <20
Saccharomyces cerevisiae <17
Streptomyces noosus 85
Solid Pressure Shear A complete and very efficient cell disintegration and general preservation of product properties is achieved by pressing a frozen cell paste through a small slit or orifice under pressure. The X-press, Hughes press and Chaikoff press all operate on this principle. Using a starting temperature of -25 to -27?C and pressure of up to 550 MPa, a change of the ice crystal state from ice I to ice III with a 20% decrease in volume will occur. This change in state, the abrasive action of the ice crystals and plastic flow through the orifice are believed to be responsible for the cell rupture obtained (70). On a single pass 90% disruption of S.cerevisiae is achieved.
These processes are efficient in rupturing a wide variety of biological material, even that of tough cell
wall structure. Lower contamination of cell fragments is achieved than in other devices (24, 46, 75).
226 S. T. L. HARRISON Semi-continuous operation of the X-press at 10 kg/h has been reponed (15). These low mass flowrates indicate that this equipment remains limited to pilot plant applications thus far.
CELL DISRUPTION BY PHYSICAL MEANS
Vapour cavities can be formed in a liquid due to a local reduction in pressure. This may be caused by a local increase in velocity (Bernoulli), rapid vibration of a boundary, ultrasonic vibrations, the separation of a liquid column or the overall reduction of the static pressure. Subsequent collapse and rebound of the cavities will occur until an increase in pressure causes their destruction. This process is
termed cavitation. Cavitation is associated with local pressure fluctuations of the order of 1000 MPa. On the collapse of cavitation bubbles, a large amount of energy is released as mechanical energy in the form of elastic waves which disintegrate into eddies. Cavitation is well-known for achieving a fine
dispersion of fragmented solid panicles or gas bubbles in a liquid and for the damage it causes to
pumps and pipework. Two methods of generating cavitation have been considered with respect to cell breakage: ultrasonic cavitation and hydrodynamic cavitation.
Doulah (20) proposed a disintegration mechanism for yeasts subjected to cavitation based on a drop breakup mechanism in hydrodynamic fields in which breakage occurs when the dynamic pressure difference across the drop exceeds its surface energy. Eddies of scale larger than the micro-organism will move it from place to place, whereas eddies of a smaller scale will impan motions of various intensities to the cell. Hence a pressure difference will be created across the cell. When this exceeds the cell wall strength, the cell will disintegrate. As cell wall strength is generally unknown, a surface? tension type force, s, was assigned to represent it:
e s(5)Esur =d where Esur is cell surface energy, 9 is a shape factor and the cell dimension (equivalent spherical dis
diameter). A largest stable cell size (dimension dm) can then be described as a function of the energy dissipation rate. While the resultant relationships illustrate the dependence on cell size shown
experimentally, the analogy of cell wall strength and surface tension is fundamentally unsound. The cell wall is a structural component gaining its strength from its chemically bonded nature. Surface tension, however, is attributed to the lower potential energy of a solvent molecule when in the bulk solvent than when in a gaseous state or immiscible solvent owing to attractive forces and hydrogen bonding between the molecules. Additionally, on the division of a droplet maintained by surface tension, a number of like entities are formed. It has been clearly shown by transmission electron microscopy of disrupted bacterial cells that this does not occur on the rupture of the peptidoglycan cell wall (36).