Concentration/fractionation of proteins by salting out

Salting out using ammonium sulfate  is one  of the  classical methods  in protein biochemistry.  Formerly  it was  widely  used  for  the  fractionation of proteins, but it is not a highly discriminating method and it is unusual to get a pure fraction, using this method.  Today it is rather used as an inexpensive way of concentrating  a protein extract,  while leaving non-protein material in solution, and any purification  with respect to  protein is generally regarded as a bonus.

1.  Why ammoniums sulfate?

Polyvalent anions are more effective at salting out than univalent anions, while polyvalent cations tend to negate the  effect of polyvalent anions. The best combination is therefore  a polyvalent  anion  with  a univalent cation.  Anions  can  be  arranged  in  a  so-called “Hofmeister series”, which  describes  their  relative  effectiveness  in  salting  out  at equivalent molar concentrations. In decreasing order of effectiveness, the  series is: citrate  > sulfate  > phosphate  > chloride  > nitrate  > thiocyanate. This series also describes a decreasing tendency for the anions to stabilize protein structure.  Citrate  and sulfate are thus “kosmotropes”, which  tend  to  stabilize  protein  structure,  while thiocyanate and  nitrate  are  “chaotropes”  which  tend  to  destabilize protein structure.  An ideal salt would, therefore,  be citrate  or  sulfate combined with a univalent  cation.  Ammonium  sulfate  is  most  popular because it meets these criteria, is available in a pure  form  at  low cost  and is highly soluble, so that high solution concentrations can be attained. The sulfate ion has been viewed in a number of ways, regarding how it salts out proteins,  including, ionic strength  effects,  kosmotropy, exclusion-crowding, dehydration,  and binding to  cationic  sites,  especially when the protein has a net  positive  charge  (denoted ZH+). All of these may play a role, depending upon the  salt concentration  and the  pH-dependent charge on the protein. Ionic strength effects.  It will be noticed that the Hofmeister series goes from multivalent to univalent ions.  This  largely reflects the fact that the Hofmeister series is based on molarity, while ionic  strength  is a factor  in salting out.  The valency of the ion has an effect on ionic strength as can be illustrated by comparing NaCl with (NH₄)₂SO₄.

Ionic strength is defined as:-

Where, c_i = concentration of each type of ion (moles/liter)

            Z_i  = charge of each type of ion.

Thus in the case of 1 M NaCl,

and for 1  M (NH₄)2SO₄,

Ionic strength effects come into play at low salt concentrations 0-0.2 M) and, as the name implies, they are not specific to  ammonium sulfate. At low ionic strength, protein solubility is at its minimum at the proteins  pI (Fig. 1).  At this pH, intramolecular electrostatic forces between oppositely  charged  side chains  are  at  a maximum,  protein conformation is maximally tightened and protein  hydration  is least.  On either side  of the  pI,  titration  of ionizable groups  leads to  a lessening of intramolecular ionic  interactions.  In consequence, protein  structure becomes more relaxed and hydration and solubility are increased.

Addition of low concentrations  of salt causes a similar  weakening  of intramolecular ionic bonds, with similar consequences of more  relaxed protein structure and greater solubility.  As shown in  Fig. 1,  addition of salt and altering of the pH, away  from  the  pI,  have  similar,  and additive, effects. The increase in solubility of protein upon addition of modest amounts of salt is known as “salting in”.

Figure 1.  “Salting in”  of proteins  the interaction of pH and ionic strength (adapted from Dennison and Lovrien  ).

Kosmotropy. At concentrations  above 0.2  M the  sulfate  ion  acts as a Hofmeister kosmotropes,  i.e.  it  stabilizes protein  structure,  and concomitantly reduces its solubility.  The effect of a kosmotropes,  in stabilizing protein structure, can be described by the reaction:-

Relaxed, open protein structure ------  compact, tight structure (more soluble, less stable)                                       (less soluble, more stable)

Kosmotropes may be described as “pushing” if they act on the left of this reaction and “pulling” if they act on the right, in either case driving the reaction to the right.

Figure 2. The effect of pH on the salting out of a protein by ammonium sulfate.

Sulfate can act as a pulling kosmotropes  by virtue  of its interaction  with protein cationic sites.  Consistent with this,  the  precipitation  of proteins is usually promoted at pH values below the pI (Fig.  2), where the protein has a maximal number of cationic sites. Reinforcing its pulling effect is the fact that the sulfate ion is divalent, and so can bind to more than  one cationic site at a time, and that it has a tetrahedral  structure, with four oxygen atoms that can hydrogen bond to multiple sites on the protein.

Sulfate also acts as a pushing kosmotropes  by virtue  of its extraordinary hydration.  By virtue  of its hydration,  the  sulfate ion can act as a dehydrating  agent and, in  its hydrated  form,  as an  exclusion-crowding agent.  The sulfate anion has 13 or  14 water molecules in its first hydration layer and possibly more in a second  layer. Consequently, in salting out at, say, 3 M ammonium  sulfate, the  sulfate  anion  will have accreted to itself 40 to 45  M out of the total of 55 M H₂O in neat water. In salting out, therefore, a large proportion of the water will be involved in hydrating the  sulfate ions and increasing their  effective  radius.  The large, hydrated,  [SO₄.(H₂O)n]₂- ions crowd and exclude the proteins, pushing them  into  tighter,  more  ordered  (less  soluble) conformations, with lower entropy.  The preferential accretion of the water molecules to the sulfate  ions excludes the  proteins  from  a proportion  of the  water (the proportion increasing with the  salt concentration),  ultimately bringing them to their solubility limit.

No other salt has the combination of properties which make ammonium sulfate so effective at salting out.  Consequently, when the word salt  is used in the  context  of salting  out,  it  invariably  means ammonium sulfate. Similarly, the term “ionic strength” is often used loosely, when what is really meant is the concentration  of ammonium sulfate.

2 . Empirical observations on protein  salting out.

Starting from  zero,  increases in salt concentration  initially  increase the solubility of the proteins, due to salting in.  With further increases in ammonium sulfate concentration, the protein solubility passes through a maximum and then decreases (Fig. 3).

The salting out relationship is described by an empirical equation;-

where,

S = protein solubility (g/l)

I =  ionic strength

fl and K_S are constants.

Figure 3. Solubility  of a  typical protein vs concentration of ammonium  sulfate.

KS, the so-called “salting out constant” (the slope of the plot in Fig. 3),  is  essentially  independent  of  temperature  and pH but varies slightly with the nature of the protein.  fl is markedly dependent upon the pH and temperature (Fig. 4) and also varies markedly with the  nature  of the protein.

Figure 4. The effect of temperature on the salting out of carboxy haemoglobin.

Note that a rise in temperature causes a decrease in fl. Note also that, since fl is in units of log S, a unit change in fl represents a ten-fold change in solubility. Therefore, a protein will be markedly less soluble at higher temperatures and in practice it is better to conduct salting out at, say, 25C rather than at 4C.  The sulfate ion is kosmotropic, so proteins are stabilized by the presence of (NH₄)₂SO₄ and a high salt concentration also inhibits microbial growth.  For these reasons,  also,  it  is less  necessary than usual to work at a low temperature.

The initial concentration of a protein in solution has a major influence on the  amount of (NH₄)₂SO₄ required to  precipitate  it. Proteins appear to fall into two categories, denoted type I and type II, depending upon how their concentration  affects  their  salting  out behaviour. For type I proteins, each protein has a characteristic precipitation curve (e.g. Fig. 5).

Figure 5. The salting out curve of carboxymyoglobin .

The lower the concentration of protein in solution, the more salt is required to  precipitate  it.  In  the  example  given  in  Fig.  5,  if carboxymyoglobin is  present  at an  initial  concentration  of 30 g/l,  it will begin to precipitate  at about 55% saturation  with  (NH₄)₂SO₄, whereas at an initial concentration  of 4 g/l, 65% saturation  is required to  begin precipitation.

Not all proteins behave in this simple way.  Type II proteins2, such as BSA and α -chymotrypsin, precipitate to an extent dependent  upon  their initial concentration, i.e. such proteins manifest a family of precipitation curves, each curve  arising from  a particular  initial  protein  concentration (Fig. 6).

Type I proteins  have a  single  precipitation  curve,  regardless  of the initial protein  concentration.  Type II proteins precipitate in a manner dependent upon their initial concentration.

Figure 6.  Differential salting out behaviour of type I and type II proteins.

Clearly, therefore,  proteins  do not precipitate  between fixed and characteristic limits of ammonium  sulfate concentration,  as is implied in much of the  older literature.  Also, there  is little  point  in  repeating  a precipitation, from the same volume and at the same ammonium sulfate saturation. Since the first precipitation will not have been quantitative, the concentration will be less if the protein is reconstituted in the same volume. To repeat the precipitation, the protein concentration should be readjusted to  the  same  value  as previously,  which  is not  always practicable. In general,  it  is  not  worth  repeating  the  precipitation  as  the cost, in terms of protein lost, is not justified by the increase in protein purity obtained.

Proteins may  be purified from a mixture by differential precipitation at different saturations of ammonium  sulfate  (e.g.  Fig. 7).  The protein solubility curve (Fig. 7)  has two steps, due to  the  precipitation  of the serum albumin, followed by the  carboxymyoglobin.  Such perfect separation is rare, however, and it is more usual to obtain mixed fractions with, possibly, only  slight enrichment of a desired protein.  By altering the protein concentration it is sometimes possible to improve the separation, e.g. by diluting the solution, the points of precipitation (the peaks in the first derivative  curve) will be moved to the  right, to  higher saturation levels.  Due to  differences  in  KS, however, the  peaks due to different proteins might move to different extents and the  separation will thus be improved.

Figure 7.  Separation of human serum albumin and carboxymyoglobin by salting out .

To summarize, in salting out the following can be manipulated;-

• pH.  It is best to  use a pH below the  pI of the  desired protein  where

precipitation is maximal.

•  Temperature.  Theoretically, the  best  temperature  is  the  highest temperature  at which the  protein  is stable.  fl decreases as the temperature is increased and there is therefore greater precipitation at higher temperatures.  For  practical  purposes,  room  temperature (25C) is adequate.

•  Protein concentration.  The difficulty here is that the direction  of any effects cannot be predicted in advance.

The resolving power of salting out is not  high and so it is now commonly used mainly as a means  of concentrating  proteins  from  dilute extracts, while leaving non-protein molecules in solution.  It is also generally used early in an isolation,  immediately  after  preparation  of the extract.

3. Three-phase partitioning (TPP)

Three-phase partitioning (TPP)  is a method in which proteins  arc salted out from  a solution  containing  a mixture  of water and t-butano. t-Butanol is  infinitely  miscible with  water but upon  addition of sufficient  ammonium  sulfate  the  solution  splits  into  two  phases,  an underlying aqueous phase  and  an  overlying  t-butanol phase.  If  protein was present in the initial solution,  three  phases  would be formed,  protein being precipitated  in a third phase  between  the  aqueous  and  t-butanol phases (Fig. 8).  The  amount  and type  of protein  precipitated  is dependent upon the ammonium  sulfate concentration, as  in conventional salting out.  Unlike  in  conventional  salting  out,  however,  the  protein precipitate is largely dehydrated and has a low salt content.  Desalting before a subsequent ion-exchange step, which is a time-consuming necessity after conventional  salting out, is therefore  generally not necessary with TPP.

Conventional salting out is effected by adding (NH₄)₂SO₄ to a purely aqueous solution of protein.  In this  case the  protein  is initially hydrated and is thus soluble, and the  addition of the  salt, serves to  dehydrate  the protein and eventually brings it to its solubility limit.

Figure 8.  Three-phase partitioning.

In TPP,  t-butanol may be first added to the aqueous solution of protein to about 20%.  It is believed that  this  results in the  protein equilibrating with the  solvent (water) and the co-solvent (t-butanol). The protein thus becomes partially hydrated and partially t-butanolated, in proportion to the  relative  abundance  of the  solvents  in the mixture.

Upon addition of (NH₄)₂SO₄, water is abstracted by the  salt ions as these become hydrated.  The salt apparently has a higher affinity for water than for t-butanol, and thus preferentially sequesters the  water.  In the absence of protein,  this results in the  solution dividing into two phases, as some of the water is made “unavailable” to the t-butanol. If protein is present, the protein equilibrates with the new proportions  of solvent and co-solvent available to  it.  Upon  addition of further (NH₄)₂S0₄, eventually the  amount  of water available to  the  protein becomes insufficient to  keep  the  protein  in solution,  and it precipitates.

At this point,  however, the  protein  will be largely t-butanolated. This results in the protein having a reduced density and so, when it precipitates with the  situation  in conventional  salting  out  where the  dehydrated protein normally sinks, indicating that it is more dense than the  solution.

In conventional salting out, as the concentration of (NH₄)₂SO₄ increases, the solution density increases and the difference in density between the solution and the precipitate decreases, until the  point  is reached where it is no  longer possible to  sediment the  precipitate.  In  the  case  of  TPP, however, the precipitate  is  less dense than  the  solution  and so,  with increasing (NH₄)₂SO₄ concentration, the precipitate floats more and more easily.

In non-aqueous environments,  a-helices are favoured and so, during TPP, as the proportion of t-butanol increases, the protein conformation may become distorted as it acquires a greater proportion  of a-helices. This distortion leads to the denaturation of many proteins, which may be a disadvantage.  On  the  other  hand,  if  the  protein  of  interest  is  able  to survive TPP,  then  it is likely that TPP  will effect a purification by the denaturation  of impurities,  as well as by its fractionating  ability.

Selective denaturation of contaminating  protein has been put to  good effect in the isolation of cathepsin D10 and a number of erythrocyte proteins", where the major problem was the presence of a large excess of haemoglobin.

To effect a fractionation of a protein  mixture by TPP,  about 20% t

butanol can be added to  the  protein  mixture  in aqueous solution and increments of (NH₄)₂SO₄ are added, the  interfacial  precipitate  being concentrated by centrifugation and removed for analysis after each increment of (NH₄)₂SO₄.

Empirically, it has been found that, in TPP;-

•  Proteins precipitate  in  order  of  their  molecular  weight,  i.e.  larger proteins precipitate before smaller proteins,  present at the same concentration.

• Proteins  are most readily  precipitated  into  the  third  phase  when they

have a positive charge,

• Proteins  are  most  soluble,  after  TPP  (i.e. denaturation  is  minimal), when TPP is done at the pI of the protein.

• The  protein  concentration  has  a  marked  influence:  generally  the greater the concentration, the more easily the protein will precipitate.

• Temperature  has little effect, in the range 0C  to 25C.

It should be noted that t-butanol is unusual in that it is an organic solvent which tends not to denature proteins.  TPP can therefore be done at room temperature,  which is fortunate  because t-butanol solidifies at about 25C,  and is most conveniently used above this temperature.

References

Dennison, C. (2002). A guide to protein isolation . School of Molecular mid Cellular Biosciences, University of Natal . Kluwer Academic Publishers new york, Boston, Dordrecht, London, Moscow .