Powder Technology 197 (2010) 54-57

Contents lists available at ScienceDirect

Powder Technology
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p ow t e c

Crystallization of porcine insulin with carbon dioxide as acidifying agent

Gisele Atsuko Medeiros Hirata, Andr Bernardo, Everson Alves Miranda
LEBp Laboratrio de Engenharia de Bioprocessos, Departamento de Processos Biotecnolgicos, Faculdade de Engenharia Qumica, Universidade Estadual de Campinas, UNICAMP,
Caixa postal 6066, CEP 13083-970 Campinas, SP, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Recent studies on the use of volatile electrolytes such as CO2 in protein precipitation showed that these
Received 25 May 2009 agents are a promising alternative to the conventional acids. This use of volatile electrolytes prevents protein
Received in revised form 17 August 2009 denaturation due to local pH extremes, and saline efuent generation is greatly reduced, as the volatile
Accepted 22 August 2009
electrolyte may be separated and recovered from solution just by pressure release. In this work, insulin was
Available online 1 September 2009
successfully crystallized in the presence of zinc using CO2 as acidifying agent. The crystals obtained were
Keywords:
rhombohedral, a common shape for porcine insulin crystals that contain zinc in their structure, and their
Carbon dioxide average size varied with the mixing applied.
Downstream 2009 Elsevier B.V. All rights reserved.
Porcine insulin
Protein crystallization

1. Introduction dissolve and dissociate in water resulting in ions (carbonate, bicarbonate,

About 70% of the products sold by the process and pharmaceutical
industries are solid [1]. In the pharmaceutical industry crystallization
is an important technological process for particle formation and over
90% of all pharmaceutical products, such as tablets, aerosols, capsules,
and suspensions, generally contain drugs in crystalline form [2].
Protein crystallization is a technique optimized over many decades to
produce crystals with better conditions for X-ray diffraction analysis.
Its potential in protein recovery and purication has recently been
revealed. The crystal form of protein products has an advantage over
the dissolved and amorphous forms, since it has longer storage life and
higher purity [3].
Frequently isoelectric precipitation is used in downstream proces-
sing for separation of proteins from aqueous solutions. This method
consists of adjustment of solution pH to the isoelectric point of the
protein (pI). At the pI, the molecule carries no net electrical charge,
reducing the electrostatic repulsion between molecules which results in
precipitation. Since the majority of proteins have a pI in the acid region,
conventional acids such as H2SO4 and HCl are commonly used for pH
adjustment in these precipitations. After precipitation, the efuent
generated must be treated before its disposal into the environment, a
complex process step that can be costly.
The use of volatile electrolytes such as CO2 is an alternative to the use
of conventional acids in the isoelectric precipitation processes. They are
called volatile electrolytes since, as exemplied in the case of CO2, they
Corresponding author. Tel.: +55 19 3521 3918; fax: +55 19 3521 3890.
E-mail address: everson@feq.unicamp.br (E.A. Miranda).
and H+) whose concentrations are dependent on the temperature and
pressure of the system, causing a reduction in pH. Under depressurization
of the system the anions CO23 and HCO3 pass from the liquid phase to the
vapor phase as CO2 and can be removed from the system and recycled in
the process. Another advantage of the use of volatile electrolytes is that
during adjustment of protein solution pH, extremes in local pH are
prevented because solubilization of the precipitating agent is slow and
more homogeneous, occurring along the whole gasliquid interface. In
conventional processes, these extreme pH values can result in the
denaturation of protein or even a reduction in precipitate purity [4].
The use of volatile electrolytes in downstream processes of proteins
is relatively recent with only a few reports in the open literature. The
published studies on protein precipitation with volatile electrolytes
were carried out with complex protein systems such as milk, soy protein
extracts [5,6], and the fractionation of protein mixtures [7,8]. The iso-
electric precipitation with carbon dioxide of a single protein, bovine
serum albumin, was investigated by Qi et al. [9]; additional investiga-
tions of the activity of some enzymes in the system formed by ethanol,
carbon dioxide, and water were presented by Yao et al. [10]. Samadi and
Husson [11] described a process of separation by precipitation with CO2
for recovering L-aspartic acid. Tashima et al. [12] were pioneers in
reporting experimental and theoretical aspects of the precipitation of a
single protein (porcine insulin) using CO2 and developed a thermody-
namic model to correlate the experimental data. No crystal formation
was reported in that work.
The objective of this work was to nd conditions for the
crystallization of a single protein using CO2 as acidifying agent, and
we successfully crystallized porcine insulin in the presence of zinc in a
pressurized CO2 stirred tank crystallizer.

0032-5910/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.powtec.2009.08.017
 G.A.M. Hirata et al. / Powder Technology 197 (2010) 54-57 55

Fig. 1. Scheme of the experimental apparatus for the crystallization of proteins with CO2.

Fig. 2. Photomicrographs of porcine insulin crystals in 50 mM NaHCO3 and 0.4 mM ZnCl2 under pressurized CO2 at 5 C. Equilibrium pH runs: R1 (magnetic stirring), pH 6.50;
R2 (magnetic stirring), pH 6.50; R3 (impeller stirring), pH 6.37; R4 (impeller stirring), pH 6.35; and R5 (impeller stirring), pH 6.46.
56 G.A.M. Hirata et al. / Powder Technology 197 (2010) 54-57

2. Materials and methods biological microscope (model Tim-108, Opton, Brazil) with a magni-
cation from 40 to 1600.
2.1. Materials

Porcine insulin (96.6% pure, 0.55% Zn) was donated by Biobrs
(Brazil). Sodium bicarbonate (100% pure) was obtained from J. T. Baker
(Mexico); carbon dioxide (>99.8% pure), from White Martins Gases
Industriais (Brazil); and zinc chloride, from Vetec Qumica Fina (Brazil).
Ultra-pure water was obtained using Milli-Q equipment from Millipore
(USA). All the other reagents were of analytical grade.
2.2. Equipment
The equipment for protein crystallization with volatile electrolytes
(Fig. 1), was the same as that used by Tashima et al. [12]. The
equipment has a cylindrical pressure vessel in stainless steel with a
340 mL internal volume, a diameter of 5.5 cm and a thermal jacket for
liquid phase temperature control (accuracy of 1.0 C). The solution
pH was measured to an accuracy of 0.03 by a high-pressure glass
electrode TBX567 (ABB, Carlson City, USA). Pressure was measured to
an accuracy of 0.15 bar by the digital transducer PSI-420 (0 to
60 bar) and indicated on the instrumentation panel PLN-2 (both
devices by Zurich Indstria e Comrcio, Brazil). The solution could be
mixed by either a magnetic stirrer (3.8 cm long with a diameter of
1 cm) or an impeller with four paddles at 45 angles with agitation
rotational speed from 120 to 720 and from 0 to 300 rpm, respectively.
The impeller was located 2.2 cm from the bottom of the reactor and it
has a diameter of 2.7 cm. The magnetic bar did not come in contact
with the surface of the bottom of the reactor during stirring since it
was kept about 1 cm from the bottom of the reactor by a thin structure
in order to minimize crystal breakage.
3. Results and discussion
The attempts to crystallize insulin at 5 mg/mL were not successful in
generating crystals due to gelation of the insulin solutions. However, a
lower insulin concentration (1.84 mg/mL) produced relatively large and
well-formed crystals (Fig. 2). Although the crystal sizes obtained in the
experiments were different, the crystal shape was similar to that of the
rhombohedral zincinsulin crystals. This rhombohedral form was also
observed by Schlichtkrull [14], who reported that crystal shape is
dependent on the species of insulin. For example, porcine and human
insulin form rhombohedra, but sheep and bovine insulin appear to be a
twin formation.
The reduction in insulin concentration in the liquid phase with
time indicated the new phase formation a solid phase observed
after separation of precipitate and supernatant by centrifugation
(Fig. 3a). During the acidifying process, the CO2 is dissociated in
solution, causing the reduction in pH, decreasing the protein solubility
and favoring the crystallization processes (Fig. 3b). These concentra-
tion and pH proles are similar to the ones reported by Tashima et al.
[12]. In that paper, the precipitation of this same insulin did not lead to
protein denaturation. Thus, no loss of activity is expected in this case.
Crystal size varied with type of mixing system, magnetic stirrer or
impeller (Table 1). In the case of the magnetic stirrer, the crystal size
was smaller than it was in the case of the impeller (average sizes of

2.3. Experimental methods: porcine insulin crystallization

The protocol for crystallization of insulin used in this work was

based on a large-scale (commercial) animal insulin protocol in which
two temperatures 15 and 5 C were used. First, insulin was
dissolved in 100 mL of 10 mM HCl and then 125 mL of 100 mM
NaHCO3 and 25 mL of water were added to produce a solution with a
nal protein concentration of 1.84 mg/mL. The solution was ltered
with a 0.2 m syringe lter Minisart 16534 (Sartorius, Germany), and
71 l of 20% zinc chloride (1.47 M) was added to the solution. The
reactor vessel previously equilibrated at 15 C was charged with this
protein solution and closed. The solution was then kept under stirring
at 120 rpm for the runs carried out with the magnetic stirrer and at
65 rpm for the runs with the impeller until the solution temperature
reached 15 C. The pH was adjusted by slow addition of CO2. The CO2
feeding was interrupted when the desired pH was reached and the
pressure (maximum of 3.5 bar) was allowed to fall to 1 bar in 1 h by
CO2 release. The solution was then kept at 15 C for 4 h longer, the
stirring was stopped, and the temperature set to 5 C for 1940 h. The
reactor was then opened after the slow release of pressure, the liquid
phase and crystals were separated by centrifugation, and the size of
crystals and the insulin concentration in solution determined.

2.4. Analytical methods

2.4.1. Determination of the protein concentration

The protein concentration in the solution was determined by
absorbance at 280 and 320 nm according to Gehle and Schgerl [13]
using a DU 640 spectrophotometer (Beckman Instruments, USA).

2.4.2. Crystal analysis
The crystal size distribution was analyzed with quasi-elastic light
scattering Mastersizer S equipment (Malvern, UK) for the size range of
0.05 to 900 m. Crystal morphology was observed with a trinocular
Fig. 3. Crystallization of porcine insulin in 50 mM NaHCO3 and 0.4 mM ZnCl2 in a CO2
atmosphere at 5 C: a) insulin concentration and b) pH of solution as functions of time.
Equilibrium pH runs: , pH 6.50 (R1, magnetic stirring); , pH 6.50 (R2, magnetic
stirring); , pH 6.37 (R3, impeller stirring); , pH 6.35 (R4, impeller stirring); and , pH
6.46 (R5, impeller stirring).
 G.A.M. Hirata et al. / Powder Technology 197 (2010) 54-57 57

Table 1
Porcine insulin crystallization with CO2 in NaHCO3 solution.

Crystallization runs R1 R2 R3 R4 R5

Mixing system Magnetic stirrer Impeller

pH5 C 6.50 6.50 6.37 6.35 6.46

t5 C (h) 19 19 19 19 40
Average size (m) 7.59 0.34 10.56 0.34 16.73 0.43 14.80 0.43 19.58 0.81
10%a 1.36 5.31 5.50 5.50 3.48
90%b 11.78 16.43 28.75 28.23 41.72
Cf (mg/mL) 0.31 0.35 0.17 0.15 0.25
Yc (%) 83 81 91 92 86

Initial Zn concentration = 30 g/mL.

t5 C: time the solution was kept at 5 C.
pH5 C: nal pH at t5 C.
Cf: nal protein concentration.
a
10% of sample particles were below the average size.
b
90% of sample particles were below the average size.
c
calculated insulin recovery for initial protein concentration of 1.84 mg/mL.

7.59 to 10.56 m and 14.80 to 19.58 m, respectively). The smaller References

crystal size in case of the samples that were stirred with a magnetic
bar could be attributed to crystal breakage due to the larger rotation of
the magnetic bar. In addition, because of the shape the magnetic bar
favors frontal impact with the crystals as opposed to the pitched blade
impeller that favors a ow of crystals over the blade. Therefore, the
difference in crystal size for both systems can be thought of as a result
of different mixing regimes. The crystals that were ground during the
5 h mixing when the magnetic bar was used grew to a smaller size
after the 19 or 40 h aging.
4. Conclusions
This work showed that CO2 as acidifying agent is able to crystallize
porcine insulin in the presence of zinc, offering a new technology for
mass crystallization of proteins. The type of mixing produced different
crystal size distributions; crystals were rhombohedral under all the
conditions used in this work. Further work is needed to identify relevant
operational variables and to evaluate their effect on nucleation and
growth kinetics.
Acknowledgments
The authors thank Fundao de Ampara Pesquisa do Estado de
So Paulo (FAPESP), Coordenao de Aperfeioamento de Pessoal de
Nvel Superior (CAPES), and Conselho Nacional de Pesquisa Cientca
e Tecnolgica (CNPq) for their nancial support and Dr. Pedro de
Alcntara Pessa Filho (Departamento de Engenharia Qumica da
Universidade de So Paulo, Brazil) for reviewing the text.
[1] G.M. van Rosmalen, S. Bermingham, D. Bruinsma, H. Kramer, S. Derenzo, M. Seckler, I. R,
E. Cekinski, M. Giulietti, TUDelft IPT lectures on industrial crystallization and
precipitation, 2004.
[2] B. Yu Shekunov, P. York, Crystallization processes in pharmaceutical technology
and drug delivery design, J. Cryst. Growth 211 (2000) 122136.
[3] S. Schmidt, D. Havekost, K. Kaiser, J. Kauling, H.J. Henzler, Crystallization for the
downstream processing of proteins, Eng. Life Sci. 5 (2005) 273276.
[4] N. Khorshid, M.M. Hossain, M.M. Farid, Precipitation of food protein using high
pressure carbon dioxide, J. Food Eng. 79 (2007) 12141220.
[5] G.W. Hoand, M. Van Es, L. van der Wielen, G.J. Witkamp, Isoelectric precipitation
of casein using high pressure CO2, Ind. Eng. Chem. Res. 38 (1999) 49194927.
[6] G.W. Hoand, A. De Rijke, R. Thiering, G.J. Witkamp, L. van der Wielen, Isoelectric
precipitation of soybean protein using carbon dioxide as a volatile acid, J. Chrom. B
743 (2000) 357368.
[7] R. Thiering, G. Hoand, N. Foster, G.J. Witkamp, L. van der Wielen, Carbon dioxide
induced soybean protein precipitation: protein fractionation, particle aggregation,
and continuous operation, Biotechnol. Prog. 17 (2001) 513521.
[8] R. Thiering, G. Hoand, N. Foster, G.J. Witkamp, L. van der Wielen, Fractionation of
soybean proteins with pressurized carbon dioxide as a volatile electrolyte,
Biotechnol. Bioeng. 73 (2001) 111.
[9] X.M. Qi, S.J. Yao, Y.X. Guan, Novel isoelectric precipitation of proteins in a
pressurized carbon dioxidewaterethanol system, Biotechnol. Prog. 20 (2004)
11761182.
[10] SJ Yao, YX Guan, WZ He, XM Qi, ZQ Zhu, Activities of several enzymes in ethanol+water
at elevated pressure of carbon dioxide, J. Chem. Eng. Data 49 (2004) 13331339.
[11] A. Samadi, S.M. Husson, Towards the recovery of L-aspartic acid from aqueous
solution: exploiting pH-dependent solubility via carbon dioxide addition, Sep. Sci.
Technol. 38 (2003) 28372850.
[12] AK Tashima, M Ottens, L van der Wielen, DE Cyntra, JR Pauli, PA Pessa Filho, EA
Miranda, Precipitation of porcine insulin with carbon dioxide, Biotechnol. Bioeng.
103 (2009) 909919.
[13] D. Gehle, E. Schgerl, Protein recovery by continuous otation, Appl. Microbiol.
Biot. 20 (1984) 133138.
[14] J. Schlichtkrull, Insulin Crystals: II. Shape of rhombohedral Zninsulin crystals in relation
to species and crystallization media, Acta Chem. Scand. 10 (1956) 14591464.