NASA Technical Reports Server (NTRS) 19980000098: Thermal Optimization of Growth and Quality in Protein Crystals

18 Pages·1.2 MB·English

by  NASA Technical Reports Server (NTRS)

/¢/'//"_'_,/_/ 2' 7 -- zo6o9s
Thermal Optimization of Growth and Quality in Protein Crystals
NASA Grant Number NAG8-1159
FINAL REPORT
Executive Summary
Experimental evidence suggests that larger and higher quality crystals can be attained in the microgravity
of space; however, the effect of growth rate on protein crystal quality is not well documented. This research is the
first step towards providing strategies to grow crystals under constant rates of growth. Controlling growth rates at a
constant value allows for direct one-to-one comparison of results obtained in microgravity and on earth. The overall
goal of the project was to control supersaturation at a constant value during protein crystal growth by varying
temperature in a predetermined manner. Applying appropriate theory requires knowledge of specific
physicochemical properties of the protein solution including the effect of supersaturation on growth rates and the
effect of temperature on protein solubility. Such measurements typically require gram quantities of protein and
many months of data acquisition. A second goal of the project applied microcalorimetry for the rapid determination
of these physicochemical properties using a minimum amount of protein. These two goals were successfully
implemented on hen egg-white lysozyme. Results of these studies are described in the attached reprints.
Reprinted from
JOURNOAFCL TAL
GROWTH
Journal of Crystal Growth 165(1996) 293-298
Enthalpy of crystallization of hen egg-white lysozyme
Constance A. Schall a.1 Edward Arnold b John M. Wiencek c,,
" Department of Chemical and Biochemical Engineering. Rutgers Unicersity. P.O. Box 909. Piscataway New Jersey 08855-090, USA
bCenter for Adcanced Biotechnology and Medicine andChemistry Department, Rutgers Unil'ersity, 679 Hoes Lane. Piscataway, New
Jersey 08854-563. USA
¢Chemical and Biochemical Engineering. Unir-ersity of Uh(to. 137 Chemirtry Building Iowa City, hm'a 52242-12 I, USA
Received 20 April 1995: accepted 15January 1996
ELSEVIER
......... CRYSTAL
GROWTH
ELSEVIER Journal of Crystal Growth 165(1996) 293- 298
of Utah
Enthalpy of crystallization of hen egg-white lysozyme
Constance A. Schall a._ Edward Arnold b, John M. Wiencek c.,
"'Delmrtment ,f Chemical and Biochemical Engineering Rutgers Umt erslty. P.O. Box 909. Piscataway. New Jerse_ 0_855-090. USA
th.
Center fi_r Adt anced Biotechnology and Medicine. andChemistry Department. Rutgers L.'nh ersity. 679 Hoes l_me. Piscatawa 3, New
st.of Technol. Jersey 08854-563. USA
Chemical and Biochemical Engineering. Unit ersio" of Uhao. 137Chemistry Building. Io_ta City. Iowa 52242-12 I.USA
Received 20April 1995: accepted 15January 1996
Abstract
nduc ripe,
The validity of the van "t Hoff calculation of crystallization enthalpy from the temperature dependence of solubility was
proven for hen egg-white lysozyme. The enthalpy of crystallization of hen egg-white lysozyme in two 0.05M acetate buffers
(5% NaCI, pH = 4.6 and 3% NaC1, pH = 5.2) was determined at 15°C using isothermal calorimetry and was found to be
fpita_. Devices) - 17.1 _+3.2 kcal/mol (5% NaCI) and - 10.5 _+2.3 kcal/mol (3% NaCI). These values were found to agree within
655936. MS 147
experimental error with the enthalpy of crystallization determined from a van 't Hoff plot of solubility.
;nd Minerals)
a-r.hi
1. Introduction recently, DeLucas and coworkers [2] and Wiencek
mMicro:_copy. and coworkers [3] have employed temperature con-
In crystallization, control of the level of supersat- trol to protein crystallization in batch microdrop
uration throughout the crystallization process is es- systems. DeLucas [2] utilizes an active feedback (i.e.
sential if the crystal size is to be optimized. In light scattering) to determine the onset of nucleation
_yslals)
,Massachusetts protein c_'stallization the supersaturation is normally and then relies on scanning a variety of temperature
MA 02139. USA
controlled by use of precipitants or pH. An alternate profiles to arrive at an optimum temperature control
approach to optimize crystallization conditions is algorithm. Wiencek [3] characterizes key features of
rtaxy))
1st. ofTechnol. through temperature control, For inorganic crystal- the system including critical nucleation temperatures,
lization, controlled cooling can result in fewer nuclei solubility and growth rates to arrive at a predeter-
and larger crystals formed in batches under tempera- mined temperature control algorithm. Both ap-
,_,0b0le22-0f2r4o8m) thearepubslcishheedPu_led upofnor ture control designed to allow for control of the proaches benefit tremendously from an accurate
only. Issues are sent by SAL nucleation rate [I]. Such control strategies require knowledge of the temperature dependence of the
Die. Airmail rates arcavailable
and subscriptions to: knowledge of the solubility as a function of tempera- protein's solubility.
ture, nucleation kinetics and growth kinetics. More In experiments reported here, calorimetry is uti-
lized to obtain phase diagram inlormation for the
,six monthss of our publi¢aliou tetragonal crystal form of a model protein, hen egg-
• Corresponding author.
white lysozyme. Utilizing the van "t Hoff relation-
ICurrent address: University of Tulsa Chemical Engineering
,E Amr.terdam. The Netherlands. Depa_ment. 600 South College. Tulsa. Oklahoma 74104-3189. ship. a single point measurement of the equilibrium
,1atJamai¢a NY 11431.
_ght and mailing in the USA by USA. concentration of protein and the enthalpy of crystal-
NTEDINTHENETHERLANDS. 0022-0248/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved
PII S0022-0248(96)00180.7
294 C.A. Schall et al. /Journal of Crystal Growth 165 (1996; 293-298
lization will yield the local temperature dependency temperature control methods of DeLucas and the bu
of the solubility of protein in solution. In turn, this coworkers [2] and Wiencek and coworkers [3] are Stock
infi)rmation c:m be used to develop strategies tbr only usefid it" the candidate protein's solubility is
using .
controlling the crystallization process using tempera- sensitive to temperature. Isothermal analysis of tile trator
ture as the control parameter. heat of crystallization will lead to a direct measure- with a
It is recognized that such information may be ment of this sensitivity to temperature as well as the obtain
obtained via solubility measurements at two distinct nature of the dependence (i.e. normal or retrograde
kysoz)
temperatures (lbllowed by van "t Hoff analysis) in a solubility behavior). Thus, the potential value of extinct
fairly simple microdevice such as the device devel- microcalorimetry in the study of protein crystalliza- 280 m:
oped by Pusey and coworkers [4,5]. An accurate tion is becoming apparent although reported utiliza-
0.2 _1
estimate of solubility requires not only growth of a tion of the technique is still minimal.
lizing ,
crystal from supersaturated solution but also dissolu- An important first step is to justify the use of the
papers
tion of crystals into an undersaturated solution. Con- van "t Hoff analysis for describing the temperature NaCI ,
vergence between the two methods (i.e. dissolution dependence of protein solubility. Although one would pH of
and crystallization) can take months by conventional not normally question such an analysis, previous betwee
methods. Pusey and coworkers [4,5] devised a micro- calorimetric studies have cast some doubt on the To
volume apparatus which utilizes a small volume of validity of such an analysis. Howard and coworkers
hen eg
crystallites in a semi-batch configuration to obtain [8] found that their solubility measurements yielded
tempe/
estimates for solubility. The method requires some van "t Hoff enthalpy of crystallization that varied tions: ,_
preknowledge of solubility since both under- and significantly from the values measured calorimetri- 5.2, b,
over-saturated solutions are required for this device. cally by Takizawa and Hayashi [9]. Since the Howard
egg-,a i
In order to ascertain whether a solution is under- or et al. solubility data [8] were not measured at the water.
over-saturated requires advanced knowledge of the same pH as the calorimetric data of Takizawa and acetate
solubility which the experiment intends to measure. Hayashi [9], the comparison of A/4_.,t required buffer
However, after one solubility measurement is com- some interpolation by Howard et al. [8]. In order to (Pha_
pleted, systematic studies away from that point be- alleviate any questions arising from methods of in- v,'ith 0
come trivial. Although the design of Pusey and terpolation or thermodynamic analysis, the present buffer.
coworkers [4,5] greatly accelerates the determination work measures AHcryst under a solubility condition with a[-
of solubilities, there are systems which will display for lysozyme as reported by Howard eta[. [8] (i.e. tion bu
very slow rates of dissolution and prevent accurate pH = 4.6, 5% NaCI). In addition, a second condition
The pl-
measurements in a reasonable timeframe. In such (i.e. pH 5.2, 3% NaCI) was tested to show the meter
cases, the use of isothermal calorimetry, requiring on general utility of the method, A series of measure- the 5fi
the order of 10 to 50 h to complete an experiment, ments were made using isothermal calorimetry to tion. F!
may prove advantageous. In any case, the present determine the enthalpy of crystallization of hen egg- ml of
study certainly justifies the use of van 't Hoff analy- white lysozyme in 5% NaCI and 0.05M acetate concel_
sis in protein systems, at least over a local tempera- buffer, final pH 4.6 and in 3% NaCI, 0.05M acetate constai
ture range. buffer, final pH 5.2 at 15°C. Solubility over a range _+0.01
There are many ways that microcalorimetry may of temperatures were determined for both of these then s_t
be used in the investigation of the protein crystalliza- conditions, with the pH 4.6, 3% solubility data being were t_
tion process. For example, recent work by Sibille used in addition to the data of Howard et al. [8] the so
and Pusey [6] has focused on the use of stopped-flow
change
techniques to assay the heat of binding between were t,
lysozyme and sodium chloride ions. In addition, 2. Materials and methods 5.2-5.:
differential scanning microcalorimetry has been uti- The
lized to detect the onset of nucleation [7] of lysozyme Hen egg-white lysozyme was purchased from
lysoz3 :
as a function of solution variables such as protein Sigma Chemicals (catalog number L-6876, lot num-
tivity )
and salt concentration. Such studies help define the ber 89F8276). The lysozyme was first dialyzed facture_
[abi[e/metastable boundary for protein systems. The against deionized water followed by dialysis against nolog?
('.A, Schall et al. / Journal of Crystal Growth 165 (/996J 293- 29A'
_sand
the buffer solution (0.05M sodium acetate, pH 4.5).
[3]are
Stock solutions of lysozyme were then concentrated
,i[ity is using a 5000 molecular weight cut-off microconcen-
of the trator (Amicon). Protein stock solution was mixed
o I
easure- with appropriate quantities of 10% NaCI in buffer to
! as the obtain a final salt concentration of 5% or 3%.
ograde Lysozymc concentrations were determined using an
L[ue of extinction coefficient of 26.35 for a Ic_csolution at
,mlliza-
280 nm [I0]. All stock solutions were passed through
utiliza-
0.2 txm s>ringe filters. The pH values of the crystal-
lizing solutions were measured using Baker PHix pH #,
of the 0 I000 200t) t0l) 0 41) I)0
papers (3.8 to 5.5). The final pH values of the 5%
,erature NaCI solutions were found to be 4.6, and the final TIME (MIN)
' would pH of the 3% NaCI solutions were found to lie Fig. I. The measured ,dgnal from the TAM calorimeter for a
revious between 5.2 and 5.5. crystallizing solution of lysozyme in 3ok NaCI. 0,05M sodium
on the acetate buffer, pH = 5.3 at 15°C,
To supplement literature data [8], the solubility of
_orkers
hen egg-white lysozyme was determined for several
_,ielded
temperatures and the following precipitant condi- ence cells were both 3 ml capacity stainless steel
,,aried
tions: 5% NaC1 with pH 4.5, and 3% NaCI with pH screw top ampoules. The reference cell was filled
imetri- 5.2, both in a 0.05M sodium acetate buffer. Hen with 3 ml of distilled water. The sample cell was
toward
egg-white lysozyme was dialyzed against deionized filled with 3 ml of a lysozyme and sodium chloride
at the
water, followed by a buffer solution of 0.05M sodium mixture (the crystallizing solution). The calorimeter
va and
acetate, pH 4.5 or 5.2. Solutions dialyzed against was operated isothermally at 15°C. The measurement
:quired buffer with pH 5.2 were also passed through a PD-10 recorded by the calorimeter was power (in g.W)
rder to
(Pharmacia LKB Biotechnology) desalting column versus time as shown in Fig. I. The total enthalpy
of in-
with 0.05M sodium acetate, pH 5.2. as the exchange change was calculated by summing the area under
,resent
buffer. The protein solutions were then combined the power versus time curve. Areas were calculated
ndition
with appropriate quantities of 10c_cNaC1 stock solu- using the Thermometric software, correcting for the
S] (i.e.
tion buffered at a pH matching the protein solution. initial baseline and baseline drift. The quantity of
ndition
The pH of the final mixture was measured with a pH lysozyme crystallized was assumed to be the differ-
ow the meter (Orion, model 720A) and found to be 4.69 for ence between the initial quantity of lysozyme in
easure- the 5% salt solution and 5.31 for the 3% salt solu- solution minus the quantity of lysozyme in solution
etry to
tion. Five 1.5 ml Eppendorf tubes were filled with I at the conclusion of the calorimetry experiment.
-'n egg-
ml of solution for each temperature and precipitant Lysozyme crystals formed primarily on the walls and
acetate
concentration. The tubes were then immersed in the bottom of the ampoule and were examined with
acetate
constant temperature baths (VWR model 1187, an optical microscope. The crystals exhibited a mor-
,t range +0.01°C) for 11 to 14 weeks. The supematant was phology typical of the tetragonal tbrm of lysozyme.
f these
then sampled and residual lysozyme concentrations
a being
were taken as the equilibrium solubility. The pH of
[81
the solutions measured with pH paper were un- 3. Results
changed at the time of solubility determination and
were found to be 4.6 for the 5% salt solutions and A summary of the experimentally determined
5.2-5.5 for the 3% salt solutions. lysozyme solubility values is given in Table 1. The
The heat signal from a crystallizing solution of data collected at 3% NaCl, pH 5.2 are not available
d from
lysozyme was recorded using the 2277 Thermal Ac- in the current literature and represent an independent
_t num- tivity Monitor (TAM) isothermal calorimeter manu- test of the validity of the van "t Hoff analysis of the
!ialyzed
factured by Thermometric (Allied Chemical Tech- data. The data collected at 5% NaCI, pH 4.6 are
against
nology, Burtonsville, MD). The measuring and refer- intended to supplement the data of Howard et al. [8].
296 C.A. Schall etal./ Journal t!f'Crx'stal Growth 165 (Iq96 J293 29_¢
Table I r 1 r r
Solubility of hen egg-while lysozyme (average of five samples at
each condition)
%
Temperature Solubilic2
(°C) (mg/ml) =.
3'7, NaCI. pH = 5.2 5 2.55 4`0.05 E
0 3.83 4`_).05
5 4.35 +0.06
5¢4-NaCI. pH = 4.6 5 1.42 +O.O3
10 1.93 +0.18 e..
N
i i "
: i
Since the isothermal instrument used in this study 0
was a shared instrument, these studies were limited 3 3.4 3.5 6
to 15°C. The data of Howard et al. [8] provided only lIT x 1000 (I/K)
two solubility data points (15 and 20°C) below the
Fig. 2. Van 't Hoff plot of lysozyme equilibrium solubili b as a
tetragonal to orthorhombic phase transition. These
function of temperature. 3c7cNaCI. O.05M sodium acetate, pH =
data points were not centered around the calorimetric 5,2 ( × )from Table 1: 5q NaC1.0.05M sodium acetate, pH = 4.6
measurement temperature and were, thus, supple- (O) from Table I: 5% NaCI. 0.05M sodium acetate, pit = 4.5
mented with the data of Table I. The two sets of (C)) from Howard et al. [8]. The slope of the solid line corre-
sponds to an enthalpy of crystallization of -8.1 4-2.7 kcal/mol
solubility data were plotted in the typical van 't Hoff
for 3q- NaCI (r = 0.9482). This compares vdth a ,.alue of - 10.5
fashion and yielded fairly straight lines as shown in
4`2.3 kcal/mol determined by calorimetry, The slope of the
Fig. 2. dashed line corresponds to an enthalpy of crystallization of - 15.I
Calorimetric results for four experiments with 4. 1.3 kcal/mol for 5"'_ NaC1 (r= 0.9894}. This comparc_, v,ith a
crystallization conditions of 5% NaC1, and final pH value of - 17.I _3.2 kcal/mol determined by calorimetry.
of 4.6 are shown in Table 2. The enthalpy of crystal-
lization for lysozyme at 15°C based on these four
experiments is - 17.1 4-3.2 kcal/mol. A z-test does tion conditions of 3% NaCI, and final pH of 5,2 is
not indicate a significant difference between the shown in Table 2. An example of the measured
enthalpy of crystallization measured calorimetrically signal from the TAM calorimeter is presented in Fig.
and the value of -15.1 + 1.3 kcal/mol regressed 1. The enthalpy of crystallization for lysozyme at
from a van 't Hoff plot of solubility versus tempera- 15°C based on the four TAM experiments is - 10.5
ture (Fig. 2). + 2.3 kcal/mol. A z-test does not indicate a signifi-
A summary of four experiments with crystalliza- cant difference between the enthalpy of crystalliza-
Table 2
Enthalpy of crystallization of lysozyme measured by isothermal calorimetry' (all solutions ',,,,ere made using O.f)5M sodium acetate buffer at
the indicated salt concentration and final pH
Buffer Initial protein Final protein - &H,,_,, Number of days
conditions conc. (mg/ml'l conc. (mg/ml) (kcal/mol) in calorimeter
5_ NaCI. pH = 4.6 41.0 +_0.8 9.8 _.+0.2 14.8 3
36.2 + 1.0 14.5 + 1.2 15.8 4
26.6 +_0.4 8.0 +_0.2 16.I 5
41.3 ± 1.l 8.6 4-0.4 21.8 4
3% NaCI. pH = 5.2 78.3 4` 1.6 46.5 ::k1.8 10.8 2
77.2 + 1.4 33.2 + 0.5 12.3 2.5
84.5 + 2.5 36.7 _.+1.9 I1.8 3
122.1 4` 5.0 33.4 4.0.5 7.2 4
L;,
C.A. Schall etal./ Journal of Crystal (;ro_lth 165 11996_ 29_¢ 29,_ 297
tion measured calorimetrically and the value of - 8.I
+ 2.7 kcal/mol regressed from a van "t Hoff plot
(Fig. 2).
t,,q
The crystals removed from the ampoules at the "" I4
conclusion of the crystallization experiments were
12
examined using an optical stereomicroscope and ex-
hibited a morphology typical of that observed for the r_
tetragonal R_rm of lysozyme crystals [II]. o
25 3 3.5 4 45 5 55
4. Discussion % NaC1
Fig. 3. Enthalpy of crystallization for hen egg-white Ivsoz,,me
First and foremost, the validity of the van "t Hoff versus salt concentration at pH = 4.5 in 50raM acetate buffl:r. The
analysis to protein solubility is apparent for this enthalpy of crystallization was calculated by' measuring the slope
of a ','an "t Hoff plot of solubility as a function of temperature
lysozyme system. Takizawa and Hayashi measured
_t= using solubility from Table I and Howard et at. [8]. The enthalpy
the enthalpy of crystallization of a lysozyme solution oi"crystallization is strongly affected b} salt ccmcentralicm under
4.6
in 3% NaCI, adjusted to a pH of 4.2, at 15°C and these conditions.
4..5
report a value of -25.1 4-5.0 kcal/mol [9]. Howard
et al. [8] found that their solubility data would pre-
JllOl
:1).5 dict an enthalpy of crystallization of - 18.9 kcal/mol
the based on interpolation of the solubility data. Al- as that depicted in Fig. 3 for pH = 4.5, one would
5.1 though this data is arguably within the error of the expect that Takizawa and Hayashi would report a
th a
measurement, a second data set displayed large dis- larger magnitude for AHcryst than the true value due
crepancies (13 _+3.0 kcal/mol by calorimetry versus to the miscalculated salt concentration. This would
5.3 kcal/mol by van 't Hoff analysis of solubility) account for the discrepancy between the results of
indicating that one of the two techniques were sus- Takizawa and Hayashi [9] and van "t Hoff analysis
2 is
pect. Obvious points of contention include the as- of Howard et al. [8].
red
sumptions of the van "t Hoff analysis (i.e. tempera- The enthalpy of crystallization of lysozyme mea-
Zig. ture independent enthalpy of crystallization) and the sured by calorimetry is not significantly different
_"at interpolation method. However, experimental proto- than the enthalpy of crystallization calculated from a
',O.5 cols of Takizawa and Hayashi [9] are also a source van 't Hoff plot of equilibrium solubilities for both
lifi- of error. The Takizawa and Hayashi value for the 3% and 5c_ NaCI solutions. This study justifies the
iza-
enthalpy of crystallization at 3% NaCI, pH 4.2 is use of such van 't Hoff analysis for the analysis of
signifcantly higher than our measured value in 5% protein solubilities in the future.
NaCI at pH 4.5 as well as the estimate of Howard et
al. [8]. Utilizing the data of Howard et al. [8] as well
as the data reported in Table l, the effect of salt
concentration on the enthalpy of crystallization at pH Acknowledgements
4.5 is shown in Fig. 3. The magnitude of the en-
thalpy of crystallization tends to increase dramati- The authors would like to thank Dr. Ajit Thakur
cally as salt concentration increases at a pH of 4.5. and BristoI-Meyers Squibb for generously allowing
Because the lysozyme in Takizawa and Hayashi's use of their Thermometric calorimeter for the collec-
experiment was not dialyzed, the salt content of the tion of the calorimetric data. C.A.S. was supported
crystallizing solution was most likely higher than the by NIH Biotechnology Training Grant fellowships
reported value of 3% due to the large amount of and by NASA GSRP fellowship NGT-51150. This
buffer salts typically found in lyophilized proteins. work was partially funded by NASA grant NAG8-
Assuming the same qualitative behavior at pH = 4.2 975.
298
References [6] L. Sibille and M.L. Pusey. Acta Cryst. 1) 50 !Iq94) 39h.
[7] P.A. Darcy and J.M. Wicncek, Charactcrizatmr. of Lysozymc
Crystallization by Micrtx:alorimctry pFe',clllt:tl at the 5()th
[I] A.G. Jones and J.W. Mullin, Chem. Eng. Sci 29. (1974) 105. Calorimetry Conference {NIST, GaithcJ_l_utg, MD). July
[2] T.L. Bray. L.J. Kiln, RP, Askew, M.D. Harrington. W.M. 19q5.
Rosenblum. W.W. Wilson and L.J. Del, ucas, J. Appl. Cryst.. [8] S.B. Howard, P.J, Twigg. J,K. Baird and li.J, M¢¢han, J.
submitted Crystal Growth q0 (,1988) 94,
[3] J.M. Wierlcek. E.V. Arnold. C.A. Schall. l.S. Bonita and [9] T. Takizawa and S. Hayashi, J, Ph)s. St_c. Jpn. 40 (1_)76)
P.A. Darcy. J. Crystal Growth. submitted. 299.
[41 E. Cacioppo. S. Munson and M.I,. Pasty. J. Crystal Growth [l()] A.J. Sophianopoulos. C.K. Rhodes, D.V,,', [[,,komh and K,F,
l lO(Iqqll_6. VanHolde. J. Biol. Chem. 237 (1t)62"1111)7
[5] M.L. Puse.x and S. M'unson, J. Crystal Growth 114 (1991) [II] M.L. Pusey, R.S. Snyder and R. Nanmimn. J Biol. Chem
385. 261 (1986) 652-.[.,
Reprinted from
TAL
JOURNAL OF
R TH
Journal of Crystal Grow!h 165 (1996) 299- 3(17
Application of temperature control strategies to the growth of hen
egg-white lysozyme crystals
Constance A. Schall "_• Jill S. Riley a Edwin Li _ Edward Arnold b
John M. Wiencek "*
" l)epartment of Chemical oral Biochemical Engineering, Rutgers Unirersitv. P.O. Bo._ 909. Pi_c,/¢mo_. New Jer._ev 08855-0909. USA
Center./or Adlanced Biotechnologv and Medicine. and Chemi._tt-v Department, R.tger_ (hm er_it_. 670 Hoe._ Lane. Piscaraway, New
Jersey 08854-563& USA
' Delmr/ment ofChemica[ ond Bio('hemi('_ff L_Leineering, Unher._ity t_flhm'a, [.77 CfJeln/strv [luihf/nk,, hm'a Ct't)', h_wo 52242- [219. USA
Received 20 April 1995: accepted 15Januar 3 1996
I!I.SEV1ER
......... CRYSTAL
GROWTH
ELSEVIER .lournal (ll Crystal Growth 165 (1996) 299-307
Application of temperature control strategies to the growth of hen
egg-white lysozyme crystals
Constance A. Schall '(_, Jill S. Riley ", Edwin Li _'.Edward Arnold b
John M. Wiencek c.,
" l)cl,artment o[Chemical and Bioclu,mical En.k,ineerint{. Rutgers Unirersity, P.O. Bo_ 909, Pi._catau'ay, New .lersev 08855-0909. If.%4
b ('enterlbr Advanced Bioteelmolocv and Medicine. amt Chemist O'Depat7ment, Rut l,,er_l,rnitersitv, 679 Hoes Lane. Piscatawav. Nen
Jersey 08854-563& USA
• l)epatTmen! of Chentical and Biochemical ICm,,ineerink'. (hficetwity o[Iowa, 137 ('hemistrv Buildito,,. Iowa City, Iowa 52242-1219, 1%'A
Received 20 April 1995: accepted 15January t996
Abstract
Solubility data were combined with mass balances and growth kinelics It) derive a temperature control algorithm which
maintains a constant level of supersalt,ralion. This conslanl supersaturation control (CSC) algorithm attempts to maximize
the size of protein crystals by maintaining the growth conditions in the metastable zone. Using hen egg-while lysozyme as a
model protein system, four temperature programming strategies were employed in seeded and unseeded syslems: the CSC
algorithm, a linear ramp derived from the CSC algorilhm, isothermal 20°C, and isothermal 4°C. Both the CSC-derived linear
and the CSC temperature programs yielded large, well-formed crystals which were significantly larger than crystals grown
isothermally at 20 and 4°C. The isothermal 4°C program resulted in poorly formed crystals due to the high initial growth
rates. The seeded systems displayed much higher levels of nucleation than the unseeded systems which is attributed lo
secondary nucleation. The results indicate that ntoderale deviations (~ 20e/c) from constant supersaturation can be tolerated.
while still producing large, well-formed crystals appropriate fl)r X-ray crystallography.
1. Introduction ture determination. In protein crystallization, super-
saturation is normally controlled through the use of
:heduled for
iishers upon Control of the level of supersaturation throughout precipitants or pH. Typically, protein in a hanging
•ent by SAL
are avaitabl¢ the crystallization process is essential if crystal size drop is equilibrated against a reservoir of precipitant
is to be optimized for the purpose of producing large, solution of higher concentration to produce protein
well-formed crystals [k)r X-ray crystallographic struc- crystals. This device oflers no control of the level of
supersaturation beyond setting the initial conditions.
r publication in contrast, Gernert et al. [I] devised a hanging drop
"Corresponding aulhor.
apparatus in which the reservoir concentration could
Ncthert_ld.$. ('urrenl address: Unixersilv of Tulsa. Chcmical Engineering
131. Department. 600 Soulh ('ollegc. Tulsa. Oklahoma 74104-3189. be changed over time and found that maintaining
_h¢ USA by USA. lower levels of supersaturation led to fewer protein
HERLANDS
(X)22-0248/96//$15.00 Copyright x, 1996 l!lsexicl Science BV. ,kill rights reserved
PII S0022-0248(96)00181-9