Biotechnol. Prog. 1990, 6,283-285
283
Precipitation of Nucleic Acids with Poly(ethy1eneimine) Richard M. Cordes, W. Bradley Sims, and Charles E. Glatz* Department of Chemical Engineering, 231 Sweeney Hall, Iowa State University, Ames, Iowa 50011
Removal of nucleic acids from cell extracts is a common early step in downstream processing for protein recovery. We report on the precipitation of nucleic acids from a homogenate of Saccharomyces cereuisiae by addition of the cationic polyelectrolyte poly(ethy1eneimine) (PEI), focusing on the effect of PEI dosage on particle size, protein loss, and extent of nucleic acid removal in both batch and continuous mode. Better than 95 % removal of nucleic acids from yeast homogenates was achieved by means of precipitation with PEI with protein losses of approximately 15% with or without previous removal of cell debris. The coprecipitated protein is predominately large molecular weight material and exhibits both low and high isoelectric points. Such treatment does not aggregate the cell debris; size distribution of the precipitated particles from a continuous precipitator is very similar to that for protein precipitation.
Introduction Removal of nucleic acids from cell extracts is a common early step in downstream processing for protein recovery. Such a step reduces viscosity of the extract and eliminates the interference of nucleic acids (NA) in subsequent separation steps. The procedure is also useful in production of nucleic acids and in the reduction of nucleic acid content of single cell protein. Removal methods include enzymatic and chemical treatments. Precipitation with polyelectrolytes has been investigated for Escherichia coli (Atkinson and Jack, 1973; Spitnik et al., 1955; Agerkvist and Enfors, 1989) and wheat germ (Jendrisak, 1987) and shown to compare favorably with other methods, on the basis of comparisons of extent of removal, protein loss, chemical requirements, and safety. The small amounts of polyelectrolytes required make them economically attractive precipitants. In these studies, little attention has been paid to the properties of the precipitate, which are important in the removal of the precipitate from the remaining extract. Here we report on the precipitation of nucleic acids from a homogenate of Saccharomyces cerevisiae by addition of the cationic polyelectrolyte poly(ethy1eneimine)(PEI), focusing on the effect of PEI dosage on particle size, protein loss, and extent of nucleic acid removal in both batch and continuous mode. We also find that the results are not greatly affected by the presence of cell debris but that PEI addition fails to flocculate the debris.
Experimental Procedures Compressed yeast (Red S t a r , Universal Foods, Milwaukee, WI) was suspended in chilled (4 "C) buffer (0.15 M NaCl and 4 mM KzHP04) in the ratio of 227 g of yeast/L of buffer (Hetherington et al., 1971) and ruptured by two passes through a Manton Gaulin 15T8TM lab homogenizer a t 55 MPa. Cell debris and unbroken yeast cells were removed by centrifugation (20000g, 4 OC, 60 min). Since the amount of extracted material increased with the pH of the homogenate and precipitations from pH 5.5 to 8 were to be studied, the pH
* Author to whom correspondence should be addressed.
was adjusted to 5.5 after homogenization to ensure the same starting material for all experiments. Under these conditions, typical concentrations of nucleic acids and protein were 1.54 and 12.6 mg/mL, respectively. Before the precipitation step, any floating lipid was removed and the pH was adjusted to the desired level (pH 6 unless otherwise noted) by addition of NaOH. PEI (Aldrich Chemical Co., Milwaukee, WI) was added as a 5 72 (w/v) solution at the precipitation pH to 400 mL of filtered cell extract in a baffled, stirred (200 rpm, estimated mean shear 40 s-l) vessel (Brown and Glatz, 1987). Some runs were made with 825 mL in a larger baffled vessel with estimated mean shear of 250 s-l. After 15 min, aliquots of the suspension were removed for assay. One sample was diluted with filtered buffer (0.1 M phosphate and 0.14 M NaCl) at the precipitation pH and ionic strength before measuring the particle size distribution on a Coulter Counter (Model TAII with population accessory, Coulter Electronics, Hialeah, FL). Another sample (100 mL) of the suspension was centrifuged (20000g) before the supernatant was assayed. The precipitate from this sample was resuspended in 100 mL of homogenate buffer, recentrifuged, and freeze-dried. The hot and cold perchloric acid extraction procedure of Herbert et al. (1971) was used to separate proteins and high molecular weight nucleic acids for assay. The RNA content was then determined by the orcinol method (Herbert et al., 1971) and the protein content by the biuret method (Doumas, 1975) after the protein-containing precipitates were dissolved in 1 M NaOH. For SDS-polyacrylamide gel electrophoresis (Laemmli, 1970) and isoelectric focusing (LKB, 1983),sample buffers were changed by dialysis and/or dilution as required. { potential measurements were made by using a ZetaMeter System 3.0 (Zeta-Meter Inc., Long Island, NY) with precipitate particles suspended in phosphate buffer.
Results and Discussion Figure 1demonstrates the selective removal of nucleic acids and the extent of coprecipitation of protein upon addition of PEI for experiments using three different
8756-7938/90/3006-0283$02.50/0 0 1990 American Chemical Society and American Institute of Chemical Engineers
Biotechnol. Prog., 1990, Vol. 6, No. 4
204
Table I. Analysis of Variances of Responsesn
100
PH
- 90 - BO -70
6\0 v
-60
-50 -40
- 30 ___D
"
T
0.0
I
I
I
I
I
1
1
0.5
1.0
1.5
2.0
2.5
3.0
3.5
2!$ E
response
dfb
F
nucleic acid removed protein removed particle mean size particulate RNA fraction
2 2 2 2
5.95 0.28 2.30 0.35
Pr 0.026
df
0.760 0.163 0.751
PEI dosage F Pr 0.35 0.31 0.00
1 1 1 1
0.68
0.572 0.595 0.978 0.434
F is the calculated ratio of treatment to experimental variances. The analysis gives the probability (Pr, based on the normal F distribution) that a variable (pH of 5.5, 6.0, 6.5; PEI dose of 1.34, 1.88 mg/g of yeast) does not affect the indicated response (low Pr implies a significant effect). A range of variables away from this (1
optimum region is presented in Figures 1 and 2.
- 20
1001
- 10
no -
I "
80 -
4.0
PEI CONCENTRATION (mg Pfvg yeast)
0
Figure 1. Effect of poly(ethy1eneimine)dosage on nucleic acid and protein removal. Precipitationswere carried out at pH 6
w
under the following conditions: 200 mL, centrifuge bottle, magnetic stirrer, X+; 400 mL, baffled vessel, 0;400 mL, baffled vessel in presence of cell debris, 0;825 mL, baffled vessel, A;825 mL, baffled vessel, continuous, V. All precipitations were batch, impeller stirred, and after cell debris removal unless otherwise noted.
70-
P
$
60-
e0
50-
-Z
40-
ii
0 303
z
vessel/agitation configurations as well as both batch and continuous modes. The selectivity and extent of removal as the plateau (97% NA removal) is approached are quite comparable to those reported for E. coli (Atkinson and Jack, 1973). The initial linear removal shows that RNA is essentially titrated from solution by complexation with PEI. A mass balance, assuming all the dry weight of the precipitate not accounted for by protein or nucleic acid to be PEI, gives an upper limit of 15-2092 of the added PEI remaining in solution in the linear region. A balance on the charged groups provided by PEI (assumed to be 63% protonated at pH 6 from the titration curve) and nucleic acids (assumed charge of -1 per nucleotide residue) shows a net negative charge that is probably offset by the coprecipitated proteins. Isoelectric focusing of the redissolved precipitate showed a range of protein components including a major band with p l of 3.5, which would be negatively charged a t the precipitation pH. Bloomfield et al. (1980) reported DNA condensation to occur at 90% of charge neutralization, while we estimate the PEI content of the precipitate to account for 80% neutralization. Measurement of particle l potential revealed no significant net charge on the precipitates. The only other observation made on the nature of the coprecipitated proteins is that SDS-gel electrophoresis showed that most of the high molecular weight proteins were precipitated. The data of Figure 1include a range in the initial yeast concentration from 60 to 227 g/L of buffer, but the results are independent of concentration when PEI doses are expressed relative to yeast concentration. Increasing pH from 6 to 8, however, did reduce the level [statements regarding significant differences are based on the analysis of variance (SAS User's Guide, 1985);results displayed in Table I] of RNA removal a t doses up to 1.50 mg/g of yeast (Figure 2), but there was no effect from pH 5.5 to 6.5. Composition of that precipitate which did form was unchanged. Titration of the PEI showed that the degree of protonation of the imine groups falls from 63 96 at pH 6 t o 48% a t p H 8, which would explain t h e lower
201 10
//
OY
0.00
0.25
0.50
0.75
1.W
1.25
1 io
PEI CONCENTRATION (mg PEI/g yeast)
Figure 2. Effect of pH on nucleic acid removal.
effectiveness. From pH 5.5 to 6.5, the percent dissociation only changes from 67 % to 5996, a change whose effect may not have been detectable. Carrying out precipitations a t ionic strengths of 0.15, 0.30, and 0.75 showed that RNA removal was not hampered except at the highest level, where yield was reduced from 95 % to 83 % a t 1.34 mg of PEI/g of yeast. Table I1 shows that the higher level of agitation in the larger precipitation vessel may have increased t h e selectivity for nucleic acid'removal. Better reagent dispersion has been reported to improve selectivity in some cases of protein fractionation by precipitation (Fisher et al., 1986). Table I1 also shows that changes in vessel conditions have a greater effect on particle size. Shear is known to limit particle size, and the higher shear of the larger vessel resulted in a significant reduction in particle size. However, in this case, even the smaller particle size material settled well. Figure 3 illustrates that, while mean sizes of particles from batch and continuous precipitations are similar, the size distributions are quite different. Since the size distribution and morphology (as viewed by scanning electron microscope) of these nucleic acid/ protein coprecipitates are similar to those for protein precipitation, it is to be expected that the mathematical model developed for the latter (Glatz et al., 1986) would be able to fit these results as well. We were also interested in the influence of cell debris, since it would be convenient to combine the solid/liquid separation steps to remove both cell debris and precipitates in a single step. The PEI might also act as a flocculant for the cell debris by interacting with phosphate groups of the cell wall mannoproteins (Shaeiwitz et al., 1989), as
Biotechnol. Prog., 1990, Vol. 6, No. 4
285
Table 11. Mixing Effects on Nucleic Acid and Protein Removal and Particle Size. % removal (SD) vessel nucleic acid protein mean size, pm 400 mLb 94.3(2.0) 18.9(1.0) 10.76 825 mLc 97.2(2.2) 13.4(1.2) 4.60 825 mLd 4.60 825 mLe 97.8(2.5) 14.5(1.2) 5.11 All data are for a dosage of 1.36 mg of PEI/g of compressed yeast. Batch; shear, 250 s-l; 15 min. Batch; shear, 250 s-l; 12.5 min. e Continuous; shear, 250 s-l; 12.2 min mean residence time. a
* Batch; shear, 40 s-l; 15min.
tated protein is predominately large molecular weight material and exhibits both low and high isoelectric points. Such treatment does not aggregate the cell debris; size distribution of the precipitated particles from a continuous precipitate is very similar to that for protein precipitation.
Acknowledgment We are gratefu!. to the National Science Foundation (Grant ECE-8514865) for support of this work. Partial support for R.M.C. was also provided by the Procter and Gamble Co. and the Engineering Research Institute of Iowa State University.
"Literature Cited
20 -
E
wz
W
P K 15-
3
3z 3
I 0
CONTINUOUS
I
10
15
5
PARTICLE DIAMETER (MICRONS)
Figure 3. Comparison of particle size distributions of batch and continuous precipitations of nucleic acids with cell debris. Both precipitations were carried out a t a mean shear rate of 250 s-1; 1.36 mg of PEI/g of compressed yeast; pH 6; 825 mL; baffled vessel.
has been reported with the use of chitosan (Agerkvist and Enfors, 1989) with E. coli homogenates. The elimination of the cell debris removal step had little or no effect on nucleic acid and protein removal (Figure 1). However, judging from particle size measurements, PEI addition did not result in any flocculation of cell debris. The latter failing, in contrast to the flocculating ability of chitosan, particularly with E. coli, is likely to be the result of the relatively low molecular weight (50 000) and branched structure of PEI, as there have been preliminary reports of yeast debris flocculation by using calcium and chitosan (Siika-aho et al., 1989).
Conclusions Better than 95 % removal of nucleic acids from yeast homogenates can be achieved by means of precipitation with PEI with protein losses of approximately 15% with or without previous removal of cell debris. The coprecipi-
Agerkvist, I.; Enfors, S.-0. Flocculation as a Separation Tool in Primary Recovery Stages. Presented a t the International IUPAC Congress, Stockholm, Sweden, August 1989;Enzyme Microb. Technol. 1989, in press. Atkinson, A,; Jack, G. W. Precipitation of Nucleic Acids with Polyethyleneimine and the Chromatography of Nucleic Acids on Immobilized Polyethyleneimine. Biochim.Biophys.Acta 1973, 308,41-52. Bloomfield, V. A.; Wilson, R. W.; Rau, D. C. Polyelectrolyte Effects in DNA Condensation by Polyamines. Biophys. Chem. 1980,11, 339-343. Brown, D. L.; Glatz, C. E. Aggregate Breakage in Protein Precipitation. Chem. Eng. Sci. 1987, 42, 1831-1839. Doumas, B. Protein Standards for Total Serum Protein Assays-A Collaborative Study. Clin. Chem. 1975,21, 1159-1166. Fisher, R. R.; Glatz, C. E.; Murphy, P. A. Effects of Mixing During Acid Addition on Fractionally Precipitated Protein. Biotechnol. Bioeng. 1986,28, 1056-1063. Glatz, C. E.; Hoare, M.; Landa-Vertiz, J. The Formation and Growth of Protein Precipitates in a Continuous Stirred Tank Reactor. AZChE J . 1986,32, 1196-1204. Herbert, D.; Phipps, P. J.; Strange, R. E. In Methods in Microbiology;Norris, J. R., Ribbons, D. W., Eds.; Academic Press: New York, 1971; Vol. 5B, pp 209-344. Hetherington, P. J.; Follows, M.; Dunnill, P.; Lilly, M. D. Release of Protein from Baker's Yeast by Disruption in an Industrial Homogenizer. Trans. Znst. Chem. Eng. 1971, 49, 142-148. Jendrisak, J. In Protein Purification: Micro t o Macro;Burgess, R., Ed.; Alan R. Liss, Inc.: New York, 1987; pp 75-97. Laemmli, U. K. Cleavage of Structural Proteins During the Assembly of the Head of Bacteriophage T4. Nature 1970,227, 680-685. LKB Ultramould Instruction Brochure; LKB-Produkter AB: Bromma, Sweden, 1983. Shaeiwitz, J. A.; Blair, J. B.; Ruaan, R.-C. Evidence that Yeast Cell Wall Debris Can Separate Proteins by Ion-Exchange During Cell Lysis. Biotechnol. Bioeng. 1989, 34, 137-140. Siika-aho, M.; Pavas, P.; Ojamo, H. Pilot Scale Purification of Yeast Homogenate Using Chitosan Flocculation. Presented at the 32nd IUPAC Congress, Stockholm, Sweden, August 1989. Spitnik, P.; Lipshitz, R.; Chargaff, E. Studies on Nucleoproteins: 111. Deoxyribonucleic Acid Complexes with Basic Polyelectrolytes and Their Fractional Extraction. J . Biol. Chem. 1955,215, 765-775. Accepted June 14, 1990. Registry No. P E I (homopolymer), 9002-98-6; PEI (SRU), 26913-06-4.
