Adsorptive and Chromatographic Separations

for Bio-molecules

PART 1

Arvind Lali
Chemical Engineering Division
Mumbai University Institute of Chemical Technology
Matunga, Mumbai 400 019

Email : [email protected]
Ongoing Projects :

Resindion SRL, Italy (Mitsubishi Chemical Corporation)

BioRad Laboratories USA
AlPharma, Europe
Cognis, Germany
DMV, Holland
Biocon India Limited
Dr. Reddy’s Laboratories
Lupin Pharma Limited
Alembic Limited
Wockhardt Limited
Advanced Biochemicals Limited
Godrej Industries

DBT, DST, DAE, DAE/BRNS, AICTE, ICMR, CSIR

Research areas of the group

Bioseparations : Adsorptive/Selective Precipitation

Molecular modeling/affinity separations
Preparative Chromatography of Bio-products
Dynamics of Solid-liquid fluidization and Expanded
Bed adsorption

Novel methods for bio-transformations

Biosensor design (FIA based)

RTD in Complex Flows

Dynamics of gas-solid fluidized beds

Group : 1 PD Fellow
15 PhD scholars
9 MS scholars
3 Research Assts
 Cost of Bioproduct  Purity of the product required

Cost components of a bio-product

- developmental cost or the cost of technology

- capital cost

- manufacturing cost

(Manufacturing cost + Capital cost) ∞ Final purity desired

Two stages of Production of a Bioproduct

(A) Bio-reactions (e.g. in a bioreactor)

-single step or multistep

(B) Isolation & Purification

or Downstream Processing (DSP)
-essentially multistep
 A typical DSP scheme : 10-20 steps

Start : Fermentation broth OR Natural source

Cell harvesting : centrifugation/ microfiltration

Cell disruption : mechanical, physical, chemical or enzymatic

Cell debris separation : centrifugation/microfiltration

Concentration steps : precipitation, ultrafiltration, extraction with

intermediate wash/recovery steps

Purification steps : ultrafiltration, adsorption/chromatography e.g.

Ion-exchange, hydrophobic interaction, affinity

End : Stabilized Product

A powerful Unit Operation:
Adsorptive/Chromatographic Purification

- Indispensable for high purity production

- Expensive technology (?)

- Lack of expertise

- Art OR Science

- Complex to model
(esp. when treating complex mixtures)
Basis of Adsorptive/Chromatographic Separations

Differential and Reversible Affinity between the Matrix and Solutes

Physical Adsorption/Chemical Adsorption

Modes of Contact/Classification of Methods :

- Batch adsorption
- Column chromatography
- Paper/Thin layer chromatography
- Radial chromatography

Other classifications:

gas/liquid/high performance liquid – chromatography

 S, qAi

L, CAi Single Stage

L, CAf
Batch Adsorption

S, qAf

CAi. L + qAi. S = CAf.L + qAf.S

Efficiency = (qAf – qAi).S/ CAi. L

qA
Thermodynamic Equilibrium : Adsorption Isotherm
Multi-component adsorption isotherms !
CA
 Separation depends on relative kd value
where kd = qAf/CAf

Relative kd = kdA/kdB

If Relative kd >> 1 : Batch Adsorption

Else, multistage adsorption : Column Adsorption

Applications of Column Separation :

- Capture (complete recovery and no handling of solids)

- Concentration from dilute streams (high adsorbent/liquid ratio)
- Resolution between closely related solutes
Consider,
Solute sample volume applied << Column capacity

Three cases in Isocratic system :

a. Solute completely adsorbed, =1
b. Solute does not adsorb at all,  =0
c. Is partially adsorbed, 0<<1

Where  = amount of solute adsorbed/total solute

Solute emerges at CV (column volumes) 1/(1-)

=0

=0.4

1.0 1.7
Operating modes in Column Chromatography

Isocratic Elution (analytical, GPC/SEC, SMB)

Frontal Adsorption/Desorption (IEC, affinity, HIC)

- Isocratic elution
- Gradient elution
- Displacement elution
Design of Adsorptive/Chromatographic
Separation Operation

Scale-up, Operation,
Selection of media Control and Validation
and basic protocol

Optimizing performance
of media and column
 d p . l .U t
Re p 


Selection of Interaction mode

Ion Exchange Chromatography

Gel Permeation/Size Exclusion Chromatography

Hydrophobic Interaction Chromatography

Reverse Phase Chromatography

Affinity Chromatography
Depends on
Covalent Chromatography Target molecule/s
And
Combinations ! Impurities !
Classification of Bioproducts as

- SMALL MOLECULES (vitamins, antibiotics, peptides)

invariably occur with macro-molecules that need to be

rid off

- MACRO-MOLECULES (enzymes, nucleic acids, sugars,

antibodies)
careful handling for bioactivity
Macro-molecules create problems due to:
Complex structures
Amphoteric properties
Structure related activity
Ion Exchange based Separations for Charged Molecules differing in their

- ionic dissociation constants, and/or

- charge density and distribution

Hydrophobic Interaction and Reverse Phase Chromatography

Based on differences in :

- hydrophobic index or solubility parameter

- density and distribution of surface hydrophobic
sites

Proteins : Role of surface amino acid residues and their accessibilities

HIC : More powerful technique than IEC !

Useful concept in bio-separations : Molecular Recognition

Concept of AFFINITY INTERACTIONS

Exploits the property of bio-molecules to selectivity recognize

and reversibly interact with complimentary molecules/bio-molecules
with certain properties/structure

+
+ +
impurities impurities
 Classification of Affinity Interactions

Affinity Interactions

Bio-specific Pseudo Bio-specific

Specific Group specific

receptor-hormones enzyme-cofactor Non-Biological
antigen-antibody lectins-glycoproteins metal ions
enzyme-substrate proteins A and G triazine dyes
enzyme-inhibitor etc. HIC ligands
Biological
amino acids
like histidine
 BASIS OF TECHNIQUE

+
ligate
matrix ligand

Tailor-made affinity matrices can be developed from commercial

adsorbents using established chemistries

Associated problems – cost and fragility !

SOLUTION  Robust Pseudo-affinity Ligands

 Chromatography as Initial Step in DSP
Selection and Design of Adsorbent Matrix

Most crucial component of the operation success

Choice between market available adsorbent/design

of adsorbents

Important parameters : pore size/distribution, bead size/shape,

surface area, site density, site chemistry,
binding capacity/rate, strength and
reversibility of interactions

Specificity of Interaction desirable

 High Relative Kd
Key Issues : Binding Capacity
Binding Strength
Specificity or Selectivity or Resolution

Both these issues are determined by

Thermodynamics, and
Hydrodynamics

Thermodynamics = Fn (ligand, ligand density, binding strength, microenvironment)

Hydrodynamics = Fn (pore size, particle size, solute size, flow velocity, mobile phase
physical properties)
Media Selection : Thermodynamic Considerations
(deals with molecular level phenomena)

Depend on properties of solutes and their variation with

environmental parameters e.g. temperature, pH, ionic strength etc.

Net +ve pI

pH
Net -ve

Charge Variation with pH

Hydrophobicity

Temperature
or
Ionic Strength

Hydrophobicity Variation with Temperature

Protein-Site Interactions :
Equilibrium (Dissn)Constants
L + Pn ↔ LPn k1
L + LPn ↔ L Pn 2 k2
L + L Pn ↔ L Pn
2 3 k3
….
L + Ln-1Pn ↔ LnPn kn

Koverall = Ke = k1 x k2 x k3 x…..x kn

Multipoint interactions necessary for good binding of large molecules

IExC = Multivalent molecules bind preferentially at low ionic strength and

desorb at high ionic strength (Example : Water Softening)
Binding Strength : Also depends upon other factors

Let,  = Ratio of Adsorbed solute to Total solute

So that  = 1/(1+K)
where K= Cs/Cm the partition coefficient

For, m + p ↔ q Ke = (m) x (p)/(q)

At equilibrium mt = total sites = m + q

pt = total protein = p + q
So that = q/pt

pt.2 – (mt + pt + Ke) + mt = 0

Typically
mt~ 0.01mM for affinity adsorbent and, 1mM for ion exchangers

Ke mM pt mM mt mM 

0.1 0.1 1.0 0.9

0.1 0.5 1.0 0.85

0.01 0.5 1.0 0.98

0.01 0.01 1.0 0.99

1.0 0.01 1.0 0.50

1.0 0.1 10.0 0.99

0.1 0.1 0.2 0.59

0.1 0.1 0.5 0.81

 Needs to be about 0.8 for chromatography which requires that

Ke < 0.1mM or Ke < 10-4 M
Binding Capacity and Ligand Density

Binding
Capacity,
mg/mL

0
Ligand Density, mM
 Binding Strength and Ligand Density

(a)

Binding
Strength,
mg/mL
(b)

0
Ligand Density, mM

(a) – Affinity, IMAC, HIC (b) -- Ion Exchange, HIC

Equilibrium Binding Capacity : Isotherms

q
mg/mL

Equlbm concn. C
 Single Component : Models

Langmuir Model :

q max C
q
q max C q Reciprocal plot
CK
CK

Freundlisch Model : n
q  kC Logarithmic plot
Multicomponent Isotherms : Models

Langmuir Bi-component Adsorption isotherms

a 1 C e1
q1 
1  b1Ce1  b2Ce 2

a 2 C e2
q2 
1  b1Ce1  b2Ce 2

Freundlich bi-component adsorption isotherms

a 1 C (b1 b11)
q1  b11 e1 b12
C ei  a 12 C e2

a 2 C (b2  b21)
q 2  b11 e2 b21
C ei  a 21C e2
Careful Observations from Isotherms : At high concentrations
Non-linear case

Isotherms can also mislead when hydrodynamic considerations

Override thermodynamic selectivities !
 Media Selection : Hydrodynamic Considerations

Dynamic Binding Capacity, Concentration and Resolution

All depend upon hydrodynamics ie the Rate Processes that are
additional to the intrinsic adsorption rate process

Rate Processes should be as fast as possible to approach equilibrium

so that maximum use of thermodynamic capacity and selectivity is made possible

Ci

Concn

Time, min
Dynamic Batch Binding Process
Four phenomena ‘spoil’ the instantaneous approach to
equilibrium situation in a column adsorption system with constant
flowing mobile phase :

a. Axial or Longitudinal Dispersion in Column

b. Diffusion through to adsorbent particle surface
c. Diffusion in along the pores to inside adsorbent surface
d. The adsorption process

In preparative chromatography the first of the above is called the

dispersive effect while the latter three are clubbed into what is called the
non-equilibrium effect

A column operation is an unsteady state operation where there is

continuous ‘Upsetting of the Equilibrium’

One advantage : Full use of isotherm capacity compared to Batch system

 Eddies and
Backmixing
Axial Flow

Film Diffusion

Pore Diffusion
Mathematical models for unsteady state rate processes

Two components of the models

1. Hydrodynamic rate processes in bulk

and in adsorbent matrix

2. Thermodynamic rate processes : kinetics

of adsorption and equilibrium

Rate models and Statistical models : Basis for emergence

of HPLC in sixties
Classification of Rate Models :

• Ideal-Linear system
• Ideal Non-linear system
• Non-ideal-Linear system
• Non-ideal-Non-linear system

Most often analysis done using linearized non-ideal systems

However, today, computing power available permits handling

non-linear systems with ease : requires solution of initial value
non-linear unsteady state simultaneous differential equations

Solution of the model gives simulation of breakthrough curve, and

Elution profiles and resultant resolutions possible on a given column,
and can form basis of design of suitable hardware and software
However, a model simulation is as good
as the accuracy of the parameters used

The parameters involved are related to :

- Relative diffusion processes in the pores of the matrix

- Description of the pore structure of given matrix

- Axial dispersion, bulk and surface flow profiles

- Kinetics of adsorption of different solutes of the mixture

In practical situations most of these are unknown !

 Use of simple Linearised system

Concept of Plate Height : From Dsitillation

Van Deemter Equation : Derived from rate model

HETP = A + B/u + C.u

Axial dispersion Solid phase

effects effects : film diffn.;
Molecular diffusion pore diffusion, and
effects adsorption

LOW HEPT larger number of theoretical plates and

thus higher efficiency in a given column height
Typical HETP variation with mobile
phase velocity

HETP

Velocity, cm/s

Practical equation
in prep systems HETP = A + C.u
Plate Height on a given media and column depends on
solute molecule

BSA

Chymotrypsin
H, cm

Glycyl tyrosine

U, cm/hr
Capacity and Resolution Power both related to
H offered by a column + media

Two Situations in a Typical Adsorption Chromatography:

1. Frontal Adsorption of the Solute till Breakthrough

2. Differential Elution of the Solutes

OBJECTIVE :

- Maximum Loading and Productivity (per unit time per column volume)
- Sharp Eluting peaks with Complete Resolution

Both require high column efficiency i.e. Low Plate Height

Free bed

Adsorption Zone, La Solute Bands

Exhausted resin
Low Plate Height, H or Large number of Plates, N per meter bed

1. Sharper breakthrough and better Capacity

2. Sharper peaks and better Resolution

H  Depends upon Hydrodynamic parameters ie rate processes

Estimation of Plate Height : Important in Resin Evaluation and Scale-up

 1.6
CE
1.4

Eluent concentration (arbitrary)

1.2

0.8

0.6

0.4

0.2
CB 0
0 5 10 15 20 25

Eluent volume, V (arbitrary units)

VB VR VE
Typical single solute breakthrough profile

BREAKTHROUGH CURVE ANALYSIS

Number of theoretical plates N given by
2
 VR 
N  16  
VE  VB 

Length of adsorption zone La

VE VB Wa
La  L Where f 
VE  (1 f )(VE VB ) W

Wa  W   and WCi (VE VB)

And  is the area under the curve in the period between
breakthrough and exhaustion

La
H 
N
 ELUTION PEAK ANALYSIS

tR

W1/2

 t 
No. of Plates (N)  5.54 R 
 w1/2 
L
Plate Height (H) 
N
tR = Maximum retention time
W1/2 = Half peak width
L = Length of bed

Note : Accurate if the peak is nearly Guassian with small tail

Resolution of the two solutes 0.54 tR1 tR2
0.48

Two solute effluent curve with 0.42

0.36
Gaussian distribution of solute

Concentration
0.3 2 2
concentration about peaks at
0.24
Mean residence time and deviations 0.18
About mean time 0.12
0.06
0
The degree of separation of two 0 2 4 6 8 10 12 14 16
Solutes usually expressed in terms Time or Volume

Of resolution defined as
Scaled Gaussian effluent curves

1  t R2  t R1 
R 12   
2 σ1  σ 2

 Where,
tR1 = Maximum peak time of 1st solute
tR2 = Maximum peak time of 2nd solute
1 = Standard deviation of 1st peak
1 = Standard deviation of 2nd peak
For two solutes peak eluting, the Resolution parameter, Rs is
defined as (Should be more than unity for complete resolution)

   1  k avg  N 

Rs       1.0

   1  k avg  1  2 

where  is the k2/k1, and

kavg is average of two capacity factors k1 and k2

Capacity Factor for Solute i given by

Vs Cs Vs
ki     Ki
Vm Cm Vm

Where, K is the Distribution or Partition Coefficient for Solute i

Typical design procedure :

1. Select the best selective resins from batch binding experiments

2. Determine isotherms/equilibirum capacities under difft conditions
3. Breakthrough experiments with pure solute and mixture
at different velocities. Estimate HETP vs velocity relation.
4. Elution at different conditions (isocratic, step gradient and
linear gradient) and determine elution H and
resolution parameters
5. From 3 and 4 establish optimum operating conditions, so
adsorbent is optimally use, and the elution resolves the
solutes
6. Determine the effect of scale-up on performance of column
Considerable experience and expertise required in :

Choice and design of Matrix

Requires understanding the different attributes of
matrices, and their possible effect on binding/elution

Choice and design of Binding, Washing, Elution,

and Regeneration conditions
All have bearing on economics of purification

Scale-up to Preparative scale and underlying hydro-

dynamic considerations
Effect of flow distribution, and axial dispersion
is most serious