Ezgfio Tm. h _ Mi
. , . ‘rivto I
. :7. :1. .332»: d."
.uv.§..!,r-:e4lv c Fol.“
1 \zlrtl {V
.\. iii): tip».
il pnvltivfio ttnclclil...
1!!! l: .
0.. .b . a. v.
14.. :51!!!-

vtq u
. . vii-5.3.“
5.1.1.: 2. 13.1....» ._

("Y

 

.. rt»; 1 4...
t. .. . 1%.?

arraitr’.
9'... 1r]...
9 A. a.

.q r 3..»

I III I}. \u\

.12.. -7. s.
r

{lisiavs 4
Y. #5....

 

 

1-??va

MICHIGAN STATE U we

W I Ill/Illllllllllf l/ff

 

 

WWII/7 F"

lWilli/IIll/ll/HI/l/I a L

3 1293 00910 9962

 

 

This is to certify that the

dissertation entitled
The Use of Column Flotation in the Downstream

Processing of Fermentation Products
presented by
Everson Alves Miranda

has been accepted towards fulfillment
of the requirements for

Ph.D. degreein Chem. Eng.

 

 

fiaw

Major professor

Date 13/?” 3

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

! LIBRARY 7
1M. chlgan Stats}
L University

*.

 

l

 

 

RETURN BOX to remove this checkout from your record.

PLACE IN
TO AVOlD FINES return on or before date due.
DATE DUE DATE DUE DATE DUE “

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, - \T’T

MSU Is An Affirmative ActiorVEqual Opportunity Institution”
3.9.1

 

 

 

a....—-—

____—-—'-"'——

#—

 

THE USE OF COLUMN FLOTATION IN THE DOWNSTREAM
PROCESSING OF FERMENTATION PRODUCTS

BY

Everson Alves Miranda

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemical Engineering

1993

ABSTRACT

THE USE OF COLUMN FLOTATION IN THE DOWNSTREAM
PROCESSING OF FERMENTATION PRODUCTS

By

Everson Alves Miranda

Flotation has been considered a unit operation with a largely
unexplored potential in biotechnology. There is a general lack of studies
concerning its application in the recovery of fermentation products. This A
work is the study of column flotation for the recovery of products of
fermentation.

Two approaches for the application of column flotation in
biotechnology were addressed in this work. First, a unique separation of a
product was accomplished by recognizing the natural floatability of a
particulate separation system previously determined. A process to
recover Bacillus sfearothermophilus rat-amylase expressed in Escherichia
coli from fermentation broth and periplasmic extract had been developed
(Summerfelt, 1988). The solid-liquid separation of the salted-out
particles of (hydroxypropyl)methylcellulose by filtration and
centrifugation was inefficient. Column flotation was able to recover
virtually 100 percent of the enzyme partitioned to the salted-out particles
with no loss of enzymatic activity. The recovery of the extracellular

enzyme directly from the fermentation broth required the elimination of

small molecular weight compounds by ultrafiltration for improved
recovery.

The second approach for application of flotation in biotechnology,
the design and preparation of a flotation system to perform separation of
proteins based on specific type of interaction between a product and a
ligand, was also developed in this work. Hydrophobic interaction between
albumin and hydrophobic groups was selected for the system developed.
First, derivatives of methylcellulose with fatty acids were prepared by
esterification. After characterization, the derivatives were tested in the
precipitation of bovine serum albumin. One of the derivatives was shown
to be efficient in the precipitation of the protein. The predictability of
the recovery of the protein by flotation was good despite low floatability

due to the large concentration of salt and protein.

References:
Summerfelt, S. T. The concentration and the purification of an a-amylase
using a methylcellulose-salt, two phase partitioning process. M. S. Thesis,

Michigan State University, 1988.

Gravity? Why fight it?

Fighting gravity is a losing proposition, and yet
that is precisely what conventional flotation does.
It fights gravity.

The column does not.

D. A. Wheeler

Acknowledgements

I would like to thank my advisor, Dr. Kris Berglund for the friendship
and freedom he provided to develop a research project of my choice.
Thanks also to the members of my committee, Dr. Oriel, Dr. Steffe, and Dr.
Worden, for their encouragement and participation in fruitful discussions.

I am in great debt to my family for their patience and support during
all these years.

I would also like to thank all my friends who made it possible to
overcome the difficult times. I will always remember them.

Finally, I would like to express my deep appreciation for the
financial support provided by the Conselho Nacional de Desenvolvimento

Cientifico e Tecnolégico - CNPq, Irazil.

TABLE OF CONTENTS

Page

LIST OF TABLES ............................................................................................................... i x
LIST .OF FIGURES .............................................................................................................. xi
CHAPTER 1 : INTRODUCTION AND OBJECTIVES .................................................. 1
1.1 Scope : Biotechnology and Downstream Processing ......................... 1
1.2 Flotation and its Potential in Biotechnology ...................................... 4
1.2.1 Flotation as a Unit Operation ......................................................... 4

1.2.2 Flotation in Biotechnology .............................................................. 7

1.2.3 Potential of Flotation in Biotechnology .................................... 9

1.3 Column Flotation : the New Concept of Flotation .............................. 13
1.4 Objective .............................................................................................................. 13
1.5 List of References ............................................................................................ 17
CHAPTER 2 : FLOTATION, AN ADSORPTIVE BUBBLE SEPARATION ................ 22
2.1 Adsorptive Bubble Separation Techniques ............................................ 22
2.1.1 Concept and Classification .............................................................. 22

2.1.2 Non-Foaming Techniques : Bubble Fractionation .................... 25

2.1.3 Foam Separation ................................................................................... 26
2.1.3.1 Foam Fractionation ............................................................. 26

2.2 Flotation and Column Flotation ................................................................. 27
2.2.1 Basic: of Flotation ............................................................................. 27

2.2.2 Column Flotation .................................................................................. 40

2.2.2.1 Operational Principles and Advantages .................... 40

2.2.2.2 Operational Parameters ................................................... 46

2.2.2.3 Modeling and Scale-Up ................................................... 50

2.2.2.4 New Concepts and Design Variations ......................... 52

2.3 List of References .......................................................................................... 55
CHAPTER 3 : RECOVERY OF A GENETICALLY ENGINEERED a-AMYLASE ....... 62
Abstract ....................................................................................................................... 62
3.1 Introduction ........................................................................................................ 62
3.1.1 Expression of Heterologous Protein ............................................. 62

vi

Page

3.1.2 Fusion Proteins and Purification Fusion .................................... 65
3.1.3 a-Amylase/Alkaline Phosphatase Fusion Proteins ............ 66
3.1.4 The Recovery Process by HPMC/Salt Partitions ..................... 69
3.1.5 The Problem ............................................................................................ 72
3.1.6 The Proposed Solution ........................................................................ 73
3.2 Material and Methods ...................................................................................... 74
3.2.1 Cellulose Ether and Other Chemicals ........................................... 74
3.2.2 Organism, Cultivation, and Broth and Periplasmic
Extract Preparation ............................................................................. 74
3.2.3 Diafiltration ........................................................................................... 75
3.2.4 Enzyme Partition and HPMC Salting-Out 75
3.2.5 Continous Flotation : Systems and Runs .................................... 77
3.2.6 Batch Flotation : System and Runs ............................................... 80
3.2.7 Enzyme Activity Assay ...................................................................... 81
3.2.8 Total Carbohydrates Assay .............................................................. 82
3.2.9 Other Experimental Methods ............................................................ 83
3.3 Results .................................................................................................................. 83
3.3.1 Experiments with Synthetic Mixtures ......................................... 83
3.3.2 Experiments with Broth Mixture and Periplasmic
Extract ....................................................................................................... 88
3.4 Discussion ........................................................................................................... 98
3.5 Conclusions ........................................................................................................ 105
3.6 List of References ............................................................................................ 107
CHAPTER 4 : AFFINITY FLOTATION BASED ON LlGAND-LIGATE
HYDROPHOBIC INTERACTION ............................................................ 1 12
Abstract ...................................................................................................................... 112
4.1 Introduction and Objectives ....................................................................... 1 12
4.2 Hydrophobic Interaction ............................................................................... 117
4.3 Hydrophobic Interaction Chromatography of Proteins .................... 121
4.4 Materials and Methods ................................................................................... 130
4.4.1 Chemicals ................................................................................................. 130
4.4.2 Methods ..................................................................................................... 130
4.4.2.1 Esterification of Methylcellulose ............................... 130
4.4.2.2 Apparent Solubility of the Derivatives .................... 131
4.4.2.3 Solutions of the Derivatives for BSA
Precipitation ........................................................................ 1 32
4.4.2.4 Salting-Out of the Derivatives .................................... 132
4.4.2.5 Flotation : System and Runs ........................................... 132
4.4.2.6 De-fatting of BSA ............................................................... 135
4.4.2.7 Precipitation of BSA with Salt .................................... 136
4.4.2.8 BSA Precipitation with Derivatives at
Different pH .......................................................................... 136

vii

Page

4.4.2.9 Time Course of Precipitation of BSA

 

 

with C4 Derivative ............................................................ 1 3 7

4.4.2.10 BSA Precipitation at Different Salt
Concentration ....................................................................... 137
4.4.2.11 Total Carbohydrates Assay ............................................ 138
4.4.2.12 Protein Assay ....................................................................... 138
4.4.2.13 Infra-red Spectra ............................................................... 138
4.5 Results ................................................................................................................ 139

4.5.1 Esterification of Methylcellulcse and

Characterization of the Derivatives .......................................... 139
4.5.2 Precipitation of BSA with Methylcellulose Derivatives 154
4.6 Discussion ..... - _- -_ ............ - 160
4.7 Conclusions ........................................................................................................ 167
4.8 List of References .......................................................................................... 169
CHAPTER 5 : GENERAL CONCLUSIONS AND RECOMMENDATIONS ................... 174
5.1 General Conclusions ..................................................................................... 174
5.2 Recommendations ........................................................................................... 175
APPENDIX .......................................................................................................................... 177

viii

CHAPTER 3

Table 3.1

Table 3.2

Table 3.3

Table 3.4

CHAPTER 4

Table 4.1

Table 4.2

Table 4.3
APPENDIX
Table A.1
Table A2
Table A3
Table A.4

Table A.5

LIST OF TABLES

Page
Carrier moieties for protein fusion ............................................. 68
Particle size distribution of HPMC particles
suspension used for experiments illustrated
in Figure 3.4 ............................................................................................. 85
Flotation of periplasmic extract and the effect of
lowering the ammonium sulfate concentration in
the washing solution ........................................................................... 95
Temperature effect on the flotation recovery
of HPMC from broth .............................................................................. 99
Comparison among flotability of the derivatives
and unsubstituted methycellulose .................................................. 153
Effect of temperature on the precipitation
of BSA with C4 derivative ................................................................. 161
Flotation of the BSA-derivative precipitates .......................... 162
Raw data off-"igure 3.4 ........................................................................ 177
Raw data of Figure 3.5 ........................................................................ 177
Raw data of Figure 3.6 ........................................................................ 178
Raw data of Figure 3.7 ........................................................................ 1 79
Raw data of Figure 3.8 ........................................................................ 1 79

ix

Table A.6

Table A.7

Table A8

Table A9

Table A.10
Table A.11
Table A.12
Table A.13
Table A.14
Table A.15
Table A.16
Table A.17
Table A.18
Table A.19

Table A.20

Page

Raw data of Figure 3.9 and Figure 3.10 ...................................... 180
Raw data of Figure 3.11 ...................................................................... 1 8 0
Raw data of Figure 3.12 ...................................................................... 1 81
Raw data of Figure 4.9 ........................................................................ 1 81
Raw data of Figure 4.10 ..................................................................... 1 81
Raw data of Figure 4.1 1 ..................................................................... 1 8 2
Raw data of Figure 4.12 ..................................................................... 1 8 3
Raw data of Figure 4.13 ...................................................................... 1 84
Raw data of Figure 4.14, Figure 4.15, and Figure 4.16 ........ 185
Raw data of Figure 4.17 ..................................................................... 1 8 7
Raw data of Figure 4.18 ..................................................................... 1 8 8
Raw data of Figure 4.19 ..................................................................... 1 8 9
Raw data of Figure 4.20 ..................................................................... 1 9 0
Raw data of Figure 4.21 ..................................................................... 1 9 0
Raw data of Figure 4.22 ..................................................................... 1 91

CHAPTER 1

Figure

Figure

Figure 1.3

Figure 1.4

1.1

1.2

CHAPTER 2

Figure
Figure
Figure
Figure

Figure

Figure

2.1

2.2

2.3

2.4

2.5

2.6

LIST OF FIGURES

Page
General flow sheet of a typical production process in
biotechnology (from Prave etal., 1987) .................................. 3
Schematic diagram of a flotation cell (from
Scmasundaran and Ananthapadmanabhan, 1987) .................. 6
General scheme of a flotation column (from
McKay et al., 1988) ............................................................................. 14
Some different types of flotation columns found in the
biotechnology literature .................................................................. 1 6
Classification of the adsorptive bubble separation
techniques (modified from Lemlich, 1972) ............................ 24
Operational modes of foam fractionation (from
Thomas and Winkler, 1977) ............................................................ 28
Illustration of particle capture mechanism where
0 <t1 < 12 < t3 ...................................................................................... 30
Contact angle and hydrophobicity (from Lemlich,
1972; Svarovsky, 1 981 ) .................................................................... 3 2
Effect of zeta potential of the bubbles and latex
particles (C) on flotation efficiency (11) in a
cationic surfactant solution, with change of pH
(from Okada etal., 1990) ................................................................. 35

Schematic representation of the double layer according
to Stern's model (from Scmasundaran, 1975) ........................ 41

x i

Figure 2.7

CHAPTER 3

Figure

Figure

Figure

Figure

Figure

Figure

3.1

Page

Froth structure. 8 is the fractional gas holdup
(from Yanatos ef al., 1986) ............................................................. 44

Examples of gene fusions used for the expression of a
recombinant protein (X). Promoters are symbolized by
arrows and transcription sequences by rectangles. The

boxes represent signal sequences (S), different affinity
“handles“ (A and B), and a domain for insertion into
membranes or cell wall (I) (according Uhlén and Moks,

1990) 67

 

Effect of ammonium sulfate in the partition of

iii-amylase (0), partition of total protein ( El ), and
salting-cut of HPMC (A). Data from Miranda (1988)

and Summerfelt (1988) ................................................................... 70

Basic design of column #2. Measurements in
millimeters ........................................................................................... 78

Response of HPMC flotation to particle size. Partition

of synthetic mixture : batch volume, 4 L; ammonium

sulfate addition flow rate, 5 and 20 mL/min for 10

and 16 pm, respectively. Continuous flotation with

no washing : column used, #1; feed flow rate,

27 mL/min; froth height, 0.5 to 3 cm ........................................ 86

Effect of ammonium sulfate concentration on the

recovery of HPMC by flotation. Salting-out of

synthetic mixture : batch volume, 1.5 L; HPMC

concentration before salt addition, 0.07 percent;

ammonium sulfate addition flow rate, 14 mL/min.
Continuous flotation with no washing : column used,

#1; feed flow rate, 36 mL/min; froth height, 1 cm ............ 87

Determination of apparent column capacity. Partition

of synthetic mixture : batch volume, 10 L; ammonium

sulfate addition flow rate, 60 mL/min. Continuous

flotation with no washing : column used, #1; froth

height,3cm -- ...... 89

 

xii

Figure 3.7

Figure 3.8

Figure 3.9

Figure 3.10

Figure 3.11

Page

Effect of washing on concentrate purity. Partition of
synthetic mixture : batch volume, 10 L; ammonium

sulfate addition flow rate, 60 mL/min; tracer, 2.0

percent copper sulfate (CuSO4.5HzO), 0.4 percent

sodium hydroxide, and 2.0 percent sodium potassium

tartrate. Continuous flotation : Column used, #1; feed

flow rate, 40 mL/min; froth height, 4 cm; washing

solution, 2.0 percent sodium potassium tartrate in

30 percent saturation ammonium sulfate ................................. 90

Effect of pH on flotation

recovery of HPMC from synthetic mixture by

batch flotation. Partition: batch volume, 150 mL;

HPMC concentration, 0.1 percent; ammonium

sulfate concentration and addition rate, 30 percent
saturation and 3.3 mL/min, respectively. Flotation:

air flow rate, 4.0 mL/min; make-up water flow rate,

1.6 mL/min; flotation time, 5 minutes. .................................... 91

Flotation response to broth pre-purification by

diafiltration. Partition of broth mixture : batch

volume, 1.5 L; ammonium sulfate addition flow rate,

31 mL/min. Continuous flotation without washing :

column used, #2; feed flow rate, 20 mL/min; froth
height,Scm ........................................................................................... 93

Correlation between actual activity recovery and

theoretical activity recovery calculated as 0.912 x

HPMC recovery. Data from experiment depicted on

Figure 3.9 ................................................................................................ 94

Effect of pH on flotation recovery of HPMC from

broth mixture by batch flotation. Partition: batch

volume, 150 mL; HPMC concentration, 0.1 percent;

ammonium sulfate concentration and addition rate,

30 percent saturation and 3.3 mL/min, respectively.
Flotation: air flow rate, 4.0 mL/min; make-up water

flow rate, 1.6 mL/min; flotation time, 5 minutes. ................ 96

xiii

Figure 3.12

CHAPTER 4

Figure

Figure
Figure
Figure

Figure

Figure

Figure

Figure

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

pH effect on recovery of HPMC particles, activity
associated with HPMC particles, and total activity.
Partition of broth mixture : batch volume, 4 L;
ammonium sulfate addition flow rate, 42 mL/min.
Continuous flotation : column used, #2; feed flow rate,
20 mL/min; washing solution flow rate, 7 mL/min;

froth height, 7 cm .............................................................................

Schematic representation of hydrophobic interaction.
The situation refers to binding of a protein to a solid
matrix as in hydrophobic chromatography

(from Mchr, 1985) .............................................................................

Reactions used to couple ligands to agarose (from

Pharmacia Fine Chemicals, 1983) ..............................................

Homologous series of alkyl agaroses which can be
prepared for use in hydrophobic chromatography

(from Shaltiel eta/., 1978) ...........................................................

Schematic illustration of how a flexible linear ligand
can accommodate itself in different hydrophobic sites

on a protein molecule (from Shaltiel ef al., 1978) ............

Effect of length of a linear ligand on the
adsorption (O) and desorption (0) of glycogen
phosphorylase b on a hydrophobic agarose

(from Shaltiel, 1978) ......................................................................

Illustration of the role of the length of the spacer
in the binding efficiency of a ligand to a protein.
a) spacer is too short; b) spacer is of right size; c)

and c') spacer is too long (from Scouten, 1981) .................

Basic design of column flotation. Measurements

in millimeters ....................................................................................

Typical spectrum of methylcellulcse derivative.
Fatty acid, butyric acid; volume of sulfuric acid,
5 uL/g of methylcellulose; temperature, 65 °C;

reaction time 15 hours ...................................................................

xiv

Page

97

.. 120

.124

.125

. 126

.126

.128

133

.. 141

Figure 4.9

Figure 4.10

Figure 4.11

Figure 4.12

Figure 4.13

Figure 4.14

Figure 4.15

Figure 4.16

Page

Effect of sulfuric acid as a catalyst on the
esterification of methylcellulose using cctancic acid;
catalyst 20 uL/g of methylcellulose; temperature, 65 °C..142

Effect of temperature and amount of sulfuric acid
on the esterification of methlcellulose using cctancic
acid ............................................................................................................... 143

Reactivity of different fatty acids with respect

of the esterification of methylcellulose. Volume

of sulfuric acid, 5 uL/g of methylcellulose;

temperature, 65 °C; reaction time 24 hours ........................... 145

Kinetics of esterification of methylcellulose by
fatty acids. Volume of sulfuric acid, 5 uL/g of
methylcellulcse; temperature, 65 °C ........................................ 146

Hydrolysis of methylcellulose during esterification
measured as decrease in yield. Volume of sulfuric acid,
5 uL/g of methylcellulcse; temperature, 65 °C .................... 147

Apparent solubility of the derivatives (C3 to 014).
Derivatives were treated according to the procedure

to dissolve methylcellulose as for the preparation of

a 1 mg/mL solution. The concentration of derivative

in the liquid phase was determined and was called

apparent solubility. ........................................................................... 148

Apparent solubility of: the derivatives (03 to C8 and

C8 to C14). Derivatives were treated according to

the procedure to dissolve methylcellulose as for the
preparation of a 1 mg/mL solution. The concentration

of derivative in the liquid phase was determined and

called apparent solubility ................................................................ 1 4 9

Differences in the extent of reaction of soluble and
insoluble fractions of different derivatives. Soluble

and insoluble fractions of the derivatives resulted

from experiment showed in Figures 4.14 and 4.15

were dried and their extent of reaction measured.

Open symbols, soluble fraction; closed symbols,

insoluble fraction. .............................................................................. 1 51

XV

Figure 4.17

Figure 4.18

Figure 4.19

Figure 4.20

Figure 4.21

Figure 4.22

Page
Salting-cut of derivatives and unsubstituted
methylcellulose. Soluble fractions of the derivatives
resulted from experiments showed in Figures 4.14 and
4.15 were salted-cut with ammonium sulfate. The
concentration given by the remaining amount of
derivatives in the solution was measured and
compared with the original concentration ................................. 152

pH precipitation curves for BSA. 0 - 60 percent

saturation of ammonium sulfate; A - no ammonium

sulfate. Protein concentration, 0.66 mg/mL; Aging time, 2
hours. The dashed line is the expected absorbence

in the case of no precipitation ......................................................... 155

Precipitation curves for BSA in the presence of

derivatives (0, C4; A, 08; 13, C12) and unsubstituted
methylcellulose (O). Cn/BSA = 2; Protein

concentration, 0.66 mg/mL; salt concentration,

60 percent saturation; aging time, 2 hours. The

dashed line is the expected absorbence in the case of

no precipitation ...................................................................................... 156

Time course for precipitation of BSA with C4

derivative. Cn/BSA = 1.5; salt concentration,

60 percent saturation; pH, 6.0. The dashed line

is the expected absorbence in the case of no

precipitation ............................................................................................ 157

Effect of ammonium sulfate concentration on the

precipitation of BSA with C4 derivative. Cn/BSA = 2;

Protein concentration, 0.66 mg/mLpH,

6.0; aging time, 1 hour ......................................................................... 158

Effect of the ratio C4/BSA on the precipitation of

BSA with C4 derivative at different ammonium sulfate
concentrations (0, 60 percent saturation; A, 55 percent
saturation; D, 50 percent saturation; or, 45 percent
saturation). Protein concentration, 0.66 mg/mL; pH, 6.0;
aging time, 1 hour .................................................................................. 159

xvi

Chapter 1

Introduction and Objectlve

1.1 - Scope: Biotechnology and Downstream Processing
Biotechnology, a science as old as history, deals with the use of
animal cells, plant cells and microorganisms to produce products useful to
humans. It was at the beginning, thousands of years ago, an unconscious
use in the production of ethanol in the form of beer or wine, curdled milk
products, cheese, and East Asian and Oriental foods. The pioneer work of
Pasteur and other scientists showed that microorganisms were capable of
metabolic reactions of practical utilization. What followed these findings
was the development of manufacturing processes for primary metabolites,
such as glycerol, citric acid, and butanol/acetone. A characteristic of
such processes was their ability to be run without “absolute” prohibition
of foreign infections. The next phase of biotechnology was the
development of processes that were run with the exclusion of unwanted
microorganisms. This opened the era for the production of compounds
formed only in very small amount by cells. Microbiological
transformations could be done and many secondary metabolites such as
vitamin B12, cortisone/cortisol, and estrogenic ovulation inhibitors could
be manufactured. The most recent step in the evolution of biotechnology
was the birth of the genetic engineering that provided a degree of control

over microorganisms never deemed possible. Products can now be

 

“tailored-made” and microorganism growth and action on substrate
optimized (Prave et al., 1987).

Biotechnology is the result of the interweaving of many different
areas of knowledge. Among them the three more relevant are
microbiology, including microbial genetics, biochemistry, and process
engineering including equipment construction.

One characteristic of biotechnology industries is the variety of the
products involved. While a typical petrochemical company produces around
10 products, a typical drug company produces more than 200 which are, at
least in part, made biochemically (Belter et al., 1988). Regardless if a
substance is produced by fermentation or extraction (from animal tissues
or plants), it is often dilute and contaminated with other bioproducts.
Concentration and purification of the desired product (or products) to a
practical, economical, and useful level is accomplished by a series of
process steps named downstream processing.

Downstream processing is carried out by the use of well-known unit
operations of chemical engineering such as liquid-liquid extraction as well
as by methods traditionally used in biochemical laboratories such as
chromatography and electrophoresis. The scale of these processes vary
from hundreds of milliliters to hundreds of cubic meters. The operations
can be continuous or batch; steady- or unsteady-state; cocurrent or
counter-current contacting. They may involve physical separations, rate
controlled separations, and equilibrium controlled separations. Figure 1.1
shows a general flow sheet for a typical production process in
biotechnology.

The viscous, heat unstable, and highly non-Newtonian feedstock, the

low product concentration, and the chemical and physical similarities

 

 

 

 

Preparation of the "—- H20
1mm :——-Mk'ngtechnolcgy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Am of salts. .. m
soirces of P and N. etc. 1
I
Prol'feraticn of the :
production stran |
I
L 5mm" 1 Inocuaticn matenal :
I
I
5W ~ II {I #~ CO,andothergases E
Aeration * Fermentation H“, I
Measueinentandcontrol —" :
I
1’ Residual culture I
aqua oriand
Filtration _... Separation of i———.-i , __...
. biomass and m
[ Product isolation 1—-—-Ferrnentaticn rescues -—-w
PTOOIJCT WWI. ‘ .
I even -—Fermentation rescues —--'
| Product ] L Residue disposal

 

 

 

Figure 1.1 - General flow sheet of a typical production process in
biotechnology (from Prave et al., 1987).

among product and contaminants make downstream processing a difficult
task. Typically, product specifications are attained only after different
separation and/or concentration methods are applied, sometimes more
than once. Product loss, space, and capital and Operating costs often make
downstream processing the major component in the economics of
biotechnological processes. In fact, a company's competitiveness may
depend on how well it accomplishes downstream processing. (Belter et al.,
1988; Dechow, 1989; Parker et al., 1990; Wheelwright, 1987; Zlokarnik,
1990)

The search for simple, efficient and low cost downstream processes
is a present challenge for the biochemical engineer (Kelly, 1987). The

proposed research is one potential answer to this challenge.

1.2 - Flotatlon and Its Potential In Biotechnology
1.2.1 - Flotatlon as a Unit Operation

Flotation is a technique mostly used in the selective separation and
concentration of particles, mainly minerals. It is based on the ability of
particles in a aqueous suspension to attach themselves to gas bubbles
which carry them up in the liquid phase due to buoyancy. The removal of
these particles as they reach the top of the liquid phase along with only a
small amount of liquid results in a concentration effect (solid-liquid
separation). If particle attachment is selective towards a specific class
of particle, enrichment takes place.

Bubble generation is done by air dispersion (the most common
procedure), by reducing air solubility with pressure reduction, or by

electrolysis. These bubble generation methods characterize what is

referred as dispersed-air flotation, dissolved-air flotation, and
electrolytic flotation, respectively (Matis and Mavros, 1991).

There are chemical agents that can change the surface properties of
particles making possible the design and control of the specificity and
yield of the separation. By the use of appropriate surfactants, particles
can have their hydrophobicity increased. These surfactants are called
collectors. Agents used to enhance or prevent interaction of the particles
with collectors are called activators or depressors, respectively. Those
agents which make the liquid foam are called frothers.

Flotation is conducted in special equipment called flotation cells.
Leja reported the existence of more than 40 different designs of such cells
which are variations of three basic types: mechanical cell, pneumatic cell,
and cyclone cell (Leja, 1982). Figure 1.2 is a scheme of a mechanical cell.
The impeller is a bubble generation device and keep particles in suspension
in order to make particle-bubble contact and attachment possible. There
is also a skimming device of some sort at the top of the cell to remove the
froth which is rich in hydrophobic particles. The other two types of cells
differ from the mechanical one mainly by the absence of an impeller. ln
pneumatic cells, feed is introduced over the top of the froth bed which is
generated by air injection through a array of tubes located under this bed.
In cyclone cells, highly pressurized feed together with air is introduced
tangentially in a cyclone feeder located inside the cell that otherwise look
similar to a mechanical cell.

The importance of flotation to the economy of the industrialized
world can be illustrated by the amount of mineral materials floated in the

United States: about 1.1 millions tones of ores and raw coal in 1985

 

. I,“ "_-.'-'.":5:'-;7- FROTH
' - . CONTAINING
z HYDROPHOBIC
8 ENTRAPPED
MATERIALS

    
      

E71: TAILS
“—2.: CONTAINING
UNFLOATED

IMPELLER MATERIAL

 

Figure 1.2 - Schematic diagram of a flotation cell (from Scmasundaran
and .“ ‘“ ‘ an...” 1987).

r

 

(United States Bureau of Mines, 1985). This primary use of flotation to
treat commodities of very low-value, ores, is confirmation of one the
most important characteristics of flotation, its low cost. Other
advantages are selectivity and versatility.

The first mention of a separation technique we could call flotation,
a copper carbonate enrichment, dates back to the 15th century (Kitchener,
1984). The first patent related to mineral flotation, which was
concerned with differences in wettability of minerals by water and oil,
was issued in 1860 (Anonymous, 1982). The golden age of the invention of
flotation was the period between 1910-1915. It was the result of small
contributions of many peeple. T. J. Hoover, in 1912, named 57 people who
had made significant contributions to the development of flotation up to
that date (Kitchener, 1984). In 1921, a Norwegian engineer by the name
Nils Padersen patented a flotation machine in the United States (Parker,

1958).

1.2.2 - Flotation In Biotechnology

The typical aqueous environment and the surface activity of many
bioproducts certainly made the extension of the use of flotation to
biotechnology a logical step 1. Besides that, there has been a significant
concern and knowledge about properties of biological media and products
such as foaming and denaturation due their undesired effects like loss In
activity and fermentation operational problems. However, natural
properties such as hydrophobicity and foaming made the application of

flotation not only possible but also viable, simple, and inexpensive. These

 

1 When looking through the biotechnology literature we must be careful in identifying
what authors called flotation. Some authors called flotation some other adsorptive bubble
separation techniques like bubble and foam separation. Description of these techniques
follows.

 

characteristics are not only desirable in flotation but also a must since
the addition of chemicals to promote hydrophobicity and foaming would be
restricted in the case of dealing with food-grade or pharmaceutical
products (Evans et al., 1970; Kalischewski et al., 1979; Kolsaridu ef al.,
1983; Vardar-Sukan, 1988).

The first patents on the use of flotation in biotechnology may have
been the processes for production of algin from seaweed by LeGloahec and
Herter in 1938 and the separation of germinating from non-germinating
seeds by Thomas Earle in 1939. The next decade saw a wide application of
flotation in the processing of biological materials. Rubber, oils, and
resins were recovered from plants, fat from wool and animal tissues, and
proteins from dairy effluents. Sugar solutions were better clarified,
starch and gluten were separated from each other, and the first flotation
of microorganisms (Koch bacillus, Escherichia coli, Staphylococcus albus,
and Schircsaccharomyces sp.) was done by Dognon and Dumontet in 1941
(Dognon and Dumontet, 1941; Gaudin, 1957; LeGloahec and Herter, 1938;
Thomas and Winkler, 1977; Turrall et al., 1943).

After the initial burst of broad application, flotation in
biotechnology was restricted to waste treatment, probably due to the fear
of loss of biological activity of fermentation products. Waste treatment
targeted the removal of proteins, lipids, antibiotics, adsorbing agents as
activated charcoal, and especially microorganisms. An advantage of
microorganisms over minerals with respect to flotation is their
homogeneity in shape and size. The disadvantage is their microscopic size
that hinders attachment to the gas bubbles. Many different bacteria were
concentrated, with or without the use of collectors (cell enrichment ratios

up to 10,000,000); spores were separated from vegetative cells and cell

debris; non-foaming mutants were separated from foaming parent yeast
cells; algae were recovered from cultures; and lead was removed from
solution by flotation of microorganisms to which it had adsorbed and
provided increase in hydrophobicity (Bishop, 1978; Cable and Schi‘irgel,
1984a; Golab and Smith, 1990; Grieves and Wang, 1966; libuchi ef al.,
1974; Rubin, 1968; Rubin et al., 1966; Thomas and Winkler, 1977).

1.2.3 - Potential of Flotation In Biotechnology

Limited interest of flotation lasted until recently when some
investigators realized that biotechnology research has concentrated on the
developing new bioreactcrs, the need for which was not evident, with
disregard to the development of more efficient and less costly
downstream process techniques. Among many techniques, flotation was
cited as one with no doubt of successful application into biotechnology
once process engineers recognize its potential (Gehle et al., 1991;
Hernainz ef al., 1987; Zlokarnik, 1990). Parallel to that, the collection of
data through the years made the fear of generalized loss of biological
activity unfounded. The eventual accumulation of research data on the
separation and concentration of an extended list of proteins showed that
there is a considerable variation in the susceptibility of different proteins
to denaturation. The loss in activity can vary from 0 to 50 percent,
depending on the specific system and conditions used (Thomas and Winkler,
1977; Uraizee and Narsimhan, 1990). Therefore, for the right system,
flotation is a viable option.

The potentially extensive application of flotation in biotechnology is
its use as a solid-liquid separation or recovery unit operation in

downstream processing, especially of fermentation products such as

10

proteins, amino-acids, antibiotics, and polysaccharides. Two approaches
to accomplish these separations can be considered. First, some products
are naturally insoluble or can be made insoluble in the media and then
floated. Microorganisms, proteins, and polysaccharides are good examples.
Second, most of the separation methods in biotechnology based on
chemical or physical interactions can be adapted for flotation by the
engineering of an appropriate carrying system. Details of these two
approaches follow.

It was said above that flotation has been proven to be an efficient
method to recover microorganisms in the treatment of waste water.
Therefore, it would be logical to attempt to recover or concentrate cells
from batch or continuous cultures. Miyazu and Yano determined that
Mycoforula japcnica can be effectively recovered from a aqueous
suspension by flotation when the cells are grown with n-parafin instead of
glucose (Miyazu and Yano, 1974). Continuous recovery of Hansenula
polymorpha and Saccharomyces cerevisiae from a culture medium at the
pilot plant level produced results better than laboratory tests at the
Institut fiJr Technische Chemie, Germany, one of the leading institutions in
the development of flotation applications in biotechnology (Gehle er al.,
1991).

Proteinaceous materials and protein precipitates also can be
sufficiently hydrophobic to attach and be lifted by air bubbles. libuchi and
co-workers reported a much better recovery of hemoglobin and casein by
particle flotation than by bubble separation when the proteins were in
solution (libuchi ef al., 1974). In Northern Ireland, flotation was studied
as a Way to recover protein from fresh sea mussels. Recoveries up to 90

percent were reported (Holland and McComiskey, 1986; Holland and

11

Shahbaz, 1986). The advantage of flotation in this case is the possibility
of simultaneous recover of insoluble and soluble protein (40 and 60
percent of the total protein, respectively) with the use of flocculants.

Polysaccharides with food and nonfood applications like xanthan,
curdlan, Ievan, and cellulose derivatives are bioproducts that have at some
point of their purification protocol an insolubilization step by the use of
specific ions, pH, alcohol, increase in ionic strength, etc. (Cadmus ef al.,
1976; Godet, 1973; Han, 1990; Harada ef al., 1968). Therefore, there are
opportunities for the substitution of centrifugation, and especially
filtration, by flotation in these processes since these products have a
tendency to form gels that drastically clog filter media. Many of these
compounds are non-ionic, a desirable characteristic when the final
objective is to have a hydrophobic particle. One example of flotation of
such a polysaccharide is the flotation of algin cited above (LeGloahec and
Heden 1938)

The second approach to recover bioproducts by flotation is the use of
a carrying system. Soluble products can be floated by engineering of a
carrying system which is insoluble (or can be insolubilized) and floatable.
The motivation to design such a system is based on the fact that the
majority of separations in biotechnology are based on the interaction
between the molecule or molecules being separated with another molecule
or group added to the system, as a solid, liquid or a species in solution, the
agent of separation. Such interactions can be non-specific as in the case
of adsorption on activated charcoal or specific such as electrostatic,
hydrophobic, or affinity interactions with substrate analogs, inhibitors,
triazine dyes, antigens, antibodies, etc. If the agent of separation is

located on the surface of a particle, we have the case of adsorption, which

12

is usually carried out in packed beds. If it is in solution and two
immiscible phases are formed, we have the case of aqueous two-phase
extraction. The formation of an insoluble complex characterizes common
precipitation or affinity precipitation. If the agent of separation is a high
molecular weight compound and physical separation is accomplished by
ultrafiltration, we have the case of affinity ultrafiltration. In general we
must have a basic interaction between the molecule being separate and the
agent of separation. The other details would dictate kinetics and the
mechanics of the operation according the physical state of the phase in
which the agent of separation is located. Therefore, if we engineered a
system in which the agent of separation is associated with a carrying
moiety that is or can be made insoluble and hydrophobic, the final complex
“product-agent of separation-carrying moiety“ can also be floated. With
the exception of few techniques like reverse osmosis, electrodyalisis, and
electrophoresis, almost all other separation methods in biotechnology have
the potential to be practiced as flotation.

The use of these separation principles in the form of flotation would
have many advantages. First, the immobilization of the product in a
particle may contribute to a protection against denaturation. Second,
when the agent of separation is in a liquid phase, flotation would combine
advantages of adsorption on a solid (large selection of adsorbent groups to
choose from) with advantages found in aqueous two-phase extraction (fast
kinetics of liquid systems).

This new focus on flotation is occurring in parallel with the
acceptance of a more efficient concept of a flotation device, the flotation

cdumn.

13

1.3 - Column Flotatlon: the New Concept of Flotation

Column flotation is the newest successful development in flotation
equipment (Figure 1.3). Its mechanism of operation is very simple. Feed
is introduced at an intermediate height in the column (approximately at
the top one third) and bubbles are generated at the bottom. Hydrophobic
particles from the feed are carried up to the froth at the end of the .
column. Wash water added at the top washes down hydrophylic particles
and liquid from the feed. The overflown froth (called concentrate)
contains the hydrophobic particles and liquid composed of wash water and
a small portion of feed liquid, if any. Hydrophilic particles are eluted with
the bottom stream (called failings), which is the largest flow in the
cmumn.

Column flotation is the present trend in the mineral industry where
it is replacing the traditional flotation cells for many reasons. The
columns are highly efficient providing higher grade and recovery than
conventional mechanical cells. They have high capacity despite their small
size and low fixed and operating costs. The equipment can be constructed
on the site and have no moving parts saving energy and maintenance. They
can be fully automated and laboratory results are easily scaled-up (Foot ef
al., 1986; Hernainz ef al., 1987; Hyde and Stojsic, 1987; Sastry, 1988;
Schubert, 1988).

1.4 - Objective
The objective of this proposed research is to study the application
of the modern column flotation technique to biotechnology processes for

the recovery of biomolecules produced by fermentation.

14

 

 

 

 

f»
CLEANING INTERFACE
m
(Purity)
I SUSPENSION
SUSPENSIONFEED fl...
is:
DIAMETER _, _
(Throughput)
COIJ.ECTION
m _
(Recovery)
GENERATOR i:
' TAILINGS

Figure 1.3 - General scheme of a flotation column (from McKay et al.,
1988i

15

The recognition of the potential of flotation has not yet been
translated into a large number of processes developed and published
research. In an extensive literature search, only a few papers dealing with
application of flotation as a recovery method of species from culture
media were found (Gehle and Schfigerl, 1984a; Gehle and Schiigerl, 1984b;
Gehle ef al., 1991; Hernainz et al., 1987). However, these investigations
dealt with the recovery of microorganisms or molecules from
fermentation waste. No work on the recovery of a biomolecule as the
major fermentation product was found. The development of new flotation
applications is necessary to validate flotation as a viable and versatile
downstream operation, especially at a commercial scale. Also, some
aspects of how flotation have been practiced and developed in the field of
biotechnology need to be addressed. Some of these issues are cited below.

Virtually no published work uses the modern concept of column
flotation. The major omission in the investigations is the use of
equipment with the washing water injection, the simple, but most
innovative characteristic of column flotation. It is the washing water
that gives this type of column the capability to promote high purification,
one of the main objectives in downstream processing.

The omission of this important characteristic, washing water, did
not lead to simpler equipment. Instead, the different equipment reported
in the literature either have the basic configuration of the old flotation
cells or are rather complicated cylindrical devices, as it can be seen by
the examples in Figure 1.4.

Another point of concern is that the majority of the flotation
processes cited above made use of the addition of collectors, frothers. and

modifiers. Those additives commonly used in the mineral industry like

16.

.94.

17:1]
' i=3

 

 

 

 

from Gehle and from Gehle
Schfirgel, 1984b at al., 1991

 

 

from Holland and from Miyazu and
McComiskey, 1986 Yano, 1974

Figure 1.4 - Some different types of flotation columns found in
the biotechnology literature

17

kerosene, fuel oil, cyanide, and xanthates would not be approved for food or
pharmaceutical applications. Flotation systems that naturally float or use
food or pharmaceutical grade additives need to be developed.

This research addressed the two potential aproaches for flotation in
biotechnology discussed in section 1.2.3. First, in the work described in
Chapter 3 the natural floatability of a separation system to recover a
product from fermentation broth was exploited: partition of a iii-amylase
to a salted-out (hydroxypropyl)methylcellulose phase from fermentation
broth. Chapter 4 is the demonstration of the viability of development of a
carrying system to perform flotation based on a specific product-Iigant
interaction. The example system consisted in modified methylcellulose as
the carrying moiety, alkyl chains as ligands, and bovine serum albumin as
the product. Chapter 2 is a background in flotation, especially column
flotation, and related techniques classified as adsorptive bubble

separation.

1.5 - List of References

Anonymous. Selective flotation: an art as well as a science. Mining
Engineering. 1982, March, 275-279.

Belter, P. A.; Cussler, E. L.; l-Iu, W.S. Bioseparaficns. Willey Interscience:
New York, 1988, p v, 1-4.

Bishop, P. Removal of powdered activated carbon from water by foam
separation. Separation Science Technology. 1978, 13(1), 47-57.

Cadmus, M. C.; Rogovin, S. P.; Burton, K. A.; Pittsley, J. E.; Knutson, C. A.;
Jeanes, A. Colonial variation in Xanthomcnas campesfris NRRL 8-1459 and
characterization of the polysaccharide from a variant strain. Canadian
Journal of Microbiology. 1976, 22, 942-952.

18

Dechow, F. J. Separation and Purification Techniques in Biotechnology;
Noyes: Park Ridge, 1989, pp 1-4.

Dognon, A.; Dumontet, A. Concentration of separation des microorganismes
par moussage. Comptes Fiendus des Séances et Mémcires de la Société de
Biologie et de ses Ellie/95.1941, 135, 884-887.

Evans, M. T. A.; Mitchell, J.; Mussellwhite, P. R.; Irons, L. The effect of the
modification of protein structure on the properties of proteins spread and
adsorbed at the air-water interface. In Surface Chemistry of Biological
Systems; Blank, M., Ed.; Plenum, 1970, pp 1-22.

Foot, D. G.; McKay, J. l; Huiatt, J. L. Column flotation of chromite and
fluorite ores. Canadian Metallurgical Ouarterly.1986, 25(1), 15-21.

Gaudin, A. M. Flotation; McGraw-Hill: New York, 1957, pp 547- 555.

Gehle, R. D.; Schiigerl, K. Penicillin recovery from aqueous solutions by
continuous foam flotation. Applied Microbiology and Biotechnology.
1984a, 19, 373-375.

Gehle, R. D.; Schugerl, K. Protein recovery by continuous flotation. Applied
Microbiology and Biotechnology. 1984b, 20, 133-138.

Gehle, R.; Sie, T-L; Kramer, T; Schiigerl, K. Continuous cultivation of
Hansenula polymorpha and Saccharomyces cerevisiae in an integrated
pilot plant reactor-flotation column-system. Journal of Biotechnology.
1991, 17, 147-154.

Godet, P. Fermentation of polysaccharides gums. Process Biochemistry.
1973, 8(1), 33-34.

Golab, Z. S.; Smith, R. W. Flotation of Streptomyces pilosus after lead
accumulation. Biotechnology and Bioengineering. 1990, 37, 955-959.

Grieves, R.B.; Wang S. -L. Foam separation of Escherichia coli with a
cationic surfactant. Biotechnology and Bioengineering. 1966, 8, 323-336.

Han, Y. Microbial levan. Applied Microbiology. 1990, 35, 171-194.

19

Harada, T.; Misaki, A.; Saito, H. Curdlan, a bacterial gel-forming B-1, 3-
glucan. Archives of Biochemistry and Biophysics. 1968, 124, 292-298.

Hernainz, F. B. C.; Garzon, M. G.; Pardo, L. E. F. La flotacibn como alternative
en la separacibn de materiales biolégicos. Ingenieria Oufmica (Madrid).
1987, 19, 539-543.

Holland, C. R.; McComiskey, P. Mussel protein recovery - batch dissolved air
flotation studies. Journal of Food Technology. 1986, 21, 763-774.

Holland, C. R.; Shabaz,M. Mussel protein recovery using dissolved air
flotation. Journal of Food Engineering. 1986 , 5, 135-151.

Hyde, G. R.; Stojsic, A. Electrochemical and column flotation technologies
can mean higher recovery at lower operating costs. Mining Engineering.
1987, July, 481-484.

libuchi, S.; Yano, T.; Torikata, Y. Adsorption of hemoglobin and casein onto
air bubbles. Agricultural and Biological Chemistry. 1974, 38(2), 395-400.

Kalischewski, K.; Bumbullis, W.; Schi'igerl, K. Foam behavior of biological
media, I. protein foams. European Journal of Applied of Microbiology and
Biotechnology. 1979, 7, 21-31.

Kotsaridu, M.; Gehle, R; Schiigerl, K. Foam behavior of biological media, IX.
pH and salt effects. European Journal of Applied of Microbiology and
Bictechnology.1983, 18, 60-63.

Kelly, R. M. General processing considerations. In Handbook of Separation
Process Technology, Rousseau, R. W., Ed.; Wiley-Interscience: New York,
1987, pp 221-222.

Kitchener, J. A. The forth flotation process: past, present, and future, in
brief. In The Scientific Basis of Flotation; Ives, K. J., Ed.; Martinus Nijhoff:
The Hauge, 1984; pp 3-51.

LeGloahec, V. C. E.; Herter, J. R. Method of treating seaweed. US patent
number 2,123,551, August 1938.

Leja, J. Surface Chemistry in Flotation. Plemun, New York, 1982.

Z)

Matis, K. A.; Mavros, P. Recovery of metals by ion flotation from diluted
aqueous solutions. Separation and Purification Science. 1991, 20(1), 1-48.

McKay, J. 0.; Foot, D. G.; Shirts, M. B. Column flotation and bubble generation
studies at the Bureau of Mines. Column Flotation '88; Sastry,K. V. 8., Ed.;
Society of Mining Engineers: Littleton. 1988, pp 173-186.

Miyazu,Y.; Yano, T. Surface character and adsorbability to air bubbles of a
hydrocarbon-grown yeast. Agricultural Biological Chemistry. 1974, 38(1),
183-188.

Parker, D. J. Flotation and washing of slimy gelatinous precipitates of
oxygenated metals. M. S. Thesis, Michigan State University, East Lansing,
MI, 1958.

Parker, D. E. ; Glatz, C. E. ; Ford, C. F. F. ; Gendel, S. M. ; Suominen, I. ;
Rougvie, M. A. Recovery of a charged-fusion protein from cell extracts by
polyelectrolyte precipitation. Biotechnology and Bioengineering. 1990, 36,
467-475.

Prave, P.; Faust, U.; Sittig, W.; Sukatsch, D. A. Ed. Basic Biotechnology. VCH,
New York, 1987.

Rubin, A. J. Microflotation: coagulation and foam separation of Aerobacter
aercgenes. Biotechnology and Bioengineering. 1968, 10, 89-98.

Rubin, A. J.; Cassel, E. A.; Henderson, 0.; Johnson, J. 0.; Lamb, J. C.
Microflotationmew low gas-flow rate foam separation technique for
bacteria and algae. Biotechnology and Bioengineering. 1966, 8, 135-151.

Sastry, K. V. S. Introduction. In Column Flotation '88; Sastry,K. V. 8., Ed;
Society of Mining Engineers: Littleton. 1988, p iii.

Schubert, H. Counter-flow flotation cells (flotation columns), present
state and current trends. Aufbreitungs-Technick. 1988, 29(6), 307-315.

Scmasundaran, P.; Ananthapadmannabhan, K. P. Bubble and foam flotation,
ore flotation. In Handbook of Separation Process Technology, Rousseau,R.
W., Ed.; Wiley-Interscience: New York, 1987, pp 775-805.

21

Summerfelt, S. T. The concentration and the purification of an a-amylase
using a methylcellulose-salt, two phase partitioning process. M. S. Thesis,
Michigan State University, 1988.

Thomas, A.; Winkler, M. A. Foam separation of biological materials. In
Topics in Enzyme and Fermentation Biotechnology, Wiseman, A., Ed; Ellis
Horwood: Sussex, 1977, pp 43-71.

Turrall, W. T.; Klasen, J.; Smedley, H. Froth flotation of rubber and resin
components of milkweed and other plants that contain resins, rubber,
batata, or related gummy substances. Canadian Journal of Research. 1943,
218(10), 195-201.

United States Bureau of Mines. Mining and quarrying in the metals and
industrial minerals industries. Bureau of Mines Minerals Yearbook, 1985.

Uraizee, F.; Narsimhan, G. Foam fractionation of proteins and enzymes: I.
applications. Enzyme and Microbial Technology. 1990, 12, 232-233.

Vardar-Sukan, F. Efficiency of natural oils as antifoaming agents in
bioprocess. Journal of Chemical Technology and Biotechnology. 1988, 43,
39-47.

Wheelwright, S. M. Designing downstream processes for large-scale
protein purification. Bictechnclogy.1987, 5, 789-793.

Zlokarnik, M. Trends and needs in bioprocess engineering. Chemical
Engineering Progress. 1990, April, 62-67.

Chapter 2

Flotatlon, an Adsorptlve Bubble Separatlon

In order to better understand flotation in biotechnology, a literature
search was conducted on this subject with special interest in its previous
applications in downstream processing. Only a small number of
publications about this use of flotation was found with the exception of
reports concerning waste treatment. However, a large number of
publications about techniques similar to flotation applied to or developed
for downstream processing were found. These techniques, together with
flotation, comprise what is called adsorptive bubble separations.
Therefore, this literature review was not restricted only to column
flotation since flotation and the related adsorption bubble separation
techniques were developed simultaneously and findings in one area were
used in the development of others.

Due to the difficulty of quantifying individual proteins in a mixture
and the simplicity of some enzymatic assays in the early days of
adsorptive bubble separations, enzymes were the compounds most used in

the development of these techniques (Thomas and Winkler, 1977).

2.1 - Adsorptlve Bubble Separation Techniques
2.1.1 - Concept and Classification
I Adsorptive bubble separation is the generic name given to

separation/concentration techniques based on surface activity (Lemlich,

22

23

1972). Species (compounds, particles, cells, precipitates, etc.) can be
adsorbed or attached to rising bubbles in a liquid phase according to their
surface activity. Transport of the attached species by the bubbles and
foam causes their concentration at the top of the liquid phase or foam. If
the adsorption/attachment is selective, enrichment takes place.

The different types of adsorptive bubble separation techniques are
shown in Figure 2.1. The first distinction to be made is on the phase which
will receive the species carried up by the bubbles. If there is no foam
involved in this reception, the technique is classified non-foaming
separation; otherwise, it is called foam separation.

Non-foaming separation is classified in two types: bubble
fractionation and solvent sublation. In bubble fractionation, the separated
species is left just at the liquid surface after bubble collapse. In solvent
sublation. the separated species are transfered either to an immiscible
liquid phase atop the lower liquid phase or to the interface between these
hNo phases.

Foam separation is divided in foam fractionation and flotation. In
the former, the species of interest, part of an homogeneous solution, is
separated and/or concentrated by adsorption on bubble surface which are
foamed out of the system. In the latter, the species is carried up by the
gas bubbles as a particle or associated with one.

Lemlich (1972) subdivided flotation in many categories. However,
for simplicity, these groups will be regrouped in only two groups: flotation
for particle separation, wherein particles are the species to be separated
(e.g., ore flotation), and separation by particle flotation, wherein specific

particles, formed in the system or not, are the vehicle by which the

Adsorptive Bubble Separation

I

Foam Separation Non-Foaming
Separation

 

Bubble Solvent
Fractionation Sublation

 

 

 

Foam Fractionation

 

Flotation

 

Flotation for Separation by Particle
Particle Separation Flotation

Figure 2.1 - Classification of the adsorptive bubble separation
techniques (modified from Lemlich, 1972).

%

species of interest are separated. In this second group, a precipitate can
be formed by addition of either any suitable chemical (precipitation
flotation) or a collector (ion and molecular flotation). Another case of
separation by particle flotation is the addition or formation of colloid
particles which adsorb the species of interest (adsorbing colloid

flotation).

2.1.2 - Non-Foaming Technlques: Bubble Fractionation

Dorman and Lemlich (1965) developed a foamless separation
technique called bubble fractionation. The objective was to devise an
adsorptive bubble separation technique applicable to liquid systems that
do not foam. The separator was a bubble column: a vertical tube with a gas
sparger at the bottom. In the case of continuous operation, inlet for the
feed at intermediate height and outlets for rich and lean streams at the
top and bottom, respectively, were installed. Monobutyl diphenyl sodium
monosulphonate was concentrated more than four times by batch runs
after several hours of bubbling.

Miyazu and Yano (1974) used a Dorman-Barker's apparatus as a
bubble fractionation separator to study cell removal from aqueous
suspension. The separator was a bubble column divided in two stages by
inverted funnels to reduce back-mixing (Figure 1.4). They compared
adsorption per unit area of bubbles and affinity of hydrocarbon-grown and
glucose-grown cells to bubbles. High recovery was achieved for
hydrocarbon-grown cells (up to 95 percent). Based on the results authors
suggested that flotation could be an effective way to recovery hydrophobic

cells.

%

Chiang and co-workers (1981) studied the separation of isoelectric
precipitated hemoglobin and oil-hemoglobin with an apparatus similar to
the one used by Miyazu and Yano (1974) cited above. The fundamental
difference was the division of the column in stages by perforated disks
instead of inverted funnels. The authors claimed that an unexpected
reverse enrichment took place under certain conditions where the
precipitate concentrated on the bottom of the separator. It seemed that a
“not-so thick layer of foam” beneath the plates retained the particles from
the upward stream of back-mixing.

Potter and co-workers (1990) developed a mathematical steady-
state model for in situ separation by bubble fractionation. They studied
the separation of proteins in the baker's yeast fermentation process using
carbon monoxide either generated by fermentation or supplied externally

as the gas phase.

2.1.3 - Foam Separatlon
2.1.3.1 - Foam Fractionation

According Charm and co-workers (1966) and Thomas and Winkler
(1977), foam fractionation was developed by W. Ostwald in 1937. Ostwald
first used this technique when he separated albumin from potato and beef
juices. This work on the separation of protein from plant juices was
continued later by other investigators (Davis et al., 1949a; Davis et al.,
1949b; Weijenberg etal., 1978).

However, in the same year, Schlitz (1937) claimed foam
fractionation development with the separation of cholic acid from its
sodium salt and hemoglobin along with antibiotics from rabbit serum. A

more complete description of the technique, that he called fractionation by

27

adsorption in foam, came later (Bader and Schlitz, 1946; Schlitz, 1946).
These subsequent authors reported the purification of choline-estearase,
sodium cholate, bile salts-saponin, lecithin, and horse serum albumin.
Concentrations up to 16-fold were observed but foaming times were high,
up to 3 days (Bader et al., 1944). London and co-workers (1954) also had
the goal ”to start the development of a simple technique to aid in in the
purification of enzymes“ when they reported the purification of urease and
catalase by foam fractionation.

Five different approaches to foam fractionation operation have been
developed (Figure 2.2). One of the most significant changes in the mode of
operation was the introduction of reflux to increase the overall separation
analogous to distillation and extraction (Lemlich and Lavi, 1961).
Concentration of the foamate increased 10-fold; however, foaming times
were still long, around 48 hours. These large processing times may be one
of the reasons why the potential applications of foam fractionation in the
large-scale concentration/purification of enzymes suggested by Charm and

co-workers (1966) did not materialize into actual commercial processes.

2.2 - Flotatlon and Column Flotation
2.2.1 - Basics of Flotation

Flotation is a technique mostly used in the selective separation and
concentration of particles, mainly minerals. It had its inception and
development in the area of mineral processing. Through the years and
presently, mineral concentration and enrichment is by far the most
important application of flotation. The vast literature in flotation is, as a

consequence, concerned about ore flotation. The best treatments of

 

 

 

 

FOAMATE FOAMATE
FOAM FOAM
0°90 FEED 000 ° BOTTOMS
e°°°e e 0 00° 0 .__.
°e 0° 0° o°o

 

 

 

 

 

 

 

 

l:
I:

Batch Continuous

 

 

 

on.,,,aorroms
O

 

Continuous stripper

sown: FOAMATE

           

   
 

  

  

    
 

000° 0
o o o°o
o 0

Continuous enricher Continuous combined

Figure 2.2 - Operational modes of foam fractionation (from Thomas and
Winkler, 1977),

E

flotation and its basic principles are the works of Gaudin (1957),
Fuerstenau (1976) and Leja (1982). Svarovsky (1981) has a less common
treatment of flotation since it approaches flotation only as a unit
operation for solid-liquid separation.

To better understand the importance of the different parameters
related to flotation, an understanding of the most basic phenomenon in
flotation, particle collection by bubbles, must come first. Figure 2.3 is a
representation of this phenomenon. It is assumed that a bubble is a a rigid
sphere travelling in the liquid were hydrophobic and hydrophilic particles
are in suspension moving towards it at a low Reynolds number. As the
particles approach the bubble following the stream lines, they begin to
slide around it. The hydrophilic particle, despite the close positioning,
slides all the way around the bubble and both, particle and bubble, continue
their trajectories. They do not interact with each other. It is said that a
collision happened and failed in producing attachment. However, the
hydrophobic particle and the bubble, are pulled close towards each other by
hydrophobic forces. The hydrophobic particle travels around the bubble
moving to the inner streamlines. Therefore, the liquid film between them
thins to a point where rupture occurs. This rupture means the formation of
a solid-gas interface. What occurs next is a rapid expansion of the gas
meniscus over the particle. If this expansion is fast enough the interaction
between the hydrophobic solid and gas bubble is strong enough to support
mechanical disturbances in form of future collisions. This way, a stable
attachment takes. place. We can say that the particle was collected by the
bubble. The time the particle travels on the surface of the bubble is known

as the sliding or contact time. The time necessary for the liquid film to

Time=0

Hydrophobic patch

fl
0
M Hydropniiippaicn

Time =11

K
M
v

Time = t 3

K
M

Figure 2.3 - Illustration of particle capture mechanism where
0 < t 1 < t 2 < t 3.

31

thin and rupture is known as the induction time (Dobby and Finch, 1986a;
Dobby and Finch, 1987; Svarovsky, 1981).

The key factor in the thinning and rupture of the liquid film between
particle and bubble is the hydrophobicity of the particle surface (even
though bubble surface charge may play a important part, according to
Okada et al., 1990). It is the degree of hydrophobicity of the particles that
at the and determine which particles, after colliding with the bubble, will
attach to it, and consequently, float.

A measurement of hydrophobicity is given by the contact angle a
measured according the scheme in Figure 2.4. Despite its popularity and
widespread use, contact angle measurement is not always a viable
technique. Problems include changes in the contact angle measured with
gas adsorption on the surface of the particle, the need of a polished
smooth surface, and the inapplicability for cases were the angle to be
measured is smaller than 15°. Recently, a bubble pick-up method was
improved as a alternative way to evaluate particle hydrophobicity. The
method measures the extent of particle-to-bubble adhesion by the angle by
which the gas bubble surface is covered with adhering particles under
static conditions (Vinke et al., 1991). The advantages of this method are
the absence of particle pretreatment, its suitability for the determination
of small angles, its applicability to porous materials, and that it is more
closely related to the conditions occuring during flotation than other
methods.

Despite its drawbacks, the thermodynamic interpretation of contact
angle has a great significance. The thermodynamic condition for the

existence of three phase contact is represented by Young's equation

32

   

LIQUID

SOLID T 7;; 78L

Contact equilibrium

 

    
 

J..-

W

/Woter

 

Contact angle for a solid with high hydrophobicity

Figure 2.4 - Contact angle and hydrophoblclty (from Lemlich, 1972;

Svarovsky, 1981 ).

3'3

'Ysg = 'Ysl + ‘Ylg COS O (2.1)

where 7 represents interfacial tension of solid-gas, solid-liquid, and
liquid-gas phases according the subscripts indication (I, g, and 3 stand for
liquid, gas, and solid, respectively). The free energy change due to

attachment AG is given by

AG = 759 - ('Ysl+7lg) (2-2)

This is an account for the energy lost due to the disappearance of solid-
quuid and liquid-gas interfaces and the energy gained with the creation of
the new one, the solid-gas interface.

Applying Young's equation results in

AG = yg|(coso-1) (2.3)

This equation, known as the Young-Dupré equation, shows that for a
finite contact angle there will be a decrease in the free energy of the
system. Contact angles over 30 degrees are usually necessary for a good
flotation (Lemlich, 1972; Svarovsky, 1981).

A more practical indicator of particle hydrophobicity is the
electrokinetic potential also known as zeta potential. A charged surface
particle generates a electrical potential gradient through the liquid phase.
The potential at the slip plane of a particle when it is moving it is known
as zeta potential. The exact position of this slip plane is not known but
the thickness of the liquid layer adhering to the solid is believed not to be

more than few angstroms (Leja, 1982). Therefore, the zeta potential is a

34

close indication of the charge at the particle surface as well as of a
bubble (Yoon and Yordan, 1986). Figure 2.5 depicts the striking relation
between the zeta potential and the floatability for some materials.

The paramount importance of bubble capture, for which the
mechanism was discussed above, relies on the fact that it determines how
operating parameters affect flotation performance. Specifically, it
determines the rate by which particle concentration changes with time in

the flotation equipment (Yoon et al., 1988). This rate of collection is given

by
dC/dt=-K[C(t)-C°.]n (2.4)

where t is time, C is particle concentration, C a. is the terminal
concentration of particles (concentration of apparently floatable material
that would remain in the flotation cell after long times), n is the order of
the rate equation, and K is the rate constant (Lemlich, 1972).

The constant K, a lumped parameter involving physical, chemical, and
surface properties of the system, is where the collection effect is
accounted by a parameter called collection efficiency, Ek, defined as the
fraction of all particles swept out by the projected area of the bubble that
collide with, attach to and remain attached to the bubble until reaching the
froth. This efficiency can be understood as the product of two other
efficiencies related to the two subprocess in the collection phenomenon:
collision and attachment. Efficiency of collision, Ec, is defined as the rate
at which particles collide with the bubble divided by the rate at which
particles flow across the projected area of the bubble. Efficiency of
attachment, E3, is defined as the fraction of all colliding particles that

reside on the bubble for a time greater than the induction time, i.e. have a

 

 

 

 

 

 

+100 '
\.
e75 '-
3‘ “5° '
5
s25 '
M
‘0 i
-25 " ‘ 5 E f—*
i4““'i:r‘""""5'l'”"5"": :H’&‘ 0

 

 

-50 1
2 4 6 8 10 12

pH

Figure 2.5 - Effect of zeta potential of the bubbles and latex particles (0
on flotation efficiency (n) in a cationic surfactant solution, with change of

pH (from Okada et al., 1990).

{3

sliding time or contact time larger than the induction time (Dobby and

Finch, 1986a; Dobby and Finch, 1987). Therefore,

Ek=Ec-Ea (2.5)

Dobby and Finch (1987) developed a fundamental model for the
capture process which includes dependence on the particle size and
density, bubble size, and liquid viscosity. Particle size affects the
collision and attachment efficiencies differently. As the diameter of the
particle increases the efficiency of collision increases due to the higher
inertia of the particle. Simultaneously, efficiency of attachment
decreases due to higher sliding velocity around the bubble. The net result
is the existence of am optimum particle size with respect to collection
efficiency which is in accord with recovery-size data reported for many
mineral systems.

Particle density markedly affects the collision efficiency through
effect on inertia in the same manner that particle size does. However, the
net result is the prevalence of the increase in the collision efficiency in
the range of 1.3 to 7.0 g/mL.

Both collision and attachment efficiencies vary inversely with
bubble size. The increase of collision efficiency with bubble size decrease
can be explained In part by the higher rigidity of small bubbles (Jameson et
al., 1977). Since the bubble interface is deformable, a particle with high
inertia may cause a deformation in the bubble surface and even bounce off
it which means no collision. In the case of the more rigid small bubbles, a
particle would have a lesser chance to slide around it and not experience

this bouncy effect (Schulze et al., 1989). The increase in attachment

37

efficiency is due to a smaller sliding velocity of smaller bubbles (bubble
rising velocity decrease with decrease in bubble size) which exceeds a
decrease in the sliding distance.

Attachment efficiency was insensitive to particle hydrophobicity
(represented by induction time) for very small particles (less than 5 pm)
due to the long sliding time of small particles. This time would be enough
for a highly hydrophobic particle as well as for a not so hydrophobic
particle to adhere to the bubble by the thinning of the liquid layer that
separates bubble from liquid. The importance of this phenomenon is the
decrease in selectivity. An example is given where separation between
particles with 40 and 15 ms of induction time changes from excellent to
virtually no separation for particle size change from more than 20pm to
less than 5 pm.

Dobby and Finch (1987) also predicted the negative effect of the
increase in viscosity on the collection efficiency due to decreasing bubble
rise and particle settling velocities and increase in induction time.

Besides the physical aspects of bubbles and particles, their surface
chemistry, especially of the particles, is of large significance in flotation.
Few materials are naturally hydrophobic enough to be floated, particularly
minerals. The solution is the use of collectors that incorporate on the
solid surface imparting hydrophobicity to the particles. Collectors are
amphipatic surfactants, R-Z type of molecules, where R and Z represent a
non-polar and a polar group, respectively. The association of the
collectors on the particle surface is the result of many different
interactions operating individually or in combination with each other.
Electrostatic interaction between charged areas on the solid and ionic

species are common. Some collectors are covalently bound to the surface

38

by chemisorption while others interact by hydrogen bonds. Species may
undergo chain to chain interaction upon adsorption that reinforces the
individual adsorption. A collector chain to particle surface interaction
may also occur. The driving force for adsorption, the free energy of
adsorption, is therefore, a summation of all the free energies related to
these interactions plus the free energy due to salvation or desolvation
attributed to these species during adsorption (Scmasundaran and
Ananthapadmannabhan, 1987).

These particle-collector interactions depend on pH, ionic strength,
collector concentration, type of solid surface, temperature, presence Of
other chemical species, etc. Therefore, each system has a different
mechanism playing the central role in the adsorption of the collector. The
many possible combinations of chemical process variables give flotation a
large degree of flexibility. As an example, Lemlich (1972) discusses the
concentration of iron ores where hematite is to be separated from quartz.
Six different processes have been developed, the first four in the list

below have been put in commercial operation:

1— Flotation of hematite using a sulfonate as the collector at

pH2tO 4.

2.- Flotation Of hematite with a fatty acid as the collector at pH

6to 8.

3- Flotation of quartz with an amine as collector at pH 6 to 7.

4— Flotation of quartz activated with calcium ions at pH 11 to 12,
using a soap as a collector together with starch to depress the

hematite.

5- Flotation af hematite with an amine as collector at pH 1.5 in the

presence of hydrochloric or sulfuric acid.

6- Flotation of hematite with hydroxamate as a collector at pH 8.5

and methylisobutylcarbinol as a frother.

The first process is based on the electrostatic interaction between
the negatively charged sulfonate with the positively charged hematite.
The quartz under these conditions is negatively charged. In the second
process, the fatty acids chemisorb only on the hematite. Flotation of
quartz in the third process is due to its electrostatic interaction with the
cationic amine (the hematite is essentially uncharged in the process
conditions). In the fourth process, starch chemisorbs on the surface of the
hematite preventing its flotation.

Collector adsorption on the surface of particles driven by
electrostatic interactions are dictated by the particle surface charge.
This charge is originated by the breaking of bonds during particle size
reduction, ions dissolution, etc. A practical measurement of this charge is
given by the electrokinetic potential also known as zeta potential. Its
measurement is based on the electrical potential gradient generated by the
charged particle surface. This potential causes a specific arrangement of
ions throughout the interface as a way of neutralizing the electrical
charge. First, there is a layer of ions, termed counter ions, that adsorb on
the interface generating a double layer of charged ions. Due to thermal
agitation the next layer is not fixed in a plane by it diffuses throughout
the liquid over a finite distance from the particle surface. The potential
at the slip plane of a particle when it is moving relatively to the water is

known as zeta potential. The exact position of this slip plane is not

40

known, but the thickness of the liquid layer adhering to the solid is
believed not to be more than few angstroms (Leja, 1982). Therefore, the
zeta potential zeta is a close indication of the charge at the particle
surface. Charge of bubbles can also be evaluated by the same principle
(Yoon and Yordan, 1986). Figure 2.6 is a schematic representation of this

electric double layer.

2.2.2 - Column Flotatlon
2.2.2.1 - Operational Principles and Advantages

Bubble columns, commonly used in chemical engineering practice as
chemical reactors, were first used as flotation devices for the treatment
of fine particle systems. The first flotation column were patented in'
Canada in the early 60's. The mineral industry seemed to be reluctant in
using the new devices, choosing to deslime the ores prior to flotation
despite the loses of valuable product in the finer materials. The main
reason for this was probably the marked difference in design and operating
philosophy between flotation column and conventional mechanical
flotation equipment. The breakthrough in North America happened only in
the beginning of the 80‘s. The necessity of treating ores of increasing
difficulty of separation and the need for minimization of losses led
industries to adopt the flotation column concept. However, since the early
60's the Russians also have been working in this new concept and the
Chinese already have columns in operation (Dobby and Finch, 1986a; Hu and
Liu, 1988; Mavros et al., 1990; Schubert, 1988).

A schematic of a flotation column showing the delineation of the
two distinct sections in the equipment, the collection or recovery zone and

the cleaning or washing zone, was presented it Figure 1.3. Also, the

41

 

 

 

 

SURFACE _
CHARGE STERN
PLANE
r...
1 e G) 9
° 6)
|
g l G)
o o
(I) I
‘ I
l e POTENTIAL-
w j c DETERMINING MS
I o“ ; O HYDRATED
2’ 1' . COUNTER IONS
; a‘ O NEGATIVE co-IONs
2 C-
DJ
j. I
O I
l l
O l
8 DISTANCE—-

Figure 2.6 - Schematic representation of the double layer according to
Stern's model (from Scmasundaran, 1975).

42

location of feed, washing water, and gas inlets, and failings and
concentrate outlets are indicated. The tailings volumetric flow rate is
slightly higher than the feed volumetric flow rate due to the introduction
of the washing water at the top. A small part of this water goes to the
concentrate (overflow) and the rest of it goes to the failings (underflow).
This imbalance between feed and failings volumetric flow rate is called
bias. The collection and cleaning sections work under countercurrent flow,
but with distinct characteristics and functions. Ideally, the column should

work in a plug flow regime (Wheeler, 1986).

In the recovery section, there is a countercurrent flow of suspension
and air bubbles generated in the bottom of the column. There is no need of
moving parts to agitate and keep particles in suspension so they can
contact the air bubbles. In conventional machines, impellers are used for
this purpose and, at the same time, to generate and/or introduce the air
bubbles in the system. However, the turbulence necessary for particle-
bubble collision is, at the same time, deleterious to the flotation
phenomenon because it causes detachment between particles and bubbles
attached in previous contact.

The descending particles in the recovery section meet a upstream of
moving bubbles during all their residence time. Therefore, particles have a
chance to collide and possibly attach to bubbles during all their way down
in the column. This means that they spend their whole time productively
in the collection zone. In conventional cells, however, after particles and
bubbles leave an active zone their velocities are lowered which reduces
the probability of collision and attachment between them. This may be
why flotation columns use up to eight times less air per volume than

mechanical cells. Moreover, the recovery zone is one hundred percent

43

available for a particle coming from the cleaning section due to undesired
accidental detachment. In contrast, the horizontal configuration of the
mechanical cells would offer a detached particle only a fraction of the
total retention time (the fraction corresponding to the last cells of the
bank). Due to the bottom location of the failings exit and to the top to
bottom liquid flow, the column never sands up.

The most unique characteristic Of the flotation column, the washing
water, is located in the cleaning section. In the conventional flotation
devices, there is a draining froth bed where non-flotable particles that
were pulled up in the wake of the rising bubbles are separated from the
bubble-attached particles. The bubbles are surrounded by large amounts of
suspension at their arrival at the liquid-froth bed interface. As the
bubbles move up in the bed, the suspension containing the non-flotable
particles drains back to the liquid phase. However, this makes the bubbles
closer to each other due to the thinning of the liquid wall between them up
to a point that they work as a filter trapping non-flotable particles among
them. In flotation columns, the washing water flowing down the cleaning
section sweeps away the tail of non-flotable particles following the
bubbles. Moreover, it thickens the liquid wall between the bubbles
preventing any mechanical entrainment, just the opposite of what happens
in the draining froth bed. This avoids even non-flotable slimes from
reaching the concentrate by washing down feed water. Consequently, the
water in the overflow is wash water.

In columns there is no single draining froth bed, but a froth
structure composed of two sections due to the injection Of washing water
at a internal point of the column (Figure 2.7). The section above the

injection point is just froth that characterizes it as a conventional

 

   

wash
water
concentrate
draining negative
froth i bias
(>030 positive
bias
packed
bubble bed
i > 0.14
expanded
bubble bed -,' interface
e <0." .5. 2 level
bubbling
zone
e (0.20

Figure 2.7 - Froth structure. 8 is the fractional gas holdup (from Yanatos,
et al., 1986).

45

draining froth bed. However, the portion under the injection point is a
more loose configuration of bubbles due to the washing water flowing
downwards, termed bubble bed. Due to the existence of a water content
gradient in this bubble bed, it can be further subdivided in two
subsections. At the bottom, right at the interface with the suspension, the
bubbling zone, is the expanded bubble bed; above it is a less loose bed
termed packed bubble bed (Yanatos ef al., 1986).

The positioning of the washing water is a valuable parameter of
operation since it affects the amount of liquid that overflows the column.
Therefore, it Partially controls how much liquid is recovered with the
floated particles. This is measured by the volumetric concentration ratio,
i.e. the ratio between the feed and the concentrate volumetric flow rates.
A high volumetric concentration ratio is a desirable characteristic of an
unit operation in downstream processing and it is almost a must if the
unit operation is applied as one of the first steps in a multiple step
purification/recovery scheme.

The most important practical consequence of these fundamental
differences between columns and conventional mechanical cells is the
higher separation/recovery performance. Wheeler (1986) reported
increase in grades for a cleaning operation of a complex copper ore in Peru
from 22.5 to 30.1 percent and, for a molybdenum circuit at Gaspé Copper
Mines, Canada, increase in recovery of 100 percent (from 40 to 80
percent). Other advantages of flotation columns over conventional devices
are ease of control, higher capacity (up to ten times faster), lower energy
consumption, less space for installation, less maintenance, and lower

operating costs.

46

2.2.2.2 - Operational Parameters
Schubert (1988) pointed out the following essential column design

and operational parameters:

- Column height L;

- Column diameter D;

Column height to column diameter ratio L/D;

Position of feed inlet;

Type of aerator;

Bubble size;

Superficial air velocity;

Superficial pulp velocity;

Volume percent of solids in the feed; and

Superficial velocity of wash water.

The five first parameters are determined by design. The rest can be
adjusted within given operational limits. Typical industrial columns are
13 m high by 0.3 to 2.0 m in diameter, either circular or square in cross-
section. The collection section is typically about 10 m. Typical operating
conditions are: superficial gas velocity, from 1 to 3 cm/s; superficial
suspension velocity, from 0.5 to 2 cm/s; bubble diameter, from 0.05 to 0.2
cm; and gas hold-up, from 10 to 25 percent (Dobby et al., 1988).

Yianatos and co-workers (1988) studied the effect of height and
diameter on column performance. They addressed the question as to why
industrial columns have about recovery section height (H) and diameter (D)
of about 10 and 1 m, respectively. Simulation studies showed the
existence of a compromise among degree of mixing and volumetric bias
rate versus volumetric gas rate. The longer the column, the smaller the

degree of mixing which leads to better performance (grade and recovery).

47

However, the gain in recovery levels off as the the ratio HID increases: due
to the fixed superficial gas velocity, the gas flow rate decreases which
causes the column to approach overloading. Their final conclusion was
that for typical operating conditions, a height to diameter ratio of 10 is a
reasonable compromise.

The feed inlet position indirectly affects the cleaning action Of the
wash water. Because the froth bed can not be longer than the cleaning
section, the feed position determines its limit. However, froth bed height
is not the only factor upon which cleaning is controlled. Yianatos and co-
workers (1987), used laboratory and large scale experiments to determine
hydraulic entrainment of fine particles in industrial columns (12 m in
height). Minimum entrainment was found for froth height of at least 1 m
as long as superficial gas velocity and bias rate were less than 1.5 cm/s
and in the range of 0.2 to 0.4 cm/s, respectively. A remarkable finding
was that the main cleaning action occurred close to the interface,
approximately 10 cm at superficial gas rates lower than 1 cm/s.

Superficial pulp velocity and pulp density (a combination of particle
density, size, and population density) determine the column capacity which
is limited by the rate of concentrate removal. The rate at which particles
can be removed depends on the individual bubble loading (mass of solids
per area of bubble surface) and bubble surface production rate (area of
bubble per time). Espinosa-Gomez and co-workers (1988a and 1988b)
studied carrying capacity in full scale columns. Based on data from six
different plants, an empirical relationship was found that relates the

carrying capacity to particle size and particle density:

03 = 0.0682 (dao) (rp) (2.6)

where Ca is the carrying capacity expressed as mass of solids to overflow
per time per unit of column crossectional area in grams/minutes/cmZ, deo
is particle size as the 80 percent passing size of the concentrate in um,
and rp is particle density in g/mL. The authors noted that there is
evidence that Ca is independent of column diameter and it has potential as
a scale-up parameter.

Superficial gas rate is, therefore, responsible for the recovery as
the determinant of bubble surface production rate. However, an upper
limit for this parameter exists above which flooding will occur (Dobby and
Finch, 1986b). Smaller bubbles and higher downward liquid velocity act to
lower this limit. For a specific bubble size the maximum collection rate
constant is Obtained while operating at a superficial gas velocity Vg at (or
close to) the maximum V9. Since the bubble size increases with increasing
gas rate, design Of gas spargers has a twofold converse objective:
producing small gas bubbles at high gas velocity.

Since bubble size is the most important physical parameter in
column flotation, the sparger is the heart of the equipment (Wheeler,
1986). As the bubbles get smaller, the surface area per unit volume
increases. Therefore, there is more surface available for bubble-particle
contact. Moreover, the larger the bubbles, the alrger the wake they cause
drawing more material (liquid and hydrophilic particles) up as the bubbles
rise. This contributes to particle entrainment and, consequently, lower
grades.

Spargers for industrial columns are made of flexible materials such
as perforated rubber or filter cloth. One singular characteristic of the
flexible spargers is a self-regulating property, with hole size increasing

in direct proportion to pressure drop across sparger. Laboratory and pilot

49

scale columns usually have rigid spargers made of ceramic or porous metal
for convenience. The choice of flexible spargers for industrial columns is
based on the decreased chance that these spargers will get plugged with
particles or salt deposits. Additionally, Rice and co-workers (1981) and
Weber and co-workers (1990) found that flexible spargers are more
efficient bubbling devices than rigid ones producing more uniform
emulsions, smaller bubbles, larger voids, and reduced dispersion
coefficients. Saxena and co-workers (1990) verified that this emulsion
uniformity is relative since they detected a segregation of bubbles through
the cross sectional area of the column, with small bubbles concentrating
towards the wall.

A study by Xu and Finch (1989) showed a modest effect of the
sparger material on the bubble diameter. Moreover, this paper brings three

other conclusions that deserve mention:

- Bubble size decreases as the column cross-section area to
sparger surface area ratio decreases. This is in accordance with

results showed by Reddy and co-workers (1987).

— Bubble diameter correlates with gas flow rate per unit area of

sparger by
db a (Rs.vg )025 (2.7)

where db is the bubble diameter, V9 is the superficial gas rate, Rs
is the ratio of column crossectional area to sparger surface
area, and, consequently, Rs . V9 is the gas flow rate per unit area

of sparger.

— As a scale-up criterion, Rs should be preserved to ensure

50

the same db and V9 in an up-scaled column.

Methods for bubble size distribution measurement include
photography, high speed cinephotography, fiber optics technique, and drift
flux analysis (Dobby et al., 1988; Saxena, et al., 1990).

2.2.2.3 - Modeling and Scale-Up

Ease of modeling and scaling-up based upon fundamental physico-
chemical principals are other advantages of flotation columns over the
conventional cells. Dobby and Finch (1986a) pointed out four reasons for
this. First, there is a well defined flow regime of particles and bubbles in
the columns. Second, there is no turbulence to be matched in laboratory
machines as is the case of mechanical cells. Third, there is a minimum
particle entrainment in the froth contrary to which happens in the draining
froth employed in the conventional cells. Fourth, the existence of a single
three-dimensional froth structure in a single device makes froth modeling
for the column less complicated than the conventional case where there is
a bank of cells each one with different froth characteristics.

Therefore, many have centered their research efforts in
understanding the basic fundamentals necessary to be incorporated in
models (Dobby and Finch, 1985; Jameson et al., 1977; Laplante et al., 1988;
Pal and Masliyah, 1989; Schulze et al., 1989; Xu, 1985) and different
models have been developed.

Dobby and Finch (1986a) developed a model based on the kinetics Of
flotation. Their work was based on the fact that for a first-order rate
process the conversion of a reactor can be predicted using the values of
the rate constant and two reactor mixing parameters: mean residence time

and dispersion number. The approach for the modeling was to develop an

51

understanding of the mixing and the kinetics of the two column sections,
collection and cleaning zones, separately. The overall model was obtained
by combining the model for these two regimes (Castillo et al., 1988; Dobby
and Finch, 1985).

Another model based on kinetics, simpler than the one above, was
used to study oil separation from aqueous emulsion. The model adequately
described the oil recovery and the order of the flotation rate was
determined to be 0.6, away from the generally accepted value of 1.0 for
mineral systems.

Sastry and Lofftus (1988) developed a general dynamic
mathematical model for column flotation based on mechanistic
representation. The model equations were normalized and dimensionless
groups directly responsible in determining the column behavior were
identified. These parameters were the normalized particle-bubble
attachment rate, the normalized particle-bubble detachment rate, the
Peclet number representing the dispersion of solid particles in the froth
phase, and the Peclet number representing the dispersion of solid particles
in the liquid phase. The resulting model equations, with no analytical
solution, were incorporated in a computer simulation.

Luttrel and co-workers (Luttrel et al., 1987; Luttrel et al., 1988;
Yoon et al., 1988) developed a basic model where transport and rate terms
were considered. The model was developed by describing the column as
divided in sections according the different flow conditions. Each section
was further divided in one or more well mixed zones. For each zone a mass
or volume balance was applied for each particulate class present. The
classes that were considered were air, unattached solids, and solids

attached to air bubbles. Particulate solids were further classified in two

52

categories, coal or mineral matter (at first, the model was developed for
column flotation of coal). The model was considered an excellent tool to

independently asses the effects of various operating parameters.

2.2.2.4 - New Concepts and Design Variations

Many variations Of flotation column have been studied and
conventional equipment, some not used as flotation machines before, has
been modified to incorporate the principles of column flotation.

Rice and co-workers (1974) used funnellike baffles inside the
column to reduce axial dispersion. The main objective was achieved and
the baffles did not present any problems regarding solids build up or
clogging. The use of baffles in bubble columns was also reported by Miyazu
and Yano (1974) and later on by libuchi and Yano (1980) and Gehle and
Schiigerl (1984).

In China, a type of column takes advantage of the two common
column geometries, round and square (Hu and Liu, 1988). These columns
have a square upper part with a round lower body. The authors listed the

following merits of the combined geometry column:

It is easy to remove the froth by scraper;
- There is more uniform pulp loading;

- A square cross-section with a side length equal to a diameter of
the round cross-section gives a larger area, thus allowing the

height to be decrease; and

- Several square columns can use a common wall, thus saving in

construction and floor space.

. Yang (1988) developed a packed column with packing similar to
plates of a heat-exchanger. The equipment is said to be ideal for cleaning
of fine particles that can not be efficiently treated in conventional
flotation columns. The key mechanism seems to be in the small passages
provided by the packing that causes an intimate particle-bubble contact.
As air moves upwards in the column it is broken in small bubbles of
relative uniform size. The author claimed that no fine bubble generator is
necessary.

A flotation column that physically is more similar to a conventional
cell due to the existence of a impeller and the feed inlet at the bottom was .
reported by Degner and co-workers (1988). The Leeds flotation column, as
it is known, consists of a series of horizontal baffles designed to strip
gangue particles from loaded bubbles as it rises to the top of the column.
This baffle system consists basically in set of two rods of different
densities. The more dense is positioned over the less dense rod. The
difference in buoyance creates a hydrostatic pinch that act on the bubbles
removing entrained hydrophilic particles from their surface.

The disadvantages of a tall column such as high ceiling requirement,
widespread concentrate and feed stream, high volume of pulp inside the
column, and cost Of pumping feed up and providing air at sufficient
pressure to overcome the hydrostatic head was addressed by Jameson
(1988) with a new column design. The height of the column could be
reduced to one fifth of the height of conventional columns by pro-mixing
air and feed in a short tube called a downcomer that is positioned through
the cleaning section. The air-feed mixture is discharged into a shallow
pool of pulp where non-attached particles disengage from bubbles and flow

downwards as tailings. Bubbles and the hydrophobic particles attached to

54

them rise to the froth bed. In other words, the collection section was
substituted by the downcomer which has approximately one tenth of the
height of the collection zone of a conventional column. Wash water inlet,
concentrate and failings withdrawal are similar to the conventional
columns.

Schneider and Weert (1988) reported the development and full scale
operation of a new flotation column called a Hydrochem column. The
column is basically a conventional column with radial centered mechanical
agitators through all its height similar to the one used by Parker (1958).
The main difference between them is that in this new concept the mixers
are designed to give to different sections of the column different degrees
of mixing. The full scale prototype column is divided in five sections: a
suspension zone for suspension of tailings to be discharged; a gas
dispersion zone where gas bubbles are generated by suspension agitation; a
collection zone where agitation is optimized to increase particle
collection; a froth zone which is the zone with the minimum agitation; and
the washing zone.

Raju and co-workers (1988) incorporated the concept of
electroflotation in column flotation that is the generation of fine and
uniform bubbles electrolytically. Three different ores were used and
control experiments with conventional columns were ran. At least for one
type of ore, chaIOOpyrite, the electrO-column was more effective.

Entrainment of magnetic gangue particles was hindered by the use of
magnetic field around part of the collection section in a new design by
Sonokilar and co-workers (1988). The column design differs from the
conventional by the absence of a froth bed and the lateral outlet for

concentrate withdrawal.

55

Ye and co-workers (1988) developed a new device by the adaptation
of a well known solid-liquid separator, the hydrocyclone. The new
flotation column, or the swirl-flow flotation column, was said ideal for
the treatment of ores with low grades due to its high capacity: 100 to 600
tons per day per cubic foot of cell volume for a nominal 5 cm x 50 cm
system compared to 1 to 2 tons per day per cubic foot of cell volume for

conventional columns and mechanical machines.

2.3 - List of References

Bader, R.; Schlitz, F.; Stacey, M. A crystalline serum muco-protein with high
choline estearase activity. Nature. 1944, 154, 183-184.

Bader, R.; Schlitz, F. Fractionation by adsorption and crystallization on
foam, part III, experiments with bile salts. Transactions of the Faraday
Society. 1946, 42, 571-576.

Castillo, D. l.; Dobby, G. S.; Finch, J. A. Fine particle separation performance
in a flotation column under conditions of heavy froth loading. In
Production and Processing of Fine Particles (Proceedings of the CIM
Conference of Metallurgists, Montreal, August, 1988); Plumpton, A., Ed.;
Pergamon, 1988.

Charm, S. E.; Moningstar, J; Matteo,C. C.; Paltiel, B. The separation and
purification of enzymes through foaming. Analytical Biochemistry. 1966,
15, 498-508.

Chiang, W. -C.; libuchi, 8.; Yano, T. Reverse enrichment in a multi-stage
bubble separator. Agricultural and Biological Chemistry. 1981, 45(11),
2469-2473.

Davis, S. G.; Fellers, C. L.; Esselen, W. B. Foam fractionation procedures in
the isolation of fruit proteins. Food Technology. 1949a, 3(6), 198-201.

Davis, S. G.; Fellers, C. L.; Esselen, W. 8. Composition and nature of apple
protein. Food Research. 19496, 14, 417-428.

56

Degner, V. R.; Sabey, J. B. Wemco/Leeds flotation column development. In
Column Flotation '88; Sastry,K. V. 8., Ed.; Society of Mining Engineers:
Littleton. 1988, pp 267-279.

Dobby, G. S.; Finch, J. A. Mixing characteristics of industrial flotation
columns. Chemical Engineering Science. 1985, 40(7), 1061-1068.

Dobby, G. S.; Finch, J. A. Flotation column scale-up and modeling. Canadian
Institute of Mining Bulletin. 1986a, 79(889), 89-96.

Dobby, G. S.; Finch, J. A. Particle collection in columns-gas rate and bubble
size effects. Canadian Metallurgical Quarterly. 1986b, 25(1), 9-13.

Dobby, G. S.; Finch, J. A. Particle size dependence in flotation derived from
a fundamental model of the capture process. International Journal of
Mineral Processing. 1987, 21, 241-260.

Dobby, G. S.; Yianatos, J. B.; Finch, J. A. Estimation of bubbles diameter in
flotation columns from drift flux analysis. Canadian Metallurgical
Quarterly 1988, 27(2), 85-90.

Dorman, D. C.; Lemlich, R. Separation of liquid mixtures by non-foaming
bubble fractionation. Nature. 1965, 207, 145-146.

Espinosa-Gomez, R.; Finch, J. A.;Yianatos, J. B.; Dobby, G. S. Flotation
column carrying capacity; particle size and density effects. Minerals
Engineering. 1988a, 1(1), 77-79.

Espinosa-Gomez, R.; Yianatos, J. B.; Finch, J. A.; Johnson, N. W. Carrying
capacity limitation in flotation columns. In Column Flotation '88; Sastry,K.
V. 8., Ed.; Society of Mining Engineers: Littleton. 1988b, pp 143-148.

Fuerstenau, M.C. Ed. Flotation, A. M. Gaudin Memorial Volume. American
Institute of Mining Engineers, New York, 1976.

Gaudin, A. M. Flotation; McGraw-Hill: New York, 1957, pp 547- 555.

Gehle, R. D.; Schligerl, K. Protein recovery by continuous flotation. Applied
Microbiology and Biotechnology. 1984, 20, 133-138.

5'7

Hu, W.; Liu, G. Design and operating experiences with flotation columns in
China. In Column Flotation '88; Sastry,K. V. 8., Ed.; Society of Mining
Engineers: Littleton. 1988, pp 35-42.

libuchi, S.; Yano, T. The performance of a Multi-stage bubble separator.
Agricultural and Biological Chemistry. 1980, 44(12), 2885-2892.

Jameson, G. J.; Nam, 8.; Young, M. M. Physical factors affecting recovery
rates in flotation. Minerals Science and Engineering. 1977, 9(3), 103-118.

Laplante, A. R.; Yianatos, J.; Finch, J.A. On the mixing characteristics Of the
collection zone in flotation columns. Column Flotation '88; Sastry, K. V. 8.,
Ed.; Society of Mining Engineers : Littleton. 1988, pp 69-79.

Leja, J. Surface Chemistry in Flotation. Plemun, New York, 1982.

Lemlich, R. ed. Adsorptive Bubble Separation Techniques. Academic, New
York, 331 p .1972.

Lemlich, R.; Lavi, E. Foam fractionation with reflux. Science. 1961, July,
191.

London, M.; Cohen, M.; Hudson, P. B. Some characteristics of enzyme foam
fractionation. Biochimica et Biophisica Acta. 1954, 13, 111-120.

Luttrel, G. H.; Adel, G. T.; Yoon, R. H. Modeling of column flotation. Society
of Mining Engineers, 1987, Preprint number 87-130.

Luttrel, G. H.; Adel, G. T.; Yoon, R. H. Hydrodynamics and mathematical
modeling of fine coal flotation. In Proceedings of XVI International Mineral
Processing Congress; Forssberg, E., Ed.; Elsevier: Amsterdam. 1988, pp
1791-1802.

Mavros, P.; Lazaridis, N. K.; Matis, K. A.; Stalidis, G. A. A study of flotation
column performance for the recovery of fine particles. Separation Science
and Technology. 1990, 25(1-2), 155-170.

Miyazu,Y.; Yano, T. Surface character and adsorbability to air bubbles of a
hydrocarbon-grown yeast. Agricultural Biological Chemistry. 1974, 38(1),
183-188.

58

Okada, K.; Akagi, Y.; Kogure, M.; Yoshioka, N. Effect on surface charges of
bubbles and fine particles on air flotation process. The Canadian Journal of
Chemical Engineering. 1990, 68, 393-399. '

Pal, R.; Masliyah, J. Flow characterization of a flotation column. The
Canadian Journal of Chemical Engineering. 1989, 67, 916- 923.

Parker, D. J. Flotation and washing of slimy gelatinous precipitates of
oxygenated metals. M. S. Thesis, Michigan State University, East Lansing,
MI, 1958.

Potter, F. J.; DeSouza, A. H. G.; Tanner, R. D. Modeling in situ protein
separation by bubble fractionation in a baker's yeast fermentation process.
Separation Science and Technology. 1990, 25(6), 673-687.

Raju, G. 8.; Prabhakar, S.; Khangoakar, P. R. Beneficiation of low grade ores
by electro-column flotation technique. In Column Flotation '88; Sastry,K. V.
8., Ed; Society of Mining Engineers: Littleton. 1988, p 293-298.

Reddy, P. S. R.; Prakash, 8.; Bhattacharyya, K. K.; Kumar, S. G.; Sastri, S. R.
S.; Narasinhan, K. S. Development and scale up of flotation column.
Proceedings of the International Symposium on Beneficiation and
Agglomeration, 2nd. 1987, pp 234-238.

Rice, R. 6.; Oliver, A. D.; Newman, J. P.; Wlies, R. J. Reduced dispersion using
baffles in column flotation. Powder Technology. 1974, 10, 201-210

Rice, R. G.; Tupperaineen, J. M. |.; Hedge, R. M.; Hedge, R. M. Dispersion and
hold-up in bubble columns - comparison of rigid and flexible spargers. The
Canadian Journal of Chemical Engineering. 1981, 59, 677-687.

Sastry, K. V. S.; Lofftus, K. D. Mathematical modeling and computer
simulation of column flotation. In Column Flotation '88; Sastry, K. V. 8.,
Ed; Society of Mining Engineers: Littleton. 1988. pp 57-68.

Saxena, A. C.; Rao, N. S.; Saxena, S. C. Bubble size distribution in bubble
columns. The Canadian Journal of Chemical Engineering. 1990, 68, 159-
161.

$

Schneider, J. C.; Weert. G. V. Design and operation of the Hydrochem
flotation column. In Column Flotation '88; Sastry,K. V. 8., Ed; Society of
Mining Engineers: Littleton. 1988, pp 287-293.

Schubert, H. Counter-flow flotation cells (flotation columns), present
state and current trends. Aufbreitungs-Technick. 1988, 29(6), 307-315.

Schulze, H. J.; Radoev, B.; Geidel, T.; Stechemesser, H.; Topfer, E.
Investigations of the collision process between particles and gas bubbles
in flotation - a theoretical analysis. International Journal of Mineral
Processing. 1989, 27(3-4), 263-278.

Schiitz, F. Adsorption on foam. Nature. 1937, 139, 629-630.

Schiitz, F. Fractionation by adsorption and crystallization on foam, part I,
methods and principles. Faraday Transactions. 1946, 42, 437-444.

Scmasundaran, P. Interfacial chemistry of particulate flotation. AlChE
Symposium Series. 1975, 71(150), 1-15.

Scmasundaran, P.; Ananthapadmannabhan, K. P. Bubble and foam flotation,
ore flotation. In Handbook of Separation Process Technology; Rousseau,R.
W., Ed.; Wiley-Interscience: New York, 1987, pp 775-805.

Sonokilar, R. L.; Mandlekar, V. A.; Gaidhani, S. B. Effect of magnetic field on
column flotation of ore containing magnetic content.ln Column Flotation
'88; Sastry,K. V. 8., Ed; Society of Mining Engineers: Littleton. 1988, pp
299-304.

Svarovsky, L. Solid-Liquid Separation. Butterworths, London,1981.

Thomas, A.; Winkler, M. A. Foam separation of biological materials. In
Topics in Enzyme and Fermentation Biotechnology, Wiseman, A., Ed; Ellis
Horwood: Sussex, 1977, pp 43-71.

Vinke, H.; Hamersma, P. J.; Fortuin, J. M. H. Particle-to-bubble adhesion in
gas/quuid/solid slurries. AlChE Journal.1991, 37(12), 1801-1809.

G)

Weber, M. E.; Loiselle, E.; Charrier, J. -M. Performance of tubular elastic
sparger. The Canadian Journal of Chemical Engineering. 1990, 68, 742-
747. '

Weijenberg, D. C.; Mulder, J. J.; Drinkenburg, A. A. H.; Stermeding, S. The
recovery of protein from potato juice waste water by foam separation.
Industrial Engineering and Chemical Process Design and Development.
1978, 17(2), 209-213.

Wheeler, D. A. Column flotation, the original column. Manuscript from a
McGill University Seminar. May, 1986.

Xu, C. Kinetic models for batch and continuous flotation in a flotation
column. Proceedings of the International Mineral Processing Congress,
15th. pp 16-27 (1985).

Xu, M.; Finch, J. A. Effect of sparger area on bubble diameter In flotation
columns. Canadian Metallurgical Quarterly. 1989, 28(1), 1-6.

Yang, D. C. A new packed column flotation system. In Column Flotation '88;
Sastry,K. V. 3., Ed; Society of Mining Engineers: Littleton. 1988, pp 257-
265.

Ye, Y.; Gopalakrishnan, S.; Pacquet, E.; Miller, J. D. Development of the air-
sparged hydrocyclone. a swirl-flow flotation column. In Column Flotation
'88; Sastry,K. V. 8., Ed; Society of Mining Engineers: Littleton. 1988, pp
305-313.

Yianatos, J. 8.; Finch, J. A.; Laplante, A. R. Holdup profile and bubble size
distribution of flotation column froth. Canadian Metallurgical Quarterly.
1986, 25(1), 23-29.

Yianatos, J. 8.; Espinosa-Gomes, R.; Finch, J. A.; Laplante, A. R.; Dobby, G. S.
Effect of column height on flotation column performance. Minerals and
Metallurgical Processing. 1987, 96(December).

Yianatos, J. B.; Finch, J. A.; Laplante, A. R. Cleaning action in column
flotation froth. Transaction of the Institute of Mining and Metallurgy.
Section C:MineraI Processing and Extractive Metallurgy. 1988, 5(1), 11-14.

61

Yoon, R.-H.; Yordan, J. L. Zeta-potential measurements on microbubbles
generated using various surfactants. Journal of Colloid and Interface
Science. 1986, 113(2), 430-438.

Yoon, R. H.; Adel, G. T.; Luttrel, G. H.; Mankosa, M. J.; Weber, A. T. Microbubble
flotation of fine particles. In Interfacial Phenomena in Biotechnology and
Materials Processing: Attia,Y. A., Moudgil, B. M., Chander, S., Eds.; Elsevier:
Amsterdam, 1988, pp 363-374.

Chapter 3

Recovery of a Genetically Engineered oI-Amylase

Abstract

A novel technique for recovery of amylases from fermentation broth
was developed by combining a unique separation of amylases with
flotation. The partition uses (hydroxypropyl)methylcellulose (HPMC), a
food grade polymer, and ammonium sulfate. First, the enzyme is removed
from the liquid phase by partition to a salted-out HPMC phase. The enzyme
containing polymer flocs are then floated from the liquid in a flotation
column by exploiting the natural floatability of these flocs. Bacillus
stearothermophilus a-amylase expressed in Escherichia coli was
recovered from fermentation broth by such a technique. Recovery of active
enzyme was as high as 90 percent with throughput as high as 94
m3/(h.m2). Floatability of the enzyme from a periplasmic extract was
higher than extracellular enzyme in the broth due to the presence of

depressors of molecular weight lower than 10,000 in the broth.

3.1 - introduction
3.1.1 - Expression of Heterologous Protein

Until the beginning of the 80's, research and commercial needs for
plant or animal proteins could not be matched in cases where relatively
large amounts were required. While some products like hemoglobin,

proteases, and insulin were naturally available in large quantities, others

62

were not. As an example, production of 80 pg of pure and biologically
active secretin required 3,000 Kg of bovine intestine for its production.
No general method for the production of these proteins was available
(Josephson and Bishop, 1988; Wilcox and Studnicka, 1988).

The development of molecular biology, including the techniques
collectively called recombinant DNA technology, made possible the
production of virtually any protein or peptide in large amounts. This is due
to the ability to identify and isolate the gene responsible for the
protein or peptide production in the original animal or plant and then
transfer it to a selected organism. This transfer is done such that the
host organism is able to use the foreign gene to produce the desired
product (heterologous protein) in a relatively controlled and predictable
way. The overall process is often termed as genetic engineering.

Despite the application of general basic concepts, the overall
process to develop a final expression system has a large empirical
component. Each gene or cDNA expression represents a peculiar case with
specific problems to be overcome (Goeddel, 1990).

There are many host organisms or expression systems in use. The
most common are the prokaryotes Escherichia coli, Bacillus subtilis,
Staphilococcus aureus, Streptomyces lividans, and the eukaryotes
Saccharomyces cerevisae and mammalian cells.

The most basic choice among these different expression systems
must be made between the type of microorganism (prokaryote or eukaryote
host) due to significant difference in their genetic machinery. Besides the
fact that eukaryotes are more complex and difficult to handle, they are
able to perform sophisticated post-translational modifications and intron

splicing that prokaryotes do not do. However, the most used expression

64

system is E. call which is linked to the development of the recombinant
DNA technology itself. E. coli physiology and genetics are much more well
understood than those of other organisms which makes for easier cloning
and high levels of the expression of heterologous proteins. E. coli can also
be grown to a high cell density with possible yields of one gram of
heterologous protein per liter of culture (Abrahmsén et al., 1986; Georgiou
etal., 1988; Gold, 1990; McGregor, 1982).

The most common locus of heterologous proteins expressed in E. coli
is in the cytoplasm. There are cases where they are transported to the
inner membrane or to the periplasmic space. Rarely are they excreted to
the medium. Secretion to the periplasm (which contains only 4 percent of
E. coli protein) or excretion to the medium would be highly desirable to
facilitate the downstream processing. In addition, extracellular release
would allow microorganisms to be grown chemostatically instead of batch
by batch (Uhlén and Moks, 1990; Wilcox and Studnicka, 1988).

When the expressed protein is confined to the cytoplasmic space,
four deleterious phenomena may occur. First, proper folding based on
disulfide bridges may not occur due to the reducing environment of the
cytoplasm resulting in an inactive protein. Second, the expressed product
may be degraded by proteases (E. coli has at least eight soluble proteolytic
activities). Third, the expressed protein may accumulate as an aggregate
of denaturated chains termed inclusion bodies. After solubilization and
purification the proteins are often in a non-active form. Complex
renaturation procedures are required to obtain an active product. And
fourth, the heterologous protein may be extremely toxic to E. coli causing
cessation of protein synthesis shortly after induction of the heterologous

gene. This toxicity may be represented by ribosome inactivation due to

65

enzymatic activity or association with membrane potential (Abrahmsén et
al., 1986; Davies, 1988; Gold, 1990; McGregor, 1982; Mitraki and King,
1989)

As in the case of traditional biotechnology, downstream processing
is a key component in the overall process of production of genetically
engineered proteins. Specific problems in separation/purification are the
frequent intracellular location, usually as inclusion bodies of non-active
protein, and hydrolysis by host proteases cited before.

It is interesting to note that one solution for the problems of
downstream processing of genetic engineered proteins lies in the
application of the same technique which made possible their expression in
the first place, recombinant DNA technology. This solution is the

expression of fusion proteins.

3.1.2 - Fusion Proteins and Purification Fusion

Fusion proteins are the result of expression of in vitro fusion of
genes or gene fragments. One of the genes or gene fragments codes for the
desired product while the other codes for a carefully chosen protein or
peptide. This protein or a peptide, a carrier, is selected for being
expressed extracellular, having a cheap, simple, and efficient purification
protocol, masking product toxicity towards the host, and/or protecting the
it from hydrolysis by the host proteases. If the fusion strategy is
successful, the carrier imparts some of its properties to the hybrid
protein. If the hybrid can be recovered or purified by exploiting the
advantageous properties previously recognized on the carrier, we have the
case of purification fusion. The product is termed the “wagon“ moiety

while the other part, the carrier, is termed "mule" moiety, purification

%

tail, purification carrier, affinity tail, affinity tag, or affinity handle. If
it is necessary, the carrier can be removed by site-specific cleavage to

generate the native gene product. The idea of such constructions came 7

from the existing natural gene fusion events such as gene duplication and
exon shuffling. Proteins seem to be tolerant to changes involving whole

domains (Uhlén and Moks, 1990).

Several fusion schemes have been devised (Figure 3.1) and a large
number of carriers identified some of which are listed in Table 3.1. Most
of these carriers were designed for purification purposes. Chromatography
(affinity, chelation, ion-exchange, etc) was generally the method of choice
but aqueous two-phase extraction has also been used (Veide et al., 1987).
However, in an extensive literature search only one example of
extracellular fusion release was found. This was the human insulin-like
growth factor fused to lgG-binding domain from staphylococcal protein A
(Moks et al., 1987). It was also the only example found of a fusion protein

processed in large scale.

3.1.3 - oI—Amylase/Alkallne Phosphatase Fusion Proteins

The study of a thermostable a-amylase from Bacillus
stearothermophilus by Oriel and Schwacha led to its recognition as a
potential purification carrier. When expressed in E. coli this enzyme was
surprisingly excreted to the medium (up to 28 percentage) in both
logarithmic and stationary phase (Oriel and Schwacha, 1988; Sen and Oriel,
1989). The study of continuous culture of the transformed E. coli showed a
extracellular activity 100 percent higher than the extracellular activity of
B. stearothermophilus batch culture and 10-fold more activity release to

the medium. The compartmentalization of the a-amylase was unaffected

_I_._ Secretion

 

 

 

,,,,,,,,, , 1'
x v» ="-- x H‘
Ins-{'12 3.1.1:, .10. ,2 12:1: ’2’. (12,2:M Win13
O

I I. C-ter-minai fusion

—
s o s s o \ s -. - --.-,.. ',
s'u's'::u :'s:\'\:: .‘ I 0‘ ' V:X' ~.' 3: - ' - "-

. a .0 ’ .‘Io
' ’ ’ U, U, :Eli’.::: {’:$’ . {ft-'50.. It...

 

 

IV . N-terrhinai fusion

"-' ' no a o a: o‘c‘o‘o‘
"3 {-‘X: ,- {-2,}. ,‘_’,:‘_ § \ s “s s s s 0‘ \

 

LSecretion-affinity fusion

2'. 2 -. ,,,,, ' _ .. _i. .'.. ."....
s\ o o no a “a no a o I ' -.'--.'-,'.‘-.‘" '-,"-"-." '.‘ ‘
fig: $3:I"- . . . H. .A ‘ ‘ ”\ ‘ L212... ‘X‘uh

 

 

VI .Secretion-insertion

 

Figure 3.1 - Examples of gene fusions used for the expression of a
recombinant protein (X). Promoters are symbolized by arrows and
transcription sequences by rectangles. The boxes represent signal
sequence (S), different affinity ”handles" (A and B), and a domain for
insertion into membranes or cell wall (I) (according Uhlen and Moks,
1990).

Table 3.1 - Carrier moieties for protein fusion

 

CARRIER

REFERBDE

 

Poly-arginine peptide

E. coli outer membrane OmpF

B-galactosidase
lgG-binding domain

Maltose binding protein

Poly-histidine peptide

Cellulose-binding domains

FIagTM peptide

Protein A

Chloramphenicol acetyltransferase
Streptavidin

Poly-glutamic acid peptide
Phosphatase-binding protein
Cystein

Protein G

Glutathione S-transferase

Brewer and Sassenfeld, 1985
Nagahari etal., 1985

Seo et al., 1988

Veide et al., 1987

Moks et al., 1987

Guan etal., 1988

Bedouelle and Duplay, 1988
Maina et al., 1988

Hochuli etal., 1988

Ong etal., 1989a

Ong etal., 1989b
Sassenfeld, 1990

Sassenfeld, 1990
Sassenfeld, 1990
Sassenfeld, 1990
Sassenfeld, 1990
Sassenfeld, 1990
Sassenfeld, 1990
Sassenfeld, 1990
Uhlén and Moks, 1990

 

Q

by dilution rate and the specific productivity increased with dilution rate
(Alexander et al., 1989).

in searching for the cause of the aforementioned release, two 5' (N
terminus) fragments of the amylase gene containing the putative
translocation sequence were fused with truncate E. coli alkaline
phosphatase. These fusions contained the amylase control and signal
sequences and the desired N-terminal fragment, followed by the alkaline
phosphatase coding region lacking its own control regions, signal
sequence, and the firSt four residues of the mature phosphatase. The
resulting hybrid protein was also excreted to the medium. Alkaline
phosphatase activity in the medium was as high as 18 percent of the total
phosphatase activity produced for the hybrid protein with the longer a-
amylase sequence (Oriel, 1989).

Recognizing the potential use of this a-amylase as a carrier of
fusion systems for expression, excretion, and purification of other
proteins and peptides and the importance of a-amylase itself (in volume,
the second largest group of enzymes produced for biotechnological purpose
according Vihinen and Mantsala, 1990), a process to recover this a-
amylase from fermentation broth was developed (Summerfelt, 1988;

Miranda and Berglund, 1990). Description of this process follows.

3.1.4 - The Recovery Process by HPMC/Salt Partition

The process for amylase recovery from fermentation broth
applicable to the amylase/phosphatase fusion above is based on partition
of the enzyme to an ammonium sulfate salted-out cellulose ether phase
previously added to the broth (Figure 3.2). These cellulose ethers can be

(hydroxypropyl)methylcellulose or methylcellulose, termed HPMC or MC,

7O

 

1 ' j y ' fi fi I v I I ‘l r Y

 

 

 

 

 

4—0

CA gafo U

352100;: B '

cu: ._ —(

53 . «

‘é’e ‘ '

a,_ 802' 2

.E-S .
'E 2 .

O-- .

275' 50'

a... . .

IS - .

L .

5‘5 40_ .,
-c>

EU) I

23 '

05-; -

L-

0.0 20' "
-‘5 -

aa- 0

3.2—0:8 b 1 A\A==A:=A_._:

0v 0 a 1‘: a a

< O 10 20 30

Ammonium sulfate concentration
(as percentage of saturation)

Figure 3.2 - Effect of ammonium sulfate in the partition of a-amylase (0),

partition of total protein ( D ), and sailing-out of HPMC (A). Data from
Miranda (1988) and Summerfelt (1988).

71

respectively. The enzyme containing particles are then pelleted by
centrifugation or filtration. The HPMC or MC polymer can be separated
from the enzyme after dissolution of the pellet and addition of ammonium
sulfate to salt-out the polymer at a lower salt concentration than at
partition. Therefore, the supernatant contains the enzyme.

The nature of the interaction between amylases and the salted-out
polymer is believed to be specific since the process works poorly for
amylase grown in medium containing starch as the carbon source. The
backbone of the polymer, cellulose, may mimic the structure of the linear
part of the amylases substrate, starch.

The process has potential applicability at large scale for many
reasons. Recovery is as high as 96 percent and requires less salt than
regular salting-out. Salt savings are 67 and 57 percent for the a-amylase
and a-amylase/phosphatase fusion, respectively. Purification for the
extracellular a-amylase was as high as 59-fold. Up to 95 percent of the
polymer can be separated from the enzyme and recycled. Polymer loading
is low: 1.1 and 2.6 mg of polymer per mg of total protein for a-amylase and
fusion, respectively. The polymer used has many advantages over polymers
used in aqueous two-phase partition like food-grade status, low cost, and
ready availability. The salt used, ammonium sulfate, does not have
negative environmental impacts when compared with the salt usually
required in these extractions, phosphates (Kula et al., 1982).

The difference between results obtained for the a-amylase and for
the fusion a-amylase/phosphatase is probably due to the use of fusion
containing only 10 percent of the whole amylase. The use of fusions with
a more complete amylase moiety may drive these results close to the ones

for a-amylase.

72

This partition process is potentially applicable not only to other
amylases and amylase fusions discussed above, but also to the purification
of other fusions whose affinity tag may bind to the HPMC. Some
candidates are the fusions of cellulose-binding domains and fusions of
maltose-binding proteins cited before (Bedouelle and Duplay,1988; Guan et
al., 1988; Maina et al., 1988; Ong et al., 1989a; Ong et al., 1989b). The
former has a specific affinity for cellulose, the back-bone of the HPMC
structure. The latter binds to maltose which is the substrate of an enzyme
also purified by the HPMC partition process, B-amylase (Miranda and

Berglund, 1990).

3.1.5 - The Problem

During development of the partition process for amylases described
above, pelleting the enzyme containing polymer phase for solid-liquid
separation was difficult. The salted-out polymer particles are better
described as light flocs made by the agglomeration of small elongated
fiberlike particles. Filtration was not efficient due to gel characteristics
of the cake formed. Filters were rapidly clogged. Centrifugation failed in
producing a compact pellet containing all flocs and a clear supernatant. A
significant number of small flocs were always present either floating in
the supernatant or attached to air bubbles at the surface of the liquid.
These phase separation problems would translate into in a combination of
larger equipment, higher processing times, loss of product, loss of
activity, and product contamination. It was concluded that an alternative
method for phase separation had to be developed in order to make the

partition process feasible for commercial scale application.

3.1.6 - The Proposed Solution

The unit operation chosen for the separation of a-amylase
containing HPMC or MC pellets from the liquid phase was flotation. Since
centrifugation and filtration, separation methods based on size and
specific gravity, were not suitable for the polymer particle recovery, we
looked at the surface properties of the polymer particles and solution. The
choice of flotation, an adsorptive bubble separation technique, was based
on the hydrophobic character of the HPMC or MC molecules and on the low
surface tension of their aqueous solutions. These polymers are the result
of hydroxypropyl and/or methyl substitutions on the cellulose that give
them hydrophobicity while preserving the original non-ionic character of
cellulose (Dow Chemical, 1988; Sarkar, 1979). Therefore, the polymer
particles could be preferentially attached to air bubbles rising in a liquid
and be lifted. In fact, this behavior was observed in some partition
experiments. At the same time, the HPMC or MC remaining In solution
lowers the surface tension of the water making the polymer a frother
(Greminger and Krumel, 1980). As a consequence, no addition of chemicals
as collectors or frothers is necessary which is important in the case of
food related or pharmaceutical products. Moreover, the amylases seemed
to be stabilized by the interaction with the polymer in a way similar to
the same effect given by immobilization of enzymes on solid matrices
(Miranda, 1988). Therefore, no loss in activity due to bubbling is expected.

The overall oI-amylase purification process can be classified as a
adsorbing colloid flotation according Wilson (1971) since the enzyme is

scavenged on a floc which is then floated out of the suspension.

74

3.2 - Materials and Methods
3.2.1 - Cellulose Ether and Other Chemicals

The cellulose ether chosen for this study was
(hydroxypropyl)methylcellulose (HPMC), type K4M supplied by Dow
Chemical, Midland. This is a product with 85,000 nominal molecular
weight and hydroxypropyl molar substitution in the range of 0.1-0.3. Stock
solutions of approximately 1 percent in concentration were prepared
according to the manufacturer's instructions: wetting and dispersing the
particles in hot water and then adding cold water to dissolve the wetted
particles (Dow Chemical, 1988).

Soluble starch (media preparation type) was obtained from Difco,

Detroit, and ammonium sulfate (ACS grade) from Columbus Chemical

Industries, Columbus. All other chemicals were reagent grade.

3.2.2 - Organism, Cultivation, and Broth and Periplasmic
Extract Preparation

The microorganism for the production of the a-amylase was the E.
coli DH1 transformed with the a-amylase gene from Bacillus
stearothermophilus (EC 217) provided by Dr. Patrick Oriel, Michigan State
University. The microorganism was grown overnight at 37 °C in LB medium
(Lennox L broth base, Gibco Laboratories) containing 50 pg/mL of
ampicillin. The cells were separated by centrifugation in a Sorvall RCSB
centrifuge at 6000 rpm for 10 min. Calcium chloride was added to the
supernatant (15 mmoles/L) for the subsequent heat treatment at 80 °C for
45 min to denature and precipitate contaminants. After cooling to 4 °C,
the precipitate was removed by centrifugation at 6000 rpm for 40 min and

sodium azide was added to the broth as a preservative (100 mg/L).

75

Periplasmic extracts were prepared by osmotic shock according to
methodology developed by Nossal and Heppel (1966) with the exception
that six times more mass of cells per volume of solution were used
throughout the procedure. Calcium addition and heat treatment were
performed as for the case of the broth. Cell and cell debris removal and
heat treatment were used to obtain a stable source of enzyme. Calcium

was added for enzyme stability and activity.

3.2.3 - Diafiltratlon

An Amicon hollow fiber cartridge system composed of a membrane
cartridge H1P10-43 with a 10,000 nominal molecular weight cut-off, 20.3
cm in length, and 0.03 m2 of surface area mounted in a DH2 holder was 2
used for diafiltration. Batch diafiltration of approximately 1 L of broth at
recirculation rate around 100 mL/min gave a filtration rate of 8 mL/min.
The buffer used for these purifications was 0.010 M sodium acetate with

0.005 M calcium chloride pH 5.25.

3.2.4 - Enzyme Partition and HPMC Salting-Out

The aqueous solution of HPMC used for enzyme partition or
partition-flotation was termed synthetic mixture while the HPMC
containing broth was termed broth mixture. Periplasmic extracts were
diluted six times with 0.020 M sodium acetate buffer pH 5.25 containing
0.005 M calcium chloride (termed herein acetate buffer) prior use in
partition experiments. Average broth activity was 8.5 units/mL (standard
deviation=1.6, n=14) and activities of the two batches of periplasmic

extract used were 596 and 237 units/mL, respectively.

76

Enzyme partitions for continuous flotation experiments were carried
out in a batch fashion at a HPMC concentration of 0.1 percent (before salt
addition) and final ammonium sulfate concentration of 30 percent
saturation by addition of a 90 percent saturation salt solution with a
peristaltic pump. Mixing was provided by a magnetic stirrer bar at a speed
that provided enough mixing without excessive bubble generation. The
reactors were cylindrical in shape and had different capacities according
to the requirement of the particular experiment (from 1.5 to 10 L).

Enzyme partitions for batch experiments were similar to the ones
described above. However, some conditions were fixed: reactor volume
was 150 mL, volume of the HPMC containing solution was 100 mL, and
volume of salt solution was 50 mL added by a syringe pump at rate of 3.3
mL/min.

Control experiments for the level of recovery of a-amylase activity
by partition at 0.1 percent HPMC and 30 percent ammonium sulfate
saturation were also carried out as batch experiments. Nine milliliters of
broth or periplasmic extract were added to 1.0 mL of 1 percent HPMC stock
solution in a test tube. During mixing in a vortex, 4.3 mL of a saturated
solution of ammonium sulfate were added drop by drop. The particles
formed were separated by centrifugation in a Sorvall centrifuge R058 for
10 min at 15,000 rpm. The pellet formed was dissolved in 9.0 mL of
acetate buffer. The dissolved pellet solution and supernatant were
analyzed for activity.

Experiments to determine the effect of pH on the salting-out of
HPMC at 30 percent saturation were conducted in batch experiments
similarly to the determination of level of a-amylase recovery by partition

described above. Nine milliliters of water were added to 1.0 mL of 1

77

percent HPMC stock solution. During mixing in a vortex, 5.0 mL of 90
percent saturated solution of ammonium sulfate were added drop by drop.
The pH of the suspension formed was then adjusted to the desired values
with ammonium hydroxide and sulfuric acid. The particles formed were
separated by centrifugation in a Sorvall centrifuge 8058 for 10 min at
15,000 rpm. Pellets were dissolved in 35 mL of water and both pellet

solution and supernatant were analyzed for total carbohydrates.

3.2.5 - Continuous Flotatlon: System and Runs

All columns used were made of pyrex glass. Spargers were rigid
cylinders made of sintered glass with pore sizes from 40 to 60 um,
diameter of 1 cm, and length of 2 cm fixed at the tip of a glass tube. Two
similar columns were used for continuous flotation experiments. The
basic design of column #2 is shown in Figure 3.3. The enlargement at the
top of the column is to avoid particle accumulation on the internal surface
of the column (Oliveira, 1990). The enlargement at the bottom
compensates for the reduction of the crossectional area due to the
presence of the sparger to avoid clogging. The cuplike device around the
top of the column collects the foam that overflows the column (the
concentrate fraction). Column #1 differs from #2 by the absence of the
top enlargement and internal part of the feed injection port and by
differences in the recovery and cleaning zone lengths, 31 and 13 cm,
respectively. The feed line consisted of a glass tip immersed into the feed
suspension, plastic tubing, and a peristaltic pump. The tailings line was a
siphon made of plastic tubing connected to the bottom of the column with
a glass tip at the other end. This glass tip end was raised up to a few

centimeters above the top of the column where it was turned downwards.

78

 

 

130

30L

 

 

 

i ”ii
117%

Figure 3.3 - Basic design of column #2. Measurements in millimeters.

8y lowering and lifting this glass tip, the tailings flow rate and,
consequently, the level of the froth inside the column, could be adjusted to
a desired level. A funnel positioned under the glass tip collected the
failings into a graduated cylinder. The gas line was a plastic tube
connecting the air reservoir with the fritted sparger. The wash line
consisted of a wash solution reservoir, a peristaltic pump, plastic tubing,
and the injection device (a syringe needle size 18 G 11/2 radially centered
and immersed approximately 0.5 cm into the column).

Continuous experiments were started by flowing air (27.5 mL/min)
and pumping suspension and wash solution into the column at desired rates
until it overflowed. The tip of the siphon for failings withdrawal was
lowered stepwise to set the froth-suspension interface to the desired
level. Once the air flow rate and froth-suspension interface level were
set to the desired values, the column was operated in a pre-run mode for a
time equivalent to the pumping of at least three recovery zone volumes of
suspension-washing solution at the current feed flow rates. The
corresponding concentrate and tailings collected were discarded. The
actual run was initiated with time recording and concentrate and tailings
collection. Concentrate was collected over water or acetate buffer for
dissolution of the HPMC particles. A graduated cylinder was used to
collect the failings. In flotation experiments with synthetic mixtures, the
failings (particles and liquid) were dissolved in water. However, in the
case of broth flotation the particles from failings were separated by
filtration on a sintered glass filter and dissolved in acetate buffer.
Synthetic and broth mixtures, concentrate, tailings liquid, and tailings

particles were analyzed for a-amylase activity and total carbohydrates.

8)

3.2.6 - Batch Flotation: System and Runs

Due to the absence in the literature of description of batch flotation
columns, a column was developed for this purpose along with a procedure
for its use. The batch column was a straight vertical tube with 66 cm in
height, 1.7 cm in diameter, internal volume of 140 mL with a bottom outlet
and overflow collector similar to the continuous columns. Design and
construction precautions were taken to minimize the ”dead volume“ below
the porous part of the sparger (the effective sparger area).

Batch flotation was initiated with the air flow through the column.
The suspension to be tested was then poured into the column until it
overflowed. Flotation time was recorded with the start of bubbling and,
simultaneously, the air flow rate was set at 4.0 mL/min. To avoid a large
drop in liquid level inside the column due to loss of liquid with the
overflown froth, a solution was continuously added at the top of the
column. This solution had the same salt concentration and pH of the
suspension under test. Its addition started at 1 min. of flotation at a rate
of 1.6 mL/min. A needle was used as the injection device and the solution
was pumped by a peristaltic pump. The concentrate overflow was
collected over water. The test was terminated with a fast drainage of the
column by opening the bottom outlet (this drained suspension was termed
failings). After HPMC particle dissolution, both concentrate and failings
solutions were used for total carbohydrate determination. In the case of
broth flotation, the volume of feed liquid that overflowed was estimated
by the ratio between absorbance at 257 nm of concentrate solution and
feed liquid (this was an empirical method using the maximum of the
absorbance spectrum of broth between 200 and 400 nm). Knowing this

volume and its total carbohydrate concentration, the amount of total

81

carbohydrate in the concentrate due to liquid overflow could be calculated.
This amount was deducted from the amount of total carbohydrate
determined in the concentrate to give the amount of HPMC particles
floated. Due to high floatability of the HPMC particles, especially in the
synthetic systems, batch flotation conditions were set to avoid total
recovery. In this way, any possible effect of variables under study on the
HPMC particle recovery could be detected.

The hydrodynamics of this column are certainly different from the
continuous column. However, the objective was to have a batch equipment
that needed small amounts of sample and time for operation rather than a
device that exactly represented column flotation. The conventional batch
devices like the Hallimond tube, bench mechanical cell, and those
developed by Fuerstenau (1964) and Partridge and Smith (1971) were
inappropriate due to complex design and requirements like particle-bubble
disengagement at the suspension-froth interface and either too low or too
high sample volume. Moreover, they are further away from duplicating the
hydrodynamics of a flotation column in continuous operation than the

column developed here.

3.2.7 - Enzyme Activity Assay

a-Amylase activity was measured using a modified iodine method
described by Yoo and co-workers (1987). One unit of activity was defined
as the of amount of substrate hydrolyzed, calculated as micromoles of
maltose, per minute under the assay conditions. Substrate solution was
1% starch in 0.1 M sodium acetate with 0.025 M calcium pH 5.25. The
hydrolysis was carried out at 70 °C by incubating 24 uL of the enzyme

solution with 238 uL of substrate solution. Incubation was terminated by

82

addition of 238 uL of 0.5 N hydrochloric acid. Five milliliters of iodine
solution (5 percent potassium iodide with 0.5 percent iodine) freshly
diluted 50 times was added to develop color. Absorbance at 555 nm was
taken with water as blank (readings termed R). The reference for color
reduction was done by adding the enzyme aliquot only after acid addition
(readings termed Ro). Control for background color of iodine solution was
prepared by substituting buffer for starch solution in the incubation
mixture (readings termed 8). Activity was determined based on the

fractional reduction of absorbance I calculated as

f = (R0 - R) / (R0 - B) (3.1)

As suggested by Yoo and co-workers (1987) and confirmed by the
results, I values between 0.20 and 0.70 were linear with respect to

enzyme concentration.

3.2.8 - Total Carbohydrates Assay

Total carbohydrates (TCH) were determined by a method based on the
one described by Dubois and co-workers (1956). HPMC can be determined
by this method because polysaccharides and their derivatives, including
the methyl ethers with free or potentially free reducing groups, give a
orange-yellow color when treated with phenol and sulfuric acid. The
procedure consists in preparing a reaction mixture with 0.2 mL of sample
and 0.2 mL of 5 percent phenol solution in 13x100 mm glass test tubes.
The addition of 1.0 mL of concentrated sulfuric acid to the surface of the
reaction mixture is performed as fast as possible to generate high
temperature during the color development period (at least 20 min). The

tubes were manually shaken and, after 10 min. the contents were mixed

83

made with water in place of sample was used as reference. Glucose was

used. as the standard.

3.2.9 - Other Experimental Methods

Different particle sizes were obtained by changing the rate of
addition of the ammonium sulfate solution and the mixing during partition.
Particle size was analyzed in a Horiba Cape-700 particle analyzer. This
equipment operates based on the sedimentation speed of particles
(gravitational or centrifugal sedimentation) which is measured by
absorbance of the particle suspension. Basic data to be fed to the analyzer
was dispersant density and viscosity and particle density. The first two
were measured using the filtrate of a suspension, the actual dispersant of
HPMC particles (1.08 g/mL and 1.18 cP, respectively). Viscosity was
measured using a capillary viscometer. However, due to the gel
characteristics of the HPMC particles, which prevent them from being
dried without destructing them, particle density was not determined.
Particle density was assumed to be the density of cellulose, 1.3 g/mL.
Bubble sizes were estimated to be in the range of 0.1 to 0.2 mm by

photographic analysis.

3.3 - Results
3.3.1 - Experiments with Synthetic Mixtures

Small variations in the salt solution flow rate and/or mixing during
partition resulted in the change of the size and morphology of salted-out
particles. At high mixing conditions and slow salt addition individual
particles of a fraction of a millimeter were formed. For a slightly faster

salt addition or lower mixing we would get elongated fibers of a few

84

millimeters which formed flocs. These irregular shapes made their size
determination by traditional methods inappropriate. The effect of particle
size on flotation of HPMC was demonstrated at two sizes where particles
were mostly individual, approximately spherical to rodlike particles, with
nominal sizes of 10 and 16 um (Table 3.2 and Figure 3.4). The differences
in conditions for production of these particles were relatively large (salt
solution flow rate of 5 and 20 mL/min, respectively). Although the
particle sizes obtained were less than one order of magnitude different
from each other, the presence of some agglomerates of filaments was
noticed in the suspension. A remarkable effect of particle size on
recovery was noticed by the different air flow rates required to maintain a
certain recovery level.

Knowing this significant effect of particle size on flotation
recovery and due to the difficulty in control and analysis of their size
distribution, care was taken to assure a constant particle size distribution
within an experiment. Each experiment was done from a single partition or
when treatments of experiments were carried out from different batches,
these batches were processed with the same conditions (volumes, mixing
and rate of salt addition). Conditions were set such that small flocs were
formed without excessive agglomeration into large flocs or elongated
strings.

Figure 3.5 shows how the recovery of l-IPMC is affected by the
ammonium sulfate concentration of the supernatant. Salt concentration
lower than 45 percent saturation allowed recovery of virtually 100
percent of the polymer. A sharp decrease in recovery takes place as

ammonium sulfate concentration increases. We qualitatively noticed a

85

Table 3.2 - Particle size distribution of HPMC particles suspension used

for experiments illustrated in Figure 3.4.

 

 

Nominal Mean Standard Deviation n
size (pm) x sd x sd

10 9.96 0.55 12.55 1.12 4

16 16.15 0.75 9.74 0.75 4

 

Particle size analyzer: Horiba Capa-700.

 

 

 

 

 

100 ' * r I T ‘ Y fifi * ' I V f a l f '
F /A"' ‘
A r A /o-ITD ‘
a . o . ,.
(U i .
E l ‘
E 95 2 1 2 .
a E . ° . I
g _ l «I
8 {‘3’ » I -
e :3 90 - -
0 o » .
E o » .
I 8’ » .
E l f: . -
m
o 85 " '
'5 > o o -
o. I 10pm ‘
a) A 16pm
g l
I A t
80 a A A J L 4 A l A A a l A a 4 l a
O 20 40 60 80

Air flow rate (mLImin)

Figure 3.4 - Response of HPMC flotation to particle size. Partition of
synthetic mixture: batch volume, 4 L; ammonium sulfate addition flow
rate, 5 and 20 mLImin for 10 and 16 um, respectively. Continuous
flotation with no washing: column used, #1; feed flow rate, 27. mLImin;
froth height, 0.5 to 3 cm.

 

 

 

 

 

.4-fi,-.4.,.-fij,.fifij,-
100$ O .
A o
(D
(D
as
E 80- ,
.73
s.- . co
C
59' .
32 6°'
09..
93,, '
00
2m 40*- .
0.01
:m
E 0
° 0
O
a, 20- o a:
0' o
a, .
as
v
0 AJAAIA.AAL-A.LIAA.AIA
4O 45 50 55 60

Ammonium sulfate concentration
(as percentage of saturation)

Figure 3.5 - Effect of ammonium sulfate concentration on the recovery of
HPMC by flotation. Salting-out of synthetic mixture: batch volume, 1.5 L;
HPMC concentration before salt addition, 0.07 percent; ammonium sulfate
addition flow rate, 14 mUmin. Continuous flotation with no washing:
column used, #1; feed flow rate, 36 mUmin; froth height, 1 cm.

 

88

decrease in the foamability of the suspension at higher salt
concentrations.

An apparent column capacity for synthetic mixtures was determined
by feeding the column at increasing flow rate (Figure 3.6). The
breakthrough point (a sharp decrease in HPMC recovery) as the feed flow
rate reached a certain value is defined as apparent column capacity. lts
nominal value was determined by fitting separate straight lines to the
lowest 7 and highest 9 feed flow rate points and finding their intercept.
The apparent capacity found was 156 mLImin which corresponds to 94
cubic meters per hour per square meter of column crossectional area. A
bubble size increase was observed with the increase of the feed flow rate.

Concentrate purification by washing of liquid from the feed out of
the froth was studied with the help of the tracer alkaline copper sulfate,
which could be detected by absorbance at 700 nm. The purification
quantified in Figure 3.7 could also be visually detected due to the strong
color of the feed liquid. Not only the change in the blue color of the froth
from treatment to treatment was noticeable, but a gradient of color was
clearly present throughout the froth.

Changes in the pH of the suspension by addition of sulfuric acid or
ammonium hydroxide did not significantly affect HPMC recovery (Figure
3.8). pH changes were limited to the range of 3.0 to 9.0 due to the
_ buffering capacity of the system. A control experiment showed that pl-l

changes do not affect solubility of HPMC in this range.

3.3.2 - Experiments with Broth Mixture and Periplasmic Extract
Continuous experiments with broth mixtures at different levels of

diafiltration showed increased recovery of HPMC with increasing

 

 

 

 

l o .
A l"
i \o
E 100 :__ _____ o_ '
33 o 0 032—5
2:53 l \
\.
In - 80" ‘
e '5 °\
C)
E g o \
§ 6OP \ .I
a \
a ’ °\
4050 ‘ L A 4100‘ A A 2150‘ A A 12004

Feed flow rate (mLImin)

Figure 3.6 - Determination of apparent column capacity. Partition of
synthetic mixture: batch volume, 10 L; ammonium sulfate addition flow
rate, 60 mUmin. Continuous flotation with no washing: column used, #1;
froth height, 3 cm.

 

0.5

U
L

0.4

0.3 b ..

0.2" ‘

0.1 ’ 0 '

 

 

 

0.0 2.0 4.0 6.0

Concentrate Contamination by Liquid From the Feed
Stream (as fractIon of concentrate volume)

Washing solution flow rate
(mUmin)

Figure 3.7 - Effect of washing on concentrate purity. Partition of
synthetic mixture: batch volume, 10 L; ammonium sulfate addition flow
rate, 60 mUmin; tracer, 2.0 percent copper sulfate (CuSO4.5H20),

0.4 percent sodium hydroxide, and 2.0 percent sodium potassium tartrate.
Continuous flotation: column used, #1; feed flow rate, 40 mLImin; froth
height, 4 cm; washing solution, 2.0 percent sodium potassium tartrate in
30 percent saturation ammonium sulfate.

91

 

HPMC recovery
(as percentage of total initial mass)

 

 

 

70 I- -I
60 n _
50 . j . 1 . l .

2.0 4.0 6.0 8.0 10.0

Suspension pH

Figure 3.8 - Effect of pH on flotation recovery of HPMC from synthetic
mixture by batch flotation. Partition: batch volume, 150 mL; HPMC
concentration, 0.1 percent; ammonium sulfate concentration and addition
rate, 30 percent saturation and 3.3 mLImin, respectively. Flotation: air
flow rate, 4.0 mUmin; make-up water flow rate, 1.6 mUmin; flotation
time, 5 minutes.

 

W

purification of the broth up to the point of total recovery (Figure 3.9).
Activity recovery followed the same pattern increasing up to the total
recovery possible by the partition (91 percent, determined by control
partition experiments). Both maximum HPMC and activity recoveries took
place when broth was diafiltrated with at least two volumes of buffer per
volume of broth. A linear relationship was observed between the recovery
of HPMC and recovery of activity with HPMC particles. Figure 3.10 is a
plot of the actual recovery of activity versus the theoretical recovery of
activity based on the recovery of HPMC particles (slope of 0.88 and
regression coefficient of 0.992). The slope deviated slightly from the
expected value of one indicating that the flotation operation recovered all
the enzyme that was associated with the particles. 2

Periplasmic extracts were much more efficiently treated than broth
with virtually 100 percent recovery of particles and total possible
activity recovery (90 percent, determined by the control partition
experiment). An attempt to reduce the salt used in the flotation was made
by lowering the salt concentration in the washing solution (Table 3.3).
However, the slight reduction from 30 to 25 percent saturation resulted in
about one third reduction in the activity recovered (from 90 to 62 percent
of total initial activity).

pH dependence of l-IPMC recovery from a broth mixture was
demonstrated by batch experiments (Figure 3.11); the lower the pH, the
higher the HPMC recovery. Continuous experiments were done to evaluate
the influence of pH on the recovery of activity (Figure 3.12). HPMC
recovery had the same behavior as in the batch flotation experiments for
the pH range from 3.0 to 7.0: HPMC recovery increased as the pH was

lowered. The lowering of the pH to 2.5 decreased the HPMC recovery

 

 

 

 

 

 

£3
E I HPMC
8 100 '- A/A- -
g .2 . é/ acnvrrv
8 ‘8' 8° °
2 - P
m .E
:5 7.2
O O
3 "' 40 ~ .
:2 “a ,
‘6
< 8
S 20 ‘ -
C
0)
o l
B
Q o J l ; L 4 l
,,, 0.0 1.0 2.0 3.0
Cu

Diafiltration level
(volume of buffer per volume of broth)

Figure 3.9 - Flotation response to broth pre-purification by diafiltration.
Partition of broth mixture: batch volume, 1.5 L; ammonium sulfate addition
flow rate, 31 mLImin. Continuous flotation without washing: column used,
#2; feed flow rate, 20 mLImin; froth height, 5 cm.

 

Activity recovery
(as percentage of total initial activity)

 

 

 

 

50 4 J A 1 4 l A J
50 60 7O 80 90

HPMC recovery x 0.912
(recovery as percentage of total initial mass)

Figure 3.10 - Correlation between actual activity recovery and theoretical
activity recovery calculated as 0.912 x HPMC recovery. Data from
experiment depicted on Figure 3.9.

r3

%

Table 3.3 - Flotation of periplasmic extract and the effect of lowering the

ammonium sulfate concentration in the washing solution.

 

 

Concentration of ammonium Activity recovery
sulfate in the washing solution (as percentage of total activity)
(as percentage of saturation) Run#1 Run#2 Average

30 91 88 90
25 62 62 62
20 45 41 43

 

Conditions: Partition of periplasmic extract: batch volume, 600 mL;
ammonium sulfate addition flow rate, 10 mL/min. Continuous flotation:
column used, #2; feed flow rate, 20.0 ml/min; washing flow rate, 10
mL/min; froth height, 7 cm.

 

 

 

 

 

A 60* v 1
(I)
(I)
m
E
— 50l- u
g <> .
a?
275 4°" ‘
88 I
2“ P a»
($3.5 30" -I
(1.8, » § i
Ice
5 2°" ( 6 ‘
2
m
0.10,. c-
(D
i! . .
0 4 l A l A J A
2.0 4.0 6.0 8.0 10.0

Suspension pH

Figure 3.11 - Effect of pH on flotation recovery of HPMC from broth
mixture by batch flotation. Partition: batch volume, 150 mL; HPMC
concentration, 0.1 percent; ammonium sulfate concentration and addition
rate, 30 percent saturation and 3.3 mLImin, respectively. Flotation: air
flow rate, 4.0 mUmin; make-up water flow rate, 1.6 mLImin; flotation
time, 5 minutes.

 

 

 

 

 

 

31m 1 I fl T i 1 fi I f I
(U
E , .
8
5230’ ‘
BE
00-0 b
928
g3: 60» -I
ab:
2:5 .
87.2
£2 40l- _
.2...
‘60 . i
<0)
0)
£3 20" 1
C
G)
e l I
Q
O.
m o A l J A 1 A 1 L l
g 20 30 4.0 5.0 60 70

Suspension pH

Figure 3.12 - pH effect on recovery of HPMC particles, activity associated
with HPMC particles, and total activity. Partition of broth mixture: batch
volume, 4 L; ammonium sulfate addition flow rate, 42 mUmin. Continuous
flotation : column used, #2; feed flow rate , 20 mUmin; washing solution
flow rate, 7 mUmin; froth height, 7 cm.

%

making the pH 3.0 a maximum. Activity recovered with the particles also
increased with pH decrease to a value at which a maximum was reached.
However, the maximum for activity recovery took place at a higher pH than
that for HPMC recovery (pH 4.0 instead of pH 3.0). Total activity recovery
was as high as 90 percent for pH equal to or higher than 4.0 with a
continuing decrease for lower values.

Continuous experiments produced a concentration ratio from 3 to 4
(ratio of volume of feed to volume of concentrate) and concentrate
contamination by the feed liquid around 30 percent.

Temperature increase resulted in a decrease in the recovery of
particles (Table 3.4). An increase in bubble size could be seen as the

temperature was raised.

3.4 - Discussion

Nearly total recovery of HPMC particles from both the synthetic
mixture and the periplasmic extract was possible. The broth mixture had a
lower, but significant floatability. However, pre-purification of the broth
made it possible to achieve total recovery of activity and HPMC particles.
Moreover, the recovery can be optimized with the increase of the air flow
rate. Typical superficial air velocity in mineral processing is in the range
of 1 to 3 cm/s while we used a value of 0.18 cm/s in these continuous
experiments (Dobby et al., 1988). The linear relationship between recovery
of HPMC particles and recovery of enzyme activity indicates that no
significant enzyme denaturation occurred due to flotation (the activity
recovered in the concentrate was proportional to the recovery of the
particles). The recovery of active enzyme/HPMC for the pH experiments

was lower than the recovery for the diafiltration experiments. This

99

Table 3.4 - Temperature effect on the flotation recovery of HPMC from

broth

 

 

Temperature (°C) HPMC recovery (as percentage)
Run#1 Run#2 Average
2510.5 19 25 22
4012.0 15 11 _ 13
561:4.0 6 5 6

 

Conditions: Partition of broth mixture: batch volume, 600 mL; ammonium
sulfate addition flow rate, 10 mL/min. Continuous flotation: column used,
#2; feed flow rate, 20.0 mI/min; washing flow rate, 10 mLImin; froth
height, 7 cm.

1%

decrease in efficiency is probably caused by hydrodynamic disengagement
due to the introduction of washing at this experiment.

The low activity recovery when the salt concentration in the
washing solution was decreased is the result of the large ratio of volume
of washing solution to volume of suspension in the froth (approximately 2
due to wash solution flow rate of 10 mL/min and concentrate overflow
around 5 mUmin). The salt concentration is close to the value for the
washing solution and partition strongly depends on the salt concentration
of the system (Figure 3).

The low floatability for the broth mixture is not surprising. The
electrokinetic properties of the solid/water interface of hydrophobic
solids are primarily or completely determined by the aqueous phase. The '
electric double layer is entirely on the water side of the interface. Also,
the behavior of the solid/water interface is similar to the behavior of the
air/ water interface (Arbiter et al, 1975). Due to the complex composition
of the broth, interactions of its components at the solid/liquid and
gas/liquid interfaces were expected. However, not all are expected to be
neutral toward floatability. The components may act as depressors,
activators, or collectors. In this specific case, a depressant effect
prevailed. The improved floatability of the broth pre-purified by
diafiltration indicates that these depressors have, or are activated by, low
molecular weight molecules (less than 10,000). These molecules are
- probably amphipatic entities of which nonpolar ends interact with the
nonpolar structure of the HPMC exposing the charged groups to the liquid.
These charged groups may impart to the particle a net electric charge
responsible for the decrease in hydrophobicity and, as a consequence,

decrease in floatability.

101

Some insight into this depressor phenomenon was provided by the pH
experiments. No alteration in floatability was noticed when the pH of
suspension of synthetic mixture was modified. However, in the case of
broth mixtures, pH changes seemed to interfere with intermolecular
interactions. The higher recovery at low pH suggested a negatively
charged molecule as the depressor which is neutralized as pH is lowered.
This depressor could be an amphiphilic molecule having its hydrophobic
moiety positioned towards the hydrophobic phases (particles and air
bubbles) and its charged moiety towrds the water phase. Therefore, this
negative charge increases both the particle affinity to water and the
repulsion from an also negatively charged air bubble. An excessive
decrease in pH would cause its protonation to a positive net charge with
decrease in floatability.

The media used in the fermentation, LB medium, is rich in peptides
and polypeptides since it contains peptone and yeast extract. Due to the
diversity of chemical character that polypeptides can have, some may act
as depressors of the HPMC particles. Polypeptides that are relatively rich
in glutamic, aspartic, and hydrophobic residues like Ieucine, isoleucine,
phenylalanine, tryptophan, and tyrosine are consistent with the depressor
pictured above, Le, a negative charged molecule at neutral to moderate
low pH with hydrophobic patches.

The effect of particle size on flotation can be explained by the
limited carrying capacity of bubbles and different specific area of
particles of different sizes. A bubble has a limited area to which
particles can attach (onlythe bottom part is effective on carrying
particles). The smaller the particles, the larger their specific surface

area which means that small particles have more surface area than large

 

102

particles for the same mass. Consequently, small particles require more
bubble area to achieve the same level of flotation; or small particles have
lower recovery for the same bubble area available (gas flow rate).

The column capacity measured is termed apparent capacity because
it is lower than the true column capacity. In the determination of apparent
capacity, the increase in the feed flow rate at a fixed particle
concentration causes an increase in the amount of HPMC fed to the column.
Simultaneously, it also decreases the residence time of particles in the
equipment which reduces their chance to be picked up by the bubbles. The
net result is lower recovery. However, this apparent capacity is a good
indicator for capacity and consequently column size for a specific
separation. Despite of the high column throughput for this system, 94
m3/(h.m2), this is not a Optimized value. The dependence of capacity on
the bubble area available for flotation makes it possible to increase the
capacity by operating at higher air flow rates as long as flooding does not
occun

The bubble size increase observed with the increase of the feed flow
rate may be due to the heavy bubble load conditions reached. As two
bubbles get close to each other, the particles attached to them seem to be
in a thin liquid film. Bubble coalescence caused by hydrophobic particles
in thin films is a known phenomenon and there are two mechanisms by
which it takes place. The hydrophobicity of the particle induces the film
thinning by the hydrophobic interaction between the solid and the gas
phase to a point that ruptures occur. Also, the particle may partially
dissolve in this film spreading its surfactant molecules over the

air/water interface. Water in the film flows along with these molecules

 

 

103

(the Marangoni effect) away from the particle causing the film to thin
around it up to the rupture point (Walstra, 1989).

Extensive washing accomplished with the synthetic mixture is the
characteristic that enables column flotation to combine the two goals of
solid recovery from liquid suspension (particle concentration and liquid
removal) in one step. Washing efficiency will vary from system to system
due to different liquid viscosity and density, bubble size, and how the
froth is packed. However, the result obtained here (Figure 3) illustrates m

the potential of this flotation column feature in contributing to the overall

purification of the partition process.

 

The level of concentrate contamination (30 percent) obtained for the
broth mixture during the pH experiments is high compared with the results.
obtained for the synthetic mixture. Also, the volumetric concentration
ratio from 3 to 4 is not overly impressive. This high contamination and
low volumetric concentration ratio are the result of a finer froth
structure composed of small bubbles. As a consequence, the liquid content
of the froth is higher than the values for synthetic mixture. Therefore,
there is a higher overflow rate out of the column. This results in a higher
rate of feed liquid in the froth to be pushed down the column by the
washing solution. The final result is a reduced concentration ratio and
purification efficiency.

This froth characteristic may be due to an increased stability
conferred to the bubbles by surfactants from the fermentation broth. It is
known that proteins or polypeptides at the air/water interface may
promote a concentration of polysaccharide frothers in this interface by
interacting with them. The result is a more stable foam due to the

increase of the thickness and mechanical strength of the film between

 

104

bubbles (Dickinson, 1989). Polypeptides and proteins in the broth may be
enhancing the action of the soluble HPMC as a frother beyond a desired
level. The use of food-grade antifoams may set foamability to a optimum
level.

The washing effect is expected to be improved in a scaled-up column
with a larger cross-sectional area due to a better distribution of the
washing solution traveling down the column. At low shear rates, foam
flowing inside a column exhibits a non-Newtonian behavior following a
Casson model. Due to a yield stress, the velocity profile is flat at the
center of the column. At the wall, there is a thin layer of liquid devoid of
gas bubbles which is characteristically Newtonian, called a skimming
layer. Therefore, the velocity profile close to the wall is linear and goes 2
through a transition towards the central region where it reaches the
plateau (Mahalingam et al., 1975). As a consequence, there may be a
uneven distribution of the washing solution throughout the cross-sectional
area of the column. The high inertia of the foam at the central part of the
column may divert the washing solution towards the wall. With most of
the washing solution flowing away from the center of the column,
excessive washing may take place at the region close to the wall at the
expense of poor washing at the center. This can result in chanelling
resulting in poor utilization of the wash solution. This phenomenon can be
called a wall effect. A column with a larger cross-sectional area will
experience a smaller wall effect due to a larger cross-sectional area-to-
perimeter ratio.

More complete insolubilization of the HPMC, the frother of the
system, at high salt concentration and temperature would contribute to

the froth instability and to the low recovery at these conditions. Larger

 

105

bubbles in the recovery zone and coalescence to even larger bubbles at the
froth bed decrease the bubble surface area available for particle pick-up in
the suspension and their transport through the froth. We should note that
HPMC has a “inverse“ solubility property, i.e., it is more soluble in water at
low temperatures than at high temperatures (Sarkar, 1979). Another way
to improve volumetric concentration ratio is by lowering the injection
point of wash water. This increses the length of the draining froth bed.
More liquid drains down the column and, consequently, a drier foam

overflows.

3.5 - Conclusions

In summary, the recovery of a-amylase by the combination of
partition to a HPMC/salt phase with subsequent solid-liquid separation in
a flotation column was shown to be feasible. Processing of synthetic
systems showed that high polymer recovery, and consequently high enzyme
activity recovery, and equipment capacity are possible with good
purification even before optimization of operational variables. The
processing of fermentation broth and periplasmic extract confirmed the
feasibility of the process, despite results not as good as for the synthetic
systems.

Scaling-up of the process seems to be straightfoward and suitable
for large scale operations due to its simplicity, efficiency, high
throughput, and low cost.

Overall, the flotation oi this biological system behaved similarly to
flotation of mineral systems. The same basic principles and mechanisms
were valid. Specifics due to the biological nature of the feedstock exist

and, consequently, solution of some problems required approaches more

 

106

restricted to biotechnological processing (e.g., diafiltration). Other
problems were addressed by procedures common to mineral processing
(e.g., pH adjustment). As it was shown for this partition process, it is
possible that other particulated systems of difficult solid/liquid
separation used in separations in biotechnology can be processed in as
flotation.

The important conclusions from the experimental results are:

a) The potential capacity of flotation columns in the treatment of

(hydroxypropyl)methylcellulose particles salted-out at 30 percent

 

saturation of ammonium sulfate is on the order of 100 m3/(h.m2). It is
independent of pH and dependent on the particle size and salt

concentration.

b) The washing capability of the column can allow the produciton of
concentrates with low contamination of feed liquid (as low as 2 to 3

percent).

c) The recovery of a-amylase directly from the fermentation broth
by flotation is feasible at reasonable capacity. The capacity can be
increased by the elimination of small molecular weight compounds (less
than 10,000) from the enzyme preparation that act depressants of the
flotation. Also, these depressant effects can be counterbalanced by pH

lowering down to pH 3.

d) The flotation column demonstrated higher capacity for the

flotation of periplasmic a-amylase than for the extracellular enzyme.

9) There is no detectable enzyme denaturation due to the partition or

the flotation steps.

 

107

3.6 - List of References .

Abrahmsén, L.; Moks, T.; Nilsson, 8.; Uhlén, M. Secretion of heterologous
gene products to the culture medium of Escherichia coli. Nucleic Acids
Research. 1986, 14(18), 7487-7501.

Alexander, P.; Oriel, P. J. Continuous culture of Escherichia coli for
extracellular production of recombinant amylase. Biotechnology
Letters.1989, 11(9), 609-614.

Dow Chemical, Midland, MI. Methocel cellulose ethers technical handbook.
Form No. 192-1062-88-JB, 1988

Arbiter, N.; Fujii, Y.; Hansen, 8.; Raja, A. Surface properties of hydrophobic
solids. AlChE Symposium Series. 1975, 71(150), 176-182.

Bedouelle, H.; Duplay, P. Production in Escherichia coli and one step
purification of bifunctional hybrid proteins which bind maltose. European
Journal of Biochemistry. 1988, 171, 541-549.

Brewer, S. J.; Sassenfeld, H. M. The purification of recombinant proteins
using C-terminal polyarginine fusions. Trends in Biotechnology. 1985,
3(5), 119-122.

Davies, J. Genetic engineering, processes and products. Trends in
Biotechnology. 1988, 6(4), S7-SI1.

Dickinson, E. Protein adsorption at liquid interfaces and the relationship to
foam stability. In Foams, Physics, Chemistry, and Structure. Wilson, A. J.,
Ed.; Springer-Verlager: Berlin, 1989, pp 39-53.

Dobby, G. 8.; Yianatos, J. 8.; Finch, J. A. Estimation of bubbles diameter in
flotation columns from drift flux analysis. Canadian Metallurgical
Quarterly 1988, 27(2), 85-90.

Dubois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Colorimetric
method for determination of sugars and related substances. Analytical
Chemistry. 1956, 28, 350-356.

 

KB

Fuerstenau, M. C. An improved micro-flotation technique. Engineering and
Mining Journal. 1964, 11, 108-109

Georgiou, G.; Wilson, D. 8. Release of perisplamic enzymes and other
physiological effects of B-lactamase overproduction in Escherichia coli.
Biotechnology and Bioengineering. 1988, 32, 741-748.

Goeddel, D. V. Systems for heterologous gene expression. Methods in
Enzymology.1990, 185, 3-7.

Gold, L. Expression of heterologous protein in Escherichia coli . Methods in
Enzymology.1990, 185, 11-15.

Greminger, G. K.; Krumel, K. L. In Handbook of Water Soluble Gums and
Resins; Davidson, R. L., Ed.; McGraw-Hill: New York, 1980, pp 3/1-3/25.

Guan, C.; Li, P.; Riggs, P. D.; lnouye, H. Vectors that facilitate the
expression and purification of foreign peptides in Escherichia coli by
fusion to maltose-binding protein. Gene. 1988, 67, 21-30.

Hochuli, E.; Iannwarth, W.; Ddbeli, H.; Gentz, R.; Stiner, D. Genetic approach
to facilitate purification of recombinant proteins with a novel metal
chelate adsorbent. Biotechnology. 1988, November, 1321-1325.

Josephson, 8.; Bishop, R. Secretion of peptides from E. coli, a production
system for the pharmaceutical industry. Trends in Biotechnology. 1988, 6,
218-224.

Kula, M. -R.; Kroner, K. H.; Husted, H. In Advances in Biochemical
Engineering, Reaction Engineering; Fietcher, A., Ed; Springer-Verlag:
Berlin,1982.

Mahalingam, R; Surati, H. S.; Brink, J. A. High-expansion foam flow
analyses. AlChE Symposium Series. 1975, 71(150), 52-58.

Maina, C. V.; Riggs, P. D.; Grandea, A. G.; Slatko, B. E.; Moran, L. S.;
Tagliamonte, J. A.; McReynolds, L. A.; Guan, C. An Escherichia coli vector to
express and purify foreign proteins by fusion to and separation from
maltose-binding protein. Gene. 1988, 74, 365-373.

was: ”K.;".
I

109

McGregor, W. C. Large-scale isolation and purification of proteins from
recombinant E. coli. Annals of New York Academy of Sciences. 1982, 231-
236.

Miranda, E. A. Recovery of Clostridium thermosulfurogenes B-amylase from
fermentation broth by two-phase partition. M. S. Thesis, Michigan State
University, East Lansing, MI, 1988.

Miranda, E. A.; Berglund, K. A.; Recovery of Clostridium thermosulfurogenes
produced B-amylase by (hydroxypropyl) methylcellulose partition.
Biotechnology Progress. 1990, 6, 214-219.

Mitraki, A.; King, J. Protein folding intermediates and inclusion body
formation. Biotechnology. 1989, 7, 690-697.

Moks, T.; Abrahmsén, L.; Osterldf, 8.; Josephson, S.; Ostling, M.; Enfors, S. -'
O.; Persson, l.; Nilsson, 8.; Uhlén, M. Large-scale affinity purification of
human insulin-like growth factor I from culture medium of Escherichia
coli. Biotechnology. 1987, 5, 379-382.

Nagahari, R.; Kanaya, 8.; Munakata, K.; Aoyagi, Y.; Mizushima, S. Secretion
into the culture medium of a foreign gene product from Escherichia coli,
use of ompF gene for secretion of human fl-endorphin. The EMBQ Journal.
1985, 4(13), 3589-3592.

Nossal, N. G.; Heppel, L. A. The release of enzymes by osmotic shock from
Escherichia coli in exponential phase. Journal of Biological Chemistry.
1966, 241(13), 3055-3062.

Oliveira, M. L. M. Personal communication. Comissao Nacional de Energia
Nuclear, Centro de Desenvolvimento de Tecnologia Nuclear, Brazil, 1990.

Ong, E.; Gilkes, N. R.; Warren, R. A. J.; Miller, R. C.; Kilburn, D. G. Enzyme
immobilization using the cellulose-binding domain of a Cellulomonas fimi
exoglucanase. Biotechnology. 1989a, 7, 604-607.

Ong, E.; Greenwood, J. M.; Gilkes, N. R.; Kilburn, D. G.; Miller, R. 0.; Warren, A.
J. The cellulose-binding domains of cellulases, tools for biotechnology.
Trends in biotechnology. 1989b, 7(9), 239-243.

 

\: T.;-1

 

 

110

Oriel, P. J. Personal communication. Michigan State University, 1989.

Oriel P.; Schwacha, A. Growth on starch and extracellular production of
thermostable amylase by E. coli. Enzyme and microbial technology. 1988,
10, 41-46.

Partridge, A. C.; Smith, G. W. Small-sample flotation testing, a new cell.
Transactions of the Institute of Mining and Metallurgy, section C. 1971,
80, C199-C200.

Sarkar, N. Thermal gelation properties of methyl and hydroxypropyl
methylcellulose. Journal of Applied Polymer Science. 1979, 24, 1073-
1087.

Sassenfeld, H. M. Engineering proteins for purification. Trends in
Biotechnology. 1990, 8, 88-93.

Sen, 8.; Oriel, P. Multiple amylase genes in two strains of Bacillus
stearothermophilus. Gene. 1989, 76, 137-144.

Seo, J.-H.; Loffler, A. l.; Bailey, J. E. A parametric study of cloned fusion
protein expression in Escherichia coli. Biotechnology and Bioengineering.
1988, 32, 725-730.

Summerfelt, S. T. The concentration and purification of an a-amylase using
a methylcellulose-salt, two phase partitioning process. M. S. Thesis,
Michigan State University, East Lansing, Ml,1988.

Uhlén, M.; Moks, T. Gene fusion for purpose of expression, an introduction.
Methods in Enzymology.1990, 185, 129-143.

Veide, A.;Strandberg, L.; Enfors, S. -O. Extraction of B-galactosidase fused
protein A in aqueous two-phase systems. Enzyme and Microbial Technology.
1987, 9, 730-738.

Vihinen, M.; Mantsala, P. Characterization of a thermostable Bacillus
stearothermophilus a-amylase . Biotechnology and Applied Biochemistry.

1 990, 12, 427-435.

 

111

Walstra, P. Principles of foam formation and stability. In Foams, Physics,
Chemistry, and Structure. Wilson, A. J., Ed.; Springer-Verlager: Berlin,
1989. Pp 1-15.

Wilcox, G.; Studnicka. Expression of foreign proteins in microorganisms.
Biotechnology and Applied Biochemistry. 1988, 10, 500-509.

Wilson, J. D. Adsorbing colloid flotation methods. In Physical Separations;
Freeman, M. P.; FitzPatrick, J. A., Ed.; Engineering Foundation: New York,
1971, pp 181-220.

Yoo, Y. J.; Hong, J.; Hatch, R. T. Comparison of a-amylase activities from

different assay methods. Biotechnology and Bioengineering. 1987, 30,
147-151.

 

Chapter 4
Affinity Flotatlon Based on Ligand-Ligate
Hydrophobic Interaction

Abstract

A flotation system was developed to perform protein separation
based on hydrophobic interaction. It consisted of a ligand covalently
attached to the carrier molecule methylcellulose, a polymer that when
salted-out by ammonium sulfate can be floated. The ligands were
saturated fatty acids of linear chains attached to the methylcellulose by
esterification. These methylcellulose derivatives were partially soluble
in water and their salting-out properties and floatability were close to
the unsubstituted methylcellulose. When the derivatives were tested for
the precipitation of bovine serum albumin, the butyric acid ligand showed
a higher efficiency than others tested. However, the floatability of the
protein-derivative particles was low due to the high salt concentration
required for binding and the high protein concentration. There is evidence
that the major driving force for the precipitation is hydrophobic

interaction.
4.1 - Introduction and Objectives

One of the most successful strategies for product recovery in

downstream processing of fermentation products, especially proteins, is

112

113

the use of affinity separations. These are separation techniques based on
affinity interactions; i.e., the result of a selective and reversible
interaction between two entities called the ligate and the ligand. The
main advantage of these techniques is the high purification attained by
selective binding of ligates to ligands. Ligands can recognize either an
entire class of ligates (general ligands) or only one or two (specific
Hgands)

Ligate-ligand interactions occur in accordance with their biological
functions. For example, coenzymes, antibodies, inhibitors, and substrates
are commonly used as ligands for protein purification (Scouten, 1981).
Some ligands such as hydrophobic groups are not classified as affinity
ligands when ligate-ligand bonding is not related to the biological activity.

The unit operation by which the separation is actually carried out
depends on the physical characteristics of the ligand or ligand-ligate
complex under the conditions of separation. Therefore, affinity
precipitation, affinity ultrafiltration, affinity aqueous two-phase
extraction, and affinity adsorption have been developed (Flygare et al.,
1982; Janson, 1984; Male et al., 1987; Mattiasson et al., 1986; Scopes,
1987). The most common affinity unit operation is the adsorption of the
ligate onto a solid matrix where the ligand isimmobilized. Despite the use
of membranes, fibers, and sponges the most popular solid matrix used, in
both laboratory and industrial scale, is small agarose beads.

In the vast majority of the adsorption processes described in the
literature, the adsorption and desorption steps are carried-out in batch
fashion with the beads contained in chromatographic columns due to the
convenience of handling beaded shaped adsorbents in such devices.

Therefore, such separation techniques are called affinity chromatography.

 

114

However, among the many cases of affinity chromatography reported in the

literature, only few cases of true chromatography (the simultaneous but

differential migration of different molecules along a packed bed) are cited

(Janson, 1984).

The drawbacks of affinity chromatographic processes are similar to
the problems found in other processes that utilize packed beads. The
difficulty In operating on a continuous basis, the need to find the optimum
flow rate (balance between high capacity and adsorption efficiency), the
consolidation of the bed at high flow rates due to mechanical weakness of
beads, and the need of a feed free of particulate material are a few of
them. The significance of these problems are magnified by the fact that
affinity chromatography is a four-stage operation including adsorption,
washing, desorption, and washing.

To circumvent these problems, systems like CARE (continuous
affinity-recycle extraction), a combination of a batch adsorption with
packed bed desorption, partition of adsorbents to aqueous two-phase
system, fluidized beds, and magnetized adsorption beads have been
suggested (Gordon et al., 1990; Janson, 1984; Pungor et al., 1987; Yang et
al., 1989; Yang and Goto 1992). To date, it seems that these techniques
have only been used on a laboratory scale.

Flotation is a potential unit operation for affinity separations
without the negatives associated with packed bed operation. The basic
requirement for affinity flotation is that the ligand-ligate complex must
be insoluble and floatable under the binding conditions. The literature is
abundant with reports of successful affinity precipitations that may suit
these requirements (Scopes, 1987). However, the potential of affinity

flotation is not restricted to known affinity precipitation systems. We

a“

“.13.":

115

can design and prepare molecules for this purpose. If the complex ligate-
ligand is not insoluble or floatable, the attachment of the ligand to a
carefully chosen carrier may produce a ligate-ligand-carrier complex that
is insoluble and floatable. Its important to understand the properties of
the carrier to a considerable depth since, ideally, its properties would
markedly be expressed by the ligand-ligate-carrier complex. Therefore,
the knowledge of carrier precipitation and flotation could be used to
predict ligand-ligate-carrier precipitation and flotation.

The advantages of flotation over adsorption on solid matrices are
many. First, flotation is a typical and a simple continuous unit operation
well know for its high throughput. By designing a system where ligate-
ligand binding precedes insolubilization, we can take advantage of the fast
kinetics typical of liquid systems. The presence of particles other than
ligate containing particles is of little concern due to the ability of the
flotation process to reject non-floatable particles in the failings stream
by the action of the wash solution. This feature also reduces the four
steps needed when packed beds are used (adsorption, washing, desorption,
and washing) to two steps (adsorption-washing and desorption-washing)
since, ideally, the washing rejects all feed liquid from the concentrate.

This chapter describes the development of an example of the affinity
flotation system discussed above composed of flotation of bovine serum
albumin (BSA) bound by hydrophobic interaction to a linear alkyl chain
attached to a floatable methylcellulose. The carrier, methylcellulose, and
type of interaction for ligand-ligate binding (hydrophobic interaction)
were selected based on the common characteristic of high salt
concentration requirement. Hydrophobic interaction, as discussed above,

is usually driven by high salt concentration. Flotation of cellulose ethers

'K.._ |f

116

was discussed in Chapter 3. A high salt concentration is also required for
insolubilization and flotation of the cellulose ether. Thus, high salt
concentration favors both the chemical and physical requirements of
flotation. Also, there is a vast literature in this field describing the
purification of a large number of different proteins, including enzymes, by
hydrophobic chromatography. Particularly, bovine serum albumin has been
extensively studied as a model system for hydrophobic chromatography.
One of serum albumin's biological functions is the transport of fatty acids
in the blood stream (Peters et al., 1973; Wichman and Anderson, 1974);
therefore, fatty acids are a widely used ligand for serum albumin.
Derivatization using linear aliphatic chains was used to impart
different degrees of hydrophobicity to methylcellulose by varying the size
and the degree of attachment. Acid catalyzed esterification by fatty
acids.was used since it is simpler than the other methods for
derivatization of agarose. The common procedure for hydrophobic
chromatography is to activate the matrix, attach a linear spacer, and then
attach the ligand to this spacer. However, a spacer may not be as
important in the case of soluble ligand-carriers as in the case of solid
matrices since the flexibility of a soluble carrier of high molecular weight
may make the ligand accessible to the ligate. The esterification of the
hydroxyl groups of the methylcellulose is not only a simpler procedure, but
also may subject the methylcellulose to fewer side-reactions. An
extensive literature search,did not reveal any examples of agarose
derivatization by esterification. It is possible that the acid conditions of
such a reaction causes degradation of the agarose beads altering their

porosity, shape, and their physical strength. As mentioned before, the

117

physical Strength of the beads is an important characteristic for

adsorbents used in packed beds.

4.2 - Hydrophobic Interaction

The term hydrophobic was first used by J. Perrin in 1905 to refer to
colloid sols that readily coagulate by addition of electrolytes (Melander
and Horvath, 1977). Hydrophobic substances are defined by Tanford (1980)
as those readily soluble in nonpolar solvents and only sparingly soluble in
water. These substances are nonpolar or have a structure primarily
nonpolar. The apparent repulsion from water is a common property and
explains the name. However, some disagree with this name claiming that
there is a significant attraction between these substances and water 6
molecules (Bull and Breese, 1980). Their association in aqueous media is
of enormous importance, especially in biological systems, where they are
responsible for the organization of living matter as a coherent entity. In
1981 about 2000 titles in the Science Citation Permuterm Index contained
the word hydrophobic (Franks, 1982).

At first, hydrophobic interaction was believed to be due to
attraction between the nonpolar moieties of the interacting species. More
recently, it has been accepted that the attraction between hydrophobic
molecules or moieties is only marginal in the overall hydrophobic effect.
The basis for the effect is believed to be the strong preference of water
molecules in interacting with other water molecules through hydrogen
bonding rather than with hydrophobic sites. This causes a rearrangement
of the water molecules around hydrophobic sites to maximize hydrogen
bonding among themselves. This high energy structure of water is lowered

by the reduction of the hydrophobic area exposed to the water as the

118

hydrophobic sites interact. In parallel, other forces like Van der Waals
forces and charge transfer bonds contribute to maintain the interaction.
The noncooperative migration of molecules of amphiphiles
(molecules bearing both polar and nonpolar moieties) containing long
hydrophobic chains (e.g., fatty acids, alcohols, and amides) to the surface
of aqueous solution at low bulk concentrations demonstrated this
preferential interaction of the water with itself (Tanford, 1980). Under
this condition, the solution surface is sparsely populated and contact
between hydrocarbon chains is negligible. The surface concentration was
found to be proportional to the bulk concentration and this would not be
expected if the amphiphile molecules had a strong attraction for
themselves. Moreover, the ratio between amphiphile surface concentration
and bulk concentration increased three fold for each CH2 group added to
the alkyl chain. Since interaction among the hydrophobic chains are
negligible, this effect of the additional CH2 must reside on the lack of
affinity of water to the hydrocarbon chain. This dominant effect of water
self-attraction in hydrophobic interaction was also demonstrated by
measurement of the free energy of attraction of two liquids (water and
hydrocarbons) that were previously exposed to air and then put in contact
with each other. This free energy of attraction is reported as the surface
tension of pure liquids or interfacial tension of hydrocarbon-water
mixtures. The free energy of attraction for water-hexane and water-
octane systems are about-40 erg/cm2. The same value was found for
attraction of the hydrocarbons to themselves. However, the free energy of
attraction of water for itself is about -144 erg/cm2. The hydrophobic
effect of dilute solutions of salt is essentially the same as that of pure

water.

119

Thermodynamically, hydrophobic interaction is seen as an
entropically driven phenomenon because of the more loose structure
attained by the water molecules that were surrounding the hydrophobic
species before interaction. The introduction of a hydrophobic species to
water causes the breaking of hydrogen bonds between water molecules so
these species can be accommodated in the bulk of the liquid. It is as if a
cavity were created. The water, which originally had a dynamic, relatively
loose structure provided by hydrogen bonds, acquires a more rigid
structure around the hydrophobic species to minimize the loss of hydrogen
bonding. This increase in the organization of the system translates into a
decrease in its entropy. Therefore, association of hydrophobic species is
expected to take place along with an increase in the entropy of the system
since it releases these more structured water molecules to their initially
more loose configuration (Figure 4.1). However, this release is
accomplished at the expense of hydrogen bonds that gives a positive
enthalpy contribution for the change (Creigthon, 1984).

The illustration of the solvation of a hydrophobic species through
the the idea of cavity formation is supported by the linear correlation
between the free energy of transfer of hydrocarbon or amphiphile
molecules of different surface areas from water to pure hydrocarbon
(Creighton, 1984; Tanford, 1980). The larger the area of the molecule
transferred, the larger is the decrease in the free energy of transfer due to
a larger restoration of original water structure for the transfer from
water of the larger molecules.

The cavity formation model with the surface area effect on the
solvation energy led to the investigation of the role of the surface tension

of the aqueous solution on the hydrophobic effect. The increase of surface

 

 

 

120

rApolarIc ~

: f; side clan mwy
e : Hydrophobtc\'g proton
lolNH, area 0"

9W” If oINi-l2 é.
I+l , 3.
NH! °°°°.°. + 2@ _-
Nhfrix °

M 1.:
new + g
pu‘hcle ° 0

Protein“ ,- .°
“W2 2, mm, Mobilized
3 ° 3 “ .‘ wafer
0 ° k Ordered wafer /' ” Hvdmphobioely molecules
° molecules adsorbed .

profe'n
—> INCREASE 0F ENTROPY AS
II 9K -9.7 from q-QIHALPERNJITBI

Figure 4.1 - Schematic representation of hydrophobic interaction.

The
situation refers to binding of a protein to a solid matrix as in hydrophobic
chromatography (from Mohr, 1985).

121

tension of water by salts was found to correlate directly with the
strength of the hydrophobic effect. The molal surface tension increment
of a salt was proposed as a measurement of its ability to affect
hydrophobic interaction and the basis for a natural Iyotropic series

(Melander and Horvath, 1977).

4.3 - Hydrophobic Interaction Chromatography of Proteins
Proteins can experience intramolecular and intermolecular
hydrophobic and hydrophilic interactions since they are amphiphilic
molecules. The side chains of the amino acids that make up the backbone
of proteins can have character ranging from the highly hydrophobic to the
highly ionic. Tryptophan and aspartic acid are good examples of these
extremes. In water soluble proteins, 25 to 30 percent of the amino acid
residues are strongly hydrophobic while 45 to 50 percent are typically
hydrophilic. The hydrophobic residues tend to bury themselves in the
interior of the protein to minimize the energy state of the molecule. In
doing this, they bring with them some hydrophilic groups due to the
primary structure of the protein: the highly hydrophilic peptide bond and
hydrophilic amino acid residues. To compensate for the lost interaction
with water, the internal hydrophilic groups must undergo hydrogen bonding
among themselves. In the case of the peptide bond, this compensation is
especially important because the hydrogen bonding between the hydrogen
of the amino group and the oxygen of the carbonyl group are responsible for
the protein secondary structure (oi-helix and the B-sheet). In contrast to
this burial of hydrophilic residues into the interior of the protein
molecule, there are always some hydrophobic amino acid residues that are

present at the surface of the molecule or in crevices. Forming patches or

 

 

 

122

found in crevices, they are sites for hydrophobic interaction with other
protein molecules or hydrophobic ligands (Creighton, 1984; Tanford, 1980).
Hydrophobic characteristics of proteins have been extensively

exploited as a way to purify these molecules by hydrophobic interaction
chromatography. Besides versatility, this technique is able to produce
high levels of purification, up to 200-fold (Bigelis and Umbarger, 1975).
Relevant literature is abundant. Review articles are available and books
about affinity chromatography always contain a chapter about it. if"

The basic procedure for the separation is shown in Figure 4.1. First,

agarose matrix is derivatized with a hydrophobic chain. The matrix is

 

packed in a column and after equilibrating with a buffer, usually with a
high salt concentration, the protein is introduced into the column.
Adsorption takes place and after washing the column, conditions (e.g., salt
concentration, aditives, pH) are changed to promote protein desorption.
The high salt concentration found in hydrophobic chromatography systems
offers important advantages to the separation in addition to increased
binding. First, the proteins are often stabilized in the solution of the salts
employed and second, bacterial growth is inhibited (Porath, 1987).

The matrices and the procedures for activation and derivatization
are in general the same as used in other affinity chromatographic
technique discussed above. Agarose is generally the matrix of choice.
Activation with cyanogen bromide is the most common procedure to
prepare the matrix for coupling with the spacer or ligand. The cyanogen
bromide reaction with the hydroxyl groups of agarose forms cyclic
imidocarbonate that can react with nucleophiles. Spacers (e.g., NH2-

(CH2)n-COOH, NH2-(CH2)n-NH2) and ligands (e.g.,R-NH2) have amino groups

 

that react with activated agarose forming N-substituted carbamate, N-

123

substituted imidocarbonate, and isourea derivatives. These isourea
functions give an ion-exchange property to the matrix which is deleterious
for some separations (Scouten, 1981).

The attachment of ligands to the spacers, besides the direct
attachment to the activated matrix, can be done by a variety of reactions
as shown in Figure 4.2. There is also a large variety of hydrophobic ligands
available as depicted in Figure 4.3. Additional series of ligands can be
obtained by using other functional groups, placing the functional group at
various position along the chain, or having more than one functional group
on the chain (Shaltiel et al., 1978). The use of linear aliphatic chains is
very common due to the possibility of changing the hydrophobicity by
changing the length of the chain and also due to the benefits of having a
flexible ligand (Figure 4.4). Shaltiel (1978) verifiedthe importance of
length of linear chains during purification of glycogen phosphorylase b
(Figure 4.5). Total retention of the protein was possible only for ligands
with more than 3 carbons in the chain. Moreover, desorption was possible
only from the ligand with 4 atoms of carbon despite the use of drastic
desorption conditions. This "irreversible“ binding has been exploited as a
way to prepare immobilized enzymes for bioconversion in fixed bed
reactors (Caldwell et al., 1975; Hofstee 1976).

Another factor concerning the degree of hydrophobicity of the matrix
is the the extent of derivatization or density of ligands on the matrix.
Jennissen (1976; 1978) demonstrated that the binding of glycogen
phosphorylase b on three different alkylamine-sepharoses (methylamine-,
ethylamine-, and butylamine-sepharose) is a function of the density of

matrix substitution. However, he found that there was a minimal density,

 

 

124

 

 

 

 

 

 

 

wtmhm I W 41 g “(ammonia
l 43 M
I
”131019.000!
I M4,
gear-Acumen Sepheroee ee #- 4MB
'01: (011.). NH; -
a cum
—: i “Ohm
l
I
Malawi. N 2'“, '2
48 4, MGBWICH.
V WW)
: 411(qu
I
(1)
no
‘2’ («Iodide MI
(3)

 

 

% -NHICH.I.M|OOICI‘1,),COOH

Figure 4.2 - Reactions used to couple ligands to agarose (from Pharmacia

Fine Chemicals, 1983).

 

 

 

”Neuter: e! e ”whom ”helm-Q)
mm mm

so”. He":

sou-c: How-as

um: Hafiz-Owe

sou-c. Weeds

are MW

sou-c. MW

”Elam!” of Series
W M
W"- Mic"
aria-m: “Win-*2
sums-coon Mg-oooi
w-q-ou has»; Iron
sou-we MLQ
sore-ems [mu-am:
was” Mire-0'

“Ho-{z Wax-{2

Figure 4.3 - Homologous series of alkyl agaroses which can be prepared for
use in hydrophobic chromatography (from Shaltiel et al., 1978).

 

 

Figure 4.4 - Schematic illustration of how a flexible linear ligand can
accommodate itself in different hydrophobic sites on a protein molecule

(from Shaltiel et al., 1978).

 

 

 

C
8 ,o\
N 0.4 — ’ \
'5 I, I R
e I
3 I

_ I _
.5 O-"lo o\a‘~—_o I
" °'° I I

o 2 4 e e

 

Carbon Atoms/Hydrocarbon Chain

Figure 4.5 - Effect of length of a linear ligand on the adsorption (0) and
desorption (0) of glycogen phosphorylase b on a hydrophobic agarose (from

Shaltiel, 197a).

 

 

 

127

called critical density, under which no binding took place. A study of the
binding of phosphorylase kinase led the author to conclude that the binding
of this enzymes does not follow a simple key-lock mechanism. Instead, it
was found that positive cooperative interaction of the substituted alkyl
ligands is responsible for the adsorption in a multipoint interaction
between the ligand and the proteins.

The role of the spacers is to provide the right steric conditions for
the binding between ligand and ligate. Therefore, some optimization of the
length of the spacer is required because the improper length of spacers can
reduce the binding due to inaccessibility of the binding site to the ligand
or vice-versa (Figure 4.6).

Protein elution, as well as adsorption, depends on the additives to
the solvent in use. As discussed above, the salts usually used for
adsorption are those strongly excluded from the protein and have large
molal surface tension increments. The salts that show little preferential
exclusion from or bind to proteins and have a low or negative molal
surface tension increment are used for desorption. However, the most
common elution strategy is the decrease of the concentration of salt
which promoted binding in first place. Compounds like ethylene glycol,
polyethylene glycol, dimethyl sulfoxide, urea, glycerol, guanidine
hydrochloride, and chaotropic salts are employed when desorption is
difficult (Arakawa and Narhi, 1991; Raymond at al., 1981). pH changes
have also been used as a part of elution strategy (Porath et al., 1973).

Despite the many applications of hydrophobic chromatography, its
detailed understanding is incomplete. In practice, the binding of
hydrophobic species is- never the result of only hydrophobic interaction.

There is always the overlapping of a medley of interactions such as

 

 

 

_A'k-

a b

.,

 

 

 

 

 

Figure 4.6 - Illustration of the role of the length of the spacer in the
binding efficiency of a ligand to a protein. a) spacer is too short; b) spacer
is of right size; c) and c') spacer is too long (from Scouten, 1981).

 

 

129

electrostatic interactions, hydrogen bonds, Van der Waals forces, and
charge transfer bonds. Also, the conformation of the protein, and possibly
of some ligands, can be altered due to binding or changes in the solvent
composition. Therefore, the determination of the character of the binding
is not a simple task (Raibaud et al.,1975). While some authors claim the
achievement of “true“ hydrophobic chromatography (Halperin at al., 1981;
Mevarech at al., 1976), others found that their separation was the result of
combination of hydrophobic and electrostatic interactions (Hofstee, 1973).
The changes in adsorption temperature and elution with ethylene
glycol are examples of how elusive the pursuit of the mechanism of
binding can be (Hjertén, 1973; Hjertén at al., 1974). Since hydrophobic
interaction is a entropy driven phenomenon, a decrease in protein binding
as the temperature is lowered is expected. Breitenbach (1976) found the
opposite when studying the separation of hemoglobins and myoglobins in
alkyl-Sheparoses. When alkyl-Sepharose was replaced by aminohexyl-
Sepharose, binding was stronger at higher temperature. However, the
author attributes the increase in binding to a temperature dependence of
the isoelectric point of the proteins. Ethylene glycol has been used to
determine the contribution of hydrophobic interaction in hydrophobic
chromatography due to its capacity for disrupting native conformation of
proteins and nucleic acids (Nemat-Gorgani and Karimian, 1982). However,
Hofstee and Otillio (1978) desorbed y-globulin from butylamine-Sepharose
and control Sepharose (unsubstituted, inactivated, cyanogen bromide-
treated agarose). The first desorption could have led one to interpret the
adsorption as a hydrophobic based phenomenon; however, the desorption

from the control indicates that this is not the case.

DEV—E

 

130

4.4 - Materials and Methods
4.4.1, - Chemicals

Methylcellulose was obtained from Dow Chemical, Midland, Ml., with
a molecular weight of 85,000 and a degree of substitution in the range of
1.6 to 1.8. Fatty acids were obtained from Aldrich, Milwaukee. IR grade
KBr was purchased from Sigma, St. Louis. Bovine serum albumin (fraction
V, heat shock) was obtained from Boehringer Mannheim Biochemicals,
Indianapolis. ACS grade ammonium sulfate was obtained from Columbus

Chemicals, Columbus. All other chemicals were reagent grade.

4.4.2 - Methods
4.4.2.1 - Esterification of Methylcellulose

A linear series of alkyl carboxylic acids were used to prepare
different derivatives of methycellulose. Propanoic (propionic), butanoic
(butyric), pentanoic (valeric), hexanoic (caproic), octanoic (caprylic),
decanoic (capric), dodecanoic (Iauric), tetradecanoic (myristic), and
hexadecanoic (palmitic) acids which contain 3, 4, 5, 6, 8,10, 12, 14 and 16
atoms of carbon, respectively, were used. in this work, derivatives of the
acids were labeled as Cn where n represents the number of atoms of
carbon of the acid used in their preparation. Unsubstituted methylcellulose
was labelled C1.

Esterification of methylcellulose with fatty acids was performed in
batches of 1 and 20 g of the polymer in glass vials. Each batch consisted
of 8 mL of fatty acids containing 5 to 80 uL of sulfuric acid per gram of
methylcellulose. The fatty acid concentration is at least five times the
stoichiometric requirement to esterify the average number of hydroxyl

groups available in the methylcellulose. The objective was to have an

 

 

131

excess of acid to shift the reaction equilibrium towards the esterification.
The reaction was carried out in an oven at temperatures from 25 to 100 °C.
After reacting from 3 to 216 hours, the mixtures were filtered in sintered
glass filters. The cake was transferred to a beaker and it was washed
with mixing with ethanol (40 mL of ethanol per 1 g of methylcellulose fed)
for 10 minutes. The ethanol washing was repeated twice and, after partial
drying on the filter, the cake was washed three times with ethyl ether
(same proportion as used for ethanol). After suction drying on the filter
the powder obtained was spread in a tray, dried at room temperature for
approximately 12 hours, and weighed.

Large batches of C4, 08, and C12 derivatives were prepared for the
flotation experiments and others that followed. Esterification times were i
4 hours for the C4 and C8 derivatives and 24 hours for the C12 derivative.
The extent of reaction was 0.43, 0.32, and 0.22 for the C4, the 08, and the

C12 derivative, respectively.

4.4.2.2 - Apparent Solubility of the Derivatives

Apparent solubility of derivatives was determined by studying 1
mg/mL solutions. These solutions were prepared according to the general
procedure for dissolving methylcellulose and other cellulose ethers:
hydrating the particles at high temperature and dissolving them at low
temperatures. One hundred milligrams of derivatives was added to 20 mL
of water at 70 to 90 °C under mixing. Twenty milliliters of cold water
was added and after mixing in a ice-water bath for one hour more water
was added to make a final volume of 100 mL. A 1.4 mL sample of the

suspension was taken and the soluble and the insoluble fraction were

 

 

 

132

separated by centrifugation (14,000 rpm, 20 minutes in a Eppendorf

centrifuge model 54150).

4.4.2.3 - Solutions of Derivatives for BSA Precipitation

Individual solutions of the derivatives for BSA precipitation were
prepared by dispersing the derivatives .(10 g for C4 and 08 and 20 g for 012
derivatives) in approximatelly 500 mL of hot water (from 70 to 90 °C)
under strong mixing. The volume was doubled with cold water and mixing
was continued for 48 hours in a ice-water bath. After this period,
centrifugation at 8,000 rpm for 20 min in a Sorvall RC—SB centrifuge

separated a gel cake and a slightly milky solution free of particles that

were used for the precipitation experiments. Concentration of solutions of

the C4 and C8 derivatives was around 8 mg/mL while solutions of 012 was

around 15 mg/mL.

4.4.2.4 - Salting-out of the Derivatives

Salting-out of the soluble fraction of the derivatives with
ammonium sulfate was done by adding different volumes of saturated salt
solution to 1 mL of the soluble fraction. After mixing and 10 minutes of

aging, the supernatant was separated and analyzed for total carbohydrates.

4.4.2.5 - Flotation: System and Runs

The flotation column was made of pyrex glass. Spargers were rigid
cylinders made of sintered glass with pore sizes ranging from 40 to 60
um, diameter of 1 cm, and length of 2 cm fixed at the tip of a glass tube.
The basic design of the column is shown on Figure 4.7. The enlargement at

the top of the column is to avoid particle accumulation on the internal

 

 

 

 

30

 

30

130

10:

300

 

 

 

 

 

  its

Figure 4.7 - Basic design of flotation column. Measurements in
millimeters.

 

134

surface of the column (Oliveira, 1990). To avoid clogging, the enlargement
at the bottom compensates for the reduction of the cross-sectional area
due to the presence of the sparger. The cup-like device around the top of
the column collects the foam that overflows the column (the concentrate
fraction). The feed line consisted of a glass tip immersed into the feed
suspension, plastic tubing, and a peristaltic pump. The tailings line was a
siphon made of plastic tubing connected to the bottom of the column with
a glass tip at the other end. This glass tip and was raised up to a few
centimeters above the top of the column where it was turned downwards.
By lowering and lifting this glass tip, the tailings flow rate and,
consequently, the height of the froth inside the column, could be adjusted
to a desired level. A funnel positioned under the glass tip collected the
tailings into a graduated cylinder. The gas line was a plastic tube
connecting the air reservoir with the fritted sparger.

Salted-out derivatives were prepared for flotation experiments
using 20 percent saturation of ammonium sulfate, salt addition of 40
mL/min, and derivative concentration of 0.3 mg/mL (before salt addition).
Flotation conditions were feed flow rate of 100 mL/min, air flow rate of 5
mL/min and froth heigth of 2 cm.

Batches of BSA precipitates were prepared for flotation
experiments by using 600 mL final volume, pH 6.0, 45 percent saturation of
ammonium sulfate, salt solution added at the rates of 10 and 50 mL/min
(for Cf and C4, respectively), Cn/BSA (mass of derivative per mass of
protein) equal to 2 for a mass of protein of 397.2 mg, and aging of 1 hour.
Flotation conditions were feed flow rate of 20 mL/min, air flow rate of 5

and 16 mL/min (for C1 and C4, respectively)

 

 

 

135

Flotation experiments, done in triplicates, were started by flowing
air and pumping suspension into the column at desired rates until it
overflowed. The tip of the siphon for tailings withdrawal was lowered
stepwise to set the froth-suspension interface to the desired level. Once
air flow rate and froth-suspension interface level were set to the desired
values, the column was operated in a pre-run mode for a time equivalent to
the pumping of at least three recovery zone volumes of suspension-
washing solution at the current feed flow rates. The corresponding
concentrate and tailings collected were discarded. The actual run was

initiated with time recording and concentrate and tailings collection.

 

Concentrate was collected in a beaker and a graduated cylinder was used
to collect the tailings. Tailings and concentrate had particles and liquid
separated by filtration in a sintered filter of 30 mL ASTM 40-60 C. The
particles were dissolved in 0.01 M sodium acetate buffer pH 6.0 and
together with the liquid from concentrate and tailings they were analyzed

for protein or total carbohydrates.

4.4.2.6 - De-fatting of BSA
To assure that the BSA was free from fatty acids, the protein was
de-fatted. De-fatting of the BSA was done according to the procedure

developed by Chen (1967). Five grams of albumin were dissolved in 50 mL

 

of water and 2.5 g of activated charcoal were added. The pH of the
suspension was adjusted to 3.0 with 4 N HCl. The suspension was kept in
ice-water for 1 hour under mixing. The particles were separated by

centrifugation (12,000 rpm for 20 minutes in a Sorvall RC-SB centrifuge).

 

The supernatant pH was adjusted to 6.0 with sodium hydroxide and sodium

acetate was added to a final concentration of 0.01 M.

136

4.4.2.7 - precipitation of BSA with Salt

Precipitation of bovine serum albumin at 60 percent saturation of
ammonium sulfate was done at different pH values. A solution of bovine
serum albumin was prepared by diluting 402 pl of a stock solution of the
protein (82.75 mg/mL) in 20 mL of 0.01M sodium acetate buffer. Thirty
milliliters of saturated solution of ammonium sulfate was added to the
protein solution at the rate of 1 mL/min through a syringe pump (Sage
Instruments model 355). The pH of the mixture was set around 9 (Orion pH
meter model 501) with sodium hydroxide and, under mixing, a sample of
approximately 1 mL was taken. The pH was lowered approximately 0.2 pH
units with sulfuric acid and the sampling was repeated. This procedure of
pH lowering and sampling was repeated until pH 3 was reached. After 90
minutes of aging, the samples were centrifuged in a Eppendorf centrifuge
at 14,000 rpm for 10 minutes and the supernatant collected. Absorbance
at 280 nm was used to evaluate the amount of protein precipitated. in a

similar way, a control was done with no ammonium sulfate.

4.4.2.8 - BSA Preclpltation with Derivatives at Different pH

The study of BSA precipitation in the presence of derivatives was
done at 60 percent saturation of ammonium sulfate and different pH
values. Final batch volumes were 50 mL due to the addition of 30 mL of
saturated ammonium sulfate solution at pH 9 to 20 mL of the derivative-
protein solution. The salt addition was done by a syringe pump at the rate
of 1 mL/min. The Cn/BSA ratio was kept constant at a value of 2 for a
mass of protein of 33.1 mg. Therefore, due to the different concentrations

of the derivative solutions, different volumes were used. The derivative-

 

 

137

protein mixture volume was increased to 20 mL by addition of 0.01 M
sodium acetate at pH 9.0. A 5 mL sample of the suspension was taken. The
pH was lowered approximately 1 pH unit and another sample was taken.
The procedure of pH lowering and sampling was repeated 3 more times
until a pH around 5.5 was reached. After two hours of aging, the samples
were filtrated in 30 mL sintered glass filter (ASTM 10-15 C) and filtrate

absorbances at 280 nm were taken.

4.4.2.9 - Time Course of Precipitation of BSA with C4 Derivative
BSA batch precipitation with a C4 derivative at a Cn/BSA ratio of
1.5 was achieved as described in section 4.4.2.8 above with the exception
of addition of the salt solution at a much faster rate (30 mL in 25 seconds
or 72 ml/min). Samples of 5 ml were taken at different times after the
addition of the salt solution, filtered in a sintered glass filter (ASTM 10-
15 C), and the absorbance at 280 nm was recorded. The Cn/BSA ratio was
changed from 2 to 1.5 to avoid extensive precipitation of the protein. The
objective was to avoid low absorbance of the filtrate at extensive

precipitation which can be significantly affected by background.

4.4.2.10 - BSA Precipitation at Different Salt Concentration
Test tubes batches of 2 mL of C4 derivative-BSA mixtures were
prepared using the same proportions as before (section 4.4.2.8), but with a
Cn/BSA ratio equal to 2. During mixing in a vortex, a saturated solution of
ammonium sulfate was added to achieve different salt concentrations.
After one hour of aging, the suspensions were filtered in a sintered glass

filter (ASTM 10-15 C) and filtrate absorbances at 280 nm were recorded.

 

 

138

4.4.2.11 - Total Carbohydrates Assay

Total carbohydrates (TCH) were determined by a method described by
Dubois and co-workers (1956). Methylcellulose can be determined by this
method because polysaccharides and their derivatives, including the
methyl ethers with free or potentially free reducing groups, give a orange-
yellow color when treated with phenol and sulfuric acid. The procedure
consists in preparing a reaction mixture with 0.2 mL of sample and 0.2 mL
of 5 percent phenol solution in 13x100 mm glass test tubes. One milliliter
of concentrated sulfuric acid is added to the surface of the reaction
mixture as fast as possible to generate a high temperature during the color
development period (at least 20 min). The tubes were manually shaken
and, after 10 min the contents mixed with a vortex. Absorbances at 488
were taken after at least 10 minutes. A blank made with water in place of

the sample was used as reference. Glucose was used as the standard.

4.4.2.12 - Protein Assay

Protein concentration was determined by two methods: by
absorbance at 280 nm and by a dye-binding assay from Bio-Rad, Richmond,
CA. Bovine serum albumin was used as the standard. Absorbance readings
were recorded with a Perkin-Elmer absorption spectrometer, model

Lambda 3A.

4.4.2.13 - infra-red Spectra

infra-red spectra were taken using the KBr technique. A few
milligrams of the sample were ground with approximately 100 mg of KBr
in a mortar. Pellets were made by pressing the mixture with a hand-press.

Liquid samples were dried directly in the mortar (100 to 200 uL) at

 

 

 

139

approximately 90 °C by a heat gun. Spectra were taken using a Perkin-
Elmer Infra-Red Fourier Transform spectrometer, model 1750 interfaced

to a Perkin-Elmer 7700 Professional Computer.

4.5 - Results
4.5.1 - Esterification of Methylcellulose and Characterization of
the Derivatives

The infra-red spectra of both carboxilic acids and esters have a
characteristic band due to the carbonyl stretch. The carbonyl stretch of
acids appears between 1700 to 1725 cm-1 while that of esters occurs
around 1740 cm-1. Esterification of a given carboxilic acid can be
verified by the appearance of the shifted carbonyl stretch of the ester.
Likewise, spectral phenomena can be utilized to detect the esterification
of methylcellulose (Morrison and Boyd, 1973).

The first verification of methylcellulose esterification was
accomplished with octanoic acid derivatives. The spectrum of octanoic
acid showed a sharp peak at 1710 cm-1- The methylcellulose spectrum did
not show a peak in this region. However, the derivative of this acid
showed a sharp peak at 1740 cm-1. Spectra obtained immediately after
mixing methylcellulose with this acid at room temperature and no catalyst
present showed a doublet with peak maxima ocurring at 1710 and 1740
cm-1. The peak at 1710 cm-1 corresponds to the free acid. The other
peak, at 1740 cm-1, corresponds to some esterification during the
preparation of the pellet, probably because of high pressure used in the
process.

An arbitrary quantity to follow the progress of the esterification of

methylcellulose was developed. it was termed “Extent of Reaction" and it

.'r—*

wan-in.- min-ti.
I

I-

‘i

 

 

 

140

was defined as the ratio between the transmittance measurements at the
characteristic maxima of the ester carbonyl group and the methylene
group, at 1740 and 2930 cm-1, respectively (Pradeau et al., 1982). Figure
4.8 shows a typical spectrum of the derivatives.

A characteristic of esterification is the effect of inorganic acids
(e.g., sulfuric or hydrochloric acids) as catalyst. The effect of sulfuric
acid as a catalyst on the esterification of methylcellulose was checked by
running two series of reaction mixtures, one containing the catalyst and
other without it (Figure 4.9). The catalytic effect of the sulfuric acid is
easely recognized by the higher extent of reaction of the reaction mixtures
which contained it.

A short optimization of the esterification was done with regard to
two variables: temperature and amount of sulfuric acid (Figure 4.10). The
use of 80 uL of sulfuric acid per gram of methylcellulose resulted in total
degradation of the methylcellulose at all temperatures studied. Also the
use of 100 °C in a reaction system with 20 uL of acid per gram of
methylcellulose caused the total degradation of the polymer. As it is
showed in Figure 4.10, a higher volume of sulfuric acid resulted in a faster
reaction. High temperatures increased the reaction rate; however, the
combination of high temperature and high amount of sulfuric acid led to
degradation of the methylcellulose.

In preparation for the production of derivatives of the different
fatty acids at different degrees of esterification, an evaluation of their
reactivities regarding the esterification was done. This evaluation

consisted in determining the extent of reaction of derivatives produced

 

 

 

141

 

 

 

 

54 l l l l l 1 4L # l L l l l l l A l l I 4
' Extent of _ Peak hei ht at 1740 cm-1
‘ Reaction (E) - Peak height at 2930 cm-1 -
47 -
g " r-
v .o. l
O
O ' .
C
‘3 ..
E .
E - -
U)
C
9 33 - ~
P I
C=O '
d .1 -
C H2 (1740 cm )
.1
25 . (2930 cm ) _
19 I I I I l I I I I I I I I I I I I I I fl
3800 3000 2000 1600 1200 300

Wavenumber (cm-1)

Figure 4.8 - Typical spectrum of methylcellulose derivative. Fatty acid,
butyric acid; volume of sulfuric acid, 5 tiL/g of methylcellulose;
temperature, 65 °C; reaction time, 15 hours.

 

 

142

 

1,0 - 1 - t a t - 1

    

Catalyst Present

0.8 . ‘
t:
.2
H
8
m 0.6 " .,
a:
h b Ii
0
E 0.4 - -
*3 Control .
LLI

0.2

 

 

 

0 10 20 3O 40

Time (hours)

Figure 4.9 - Effect of sulfuric acid as a catalyst on the esterification of
methylcellulose using octanoic acid; catalyst 20 uL/g of methylcellulose;
temperature, 65 °C.

 

 

143

 

 

 

 

100 -c

1.0 ~
5 03 -
8 70 ‘C
G
O
i: 0.6 -
.6 50 C
E 0.4 r
X
in

02 - 25 -c ‘ L ‘

0. - ‘

"ii 5 10
Time (hours) i

5 tiL of sulfuric acid I 1 g of methylcellulose

 

 

 

 

1.0
70 -c

C

0.8 -
g 50 -c
O
8
c: 0.6 -
‘6
E o -
o .4
3'2
u:

0.2 -

as 'c
0.‘ W
(l 5 10
Time (hours)

20 uL of sulfuric acid / 1 g of methylcellulose

Figure 4.10 - Effect of temperature and amount of sulfuric acid on the
esterification of methylcellulose using octanoic acid.

144

under the same conditions (Figure 4.11). The longer the linear chain, the
lower was the esterification of the methylcellulose. Based on these
results, an experiment was designed to produce derivatives of the
different fatty acids at different extents of reaction (Figure 4.12). The
faster reaction rate of the smaller acids was again clearly evident in
these experimental results. Evaluation of degradation of the
methylcellulose by the reaction conditions was done by determining the
yield (mass of derivative after washing and drying). instead of a weight
gain, expected if esterification was the only reaction to take place, a
weight loss took place in almost all preparations. The smaller fatty acids
showed a much larger weight loss than the longer ones (Figure 4.13).

Characterization of the derivatives was based on their solubility in
water, salting-out with ammonium sulfate, and floatability. The
importance in determining derivative solubility is the desire to use a
carrier-ligand system soluble in water. Faster kinetics are expected when
liquid phases are used for ligand-ligate binding. Note that methylcellulose
is very soluble in water with concentrations as high as 100 mg/mL having
been reported (Dow Chemical, 1988). However, the attachment of a
hydrophobic chain should decrease this solubility. This was confirmed by
results of the measurement of the apparent solubility of the derivatives
previously made (Figure 4.14). Since the maximum concentration
achievable was 1 mg/mL, the apparent solubility is also a measurement of
the fractional soluble mass. An inverse dependence of the apparent
solubility on the extent of reaction was noted.

Figure 4.14 was divided into two plots as shown in Figure 4.15.
Three groups of derivatives with close apparent solubility dependencies on

the extent of reaction are evident. Derivatives of C3 and C4 form one

 

145

 

 

 

 

i O t
1.5“ ‘
c b O O 0 j
.9
"" O
3 ° 8
2 1.0: '
“5 t 6
o—a L 6 ,
E
g .
UJ 0'5. e .
i . 9 o
0.0 A J#J . 1 4 L . r . 4 l j)
0 5 1O 15

Number of carbons in the linear chain

Figure 4.11 - Reactivity of different fatty acids with respect to the

esterification of methylcellulose.

Volume of sulfuric acid, 5 uL/g of

methylcellulose; temperature, 65 °C; reaction time 24 hours.

 

 

146

 

 

Extent of reaction

 

 

 

 

0.0" A l n l A l n I
0 20 40 60 80

Reaction time (hours)

Figure 4.12 - Kinetics of esterification of methylcellulose by fatty acids.
Volume of sulfuric acid. 5 ng of methylcellulose; temperature, 65 °C.

 

 

147

 

1

Yield (9)

 

 

 

 

 

 

- . a t - I —,
1.1 . v

 

Yield (9)

 

+ C12 cl
_O—' C14

0.7 *'

 

L - l A

0 A 20 40 60 80 100

 

 

0.6

Time (hours)

Figure 4.13 - Hydrolysis of methylcellulose during esterification measured
as decrease in yield. Volume of sulfuric acid, 5 ung of methylcellulose;
temperature, 65 °C.

 

 

148

 

 

 

 

 

 

1. P -l
A 00 o D A A
Te‘ i ‘3 ° ‘
E» 080 - ~ 1
°
V n A e 0 C3 1...:
> _ O 4 A C4
E 060 9 a ‘3 °C, 0A 0 cs ‘
g a A ' 08
3 ‘fi A . A C10
0) 0.40 ’ . A o o I C12
E ‘ . % AA A a C14
93 W o
a 0.20 - It 5 D ‘
5% ' n‘ ‘3: ‘
000 - l L L Em . I - 1
' 0.0 0.2 0.4 0.6 0.8 1.0 1.2

Extent of reaction

Figure 4.14 - Apparent solubility of the derivatives (03 to 014).
Derivatives were treated according to the procedure to dissolve
methylcellulose as for the preparation of a 1 mgImL solution. The
concentration of derivative in the liquid phase was determined and
was called apparent solubility.

16

 

 

 

 

 

 

A 1.0. 0 C3 .

e . to c.

\ \ \ o —-—-n cs

a \ non-Ice...

E 0.8~ - \\ 0 cs

V \

>~ .

g

:5 05'

2

O

a: 0.4- o <

a

C

O

a .' <

3.02 -n . 1-

'< .k’nn ‘ 3’
00 ‘ ‘ ‘ f

 

02 0A 05 03 1D 12

Extent of Reaction

 

 

 

 

1.0
3‘ ? o a
E 0.8- o 010
CE” A C12
V D D C14
3?; 0.6 ' \d ‘
g
8 0.4 - 13 1
‘5
co '3. .
587' 0.2 - :0
° iii ‘
< m
°'%.o o 2 0.4 0.6 0.8

Extent of Reaction

Figure 4.15 - Apparent solubility of the derivatives (C3 to CB and C8 to
C14). Derivatives were treated according to the procedure to dissolve
methylcellulose as for the preparation of a 1 mg/mL solution. The
concentration of derivative in the liquid phase was determined and
was called apparent solubility.

150

group, 06 and C8 form a second one, and 010, C12, and C14 form a third
one. instead of studying the derivatives of the seven fatty acids, one
member from each one of the three groups was chosen for the subsequent
experiments. The objective was to reduce the number of experiments by
studying derivatives that are representatives of the entire range of
derivatives. C4, C8, and C12 derivatives were used in the subsequent
experiments.

Due to the objective of using a soluble derivative for the protein
binding applications, the extent of reaction of the bulk derivatives and
soluble and insoluble fractions were compared (Figure 4.16). The
existence of a limiting extent of reaction for solubility was evident. The
longer the fatty acid chain, the lower is this limiting extent of reaction.

Besides the knowledge on the solubility of the derivatives and their
degree of esterification, another important solubility related behavior is
their insolubilization with salts. Therefore, the soluble fraction of the
C4, C8, and 012 derivatives were salted-out with ammonium sulfate
parallel to salting-out of unsubstituted methylcellulose (Figure 4.17). The
three derivatives have salting-out properties very close to each other.
However, salting-out started at lower salt concentration. than the
unsubstituted methylcellulose although the minimum concentration of
ammonium sulfate necessary to salt-out virtually all material is
approximately the same (around 20 percent saturation).

As the. last characterization of the derivatives, their floatability
was compared to the floatability of the unsubstituted methylcellulose
under similar conditions (Table 4.1). C4 and 08 derivatives showed the
same level of recoveries by flotation as the unsubstituted methylcellulose

while 012 derivative showed a much lower recovery.

 

 

 

151

 

1.0- A

lb.
0
00

0.8 ' 7

_ a?
o

0.6
0.4 - ‘ -

0.2 - 4° -
[II

Extent of reaction of
soluble or insoluble fraction

 

 

 

0'%.o 0.2 0.4 0.6 0.8 1.0 1.2

Extent of reaction of bulk derivative

Figure 4.16 - Differences in the extent of reaction of soluble and insoluble
fractions of different derivatives. Soluble and insoluble fractions of the
derivatives resulted from experiment showed in Figures 4.14 and 4.15
were dried and their extent of reaction measured. Open symbols, soluble
fraction; closed symbols, insoluble fraction.

 

 

 

152

 

 

 

 

 

:3 A 0C4
2° 408
OE 9 0012
mm 80" ‘le .
£70 0 °
OE r
C
(5:: 60l- ..
.g: r ‘
D)
Em 40l- .
E
20>
-—0
‘55 20)- .,
.20-
50) . f .
mg \a:&
04 41 4 14 a4axaa
0 5 10 15 20

Ammonium sulfate concentration
(as percentage of saturation)

Figure 4.17 - Salting-out of derivatives and unsubstituted methylcellulose.

Soluble fractions of the derivatives resulting from experiment showed in
Figures 4.14 and 4. 15 were salted-out with ammonium sulfate. The
concentration given by the remaining amount of derivatives in the solution
was measured and compared with the original concentration

 

 

 

153

Table 4.1 - Comparison among floatability of the derivatives and

unsubstituted methylcellulose (C1).

 

 

DERIVATIVE X s 11
C1 66.6 1.3 3
C4 70.0 3.6 3
08 62.0 4.5 3
C12 18.1 1.2 3

 

X = average recovery; 3 = standard deviation; n = number of treatments

Conditions - Batch partition: derivative concentration before salt addition,
0.3 mg/mL; salt concentration, 20 percent saturation. Continuous
flotation: feed flow rate, 100 mL/min; air flow rate, 5 mL/min

 

 

 

154

4.5.2 - Precipitation of BSA with Methylcellulose Derivatives
C4, 08, and 012 derivatives were tested for the precipitation of BSA
at 60 percent saturation of ammonium sulfate. First, due to the high salt
concentration, control of precipitation of ISA at different pH with and
without salt was necessary (Figure 4.18). BSA showed an extensive

precipitation at pH values lower than 4.7 . Therefore a lower limit for the

 

precipitation in the presence of the derivatives was set at pH 5.5. A

 

higher limit was set at pH 9 due to the strong buffer capacity of the
suspension at high pH values.

A modest enhancement of BSA precipitation was realized for all
derivatives except great enhancements of the C4 derivative (Figure 4.19).
Virtually all BSA was precipitated when 04 derivative was present. For
all three derivatives and unsubstituted methylcellulose, little protein
precipitation was noted as the pH was lowered from around 9 to around
5.5. Due to the superior capability of the C4 derivative in the precipitation
of BSA, it was the only one selected to be used for the experiments in
which aging time, salt and protein concentration, temperature, and

floatability were studied.

 

The kinetics of the precipitation were estimated by following the
amount of protein precipitated out of solution with time (Figure 4.20). The
precipitation seemed to take place in a short time period relative to the
experimental time. Based on this finding, a one hour aging time was used
for the subsequent experiments.

Changes in salt concentration strongly affected the precipitation of
the BSA with the C4 derivative (Figures 4.21 and 4.22). The higher the salt
concentration, the more complete was the protein precipitation. The

concentration of the C4 derivative had a similar effect on the BSA

155

 

 

 

 

 

Eo_4o-..,...,.--,f.1
C O A A0 1
o _— — — — —§o —o o— —-
a l E a 8A 6A3 $2 0A ‘
.. 0.30- o -
CU ’ .
0) 00
0 i
c I 1
(U
.9 0.20- .
o ’ o d
(D .
.Q 0 .
<0 * o 1'
E . -
S 0310’ o ‘
N
C
I— , .
3- ' 3° Q69 .
a 0.00 1D. . . L4 . . l . x .
.0 4.0 6.0 8.0 10.0

 

Figure 4.18 - pH precipitation curves for BSA. 0 - 60 percent saturation of
ammonium sulfate; A - no ammonium sulfate. Protein concentration, 0.66
mgImL; aging time, 2 hours. The dashed line is the expected absorbance in
the case of no precipitation. ‘

156

 

 

 

 

E . .
c 0.30 - .l
O . ,
(D
N b 1
T6 ' I
m b o o o I
o 0020 I. o o D D -
c ' A E A A -
It! A El
.0 l:
8 r 1
8 . .
cd 0.10 ' 1
.9 ‘ i
<0
.2 .
:- _ 0
LL 0

0.00 - ' ' a L . ' - .

5.0 6.0 7.0 8.0 9.0

Suspension pH

Figure 4.19 - Precipitation curves for BSA in the presence of derivatives
(0, 04; A, 08;l:, 012) and unsubstituted methylcellulose (O). CnIBSA = 2;
protein concentration, 0.66 lmglmL; salt concentration, 60 percent
saturation; aging time, 2 hours. The dashed line is the expected
absorbance in the case of no precipitation.

 

 

 

 

 

 

 

 

 

g . .
o 0.30 ‘
a) 1
N P d
13' _ I
8 0.20 - . -
C l . (
m u
.0
I—
O .
"’ ‘ i
‘8 0 10 1
8 I? § 0 Q) 0 00
(B
:5 .
E . l
0.00 4 l 4 l 4 l a l
0 1 0' 20 30 40

Time course of precipitation (minutes)

Figure 4.20 - Time course for precipitation of BSA with 04 derivative.
Time recording started after addition of the‘ salt solution. CnIBSA = 1.5
(decreased from 2 to avoid supernatants with low absorbance at 280 nm);
protein concentration, 0.66 mgImL; salt concentration, 60 percent
saturation; pH, 6.0. The dashed line is the expected absorbance in the case
of no precipitation.

 

 

 

 

 

 

E 100 e . T .
3
E
c
0
=2 80 o 8 ~
2 to
CE » -
”’5 . g .
.E m 60" o "
m: . o .
.2: . o t
.E 0 ’ o 1
c0 01 40' o
E O) P 8 1
9: ~ 1
g 8 20. 0'
h
an a i i
. o‘
m b u
m 0 n L A J n L
V 30 40 50 60

Ammonium sulfate concentration
(as percentage of saturation)

Figure 4.21 - Effect of ammonium sulfate concentration on the
precipitation of BSA with C4 derivative. CnIBSA = 2; protein
concentration, 0.66 mgImL; pH, 6.0; aging time, 1 hour.

 

8

 

 

 

 

l
J A l A I n I 1 J A I

O a
0.0 1.0 2.0 3.0 4.0 5.0 6.0

C4IBSA
(mass of C4 derivative per mass of BSA)

BSA prec1pltated
(as percentage of the initial amount)

Figure 4.22 - Effect of the ratio C4IBSA on the precipitation of BSA with
C4 derivative at different ammonium sulfate concentrations (0, 60 percent
saturaton; A, 55 percent saturation; D , 50 percent saturation; at, 45 percent
saturation). Protein concentration, 0.66 mglmL; pH, 6.0; aging time, 1
houn

 

 

 

1m

precipitation. However, there was less dependence of the precipitation on
the derivative concentration as the salt concentration was lowered from
60 to 45 percent saturation.

The last precipitation variable studied was aging temperature
wherein precipitated batches were aged at three temperatures (Table 4.2).
The results showed an increase in precipitation with an increase in
temperature. At the same time, virtually no derivative was detected in I
the filtrate (data not shown). d

Flotation experiments (Table 4.3) first targeted the effect of the

 

BSA on the flotation recovery of unsubstituted methylcellulose. The l H
results showed a reduction in the recovery from 99 to 25 percent when #1
BSA was added to the system. Subsequently, it was estimated whether I

total protein recovery would correspond to the amount predicted by

precipitation with the derivative and flotation recovery combined. BSA

precipitation under this conditions is 35 percent according data from

Figure 4.22. The flotation experiments showed a recovery of derivative

particles in the concentrate of 27 percent (as percentage of total mass of

derivative fed). The recovery of protein associated with these particles

was 13 percent as percentage of total protein fed (analyzed by the dye

method). This 27 percent recovery of derivative by flotation and the 35

percent recovery of protein by precipitation combine to predict a protein

recovery by the combination of precipitation and flotation of 9 percent.

The experimental value found was slightly higher, 13 percent.

4.6 - Discussion
Esterification as a route to derivatize methylcellulose with linear

chains of carboxylic acids proved feasible. As typical in esterification,

 

 

161

Table 4.2 - Effect of temperature on the precipitation of BSA with C4

denvaflve

 

Temperature (°C) Filtrate Absorbance at 280 nm'

 

0 0.292
24 0.244
55 0.169

 

'Average of three treatments

Cn/BSA, 1 (decreased from 2 to avoid filtrate with low absorbance at 280
nm); protein concentration, 0.66 mglmL; pH, 6.0; salt concentration, 50
percent saturation; aging, 1 hour; expected absorbance at 280 nm for no
precipitation, 0.338.

 

162

Table 4.3 - Flotation 0f the BSA-derivative precipitates'

 

 

 

Precipitation Flotation Recovery (1)
Run Recovery“) Derivative(2) BSA X(2)
C4 + BSA 35 27 13 ‘ 9
------------------------------------------------------------------- ed.
01 - - 99 - - ‘ '
C1 + BSA - - 25 - - - -

 

' Average of three treatments

Precipitation Recovery (1) is the expected percentage of protein
precipitated at the experimental conditions given below and acccording to
Figure 4.22.

Flotation Recovery Derivative(2) gives the derivative particles recovered
by flotation as a percentage of total initial amount.

(1)x(2) is the predicted recovery of BSA as a result of a combination of
precipitation and flotation.

Conditions:

Precipitation: Cn/BSA = 2; protein concentration, 0.66 mglmL; 45 percent
saturation of ammonium sulfate; pH = 6. Flotation: feed flow rate, 20
mL/min; air flow rate, 5 and 16 mL/min (01 and 04 derivative,
respectively).

 

163

the catalytic effect of an inorganic acid (sulfuric acid) was observed. The
presence of degradation of the methylcellulose by parallel reactions was
also evident. However, mild conditions provided sufficient kinetics and
yield.

The dimensionless parameter extent of reaction proved to be a
Simple and useful tool in following the progress of the reaction. A weak
point of the method is that the reference peak for the measurement of the
carbonyl stretch peak (stretch of the methylene groups) also increases as
the reaction progresses: the methylene groups are the majority of the
components of the added chain to the methylcellulose. However, the effect
on the extent of reaction seemed to be minimal probably due to the
originally relatively high degree of substitution of the methylcellulose
(1.6 to 1.8 where the possible maximum is 3.0) and, consequently, the low
esterification levels achieved. Therefore, the methylene groups introduced
may have been significantly less in number than the ones present in the
structure of the unsubstituted methylcellulose.

The different reactivity of the acids (the Shorter chain acids being
more reactive in terms of producing derivatives with higher extent of
reaction and giving lower yield due to degradation) was expected since the
shorter chain acids tend to react more vigorously than the longer bulkier
ones. Large groups attached to the carboxylic groups are known to create
steric problems and lower the reactivity of organic acids in esterification
(Morrison and Boyd, 1973).

The low apparent solubility of the derivatives was the result of
increasing of the hydrophobicity of the methylcellulose by attaching
hydrophobic groups to it. The spreading of the apparent solubility curves

(Figure 4.15) should be analyzed in terms of increased hydrophobicity of

 

 

164

the derivative molecule. The reaction of fatty acids of different sizes
with methylcellulose produce different derivatives. These derivatives
differ in the hydrophobicity due their different number of carbons in the
chain added. This is noted by the position of the curves which is, from top
to bottom, 03, 04, C6, and 08. However, as the ligand increases in size,
hydrophobicity added by a few more carbons in the chain is not as
significant as when the derivative was much shorter. As a consequence,
the difference in hydrophobicity between the derivatives of same extent of
reaction is negligible and the apparent solubility curves overlap as in the

case of 010, C12, and C14 derivatives. The limiting extent of reaction of

 

the soluble fractions of the different derivatives is another example of the
lower hydrophobicity of the Shorter chains. The Shorter the fatty acid I
used, the higher the extent of reaction of the soluble fraction. The values

of these extent of reaction limits (approximately 0.60, 0.25, 0.10 for C4,
08, and 08 derivative, respectively) are Spaced in the same pattern of the
apparent precipitation curves.

Another possible factor in the differing behavior of the derivatives
is the modification of the methylcellulose caused by reactions other than
esterification. One of these reactions is the acid hydrolysis of the
methylcellulose. The different yields of derivatives obtained with the
different fatty acids indicates different degrees of hydrolysis. This
results in derivatives of different molecular weight distributions that
may influence their solubility and binding characteristics.

The precipitation of BSA with the derivatives, especially the C4

derivative, does not seem to be a simple phenomenon. The BSA removal

 

from solution only occurs at salt concentrations higher than the minimum

concentration for the salting-out of the derivatives. it seems that the

165

protein adsorbs to the salted-out derivative. However, the Cn/BSA ratios
used are high enough to discard the concept of adsorption in this case.

The better effectiveness of the 04 derivative may not be due only to
the right size of ligand, but it may also be due to the large extent of
reaction achieved with butyric acid. This would provide not only the
conditions for the right solvation environment, but also a density of
substitution larger than the critical density necessary for the binding to
occur as previously discussed (Jennissen, 1976 and .1978).

The separation required a high salt concentration to significantly
precipitate BSA contrary to most reports in the literature where binding
to chromatographic matrices is achieved at low ionic strength (Hofstee
and Otillio, 1978; Nemati-Gogani and Karimian, 1982; Wichman and
Anderson, 1974; Peters et al., 1973). However, _Pahlman and co-workers
(1977) reported the need of high salt concentration for the binding of the
protein to pentyl-Sepharose gel (ionic strength as high as 3 M). These
authors correlated the decrease in the binding with changes in the
conformation of the BSA when different salts were used in the protein
solution. Therefore, it is possible that the BSA binds to the C4 derivative
at low ionic strength, but the increase in salt concentration to precipitate
the derivative disrupts the binding due to conformational changes in the
protein molecule. Parallel to that, we must consider also changes in the
conformation of the derivative. At low ionic strength, the hydrophobic
chains may be available for binding. As the salt concentration increases,
not only does the protein conformation change, but as the derivative iS
precipitated, its conformation may also change. The hydrophobic n-propyl

chain of the reacted butyric acid may preferentially interact among

 

 

 

 

166

themselves. Only at high salt concentration can the conditions be reversed
enough for binding to occur.

Despite the murky picture of the overall phenomenon, there is
evidence that hydrophobic interaction plays a role in the separation. First,
the precipitation is not very sensitive to pH in the range studied indicating
that electrostatic forces do not play a major part in the precipitation.
Second, the dependence of the precipitation on the ammonium sulfate
concentration is typical of hydrophobic based phenomena. Third, the
temperature experiments showed an increased precipitation as the
temperature was raised, a behavior related to the principle that
hydrophobic interactions are entropically driven. The complete
insolubilization of the C4 derivative at 60 percent saturation of
ammonium sulfate ruled out the possibility that the increase in BSA
precipitation is due to a higher insolubilization of the derivative at higher
temperatures.

The flotation results Showed that the derivatization does not affect
floatability Significantly, expect for the C12 derivative. The large
reaction time for the production of this derivative (24 hours compared to 4
hours for the others) and, consequently, the long exposure to sulfuric acid
may be the reason for this different behavior.

The floatability of the BSA precipitated with C4 derivative was not
high. The presence of the BSA in a high concentration, as it was seen 'in
the flotation of the unsubstituted methylcellulose, certainly contributed
to the expected low floatability due to the high salt concentration. As it
was seen in Chapter 3, the floatability of a very Similar polymer,
(hydroxypropyl)methylcellulose, iS strongly affected by the increase of

ammonium sulfate from more than 40 percent saturation.

 

 

 

167

However, the recovery of BSA with the particles floated was, within
a margin of error, the percentage predicted by the combined recoveries of
precipitation and flotation. This result demonstrated the feasibility of
the system developed. A known amount of protein was precipitated with a
chemically engineered molecule. The flotation step did not affect the

distribution of the BSA between solid and liquid phases. As long as the

 

particle recovery by flotation is known, the combined recovery of
precipitation and flotation can be predicted. The prediction of the
flotation step is complicated in this case due to the high concentration of
protein. However, its dependence on bold changes in the system is a
characteristic of the flotation system.

In summary, the development of the proposed flotation system to
perform protein separation by hydrophobic interaction between a ligand
and a ligate was accomplished although the protein chosen required

conditions that required the system to be run at suboptimal conditions.

4.7 - Conclusions

Besides the overall conclusion that the design and production of a
flotation system to perform protein separation by flotation based on
protein-ligand hydrophobic interaction is possible there are other
conclusions from this study. They are as follows:

a) Direct esterification of methylcellulose by fatty acids can be
performed at mild temperature (around 65 °C). Sulfuric acid is an

efficient catalyst for this reaction.

 

168

b) The longer is the fatty acid, the Slower is the reaction with the

methylcellulose.

c) The longer the fatty acid, the smaller is the soluble fraction of

its derivative fora fixed extent of reaction.

d) the longer the fatty acids, the lower is the extent of reaction of

the soluble fraction of its derivative.

e) The derivatives are salted-out from aqueous solutions at lower

ammonium sulfate concentration than the unsubstituted methylcellulose.

f) The floatability of derivatives of butyric and octanoic acids are
similar to the unsubstituted methylcellulose while the dodecanoic acid has

a much lower floatability.

g) The butyric acid derivative promotes extensive precipitation of

the bovine serum albumin at 60 percent saturation of ammonium sulfate.

h) pH does not significantly affect the precipitation of bovine serum

albumin by the derivatives used.

i) The precipitation with C4 derivative of bovine serum albumin
takes place in less than one hour.
j) The precipitation of bovine serum albumin with the butyric acid

derivative is dependent on the salt concentration and on the ratio mass of
derivative to mass of protein.

k) The presence of bovine serum albumin iS deleterious to the
floatability of the unsubstituted methylcellulose.

 

 

 

1w

' l) The interaction that causes the precipitation of bovine serum
albumin seems to be at least partially of a hydrophobic nature.

4.8 - List of References

Arakawa, T.; Nahri, L. O. Solvent modulation in hydrophobic interaction
chromatography. Biotechnology and Applied Biochemistry. 1991, 13, 151-
172.

Breitenbach, M. Hydrophobic chromatography of catalases and hemoglobins
on a series of alkyl-substituted sepharoses. ln Protides of the Biological ‘ ‘1
Fluids. Peelers, H.; Ed.; Pergamon Press: Oxford, 1976, pp 687-690.

Bull, H. B.; Breeese, K. Protein solubility and the lyotropic series of ions.
Archives of Biochemistry and Biophysics. 1980, 202(1), 116-120.

Caldwell, K. D.; Axen, Rolf, Porath, J. Utilization of hydrophobic interaction 1
for the formation of an enzyme reactor bed. Biotechnology and
Bioengineering. 1975, 17, 613-616.

Chen R. Removal of fatty acids from serum albumin by charcoal treatment.
The Journal of Biological Chemistry. 1967, 242(2), 173-181.

Creigthon, T. E. Proteins, Structures and Molecular Principles. W. H.
Freeeman, New York, 1982.

Dow Chemical. Methocel cellulose ethers, techincal handbook. Form 192-
1062-88-JB, 1988.

Gorgon, N. F.; Tsujimura, H.; Cooney, C. L. Optimization and simulation of
continuous affinity—recycle extraction. Bioseparation.1990, 1, 9-21.

Flygare, 8.; Mannsson', Larsson, P.- O.; Mosbach, K. Affinity precipitation of
enzymes. Applied Biochemistry and Biotechnology. 1982, 7, 59-61.

Franks, F. Hydrophobic interactions, a historical perspective. Faraday
Symposia of the Chemical Society. 1982, 17, 7-10.

 

170

Hjertén, S. Some general aspects of hydrophobic interaction
chromatography. Journal of Chromatography. 1973, 87, 325-331.

Hjertén, S.; Rosengren, J.; Pahlman, S. Hydrophobic interaction
chromatography, the synthesis and the use of some alkyl and aryl
derivatives of agarose. Journal of Chromatography. 1974, 101, 281-288.

Hofstee, B. H. .l. Hydrophobic effects in adsorptive protein chromatography.
Journal of Macromolecular Science-Chemistry.1976, A10(1&2), 111-147.

Hofstee, B. H. J.; Otillio, N. F. Non-ionic adsorption chromatography of
proteins. Journal of Chromatography. 1978, 159, 57-69.

Janson, J.- 0. Large-scale affinity purification, state of the art and future
prospects. Trends in Biotechnology. 1984 , 2(2), 31-38.

 

Jennissen, H. Evidence for negative cooperativity in the adsorption of
phosphorylase b on hydrophobic agaroses. Biochemistry. 1976, 15(56),
5683-5692.

Jennissen, H. P. Cooperative phenomena in the adsorption of the
phophorylase kinase and phosphorylase b on hydrophobic agarose. ln
Hydrophobic, ion exchange, and affinity methods. Epton, R.; Ed.; Ellis
Horwood: Chichester, 1978, pp 42-52.

Male, K. B.; Luong, J. H. T.; Nguyen, A. L. Studies on the application of a
newly synthesized polymer for trypsin purification. Enzyme and Microbial
Technology. 1987, 9(6), 374-378.

Mattiasson, B.; Ling, T. G. i. Ultrafiltration affinity purification, a process
for large-scale biospecific separations. In Membrane Separations
Technology. McGregor, W. 0., Ed.; Marcel Dekker: Basel, 1986, pp 99-114.

Melander, W.; Horvath, C. Salts effects on hydrophobic interactions in
precipitation and chromatography of proteins, an interpretation of the
lyotropic series. Archives of Biochemistry and Biophysics. 1977, 183,
200-215.

 

171

Mevarech, M.; Leicht, W.; Werber, M. M. Hydrophobic chromatography and
fractionation of enzymes from extremely halophilic bacteria using
decreasing concentration gradients of ammonium sulfate. Biochemistry.
1976, 15(11), 2383-2387.

Mohr, Peter. Affinity Chromatography. Marcel Dekker, New York, 1985.

Morrison, R.; Boyd, R. Ouimica Orgénica, 7 Ed. Fundagao Calouste Gulbenkian,
Lisboa, 1973, p. 734..

Nemat-Gorgani, M.; Karimian, K. Non-ionic adsorptive immobilization of
proteins to palmityl-substituted Shepharose 43. European Journal of
Biochemistry. 1982, 123, 601-610.

Pahlman, S.; Rosengren, J.; Hjertén, S. Hydrophobic interaction
chromatography on uncharged Sepharose Derivatives. Journal of
Chromatography. 1977, 131, 99-108.

Peters, T.; Taniuchi, H.; Anfinsen, C. B. Affinity chromatography of albumin
with fatty acids immobilized on agarose. The Journal of Biological
Chemistry. 1973, 248(7), 2447-2451.

Pharmacia Fine Chemicals. Affinity Chromatography. 1983

Porath, J; Sundberg, L; Fornstedt, N.; Olsson, I. Salting-out in amphiphilic
gels as a new approach to hydrophobic adsorption. Nature. 1973, 245, 465-
466.

Porath, J. Metal ion-hydrophobic, thiophilic and «electron governed
interactions and their application on salt-promoted protein adsorption
chromatography. Biotechnology Progress. 1987, 3(1), 14-21.

Pradeau, 0.; Bellenger, P.; Hamon, M. Methods d'intentification de quelques
epaississants glucidiques industriels. Anna/es Pharmaceutiques
Francaises. 1982, 40(6), 555-566.

Pungor, E.; Ateyan, N. 8.; Gordon, N. F.; Cooney, C. L. Continuous affinity-

 

 

 

172

recycle extraction, a novel protein separation technique. Biotechnology.
1987, 5, 604-608.

Flaibaud. 0.; Hogberg-Raibaud, A.; Goldberg, M. E. Purification of E. coli
enzymes by chromatography on amphiphilic gels. FEBS Letters 1975,
50(2), 130-134.

Raymond, J.; Azanza, J. -L.; Fotso, M. Hydrophobic interaction
chromatography, a new method for sunflower protein purification. Journal
of Chromatography. 1981, 212, 199-209.

Shaltiel, S. Hydrophobic chromatography. In Hydrophobic, ion exchange, and
affinity methods. Epton, R.; Ed.; Ellis Horwood: Chichester, 1978, pp 13-41.

Shaltiel, S.; Halperin, G.; Er-el, Z.; Finfelstein, T-.; Amsterdam, A,
Homologous series of hydrocarbon-coated agarose in hydrophobic
chromatography. In Affinity Chromatography. Hoffmann-Ostenhof, 0.;
Breitenbach, M.; Koller, F.; Kraft, D.; Scheiner, 0., Eds.; Pergamon: Oxford,
1978, pp 141-160.

Scopes, Ft. K. Dye-ligands and multifunctional adosrbents, an empirical
approach to affinity chromatography. Analytical Biochemistry. 1987, 165,
235-246.

Scouten, W. H. Affinity Chromatography, Bioselective Adsorption on Inert
Matrices. Wiley-lnterscience, New York,1981.

Tanford, C. The hydrophobic effect, formation of micelles and biological
membranes. John Wiley and Sons, New York, 1980.

Wichman, A.; Anderson, L. -0. Purification of human serum albumin by
affinity chromatography. Biochimica et Biophysica Acta.1974, 372, 218-
224.

Yang, B. -L.; Goto, S. Consecutive reaction system combined with batchwise
adsorption and columnwise elution. Journal of Chemical Engineering of
Japan. 1992, 25(2), 229-231.

 

 

173

Yang, B. -L.; Goto, M.; Goto, S. Enzyme purification by affinity
chromatography combined with batchwise adsorption and columnwise
‘elution. Journal of Chemical Engineering of Japan. 1989, 22(5), 532-537.

 

 

Chapter 5

General Conclusions and Recommendatlons

5.1 - General Conclusion

The general conclusion of this dissertation is that flotation can be
applied to the recovery of fermentation products. This work explored and
showed the feasibility of the two routes through which this can be
accomplished. First, was the use of polymer-protein precipitates already
in use in the separation of proteins. The natural floatability of these
precipitates make the development of the flotation processes a relatively

easy task. The partition of the protein to a particulate system seems to

protect the protein from denaturation and loss of activity during flotation.

The recovery of a-amylase from fermentation broth and periplasmic
extract by partition to (hydroxypropyl)methylcellulose and subsequent
flotation were the example of such a route.

Second, a flotation system was designed to recover products
according to a specific interaction in which they participate. The
chemical engineering of a floatable molecule containing a ligand that is
also floatable with the product bound to it was the core of this work. The
feasibility of such an approach to flotation in biotechnology was
exemplified by the production of fatty acids derivatives of
methylcellulose and their use in the precipitation and flotation of bovine

serum albumin.

174

Ll

 

1

 

 

175

5.2 - Recommendations

'As said in the beginning of this dissertation, little has been done in
the field of flotation applied in the recovery of biomolecules from
fermentation broth. Therefore, there are many areas were future work can
contribute to the expansion of flotation in biotechnology.

In terms of the experimental work done in the recovery of a-
amylase, the identification of the molecules that have the depressant
effect on flotation would be helpful in improving the process as well in
helping to understand the flotation of (hydroxypropyl)methylcellulose.
Since this polymer showed good qualities as a flotation carrier, it is
important to optimize its flotation regarding particle size distribution.
Due to the better floatability of the flocs of twisted filaments of salted-
out polymer, a better understanding of the filaments and floc formation is
necessary. Followed by the study of how they are attached to air bubbles
and lifted, an optimization of the flotation of these particles can be done.

The production of derivatives of methylcellulose needs optimization
in few areas. According to the protocol used, large amounts of fatty acids
are wasted and it is difficult to assure homogeneity in the reaction
mixture. The use of an inert solvent for the fatty acid can improve both of
these problems. The washing of the derivatives needs to be changed after
the optimization of the esterification reaction. The lower load of fatty
acids to be washed will result in lower solvent requirements. Also, one
must investigate if the different derivatives have significantly different
molecular weight distributions due to acid hydrolysis. If so, the effect of
molecular weight on the properties of the derivatives should be

investigated.

 

 

 

176

Looking now in broader terms, there is work needed in the two
routes for the application of flotation in biotechnology. In the area of
exploring the natural floatability of protein-polymer precipitates known
to biotechnology, a screening of these precipitates regarding their
floatability is lacking. Such a study would not only identify precipitates
that are floatable but also develop basic knowledge of which conditions
and which characteristics protein precipitates must have to be floatable.
In the design of flotation systems, there is a need for development and
characterization of carriers. The literature is abundant with descriptions
of ligands and conditions for their binding with a variety of proteins.
Therefore, these carriers must be developed with this in mind. They must
be able to be insolubilized in a variety of conditions to match with the
ligand-protein binding conditions. Parallel to this, the attachment of
different ligands to these new carriers would open the door for the

purification of a variety of different proteins by flotation.

 

 

 

 

APPENDIX

 

177

Raw data used in the preparation of tables and figures.

Table A.1 - Raw data of Figure 3.4.

 

 

 

Air flow rate HPMC recovery (as percentage of initial amonut)
(mL/min) 10 pm 16 pm

84.8 98.0 -

76.8 98.4 -
61.3 96.7 - t
50.5 93.2 99.0 L
32.5 84.0 98.2 .
15.2 - 94.4 '
10.6 - 93.2 I
7.8 - 86.1

 

 

Partition of synthetic mixture: batch volume, 4 L; ammonium sulfate
addition rate, 5 and 20 mL/min for 10 and 16 pm, respectively. Continuous
flotation with no washing: column used, #1; feed flow rate, 27 mL/min;
froth height, 0.5 to 3 cm.

Table A2 - Raw data of Figure 3.5.

 

Ammonium sulfate concentration HPMC recovery
(percent of saturation (as percentage of initial amount)

 

40.0 101.5
45.0 94.8
45.0 100.6
50.0 68.4
50.0 68.5
55.0 16.6
55.0 24.6
60.0 13.3
60.0 29.1

 

Salting-out of synthetic mixture: batch volume, 1.5 L; HPMC concentration
before salt addition, 0.07 percent; ammonium sulfate addition flow rate,

14 mL/min. Continuous flotation with no washing: column used, #1; feed
flow rate, 36 mL/min; froth height, 1 cm.

 

178

Table A3 - Raw data of Figure 3.6.

 

 

Feed flow rate HPMC recovery
(mL/min) (as percentage of initial amount)
67.3 98.0
99.0 96.2
150.1 94.8
153.5 103.2
164.9 108.5
165.5 94.8
168.0 99.4
170.0 65.5
175.6 76.9
178.2 82.7
201.3 50.7

 

 

Partition of synthetic mixture: batch volume, 10 L; ammonium sulfate
addition flow rate, 60 mL/min. Continuous flotation with no washing:
column used, #1; froth height, 3 cm.

 

179

Table A.4 — Raw data of Figure 3.7.

 

Wash water flow rate Concentrate contamination by liquid from the

 

(mL/min) feed stream (as fraction of concentrate volume)
0.90 0.410
1.25 0.275
1.90 0.100
3.10 0.070
4.70 0.020
6.80 0.010
0.80 0.470 _I_.
0.90 0.450
1.30 0.280
1.90 0.150
3.10 0.080
4.70 0.030
6.80 0.040

 

 

Partition of synthetic mixture: batch volume, 10 L; ammonium sulfate
addition flow rate, 60 mL/min; tracer, 2.0 percent copper sulfate
(CuSO4.5H20), 0.4 percent sodium hydroxide, and 2.0 percent sodium
potassium tartrate. Continuous flotation: column used,#1; feed flow rate,
40 mL/min; froth height, 4 cm; washing solution, 2.0 percent sodium
potassium tartrate in 30 percent saturation ammonium sulfate.

Table A5 - Raw data of Figure 3.8.

 

 

p—l HPMC recovery" Standard deviation
(as percentage of initial amount)
3.0 81.0 1.6
4.0 81.6 3.0
4.8 77.8 2.8
6.0 80.5 1.7
7.0 81.7 2.0
8.0 79.9 1.9
9.0 79.7 2.4

 

‘Average of three treatments.

Partition: batch volume, 150 mL; HPMC concentration, 0.1 percent;
ammonium sulfate concentration and addition rate, 30 percent saturation
and 3.3 mL/min, respectively. Flotation: air flow rate, 4.0 mL/min; make-
up water flow rate, 1.6 mL/min; flotation time, 5 minutes.

 

180

Table A.6 - Flaw data of Figure 3.9 and Figure 3.10.

 

 

Diafiltration level Recovery (as percentage of initial amounts)
Activity HPlVC HPMC x 0.912
0 58.4 58.9 53.7
0 60.1 66.0 60.2
1 81.6 88.0 80.3
1 79.4 84.3 76.9
2 78.1 84.4 77.0
2 86.4 96.7 88.2
2 88.0 98.5 89.8
3 90.2 99.7 90.9

 

Partition of broth mixture: batch volume, 1.5 L; ammonium sulfate addition
flow rate, 31 mL/min. Continuous flotation without washing: column
used, #2; feed flow rate, 20 mL/min; froth height, 5 cm.

 

Table A.7 - Raw data of Figure 3.11.

 

p—l HPMC recovery' Standard deviation
(as percentageof initial amounts)

 

3.0 46.3 12.9
4.0 24.0 5.6
5.0 22.7 3.5
6.9 19.3 3.4
9.0 19.5 0.9

 

'Average of three treatments.

Partition: batch volume, 150 mL; HPMC concentration, 0.1 percent;
ammonium sulfate concentration and addition rate, 30 percent saturation
and 3.3 mL/min, respectively. Flotation: air flow rate, 4.0 mL/min; make-
up water flow rate, 1.6 mL/min; flotation time, 5 minutes.

 

181

Table A8 - Raw data of Figure 3.12.

 

 

p-l HPMC recovered' Activity recovered’
with particles Total

7.0 26.4 41.4 88.5
6.0 27.2 41.3 93.5
5.0 33.7 42.3 90.8
4.0 55.6 61.1 89.0
3.0 63.4 29.5 46.3
2.5 53.4 18.8 31.3

 

' As percentage of initial amounts.
Partition of broth mixture:

flow rate, 42

Table A9 - Raw data of Figure 4.9.

Average of two treatments.
ammonium sulfate addition
mL/min. Continuous flotation: column used #2; feed flow
rate, 20 mL/min; washing solution flow rate, 7 mL/min;froth height, 7 cm.

batch volume, 4 L;

 

Reaction time

Extent of reaction of mixtures

 

(hours) with catalyst without catalyst
5 0.59 0.07
17 0.92 0.14
27 0.91 0.23
41 0.94 0.27

 

 

Fatty acid, octanoic acid; catalyst 20 pL/g of methylcellulose;
temperature, 65 °C.

Table A.10 - Raw data of Figure 4.10.

 

Time Extent of reaction

 

5 pL acid/g methylcellulose 20 pL acid/g methylcellulose

 

 

 

25 °C 50 °C 70 °C 100 °C 25 °C 50 °C 70 °C

4 0.04 0.30 0.51 0.97 0.04 0.63 0.74
9 0.10 0.54 0.81 1.02 0.08 0.74 0.87
12 0.19 1.01 0.99 1.07 0.06 0.75 0.87

 

Fatty acid used: octanoic acid.

182

Table A.11 - Raw data of Figure 4.11.

 

 

 

Number of carbons in the chain Extent of reaction
16 0.02
16 0.08
14 0.13
14 0.11
12 0.42
12 0.45
10 0.71
10 0.73

8 0.84
8 0.86
6 1.10
6 1.05
5 1.37
5 1.17
4 1.13
4 1.34
3 1.70
3 1.34

 

Volume of sulfuric acid, 5 pL/g of methylcellulose; temperature, 65 °C;
reaction time, 24 hours.

I ”it—-

183

Table A.12 - Raw data of Figure 4.12.

 

 

 

 

Time Extent of reaction of the different derivatives'
(hours)
C3 C4 C6 08 C10 C12 C14

3 0.58 0.46 0.32 - - - -

6 0.82 - 0.49 - - - -

9 0.79 0.84 0.63 - - - -
12 0.96 0.92 0.74 0.62 - - -
15 1.13 1.00 0.81 - - - -

8 - - - 0.48 - - -
10 - - - 0.55 - - -
14 - - - 0.64 0.33 — -
17 - - - 0.67 - - -
20 - - - 0.73 - - -
23 - - - 0.74 - - -
22 - - - - 0.45 0.18 0.12
36 - - - - 0.57 0.35 -
61 - - - - 0.68 0.47 -
85 - - - - 0.73 0.56 0.31
65 - - - - - - 0.22
'Average of two treatments.

Volume of sulfuric acid, 5 pUg of methylcellulose; temperature, 65 °C.

 

 

Table A.13 - Raw data of Figure 4.13.

184

 

 

 

Time Yield of the different derivatives (g)*
(hours)
C3 C4 C6 C8 C10 012 C14

3 0.938 1.024 1.016 - - - _

6 0.82 0.944 1.003 - - - -

9 0.727 0.864 1.041 - - - -

12 0.678 0.826 1.059 1.048 - - -

15 0.714 0.807 1.003 - - - -

8 - - - 1.026 - - - l
10 - - - 1.042 - - - g
14 - - - 1.045 1.011 - - g
17 - - - 1.041 - - - '
20 - - - 1.035 - - -

23 - - - 1.029 - - - E
22 - - - - 1.026 0.984 0 948

36 - - - - 1.015 0.978 - ‘

61 - - - - 0.994 0.974 -

65 - - - - - - 0.897

85 - - - - 0.953 0.936 0.833

 

 

‘Average of two treatments.
Volume of sulfuric acid, 5 pL/g

of methylcellulose;

temperature. 65 °C.

185

Table A.14 - Raw data of Figure 4.14, Figure 4.15 and Figure 4.16

 

 

Derivative E bulk E soluble E insoluble Apparent
solubility (mg/mL)
0.57 - - 0.744
0.61 - - 0.872
0.82 - - 0.670
0.82 - - 0.681
0.66 - - 0.614
03 0.91 - - 0.582
0.99 - - 0.536
0.93 - - 0.541
1.07 - - 0.333
1.18 - - 0.378
0.42 - - 0.943
0.49 - - 0.970
0.72 0.46 0.94 0.715
1.01 - - 0.524
0.84 0.65 0.91 0.441
C4 0.83 0.59 0.94 0.483
0.94 0.59 1.05 0.301
0.90 0.57 1.09 0.389
0.97 0.63 1.05 0.288
1.03 0.61 1.09 0.293
0.28 - - 0.954
0.36 - - 0.952
0.46 - - 0.560
0.52 - - 0.580
0.63 - - 0.297
C6 0.62 - - 0.311
0.76 - - 0.149
0.71 - - 0.141
0.79 - - 0.062
0.82 - - 0.045

 

186

Table A.14 - cont'n

 

Derivative E bulk E soluble E insoluble Apparent
solubility (mg/mL)

 

 

 

0.47 0.21 0.53 0.365
0.49 0.22 0.48 0.359
0.55 - - 0.290
0.54 - - 0.291
0.60 - - 0.231
0.63 - - 0.181
C8 0.65 0.26 0.60 0.179
0.63 0.22 0.62 0.189
0.68 - - 0.114
0.66 - - 0.117
0.70 - - 0.085
0.75 - - 0.089
0.74 0.25 0.72 0.064
0.73 0.22 0.66 0.060
0 36 - - 0.461
0 30 - - 0.442
0 50 - - 0.195
0 40 - - 0.294
0.52 - - 0.135
C10 0.62 - - 0.044
0 66 - - 0.024
0 69 - - 0.014
0 74 - - 0.011
0 72 - - 0.014
0 25 0 14 0.28 0 589
0 12 0 12 0.30 0 865
0.35 0.14 0.45 0.423
C12 0.35 - - 0.418
0 45 0 13 0.56 0 227
0 48 0 13 0.50 0 188
0 59 0 11 0.67 0.040
0 53 - - 0.032
0 59 - - 0.013

 

Table A.14 - cont'n

 

Derivative E bulk E soluble E insoluble Apparent
‘ solubility (mg/mL)

 

0.13 - - 0.902
0.11 - - 0.909
0.24 - - 0.573
0.20 - - 0.690
0.36 - - 0.438
C14 0.26 - - 0.512
0.47 - - 0.149
0.43 - - 0.249
0.50 - - 0.077
0.50 - - 0.051

 

E soluble and E insoluble stand for extent of reaction of the soluble and the
insoluble fractions of the derivatives, respectively.

Table A.15 - Raw data of Figure 4.17.

 

Salt concen- Derivative precipitated (as percentage of initial amounts)

 

 

tration'
C4 C8 C12 C1
4 81.3 67.0 85.5 93.4
4 84.7 68.7 86.3 100.6
8 36.0 32.2 37.3 95.5
8 34.1 32.8 38.0 96.6
12 10.2 17.0 21.0 16.2
12 10.9 17.7 21.6 17.9
16 3.8 8.8 11.9 4.8
16 4.5 9.3 12.6 3.8
20 2.0 5.4 7.0 2.2
20 2.9 6.1 7.8 2.9

 

Soluble fractions of the derivatives resulted from experiment showed in
Figures 4.14 and 4.15 were salted-out with ammonium sulfate. The
concentration given by the remaining amount of derivatives in the solution
was measured and compared with the original concentration

 

 

188

Table A.16 - Raw data of Figure 4.18.

 

 

 

d-I Supernatant absorbance at 280 nm
Ammonium sulfate Buffer
8.89 0.330 -
8.50 0.340 0.359
8.05 0.336 0.373
7.50 0.337 0.325
7.10 0.330 0.343
8.80 0.348 -
7.30 0.337 -
6.80 0.337 -
5.90 0.336 0.333
5.40 0.324 -
5.20 0.300 0.335
5.05 0.252 0.335
4.85 0.150 0.322
4.65 0.032 -
4.50 0.012 -
4.15 0.004 0.321
3.50 0.001 0.325
3.05 0.007 0.322
7.30 0.335 -
6.60 0.344 0.341
5.70 0.346 -
5.40 0.346 0.329
5.25 0.338 -
5.10 0.332 0.330
5.00 0.324 -
4.90 0.315 0.323
4.80 0.258 -
4.70 0.138 0.322
4.75 0.178 0.332
4.65 0.086 -
4.55 0.020 0.324
4.50 0.020 -
4.40 0.017 0.320
4.25 0.020 -
3.50 0.022 0.325
3.05 0.016 0.322
7.00 - 0.335

 

 

189

Table A.16 - Cont'n

 

 

 

p-l Supernatant absorbance at 280 nm
Ammonium sulfate Buffer
3.60 - 0.318
2.90 - 0.316
9.10 - 0.336
5.60 - 0.333
4.15 - 0.329

 

Ammonium sulfate concentration, 60 percent saturation. Protein
concentration, 0.66 mglmL; aging time, 2 hours.

Table A.17 - Raw data of Figure 4.19.

 

 

 

pH

Filtrate absorbance at 280 nm

 

C1 C4 CB C12

 

8.30
7.93
7.03
6.23
5.68
8.70
7.83
6.93
6.15
5.68
8.58
7.88
6.95
6.18
5.68
8.38
7.85
6.98
6.23
5.65

0.210
0.218
0.199
0.190
0.210
0.024
0.009
0.005
0.013
0.007
0.180
0.182
0.182
0.176
0.165
0.191
0.191
0.186
0.166
0.153

 

Cn/BSA = 2; protein concentration, 0.66 mglmL; salt concentration, 60
percent saturation; aging time, 2 hours. The dashed line is the expected
absorbance in the case of no precipitation.

190

Table A.18 - Raw data of Figure 4.20.

 

 

Time (hours) Filtrate absorbance at 280 nm
0 0.097
0 0.088
5 0.077
5 0.081

10 0.079
10 0.078
16 0.087
16 0.081
24 0.071
24 0.071
40 0.069
40 0.081

 

Cn/BSA = 1.5 (decreased form 2 to avoid supernatants with low absorbance
at 280 nm); protein concentration, 0.66 mglmL; salt concentration, 60
percent saturation; pH, 6.0.

Table A.19 - Raw data of Figure 4.21.

 

 

Ammonium sulfate concentration Protein precipitated
(as percentage of saturation) (as percentage of initial amount)
30 85.1
30 76.2
35 77.1
35 76.5
40 81.3
40 78.9
45 58.5
45 62.5
45 67.5
50 49.3
50 55.1
50 43.5
55 33.8
55 31.4
55 37.1
60 18.1
60 9.1

 

Cn/BSA = 2; protein concentration, 0.66 mglmL; pH, 6.0; aging time, 1 hour.

191

Table A.20 - Raw data of Figure 4.22.

 

 

Ammonium sulfate Protein precipitated
concentration C4IBSA (as percentage of
(as percentage of saturation) initial amount)

0.8 68.0

0.8 71.3

0.8 68.9

0.5 65.1

60 0.5 47.6
0.5 55.6

0.2 29.9

0.2 42.0

0 2 39.3

2 5 71.4

2 5 68.8

2 5 68.5

1.0 55.5

55 1.0 50.3
1 0 47.4

0 5 37.5

0 5 38.0

0 5 44.3

5 0 72.1

5 0 63.3

5 0 66.4

2.0 48.6

50 2.0 50.5
2 0 50.3

1 0 43.0

1 0 43.4

1 0 42.2

5 8 44.4

5 8 42.7

5 8 46.2

4.0 43.6

45 4.0 40.1

4 0 40.1

2 0 35.4

2 0 35.4

2 0 32.8

 

Protein concentration, 0.66 mglmL; pH, 6.0; aging time, 1 hour.

1'1

GRN STR

HICHI
"MWWMWW
31293

N
0

TE UNIV. LIBRQRIES
WWWWMWWWMM
09109962